bpf, riscv: add BPF JIT for RV64G
This commit adds a BPF JIT for RV64G.
The JIT is a two-pass JIT, and has a dynamic prolog/epilogue (similar
to the MIPS64 BPF JIT) instead of static ones (e.g. x86_64).
At the moment the RISC-V Linux port does not support
CONFIG_HAVE_KPROBES, which means that CONFIG_BPF_EVENTS is not
supported. Thus, no tests involving BPF_PROG_TYPE_TRACEPOINT,
BPF_PROG_TYPE_PERF_EVENT, BPF_PROG_TYPE_KPROBE and
BPF_PROG_TYPE_RAW_TRACEPOINT passes.
The implementation does not support "far branching" (>4KiB).
Test results:
# modprobe test_bpf
test_bpf: Summary: 378 PASSED, 0 FAILED, [366/366 JIT'ed]
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier
...
Summary: 761 PASSED, 507 SKIPPED, 2 FAILED
Note that "test_verifier" was run with one build with
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y and one without, otherwise
many of the the tests that require unaligned access were skipped.
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
0
No CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
59
The two failing test_verifier tests are:
"ld_abs: vlan + abs, test 1"
"ld_abs: jump around ld_abs"
This is due to that "far branching" involved in those tests.
All tests where done on QEMU (QEMU emulator version 3.1.50
(v3.1.0-688-g8ae951fbc106)).
Signed-off-by: Björn Töpel <bjorn.topel@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2019-02-05 05:41:22 -07:00
|
|
|
// SPDX-License-Identifier: GPL-2.0
|
|
|
|
/* BPF JIT compiler for RV64G
|
|
|
|
*
|
|
|
|
* Copyright(c) 2019 Björn Töpel <bjorn.topel@gmail.com>
|
|
|
|
*
|
|
|
|
*/
|
|
|
|
|
2021-10-27 04:18:22 -07:00
|
|
|
#include <linux/bitfield.h>
|
bpf, riscv: add BPF JIT for RV64G
This commit adds a BPF JIT for RV64G.
The JIT is a two-pass JIT, and has a dynamic prolog/epilogue (similar
to the MIPS64 BPF JIT) instead of static ones (e.g. x86_64).
At the moment the RISC-V Linux port does not support
CONFIG_HAVE_KPROBES, which means that CONFIG_BPF_EVENTS is not
supported. Thus, no tests involving BPF_PROG_TYPE_TRACEPOINT,
BPF_PROG_TYPE_PERF_EVENT, BPF_PROG_TYPE_KPROBE and
BPF_PROG_TYPE_RAW_TRACEPOINT passes.
The implementation does not support "far branching" (>4KiB).
Test results:
# modprobe test_bpf
test_bpf: Summary: 378 PASSED, 0 FAILED, [366/366 JIT'ed]
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier
...
Summary: 761 PASSED, 507 SKIPPED, 2 FAILED
Note that "test_verifier" was run with one build with
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y and one without, otherwise
many of the the tests that require unaligned access were skipped.
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
0
No CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
59
The two failing test_verifier tests are:
"ld_abs: vlan + abs, test 1"
"ld_abs: jump around ld_abs"
This is due to that "far branching" involved in those tests.
All tests where done on QEMU (QEMU emulator version 3.1.50
(v3.1.0-688-g8ae951fbc106)).
Signed-off-by: Björn Töpel <bjorn.topel@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2019-02-05 05:41:22 -07:00
|
|
|
#include <linux/bpf.h>
|
|
|
|
#include <linux/filter.h>
|
2023-02-15 06:52:04 -07:00
|
|
|
#include <linux/memory.h>
|
|
|
|
#include <linux/stop_machine.h>
|
2023-02-27 00:20:16 -07:00
|
|
|
#include <asm/patch.h>
|
2024-03-03 10:02:07 -07:00
|
|
|
#include <asm/cfi.h>
|
2024-05-02 08:18:51 -07:00
|
|
|
#include <asm/percpu.h>
|
2020-03-04 22:02:04 -07:00
|
|
|
#include "bpf_jit.h"
|
bpf, riscv: add BPF JIT for RV64G
This commit adds a BPF JIT for RV64G.
The JIT is a two-pass JIT, and has a dynamic prolog/epilogue (similar
to the MIPS64 BPF JIT) instead of static ones (e.g. x86_64).
At the moment the RISC-V Linux port does not support
CONFIG_HAVE_KPROBES, which means that CONFIG_BPF_EVENTS is not
supported. Thus, no tests involving BPF_PROG_TYPE_TRACEPOINT,
BPF_PROG_TYPE_PERF_EVENT, BPF_PROG_TYPE_KPROBE and
BPF_PROG_TYPE_RAW_TRACEPOINT passes.
The implementation does not support "far branching" (>4KiB).
Test results:
# modprobe test_bpf
test_bpf: Summary: 378 PASSED, 0 FAILED, [366/366 JIT'ed]
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier
...
Summary: 761 PASSED, 507 SKIPPED, 2 FAILED
Note that "test_verifier" was run with one build with
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y and one without, otherwise
many of the the tests that require unaligned access were skipped.
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
0
No CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
59
The two failing test_verifier tests are:
"ld_abs: vlan + abs, test 1"
"ld_abs: jump around ld_abs"
This is due to that "far branching" involved in those tests.
All tests where done on QEMU (QEMU emulator version 3.1.50
(v3.1.0-688-g8ae951fbc106)).
Signed-off-by: Björn Töpel <bjorn.topel@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2019-02-05 05:41:22 -07:00
|
|
|
|
2023-07-21 03:06:27 -07:00
|
|
|
#define RV_FENTRY_NINSNS 2
|
|
|
|
|
bpf, riscv: add BPF JIT for RV64G
This commit adds a BPF JIT for RV64G.
The JIT is a two-pass JIT, and has a dynamic prolog/epilogue (similar
to the MIPS64 BPF JIT) instead of static ones (e.g. x86_64).
At the moment the RISC-V Linux port does not support
CONFIG_HAVE_KPROBES, which means that CONFIG_BPF_EVENTS is not
supported. Thus, no tests involving BPF_PROG_TYPE_TRACEPOINT,
BPF_PROG_TYPE_PERF_EVENT, BPF_PROG_TYPE_KPROBE and
BPF_PROG_TYPE_RAW_TRACEPOINT passes.
The implementation does not support "far branching" (>4KiB).
Test results:
# modprobe test_bpf
test_bpf: Summary: 378 PASSED, 0 FAILED, [366/366 JIT'ed]
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier
...
Summary: 761 PASSED, 507 SKIPPED, 2 FAILED
Note that "test_verifier" was run with one build with
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y and one without, otherwise
many of the the tests that require unaligned access were skipped.
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
0
No CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
59
The two failing test_verifier tests are:
"ld_abs: vlan + abs, test 1"
"ld_abs: jump around ld_abs"
This is due to that "far branching" involved in those tests.
All tests where done on QEMU (QEMU emulator version 3.1.50
(v3.1.0-688-g8ae951fbc106)).
Signed-off-by: Björn Töpel <bjorn.topel@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2019-02-05 05:41:22 -07:00
|
|
|
#define RV_REG_TCC RV_REG_A6
|
|
|
|
#define RV_REG_TCC_SAVED RV_REG_S6 /* Store A6 in S6 if program do calls */
|
2024-04-04 04:42:02 -07:00
|
|
|
#define RV_REG_ARENA RV_REG_S7 /* For storing arena_vm_start */
|
bpf, riscv: add BPF JIT for RV64G
This commit adds a BPF JIT for RV64G.
The JIT is a two-pass JIT, and has a dynamic prolog/epilogue (similar
to the MIPS64 BPF JIT) instead of static ones (e.g. x86_64).
At the moment the RISC-V Linux port does not support
CONFIG_HAVE_KPROBES, which means that CONFIG_BPF_EVENTS is not
supported. Thus, no tests involving BPF_PROG_TYPE_TRACEPOINT,
BPF_PROG_TYPE_PERF_EVENT, BPF_PROG_TYPE_KPROBE and
BPF_PROG_TYPE_RAW_TRACEPOINT passes.
The implementation does not support "far branching" (>4KiB).
Test results:
# modprobe test_bpf
test_bpf: Summary: 378 PASSED, 0 FAILED, [366/366 JIT'ed]
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier
...
Summary: 761 PASSED, 507 SKIPPED, 2 FAILED
Note that "test_verifier" was run with one build with
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y and one without, otherwise
many of the the tests that require unaligned access were skipped.
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
0
No CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
59
The two failing test_verifier tests are:
"ld_abs: vlan + abs, test 1"
"ld_abs: jump around ld_abs"
This is due to that "far branching" involved in those tests.
All tests where done on QEMU (QEMU emulator version 3.1.50
(v3.1.0-688-g8ae951fbc106)).
Signed-off-by: Björn Töpel <bjorn.topel@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2019-02-05 05:41:22 -07:00
|
|
|
|
|
|
|
static const int regmap[] = {
|
|
|
|
[BPF_REG_0] = RV_REG_A5,
|
|
|
|
[BPF_REG_1] = RV_REG_A0,
|
|
|
|
[BPF_REG_2] = RV_REG_A1,
|
|
|
|
[BPF_REG_3] = RV_REG_A2,
|
|
|
|
[BPF_REG_4] = RV_REG_A3,
|
|
|
|
[BPF_REG_5] = RV_REG_A4,
|
|
|
|
[BPF_REG_6] = RV_REG_S1,
|
|
|
|
[BPF_REG_7] = RV_REG_S2,
|
|
|
|
[BPF_REG_8] = RV_REG_S3,
|
|
|
|
[BPF_REG_9] = RV_REG_S4,
|
|
|
|
[BPF_REG_FP] = RV_REG_S5,
|
|
|
|
[BPF_REG_AX] = RV_REG_T0,
|
|
|
|
};
|
|
|
|
|
2021-10-27 04:18:22 -07:00
|
|
|
static const int pt_regmap[] = {
|
|
|
|
[RV_REG_A0] = offsetof(struct pt_regs, a0),
|
|
|
|
[RV_REG_A1] = offsetof(struct pt_regs, a1),
|
|
|
|
[RV_REG_A2] = offsetof(struct pt_regs, a2),
|
|
|
|
[RV_REG_A3] = offsetof(struct pt_regs, a3),
|
|
|
|
[RV_REG_A4] = offsetof(struct pt_regs, a4),
|
|
|
|
[RV_REG_A5] = offsetof(struct pt_regs, a5),
|
|
|
|
[RV_REG_S1] = offsetof(struct pt_regs, s1),
|
|
|
|
[RV_REG_S2] = offsetof(struct pt_regs, s2),
|
|
|
|
[RV_REG_S3] = offsetof(struct pt_regs, s3),
|
|
|
|
[RV_REG_S4] = offsetof(struct pt_regs, s4),
|
|
|
|
[RV_REG_S5] = offsetof(struct pt_regs, s5),
|
|
|
|
[RV_REG_T0] = offsetof(struct pt_regs, t0),
|
|
|
|
};
|
|
|
|
|
bpf, riscv: add BPF JIT for RV64G
This commit adds a BPF JIT for RV64G.
The JIT is a two-pass JIT, and has a dynamic prolog/epilogue (similar
to the MIPS64 BPF JIT) instead of static ones (e.g. x86_64).
At the moment the RISC-V Linux port does not support
CONFIG_HAVE_KPROBES, which means that CONFIG_BPF_EVENTS is not
supported. Thus, no tests involving BPF_PROG_TYPE_TRACEPOINT,
BPF_PROG_TYPE_PERF_EVENT, BPF_PROG_TYPE_KPROBE and
BPF_PROG_TYPE_RAW_TRACEPOINT passes.
The implementation does not support "far branching" (>4KiB).
Test results:
# modprobe test_bpf
test_bpf: Summary: 378 PASSED, 0 FAILED, [366/366 JIT'ed]
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier
...
Summary: 761 PASSED, 507 SKIPPED, 2 FAILED
Note that "test_verifier" was run with one build with
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y and one without, otherwise
many of the the tests that require unaligned access were skipped.
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
0
No CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
59
The two failing test_verifier tests are:
"ld_abs: vlan + abs, test 1"
"ld_abs: jump around ld_abs"
This is due to that "far branching" involved in those tests.
All tests where done on QEMU (QEMU emulator version 3.1.50
(v3.1.0-688-g8ae951fbc106)).
Signed-off-by: Björn Töpel <bjorn.topel@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2019-02-05 05:41:22 -07:00
|
|
|
enum {
|
|
|
|
RV_CTX_F_SEEN_TAIL_CALL = 0,
|
|
|
|
RV_CTX_F_SEEN_CALL = RV_REG_RA,
|
|
|
|
RV_CTX_F_SEEN_S1 = RV_REG_S1,
|
|
|
|
RV_CTX_F_SEEN_S2 = RV_REG_S2,
|
|
|
|
RV_CTX_F_SEEN_S3 = RV_REG_S3,
|
|
|
|
RV_CTX_F_SEEN_S4 = RV_REG_S4,
|
|
|
|
RV_CTX_F_SEEN_S5 = RV_REG_S5,
|
|
|
|
RV_CTX_F_SEEN_S6 = RV_REG_S6,
|
|
|
|
};
|
|
|
|
|
|
|
|
static u8 bpf_to_rv_reg(int bpf_reg, struct rv_jit_context *ctx)
|
|
|
|
{
|
|
|
|
u8 reg = regmap[bpf_reg];
|
|
|
|
|
|
|
|
switch (reg) {
|
|
|
|
case RV_CTX_F_SEEN_S1:
|
|
|
|
case RV_CTX_F_SEEN_S2:
|
|
|
|
case RV_CTX_F_SEEN_S3:
|
|
|
|
case RV_CTX_F_SEEN_S4:
|
|
|
|
case RV_CTX_F_SEEN_S5:
|
|
|
|
case RV_CTX_F_SEEN_S6:
|
|
|
|
__set_bit(reg, &ctx->flags);
|
|
|
|
}
|
|
|
|
return reg;
|
|
|
|
};
|
|
|
|
|
|
|
|
static bool seen_reg(int reg, struct rv_jit_context *ctx)
|
|
|
|
{
|
|
|
|
switch (reg) {
|
|
|
|
case RV_CTX_F_SEEN_CALL:
|
|
|
|
case RV_CTX_F_SEEN_S1:
|
|
|
|
case RV_CTX_F_SEEN_S2:
|
|
|
|
case RV_CTX_F_SEEN_S3:
|
|
|
|
case RV_CTX_F_SEEN_S4:
|
|
|
|
case RV_CTX_F_SEEN_S5:
|
|
|
|
case RV_CTX_F_SEEN_S6:
|
|
|
|
return test_bit(reg, &ctx->flags);
|
|
|
|
}
|
|
|
|
return false;
|
|
|
|
}
|
|
|
|
|
2019-12-16 02:13:35 -07:00
|
|
|
static void mark_fp(struct rv_jit_context *ctx)
|
|
|
|
{
|
|
|
|
__set_bit(RV_CTX_F_SEEN_S5, &ctx->flags);
|
|
|
|
}
|
|
|
|
|
bpf, riscv: add BPF JIT for RV64G
This commit adds a BPF JIT for RV64G.
The JIT is a two-pass JIT, and has a dynamic prolog/epilogue (similar
to the MIPS64 BPF JIT) instead of static ones (e.g. x86_64).
At the moment the RISC-V Linux port does not support
CONFIG_HAVE_KPROBES, which means that CONFIG_BPF_EVENTS is not
supported. Thus, no tests involving BPF_PROG_TYPE_TRACEPOINT,
BPF_PROG_TYPE_PERF_EVENT, BPF_PROG_TYPE_KPROBE and
BPF_PROG_TYPE_RAW_TRACEPOINT passes.
The implementation does not support "far branching" (>4KiB).
Test results:
# modprobe test_bpf
test_bpf: Summary: 378 PASSED, 0 FAILED, [366/366 JIT'ed]
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier
...
Summary: 761 PASSED, 507 SKIPPED, 2 FAILED
Note that "test_verifier" was run with one build with
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y and one without, otherwise
many of the the tests that require unaligned access were skipped.
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
0
No CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
59
The two failing test_verifier tests are:
"ld_abs: vlan + abs, test 1"
"ld_abs: jump around ld_abs"
This is due to that "far branching" involved in those tests.
All tests where done on QEMU (QEMU emulator version 3.1.50
(v3.1.0-688-g8ae951fbc106)).
Signed-off-by: Björn Töpel <bjorn.topel@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2019-02-05 05:41:22 -07:00
|
|
|
static void mark_call(struct rv_jit_context *ctx)
|
|
|
|
{
|
|
|
|
__set_bit(RV_CTX_F_SEEN_CALL, &ctx->flags);
|
|
|
|
}
|
|
|
|
|
|
|
|
static bool seen_call(struct rv_jit_context *ctx)
|
|
|
|
{
|
|
|
|
return test_bit(RV_CTX_F_SEEN_CALL, &ctx->flags);
|
|
|
|
}
|
|
|
|
|
|
|
|
static void mark_tail_call(struct rv_jit_context *ctx)
|
|
|
|
{
|
|
|
|
__set_bit(RV_CTX_F_SEEN_TAIL_CALL, &ctx->flags);
|
|
|
|
}
|
|
|
|
|
|
|
|
static bool seen_tail_call(struct rv_jit_context *ctx)
|
|
|
|
{
|
|
|
|
return test_bit(RV_CTX_F_SEEN_TAIL_CALL, &ctx->flags);
|
|
|
|
}
|
|
|
|
|
|
|
|
static u8 rv_tail_call_reg(struct rv_jit_context *ctx)
|
|
|
|
{
|
|
|
|
mark_tail_call(ctx);
|
|
|
|
|
|
|
|
if (seen_call(ctx)) {
|
|
|
|
__set_bit(RV_CTX_F_SEEN_S6, &ctx->flags);
|
|
|
|
return RV_REG_S6;
|
|
|
|
}
|
|
|
|
return RV_REG_A6;
|
|
|
|
}
|
|
|
|
|
|
|
|
static bool is_32b_int(s64 val)
|
|
|
|
{
|
|
|
|
return -(1L << 31) <= val && val < (1L << 31);
|
|
|
|
}
|
|
|
|
|
2020-04-06 15:16:04 -07:00
|
|
|
static bool in_auipc_jalr_range(s64 val)
|
|
|
|
{
|
|
|
|
/*
|
|
|
|
* auipc+jalr can reach any signed PC-relative offset in the range
|
|
|
|
* [-2^31 - 2^11, 2^31 - 2^11).
|
|
|
|
*/
|
|
|
|
return (-(1L << 31) - (1L << 11)) <= val &&
|
|
|
|
val < ((1L << 31) - (1L << 11));
|
|
|
|
}
|
|
|
|
|
2024-01-15 06:12:32 -07:00
|
|
|
/* Modify rd pointer to alternate reg to avoid corrupting original reg */
|
|
|
|
static void emit_sextw_alt(u8 *rd, u8 ra, struct rv_jit_context *ctx)
|
|
|
|
{
|
|
|
|
emit_sextw(ra, *rd, ctx);
|
|
|
|
*rd = ra;
|
|
|
|
}
|
|
|
|
|
|
|
|
static void emit_zextw_alt(u8 *rd, u8 ra, struct rv_jit_context *ctx)
|
|
|
|
{
|
|
|
|
emit_zextw(ra, *rd, ctx);
|
|
|
|
*rd = ra;
|
|
|
|
}
|
|
|
|
|
2022-12-06 02:14:10 -07:00
|
|
|
/* Emit fixed-length instructions for address */
|
|
|
|
static int emit_addr(u8 rd, u64 addr, bool extra_pass, struct rv_jit_context *ctx)
|
|
|
|
{
|
2023-08-31 06:12:29 -07:00
|
|
|
/*
|
|
|
|
* Use the ro_insns(RX) to calculate the offset as the BPF program will
|
|
|
|
* finally run from this memory region.
|
|
|
|
*/
|
|
|
|
u64 ip = (u64)(ctx->ro_insns + ctx->ninsns);
|
2022-12-06 02:14:10 -07:00
|
|
|
s64 off = addr - ip;
|
|
|
|
s64 upper = (off + (1 << 11)) >> 12;
|
|
|
|
s64 lower = off & 0xfff;
|
|
|
|
|
|
|
|
if (extra_pass && !in_auipc_jalr_range(off)) {
|
|
|
|
pr_err("bpf-jit: target offset 0x%llx is out of range\n", off);
|
|
|
|
return -ERANGE;
|
|
|
|
}
|
|
|
|
|
|
|
|
emit(rv_auipc(rd, upper), ctx);
|
|
|
|
emit(rv_addi(rd, rd, lower), ctx);
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
/* Emit variable-length instructions for 32-bit and 64-bit imm */
|
bpf, riscv: add BPF JIT for RV64G
This commit adds a BPF JIT for RV64G.
The JIT is a two-pass JIT, and has a dynamic prolog/epilogue (similar
to the MIPS64 BPF JIT) instead of static ones (e.g. x86_64).
At the moment the RISC-V Linux port does not support
CONFIG_HAVE_KPROBES, which means that CONFIG_BPF_EVENTS is not
supported. Thus, no tests involving BPF_PROG_TYPE_TRACEPOINT,
BPF_PROG_TYPE_PERF_EVENT, BPF_PROG_TYPE_KPROBE and
BPF_PROG_TYPE_RAW_TRACEPOINT passes.
The implementation does not support "far branching" (>4KiB).
Test results:
# modprobe test_bpf
test_bpf: Summary: 378 PASSED, 0 FAILED, [366/366 JIT'ed]
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier
...
Summary: 761 PASSED, 507 SKIPPED, 2 FAILED
Note that "test_verifier" was run with one build with
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y and one without, otherwise
many of the the tests that require unaligned access were skipped.
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
0
No CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
59
The two failing test_verifier tests are:
"ld_abs: vlan + abs, test 1"
"ld_abs: jump around ld_abs"
This is due to that "far branching" involved in those tests.
All tests where done on QEMU (QEMU emulator version 3.1.50
(v3.1.0-688-g8ae951fbc106)).
Signed-off-by: Björn Töpel <bjorn.topel@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2019-02-05 05:41:22 -07:00
|
|
|
static void emit_imm(u8 rd, s64 val, struct rv_jit_context *ctx)
|
|
|
|
{
|
|
|
|
/* Note that the immediate from the add is sign-extended,
|
|
|
|
* which means that we need to compensate this by adding 2^12,
|
|
|
|
* when the 12th bit is set. A simpler way of doing this, and
|
|
|
|
* getting rid of the check, is to just add 2**11 before the
|
|
|
|
* shift. The "Loading a 32-Bit constant" example from the
|
|
|
|
* "Computer Organization and Design, RISC-V edition" book by
|
|
|
|
* Patterson/Hennessy highlights this fact.
|
|
|
|
*
|
|
|
|
* This also means that we need to process LSB to MSB.
|
|
|
|
*/
|
bpf, riscv: Use compressed instructions in the rv64 JIT
This patch uses the RVC support and encodings from bpf_jit.h to optimize
the rv64 jit.
The optimizations work by replacing emit(rv_X(...)) with a call to a
helper function emit_X, which will emit a compressed version of the
instruction when possible, and when RVC is enabled.
The JIT continues to pass all tests in lib/test_bpf.c, and introduces
no new failures to test_verifier; both with and without RVC being enabled.
Most changes are straightforward replacements of emit(rv_X(...), ctx)
with emit_X(..., ctx), with the following exceptions bearing mention;
* Change emit_imm to sign-extend the value in "lower", since the
checks for RVC (and the instructions themselves) treat the value as
signed. Otherwise, small negative immediates will not be recognized as
encodable using an RVC instruction. For example, without this change,
emit_imm(rd, -1, ctx) would cause lower to become 4095, which is not a
6b int even though a "c.li rd, -1" instruction suffices.
* For {BPF_MOV,BPF_ADD} BPF_X, drop using addiw,addw in the 32-bit
cases since the values are zero-extended into the upper 32 bits in
the following instructions anyways, and the addition commutes with
zero-extension. (BPF_SUB BPF_X must still use subw since subtraction
does not commute with zero-extension.)
This patch avoids optimizing branches and jumps to use RVC instructions
since surrounding code often makes assumptions about the sizes of
emitted instructions. Optimizing these will require changing these
functions (e.g., emit_branch) to dynamically compute jump offsets.
The following are examples of the JITed code for the verifier selftest
"direct packet read test#3 for CGROUP_SKB OK", without and with RVC
enabled, respectively. The former uses 178 bytes, and the latter uses 112,
for a ~37% reduction in code size for this example.
Without RVC:
0: 02000813 addi a6,zero,32
4: fd010113 addi sp,sp,-48
8: 02813423 sd s0,40(sp)
c: 02913023 sd s1,32(sp)
10: 01213c23 sd s2,24(sp)
14: 01313823 sd s3,16(sp)
18: 01413423 sd s4,8(sp)
1c: 03010413 addi s0,sp,48
20: 03056683 lwu a3,48(a0)
24: 02069693 slli a3,a3,0x20
28: 0206d693 srli a3,a3,0x20
2c: 03456703 lwu a4,52(a0)
30: 02071713 slli a4,a4,0x20
34: 02075713 srli a4,a4,0x20
38: 03856483 lwu s1,56(a0)
3c: 02049493 slli s1,s1,0x20
40: 0204d493 srli s1,s1,0x20
44: 03c56903 lwu s2,60(a0)
48: 02091913 slli s2,s2,0x20
4c: 02095913 srli s2,s2,0x20
50: 04056983 lwu s3,64(a0)
54: 02099993 slli s3,s3,0x20
58: 0209d993 srli s3,s3,0x20
5c: 09056a03 lwu s4,144(a0)
60: 020a1a13 slli s4,s4,0x20
64: 020a5a13 srli s4,s4,0x20
68: 00900313 addi t1,zero,9
6c: 006a7463 bgeu s4,t1,0x74
70: 00000a13 addi s4,zero,0
74: 02d52823 sw a3,48(a0)
78: 02e52a23 sw a4,52(a0)
7c: 02952c23 sw s1,56(a0)
80: 03252e23 sw s2,60(a0)
84: 05352023 sw s3,64(a0)
88: 00000793 addi a5,zero,0
8c: 02813403 ld s0,40(sp)
90: 02013483 ld s1,32(sp)
94: 01813903 ld s2,24(sp)
98: 01013983 ld s3,16(sp)
9c: 00813a03 ld s4,8(sp)
a0: 03010113 addi sp,sp,48
a4: 00078513 addi a0,a5,0
a8: 00008067 jalr zero,0(ra)
With RVC:
0: 02000813 addi a6,zero,32
4: 7179 c.addi16sp sp,-48
6: f422 c.sdsp s0,40(sp)
8: f026 c.sdsp s1,32(sp)
a: ec4a c.sdsp s2,24(sp)
c: e84e c.sdsp s3,16(sp)
e: e452 c.sdsp s4,8(sp)
10: 1800 c.addi4spn s0,sp,48
12: 03056683 lwu a3,48(a0)
16: 1682 c.slli a3,0x20
18: 9281 c.srli a3,0x20
1a: 03456703 lwu a4,52(a0)
1e: 1702 c.slli a4,0x20
20: 9301 c.srli a4,0x20
22: 03856483 lwu s1,56(a0)
26: 1482 c.slli s1,0x20
28: 9081 c.srli s1,0x20
2a: 03c56903 lwu s2,60(a0)
2e: 1902 c.slli s2,0x20
30: 02095913 srli s2,s2,0x20
34: 04056983 lwu s3,64(a0)
38: 1982 c.slli s3,0x20
3a: 0209d993 srli s3,s3,0x20
3e: 09056a03 lwu s4,144(a0)
42: 1a02 c.slli s4,0x20
44: 020a5a13 srli s4,s4,0x20
48: 4325 c.li t1,9
4a: 006a7363 bgeu s4,t1,0x50
4e: 4a01 c.li s4,0
50: d914 c.sw a3,48(a0)
52: d958 c.sw a4,52(a0)
54: dd04 c.sw s1,56(a0)
56: 03252e23 sw s2,60(a0)
5a: 05352023 sw s3,64(a0)
5e: 4781 c.li a5,0
60: 7422 c.ldsp s0,40(sp)
62: 7482 c.ldsp s1,32(sp)
64: 6962 c.ldsp s2,24(sp)
66: 69c2 c.ldsp s3,16(sp)
68: 6a22 c.ldsp s4,8(sp)
6a: 6145 c.addi16sp sp,48
6c: 853e c.mv a0,a5
6e: 8082 c.jr ra
Signed-off-by: Luke Nelson <luke.r.nels@gmail.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Cc: Björn Töpel <bjorn.topel@gmail.com>
Link: https://lore.kernel.org/bpf/20200721025241.8077-4-luke.r.nels@gmail.com
2020-07-20 19:52:40 -07:00
|
|
|
s64 upper = (val + (1 << 11)) >> 12;
|
|
|
|
/* Sign-extend lower 12 bits to 64 bits since immediates for li, addiw,
|
|
|
|
* and addi are signed and RVC checks will perform signed comparisons.
|
|
|
|
*/
|
|
|
|
s64 lower = ((val & 0xfff) << 52) >> 52;
|
bpf, riscv: add BPF JIT for RV64G
This commit adds a BPF JIT for RV64G.
The JIT is a two-pass JIT, and has a dynamic prolog/epilogue (similar
to the MIPS64 BPF JIT) instead of static ones (e.g. x86_64).
At the moment the RISC-V Linux port does not support
CONFIG_HAVE_KPROBES, which means that CONFIG_BPF_EVENTS is not
supported. Thus, no tests involving BPF_PROG_TYPE_TRACEPOINT,
BPF_PROG_TYPE_PERF_EVENT, BPF_PROG_TYPE_KPROBE and
BPF_PROG_TYPE_RAW_TRACEPOINT passes.
The implementation does not support "far branching" (>4KiB).
Test results:
# modprobe test_bpf
test_bpf: Summary: 378 PASSED, 0 FAILED, [366/366 JIT'ed]
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier
...
Summary: 761 PASSED, 507 SKIPPED, 2 FAILED
Note that "test_verifier" was run with one build with
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y and one without, otherwise
many of the the tests that require unaligned access were skipped.
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
0
No CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
59
The two failing test_verifier tests are:
"ld_abs: vlan + abs, test 1"
"ld_abs: jump around ld_abs"
This is due to that "far branching" involved in those tests.
All tests where done on QEMU (QEMU emulator version 3.1.50
(v3.1.0-688-g8ae951fbc106)).
Signed-off-by: Björn Töpel <bjorn.topel@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2019-02-05 05:41:22 -07:00
|
|
|
int shift;
|
|
|
|
|
|
|
|
if (is_32b_int(val)) {
|
|
|
|
if (upper)
|
bpf, riscv: Use compressed instructions in the rv64 JIT
This patch uses the RVC support and encodings from bpf_jit.h to optimize
the rv64 jit.
The optimizations work by replacing emit(rv_X(...)) with a call to a
helper function emit_X, which will emit a compressed version of the
instruction when possible, and when RVC is enabled.
The JIT continues to pass all tests in lib/test_bpf.c, and introduces
no new failures to test_verifier; both with and without RVC being enabled.
Most changes are straightforward replacements of emit(rv_X(...), ctx)
with emit_X(..., ctx), with the following exceptions bearing mention;
* Change emit_imm to sign-extend the value in "lower", since the
checks for RVC (and the instructions themselves) treat the value as
signed. Otherwise, small negative immediates will not be recognized as
encodable using an RVC instruction. For example, without this change,
emit_imm(rd, -1, ctx) would cause lower to become 4095, which is not a
6b int even though a "c.li rd, -1" instruction suffices.
* For {BPF_MOV,BPF_ADD} BPF_X, drop using addiw,addw in the 32-bit
cases since the values are zero-extended into the upper 32 bits in
the following instructions anyways, and the addition commutes with
zero-extension. (BPF_SUB BPF_X must still use subw since subtraction
does not commute with zero-extension.)
This patch avoids optimizing branches and jumps to use RVC instructions
since surrounding code often makes assumptions about the sizes of
emitted instructions. Optimizing these will require changing these
functions (e.g., emit_branch) to dynamically compute jump offsets.
The following are examples of the JITed code for the verifier selftest
"direct packet read test#3 for CGROUP_SKB OK", without and with RVC
enabled, respectively. The former uses 178 bytes, and the latter uses 112,
for a ~37% reduction in code size for this example.
Without RVC:
0: 02000813 addi a6,zero,32
4: fd010113 addi sp,sp,-48
8: 02813423 sd s0,40(sp)
c: 02913023 sd s1,32(sp)
10: 01213c23 sd s2,24(sp)
14: 01313823 sd s3,16(sp)
18: 01413423 sd s4,8(sp)
1c: 03010413 addi s0,sp,48
20: 03056683 lwu a3,48(a0)
24: 02069693 slli a3,a3,0x20
28: 0206d693 srli a3,a3,0x20
2c: 03456703 lwu a4,52(a0)
30: 02071713 slli a4,a4,0x20
34: 02075713 srli a4,a4,0x20
38: 03856483 lwu s1,56(a0)
3c: 02049493 slli s1,s1,0x20
40: 0204d493 srli s1,s1,0x20
44: 03c56903 lwu s2,60(a0)
48: 02091913 slli s2,s2,0x20
4c: 02095913 srli s2,s2,0x20
50: 04056983 lwu s3,64(a0)
54: 02099993 slli s3,s3,0x20
58: 0209d993 srli s3,s3,0x20
5c: 09056a03 lwu s4,144(a0)
60: 020a1a13 slli s4,s4,0x20
64: 020a5a13 srli s4,s4,0x20
68: 00900313 addi t1,zero,9
6c: 006a7463 bgeu s4,t1,0x74
70: 00000a13 addi s4,zero,0
74: 02d52823 sw a3,48(a0)
78: 02e52a23 sw a4,52(a0)
7c: 02952c23 sw s1,56(a0)
80: 03252e23 sw s2,60(a0)
84: 05352023 sw s3,64(a0)
88: 00000793 addi a5,zero,0
8c: 02813403 ld s0,40(sp)
90: 02013483 ld s1,32(sp)
94: 01813903 ld s2,24(sp)
98: 01013983 ld s3,16(sp)
9c: 00813a03 ld s4,8(sp)
a0: 03010113 addi sp,sp,48
a4: 00078513 addi a0,a5,0
a8: 00008067 jalr zero,0(ra)
With RVC:
0: 02000813 addi a6,zero,32
4: 7179 c.addi16sp sp,-48
6: f422 c.sdsp s0,40(sp)
8: f026 c.sdsp s1,32(sp)
a: ec4a c.sdsp s2,24(sp)
c: e84e c.sdsp s3,16(sp)
e: e452 c.sdsp s4,8(sp)
10: 1800 c.addi4spn s0,sp,48
12: 03056683 lwu a3,48(a0)
16: 1682 c.slli a3,0x20
18: 9281 c.srli a3,0x20
1a: 03456703 lwu a4,52(a0)
1e: 1702 c.slli a4,0x20
20: 9301 c.srli a4,0x20
22: 03856483 lwu s1,56(a0)
26: 1482 c.slli s1,0x20
28: 9081 c.srli s1,0x20
2a: 03c56903 lwu s2,60(a0)
2e: 1902 c.slli s2,0x20
30: 02095913 srli s2,s2,0x20
34: 04056983 lwu s3,64(a0)
38: 1982 c.slli s3,0x20
3a: 0209d993 srli s3,s3,0x20
3e: 09056a03 lwu s4,144(a0)
42: 1a02 c.slli s4,0x20
44: 020a5a13 srli s4,s4,0x20
48: 4325 c.li t1,9
4a: 006a7363 bgeu s4,t1,0x50
4e: 4a01 c.li s4,0
50: d914 c.sw a3,48(a0)
52: d958 c.sw a4,52(a0)
54: dd04 c.sw s1,56(a0)
56: 03252e23 sw s2,60(a0)
5a: 05352023 sw s3,64(a0)
5e: 4781 c.li a5,0
60: 7422 c.ldsp s0,40(sp)
62: 7482 c.ldsp s1,32(sp)
64: 6962 c.ldsp s2,24(sp)
66: 69c2 c.ldsp s3,16(sp)
68: 6a22 c.ldsp s4,8(sp)
6a: 6145 c.addi16sp sp,48
6c: 853e c.mv a0,a5
6e: 8082 c.jr ra
Signed-off-by: Luke Nelson <luke.r.nels@gmail.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Cc: Björn Töpel <bjorn.topel@gmail.com>
Link: https://lore.kernel.org/bpf/20200721025241.8077-4-luke.r.nels@gmail.com
2020-07-20 19:52:40 -07:00
|
|
|
emit_lui(rd, upper, ctx);
|
bpf, riscv: add BPF JIT for RV64G
This commit adds a BPF JIT for RV64G.
The JIT is a two-pass JIT, and has a dynamic prolog/epilogue (similar
to the MIPS64 BPF JIT) instead of static ones (e.g. x86_64).
At the moment the RISC-V Linux port does not support
CONFIG_HAVE_KPROBES, which means that CONFIG_BPF_EVENTS is not
supported. Thus, no tests involving BPF_PROG_TYPE_TRACEPOINT,
BPF_PROG_TYPE_PERF_EVENT, BPF_PROG_TYPE_KPROBE and
BPF_PROG_TYPE_RAW_TRACEPOINT passes.
The implementation does not support "far branching" (>4KiB).
Test results:
# modprobe test_bpf
test_bpf: Summary: 378 PASSED, 0 FAILED, [366/366 JIT'ed]
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier
...
Summary: 761 PASSED, 507 SKIPPED, 2 FAILED
Note that "test_verifier" was run with one build with
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y and one without, otherwise
many of the the tests that require unaligned access were skipped.
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
0
No CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
59
The two failing test_verifier tests are:
"ld_abs: vlan + abs, test 1"
"ld_abs: jump around ld_abs"
This is due to that "far branching" involved in those tests.
All tests where done on QEMU (QEMU emulator version 3.1.50
(v3.1.0-688-g8ae951fbc106)).
Signed-off-by: Björn Töpel <bjorn.topel@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2019-02-05 05:41:22 -07:00
|
|
|
|
|
|
|
if (!upper) {
|
bpf, riscv: Use compressed instructions in the rv64 JIT
This patch uses the RVC support and encodings from bpf_jit.h to optimize
the rv64 jit.
The optimizations work by replacing emit(rv_X(...)) with a call to a
helper function emit_X, which will emit a compressed version of the
instruction when possible, and when RVC is enabled.
The JIT continues to pass all tests in lib/test_bpf.c, and introduces
no new failures to test_verifier; both with and without RVC being enabled.
Most changes are straightforward replacements of emit(rv_X(...), ctx)
with emit_X(..., ctx), with the following exceptions bearing mention;
* Change emit_imm to sign-extend the value in "lower", since the
checks for RVC (and the instructions themselves) treat the value as
signed. Otherwise, small negative immediates will not be recognized as
encodable using an RVC instruction. For example, without this change,
emit_imm(rd, -1, ctx) would cause lower to become 4095, which is not a
6b int even though a "c.li rd, -1" instruction suffices.
* For {BPF_MOV,BPF_ADD} BPF_X, drop using addiw,addw in the 32-bit
cases since the values are zero-extended into the upper 32 bits in
the following instructions anyways, and the addition commutes with
zero-extension. (BPF_SUB BPF_X must still use subw since subtraction
does not commute with zero-extension.)
This patch avoids optimizing branches and jumps to use RVC instructions
since surrounding code often makes assumptions about the sizes of
emitted instructions. Optimizing these will require changing these
functions (e.g., emit_branch) to dynamically compute jump offsets.
The following are examples of the JITed code for the verifier selftest
"direct packet read test#3 for CGROUP_SKB OK", without and with RVC
enabled, respectively. The former uses 178 bytes, and the latter uses 112,
for a ~37% reduction in code size for this example.
Without RVC:
0: 02000813 addi a6,zero,32
4: fd010113 addi sp,sp,-48
8: 02813423 sd s0,40(sp)
c: 02913023 sd s1,32(sp)
10: 01213c23 sd s2,24(sp)
14: 01313823 sd s3,16(sp)
18: 01413423 sd s4,8(sp)
1c: 03010413 addi s0,sp,48
20: 03056683 lwu a3,48(a0)
24: 02069693 slli a3,a3,0x20
28: 0206d693 srli a3,a3,0x20
2c: 03456703 lwu a4,52(a0)
30: 02071713 slli a4,a4,0x20
34: 02075713 srli a4,a4,0x20
38: 03856483 lwu s1,56(a0)
3c: 02049493 slli s1,s1,0x20
40: 0204d493 srli s1,s1,0x20
44: 03c56903 lwu s2,60(a0)
48: 02091913 slli s2,s2,0x20
4c: 02095913 srli s2,s2,0x20
50: 04056983 lwu s3,64(a0)
54: 02099993 slli s3,s3,0x20
58: 0209d993 srli s3,s3,0x20
5c: 09056a03 lwu s4,144(a0)
60: 020a1a13 slli s4,s4,0x20
64: 020a5a13 srli s4,s4,0x20
68: 00900313 addi t1,zero,9
6c: 006a7463 bgeu s4,t1,0x74
70: 00000a13 addi s4,zero,0
74: 02d52823 sw a3,48(a0)
78: 02e52a23 sw a4,52(a0)
7c: 02952c23 sw s1,56(a0)
80: 03252e23 sw s2,60(a0)
84: 05352023 sw s3,64(a0)
88: 00000793 addi a5,zero,0
8c: 02813403 ld s0,40(sp)
90: 02013483 ld s1,32(sp)
94: 01813903 ld s2,24(sp)
98: 01013983 ld s3,16(sp)
9c: 00813a03 ld s4,8(sp)
a0: 03010113 addi sp,sp,48
a4: 00078513 addi a0,a5,0
a8: 00008067 jalr zero,0(ra)
With RVC:
0: 02000813 addi a6,zero,32
4: 7179 c.addi16sp sp,-48
6: f422 c.sdsp s0,40(sp)
8: f026 c.sdsp s1,32(sp)
a: ec4a c.sdsp s2,24(sp)
c: e84e c.sdsp s3,16(sp)
e: e452 c.sdsp s4,8(sp)
10: 1800 c.addi4spn s0,sp,48
12: 03056683 lwu a3,48(a0)
16: 1682 c.slli a3,0x20
18: 9281 c.srli a3,0x20
1a: 03456703 lwu a4,52(a0)
1e: 1702 c.slli a4,0x20
20: 9301 c.srli a4,0x20
22: 03856483 lwu s1,56(a0)
26: 1482 c.slli s1,0x20
28: 9081 c.srli s1,0x20
2a: 03c56903 lwu s2,60(a0)
2e: 1902 c.slli s2,0x20
30: 02095913 srli s2,s2,0x20
34: 04056983 lwu s3,64(a0)
38: 1982 c.slli s3,0x20
3a: 0209d993 srli s3,s3,0x20
3e: 09056a03 lwu s4,144(a0)
42: 1a02 c.slli s4,0x20
44: 020a5a13 srli s4,s4,0x20
48: 4325 c.li t1,9
4a: 006a7363 bgeu s4,t1,0x50
4e: 4a01 c.li s4,0
50: d914 c.sw a3,48(a0)
52: d958 c.sw a4,52(a0)
54: dd04 c.sw s1,56(a0)
56: 03252e23 sw s2,60(a0)
5a: 05352023 sw s3,64(a0)
5e: 4781 c.li a5,0
60: 7422 c.ldsp s0,40(sp)
62: 7482 c.ldsp s1,32(sp)
64: 6962 c.ldsp s2,24(sp)
66: 69c2 c.ldsp s3,16(sp)
68: 6a22 c.ldsp s4,8(sp)
6a: 6145 c.addi16sp sp,48
6c: 853e c.mv a0,a5
6e: 8082 c.jr ra
Signed-off-by: Luke Nelson <luke.r.nels@gmail.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Cc: Björn Töpel <bjorn.topel@gmail.com>
Link: https://lore.kernel.org/bpf/20200721025241.8077-4-luke.r.nels@gmail.com
2020-07-20 19:52:40 -07:00
|
|
|
emit_li(rd, lower, ctx);
|
bpf, riscv: add BPF JIT for RV64G
This commit adds a BPF JIT for RV64G.
The JIT is a two-pass JIT, and has a dynamic prolog/epilogue (similar
to the MIPS64 BPF JIT) instead of static ones (e.g. x86_64).
At the moment the RISC-V Linux port does not support
CONFIG_HAVE_KPROBES, which means that CONFIG_BPF_EVENTS is not
supported. Thus, no tests involving BPF_PROG_TYPE_TRACEPOINT,
BPF_PROG_TYPE_PERF_EVENT, BPF_PROG_TYPE_KPROBE and
BPF_PROG_TYPE_RAW_TRACEPOINT passes.
The implementation does not support "far branching" (>4KiB).
Test results:
# modprobe test_bpf
test_bpf: Summary: 378 PASSED, 0 FAILED, [366/366 JIT'ed]
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier
...
Summary: 761 PASSED, 507 SKIPPED, 2 FAILED
Note that "test_verifier" was run with one build with
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y and one without, otherwise
many of the the tests that require unaligned access were skipped.
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
0
No CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
59
The two failing test_verifier tests are:
"ld_abs: vlan + abs, test 1"
"ld_abs: jump around ld_abs"
This is due to that "far branching" involved in those tests.
All tests where done on QEMU (QEMU emulator version 3.1.50
(v3.1.0-688-g8ae951fbc106)).
Signed-off-by: Björn Töpel <bjorn.topel@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2019-02-05 05:41:22 -07:00
|
|
|
return;
|
|
|
|
}
|
|
|
|
|
bpf, riscv: Use compressed instructions in the rv64 JIT
This patch uses the RVC support and encodings from bpf_jit.h to optimize
the rv64 jit.
The optimizations work by replacing emit(rv_X(...)) with a call to a
helper function emit_X, which will emit a compressed version of the
instruction when possible, and when RVC is enabled.
The JIT continues to pass all tests in lib/test_bpf.c, and introduces
no new failures to test_verifier; both with and without RVC being enabled.
Most changes are straightforward replacements of emit(rv_X(...), ctx)
with emit_X(..., ctx), with the following exceptions bearing mention;
* Change emit_imm to sign-extend the value in "lower", since the
checks for RVC (and the instructions themselves) treat the value as
signed. Otherwise, small negative immediates will not be recognized as
encodable using an RVC instruction. For example, without this change,
emit_imm(rd, -1, ctx) would cause lower to become 4095, which is not a
6b int even though a "c.li rd, -1" instruction suffices.
* For {BPF_MOV,BPF_ADD} BPF_X, drop using addiw,addw in the 32-bit
cases since the values are zero-extended into the upper 32 bits in
the following instructions anyways, and the addition commutes with
zero-extension. (BPF_SUB BPF_X must still use subw since subtraction
does not commute with zero-extension.)
This patch avoids optimizing branches and jumps to use RVC instructions
since surrounding code often makes assumptions about the sizes of
emitted instructions. Optimizing these will require changing these
functions (e.g., emit_branch) to dynamically compute jump offsets.
The following are examples of the JITed code for the verifier selftest
"direct packet read test#3 for CGROUP_SKB OK", without and with RVC
enabled, respectively. The former uses 178 bytes, and the latter uses 112,
for a ~37% reduction in code size for this example.
Without RVC:
0: 02000813 addi a6,zero,32
4: fd010113 addi sp,sp,-48
8: 02813423 sd s0,40(sp)
c: 02913023 sd s1,32(sp)
10: 01213c23 sd s2,24(sp)
14: 01313823 sd s3,16(sp)
18: 01413423 sd s4,8(sp)
1c: 03010413 addi s0,sp,48
20: 03056683 lwu a3,48(a0)
24: 02069693 slli a3,a3,0x20
28: 0206d693 srli a3,a3,0x20
2c: 03456703 lwu a4,52(a0)
30: 02071713 slli a4,a4,0x20
34: 02075713 srli a4,a4,0x20
38: 03856483 lwu s1,56(a0)
3c: 02049493 slli s1,s1,0x20
40: 0204d493 srli s1,s1,0x20
44: 03c56903 lwu s2,60(a0)
48: 02091913 slli s2,s2,0x20
4c: 02095913 srli s2,s2,0x20
50: 04056983 lwu s3,64(a0)
54: 02099993 slli s3,s3,0x20
58: 0209d993 srli s3,s3,0x20
5c: 09056a03 lwu s4,144(a0)
60: 020a1a13 slli s4,s4,0x20
64: 020a5a13 srli s4,s4,0x20
68: 00900313 addi t1,zero,9
6c: 006a7463 bgeu s4,t1,0x74
70: 00000a13 addi s4,zero,0
74: 02d52823 sw a3,48(a0)
78: 02e52a23 sw a4,52(a0)
7c: 02952c23 sw s1,56(a0)
80: 03252e23 sw s2,60(a0)
84: 05352023 sw s3,64(a0)
88: 00000793 addi a5,zero,0
8c: 02813403 ld s0,40(sp)
90: 02013483 ld s1,32(sp)
94: 01813903 ld s2,24(sp)
98: 01013983 ld s3,16(sp)
9c: 00813a03 ld s4,8(sp)
a0: 03010113 addi sp,sp,48
a4: 00078513 addi a0,a5,0
a8: 00008067 jalr zero,0(ra)
With RVC:
0: 02000813 addi a6,zero,32
4: 7179 c.addi16sp sp,-48
6: f422 c.sdsp s0,40(sp)
8: f026 c.sdsp s1,32(sp)
a: ec4a c.sdsp s2,24(sp)
c: e84e c.sdsp s3,16(sp)
e: e452 c.sdsp s4,8(sp)
10: 1800 c.addi4spn s0,sp,48
12: 03056683 lwu a3,48(a0)
16: 1682 c.slli a3,0x20
18: 9281 c.srli a3,0x20
1a: 03456703 lwu a4,52(a0)
1e: 1702 c.slli a4,0x20
20: 9301 c.srli a4,0x20
22: 03856483 lwu s1,56(a0)
26: 1482 c.slli s1,0x20
28: 9081 c.srli s1,0x20
2a: 03c56903 lwu s2,60(a0)
2e: 1902 c.slli s2,0x20
30: 02095913 srli s2,s2,0x20
34: 04056983 lwu s3,64(a0)
38: 1982 c.slli s3,0x20
3a: 0209d993 srli s3,s3,0x20
3e: 09056a03 lwu s4,144(a0)
42: 1a02 c.slli s4,0x20
44: 020a5a13 srli s4,s4,0x20
48: 4325 c.li t1,9
4a: 006a7363 bgeu s4,t1,0x50
4e: 4a01 c.li s4,0
50: d914 c.sw a3,48(a0)
52: d958 c.sw a4,52(a0)
54: dd04 c.sw s1,56(a0)
56: 03252e23 sw s2,60(a0)
5a: 05352023 sw s3,64(a0)
5e: 4781 c.li a5,0
60: 7422 c.ldsp s0,40(sp)
62: 7482 c.ldsp s1,32(sp)
64: 6962 c.ldsp s2,24(sp)
66: 69c2 c.ldsp s3,16(sp)
68: 6a22 c.ldsp s4,8(sp)
6a: 6145 c.addi16sp sp,48
6c: 853e c.mv a0,a5
6e: 8082 c.jr ra
Signed-off-by: Luke Nelson <luke.r.nels@gmail.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Cc: Björn Töpel <bjorn.topel@gmail.com>
Link: https://lore.kernel.org/bpf/20200721025241.8077-4-luke.r.nels@gmail.com
2020-07-20 19:52:40 -07:00
|
|
|
emit_addiw(rd, rd, lower, ctx);
|
bpf, riscv: add BPF JIT for RV64G
This commit adds a BPF JIT for RV64G.
The JIT is a two-pass JIT, and has a dynamic prolog/epilogue (similar
to the MIPS64 BPF JIT) instead of static ones (e.g. x86_64).
At the moment the RISC-V Linux port does not support
CONFIG_HAVE_KPROBES, which means that CONFIG_BPF_EVENTS is not
supported. Thus, no tests involving BPF_PROG_TYPE_TRACEPOINT,
BPF_PROG_TYPE_PERF_EVENT, BPF_PROG_TYPE_KPROBE and
BPF_PROG_TYPE_RAW_TRACEPOINT passes.
The implementation does not support "far branching" (>4KiB).
Test results:
# modprobe test_bpf
test_bpf: Summary: 378 PASSED, 0 FAILED, [366/366 JIT'ed]
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier
...
Summary: 761 PASSED, 507 SKIPPED, 2 FAILED
Note that "test_verifier" was run with one build with
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y and one without, otherwise
many of the the tests that require unaligned access were skipped.
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
0
No CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
59
The two failing test_verifier tests are:
"ld_abs: vlan + abs, test 1"
"ld_abs: jump around ld_abs"
This is due to that "far branching" involved in those tests.
All tests where done on QEMU (QEMU emulator version 3.1.50
(v3.1.0-688-g8ae951fbc106)).
Signed-off-by: Björn Töpel <bjorn.topel@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2019-02-05 05:41:22 -07:00
|
|
|
return;
|
|
|
|
}
|
|
|
|
|
|
|
|
shift = __ffs(upper);
|
|
|
|
upper >>= shift;
|
|
|
|
shift += 12;
|
|
|
|
|
|
|
|
emit_imm(rd, upper, ctx);
|
|
|
|
|
bpf, riscv: Use compressed instructions in the rv64 JIT
This patch uses the RVC support and encodings from bpf_jit.h to optimize
the rv64 jit.
The optimizations work by replacing emit(rv_X(...)) with a call to a
helper function emit_X, which will emit a compressed version of the
instruction when possible, and when RVC is enabled.
The JIT continues to pass all tests in lib/test_bpf.c, and introduces
no new failures to test_verifier; both with and without RVC being enabled.
Most changes are straightforward replacements of emit(rv_X(...), ctx)
with emit_X(..., ctx), with the following exceptions bearing mention;
* Change emit_imm to sign-extend the value in "lower", since the
checks for RVC (and the instructions themselves) treat the value as
signed. Otherwise, small negative immediates will not be recognized as
encodable using an RVC instruction. For example, without this change,
emit_imm(rd, -1, ctx) would cause lower to become 4095, which is not a
6b int even though a "c.li rd, -1" instruction suffices.
* For {BPF_MOV,BPF_ADD} BPF_X, drop using addiw,addw in the 32-bit
cases since the values are zero-extended into the upper 32 bits in
the following instructions anyways, and the addition commutes with
zero-extension. (BPF_SUB BPF_X must still use subw since subtraction
does not commute with zero-extension.)
This patch avoids optimizing branches and jumps to use RVC instructions
since surrounding code often makes assumptions about the sizes of
emitted instructions. Optimizing these will require changing these
functions (e.g., emit_branch) to dynamically compute jump offsets.
The following are examples of the JITed code for the verifier selftest
"direct packet read test#3 for CGROUP_SKB OK", without and with RVC
enabled, respectively. The former uses 178 bytes, and the latter uses 112,
for a ~37% reduction in code size for this example.
Without RVC:
0: 02000813 addi a6,zero,32
4: fd010113 addi sp,sp,-48
8: 02813423 sd s0,40(sp)
c: 02913023 sd s1,32(sp)
10: 01213c23 sd s2,24(sp)
14: 01313823 sd s3,16(sp)
18: 01413423 sd s4,8(sp)
1c: 03010413 addi s0,sp,48
20: 03056683 lwu a3,48(a0)
24: 02069693 slli a3,a3,0x20
28: 0206d693 srli a3,a3,0x20
2c: 03456703 lwu a4,52(a0)
30: 02071713 slli a4,a4,0x20
34: 02075713 srli a4,a4,0x20
38: 03856483 lwu s1,56(a0)
3c: 02049493 slli s1,s1,0x20
40: 0204d493 srli s1,s1,0x20
44: 03c56903 lwu s2,60(a0)
48: 02091913 slli s2,s2,0x20
4c: 02095913 srli s2,s2,0x20
50: 04056983 lwu s3,64(a0)
54: 02099993 slli s3,s3,0x20
58: 0209d993 srli s3,s3,0x20
5c: 09056a03 lwu s4,144(a0)
60: 020a1a13 slli s4,s4,0x20
64: 020a5a13 srli s4,s4,0x20
68: 00900313 addi t1,zero,9
6c: 006a7463 bgeu s4,t1,0x74
70: 00000a13 addi s4,zero,0
74: 02d52823 sw a3,48(a0)
78: 02e52a23 sw a4,52(a0)
7c: 02952c23 sw s1,56(a0)
80: 03252e23 sw s2,60(a0)
84: 05352023 sw s3,64(a0)
88: 00000793 addi a5,zero,0
8c: 02813403 ld s0,40(sp)
90: 02013483 ld s1,32(sp)
94: 01813903 ld s2,24(sp)
98: 01013983 ld s3,16(sp)
9c: 00813a03 ld s4,8(sp)
a0: 03010113 addi sp,sp,48
a4: 00078513 addi a0,a5,0
a8: 00008067 jalr zero,0(ra)
With RVC:
0: 02000813 addi a6,zero,32
4: 7179 c.addi16sp sp,-48
6: f422 c.sdsp s0,40(sp)
8: f026 c.sdsp s1,32(sp)
a: ec4a c.sdsp s2,24(sp)
c: e84e c.sdsp s3,16(sp)
e: e452 c.sdsp s4,8(sp)
10: 1800 c.addi4spn s0,sp,48
12: 03056683 lwu a3,48(a0)
16: 1682 c.slli a3,0x20
18: 9281 c.srli a3,0x20
1a: 03456703 lwu a4,52(a0)
1e: 1702 c.slli a4,0x20
20: 9301 c.srli a4,0x20
22: 03856483 lwu s1,56(a0)
26: 1482 c.slli s1,0x20
28: 9081 c.srli s1,0x20
2a: 03c56903 lwu s2,60(a0)
2e: 1902 c.slli s2,0x20
30: 02095913 srli s2,s2,0x20
34: 04056983 lwu s3,64(a0)
38: 1982 c.slli s3,0x20
3a: 0209d993 srli s3,s3,0x20
3e: 09056a03 lwu s4,144(a0)
42: 1a02 c.slli s4,0x20
44: 020a5a13 srli s4,s4,0x20
48: 4325 c.li t1,9
4a: 006a7363 bgeu s4,t1,0x50
4e: 4a01 c.li s4,0
50: d914 c.sw a3,48(a0)
52: d958 c.sw a4,52(a0)
54: dd04 c.sw s1,56(a0)
56: 03252e23 sw s2,60(a0)
5a: 05352023 sw s3,64(a0)
5e: 4781 c.li a5,0
60: 7422 c.ldsp s0,40(sp)
62: 7482 c.ldsp s1,32(sp)
64: 6962 c.ldsp s2,24(sp)
66: 69c2 c.ldsp s3,16(sp)
68: 6a22 c.ldsp s4,8(sp)
6a: 6145 c.addi16sp sp,48
6c: 853e c.mv a0,a5
6e: 8082 c.jr ra
Signed-off-by: Luke Nelson <luke.r.nels@gmail.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Cc: Björn Töpel <bjorn.topel@gmail.com>
Link: https://lore.kernel.org/bpf/20200721025241.8077-4-luke.r.nels@gmail.com
2020-07-20 19:52:40 -07:00
|
|
|
emit_slli(rd, rd, shift, ctx);
|
bpf, riscv: add BPF JIT for RV64G
This commit adds a BPF JIT for RV64G.
The JIT is a two-pass JIT, and has a dynamic prolog/epilogue (similar
to the MIPS64 BPF JIT) instead of static ones (e.g. x86_64).
At the moment the RISC-V Linux port does not support
CONFIG_HAVE_KPROBES, which means that CONFIG_BPF_EVENTS is not
supported. Thus, no tests involving BPF_PROG_TYPE_TRACEPOINT,
BPF_PROG_TYPE_PERF_EVENT, BPF_PROG_TYPE_KPROBE and
BPF_PROG_TYPE_RAW_TRACEPOINT passes.
The implementation does not support "far branching" (>4KiB).
Test results:
# modprobe test_bpf
test_bpf: Summary: 378 PASSED, 0 FAILED, [366/366 JIT'ed]
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier
...
Summary: 761 PASSED, 507 SKIPPED, 2 FAILED
Note that "test_verifier" was run with one build with
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y and one without, otherwise
many of the the tests that require unaligned access were skipped.
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
0
No CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
59
The two failing test_verifier tests are:
"ld_abs: vlan + abs, test 1"
"ld_abs: jump around ld_abs"
This is due to that "far branching" involved in those tests.
All tests where done on QEMU (QEMU emulator version 3.1.50
(v3.1.0-688-g8ae951fbc106)).
Signed-off-by: Björn Töpel <bjorn.topel@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2019-02-05 05:41:22 -07:00
|
|
|
if (lower)
|
bpf, riscv: Use compressed instructions in the rv64 JIT
This patch uses the RVC support and encodings from bpf_jit.h to optimize
the rv64 jit.
The optimizations work by replacing emit(rv_X(...)) with a call to a
helper function emit_X, which will emit a compressed version of the
instruction when possible, and when RVC is enabled.
The JIT continues to pass all tests in lib/test_bpf.c, and introduces
no new failures to test_verifier; both with and without RVC being enabled.
Most changes are straightforward replacements of emit(rv_X(...), ctx)
with emit_X(..., ctx), with the following exceptions bearing mention;
* Change emit_imm to sign-extend the value in "lower", since the
checks for RVC (and the instructions themselves) treat the value as
signed. Otherwise, small negative immediates will not be recognized as
encodable using an RVC instruction. For example, without this change,
emit_imm(rd, -1, ctx) would cause lower to become 4095, which is not a
6b int even though a "c.li rd, -1" instruction suffices.
* For {BPF_MOV,BPF_ADD} BPF_X, drop using addiw,addw in the 32-bit
cases since the values are zero-extended into the upper 32 bits in
the following instructions anyways, and the addition commutes with
zero-extension. (BPF_SUB BPF_X must still use subw since subtraction
does not commute with zero-extension.)
This patch avoids optimizing branches and jumps to use RVC instructions
since surrounding code often makes assumptions about the sizes of
emitted instructions. Optimizing these will require changing these
functions (e.g., emit_branch) to dynamically compute jump offsets.
The following are examples of the JITed code for the verifier selftest
"direct packet read test#3 for CGROUP_SKB OK", without and with RVC
enabled, respectively. The former uses 178 bytes, and the latter uses 112,
for a ~37% reduction in code size for this example.
Without RVC:
0: 02000813 addi a6,zero,32
4: fd010113 addi sp,sp,-48
8: 02813423 sd s0,40(sp)
c: 02913023 sd s1,32(sp)
10: 01213c23 sd s2,24(sp)
14: 01313823 sd s3,16(sp)
18: 01413423 sd s4,8(sp)
1c: 03010413 addi s0,sp,48
20: 03056683 lwu a3,48(a0)
24: 02069693 slli a3,a3,0x20
28: 0206d693 srli a3,a3,0x20
2c: 03456703 lwu a4,52(a0)
30: 02071713 slli a4,a4,0x20
34: 02075713 srli a4,a4,0x20
38: 03856483 lwu s1,56(a0)
3c: 02049493 slli s1,s1,0x20
40: 0204d493 srli s1,s1,0x20
44: 03c56903 lwu s2,60(a0)
48: 02091913 slli s2,s2,0x20
4c: 02095913 srli s2,s2,0x20
50: 04056983 lwu s3,64(a0)
54: 02099993 slli s3,s3,0x20
58: 0209d993 srli s3,s3,0x20
5c: 09056a03 lwu s4,144(a0)
60: 020a1a13 slli s4,s4,0x20
64: 020a5a13 srli s4,s4,0x20
68: 00900313 addi t1,zero,9
6c: 006a7463 bgeu s4,t1,0x74
70: 00000a13 addi s4,zero,0
74: 02d52823 sw a3,48(a0)
78: 02e52a23 sw a4,52(a0)
7c: 02952c23 sw s1,56(a0)
80: 03252e23 sw s2,60(a0)
84: 05352023 sw s3,64(a0)
88: 00000793 addi a5,zero,0
8c: 02813403 ld s0,40(sp)
90: 02013483 ld s1,32(sp)
94: 01813903 ld s2,24(sp)
98: 01013983 ld s3,16(sp)
9c: 00813a03 ld s4,8(sp)
a0: 03010113 addi sp,sp,48
a4: 00078513 addi a0,a5,0
a8: 00008067 jalr zero,0(ra)
With RVC:
0: 02000813 addi a6,zero,32
4: 7179 c.addi16sp sp,-48
6: f422 c.sdsp s0,40(sp)
8: f026 c.sdsp s1,32(sp)
a: ec4a c.sdsp s2,24(sp)
c: e84e c.sdsp s3,16(sp)
e: e452 c.sdsp s4,8(sp)
10: 1800 c.addi4spn s0,sp,48
12: 03056683 lwu a3,48(a0)
16: 1682 c.slli a3,0x20
18: 9281 c.srli a3,0x20
1a: 03456703 lwu a4,52(a0)
1e: 1702 c.slli a4,0x20
20: 9301 c.srli a4,0x20
22: 03856483 lwu s1,56(a0)
26: 1482 c.slli s1,0x20
28: 9081 c.srli s1,0x20
2a: 03c56903 lwu s2,60(a0)
2e: 1902 c.slli s2,0x20
30: 02095913 srli s2,s2,0x20
34: 04056983 lwu s3,64(a0)
38: 1982 c.slli s3,0x20
3a: 0209d993 srli s3,s3,0x20
3e: 09056a03 lwu s4,144(a0)
42: 1a02 c.slli s4,0x20
44: 020a5a13 srli s4,s4,0x20
48: 4325 c.li t1,9
4a: 006a7363 bgeu s4,t1,0x50
4e: 4a01 c.li s4,0
50: d914 c.sw a3,48(a0)
52: d958 c.sw a4,52(a0)
54: dd04 c.sw s1,56(a0)
56: 03252e23 sw s2,60(a0)
5a: 05352023 sw s3,64(a0)
5e: 4781 c.li a5,0
60: 7422 c.ldsp s0,40(sp)
62: 7482 c.ldsp s1,32(sp)
64: 6962 c.ldsp s2,24(sp)
66: 69c2 c.ldsp s3,16(sp)
68: 6a22 c.ldsp s4,8(sp)
6a: 6145 c.addi16sp sp,48
6c: 853e c.mv a0,a5
6e: 8082 c.jr ra
Signed-off-by: Luke Nelson <luke.r.nels@gmail.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Cc: Björn Töpel <bjorn.topel@gmail.com>
Link: https://lore.kernel.org/bpf/20200721025241.8077-4-luke.r.nels@gmail.com
2020-07-20 19:52:40 -07:00
|
|
|
emit_addi(rd, rd, lower, ctx);
|
bpf, riscv: add BPF JIT for RV64G
This commit adds a BPF JIT for RV64G.
The JIT is a two-pass JIT, and has a dynamic prolog/epilogue (similar
to the MIPS64 BPF JIT) instead of static ones (e.g. x86_64).
At the moment the RISC-V Linux port does not support
CONFIG_HAVE_KPROBES, which means that CONFIG_BPF_EVENTS is not
supported. Thus, no tests involving BPF_PROG_TYPE_TRACEPOINT,
BPF_PROG_TYPE_PERF_EVENT, BPF_PROG_TYPE_KPROBE and
BPF_PROG_TYPE_RAW_TRACEPOINT passes.
The implementation does not support "far branching" (>4KiB).
Test results:
# modprobe test_bpf
test_bpf: Summary: 378 PASSED, 0 FAILED, [366/366 JIT'ed]
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier
...
Summary: 761 PASSED, 507 SKIPPED, 2 FAILED
Note that "test_verifier" was run with one build with
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y and one without, otherwise
many of the the tests that require unaligned access were skipped.
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
0
No CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
59
The two failing test_verifier tests are:
"ld_abs: vlan + abs, test 1"
"ld_abs: jump around ld_abs"
This is due to that "far branching" involved in those tests.
All tests where done on QEMU (QEMU emulator version 3.1.50
(v3.1.0-688-g8ae951fbc106)).
Signed-off-by: Björn Töpel <bjorn.topel@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2019-02-05 05:41:22 -07:00
|
|
|
}
|
|
|
|
|
2019-12-16 02:13:39 -07:00
|
|
|
static void __build_epilogue(bool is_tail_call, struct rv_jit_context *ctx)
|
bpf, riscv: add BPF JIT for RV64G
This commit adds a BPF JIT for RV64G.
The JIT is a two-pass JIT, and has a dynamic prolog/epilogue (similar
to the MIPS64 BPF JIT) instead of static ones (e.g. x86_64).
At the moment the RISC-V Linux port does not support
CONFIG_HAVE_KPROBES, which means that CONFIG_BPF_EVENTS is not
supported. Thus, no tests involving BPF_PROG_TYPE_TRACEPOINT,
BPF_PROG_TYPE_PERF_EVENT, BPF_PROG_TYPE_KPROBE and
BPF_PROG_TYPE_RAW_TRACEPOINT passes.
The implementation does not support "far branching" (>4KiB).
Test results:
# modprobe test_bpf
test_bpf: Summary: 378 PASSED, 0 FAILED, [366/366 JIT'ed]
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier
...
Summary: 761 PASSED, 507 SKIPPED, 2 FAILED
Note that "test_verifier" was run with one build with
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y and one without, otherwise
many of the the tests that require unaligned access were skipped.
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
0
No CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
59
The two failing test_verifier tests are:
"ld_abs: vlan + abs, test 1"
"ld_abs: jump around ld_abs"
This is due to that "far branching" involved in those tests.
All tests where done on QEMU (QEMU emulator version 3.1.50
(v3.1.0-688-g8ae951fbc106)).
Signed-off-by: Björn Töpel <bjorn.topel@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2019-02-05 05:41:22 -07:00
|
|
|
{
|
|
|
|
int stack_adjust = ctx->stack_size, store_offset = stack_adjust - 8;
|
|
|
|
|
|
|
|
if (seen_reg(RV_REG_RA, ctx)) {
|
bpf, riscv: Use compressed instructions in the rv64 JIT
This patch uses the RVC support and encodings from bpf_jit.h to optimize
the rv64 jit.
The optimizations work by replacing emit(rv_X(...)) with a call to a
helper function emit_X, which will emit a compressed version of the
instruction when possible, and when RVC is enabled.
The JIT continues to pass all tests in lib/test_bpf.c, and introduces
no new failures to test_verifier; both with and without RVC being enabled.
Most changes are straightforward replacements of emit(rv_X(...), ctx)
with emit_X(..., ctx), with the following exceptions bearing mention;
* Change emit_imm to sign-extend the value in "lower", since the
checks for RVC (and the instructions themselves) treat the value as
signed. Otherwise, small negative immediates will not be recognized as
encodable using an RVC instruction. For example, without this change,
emit_imm(rd, -1, ctx) would cause lower to become 4095, which is not a
6b int even though a "c.li rd, -1" instruction suffices.
* For {BPF_MOV,BPF_ADD} BPF_X, drop using addiw,addw in the 32-bit
cases since the values are zero-extended into the upper 32 bits in
the following instructions anyways, and the addition commutes with
zero-extension. (BPF_SUB BPF_X must still use subw since subtraction
does not commute with zero-extension.)
This patch avoids optimizing branches and jumps to use RVC instructions
since surrounding code often makes assumptions about the sizes of
emitted instructions. Optimizing these will require changing these
functions (e.g., emit_branch) to dynamically compute jump offsets.
The following are examples of the JITed code for the verifier selftest
"direct packet read test#3 for CGROUP_SKB OK", without and with RVC
enabled, respectively. The former uses 178 bytes, and the latter uses 112,
for a ~37% reduction in code size for this example.
Without RVC:
0: 02000813 addi a6,zero,32
4: fd010113 addi sp,sp,-48
8: 02813423 sd s0,40(sp)
c: 02913023 sd s1,32(sp)
10: 01213c23 sd s2,24(sp)
14: 01313823 sd s3,16(sp)
18: 01413423 sd s4,8(sp)
1c: 03010413 addi s0,sp,48
20: 03056683 lwu a3,48(a0)
24: 02069693 slli a3,a3,0x20
28: 0206d693 srli a3,a3,0x20
2c: 03456703 lwu a4,52(a0)
30: 02071713 slli a4,a4,0x20
34: 02075713 srli a4,a4,0x20
38: 03856483 lwu s1,56(a0)
3c: 02049493 slli s1,s1,0x20
40: 0204d493 srli s1,s1,0x20
44: 03c56903 lwu s2,60(a0)
48: 02091913 slli s2,s2,0x20
4c: 02095913 srli s2,s2,0x20
50: 04056983 lwu s3,64(a0)
54: 02099993 slli s3,s3,0x20
58: 0209d993 srli s3,s3,0x20
5c: 09056a03 lwu s4,144(a0)
60: 020a1a13 slli s4,s4,0x20
64: 020a5a13 srli s4,s4,0x20
68: 00900313 addi t1,zero,9
6c: 006a7463 bgeu s4,t1,0x74
70: 00000a13 addi s4,zero,0
74: 02d52823 sw a3,48(a0)
78: 02e52a23 sw a4,52(a0)
7c: 02952c23 sw s1,56(a0)
80: 03252e23 sw s2,60(a0)
84: 05352023 sw s3,64(a0)
88: 00000793 addi a5,zero,0
8c: 02813403 ld s0,40(sp)
90: 02013483 ld s1,32(sp)
94: 01813903 ld s2,24(sp)
98: 01013983 ld s3,16(sp)
9c: 00813a03 ld s4,8(sp)
a0: 03010113 addi sp,sp,48
a4: 00078513 addi a0,a5,0
a8: 00008067 jalr zero,0(ra)
With RVC:
0: 02000813 addi a6,zero,32
4: 7179 c.addi16sp sp,-48
6: f422 c.sdsp s0,40(sp)
8: f026 c.sdsp s1,32(sp)
a: ec4a c.sdsp s2,24(sp)
c: e84e c.sdsp s3,16(sp)
e: e452 c.sdsp s4,8(sp)
10: 1800 c.addi4spn s0,sp,48
12: 03056683 lwu a3,48(a0)
16: 1682 c.slli a3,0x20
18: 9281 c.srli a3,0x20
1a: 03456703 lwu a4,52(a0)
1e: 1702 c.slli a4,0x20
20: 9301 c.srli a4,0x20
22: 03856483 lwu s1,56(a0)
26: 1482 c.slli s1,0x20
28: 9081 c.srli s1,0x20
2a: 03c56903 lwu s2,60(a0)
2e: 1902 c.slli s2,0x20
30: 02095913 srli s2,s2,0x20
34: 04056983 lwu s3,64(a0)
38: 1982 c.slli s3,0x20
3a: 0209d993 srli s3,s3,0x20
3e: 09056a03 lwu s4,144(a0)
42: 1a02 c.slli s4,0x20
44: 020a5a13 srli s4,s4,0x20
48: 4325 c.li t1,9
4a: 006a7363 bgeu s4,t1,0x50
4e: 4a01 c.li s4,0
50: d914 c.sw a3,48(a0)
52: d958 c.sw a4,52(a0)
54: dd04 c.sw s1,56(a0)
56: 03252e23 sw s2,60(a0)
5a: 05352023 sw s3,64(a0)
5e: 4781 c.li a5,0
60: 7422 c.ldsp s0,40(sp)
62: 7482 c.ldsp s1,32(sp)
64: 6962 c.ldsp s2,24(sp)
66: 69c2 c.ldsp s3,16(sp)
68: 6a22 c.ldsp s4,8(sp)
6a: 6145 c.addi16sp sp,48
6c: 853e c.mv a0,a5
6e: 8082 c.jr ra
Signed-off-by: Luke Nelson <luke.r.nels@gmail.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Cc: Björn Töpel <bjorn.topel@gmail.com>
Link: https://lore.kernel.org/bpf/20200721025241.8077-4-luke.r.nels@gmail.com
2020-07-20 19:52:40 -07:00
|
|
|
emit_ld(RV_REG_RA, store_offset, RV_REG_SP, ctx);
|
bpf, riscv: add BPF JIT for RV64G
This commit adds a BPF JIT for RV64G.
The JIT is a two-pass JIT, and has a dynamic prolog/epilogue (similar
to the MIPS64 BPF JIT) instead of static ones (e.g. x86_64).
At the moment the RISC-V Linux port does not support
CONFIG_HAVE_KPROBES, which means that CONFIG_BPF_EVENTS is not
supported. Thus, no tests involving BPF_PROG_TYPE_TRACEPOINT,
BPF_PROG_TYPE_PERF_EVENT, BPF_PROG_TYPE_KPROBE and
BPF_PROG_TYPE_RAW_TRACEPOINT passes.
The implementation does not support "far branching" (>4KiB).
Test results:
# modprobe test_bpf
test_bpf: Summary: 378 PASSED, 0 FAILED, [366/366 JIT'ed]
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier
...
Summary: 761 PASSED, 507 SKIPPED, 2 FAILED
Note that "test_verifier" was run with one build with
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y and one without, otherwise
many of the the tests that require unaligned access were skipped.
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
0
No CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
59
The two failing test_verifier tests are:
"ld_abs: vlan + abs, test 1"
"ld_abs: jump around ld_abs"
This is due to that "far branching" involved in those tests.
All tests where done on QEMU (QEMU emulator version 3.1.50
(v3.1.0-688-g8ae951fbc106)).
Signed-off-by: Björn Töpel <bjorn.topel@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2019-02-05 05:41:22 -07:00
|
|
|
store_offset -= 8;
|
|
|
|
}
|
bpf, riscv: Use compressed instructions in the rv64 JIT
This patch uses the RVC support and encodings from bpf_jit.h to optimize
the rv64 jit.
The optimizations work by replacing emit(rv_X(...)) with a call to a
helper function emit_X, which will emit a compressed version of the
instruction when possible, and when RVC is enabled.
The JIT continues to pass all tests in lib/test_bpf.c, and introduces
no new failures to test_verifier; both with and without RVC being enabled.
Most changes are straightforward replacements of emit(rv_X(...), ctx)
with emit_X(..., ctx), with the following exceptions bearing mention;
* Change emit_imm to sign-extend the value in "lower", since the
checks for RVC (and the instructions themselves) treat the value as
signed. Otherwise, small negative immediates will not be recognized as
encodable using an RVC instruction. For example, without this change,
emit_imm(rd, -1, ctx) would cause lower to become 4095, which is not a
6b int even though a "c.li rd, -1" instruction suffices.
* For {BPF_MOV,BPF_ADD} BPF_X, drop using addiw,addw in the 32-bit
cases since the values are zero-extended into the upper 32 bits in
the following instructions anyways, and the addition commutes with
zero-extension. (BPF_SUB BPF_X must still use subw since subtraction
does not commute with zero-extension.)
This patch avoids optimizing branches and jumps to use RVC instructions
since surrounding code often makes assumptions about the sizes of
emitted instructions. Optimizing these will require changing these
functions (e.g., emit_branch) to dynamically compute jump offsets.
The following are examples of the JITed code for the verifier selftest
"direct packet read test#3 for CGROUP_SKB OK", without and with RVC
enabled, respectively. The former uses 178 bytes, and the latter uses 112,
for a ~37% reduction in code size for this example.
Without RVC:
0: 02000813 addi a6,zero,32
4: fd010113 addi sp,sp,-48
8: 02813423 sd s0,40(sp)
c: 02913023 sd s1,32(sp)
10: 01213c23 sd s2,24(sp)
14: 01313823 sd s3,16(sp)
18: 01413423 sd s4,8(sp)
1c: 03010413 addi s0,sp,48
20: 03056683 lwu a3,48(a0)
24: 02069693 slli a3,a3,0x20
28: 0206d693 srli a3,a3,0x20
2c: 03456703 lwu a4,52(a0)
30: 02071713 slli a4,a4,0x20
34: 02075713 srli a4,a4,0x20
38: 03856483 lwu s1,56(a0)
3c: 02049493 slli s1,s1,0x20
40: 0204d493 srli s1,s1,0x20
44: 03c56903 lwu s2,60(a0)
48: 02091913 slli s2,s2,0x20
4c: 02095913 srli s2,s2,0x20
50: 04056983 lwu s3,64(a0)
54: 02099993 slli s3,s3,0x20
58: 0209d993 srli s3,s3,0x20
5c: 09056a03 lwu s4,144(a0)
60: 020a1a13 slli s4,s4,0x20
64: 020a5a13 srli s4,s4,0x20
68: 00900313 addi t1,zero,9
6c: 006a7463 bgeu s4,t1,0x74
70: 00000a13 addi s4,zero,0
74: 02d52823 sw a3,48(a0)
78: 02e52a23 sw a4,52(a0)
7c: 02952c23 sw s1,56(a0)
80: 03252e23 sw s2,60(a0)
84: 05352023 sw s3,64(a0)
88: 00000793 addi a5,zero,0
8c: 02813403 ld s0,40(sp)
90: 02013483 ld s1,32(sp)
94: 01813903 ld s2,24(sp)
98: 01013983 ld s3,16(sp)
9c: 00813a03 ld s4,8(sp)
a0: 03010113 addi sp,sp,48
a4: 00078513 addi a0,a5,0
a8: 00008067 jalr zero,0(ra)
With RVC:
0: 02000813 addi a6,zero,32
4: 7179 c.addi16sp sp,-48
6: f422 c.sdsp s0,40(sp)
8: f026 c.sdsp s1,32(sp)
a: ec4a c.sdsp s2,24(sp)
c: e84e c.sdsp s3,16(sp)
e: e452 c.sdsp s4,8(sp)
10: 1800 c.addi4spn s0,sp,48
12: 03056683 lwu a3,48(a0)
16: 1682 c.slli a3,0x20
18: 9281 c.srli a3,0x20
1a: 03456703 lwu a4,52(a0)
1e: 1702 c.slli a4,0x20
20: 9301 c.srli a4,0x20
22: 03856483 lwu s1,56(a0)
26: 1482 c.slli s1,0x20
28: 9081 c.srli s1,0x20
2a: 03c56903 lwu s2,60(a0)
2e: 1902 c.slli s2,0x20
30: 02095913 srli s2,s2,0x20
34: 04056983 lwu s3,64(a0)
38: 1982 c.slli s3,0x20
3a: 0209d993 srli s3,s3,0x20
3e: 09056a03 lwu s4,144(a0)
42: 1a02 c.slli s4,0x20
44: 020a5a13 srli s4,s4,0x20
48: 4325 c.li t1,9
4a: 006a7363 bgeu s4,t1,0x50
4e: 4a01 c.li s4,0
50: d914 c.sw a3,48(a0)
52: d958 c.sw a4,52(a0)
54: dd04 c.sw s1,56(a0)
56: 03252e23 sw s2,60(a0)
5a: 05352023 sw s3,64(a0)
5e: 4781 c.li a5,0
60: 7422 c.ldsp s0,40(sp)
62: 7482 c.ldsp s1,32(sp)
64: 6962 c.ldsp s2,24(sp)
66: 69c2 c.ldsp s3,16(sp)
68: 6a22 c.ldsp s4,8(sp)
6a: 6145 c.addi16sp sp,48
6c: 853e c.mv a0,a5
6e: 8082 c.jr ra
Signed-off-by: Luke Nelson <luke.r.nels@gmail.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Cc: Björn Töpel <bjorn.topel@gmail.com>
Link: https://lore.kernel.org/bpf/20200721025241.8077-4-luke.r.nels@gmail.com
2020-07-20 19:52:40 -07:00
|
|
|
emit_ld(RV_REG_FP, store_offset, RV_REG_SP, ctx);
|
bpf, riscv: add BPF JIT for RV64G
This commit adds a BPF JIT for RV64G.
The JIT is a two-pass JIT, and has a dynamic prolog/epilogue (similar
to the MIPS64 BPF JIT) instead of static ones (e.g. x86_64).
At the moment the RISC-V Linux port does not support
CONFIG_HAVE_KPROBES, which means that CONFIG_BPF_EVENTS is not
supported. Thus, no tests involving BPF_PROG_TYPE_TRACEPOINT,
BPF_PROG_TYPE_PERF_EVENT, BPF_PROG_TYPE_KPROBE and
BPF_PROG_TYPE_RAW_TRACEPOINT passes.
The implementation does not support "far branching" (>4KiB).
Test results:
# modprobe test_bpf
test_bpf: Summary: 378 PASSED, 0 FAILED, [366/366 JIT'ed]
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier
...
Summary: 761 PASSED, 507 SKIPPED, 2 FAILED
Note that "test_verifier" was run with one build with
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y and one without, otherwise
many of the the tests that require unaligned access were skipped.
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
0
No CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
59
The two failing test_verifier tests are:
"ld_abs: vlan + abs, test 1"
"ld_abs: jump around ld_abs"
This is due to that "far branching" involved in those tests.
All tests where done on QEMU (QEMU emulator version 3.1.50
(v3.1.0-688-g8ae951fbc106)).
Signed-off-by: Björn Töpel <bjorn.topel@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2019-02-05 05:41:22 -07:00
|
|
|
store_offset -= 8;
|
|
|
|
if (seen_reg(RV_REG_S1, ctx)) {
|
bpf, riscv: Use compressed instructions in the rv64 JIT
This patch uses the RVC support and encodings from bpf_jit.h to optimize
the rv64 jit.
The optimizations work by replacing emit(rv_X(...)) with a call to a
helper function emit_X, which will emit a compressed version of the
instruction when possible, and when RVC is enabled.
The JIT continues to pass all tests in lib/test_bpf.c, and introduces
no new failures to test_verifier; both with and without RVC being enabled.
Most changes are straightforward replacements of emit(rv_X(...), ctx)
with emit_X(..., ctx), with the following exceptions bearing mention;
* Change emit_imm to sign-extend the value in "lower", since the
checks for RVC (and the instructions themselves) treat the value as
signed. Otherwise, small negative immediates will not be recognized as
encodable using an RVC instruction. For example, without this change,
emit_imm(rd, -1, ctx) would cause lower to become 4095, which is not a
6b int even though a "c.li rd, -1" instruction suffices.
* For {BPF_MOV,BPF_ADD} BPF_X, drop using addiw,addw in the 32-bit
cases since the values are zero-extended into the upper 32 bits in
the following instructions anyways, and the addition commutes with
zero-extension. (BPF_SUB BPF_X must still use subw since subtraction
does not commute with zero-extension.)
This patch avoids optimizing branches and jumps to use RVC instructions
since surrounding code often makes assumptions about the sizes of
emitted instructions. Optimizing these will require changing these
functions (e.g., emit_branch) to dynamically compute jump offsets.
The following are examples of the JITed code for the verifier selftest
"direct packet read test#3 for CGROUP_SKB OK", without and with RVC
enabled, respectively. The former uses 178 bytes, and the latter uses 112,
for a ~37% reduction in code size for this example.
Without RVC:
0: 02000813 addi a6,zero,32
4: fd010113 addi sp,sp,-48
8: 02813423 sd s0,40(sp)
c: 02913023 sd s1,32(sp)
10: 01213c23 sd s2,24(sp)
14: 01313823 sd s3,16(sp)
18: 01413423 sd s4,8(sp)
1c: 03010413 addi s0,sp,48
20: 03056683 lwu a3,48(a0)
24: 02069693 slli a3,a3,0x20
28: 0206d693 srli a3,a3,0x20
2c: 03456703 lwu a4,52(a0)
30: 02071713 slli a4,a4,0x20
34: 02075713 srli a4,a4,0x20
38: 03856483 lwu s1,56(a0)
3c: 02049493 slli s1,s1,0x20
40: 0204d493 srli s1,s1,0x20
44: 03c56903 lwu s2,60(a0)
48: 02091913 slli s2,s2,0x20
4c: 02095913 srli s2,s2,0x20
50: 04056983 lwu s3,64(a0)
54: 02099993 slli s3,s3,0x20
58: 0209d993 srli s3,s3,0x20
5c: 09056a03 lwu s4,144(a0)
60: 020a1a13 slli s4,s4,0x20
64: 020a5a13 srli s4,s4,0x20
68: 00900313 addi t1,zero,9
6c: 006a7463 bgeu s4,t1,0x74
70: 00000a13 addi s4,zero,0
74: 02d52823 sw a3,48(a0)
78: 02e52a23 sw a4,52(a0)
7c: 02952c23 sw s1,56(a0)
80: 03252e23 sw s2,60(a0)
84: 05352023 sw s3,64(a0)
88: 00000793 addi a5,zero,0
8c: 02813403 ld s0,40(sp)
90: 02013483 ld s1,32(sp)
94: 01813903 ld s2,24(sp)
98: 01013983 ld s3,16(sp)
9c: 00813a03 ld s4,8(sp)
a0: 03010113 addi sp,sp,48
a4: 00078513 addi a0,a5,0
a8: 00008067 jalr zero,0(ra)
With RVC:
0: 02000813 addi a6,zero,32
4: 7179 c.addi16sp sp,-48
6: f422 c.sdsp s0,40(sp)
8: f026 c.sdsp s1,32(sp)
a: ec4a c.sdsp s2,24(sp)
c: e84e c.sdsp s3,16(sp)
e: e452 c.sdsp s4,8(sp)
10: 1800 c.addi4spn s0,sp,48
12: 03056683 lwu a3,48(a0)
16: 1682 c.slli a3,0x20
18: 9281 c.srli a3,0x20
1a: 03456703 lwu a4,52(a0)
1e: 1702 c.slli a4,0x20
20: 9301 c.srli a4,0x20
22: 03856483 lwu s1,56(a0)
26: 1482 c.slli s1,0x20
28: 9081 c.srli s1,0x20
2a: 03c56903 lwu s2,60(a0)
2e: 1902 c.slli s2,0x20
30: 02095913 srli s2,s2,0x20
34: 04056983 lwu s3,64(a0)
38: 1982 c.slli s3,0x20
3a: 0209d993 srli s3,s3,0x20
3e: 09056a03 lwu s4,144(a0)
42: 1a02 c.slli s4,0x20
44: 020a5a13 srli s4,s4,0x20
48: 4325 c.li t1,9
4a: 006a7363 bgeu s4,t1,0x50
4e: 4a01 c.li s4,0
50: d914 c.sw a3,48(a0)
52: d958 c.sw a4,52(a0)
54: dd04 c.sw s1,56(a0)
56: 03252e23 sw s2,60(a0)
5a: 05352023 sw s3,64(a0)
5e: 4781 c.li a5,0
60: 7422 c.ldsp s0,40(sp)
62: 7482 c.ldsp s1,32(sp)
64: 6962 c.ldsp s2,24(sp)
66: 69c2 c.ldsp s3,16(sp)
68: 6a22 c.ldsp s4,8(sp)
6a: 6145 c.addi16sp sp,48
6c: 853e c.mv a0,a5
6e: 8082 c.jr ra
Signed-off-by: Luke Nelson <luke.r.nels@gmail.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Cc: Björn Töpel <bjorn.topel@gmail.com>
Link: https://lore.kernel.org/bpf/20200721025241.8077-4-luke.r.nels@gmail.com
2020-07-20 19:52:40 -07:00
|
|
|
emit_ld(RV_REG_S1, store_offset, RV_REG_SP, ctx);
|
bpf, riscv: add BPF JIT for RV64G
This commit adds a BPF JIT for RV64G.
The JIT is a two-pass JIT, and has a dynamic prolog/epilogue (similar
to the MIPS64 BPF JIT) instead of static ones (e.g. x86_64).
At the moment the RISC-V Linux port does not support
CONFIG_HAVE_KPROBES, which means that CONFIG_BPF_EVENTS is not
supported. Thus, no tests involving BPF_PROG_TYPE_TRACEPOINT,
BPF_PROG_TYPE_PERF_EVENT, BPF_PROG_TYPE_KPROBE and
BPF_PROG_TYPE_RAW_TRACEPOINT passes.
The implementation does not support "far branching" (>4KiB).
Test results:
# modprobe test_bpf
test_bpf: Summary: 378 PASSED, 0 FAILED, [366/366 JIT'ed]
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier
...
Summary: 761 PASSED, 507 SKIPPED, 2 FAILED
Note that "test_verifier" was run with one build with
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y and one without, otherwise
many of the the tests that require unaligned access were skipped.
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
0
No CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
59
The two failing test_verifier tests are:
"ld_abs: vlan + abs, test 1"
"ld_abs: jump around ld_abs"
This is due to that "far branching" involved in those tests.
All tests where done on QEMU (QEMU emulator version 3.1.50
(v3.1.0-688-g8ae951fbc106)).
Signed-off-by: Björn Töpel <bjorn.topel@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2019-02-05 05:41:22 -07:00
|
|
|
store_offset -= 8;
|
|
|
|
}
|
|
|
|
if (seen_reg(RV_REG_S2, ctx)) {
|
bpf, riscv: Use compressed instructions in the rv64 JIT
This patch uses the RVC support and encodings from bpf_jit.h to optimize
the rv64 jit.
The optimizations work by replacing emit(rv_X(...)) with a call to a
helper function emit_X, which will emit a compressed version of the
instruction when possible, and when RVC is enabled.
The JIT continues to pass all tests in lib/test_bpf.c, and introduces
no new failures to test_verifier; both with and without RVC being enabled.
Most changes are straightforward replacements of emit(rv_X(...), ctx)
with emit_X(..., ctx), with the following exceptions bearing mention;
* Change emit_imm to sign-extend the value in "lower", since the
checks for RVC (and the instructions themselves) treat the value as
signed. Otherwise, small negative immediates will not be recognized as
encodable using an RVC instruction. For example, without this change,
emit_imm(rd, -1, ctx) would cause lower to become 4095, which is not a
6b int even though a "c.li rd, -1" instruction suffices.
* For {BPF_MOV,BPF_ADD} BPF_X, drop using addiw,addw in the 32-bit
cases since the values are zero-extended into the upper 32 bits in
the following instructions anyways, and the addition commutes with
zero-extension. (BPF_SUB BPF_X must still use subw since subtraction
does not commute with zero-extension.)
This patch avoids optimizing branches and jumps to use RVC instructions
since surrounding code often makes assumptions about the sizes of
emitted instructions. Optimizing these will require changing these
functions (e.g., emit_branch) to dynamically compute jump offsets.
The following are examples of the JITed code for the verifier selftest
"direct packet read test#3 for CGROUP_SKB OK", without and with RVC
enabled, respectively. The former uses 178 bytes, and the latter uses 112,
for a ~37% reduction in code size for this example.
Without RVC:
0: 02000813 addi a6,zero,32
4: fd010113 addi sp,sp,-48
8: 02813423 sd s0,40(sp)
c: 02913023 sd s1,32(sp)
10: 01213c23 sd s2,24(sp)
14: 01313823 sd s3,16(sp)
18: 01413423 sd s4,8(sp)
1c: 03010413 addi s0,sp,48
20: 03056683 lwu a3,48(a0)
24: 02069693 slli a3,a3,0x20
28: 0206d693 srli a3,a3,0x20
2c: 03456703 lwu a4,52(a0)
30: 02071713 slli a4,a4,0x20
34: 02075713 srli a4,a4,0x20
38: 03856483 lwu s1,56(a0)
3c: 02049493 slli s1,s1,0x20
40: 0204d493 srli s1,s1,0x20
44: 03c56903 lwu s2,60(a0)
48: 02091913 slli s2,s2,0x20
4c: 02095913 srli s2,s2,0x20
50: 04056983 lwu s3,64(a0)
54: 02099993 slli s3,s3,0x20
58: 0209d993 srli s3,s3,0x20
5c: 09056a03 lwu s4,144(a0)
60: 020a1a13 slli s4,s4,0x20
64: 020a5a13 srli s4,s4,0x20
68: 00900313 addi t1,zero,9
6c: 006a7463 bgeu s4,t1,0x74
70: 00000a13 addi s4,zero,0
74: 02d52823 sw a3,48(a0)
78: 02e52a23 sw a4,52(a0)
7c: 02952c23 sw s1,56(a0)
80: 03252e23 sw s2,60(a0)
84: 05352023 sw s3,64(a0)
88: 00000793 addi a5,zero,0
8c: 02813403 ld s0,40(sp)
90: 02013483 ld s1,32(sp)
94: 01813903 ld s2,24(sp)
98: 01013983 ld s3,16(sp)
9c: 00813a03 ld s4,8(sp)
a0: 03010113 addi sp,sp,48
a4: 00078513 addi a0,a5,0
a8: 00008067 jalr zero,0(ra)
With RVC:
0: 02000813 addi a6,zero,32
4: 7179 c.addi16sp sp,-48
6: f422 c.sdsp s0,40(sp)
8: f026 c.sdsp s1,32(sp)
a: ec4a c.sdsp s2,24(sp)
c: e84e c.sdsp s3,16(sp)
e: e452 c.sdsp s4,8(sp)
10: 1800 c.addi4spn s0,sp,48
12: 03056683 lwu a3,48(a0)
16: 1682 c.slli a3,0x20
18: 9281 c.srli a3,0x20
1a: 03456703 lwu a4,52(a0)
1e: 1702 c.slli a4,0x20
20: 9301 c.srli a4,0x20
22: 03856483 lwu s1,56(a0)
26: 1482 c.slli s1,0x20
28: 9081 c.srli s1,0x20
2a: 03c56903 lwu s2,60(a0)
2e: 1902 c.slli s2,0x20
30: 02095913 srli s2,s2,0x20
34: 04056983 lwu s3,64(a0)
38: 1982 c.slli s3,0x20
3a: 0209d993 srli s3,s3,0x20
3e: 09056a03 lwu s4,144(a0)
42: 1a02 c.slli s4,0x20
44: 020a5a13 srli s4,s4,0x20
48: 4325 c.li t1,9
4a: 006a7363 bgeu s4,t1,0x50
4e: 4a01 c.li s4,0
50: d914 c.sw a3,48(a0)
52: d958 c.sw a4,52(a0)
54: dd04 c.sw s1,56(a0)
56: 03252e23 sw s2,60(a0)
5a: 05352023 sw s3,64(a0)
5e: 4781 c.li a5,0
60: 7422 c.ldsp s0,40(sp)
62: 7482 c.ldsp s1,32(sp)
64: 6962 c.ldsp s2,24(sp)
66: 69c2 c.ldsp s3,16(sp)
68: 6a22 c.ldsp s4,8(sp)
6a: 6145 c.addi16sp sp,48
6c: 853e c.mv a0,a5
6e: 8082 c.jr ra
Signed-off-by: Luke Nelson <luke.r.nels@gmail.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Cc: Björn Töpel <bjorn.topel@gmail.com>
Link: https://lore.kernel.org/bpf/20200721025241.8077-4-luke.r.nels@gmail.com
2020-07-20 19:52:40 -07:00
|
|
|
emit_ld(RV_REG_S2, store_offset, RV_REG_SP, ctx);
|
bpf, riscv: add BPF JIT for RV64G
This commit adds a BPF JIT for RV64G.
The JIT is a two-pass JIT, and has a dynamic prolog/epilogue (similar
to the MIPS64 BPF JIT) instead of static ones (e.g. x86_64).
At the moment the RISC-V Linux port does not support
CONFIG_HAVE_KPROBES, which means that CONFIG_BPF_EVENTS is not
supported. Thus, no tests involving BPF_PROG_TYPE_TRACEPOINT,
BPF_PROG_TYPE_PERF_EVENT, BPF_PROG_TYPE_KPROBE and
BPF_PROG_TYPE_RAW_TRACEPOINT passes.
The implementation does not support "far branching" (>4KiB).
Test results:
# modprobe test_bpf
test_bpf: Summary: 378 PASSED, 0 FAILED, [366/366 JIT'ed]
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier
...
Summary: 761 PASSED, 507 SKIPPED, 2 FAILED
Note that "test_verifier" was run with one build with
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y and one without, otherwise
many of the the tests that require unaligned access were skipped.
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
0
No CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
59
The two failing test_verifier tests are:
"ld_abs: vlan + abs, test 1"
"ld_abs: jump around ld_abs"
This is due to that "far branching" involved in those tests.
All tests where done on QEMU (QEMU emulator version 3.1.50
(v3.1.0-688-g8ae951fbc106)).
Signed-off-by: Björn Töpel <bjorn.topel@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2019-02-05 05:41:22 -07:00
|
|
|
store_offset -= 8;
|
|
|
|
}
|
|
|
|
if (seen_reg(RV_REG_S3, ctx)) {
|
bpf, riscv: Use compressed instructions in the rv64 JIT
This patch uses the RVC support and encodings from bpf_jit.h to optimize
the rv64 jit.
The optimizations work by replacing emit(rv_X(...)) with a call to a
helper function emit_X, which will emit a compressed version of the
instruction when possible, and when RVC is enabled.
The JIT continues to pass all tests in lib/test_bpf.c, and introduces
no new failures to test_verifier; both with and without RVC being enabled.
Most changes are straightforward replacements of emit(rv_X(...), ctx)
with emit_X(..., ctx), with the following exceptions bearing mention;
* Change emit_imm to sign-extend the value in "lower", since the
checks for RVC (and the instructions themselves) treat the value as
signed. Otherwise, small negative immediates will not be recognized as
encodable using an RVC instruction. For example, without this change,
emit_imm(rd, -1, ctx) would cause lower to become 4095, which is not a
6b int even though a "c.li rd, -1" instruction suffices.
* For {BPF_MOV,BPF_ADD} BPF_X, drop using addiw,addw in the 32-bit
cases since the values are zero-extended into the upper 32 bits in
the following instructions anyways, and the addition commutes with
zero-extension. (BPF_SUB BPF_X must still use subw since subtraction
does not commute with zero-extension.)
This patch avoids optimizing branches and jumps to use RVC instructions
since surrounding code often makes assumptions about the sizes of
emitted instructions. Optimizing these will require changing these
functions (e.g., emit_branch) to dynamically compute jump offsets.
The following are examples of the JITed code for the verifier selftest
"direct packet read test#3 for CGROUP_SKB OK", without and with RVC
enabled, respectively. The former uses 178 bytes, and the latter uses 112,
for a ~37% reduction in code size for this example.
Without RVC:
0: 02000813 addi a6,zero,32
4: fd010113 addi sp,sp,-48
8: 02813423 sd s0,40(sp)
c: 02913023 sd s1,32(sp)
10: 01213c23 sd s2,24(sp)
14: 01313823 sd s3,16(sp)
18: 01413423 sd s4,8(sp)
1c: 03010413 addi s0,sp,48
20: 03056683 lwu a3,48(a0)
24: 02069693 slli a3,a3,0x20
28: 0206d693 srli a3,a3,0x20
2c: 03456703 lwu a4,52(a0)
30: 02071713 slli a4,a4,0x20
34: 02075713 srli a4,a4,0x20
38: 03856483 lwu s1,56(a0)
3c: 02049493 slli s1,s1,0x20
40: 0204d493 srli s1,s1,0x20
44: 03c56903 lwu s2,60(a0)
48: 02091913 slli s2,s2,0x20
4c: 02095913 srli s2,s2,0x20
50: 04056983 lwu s3,64(a0)
54: 02099993 slli s3,s3,0x20
58: 0209d993 srli s3,s3,0x20
5c: 09056a03 lwu s4,144(a0)
60: 020a1a13 slli s4,s4,0x20
64: 020a5a13 srli s4,s4,0x20
68: 00900313 addi t1,zero,9
6c: 006a7463 bgeu s4,t1,0x74
70: 00000a13 addi s4,zero,0
74: 02d52823 sw a3,48(a0)
78: 02e52a23 sw a4,52(a0)
7c: 02952c23 sw s1,56(a0)
80: 03252e23 sw s2,60(a0)
84: 05352023 sw s3,64(a0)
88: 00000793 addi a5,zero,0
8c: 02813403 ld s0,40(sp)
90: 02013483 ld s1,32(sp)
94: 01813903 ld s2,24(sp)
98: 01013983 ld s3,16(sp)
9c: 00813a03 ld s4,8(sp)
a0: 03010113 addi sp,sp,48
a4: 00078513 addi a0,a5,0
a8: 00008067 jalr zero,0(ra)
With RVC:
0: 02000813 addi a6,zero,32
4: 7179 c.addi16sp sp,-48
6: f422 c.sdsp s0,40(sp)
8: f026 c.sdsp s1,32(sp)
a: ec4a c.sdsp s2,24(sp)
c: e84e c.sdsp s3,16(sp)
e: e452 c.sdsp s4,8(sp)
10: 1800 c.addi4spn s0,sp,48
12: 03056683 lwu a3,48(a0)
16: 1682 c.slli a3,0x20
18: 9281 c.srli a3,0x20
1a: 03456703 lwu a4,52(a0)
1e: 1702 c.slli a4,0x20
20: 9301 c.srli a4,0x20
22: 03856483 lwu s1,56(a0)
26: 1482 c.slli s1,0x20
28: 9081 c.srli s1,0x20
2a: 03c56903 lwu s2,60(a0)
2e: 1902 c.slli s2,0x20
30: 02095913 srli s2,s2,0x20
34: 04056983 lwu s3,64(a0)
38: 1982 c.slli s3,0x20
3a: 0209d993 srli s3,s3,0x20
3e: 09056a03 lwu s4,144(a0)
42: 1a02 c.slli s4,0x20
44: 020a5a13 srli s4,s4,0x20
48: 4325 c.li t1,9
4a: 006a7363 bgeu s4,t1,0x50
4e: 4a01 c.li s4,0
50: d914 c.sw a3,48(a0)
52: d958 c.sw a4,52(a0)
54: dd04 c.sw s1,56(a0)
56: 03252e23 sw s2,60(a0)
5a: 05352023 sw s3,64(a0)
5e: 4781 c.li a5,0
60: 7422 c.ldsp s0,40(sp)
62: 7482 c.ldsp s1,32(sp)
64: 6962 c.ldsp s2,24(sp)
66: 69c2 c.ldsp s3,16(sp)
68: 6a22 c.ldsp s4,8(sp)
6a: 6145 c.addi16sp sp,48
6c: 853e c.mv a0,a5
6e: 8082 c.jr ra
Signed-off-by: Luke Nelson <luke.r.nels@gmail.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Cc: Björn Töpel <bjorn.topel@gmail.com>
Link: https://lore.kernel.org/bpf/20200721025241.8077-4-luke.r.nels@gmail.com
2020-07-20 19:52:40 -07:00
|
|
|
emit_ld(RV_REG_S3, store_offset, RV_REG_SP, ctx);
|
bpf, riscv: add BPF JIT for RV64G
This commit adds a BPF JIT for RV64G.
The JIT is a two-pass JIT, and has a dynamic prolog/epilogue (similar
to the MIPS64 BPF JIT) instead of static ones (e.g. x86_64).
At the moment the RISC-V Linux port does not support
CONFIG_HAVE_KPROBES, which means that CONFIG_BPF_EVENTS is not
supported. Thus, no tests involving BPF_PROG_TYPE_TRACEPOINT,
BPF_PROG_TYPE_PERF_EVENT, BPF_PROG_TYPE_KPROBE and
BPF_PROG_TYPE_RAW_TRACEPOINT passes.
The implementation does not support "far branching" (>4KiB).
Test results:
# modprobe test_bpf
test_bpf: Summary: 378 PASSED, 0 FAILED, [366/366 JIT'ed]
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier
...
Summary: 761 PASSED, 507 SKIPPED, 2 FAILED
Note that "test_verifier" was run with one build with
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y and one without, otherwise
many of the the tests that require unaligned access were skipped.
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
0
No CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
59
The two failing test_verifier tests are:
"ld_abs: vlan + abs, test 1"
"ld_abs: jump around ld_abs"
This is due to that "far branching" involved in those tests.
All tests where done on QEMU (QEMU emulator version 3.1.50
(v3.1.0-688-g8ae951fbc106)).
Signed-off-by: Björn Töpel <bjorn.topel@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2019-02-05 05:41:22 -07:00
|
|
|
store_offset -= 8;
|
|
|
|
}
|
|
|
|
if (seen_reg(RV_REG_S4, ctx)) {
|
bpf, riscv: Use compressed instructions in the rv64 JIT
This patch uses the RVC support and encodings from bpf_jit.h to optimize
the rv64 jit.
The optimizations work by replacing emit(rv_X(...)) with a call to a
helper function emit_X, which will emit a compressed version of the
instruction when possible, and when RVC is enabled.
The JIT continues to pass all tests in lib/test_bpf.c, and introduces
no new failures to test_verifier; both with and without RVC being enabled.
Most changes are straightforward replacements of emit(rv_X(...), ctx)
with emit_X(..., ctx), with the following exceptions bearing mention;
* Change emit_imm to sign-extend the value in "lower", since the
checks for RVC (and the instructions themselves) treat the value as
signed. Otherwise, small negative immediates will not be recognized as
encodable using an RVC instruction. For example, without this change,
emit_imm(rd, -1, ctx) would cause lower to become 4095, which is not a
6b int even though a "c.li rd, -1" instruction suffices.
* For {BPF_MOV,BPF_ADD} BPF_X, drop using addiw,addw in the 32-bit
cases since the values are zero-extended into the upper 32 bits in
the following instructions anyways, and the addition commutes with
zero-extension. (BPF_SUB BPF_X must still use subw since subtraction
does not commute with zero-extension.)
This patch avoids optimizing branches and jumps to use RVC instructions
since surrounding code often makes assumptions about the sizes of
emitted instructions. Optimizing these will require changing these
functions (e.g., emit_branch) to dynamically compute jump offsets.
The following are examples of the JITed code for the verifier selftest
"direct packet read test#3 for CGROUP_SKB OK", without and with RVC
enabled, respectively. The former uses 178 bytes, and the latter uses 112,
for a ~37% reduction in code size for this example.
Without RVC:
0: 02000813 addi a6,zero,32
4: fd010113 addi sp,sp,-48
8: 02813423 sd s0,40(sp)
c: 02913023 sd s1,32(sp)
10: 01213c23 sd s2,24(sp)
14: 01313823 sd s3,16(sp)
18: 01413423 sd s4,8(sp)
1c: 03010413 addi s0,sp,48
20: 03056683 lwu a3,48(a0)
24: 02069693 slli a3,a3,0x20
28: 0206d693 srli a3,a3,0x20
2c: 03456703 lwu a4,52(a0)
30: 02071713 slli a4,a4,0x20
34: 02075713 srli a4,a4,0x20
38: 03856483 lwu s1,56(a0)
3c: 02049493 slli s1,s1,0x20
40: 0204d493 srli s1,s1,0x20
44: 03c56903 lwu s2,60(a0)
48: 02091913 slli s2,s2,0x20
4c: 02095913 srli s2,s2,0x20
50: 04056983 lwu s3,64(a0)
54: 02099993 slli s3,s3,0x20
58: 0209d993 srli s3,s3,0x20
5c: 09056a03 lwu s4,144(a0)
60: 020a1a13 slli s4,s4,0x20
64: 020a5a13 srli s4,s4,0x20
68: 00900313 addi t1,zero,9
6c: 006a7463 bgeu s4,t1,0x74
70: 00000a13 addi s4,zero,0
74: 02d52823 sw a3,48(a0)
78: 02e52a23 sw a4,52(a0)
7c: 02952c23 sw s1,56(a0)
80: 03252e23 sw s2,60(a0)
84: 05352023 sw s3,64(a0)
88: 00000793 addi a5,zero,0
8c: 02813403 ld s0,40(sp)
90: 02013483 ld s1,32(sp)
94: 01813903 ld s2,24(sp)
98: 01013983 ld s3,16(sp)
9c: 00813a03 ld s4,8(sp)
a0: 03010113 addi sp,sp,48
a4: 00078513 addi a0,a5,0
a8: 00008067 jalr zero,0(ra)
With RVC:
0: 02000813 addi a6,zero,32
4: 7179 c.addi16sp sp,-48
6: f422 c.sdsp s0,40(sp)
8: f026 c.sdsp s1,32(sp)
a: ec4a c.sdsp s2,24(sp)
c: e84e c.sdsp s3,16(sp)
e: e452 c.sdsp s4,8(sp)
10: 1800 c.addi4spn s0,sp,48
12: 03056683 lwu a3,48(a0)
16: 1682 c.slli a3,0x20
18: 9281 c.srli a3,0x20
1a: 03456703 lwu a4,52(a0)
1e: 1702 c.slli a4,0x20
20: 9301 c.srli a4,0x20
22: 03856483 lwu s1,56(a0)
26: 1482 c.slli s1,0x20
28: 9081 c.srli s1,0x20
2a: 03c56903 lwu s2,60(a0)
2e: 1902 c.slli s2,0x20
30: 02095913 srli s2,s2,0x20
34: 04056983 lwu s3,64(a0)
38: 1982 c.slli s3,0x20
3a: 0209d993 srli s3,s3,0x20
3e: 09056a03 lwu s4,144(a0)
42: 1a02 c.slli s4,0x20
44: 020a5a13 srli s4,s4,0x20
48: 4325 c.li t1,9
4a: 006a7363 bgeu s4,t1,0x50
4e: 4a01 c.li s4,0
50: d914 c.sw a3,48(a0)
52: d958 c.sw a4,52(a0)
54: dd04 c.sw s1,56(a0)
56: 03252e23 sw s2,60(a0)
5a: 05352023 sw s3,64(a0)
5e: 4781 c.li a5,0
60: 7422 c.ldsp s0,40(sp)
62: 7482 c.ldsp s1,32(sp)
64: 6962 c.ldsp s2,24(sp)
66: 69c2 c.ldsp s3,16(sp)
68: 6a22 c.ldsp s4,8(sp)
6a: 6145 c.addi16sp sp,48
6c: 853e c.mv a0,a5
6e: 8082 c.jr ra
Signed-off-by: Luke Nelson <luke.r.nels@gmail.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Cc: Björn Töpel <bjorn.topel@gmail.com>
Link: https://lore.kernel.org/bpf/20200721025241.8077-4-luke.r.nels@gmail.com
2020-07-20 19:52:40 -07:00
|
|
|
emit_ld(RV_REG_S4, store_offset, RV_REG_SP, ctx);
|
bpf, riscv: add BPF JIT for RV64G
This commit adds a BPF JIT for RV64G.
The JIT is a two-pass JIT, and has a dynamic prolog/epilogue (similar
to the MIPS64 BPF JIT) instead of static ones (e.g. x86_64).
At the moment the RISC-V Linux port does not support
CONFIG_HAVE_KPROBES, which means that CONFIG_BPF_EVENTS is not
supported. Thus, no tests involving BPF_PROG_TYPE_TRACEPOINT,
BPF_PROG_TYPE_PERF_EVENT, BPF_PROG_TYPE_KPROBE and
BPF_PROG_TYPE_RAW_TRACEPOINT passes.
The implementation does not support "far branching" (>4KiB).
Test results:
# modprobe test_bpf
test_bpf: Summary: 378 PASSED, 0 FAILED, [366/366 JIT'ed]
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier
...
Summary: 761 PASSED, 507 SKIPPED, 2 FAILED
Note that "test_verifier" was run with one build with
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y and one without, otherwise
many of the the tests that require unaligned access were skipped.
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
0
No CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
59
The two failing test_verifier tests are:
"ld_abs: vlan + abs, test 1"
"ld_abs: jump around ld_abs"
This is due to that "far branching" involved in those tests.
All tests where done on QEMU (QEMU emulator version 3.1.50
(v3.1.0-688-g8ae951fbc106)).
Signed-off-by: Björn Töpel <bjorn.topel@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2019-02-05 05:41:22 -07:00
|
|
|
store_offset -= 8;
|
|
|
|
}
|
|
|
|
if (seen_reg(RV_REG_S5, ctx)) {
|
bpf, riscv: Use compressed instructions in the rv64 JIT
This patch uses the RVC support and encodings from bpf_jit.h to optimize
the rv64 jit.
The optimizations work by replacing emit(rv_X(...)) with a call to a
helper function emit_X, which will emit a compressed version of the
instruction when possible, and when RVC is enabled.
The JIT continues to pass all tests in lib/test_bpf.c, and introduces
no new failures to test_verifier; both with and without RVC being enabled.
Most changes are straightforward replacements of emit(rv_X(...), ctx)
with emit_X(..., ctx), with the following exceptions bearing mention;
* Change emit_imm to sign-extend the value in "lower", since the
checks for RVC (and the instructions themselves) treat the value as
signed. Otherwise, small negative immediates will not be recognized as
encodable using an RVC instruction. For example, without this change,
emit_imm(rd, -1, ctx) would cause lower to become 4095, which is not a
6b int even though a "c.li rd, -1" instruction suffices.
* For {BPF_MOV,BPF_ADD} BPF_X, drop using addiw,addw in the 32-bit
cases since the values are zero-extended into the upper 32 bits in
the following instructions anyways, and the addition commutes with
zero-extension. (BPF_SUB BPF_X must still use subw since subtraction
does not commute with zero-extension.)
This patch avoids optimizing branches and jumps to use RVC instructions
since surrounding code often makes assumptions about the sizes of
emitted instructions. Optimizing these will require changing these
functions (e.g., emit_branch) to dynamically compute jump offsets.
The following are examples of the JITed code for the verifier selftest
"direct packet read test#3 for CGROUP_SKB OK", without and with RVC
enabled, respectively. The former uses 178 bytes, and the latter uses 112,
for a ~37% reduction in code size for this example.
Without RVC:
0: 02000813 addi a6,zero,32
4: fd010113 addi sp,sp,-48
8: 02813423 sd s0,40(sp)
c: 02913023 sd s1,32(sp)
10: 01213c23 sd s2,24(sp)
14: 01313823 sd s3,16(sp)
18: 01413423 sd s4,8(sp)
1c: 03010413 addi s0,sp,48
20: 03056683 lwu a3,48(a0)
24: 02069693 slli a3,a3,0x20
28: 0206d693 srli a3,a3,0x20
2c: 03456703 lwu a4,52(a0)
30: 02071713 slli a4,a4,0x20
34: 02075713 srli a4,a4,0x20
38: 03856483 lwu s1,56(a0)
3c: 02049493 slli s1,s1,0x20
40: 0204d493 srli s1,s1,0x20
44: 03c56903 lwu s2,60(a0)
48: 02091913 slli s2,s2,0x20
4c: 02095913 srli s2,s2,0x20
50: 04056983 lwu s3,64(a0)
54: 02099993 slli s3,s3,0x20
58: 0209d993 srli s3,s3,0x20
5c: 09056a03 lwu s4,144(a0)
60: 020a1a13 slli s4,s4,0x20
64: 020a5a13 srli s4,s4,0x20
68: 00900313 addi t1,zero,9
6c: 006a7463 bgeu s4,t1,0x74
70: 00000a13 addi s4,zero,0
74: 02d52823 sw a3,48(a0)
78: 02e52a23 sw a4,52(a0)
7c: 02952c23 sw s1,56(a0)
80: 03252e23 sw s2,60(a0)
84: 05352023 sw s3,64(a0)
88: 00000793 addi a5,zero,0
8c: 02813403 ld s0,40(sp)
90: 02013483 ld s1,32(sp)
94: 01813903 ld s2,24(sp)
98: 01013983 ld s3,16(sp)
9c: 00813a03 ld s4,8(sp)
a0: 03010113 addi sp,sp,48
a4: 00078513 addi a0,a5,0
a8: 00008067 jalr zero,0(ra)
With RVC:
0: 02000813 addi a6,zero,32
4: 7179 c.addi16sp sp,-48
6: f422 c.sdsp s0,40(sp)
8: f026 c.sdsp s1,32(sp)
a: ec4a c.sdsp s2,24(sp)
c: e84e c.sdsp s3,16(sp)
e: e452 c.sdsp s4,8(sp)
10: 1800 c.addi4spn s0,sp,48
12: 03056683 lwu a3,48(a0)
16: 1682 c.slli a3,0x20
18: 9281 c.srli a3,0x20
1a: 03456703 lwu a4,52(a0)
1e: 1702 c.slli a4,0x20
20: 9301 c.srli a4,0x20
22: 03856483 lwu s1,56(a0)
26: 1482 c.slli s1,0x20
28: 9081 c.srli s1,0x20
2a: 03c56903 lwu s2,60(a0)
2e: 1902 c.slli s2,0x20
30: 02095913 srli s2,s2,0x20
34: 04056983 lwu s3,64(a0)
38: 1982 c.slli s3,0x20
3a: 0209d993 srli s3,s3,0x20
3e: 09056a03 lwu s4,144(a0)
42: 1a02 c.slli s4,0x20
44: 020a5a13 srli s4,s4,0x20
48: 4325 c.li t1,9
4a: 006a7363 bgeu s4,t1,0x50
4e: 4a01 c.li s4,0
50: d914 c.sw a3,48(a0)
52: d958 c.sw a4,52(a0)
54: dd04 c.sw s1,56(a0)
56: 03252e23 sw s2,60(a0)
5a: 05352023 sw s3,64(a0)
5e: 4781 c.li a5,0
60: 7422 c.ldsp s0,40(sp)
62: 7482 c.ldsp s1,32(sp)
64: 6962 c.ldsp s2,24(sp)
66: 69c2 c.ldsp s3,16(sp)
68: 6a22 c.ldsp s4,8(sp)
6a: 6145 c.addi16sp sp,48
6c: 853e c.mv a0,a5
6e: 8082 c.jr ra
Signed-off-by: Luke Nelson <luke.r.nels@gmail.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Cc: Björn Töpel <bjorn.topel@gmail.com>
Link: https://lore.kernel.org/bpf/20200721025241.8077-4-luke.r.nels@gmail.com
2020-07-20 19:52:40 -07:00
|
|
|
emit_ld(RV_REG_S5, store_offset, RV_REG_SP, ctx);
|
bpf, riscv: add BPF JIT for RV64G
This commit adds a BPF JIT for RV64G.
The JIT is a two-pass JIT, and has a dynamic prolog/epilogue (similar
to the MIPS64 BPF JIT) instead of static ones (e.g. x86_64).
At the moment the RISC-V Linux port does not support
CONFIG_HAVE_KPROBES, which means that CONFIG_BPF_EVENTS is not
supported. Thus, no tests involving BPF_PROG_TYPE_TRACEPOINT,
BPF_PROG_TYPE_PERF_EVENT, BPF_PROG_TYPE_KPROBE and
BPF_PROG_TYPE_RAW_TRACEPOINT passes.
The implementation does not support "far branching" (>4KiB).
Test results:
# modprobe test_bpf
test_bpf: Summary: 378 PASSED, 0 FAILED, [366/366 JIT'ed]
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier
...
Summary: 761 PASSED, 507 SKIPPED, 2 FAILED
Note that "test_verifier" was run with one build with
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y and one without, otherwise
many of the the tests that require unaligned access were skipped.
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
0
No CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
59
The two failing test_verifier tests are:
"ld_abs: vlan + abs, test 1"
"ld_abs: jump around ld_abs"
This is due to that "far branching" involved in those tests.
All tests where done on QEMU (QEMU emulator version 3.1.50
(v3.1.0-688-g8ae951fbc106)).
Signed-off-by: Björn Töpel <bjorn.topel@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2019-02-05 05:41:22 -07:00
|
|
|
store_offset -= 8;
|
|
|
|
}
|
|
|
|
if (seen_reg(RV_REG_S6, ctx)) {
|
bpf, riscv: Use compressed instructions in the rv64 JIT
This patch uses the RVC support and encodings from bpf_jit.h to optimize
the rv64 jit.
The optimizations work by replacing emit(rv_X(...)) with a call to a
helper function emit_X, which will emit a compressed version of the
instruction when possible, and when RVC is enabled.
The JIT continues to pass all tests in lib/test_bpf.c, and introduces
no new failures to test_verifier; both with and without RVC being enabled.
Most changes are straightforward replacements of emit(rv_X(...), ctx)
with emit_X(..., ctx), with the following exceptions bearing mention;
* Change emit_imm to sign-extend the value in "lower", since the
checks for RVC (and the instructions themselves) treat the value as
signed. Otherwise, small negative immediates will not be recognized as
encodable using an RVC instruction. For example, without this change,
emit_imm(rd, -1, ctx) would cause lower to become 4095, which is not a
6b int even though a "c.li rd, -1" instruction suffices.
* For {BPF_MOV,BPF_ADD} BPF_X, drop using addiw,addw in the 32-bit
cases since the values are zero-extended into the upper 32 bits in
the following instructions anyways, and the addition commutes with
zero-extension. (BPF_SUB BPF_X must still use subw since subtraction
does not commute with zero-extension.)
This patch avoids optimizing branches and jumps to use RVC instructions
since surrounding code often makes assumptions about the sizes of
emitted instructions. Optimizing these will require changing these
functions (e.g., emit_branch) to dynamically compute jump offsets.
The following are examples of the JITed code for the verifier selftest
"direct packet read test#3 for CGROUP_SKB OK", without and with RVC
enabled, respectively. The former uses 178 bytes, and the latter uses 112,
for a ~37% reduction in code size for this example.
Without RVC:
0: 02000813 addi a6,zero,32
4: fd010113 addi sp,sp,-48
8: 02813423 sd s0,40(sp)
c: 02913023 sd s1,32(sp)
10: 01213c23 sd s2,24(sp)
14: 01313823 sd s3,16(sp)
18: 01413423 sd s4,8(sp)
1c: 03010413 addi s0,sp,48
20: 03056683 lwu a3,48(a0)
24: 02069693 slli a3,a3,0x20
28: 0206d693 srli a3,a3,0x20
2c: 03456703 lwu a4,52(a0)
30: 02071713 slli a4,a4,0x20
34: 02075713 srli a4,a4,0x20
38: 03856483 lwu s1,56(a0)
3c: 02049493 slli s1,s1,0x20
40: 0204d493 srli s1,s1,0x20
44: 03c56903 lwu s2,60(a0)
48: 02091913 slli s2,s2,0x20
4c: 02095913 srli s2,s2,0x20
50: 04056983 lwu s3,64(a0)
54: 02099993 slli s3,s3,0x20
58: 0209d993 srli s3,s3,0x20
5c: 09056a03 lwu s4,144(a0)
60: 020a1a13 slli s4,s4,0x20
64: 020a5a13 srli s4,s4,0x20
68: 00900313 addi t1,zero,9
6c: 006a7463 bgeu s4,t1,0x74
70: 00000a13 addi s4,zero,0
74: 02d52823 sw a3,48(a0)
78: 02e52a23 sw a4,52(a0)
7c: 02952c23 sw s1,56(a0)
80: 03252e23 sw s2,60(a0)
84: 05352023 sw s3,64(a0)
88: 00000793 addi a5,zero,0
8c: 02813403 ld s0,40(sp)
90: 02013483 ld s1,32(sp)
94: 01813903 ld s2,24(sp)
98: 01013983 ld s3,16(sp)
9c: 00813a03 ld s4,8(sp)
a0: 03010113 addi sp,sp,48
a4: 00078513 addi a0,a5,0
a8: 00008067 jalr zero,0(ra)
With RVC:
0: 02000813 addi a6,zero,32
4: 7179 c.addi16sp sp,-48
6: f422 c.sdsp s0,40(sp)
8: f026 c.sdsp s1,32(sp)
a: ec4a c.sdsp s2,24(sp)
c: e84e c.sdsp s3,16(sp)
e: e452 c.sdsp s4,8(sp)
10: 1800 c.addi4spn s0,sp,48
12: 03056683 lwu a3,48(a0)
16: 1682 c.slli a3,0x20
18: 9281 c.srli a3,0x20
1a: 03456703 lwu a4,52(a0)
1e: 1702 c.slli a4,0x20
20: 9301 c.srli a4,0x20
22: 03856483 lwu s1,56(a0)
26: 1482 c.slli s1,0x20
28: 9081 c.srli s1,0x20
2a: 03c56903 lwu s2,60(a0)
2e: 1902 c.slli s2,0x20
30: 02095913 srli s2,s2,0x20
34: 04056983 lwu s3,64(a0)
38: 1982 c.slli s3,0x20
3a: 0209d993 srli s3,s3,0x20
3e: 09056a03 lwu s4,144(a0)
42: 1a02 c.slli s4,0x20
44: 020a5a13 srli s4,s4,0x20
48: 4325 c.li t1,9
4a: 006a7363 bgeu s4,t1,0x50
4e: 4a01 c.li s4,0
50: d914 c.sw a3,48(a0)
52: d958 c.sw a4,52(a0)
54: dd04 c.sw s1,56(a0)
56: 03252e23 sw s2,60(a0)
5a: 05352023 sw s3,64(a0)
5e: 4781 c.li a5,0
60: 7422 c.ldsp s0,40(sp)
62: 7482 c.ldsp s1,32(sp)
64: 6962 c.ldsp s2,24(sp)
66: 69c2 c.ldsp s3,16(sp)
68: 6a22 c.ldsp s4,8(sp)
6a: 6145 c.addi16sp sp,48
6c: 853e c.mv a0,a5
6e: 8082 c.jr ra
Signed-off-by: Luke Nelson <luke.r.nels@gmail.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Cc: Björn Töpel <bjorn.topel@gmail.com>
Link: https://lore.kernel.org/bpf/20200721025241.8077-4-luke.r.nels@gmail.com
2020-07-20 19:52:40 -07:00
|
|
|
emit_ld(RV_REG_S6, store_offset, RV_REG_SP, ctx);
|
bpf, riscv: add BPF JIT for RV64G
This commit adds a BPF JIT for RV64G.
The JIT is a two-pass JIT, and has a dynamic prolog/epilogue (similar
to the MIPS64 BPF JIT) instead of static ones (e.g. x86_64).
At the moment the RISC-V Linux port does not support
CONFIG_HAVE_KPROBES, which means that CONFIG_BPF_EVENTS is not
supported. Thus, no tests involving BPF_PROG_TYPE_TRACEPOINT,
BPF_PROG_TYPE_PERF_EVENT, BPF_PROG_TYPE_KPROBE and
BPF_PROG_TYPE_RAW_TRACEPOINT passes.
The implementation does not support "far branching" (>4KiB).
Test results:
# modprobe test_bpf
test_bpf: Summary: 378 PASSED, 0 FAILED, [366/366 JIT'ed]
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier
...
Summary: 761 PASSED, 507 SKIPPED, 2 FAILED
Note that "test_verifier" was run with one build with
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y and one without, otherwise
many of the the tests that require unaligned access were skipped.
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
0
No CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
59
The two failing test_verifier tests are:
"ld_abs: vlan + abs, test 1"
"ld_abs: jump around ld_abs"
This is due to that "far branching" involved in those tests.
All tests where done on QEMU (QEMU emulator version 3.1.50
(v3.1.0-688-g8ae951fbc106)).
Signed-off-by: Björn Töpel <bjorn.topel@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2019-02-05 05:41:22 -07:00
|
|
|
store_offset -= 8;
|
|
|
|
}
|
2024-04-04 04:42:02 -07:00
|
|
|
if (ctx->arena_vm_start) {
|
|
|
|
emit_ld(RV_REG_ARENA, store_offset, RV_REG_SP, ctx);
|
|
|
|
store_offset -= 8;
|
|
|
|
}
|
bpf, riscv: add BPF JIT for RV64G
This commit adds a BPF JIT for RV64G.
The JIT is a two-pass JIT, and has a dynamic prolog/epilogue (similar
to the MIPS64 BPF JIT) instead of static ones (e.g. x86_64).
At the moment the RISC-V Linux port does not support
CONFIG_HAVE_KPROBES, which means that CONFIG_BPF_EVENTS is not
supported. Thus, no tests involving BPF_PROG_TYPE_TRACEPOINT,
BPF_PROG_TYPE_PERF_EVENT, BPF_PROG_TYPE_KPROBE and
BPF_PROG_TYPE_RAW_TRACEPOINT passes.
The implementation does not support "far branching" (>4KiB).
Test results:
# modprobe test_bpf
test_bpf: Summary: 378 PASSED, 0 FAILED, [366/366 JIT'ed]
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier
...
Summary: 761 PASSED, 507 SKIPPED, 2 FAILED
Note that "test_verifier" was run with one build with
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y and one without, otherwise
many of the the tests that require unaligned access were skipped.
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
0
No CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
59
The two failing test_verifier tests are:
"ld_abs: vlan + abs, test 1"
"ld_abs: jump around ld_abs"
This is due to that "far branching" involved in those tests.
All tests where done on QEMU (QEMU emulator version 3.1.50
(v3.1.0-688-g8ae951fbc106)).
Signed-off-by: Björn Töpel <bjorn.topel@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2019-02-05 05:41:22 -07:00
|
|
|
|
bpf, riscv: Use compressed instructions in the rv64 JIT
This patch uses the RVC support and encodings from bpf_jit.h to optimize
the rv64 jit.
The optimizations work by replacing emit(rv_X(...)) with a call to a
helper function emit_X, which will emit a compressed version of the
instruction when possible, and when RVC is enabled.
The JIT continues to pass all tests in lib/test_bpf.c, and introduces
no new failures to test_verifier; both with and without RVC being enabled.
Most changes are straightforward replacements of emit(rv_X(...), ctx)
with emit_X(..., ctx), with the following exceptions bearing mention;
* Change emit_imm to sign-extend the value in "lower", since the
checks for RVC (and the instructions themselves) treat the value as
signed. Otherwise, small negative immediates will not be recognized as
encodable using an RVC instruction. For example, without this change,
emit_imm(rd, -1, ctx) would cause lower to become 4095, which is not a
6b int even though a "c.li rd, -1" instruction suffices.
* For {BPF_MOV,BPF_ADD} BPF_X, drop using addiw,addw in the 32-bit
cases since the values are zero-extended into the upper 32 bits in
the following instructions anyways, and the addition commutes with
zero-extension. (BPF_SUB BPF_X must still use subw since subtraction
does not commute with zero-extension.)
This patch avoids optimizing branches and jumps to use RVC instructions
since surrounding code often makes assumptions about the sizes of
emitted instructions. Optimizing these will require changing these
functions (e.g., emit_branch) to dynamically compute jump offsets.
The following are examples of the JITed code for the verifier selftest
"direct packet read test#3 for CGROUP_SKB OK", without and with RVC
enabled, respectively. The former uses 178 bytes, and the latter uses 112,
for a ~37% reduction in code size for this example.
Without RVC:
0: 02000813 addi a6,zero,32
4: fd010113 addi sp,sp,-48
8: 02813423 sd s0,40(sp)
c: 02913023 sd s1,32(sp)
10: 01213c23 sd s2,24(sp)
14: 01313823 sd s3,16(sp)
18: 01413423 sd s4,8(sp)
1c: 03010413 addi s0,sp,48
20: 03056683 lwu a3,48(a0)
24: 02069693 slli a3,a3,0x20
28: 0206d693 srli a3,a3,0x20
2c: 03456703 lwu a4,52(a0)
30: 02071713 slli a4,a4,0x20
34: 02075713 srli a4,a4,0x20
38: 03856483 lwu s1,56(a0)
3c: 02049493 slli s1,s1,0x20
40: 0204d493 srli s1,s1,0x20
44: 03c56903 lwu s2,60(a0)
48: 02091913 slli s2,s2,0x20
4c: 02095913 srli s2,s2,0x20
50: 04056983 lwu s3,64(a0)
54: 02099993 slli s3,s3,0x20
58: 0209d993 srli s3,s3,0x20
5c: 09056a03 lwu s4,144(a0)
60: 020a1a13 slli s4,s4,0x20
64: 020a5a13 srli s4,s4,0x20
68: 00900313 addi t1,zero,9
6c: 006a7463 bgeu s4,t1,0x74
70: 00000a13 addi s4,zero,0
74: 02d52823 sw a3,48(a0)
78: 02e52a23 sw a4,52(a0)
7c: 02952c23 sw s1,56(a0)
80: 03252e23 sw s2,60(a0)
84: 05352023 sw s3,64(a0)
88: 00000793 addi a5,zero,0
8c: 02813403 ld s0,40(sp)
90: 02013483 ld s1,32(sp)
94: 01813903 ld s2,24(sp)
98: 01013983 ld s3,16(sp)
9c: 00813a03 ld s4,8(sp)
a0: 03010113 addi sp,sp,48
a4: 00078513 addi a0,a5,0
a8: 00008067 jalr zero,0(ra)
With RVC:
0: 02000813 addi a6,zero,32
4: 7179 c.addi16sp sp,-48
6: f422 c.sdsp s0,40(sp)
8: f026 c.sdsp s1,32(sp)
a: ec4a c.sdsp s2,24(sp)
c: e84e c.sdsp s3,16(sp)
e: e452 c.sdsp s4,8(sp)
10: 1800 c.addi4spn s0,sp,48
12: 03056683 lwu a3,48(a0)
16: 1682 c.slli a3,0x20
18: 9281 c.srli a3,0x20
1a: 03456703 lwu a4,52(a0)
1e: 1702 c.slli a4,0x20
20: 9301 c.srli a4,0x20
22: 03856483 lwu s1,56(a0)
26: 1482 c.slli s1,0x20
28: 9081 c.srli s1,0x20
2a: 03c56903 lwu s2,60(a0)
2e: 1902 c.slli s2,0x20
30: 02095913 srli s2,s2,0x20
34: 04056983 lwu s3,64(a0)
38: 1982 c.slli s3,0x20
3a: 0209d993 srli s3,s3,0x20
3e: 09056a03 lwu s4,144(a0)
42: 1a02 c.slli s4,0x20
44: 020a5a13 srli s4,s4,0x20
48: 4325 c.li t1,9
4a: 006a7363 bgeu s4,t1,0x50
4e: 4a01 c.li s4,0
50: d914 c.sw a3,48(a0)
52: d958 c.sw a4,52(a0)
54: dd04 c.sw s1,56(a0)
56: 03252e23 sw s2,60(a0)
5a: 05352023 sw s3,64(a0)
5e: 4781 c.li a5,0
60: 7422 c.ldsp s0,40(sp)
62: 7482 c.ldsp s1,32(sp)
64: 6962 c.ldsp s2,24(sp)
66: 69c2 c.ldsp s3,16(sp)
68: 6a22 c.ldsp s4,8(sp)
6a: 6145 c.addi16sp sp,48
6c: 853e c.mv a0,a5
6e: 8082 c.jr ra
Signed-off-by: Luke Nelson <luke.r.nels@gmail.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Cc: Björn Töpel <bjorn.topel@gmail.com>
Link: https://lore.kernel.org/bpf/20200721025241.8077-4-luke.r.nels@gmail.com
2020-07-20 19:52:40 -07:00
|
|
|
emit_addi(RV_REG_SP, RV_REG_SP, stack_adjust, ctx);
|
bpf, riscv: add BPF JIT for RV64G
This commit adds a BPF JIT for RV64G.
The JIT is a two-pass JIT, and has a dynamic prolog/epilogue (similar
to the MIPS64 BPF JIT) instead of static ones (e.g. x86_64).
At the moment the RISC-V Linux port does not support
CONFIG_HAVE_KPROBES, which means that CONFIG_BPF_EVENTS is not
supported. Thus, no tests involving BPF_PROG_TYPE_TRACEPOINT,
BPF_PROG_TYPE_PERF_EVENT, BPF_PROG_TYPE_KPROBE and
BPF_PROG_TYPE_RAW_TRACEPOINT passes.
The implementation does not support "far branching" (>4KiB).
Test results:
# modprobe test_bpf
test_bpf: Summary: 378 PASSED, 0 FAILED, [366/366 JIT'ed]
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier
...
Summary: 761 PASSED, 507 SKIPPED, 2 FAILED
Note that "test_verifier" was run with one build with
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y and one without, otherwise
many of the the tests that require unaligned access were skipped.
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
0
No CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
59
The two failing test_verifier tests are:
"ld_abs: vlan + abs, test 1"
"ld_abs: jump around ld_abs"
This is due to that "far branching" involved in those tests.
All tests where done on QEMU (QEMU emulator version 3.1.50
(v3.1.0-688-g8ae951fbc106)).
Signed-off-by: Björn Töpel <bjorn.topel@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2019-02-05 05:41:22 -07:00
|
|
|
/* Set return value. */
|
2019-12-16 02:13:39 -07:00
|
|
|
if (!is_tail_call)
|
riscv, bpf: Sign-extend return values
The RISC-V architecture does not expose sub-registers, and hold all
32-bit values in a sign-extended format [1] [2]:
| The compiler and calling convention maintain an invariant that all
| 32-bit values are held in a sign-extended format in 64-bit
| registers. Even 32-bit unsigned integers extend bit 31 into bits
| 63 through 32. Consequently, conversion between unsigned and
| signed 32-bit integers is a no-op, as is conversion from a signed
| 32-bit integer to a signed 64-bit integer.
While BPF, on the other hand, exposes sub-registers, and use
zero-extension (similar to arm64/x86).
This has led to some subtle bugs, where a BPF JITted program has not
sign-extended the a0 register (return value in RISC-V land), passed
the return value up the kernel, e.g.:
| int from_bpf(void);
|
| long foo(void)
| {
| return from_bpf();
| }
Here, a0 would be 0xffff_ffff, instead of the expected
0xffff_ffff_ffff_ffff.
Internally, the RISC-V JIT uses a5 as a dedicated register for BPF
return values.
Keep a5 zero-extended, but explicitly sign-extend a0 (which is used
outside BPF land). Now that a0 (RISC-V ABI) and a5 (BPF ABI) differs,
a0 is only moved to a5 for non-BPF native calls (BPF_PSEUDO_CALL).
Fixes: 2353ecc6f91f ("bpf, riscv: add BPF JIT for RV64G")
Signed-off-by: Björn Töpel <bjorn@rivosinc.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Link: https://github.com/riscv/riscv-isa-manual/releases/download/riscv-isa-release-056b6ff-2023-10-02/unpriv-isa-asciidoc.pdf # [2]
Link: https://github.com/riscv-non-isa/riscv-elf-psabi-doc/releases/download/draft-20230929-e5c800e661a53efe3c2678d71a306323b60eb13b/riscv-abi.pdf # [2]
Link: https://lore.kernel.org/bpf/20231004120706.52848-2-bjorn@kernel.org
2023-10-04 05:07:05 -07:00
|
|
|
emit_addiw(RV_REG_A0, RV_REG_A5, 0, ctx);
|
bpf, riscv: Use compressed instructions in the rv64 JIT
This patch uses the RVC support and encodings from bpf_jit.h to optimize
the rv64 jit.
The optimizations work by replacing emit(rv_X(...)) with a call to a
helper function emit_X, which will emit a compressed version of the
instruction when possible, and when RVC is enabled.
The JIT continues to pass all tests in lib/test_bpf.c, and introduces
no new failures to test_verifier; both with and without RVC being enabled.
Most changes are straightforward replacements of emit(rv_X(...), ctx)
with emit_X(..., ctx), with the following exceptions bearing mention;
* Change emit_imm to sign-extend the value in "lower", since the
checks for RVC (and the instructions themselves) treat the value as
signed. Otherwise, small negative immediates will not be recognized as
encodable using an RVC instruction. For example, without this change,
emit_imm(rd, -1, ctx) would cause lower to become 4095, which is not a
6b int even though a "c.li rd, -1" instruction suffices.
* For {BPF_MOV,BPF_ADD} BPF_X, drop using addiw,addw in the 32-bit
cases since the values are zero-extended into the upper 32 bits in
the following instructions anyways, and the addition commutes with
zero-extension. (BPF_SUB BPF_X must still use subw since subtraction
does not commute with zero-extension.)
This patch avoids optimizing branches and jumps to use RVC instructions
since surrounding code often makes assumptions about the sizes of
emitted instructions. Optimizing these will require changing these
functions (e.g., emit_branch) to dynamically compute jump offsets.
The following are examples of the JITed code for the verifier selftest
"direct packet read test#3 for CGROUP_SKB OK", without and with RVC
enabled, respectively. The former uses 178 bytes, and the latter uses 112,
for a ~37% reduction in code size for this example.
Without RVC:
0: 02000813 addi a6,zero,32
4: fd010113 addi sp,sp,-48
8: 02813423 sd s0,40(sp)
c: 02913023 sd s1,32(sp)
10: 01213c23 sd s2,24(sp)
14: 01313823 sd s3,16(sp)
18: 01413423 sd s4,8(sp)
1c: 03010413 addi s0,sp,48
20: 03056683 lwu a3,48(a0)
24: 02069693 slli a3,a3,0x20
28: 0206d693 srli a3,a3,0x20
2c: 03456703 lwu a4,52(a0)
30: 02071713 slli a4,a4,0x20
34: 02075713 srli a4,a4,0x20
38: 03856483 lwu s1,56(a0)
3c: 02049493 slli s1,s1,0x20
40: 0204d493 srli s1,s1,0x20
44: 03c56903 lwu s2,60(a0)
48: 02091913 slli s2,s2,0x20
4c: 02095913 srli s2,s2,0x20
50: 04056983 lwu s3,64(a0)
54: 02099993 slli s3,s3,0x20
58: 0209d993 srli s3,s3,0x20
5c: 09056a03 lwu s4,144(a0)
60: 020a1a13 slli s4,s4,0x20
64: 020a5a13 srli s4,s4,0x20
68: 00900313 addi t1,zero,9
6c: 006a7463 bgeu s4,t1,0x74
70: 00000a13 addi s4,zero,0
74: 02d52823 sw a3,48(a0)
78: 02e52a23 sw a4,52(a0)
7c: 02952c23 sw s1,56(a0)
80: 03252e23 sw s2,60(a0)
84: 05352023 sw s3,64(a0)
88: 00000793 addi a5,zero,0
8c: 02813403 ld s0,40(sp)
90: 02013483 ld s1,32(sp)
94: 01813903 ld s2,24(sp)
98: 01013983 ld s3,16(sp)
9c: 00813a03 ld s4,8(sp)
a0: 03010113 addi sp,sp,48
a4: 00078513 addi a0,a5,0
a8: 00008067 jalr zero,0(ra)
With RVC:
0: 02000813 addi a6,zero,32
4: 7179 c.addi16sp sp,-48
6: f422 c.sdsp s0,40(sp)
8: f026 c.sdsp s1,32(sp)
a: ec4a c.sdsp s2,24(sp)
c: e84e c.sdsp s3,16(sp)
e: e452 c.sdsp s4,8(sp)
10: 1800 c.addi4spn s0,sp,48
12: 03056683 lwu a3,48(a0)
16: 1682 c.slli a3,0x20
18: 9281 c.srli a3,0x20
1a: 03456703 lwu a4,52(a0)
1e: 1702 c.slli a4,0x20
20: 9301 c.srli a4,0x20
22: 03856483 lwu s1,56(a0)
26: 1482 c.slli s1,0x20
28: 9081 c.srli s1,0x20
2a: 03c56903 lwu s2,60(a0)
2e: 1902 c.slli s2,0x20
30: 02095913 srli s2,s2,0x20
34: 04056983 lwu s3,64(a0)
38: 1982 c.slli s3,0x20
3a: 0209d993 srli s3,s3,0x20
3e: 09056a03 lwu s4,144(a0)
42: 1a02 c.slli s4,0x20
44: 020a5a13 srli s4,s4,0x20
48: 4325 c.li t1,9
4a: 006a7363 bgeu s4,t1,0x50
4e: 4a01 c.li s4,0
50: d914 c.sw a3,48(a0)
52: d958 c.sw a4,52(a0)
54: dd04 c.sw s1,56(a0)
56: 03252e23 sw s2,60(a0)
5a: 05352023 sw s3,64(a0)
5e: 4781 c.li a5,0
60: 7422 c.ldsp s0,40(sp)
62: 7482 c.ldsp s1,32(sp)
64: 6962 c.ldsp s2,24(sp)
66: 69c2 c.ldsp s3,16(sp)
68: 6a22 c.ldsp s4,8(sp)
6a: 6145 c.addi16sp sp,48
6c: 853e c.mv a0,a5
6e: 8082 c.jr ra
Signed-off-by: Luke Nelson <luke.r.nels@gmail.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Cc: Björn Töpel <bjorn.topel@gmail.com>
Link: https://lore.kernel.org/bpf/20200721025241.8077-4-luke.r.nels@gmail.com
2020-07-20 19:52:40 -07:00
|
|
|
emit_jalr(RV_REG_ZERO, is_tail_call ? RV_REG_T3 : RV_REG_RA,
|
2023-07-21 03:06:27 -07:00
|
|
|
is_tail_call ? (RV_FENTRY_NINSNS + 1) * 4 : 0, /* skip reserved nops and TCC init */
|
bpf, riscv: Use compressed instructions in the rv64 JIT
This patch uses the RVC support and encodings from bpf_jit.h to optimize
the rv64 jit.
The optimizations work by replacing emit(rv_X(...)) with a call to a
helper function emit_X, which will emit a compressed version of the
instruction when possible, and when RVC is enabled.
The JIT continues to pass all tests in lib/test_bpf.c, and introduces
no new failures to test_verifier; both with and without RVC being enabled.
Most changes are straightforward replacements of emit(rv_X(...), ctx)
with emit_X(..., ctx), with the following exceptions bearing mention;
* Change emit_imm to sign-extend the value in "lower", since the
checks for RVC (and the instructions themselves) treat the value as
signed. Otherwise, small negative immediates will not be recognized as
encodable using an RVC instruction. For example, without this change,
emit_imm(rd, -1, ctx) would cause lower to become 4095, which is not a
6b int even though a "c.li rd, -1" instruction suffices.
* For {BPF_MOV,BPF_ADD} BPF_X, drop using addiw,addw in the 32-bit
cases since the values are zero-extended into the upper 32 bits in
the following instructions anyways, and the addition commutes with
zero-extension. (BPF_SUB BPF_X must still use subw since subtraction
does not commute with zero-extension.)
This patch avoids optimizing branches and jumps to use RVC instructions
since surrounding code often makes assumptions about the sizes of
emitted instructions. Optimizing these will require changing these
functions (e.g., emit_branch) to dynamically compute jump offsets.
The following are examples of the JITed code for the verifier selftest
"direct packet read test#3 for CGROUP_SKB OK", without and with RVC
enabled, respectively. The former uses 178 bytes, and the latter uses 112,
for a ~37% reduction in code size for this example.
Without RVC:
0: 02000813 addi a6,zero,32
4: fd010113 addi sp,sp,-48
8: 02813423 sd s0,40(sp)
c: 02913023 sd s1,32(sp)
10: 01213c23 sd s2,24(sp)
14: 01313823 sd s3,16(sp)
18: 01413423 sd s4,8(sp)
1c: 03010413 addi s0,sp,48
20: 03056683 lwu a3,48(a0)
24: 02069693 slli a3,a3,0x20
28: 0206d693 srli a3,a3,0x20
2c: 03456703 lwu a4,52(a0)
30: 02071713 slli a4,a4,0x20
34: 02075713 srli a4,a4,0x20
38: 03856483 lwu s1,56(a0)
3c: 02049493 slli s1,s1,0x20
40: 0204d493 srli s1,s1,0x20
44: 03c56903 lwu s2,60(a0)
48: 02091913 slli s2,s2,0x20
4c: 02095913 srli s2,s2,0x20
50: 04056983 lwu s3,64(a0)
54: 02099993 slli s3,s3,0x20
58: 0209d993 srli s3,s3,0x20
5c: 09056a03 lwu s4,144(a0)
60: 020a1a13 slli s4,s4,0x20
64: 020a5a13 srli s4,s4,0x20
68: 00900313 addi t1,zero,9
6c: 006a7463 bgeu s4,t1,0x74
70: 00000a13 addi s4,zero,0
74: 02d52823 sw a3,48(a0)
78: 02e52a23 sw a4,52(a0)
7c: 02952c23 sw s1,56(a0)
80: 03252e23 sw s2,60(a0)
84: 05352023 sw s3,64(a0)
88: 00000793 addi a5,zero,0
8c: 02813403 ld s0,40(sp)
90: 02013483 ld s1,32(sp)
94: 01813903 ld s2,24(sp)
98: 01013983 ld s3,16(sp)
9c: 00813a03 ld s4,8(sp)
a0: 03010113 addi sp,sp,48
a4: 00078513 addi a0,a5,0
a8: 00008067 jalr zero,0(ra)
With RVC:
0: 02000813 addi a6,zero,32
4: 7179 c.addi16sp sp,-48
6: f422 c.sdsp s0,40(sp)
8: f026 c.sdsp s1,32(sp)
a: ec4a c.sdsp s2,24(sp)
c: e84e c.sdsp s3,16(sp)
e: e452 c.sdsp s4,8(sp)
10: 1800 c.addi4spn s0,sp,48
12: 03056683 lwu a3,48(a0)
16: 1682 c.slli a3,0x20
18: 9281 c.srli a3,0x20
1a: 03456703 lwu a4,52(a0)
1e: 1702 c.slli a4,0x20
20: 9301 c.srli a4,0x20
22: 03856483 lwu s1,56(a0)
26: 1482 c.slli s1,0x20
28: 9081 c.srli s1,0x20
2a: 03c56903 lwu s2,60(a0)
2e: 1902 c.slli s2,0x20
30: 02095913 srli s2,s2,0x20
34: 04056983 lwu s3,64(a0)
38: 1982 c.slli s3,0x20
3a: 0209d993 srli s3,s3,0x20
3e: 09056a03 lwu s4,144(a0)
42: 1a02 c.slli s4,0x20
44: 020a5a13 srli s4,s4,0x20
48: 4325 c.li t1,9
4a: 006a7363 bgeu s4,t1,0x50
4e: 4a01 c.li s4,0
50: d914 c.sw a3,48(a0)
52: d958 c.sw a4,52(a0)
54: dd04 c.sw s1,56(a0)
56: 03252e23 sw s2,60(a0)
5a: 05352023 sw s3,64(a0)
5e: 4781 c.li a5,0
60: 7422 c.ldsp s0,40(sp)
62: 7482 c.ldsp s1,32(sp)
64: 6962 c.ldsp s2,24(sp)
66: 69c2 c.ldsp s3,16(sp)
68: 6a22 c.ldsp s4,8(sp)
6a: 6145 c.addi16sp sp,48
6c: 853e c.mv a0,a5
6e: 8082 c.jr ra
Signed-off-by: Luke Nelson <luke.r.nels@gmail.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Cc: Björn Töpel <bjorn.topel@gmail.com>
Link: https://lore.kernel.org/bpf/20200721025241.8077-4-luke.r.nels@gmail.com
2020-07-20 19:52:40 -07:00
|
|
|
ctx);
|
bpf, riscv: add BPF JIT for RV64G
This commit adds a BPF JIT for RV64G.
The JIT is a two-pass JIT, and has a dynamic prolog/epilogue (similar
to the MIPS64 BPF JIT) instead of static ones (e.g. x86_64).
At the moment the RISC-V Linux port does not support
CONFIG_HAVE_KPROBES, which means that CONFIG_BPF_EVENTS is not
supported. Thus, no tests involving BPF_PROG_TYPE_TRACEPOINT,
BPF_PROG_TYPE_PERF_EVENT, BPF_PROG_TYPE_KPROBE and
BPF_PROG_TYPE_RAW_TRACEPOINT passes.
The implementation does not support "far branching" (>4KiB).
Test results:
# modprobe test_bpf
test_bpf: Summary: 378 PASSED, 0 FAILED, [366/366 JIT'ed]
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier
...
Summary: 761 PASSED, 507 SKIPPED, 2 FAILED
Note that "test_verifier" was run with one build with
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y and one without, otherwise
many of the the tests that require unaligned access were skipped.
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
0
No CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
59
The two failing test_verifier tests are:
"ld_abs: vlan + abs, test 1"
"ld_abs: jump around ld_abs"
This is due to that "far branching" involved in those tests.
All tests where done on QEMU (QEMU emulator version 3.1.50
(v3.1.0-688-g8ae951fbc106)).
Signed-off-by: Björn Töpel <bjorn.topel@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2019-02-05 05:41:22 -07:00
|
|
|
}
|
|
|
|
|
2019-12-16 02:13:36 -07:00
|
|
|
static void emit_bcc(u8 cond, u8 rd, u8 rs, int rvoff,
|
|
|
|
struct rv_jit_context *ctx)
|
|
|
|
{
|
|
|
|
switch (cond) {
|
|
|
|
case BPF_JEQ:
|
|
|
|
emit(rv_beq(rd, rs, rvoff >> 1), ctx);
|
|
|
|
return;
|
|
|
|
case BPF_JGT:
|
|
|
|
emit(rv_bltu(rs, rd, rvoff >> 1), ctx);
|
|
|
|
return;
|
|
|
|
case BPF_JLT:
|
|
|
|
emit(rv_bltu(rd, rs, rvoff >> 1), ctx);
|
|
|
|
return;
|
|
|
|
case BPF_JGE:
|
|
|
|
emit(rv_bgeu(rd, rs, rvoff >> 1), ctx);
|
|
|
|
return;
|
|
|
|
case BPF_JLE:
|
|
|
|
emit(rv_bgeu(rs, rd, rvoff >> 1), ctx);
|
|
|
|
return;
|
|
|
|
case BPF_JNE:
|
|
|
|
emit(rv_bne(rd, rs, rvoff >> 1), ctx);
|
|
|
|
return;
|
|
|
|
case BPF_JSGT:
|
|
|
|
emit(rv_blt(rs, rd, rvoff >> 1), ctx);
|
|
|
|
return;
|
|
|
|
case BPF_JSLT:
|
|
|
|
emit(rv_blt(rd, rs, rvoff >> 1), ctx);
|
|
|
|
return;
|
|
|
|
case BPF_JSGE:
|
|
|
|
emit(rv_bge(rd, rs, rvoff >> 1), ctx);
|
|
|
|
return;
|
|
|
|
case BPF_JSLE:
|
|
|
|
emit(rv_bge(rs, rd, rvoff >> 1), ctx);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
static void emit_branch(u8 cond, u8 rd, u8 rs, int rvoff,
|
|
|
|
struct rv_jit_context *ctx)
|
|
|
|
{
|
|
|
|
s64 upper, lower;
|
|
|
|
|
|
|
|
if (is_13b_int(rvoff)) {
|
|
|
|
emit_bcc(cond, rd, rs, rvoff, ctx);
|
|
|
|
return;
|
|
|
|
}
|
|
|
|
|
|
|
|
/* Adjust for jal */
|
|
|
|
rvoff -= 4;
|
|
|
|
|
|
|
|
/* Transform, e.g.:
|
|
|
|
* bne rd,rs,foo
|
|
|
|
* to
|
|
|
|
* beq rd,rs,<.L1>
|
|
|
|
* (auipc foo)
|
|
|
|
* jal(r) foo
|
|
|
|
* .L1
|
|
|
|
*/
|
|
|
|
cond = invert_bpf_cond(cond);
|
|
|
|
if (is_21b_int(rvoff)) {
|
|
|
|
emit_bcc(cond, rd, rs, 8, ctx);
|
|
|
|
emit(rv_jal(RV_REG_ZERO, rvoff >> 1), ctx);
|
|
|
|
return;
|
|
|
|
}
|
|
|
|
|
|
|
|
/* 32b No need for an additional rvoff adjustment, since we
|
|
|
|
* get that from the auipc at PC', where PC = PC' + 4.
|
|
|
|
*/
|
|
|
|
upper = (rvoff + (1 << 11)) >> 12;
|
|
|
|
lower = rvoff & 0xfff;
|
|
|
|
|
|
|
|
emit_bcc(cond, rd, rs, 12, ctx);
|
|
|
|
emit(rv_auipc(RV_REG_T1, upper), ctx);
|
|
|
|
emit(rv_jalr(RV_REG_ZERO, RV_REG_T1, lower), ctx);
|
bpf, riscv: add BPF JIT for RV64G
This commit adds a BPF JIT for RV64G.
The JIT is a two-pass JIT, and has a dynamic prolog/epilogue (similar
to the MIPS64 BPF JIT) instead of static ones (e.g. x86_64).
At the moment the RISC-V Linux port does not support
CONFIG_HAVE_KPROBES, which means that CONFIG_BPF_EVENTS is not
supported. Thus, no tests involving BPF_PROG_TYPE_TRACEPOINT,
BPF_PROG_TYPE_PERF_EVENT, BPF_PROG_TYPE_KPROBE and
BPF_PROG_TYPE_RAW_TRACEPOINT passes.
The implementation does not support "far branching" (>4KiB).
Test results:
# modprobe test_bpf
test_bpf: Summary: 378 PASSED, 0 FAILED, [366/366 JIT'ed]
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier
...
Summary: 761 PASSED, 507 SKIPPED, 2 FAILED
Note that "test_verifier" was run with one build with
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y and one without, otherwise
many of the the tests that require unaligned access were skipped.
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
0
No CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
59
The two failing test_verifier tests are:
"ld_abs: vlan + abs, test 1"
"ld_abs: jump around ld_abs"
This is due to that "far branching" involved in those tests.
All tests where done on QEMU (QEMU emulator version 3.1.50
(v3.1.0-688-g8ae951fbc106)).
Signed-off-by: Björn Töpel <bjorn.topel@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2019-02-05 05:41:22 -07:00
|
|
|
}
|
|
|
|
|
|
|
|
static int emit_bpf_tail_call(int insn, struct rv_jit_context *ctx)
|
|
|
|
{
|
|
|
|
int tc_ninsn, off, start_insn = ctx->ninsns;
|
|
|
|
u8 tcc = rv_tail_call_reg(ctx);
|
|
|
|
|
|
|
|
/* a0: &ctx
|
|
|
|
* a1: &array
|
|
|
|
* a2: index
|
|
|
|
*
|
|
|
|
* if (index >= array->map.max_entries)
|
|
|
|
* goto out;
|
|
|
|
*/
|
|
|
|
tc_ninsn = insn ? ctx->offset[insn] - ctx->offset[insn - 1] :
|
|
|
|
ctx->offset[0];
|
2024-01-15 06:12:31 -07:00
|
|
|
emit_zextw(RV_REG_A2, RV_REG_A2, ctx);
|
bpf, riscv: add BPF JIT for RV64G
This commit adds a BPF JIT for RV64G.
The JIT is a two-pass JIT, and has a dynamic prolog/epilogue (similar
to the MIPS64 BPF JIT) instead of static ones (e.g. x86_64).
At the moment the RISC-V Linux port does not support
CONFIG_HAVE_KPROBES, which means that CONFIG_BPF_EVENTS is not
supported. Thus, no tests involving BPF_PROG_TYPE_TRACEPOINT,
BPF_PROG_TYPE_PERF_EVENT, BPF_PROG_TYPE_KPROBE and
BPF_PROG_TYPE_RAW_TRACEPOINT passes.
The implementation does not support "far branching" (>4KiB).
Test results:
# modprobe test_bpf
test_bpf: Summary: 378 PASSED, 0 FAILED, [366/366 JIT'ed]
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier
...
Summary: 761 PASSED, 507 SKIPPED, 2 FAILED
Note that "test_verifier" was run with one build with
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y and one without, otherwise
many of the the tests that require unaligned access were skipped.
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
0
No CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
59
The two failing test_verifier tests are:
"ld_abs: vlan + abs, test 1"
"ld_abs: jump around ld_abs"
This is due to that "far branching" involved in those tests.
All tests where done on QEMU (QEMU emulator version 3.1.50
(v3.1.0-688-g8ae951fbc106)).
Signed-off-by: Björn Töpel <bjorn.topel@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2019-02-05 05:41:22 -07:00
|
|
|
|
|
|
|
off = offsetof(struct bpf_array, map.max_entries);
|
|
|
|
if (is_12b_check(off, insn))
|
|
|
|
return -1;
|
|
|
|
emit(rv_lwu(RV_REG_T1, off, RV_REG_A1), ctx);
|
2020-07-20 19:52:38 -07:00
|
|
|
off = ninsns_rvoff(tc_ninsn - (ctx->ninsns - start_insn));
|
2019-12-16 02:13:37 -07:00
|
|
|
emit_branch(BPF_JGE, RV_REG_A2, RV_REG_T1, off, ctx);
|
bpf, riscv: add BPF JIT for RV64G
This commit adds a BPF JIT for RV64G.
The JIT is a two-pass JIT, and has a dynamic prolog/epilogue (similar
to the MIPS64 BPF JIT) instead of static ones (e.g. x86_64).
At the moment the RISC-V Linux port does not support
CONFIG_HAVE_KPROBES, which means that CONFIG_BPF_EVENTS is not
supported. Thus, no tests involving BPF_PROG_TYPE_TRACEPOINT,
BPF_PROG_TYPE_PERF_EVENT, BPF_PROG_TYPE_KPROBE and
BPF_PROG_TYPE_RAW_TRACEPOINT passes.
The implementation does not support "far branching" (>4KiB).
Test results:
# modprobe test_bpf
test_bpf: Summary: 378 PASSED, 0 FAILED, [366/366 JIT'ed]
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier
...
Summary: 761 PASSED, 507 SKIPPED, 2 FAILED
Note that "test_verifier" was run with one build with
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y and one without, otherwise
many of the the tests that require unaligned access were skipped.
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
0
No CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
59
The two failing test_verifier tests are:
"ld_abs: vlan + abs, test 1"
"ld_abs: jump around ld_abs"
This is due to that "far branching" involved in those tests.
All tests where done on QEMU (QEMU emulator version 3.1.50
(v3.1.0-688-g8ae951fbc106)).
Signed-off-by: Björn Töpel <bjorn.topel@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2019-02-05 05:41:22 -07:00
|
|
|
|
bpf: Change value of MAX_TAIL_CALL_CNT from 32 to 33
In the current code, the actual max tail call count is 33 which is greater
than MAX_TAIL_CALL_CNT (defined as 32). The actual limit is not consistent
with the meaning of MAX_TAIL_CALL_CNT and thus confusing at first glance.
We can see the historical evolution from commit 04fd61ab36ec ("bpf: allow
bpf programs to tail-call other bpf programs") and commit f9dabe016b63
("bpf: Undo off-by-one in interpreter tail call count limit"). In order
to avoid changing existing behavior, the actual limit is 33 now, this is
reasonable.
After commit 874be05f525e ("bpf, tests: Add tail call test suite"), we can
see there exists failed testcase.
On all archs when CONFIG_BPF_JIT_ALWAYS_ON is not set:
# echo 0 > /proc/sys/net/core/bpf_jit_enable
# modprobe test_bpf
# dmesg | grep -w FAIL
Tail call error path, max count reached jited:0 ret 34 != 33 FAIL
On some archs:
# echo 1 > /proc/sys/net/core/bpf_jit_enable
# modprobe test_bpf
# dmesg | grep -w FAIL
Tail call error path, max count reached jited:1 ret 34 != 33 FAIL
Although the above failed testcase has been fixed in commit 18935a72eb25
("bpf/tests: Fix error in tail call limit tests"), it would still be good
to change the value of MAX_TAIL_CALL_CNT from 32 to 33 to make the code
more readable.
The 32-bit x86 JIT was using a limit of 32, just fix the wrong comments and
limit to 33 tail calls as the constant MAX_TAIL_CALL_CNT updated. For the
mips64 JIT, use "ori" instead of "addiu" as suggested by Johan Almbladh.
For the riscv JIT, use RV_REG_TCC directly to save one register move as
suggested by Björn Töpel. For the other implementations, no function changes,
it does not change the current limit 33, the new value of MAX_TAIL_CALL_CNT
can reflect the actual max tail call count, the related tail call testcases
in test_bpf module and selftests can work well for the interpreter and the
JIT.
Here are the test results on x86_64:
# uname -m
x86_64
# echo 0 > /proc/sys/net/core/bpf_jit_enable
# modprobe test_bpf test_suite=test_tail_calls
# dmesg | tail -1
test_bpf: test_tail_calls: Summary: 8 PASSED, 0 FAILED, [0/8 JIT'ed]
# rmmod test_bpf
# echo 1 > /proc/sys/net/core/bpf_jit_enable
# modprobe test_bpf test_suite=test_tail_calls
# dmesg | tail -1
test_bpf: test_tail_calls: Summary: 8 PASSED, 0 FAILED, [8/8 JIT'ed]
# rmmod test_bpf
# ./test_progs -t tailcalls
#142 tailcalls:OK
Summary: 1/11 PASSED, 0 SKIPPED, 0 FAILED
Signed-off-by: Tiezhu Yang <yangtiezhu@loongson.cn>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Tested-by: Johan Almbladh <johan.almbladh@anyfinetworks.com>
Tested-by: Ilya Leoshkevich <iii@linux.ibm.com>
Acked-by: Björn Töpel <bjorn@kernel.org>
Acked-by: Johan Almbladh <johan.almbladh@anyfinetworks.com>
Acked-by: Ilya Leoshkevich <iii@linux.ibm.com>
Link: https://lore.kernel.org/bpf/1636075800-3264-1-git-send-email-yangtiezhu@loongson.cn
2021-11-04 18:30:00 -07:00
|
|
|
/* if (--TCC < 0)
|
bpf, riscv: add BPF JIT for RV64G
This commit adds a BPF JIT for RV64G.
The JIT is a two-pass JIT, and has a dynamic prolog/epilogue (similar
to the MIPS64 BPF JIT) instead of static ones (e.g. x86_64).
At the moment the RISC-V Linux port does not support
CONFIG_HAVE_KPROBES, which means that CONFIG_BPF_EVENTS is not
supported. Thus, no tests involving BPF_PROG_TYPE_TRACEPOINT,
BPF_PROG_TYPE_PERF_EVENT, BPF_PROG_TYPE_KPROBE and
BPF_PROG_TYPE_RAW_TRACEPOINT passes.
The implementation does not support "far branching" (>4KiB).
Test results:
# modprobe test_bpf
test_bpf: Summary: 378 PASSED, 0 FAILED, [366/366 JIT'ed]
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier
...
Summary: 761 PASSED, 507 SKIPPED, 2 FAILED
Note that "test_verifier" was run with one build with
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y and one without, otherwise
many of the the tests that require unaligned access were skipped.
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
0
No CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
59
The two failing test_verifier tests are:
"ld_abs: vlan + abs, test 1"
"ld_abs: jump around ld_abs"
This is due to that "far branching" involved in those tests.
All tests where done on QEMU (QEMU emulator version 3.1.50
(v3.1.0-688-g8ae951fbc106)).
Signed-off-by: Björn Töpel <bjorn.topel@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2019-02-05 05:41:22 -07:00
|
|
|
* goto out;
|
|
|
|
*/
|
bpf: Change value of MAX_TAIL_CALL_CNT from 32 to 33
In the current code, the actual max tail call count is 33 which is greater
than MAX_TAIL_CALL_CNT (defined as 32). The actual limit is not consistent
with the meaning of MAX_TAIL_CALL_CNT and thus confusing at first glance.
We can see the historical evolution from commit 04fd61ab36ec ("bpf: allow
bpf programs to tail-call other bpf programs") and commit f9dabe016b63
("bpf: Undo off-by-one in interpreter tail call count limit"). In order
to avoid changing existing behavior, the actual limit is 33 now, this is
reasonable.
After commit 874be05f525e ("bpf, tests: Add tail call test suite"), we can
see there exists failed testcase.
On all archs when CONFIG_BPF_JIT_ALWAYS_ON is not set:
# echo 0 > /proc/sys/net/core/bpf_jit_enable
# modprobe test_bpf
# dmesg | grep -w FAIL
Tail call error path, max count reached jited:0 ret 34 != 33 FAIL
On some archs:
# echo 1 > /proc/sys/net/core/bpf_jit_enable
# modprobe test_bpf
# dmesg | grep -w FAIL
Tail call error path, max count reached jited:1 ret 34 != 33 FAIL
Although the above failed testcase has been fixed in commit 18935a72eb25
("bpf/tests: Fix error in tail call limit tests"), it would still be good
to change the value of MAX_TAIL_CALL_CNT from 32 to 33 to make the code
more readable.
The 32-bit x86 JIT was using a limit of 32, just fix the wrong comments and
limit to 33 tail calls as the constant MAX_TAIL_CALL_CNT updated. For the
mips64 JIT, use "ori" instead of "addiu" as suggested by Johan Almbladh.
For the riscv JIT, use RV_REG_TCC directly to save one register move as
suggested by Björn Töpel. For the other implementations, no function changes,
it does not change the current limit 33, the new value of MAX_TAIL_CALL_CNT
can reflect the actual max tail call count, the related tail call testcases
in test_bpf module and selftests can work well for the interpreter and the
JIT.
Here are the test results on x86_64:
# uname -m
x86_64
# echo 0 > /proc/sys/net/core/bpf_jit_enable
# modprobe test_bpf test_suite=test_tail_calls
# dmesg | tail -1
test_bpf: test_tail_calls: Summary: 8 PASSED, 0 FAILED, [0/8 JIT'ed]
# rmmod test_bpf
# echo 1 > /proc/sys/net/core/bpf_jit_enable
# modprobe test_bpf test_suite=test_tail_calls
# dmesg | tail -1
test_bpf: test_tail_calls: Summary: 8 PASSED, 0 FAILED, [8/8 JIT'ed]
# rmmod test_bpf
# ./test_progs -t tailcalls
#142 tailcalls:OK
Summary: 1/11 PASSED, 0 SKIPPED, 0 FAILED
Signed-off-by: Tiezhu Yang <yangtiezhu@loongson.cn>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Tested-by: Johan Almbladh <johan.almbladh@anyfinetworks.com>
Tested-by: Ilya Leoshkevich <iii@linux.ibm.com>
Acked-by: Björn Töpel <bjorn@kernel.org>
Acked-by: Johan Almbladh <johan.almbladh@anyfinetworks.com>
Acked-by: Ilya Leoshkevich <iii@linux.ibm.com>
Link: https://lore.kernel.org/bpf/1636075800-3264-1-git-send-email-yangtiezhu@loongson.cn
2021-11-04 18:30:00 -07:00
|
|
|
emit_addi(RV_REG_TCC, tcc, -1, ctx);
|
2020-07-20 19:52:38 -07:00
|
|
|
off = ninsns_rvoff(tc_ninsn - (ctx->ninsns - start_insn));
|
bpf: Change value of MAX_TAIL_CALL_CNT from 32 to 33
In the current code, the actual max tail call count is 33 which is greater
than MAX_TAIL_CALL_CNT (defined as 32). The actual limit is not consistent
with the meaning of MAX_TAIL_CALL_CNT and thus confusing at first glance.
We can see the historical evolution from commit 04fd61ab36ec ("bpf: allow
bpf programs to tail-call other bpf programs") and commit f9dabe016b63
("bpf: Undo off-by-one in interpreter tail call count limit"). In order
to avoid changing existing behavior, the actual limit is 33 now, this is
reasonable.
After commit 874be05f525e ("bpf, tests: Add tail call test suite"), we can
see there exists failed testcase.
On all archs when CONFIG_BPF_JIT_ALWAYS_ON is not set:
# echo 0 > /proc/sys/net/core/bpf_jit_enable
# modprobe test_bpf
# dmesg | grep -w FAIL
Tail call error path, max count reached jited:0 ret 34 != 33 FAIL
On some archs:
# echo 1 > /proc/sys/net/core/bpf_jit_enable
# modprobe test_bpf
# dmesg | grep -w FAIL
Tail call error path, max count reached jited:1 ret 34 != 33 FAIL
Although the above failed testcase has been fixed in commit 18935a72eb25
("bpf/tests: Fix error in tail call limit tests"), it would still be good
to change the value of MAX_TAIL_CALL_CNT from 32 to 33 to make the code
more readable.
The 32-bit x86 JIT was using a limit of 32, just fix the wrong comments and
limit to 33 tail calls as the constant MAX_TAIL_CALL_CNT updated. For the
mips64 JIT, use "ori" instead of "addiu" as suggested by Johan Almbladh.
For the riscv JIT, use RV_REG_TCC directly to save one register move as
suggested by Björn Töpel. For the other implementations, no function changes,
it does not change the current limit 33, the new value of MAX_TAIL_CALL_CNT
can reflect the actual max tail call count, the related tail call testcases
in test_bpf module and selftests can work well for the interpreter and the
JIT.
Here are the test results on x86_64:
# uname -m
x86_64
# echo 0 > /proc/sys/net/core/bpf_jit_enable
# modprobe test_bpf test_suite=test_tail_calls
# dmesg | tail -1
test_bpf: test_tail_calls: Summary: 8 PASSED, 0 FAILED, [0/8 JIT'ed]
# rmmod test_bpf
# echo 1 > /proc/sys/net/core/bpf_jit_enable
# modprobe test_bpf test_suite=test_tail_calls
# dmesg | tail -1
test_bpf: test_tail_calls: Summary: 8 PASSED, 0 FAILED, [8/8 JIT'ed]
# rmmod test_bpf
# ./test_progs -t tailcalls
#142 tailcalls:OK
Summary: 1/11 PASSED, 0 SKIPPED, 0 FAILED
Signed-off-by: Tiezhu Yang <yangtiezhu@loongson.cn>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Tested-by: Johan Almbladh <johan.almbladh@anyfinetworks.com>
Tested-by: Ilya Leoshkevich <iii@linux.ibm.com>
Acked-by: Björn Töpel <bjorn@kernel.org>
Acked-by: Johan Almbladh <johan.almbladh@anyfinetworks.com>
Acked-by: Ilya Leoshkevich <iii@linux.ibm.com>
Link: https://lore.kernel.org/bpf/1636075800-3264-1-git-send-email-yangtiezhu@loongson.cn
2021-11-04 18:30:00 -07:00
|
|
|
emit_branch(BPF_JSLT, RV_REG_TCC, RV_REG_ZERO, off, ctx);
|
bpf, riscv: add BPF JIT for RV64G
This commit adds a BPF JIT for RV64G.
The JIT is a two-pass JIT, and has a dynamic prolog/epilogue (similar
to the MIPS64 BPF JIT) instead of static ones (e.g. x86_64).
At the moment the RISC-V Linux port does not support
CONFIG_HAVE_KPROBES, which means that CONFIG_BPF_EVENTS is not
supported. Thus, no tests involving BPF_PROG_TYPE_TRACEPOINT,
BPF_PROG_TYPE_PERF_EVENT, BPF_PROG_TYPE_KPROBE and
BPF_PROG_TYPE_RAW_TRACEPOINT passes.
The implementation does not support "far branching" (>4KiB).
Test results:
# modprobe test_bpf
test_bpf: Summary: 378 PASSED, 0 FAILED, [366/366 JIT'ed]
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier
...
Summary: 761 PASSED, 507 SKIPPED, 2 FAILED
Note that "test_verifier" was run with one build with
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y and one without, otherwise
many of the the tests that require unaligned access were skipped.
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
0
No CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
59
The two failing test_verifier tests are:
"ld_abs: vlan + abs, test 1"
"ld_abs: jump around ld_abs"
This is due to that "far branching" involved in those tests.
All tests where done on QEMU (QEMU emulator version 3.1.50
(v3.1.0-688-g8ae951fbc106)).
Signed-off-by: Björn Töpel <bjorn.topel@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2019-02-05 05:41:22 -07:00
|
|
|
|
|
|
|
/* prog = array->ptrs[index];
|
|
|
|
* if (!prog)
|
|
|
|
* goto out;
|
|
|
|
*/
|
bpf, riscv: Use compressed instructions in the rv64 JIT
This patch uses the RVC support and encodings from bpf_jit.h to optimize
the rv64 jit.
The optimizations work by replacing emit(rv_X(...)) with a call to a
helper function emit_X, which will emit a compressed version of the
instruction when possible, and when RVC is enabled.
The JIT continues to pass all tests in lib/test_bpf.c, and introduces
no new failures to test_verifier; both with and without RVC being enabled.
Most changes are straightforward replacements of emit(rv_X(...), ctx)
with emit_X(..., ctx), with the following exceptions bearing mention;
* Change emit_imm to sign-extend the value in "lower", since the
checks for RVC (and the instructions themselves) treat the value as
signed. Otherwise, small negative immediates will not be recognized as
encodable using an RVC instruction. For example, without this change,
emit_imm(rd, -1, ctx) would cause lower to become 4095, which is not a
6b int even though a "c.li rd, -1" instruction suffices.
* For {BPF_MOV,BPF_ADD} BPF_X, drop using addiw,addw in the 32-bit
cases since the values are zero-extended into the upper 32 bits in
the following instructions anyways, and the addition commutes with
zero-extension. (BPF_SUB BPF_X must still use subw since subtraction
does not commute with zero-extension.)
This patch avoids optimizing branches and jumps to use RVC instructions
since surrounding code often makes assumptions about the sizes of
emitted instructions. Optimizing these will require changing these
functions (e.g., emit_branch) to dynamically compute jump offsets.
The following are examples of the JITed code for the verifier selftest
"direct packet read test#3 for CGROUP_SKB OK", without and with RVC
enabled, respectively. The former uses 178 bytes, and the latter uses 112,
for a ~37% reduction in code size for this example.
Without RVC:
0: 02000813 addi a6,zero,32
4: fd010113 addi sp,sp,-48
8: 02813423 sd s0,40(sp)
c: 02913023 sd s1,32(sp)
10: 01213c23 sd s2,24(sp)
14: 01313823 sd s3,16(sp)
18: 01413423 sd s4,8(sp)
1c: 03010413 addi s0,sp,48
20: 03056683 lwu a3,48(a0)
24: 02069693 slli a3,a3,0x20
28: 0206d693 srli a3,a3,0x20
2c: 03456703 lwu a4,52(a0)
30: 02071713 slli a4,a4,0x20
34: 02075713 srli a4,a4,0x20
38: 03856483 lwu s1,56(a0)
3c: 02049493 slli s1,s1,0x20
40: 0204d493 srli s1,s1,0x20
44: 03c56903 lwu s2,60(a0)
48: 02091913 slli s2,s2,0x20
4c: 02095913 srli s2,s2,0x20
50: 04056983 lwu s3,64(a0)
54: 02099993 slli s3,s3,0x20
58: 0209d993 srli s3,s3,0x20
5c: 09056a03 lwu s4,144(a0)
60: 020a1a13 slli s4,s4,0x20
64: 020a5a13 srli s4,s4,0x20
68: 00900313 addi t1,zero,9
6c: 006a7463 bgeu s4,t1,0x74
70: 00000a13 addi s4,zero,0
74: 02d52823 sw a3,48(a0)
78: 02e52a23 sw a4,52(a0)
7c: 02952c23 sw s1,56(a0)
80: 03252e23 sw s2,60(a0)
84: 05352023 sw s3,64(a0)
88: 00000793 addi a5,zero,0
8c: 02813403 ld s0,40(sp)
90: 02013483 ld s1,32(sp)
94: 01813903 ld s2,24(sp)
98: 01013983 ld s3,16(sp)
9c: 00813a03 ld s4,8(sp)
a0: 03010113 addi sp,sp,48
a4: 00078513 addi a0,a5,0
a8: 00008067 jalr zero,0(ra)
With RVC:
0: 02000813 addi a6,zero,32
4: 7179 c.addi16sp sp,-48
6: f422 c.sdsp s0,40(sp)
8: f026 c.sdsp s1,32(sp)
a: ec4a c.sdsp s2,24(sp)
c: e84e c.sdsp s3,16(sp)
e: e452 c.sdsp s4,8(sp)
10: 1800 c.addi4spn s0,sp,48
12: 03056683 lwu a3,48(a0)
16: 1682 c.slli a3,0x20
18: 9281 c.srli a3,0x20
1a: 03456703 lwu a4,52(a0)
1e: 1702 c.slli a4,0x20
20: 9301 c.srli a4,0x20
22: 03856483 lwu s1,56(a0)
26: 1482 c.slli s1,0x20
28: 9081 c.srli s1,0x20
2a: 03c56903 lwu s2,60(a0)
2e: 1902 c.slli s2,0x20
30: 02095913 srli s2,s2,0x20
34: 04056983 lwu s3,64(a0)
38: 1982 c.slli s3,0x20
3a: 0209d993 srli s3,s3,0x20
3e: 09056a03 lwu s4,144(a0)
42: 1a02 c.slli s4,0x20
44: 020a5a13 srli s4,s4,0x20
48: 4325 c.li t1,9
4a: 006a7363 bgeu s4,t1,0x50
4e: 4a01 c.li s4,0
50: d914 c.sw a3,48(a0)
52: d958 c.sw a4,52(a0)
54: dd04 c.sw s1,56(a0)
56: 03252e23 sw s2,60(a0)
5a: 05352023 sw s3,64(a0)
5e: 4781 c.li a5,0
60: 7422 c.ldsp s0,40(sp)
62: 7482 c.ldsp s1,32(sp)
64: 6962 c.ldsp s2,24(sp)
66: 69c2 c.ldsp s3,16(sp)
68: 6a22 c.ldsp s4,8(sp)
6a: 6145 c.addi16sp sp,48
6c: 853e c.mv a0,a5
6e: 8082 c.jr ra
Signed-off-by: Luke Nelson <luke.r.nels@gmail.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Cc: Björn Töpel <bjorn.topel@gmail.com>
Link: https://lore.kernel.org/bpf/20200721025241.8077-4-luke.r.nels@gmail.com
2020-07-20 19:52:40 -07:00
|
|
|
emit_slli(RV_REG_T2, RV_REG_A2, 3, ctx);
|
|
|
|
emit_add(RV_REG_T2, RV_REG_T2, RV_REG_A1, ctx);
|
bpf, riscv: add BPF JIT for RV64G
This commit adds a BPF JIT for RV64G.
The JIT is a two-pass JIT, and has a dynamic prolog/epilogue (similar
to the MIPS64 BPF JIT) instead of static ones (e.g. x86_64).
At the moment the RISC-V Linux port does not support
CONFIG_HAVE_KPROBES, which means that CONFIG_BPF_EVENTS is not
supported. Thus, no tests involving BPF_PROG_TYPE_TRACEPOINT,
BPF_PROG_TYPE_PERF_EVENT, BPF_PROG_TYPE_KPROBE and
BPF_PROG_TYPE_RAW_TRACEPOINT passes.
The implementation does not support "far branching" (>4KiB).
Test results:
# modprobe test_bpf
test_bpf: Summary: 378 PASSED, 0 FAILED, [366/366 JIT'ed]
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier
...
Summary: 761 PASSED, 507 SKIPPED, 2 FAILED
Note that "test_verifier" was run with one build with
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y and one without, otherwise
many of the the tests that require unaligned access were skipped.
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
0
No CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
59
The two failing test_verifier tests are:
"ld_abs: vlan + abs, test 1"
"ld_abs: jump around ld_abs"
This is due to that "far branching" involved in those tests.
All tests where done on QEMU (QEMU emulator version 3.1.50
(v3.1.0-688-g8ae951fbc106)).
Signed-off-by: Björn Töpel <bjorn.topel@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2019-02-05 05:41:22 -07:00
|
|
|
off = offsetof(struct bpf_array, ptrs);
|
|
|
|
if (is_12b_check(off, insn))
|
|
|
|
return -1;
|
bpf, riscv: Use compressed instructions in the rv64 JIT
This patch uses the RVC support and encodings from bpf_jit.h to optimize
the rv64 jit.
The optimizations work by replacing emit(rv_X(...)) with a call to a
helper function emit_X, which will emit a compressed version of the
instruction when possible, and when RVC is enabled.
The JIT continues to pass all tests in lib/test_bpf.c, and introduces
no new failures to test_verifier; both with and without RVC being enabled.
Most changes are straightforward replacements of emit(rv_X(...), ctx)
with emit_X(..., ctx), with the following exceptions bearing mention;
* Change emit_imm to sign-extend the value in "lower", since the
checks for RVC (and the instructions themselves) treat the value as
signed. Otherwise, small negative immediates will not be recognized as
encodable using an RVC instruction. For example, without this change,
emit_imm(rd, -1, ctx) would cause lower to become 4095, which is not a
6b int even though a "c.li rd, -1" instruction suffices.
* For {BPF_MOV,BPF_ADD} BPF_X, drop using addiw,addw in the 32-bit
cases since the values are zero-extended into the upper 32 bits in
the following instructions anyways, and the addition commutes with
zero-extension. (BPF_SUB BPF_X must still use subw since subtraction
does not commute with zero-extension.)
This patch avoids optimizing branches and jumps to use RVC instructions
since surrounding code often makes assumptions about the sizes of
emitted instructions. Optimizing these will require changing these
functions (e.g., emit_branch) to dynamically compute jump offsets.
The following are examples of the JITed code for the verifier selftest
"direct packet read test#3 for CGROUP_SKB OK", without and with RVC
enabled, respectively. The former uses 178 bytes, and the latter uses 112,
for a ~37% reduction in code size for this example.
Without RVC:
0: 02000813 addi a6,zero,32
4: fd010113 addi sp,sp,-48
8: 02813423 sd s0,40(sp)
c: 02913023 sd s1,32(sp)
10: 01213c23 sd s2,24(sp)
14: 01313823 sd s3,16(sp)
18: 01413423 sd s4,8(sp)
1c: 03010413 addi s0,sp,48
20: 03056683 lwu a3,48(a0)
24: 02069693 slli a3,a3,0x20
28: 0206d693 srli a3,a3,0x20
2c: 03456703 lwu a4,52(a0)
30: 02071713 slli a4,a4,0x20
34: 02075713 srli a4,a4,0x20
38: 03856483 lwu s1,56(a0)
3c: 02049493 slli s1,s1,0x20
40: 0204d493 srli s1,s1,0x20
44: 03c56903 lwu s2,60(a0)
48: 02091913 slli s2,s2,0x20
4c: 02095913 srli s2,s2,0x20
50: 04056983 lwu s3,64(a0)
54: 02099993 slli s3,s3,0x20
58: 0209d993 srli s3,s3,0x20
5c: 09056a03 lwu s4,144(a0)
60: 020a1a13 slli s4,s4,0x20
64: 020a5a13 srli s4,s4,0x20
68: 00900313 addi t1,zero,9
6c: 006a7463 bgeu s4,t1,0x74
70: 00000a13 addi s4,zero,0
74: 02d52823 sw a3,48(a0)
78: 02e52a23 sw a4,52(a0)
7c: 02952c23 sw s1,56(a0)
80: 03252e23 sw s2,60(a0)
84: 05352023 sw s3,64(a0)
88: 00000793 addi a5,zero,0
8c: 02813403 ld s0,40(sp)
90: 02013483 ld s1,32(sp)
94: 01813903 ld s2,24(sp)
98: 01013983 ld s3,16(sp)
9c: 00813a03 ld s4,8(sp)
a0: 03010113 addi sp,sp,48
a4: 00078513 addi a0,a5,0
a8: 00008067 jalr zero,0(ra)
With RVC:
0: 02000813 addi a6,zero,32
4: 7179 c.addi16sp sp,-48
6: f422 c.sdsp s0,40(sp)
8: f026 c.sdsp s1,32(sp)
a: ec4a c.sdsp s2,24(sp)
c: e84e c.sdsp s3,16(sp)
e: e452 c.sdsp s4,8(sp)
10: 1800 c.addi4spn s0,sp,48
12: 03056683 lwu a3,48(a0)
16: 1682 c.slli a3,0x20
18: 9281 c.srli a3,0x20
1a: 03456703 lwu a4,52(a0)
1e: 1702 c.slli a4,0x20
20: 9301 c.srli a4,0x20
22: 03856483 lwu s1,56(a0)
26: 1482 c.slli s1,0x20
28: 9081 c.srli s1,0x20
2a: 03c56903 lwu s2,60(a0)
2e: 1902 c.slli s2,0x20
30: 02095913 srli s2,s2,0x20
34: 04056983 lwu s3,64(a0)
38: 1982 c.slli s3,0x20
3a: 0209d993 srli s3,s3,0x20
3e: 09056a03 lwu s4,144(a0)
42: 1a02 c.slli s4,0x20
44: 020a5a13 srli s4,s4,0x20
48: 4325 c.li t1,9
4a: 006a7363 bgeu s4,t1,0x50
4e: 4a01 c.li s4,0
50: d914 c.sw a3,48(a0)
52: d958 c.sw a4,52(a0)
54: dd04 c.sw s1,56(a0)
56: 03252e23 sw s2,60(a0)
5a: 05352023 sw s3,64(a0)
5e: 4781 c.li a5,0
60: 7422 c.ldsp s0,40(sp)
62: 7482 c.ldsp s1,32(sp)
64: 6962 c.ldsp s2,24(sp)
66: 69c2 c.ldsp s3,16(sp)
68: 6a22 c.ldsp s4,8(sp)
6a: 6145 c.addi16sp sp,48
6c: 853e c.mv a0,a5
6e: 8082 c.jr ra
Signed-off-by: Luke Nelson <luke.r.nels@gmail.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Cc: Björn Töpel <bjorn.topel@gmail.com>
Link: https://lore.kernel.org/bpf/20200721025241.8077-4-luke.r.nels@gmail.com
2020-07-20 19:52:40 -07:00
|
|
|
emit_ld(RV_REG_T2, off, RV_REG_T2, ctx);
|
2020-07-20 19:52:38 -07:00
|
|
|
off = ninsns_rvoff(tc_ninsn - (ctx->ninsns - start_insn));
|
2019-12-16 02:13:37 -07:00
|
|
|
emit_branch(BPF_JEQ, RV_REG_T2, RV_REG_ZERO, off, ctx);
|
bpf, riscv: add BPF JIT for RV64G
This commit adds a BPF JIT for RV64G.
The JIT is a two-pass JIT, and has a dynamic prolog/epilogue (similar
to the MIPS64 BPF JIT) instead of static ones (e.g. x86_64).
At the moment the RISC-V Linux port does not support
CONFIG_HAVE_KPROBES, which means that CONFIG_BPF_EVENTS is not
supported. Thus, no tests involving BPF_PROG_TYPE_TRACEPOINT,
BPF_PROG_TYPE_PERF_EVENT, BPF_PROG_TYPE_KPROBE and
BPF_PROG_TYPE_RAW_TRACEPOINT passes.
The implementation does not support "far branching" (>4KiB).
Test results:
# modprobe test_bpf
test_bpf: Summary: 378 PASSED, 0 FAILED, [366/366 JIT'ed]
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier
...
Summary: 761 PASSED, 507 SKIPPED, 2 FAILED
Note that "test_verifier" was run with one build with
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y and one without, otherwise
many of the the tests that require unaligned access were skipped.
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
0
No CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
59
The two failing test_verifier tests are:
"ld_abs: vlan + abs, test 1"
"ld_abs: jump around ld_abs"
This is due to that "far branching" involved in those tests.
All tests where done on QEMU (QEMU emulator version 3.1.50
(v3.1.0-688-g8ae951fbc106)).
Signed-off-by: Björn Töpel <bjorn.topel@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2019-02-05 05:41:22 -07:00
|
|
|
|
|
|
|
/* goto *(prog->bpf_func + 4); */
|
|
|
|
off = offsetof(struct bpf_prog, bpf_func);
|
|
|
|
if (is_12b_check(off, insn))
|
|
|
|
return -1;
|
bpf, riscv: Use compressed instructions in the rv64 JIT
This patch uses the RVC support and encodings from bpf_jit.h to optimize
the rv64 jit.
The optimizations work by replacing emit(rv_X(...)) with a call to a
helper function emit_X, which will emit a compressed version of the
instruction when possible, and when RVC is enabled.
The JIT continues to pass all tests in lib/test_bpf.c, and introduces
no new failures to test_verifier; both with and without RVC being enabled.
Most changes are straightforward replacements of emit(rv_X(...), ctx)
with emit_X(..., ctx), with the following exceptions bearing mention;
* Change emit_imm to sign-extend the value in "lower", since the
checks for RVC (and the instructions themselves) treat the value as
signed. Otherwise, small negative immediates will not be recognized as
encodable using an RVC instruction. For example, without this change,
emit_imm(rd, -1, ctx) would cause lower to become 4095, which is not a
6b int even though a "c.li rd, -1" instruction suffices.
* For {BPF_MOV,BPF_ADD} BPF_X, drop using addiw,addw in the 32-bit
cases since the values are zero-extended into the upper 32 bits in
the following instructions anyways, and the addition commutes with
zero-extension. (BPF_SUB BPF_X must still use subw since subtraction
does not commute with zero-extension.)
This patch avoids optimizing branches and jumps to use RVC instructions
since surrounding code often makes assumptions about the sizes of
emitted instructions. Optimizing these will require changing these
functions (e.g., emit_branch) to dynamically compute jump offsets.
The following are examples of the JITed code for the verifier selftest
"direct packet read test#3 for CGROUP_SKB OK", without and with RVC
enabled, respectively. The former uses 178 bytes, and the latter uses 112,
for a ~37% reduction in code size for this example.
Without RVC:
0: 02000813 addi a6,zero,32
4: fd010113 addi sp,sp,-48
8: 02813423 sd s0,40(sp)
c: 02913023 sd s1,32(sp)
10: 01213c23 sd s2,24(sp)
14: 01313823 sd s3,16(sp)
18: 01413423 sd s4,8(sp)
1c: 03010413 addi s0,sp,48
20: 03056683 lwu a3,48(a0)
24: 02069693 slli a3,a3,0x20
28: 0206d693 srli a3,a3,0x20
2c: 03456703 lwu a4,52(a0)
30: 02071713 slli a4,a4,0x20
34: 02075713 srli a4,a4,0x20
38: 03856483 lwu s1,56(a0)
3c: 02049493 slli s1,s1,0x20
40: 0204d493 srli s1,s1,0x20
44: 03c56903 lwu s2,60(a0)
48: 02091913 slli s2,s2,0x20
4c: 02095913 srli s2,s2,0x20
50: 04056983 lwu s3,64(a0)
54: 02099993 slli s3,s3,0x20
58: 0209d993 srli s3,s3,0x20
5c: 09056a03 lwu s4,144(a0)
60: 020a1a13 slli s4,s4,0x20
64: 020a5a13 srli s4,s4,0x20
68: 00900313 addi t1,zero,9
6c: 006a7463 bgeu s4,t1,0x74
70: 00000a13 addi s4,zero,0
74: 02d52823 sw a3,48(a0)
78: 02e52a23 sw a4,52(a0)
7c: 02952c23 sw s1,56(a0)
80: 03252e23 sw s2,60(a0)
84: 05352023 sw s3,64(a0)
88: 00000793 addi a5,zero,0
8c: 02813403 ld s0,40(sp)
90: 02013483 ld s1,32(sp)
94: 01813903 ld s2,24(sp)
98: 01013983 ld s3,16(sp)
9c: 00813a03 ld s4,8(sp)
a0: 03010113 addi sp,sp,48
a4: 00078513 addi a0,a5,0
a8: 00008067 jalr zero,0(ra)
With RVC:
0: 02000813 addi a6,zero,32
4: 7179 c.addi16sp sp,-48
6: f422 c.sdsp s0,40(sp)
8: f026 c.sdsp s1,32(sp)
a: ec4a c.sdsp s2,24(sp)
c: e84e c.sdsp s3,16(sp)
e: e452 c.sdsp s4,8(sp)
10: 1800 c.addi4spn s0,sp,48
12: 03056683 lwu a3,48(a0)
16: 1682 c.slli a3,0x20
18: 9281 c.srli a3,0x20
1a: 03456703 lwu a4,52(a0)
1e: 1702 c.slli a4,0x20
20: 9301 c.srli a4,0x20
22: 03856483 lwu s1,56(a0)
26: 1482 c.slli s1,0x20
28: 9081 c.srli s1,0x20
2a: 03c56903 lwu s2,60(a0)
2e: 1902 c.slli s2,0x20
30: 02095913 srli s2,s2,0x20
34: 04056983 lwu s3,64(a0)
38: 1982 c.slli s3,0x20
3a: 0209d993 srli s3,s3,0x20
3e: 09056a03 lwu s4,144(a0)
42: 1a02 c.slli s4,0x20
44: 020a5a13 srli s4,s4,0x20
48: 4325 c.li t1,9
4a: 006a7363 bgeu s4,t1,0x50
4e: 4a01 c.li s4,0
50: d914 c.sw a3,48(a0)
52: d958 c.sw a4,52(a0)
54: dd04 c.sw s1,56(a0)
56: 03252e23 sw s2,60(a0)
5a: 05352023 sw s3,64(a0)
5e: 4781 c.li a5,0
60: 7422 c.ldsp s0,40(sp)
62: 7482 c.ldsp s1,32(sp)
64: 6962 c.ldsp s2,24(sp)
66: 69c2 c.ldsp s3,16(sp)
68: 6a22 c.ldsp s4,8(sp)
6a: 6145 c.addi16sp sp,48
6c: 853e c.mv a0,a5
6e: 8082 c.jr ra
Signed-off-by: Luke Nelson <luke.r.nels@gmail.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Cc: Björn Töpel <bjorn.topel@gmail.com>
Link: https://lore.kernel.org/bpf/20200721025241.8077-4-luke.r.nels@gmail.com
2020-07-20 19:52:40 -07:00
|
|
|
emit_ld(RV_REG_T3, off, RV_REG_T2, ctx);
|
2019-12-16 02:13:39 -07:00
|
|
|
__build_epilogue(true, ctx);
|
bpf, riscv: add BPF JIT for RV64G
This commit adds a BPF JIT for RV64G.
The JIT is a two-pass JIT, and has a dynamic prolog/epilogue (similar
to the MIPS64 BPF JIT) instead of static ones (e.g. x86_64).
At the moment the RISC-V Linux port does not support
CONFIG_HAVE_KPROBES, which means that CONFIG_BPF_EVENTS is not
supported. Thus, no tests involving BPF_PROG_TYPE_TRACEPOINT,
BPF_PROG_TYPE_PERF_EVENT, BPF_PROG_TYPE_KPROBE and
BPF_PROG_TYPE_RAW_TRACEPOINT passes.
The implementation does not support "far branching" (>4KiB).
Test results:
# modprobe test_bpf
test_bpf: Summary: 378 PASSED, 0 FAILED, [366/366 JIT'ed]
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier
...
Summary: 761 PASSED, 507 SKIPPED, 2 FAILED
Note that "test_verifier" was run with one build with
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y and one without, otherwise
many of the the tests that require unaligned access were skipped.
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
0
No CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
59
The two failing test_verifier tests are:
"ld_abs: vlan + abs, test 1"
"ld_abs: jump around ld_abs"
This is due to that "far branching" involved in those tests.
All tests where done on QEMU (QEMU emulator version 3.1.50
(v3.1.0-688-g8ae951fbc106)).
Signed-off-by: Björn Töpel <bjorn.topel@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2019-02-05 05:41:22 -07:00
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
static void init_regs(u8 *rd, u8 *rs, const struct bpf_insn *insn,
|
|
|
|
struct rv_jit_context *ctx)
|
|
|
|
{
|
|
|
|
u8 code = insn->code;
|
|
|
|
|
|
|
|
switch (code) {
|
|
|
|
case BPF_JMP | BPF_JA:
|
|
|
|
case BPF_JMP | BPF_CALL:
|
|
|
|
case BPF_JMP | BPF_EXIT:
|
|
|
|
case BPF_JMP | BPF_TAIL_CALL:
|
|
|
|
break;
|
|
|
|
default:
|
|
|
|
*rd = bpf_to_rv_reg(insn->dst_reg, ctx);
|
|
|
|
}
|
|
|
|
|
|
|
|
if (code & (BPF_ALU | BPF_X) || code & (BPF_ALU64 | BPF_X) ||
|
|
|
|
code & (BPF_JMP | BPF_X) || code & (BPF_JMP32 | BPF_X) ||
|
|
|
|
code & BPF_LDX || code & BPF_STX)
|
|
|
|
*rs = bpf_to_rv_reg(insn->src_reg, ctx);
|
|
|
|
}
|
|
|
|
|
2023-02-15 06:52:03 -07:00
|
|
|
static int emit_jump_and_link(u8 rd, s64 rvoff, bool fixed_addr,
|
2020-04-06 15:16:04 -07:00
|
|
|
struct rv_jit_context *ctx)
|
2019-12-16 02:13:38 -07:00
|
|
|
{
|
|
|
|
s64 upper, lower;
|
|
|
|
|
2023-02-15 06:52:03 -07:00
|
|
|
if (rvoff && fixed_addr && is_21b_int(rvoff)) {
|
2019-12-16 02:13:38 -07:00
|
|
|
emit(rv_jal(rd, rvoff >> 1), ctx);
|
2020-04-06 15:16:04 -07:00
|
|
|
return 0;
|
|
|
|
} else if (in_auipc_jalr_range(rvoff)) {
|
|
|
|
upper = (rvoff + (1 << 11)) >> 12;
|
|
|
|
lower = rvoff & 0xfff;
|
|
|
|
emit(rv_auipc(RV_REG_T1, upper), ctx);
|
|
|
|
emit(rv_jalr(rd, RV_REG_T1, lower), ctx);
|
|
|
|
return 0;
|
2019-12-16 02:13:38 -07:00
|
|
|
}
|
|
|
|
|
2020-04-06 15:16:04 -07:00
|
|
|
pr_err("bpf-jit: target offset 0x%llx is out of range\n", rvoff);
|
|
|
|
return -ERANGE;
|
2019-12-16 02:13:38 -07:00
|
|
|
}
|
|
|
|
|
2019-12-16 02:13:36 -07:00
|
|
|
static bool is_signed_bpf_cond(u8 cond)
|
|
|
|
{
|
|
|
|
return cond == BPF_JSGT || cond == BPF_JSLT ||
|
|
|
|
cond == BPF_JSGE || cond == BPF_JSLE;
|
|
|
|
}
|
|
|
|
|
2023-02-15 06:52:03 -07:00
|
|
|
static int emit_call(u64 addr, bool fixed_addr, struct rv_jit_context *ctx)
|
2019-12-16 02:13:41 -07:00
|
|
|
{
|
|
|
|
s64 off = 0;
|
|
|
|
u64 ip;
|
|
|
|
|
2023-08-31 06:12:29 -07:00
|
|
|
if (addr && ctx->insns && ctx->ro_insns) {
|
|
|
|
/*
|
|
|
|
* Use the ro_insns(RX) to calculate the offset as the BPF
|
|
|
|
* program will finally run from this memory region.
|
|
|
|
*/
|
|
|
|
ip = (u64)(long)(ctx->ro_insns + ctx->ninsns);
|
2019-12-16 02:13:41 -07:00
|
|
|
off = addr - ip;
|
|
|
|
}
|
|
|
|
|
2023-02-15 06:52:03 -07:00
|
|
|
return emit_jump_and_link(RV_REG_RA, off, fixed_addr, ctx);
|
2019-12-16 02:13:41 -07:00
|
|
|
}
|
|
|
|
|
2024-03-03 10:02:07 -07:00
|
|
|
static inline void emit_kcfi(u32 hash, struct rv_jit_context *ctx)
|
|
|
|
{
|
|
|
|
if (IS_ENABLED(CONFIG_CFI_CLANG))
|
|
|
|
emit(hash, ctx);
|
|
|
|
}
|
|
|
|
|
2022-04-10 03:12:46 -07:00
|
|
|
static void emit_atomic(u8 rd, u8 rs, s16 off, s32 imm, bool is64,
|
|
|
|
struct rv_jit_context *ctx)
|
|
|
|
{
|
|
|
|
u8 r0;
|
|
|
|
int jmp_offset;
|
|
|
|
|
|
|
|
if (off) {
|
|
|
|
if (is_12b_int(off)) {
|
|
|
|
emit_addi(RV_REG_T1, rd, off, ctx);
|
|
|
|
} else {
|
|
|
|
emit_imm(RV_REG_T1, off, ctx);
|
|
|
|
emit_add(RV_REG_T1, RV_REG_T1, rd, ctx);
|
|
|
|
}
|
|
|
|
rd = RV_REG_T1;
|
|
|
|
}
|
|
|
|
|
|
|
|
switch (imm) {
|
|
|
|
/* lock *(u32/u64 *)(dst_reg + off16) <op>= src_reg */
|
|
|
|
case BPF_ADD:
|
|
|
|
emit(is64 ? rv_amoadd_d(RV_REG_ZERO, rs, rd, 0, 0) :
|
|
|
|
rv_amoadd_w(RV_REG_ZERO, rs, rd, 0, 0), ctx);
|
|
|
|
break;
|
|
|
|
case BPF_AND:
|
|
|
|
emit(is64 ? rv_amoand_d(RV_REG_ZERO, rs, rd, 0, 0) :
|
|
|
|
rv_amoand_w(RV_REG_ZERO, rs, rd, 0, 0), ctx);
|
|
|
|
break;
|
|
|
|
case BPF_OR:
|
|
|
|
emit(is64 ? rv_amoor_d(RV_REG_ZERO, rs, rd, 0, 0) :
|
|
|
|
rv_amoor_w(RV_REG_ZERO, rs, rd, 0, 0), ctx);
|
|
|
|
break;
|
|
|
|
case BPF_XOR:
|
|
|
|
emit(is64 ? rv_amoxor_d(RV_REG_ZERO, rs, rd, 0, 0) :
|
|
|
|
rv_amoxor_w(RV_REG_ZERO, rs, rd, 0, 0), ctx);
|
|
|
|
break;
|
|
|
|
/* src_reg = atomic_fetch_<op>(dst_reg + off16, src_reg) */
|
|
|
|
case BPF_ADD | BPF_FETCH:
|
|
|
|
emit(is64 ? rv_amoadd_d(rs, rs, rd, 0, 0) :
|
|
|
|
rv_amoadd_w(rs, rs, rd, 0, 0), ctx);
|
|
|
|
if (!is64)
|
2024-01-15 06:12:31 -07:00
|
|
|
emit_zextw(rs, rs, ctx);
|
2022-04-10 03:12:46 -07:00
|
|
|
break;
|
|
|
|
case BPF_AND | BPF_FETCH:
|
|
|
|
emit(is64 ? rv_amoand_d(rs, rs, rd, 0, 0) :
|
|
|
|
rv_amoand_w(rs, rs, rd, 0, 0), ctx);
|
|
|
|
if (!is64)
|
2024-01-15 06:12:31 -07:00
|
|
|
emit_zextw(rs, rs, ctx);
|
2022-04-10 03:12:46 -07:00
|
|
|
break;
|
|
|
|
case BPF_OR | BPF_FETCH:
|
|
|
|
emit(is64 ? rv_amoor_d(rs, rs, rd, 0, 0) :
|
|
|
|
rv_amoor_w(rs, rs, rd, 0, 0), ctx);
|
|
|
|
if (!is64)
|
2024-01-15 06:12:31 -07:00
|
|
|
emit_zextw(rs, rs, ctx);
|
2022-04-10 03:12:46 -07:00
|
|
|
break;
|
|
|
|
case BPF_XOR | BPF_FETCH:
|
|
|
|
emit(is64 ? rv_amoxor_d(rs, rs, rd, 0, 0) :
|
|
|
|
rv_amoxor_w(rs, rs, rd, 0, 0), ctx);
|
|
|
|
if (!is64)
|
2024-01-15 06:12:31 -07:00
|
|
|
emit_zextw(rs, rs, ctx);
|
2022-04-10 03:12:46 -07:00
|
|
|
break;
|
|
|
|
/* src_reg = atomic_xchg(dst_reg + off16, src_reg); */
|
|
|
|
case BPF_XCHG:
|
|
|
|
emit(is64 ? rv_amoswap_d(rs, rs, rd, 0, 0) :
|
|
|
|
rv_amoswap_w(rs, rs, rd, 0, 0), ctx);
|
|
|
|
if (!is64)
|
2024-01-15 06:12:31 -07:00
|
|
|
emit_zextw(rs, rs, ctx);
|
2022-04-10 03:12:46 -07:00
|
|
|
break;
|
|
|
|
/* r0 = atomic_cmpxchg(dst_reg + off16, r0, src_reg); */
|
|
|
|
case BPF_CMPXCHG:
|
|
|
|
r0 = bpf_to_rv_reg(BPF_REG_0, ctx);
|
|
|
|
emit(is64 ? rv_addi(RV_REG_T2, r0, 0) :
|
|
|
|
rv_addiw(RV_REG_T2, r0, 0), ctx);
|
|
|
|
emit(is64 ? rv_lr_d(r0, 0, rd, 0, 0) :
|
|
|
|
rv_lr_w(r0, 0, rd, 0, 0), ctx);
|
|
|
|
jmp_offset = ninsns_rvoff(8);
|
|
|
|
emit(rv_bne(RV_REG_T2, r0, jmp_offset >> 1), ctx);
|
|
|
|
emit(is64 ? rv_sc_d(RV_REG_T3, rs, rd, 0, 0) :
|
|
|
|
rv_sc_w(RV_REG_T3, rs, rd, 0, 0), ctx);
|
|
|
|
jmp_offset = ninsns_rvoff(-6);
|
|
|
|
emit(rv_bne(RV_REG_T3, 0, jmp_offset >> 1), ctx);
|
|
|
|
emit(rv_fence(0x3, 0x3), ctx);
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2021-10-27 04:18:22 -07:00
|
|
|
#define BPF_FIXUP_OFFSET_MASK GENMASK(26, 0)
|
|
|
|
#define BPF_FIXUP_REG_MASK GENMASK(31, 27)
|
2024-04-04 04:42:02 -07:00
|
|
|
#define REG_DONT_CLEAR_MARKER 0 /* RV_REG_ZERO unused in pt_regmap */
|
2021-10-27 04:18:22 -07:00
|
|
|
|
2021-11-18 04:26:05 -07:00
|
|
|
bool ex_handler_bpf(const struct exception_table_entry *ex,
|
|
|
|
struct pt_regs *regs)
|
2021-10-27 04:18:22 -07:00
|
|
|
{
|
|
|
|
off_t offset = FIELD_GET(BPF_FIXUP_OFFSET_MASK, ex->fixup);
|
|
|
|
int regs_offset = FIELD_GET(BPF_FIXUP_REG_MASK, ex->fixup);
|
|
|
|
|
2024-04-04 04:42:02 -07:00
|
|
|
if (regs_offset != REG_DONT_CLEAR_MARKER)
|
|
|
|
*(unsigned long *)((void *)regs + pt_regmap[regs_offset]) = 0;
|
2021-10-27 04:18:22 -07:00
|
|
|
regs->epc = (unsigned long)&ex->fixup - offset;
|
|
|
|
|
2021-11-18 04:24:14 -07:00
|
|
|
return true;
|
2021-10-27 04:18:22 -07:00
|
|
|
}
|
|
|
|
|
|
|
|
/* For accesses to BTF pointers, add an entry to the exception table */
|
|
|
|
static int add_exception_handler(const struct bpf_insn *insn,
|
|
|
|
struct rv_jit_context *ctx,
|
|
|
|
int dst_reg, int insn_len)
|
|
|
|
{
|
|
|
|
struct exception_table_entry *ex;
|
|
|
|
unsigned long pc;
|
2023-08-31 06:12:29 -07:00
|
|
|
off_t ins_offset;
|
|
|
|
off_t fixup_offset;
|
2021-10-27 04:18:22 -07:00
|
|
|
|
2023-08-31 06:12:29 -07:00
|
|
|
if (!ctx->insns || !ctx->ro_insns || !ctx->prog->aux->extable ||
|
2024-04-04 04:42:02 -07:00
|
|
|
(BPF_MODE(insn->code) != BPF_PROBE_MEM && BPF_MODE(insn->code) != BPF_PROBE_MEMSX &&
|
|
|
|
BPF_MODE(insn->code) != BPF_PROBE_MEM32))
|
2021-10-27 04:18:22 -07:00
|
|
|
return 0;
|
|
|
|
|
|
|
|
if (WARN_ON_ONCE(ctx->nexentries >= ctx->prog->aux->num_exentries))
|
|
|
|
return -EINVAL;
|
|
|
|
|
|
|
|
if (WARN_ON_ONCE(insn_len > ctx->ninsns))
|
|
|
|
return -EINVAL;
|
|
|
|
|
|
|
|
if (WARN_ON_ONCE(!rvc_enabled() && insn_len == 1))
|
|
|
|
return -EINVAL;
|
|
|
|
|
|
|
|
ex = &ctx->prog->aux->extable[ctx->nexentries];
|
2023-08-31 06:12:29 -07:00
|
|
|
pc = (unsigned long)&ctx->ro_insns[ctx->ninsns - insn_len];
|
2021-10-27 04:18:22 -07:00
|
|
|
|
2023-08-31 06:12:29 -07:00
|
|
|
/*
|
|
|
|
* This is the relative offset of the instruction that may fault from
|
|
|
|
* the exception table itself. This will be written to the exception
|
|
|
|
* table and if this instruction faults, the destination register will
|
|
|
|
* be set to '0' and the execution will jump to the next instruction.
|
|
|
|
*/
|
|
|
|
ins_offset = pc - (long)&ex->insn;
|
|
|
|
if (WARN_ON_ONCE(ins_offset >= 0 || ins_offset < INT_MIN))
|
2021-10-27 04:18:22 -07:00
|
|
|
return -ERANGE;
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Since the extable follows the program, the fixup offset is always
|
|
|
|
* negative and limited to BPF_JIT_REGION_SIZE. Store a positive value
|
|
|
|
* to keep things simple, and put the destination register in the upper
|
|
|
|
* bits. We don't need to worry about buildtime or runtime sort
|
|
|
|
* modifying the upper bits because the table is already sorted, and
|
|
|
|
* isn't part of the main exception table.
|
2023-08-31 06:12:29 -07:00
|
|
|
*
|
|
|
|
* The fixup_offset is set to the next instruction from the instruction
|
|
|
|
* that may fault. The execution will jump to this after handling the
|
|
|
|
* fault.
|
2021-10-27 04:18:22 -07:00
|
|
|
*/
|
2023-08-31 06:12:29 -07:00
|
|
|
fixup_offset = (long)&ex->fixup - (pc + insn_len * sizeof(u16));
|
|
|
|
if (!FIELD_FIT(BPF_FIXUP_OFFSET_MASK, fixup_offset))
|
2021-10-27 04:18:22 -07:00
|
|
|
return -ERANGE;
|
|
|
|
|
2023-08-31 06:12:29 -07:00
|
|
|
/*
|
|
|
|
* The offsets above have been calculated using the RO buffer but we
|
|
|
|
* need to use the R/W buffer for writes.
|
|
|
|
* switch ex to rw buffer for writing.
|
|
|
|
*/
|
|
|
|
ex = (void *)ctx->insns + ((void *)ex - (void *)ctx->ro_insns);
|
|
|
|
|
|
|
|
ex->insn = ins_offset;
|
|
|
|
|
|
|
|
ex->fixup = FIELD_PREP(BPF_FIXUP_OFFSET_MASK, fixup_offset) |
|
2021-10-27 04:18:22 -07:00
|
|
|
FIELD_PREP(BPF_FIXUP_REG_MASK, dst_reg);
|
2021-11-18 04:26:05 -07:00
|
|
|
ex->type = EX_TYPE_BPF;
|
2021-10-27 04:18:22 -07:00
|
|
|
|
|
|
|
ctx->nexentries++;
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2023-07-21 03:06:27 -07:00
|
|
|
static int gen_jump_or_nops(void *target, void *ip, u32 *insns, bool is_call)
|
2023-02-15 06:52:04 -07:00
|
|
|
{
|
|
|
|
s64 rvoff;
|
|
|
|
struct rv_jit_context ctx;
|
|
|
|
|
|
|
|
ctx.ninsns = 0;
|
|
|
|
ctx.insns = (u16 *)insns;
|
|
|
|
|
|
|
|
if (!target) {
|
|
|
|
emit(rv_nop(), &ctx);
|
|
|
|
emit(rv_nop(), &ctx);
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
rvoff = (s64)(target - ip);
|
2023-07-21 03:06:27 -07:00
|
|
|
return emit_jump_and_link(is_call ? RV_REG_T0 : RV_REG_ZERO, rvoff, false, &ctx);
|
2023-02-15 06:52:04 -07:00
|
|
|
}
|
|
|
|
|
|
|
|
int bpf_arch_text_poke(void *ip, enum bpf_text_poke_type poke_type,
|
|
|
|
void *old_addr, void *new_addr)
|
|
|
|
{
|
2023-07-21 03:06:27 -07:00
|
|
|
u32 old_insns[RV_FENTRY_NINSNS], new_insns[RV_FENTRY_NINSNS];
|
2023-02-15 06:52:04 -07:00
|
|
|
bool is_call = poke_type == BPF_MOD_CALL;
|
|
|
|
int ret;
|
|
|
|
|
2023-07-21 03:06:27 -07:00
|
|
|
if (!is_kernel_text((unsigned long)ip) &&
|
|
|
|
!is_bpf_text_address((unsigned long)ip))
|
2023-02-15 06:52:04 -07:00
|
|
|
return -ENOTSUPP;
|
|
|
|
|
2023-07-21 03:06:27 -07:00
|
|
|
ret = gen_jump_or_nops(old_addr, ip, old_insns, is_call);
|
2023-02-15 06:52:04 -07:00
|
|
|
if (ret)
|
|
|
|
return ret;
|
|
|
|
|
2023-07-21 03:06:27 -07:00
|
|
|
if (memcmp(ip, old_insns, RV_FENTRY_NINSNS * 4))
|
2023-02-15 06:52:04 -07:00
|
|
|
return -EFAULT;
|
|
|
|
|
2023-07-21 03:06:27 -07:00
|
|
|
ret = gen_jump_or_nops(new_addr, ip, new_insns, is_call);
|
2023-02-15 06:52:04 -07:00
|
|
|
if (ret)
|
|
|
|
return ret;
|
|
|
|
|
|
|
|
cpus_read_lock();
|
|
|
|
mutex_lock(&text_mutex);
|
2023-07-21 03:06:27 -07:00
|
|
|
if (memcmp(ip, new_insns, RV_FENTRY_NINSNS * 4))
|
|
|
|
ret = patch_text(ip, new_insns, RV_FENTRY_NINSNS);
|
2023-02-15 06:52:04 -07:00
|
|
|
mutex_unlock(&text_mutex);
|
|
|
|
cpus_read_unlock();
|
|
|
|
|
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
2023-02-15 06:52:05 -07:00
|
|
|
static void store_args(int nregs, int args_off, struct rv_jit_context *ctx)
|
|
|
|
{
|
|
|
|
int i;
|
|
|
|
|
|
|
|
for (i = 0; i < nregs; i++) {
|
|
|
|
emit_sd(RV_REG_FP, -args_off, RV_REG_A0 + i, ctx);
|
|
|
|
args_off -= 8;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
static void restore_args(int nregs, int args_off, struct rv_jit_context *ctx)
|
|
|
|
{
|
|
|
|
int i;
|
|
|
|
|
|
|
|
for (i = 0; i < nregs; i++) {
|
|
|
|
emit_ld(RV_REG_A0 + i, -args_off, RV_REG_FP, ctx);
|
|
|
|
args_off -= 8;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
static int invoke_bpf_prog(struct bpf_tramp_link *l, int args_off, int retval_off,
|
|
|
|
int run_ctx_off, bool save_ret, struct rv_jit_context *ctx)
|
|
|
|
{
|
|
|
|
int ret, branch_off;
|
|
|
|
struct bpf_prog *p = l->link.prog;
|
|
|
|
int cookie_off = offsetof(struct bpf_tramp_run_ctx, bpf_cookie);
|
|
|
|
|
|
|
|
if (l->cookie) {
|
|
|
|
emit_imm(RV_REG_T1, l->cookie, ctx);
|
|
|
|
emit_sd(RV_REG_FP, -run_ctx_off + cookie_off, RV_REG_T1, ctx);
|
|
|
|
} else {
|
|
|
|
emit_sd(RV_REG_FP, -run_ctx_off + cookie_off, RV_REG_ZERO, ctx);
|
|
|
|
}
|
|
|
|
|
|
|
|
/* arg1: prog */
|
|
|
|
emit_imm(RV_REG_A0, (const s64)p, ctx);
|
|
|
|
/* arg2: &run_ctx */
|
|
|
|
emit_addi(RV_REG_A1, RV_REG_FP, -run_ctx_off, ctx);
|
|
|
|
ret = emit_call((const u64)bpf_trampoline_enter(p), true, ctx);
|
|
|
|
if (ret)
|
|
|
|
return ret;
|
|
|
|
|
|
|
|
/* if (__bpf_prog_enter(prog) == 0)
|
|
|
|
* goto skip_exec_of_prog;
|
|
|
|
*/
|
|
|
|
branch_off = ctx->ninsns;
|
|
|
|
/* nop reserved for conditional jump */
|
|
|
|
emit(rv_nop(), ctx);
|
|
|
|
|
|
|
|
/* store prog start time */
|
|
|
|
emit_mv(RV_REG_S1, RV_REG_A0, ctx);
|
|
|
|
|
|
|
|
/* arg1: &args_off */
|
|
|
|
emit_addi(RV_REG_A0, RV_REG_FP, -args_off, ctx);
|
|
|
|
if (!p->jited)
|
|
|
|
/* arg2: progs[i]->insnsi for interpreter */
|
|
|
|
emit_imm(RV_REG_A1, (const s64)p->insnsi, ctx);
|
|
|
|
ret = emit_call((const u64)p->bpf_func, true, ctx);
|
|
|
|
if (ret)
|
|
|
|
return ret;
|
|
|
|
|
2023-10-04 05:07:06 -07:00
|
|
|
if (save_ret) {
|
|
|
|
emit_sd(RV_REG_FP, -retval_off, RV_REG_A0, ctx);
|
|
|
|
emit_sd(RV_REG_FP, -(retval_off - 8), regmap[BPF_REG_0], ctx);
|
|
|
|
}
|
2023-02-15 06:52:05 -07:00
|
|
|
|
|
|
|
/* update branch with beqz */
|
|
|
|
if (ctx->insns) {
|
|
|
|
int offset = ninsns_rvoff(ctx->ninsns - branch_off);
|
|
|
|
u32 insn = rv_beq(RV_REG_A0, RV_REG_ZERO, offset >> 1);
|
|
|
|
*(u32 *)(ctx->insns + branch_off) = insn;
|
|
|
|
}
|
|
|
|
|
|
|
|
/* arg1: prog */
|
|
|
|
emit_imm(RV_REG_A0, (const s64)p, ctx);
|
|
|
|
/* arg2: prog start time */
|
|
|
|
emit_mv(RV_REG_A1, RV_REG_S1, ctx);
|
|
|
|
/* arg3: &run_ctx */
|
|
|
|
emit_addi(RV_REG_A2, RV_REG_FP, -run_ctx_off, ctx);
|
|
|
|
ret = emit_call((const u64)bpf_trampoline_exit(p), true, ctx);
|
|
|
|
|
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
|
|
|
static int __arch_prepare_bpf_trampoline(struct bpf_tramp_image *im,
|
|
|
|
const struct btf_func_model *m,
|
|
|
|
struct bpf_tramp_links *tlinks,
|
|
|
|
void *func_addr, u32 flags,
|
|
|
|
struct rv_jit_context *ctx)
|
|
|
|
{
|
|
|
|
int i, ret, offset;
|
|
|
|
int *branches_off = NULL;
|
|
|
|
int stack_size = 0, nregs = m->nr_args;
|
2023-07-21 03:06:27 -07:00
|
|
|
int retval_off, args_off, nregs_off, ip_off, run_ctx_off, sreg_off;
|
2023-02-15 06:52:05 -07:00
|
|
|
struct bpf_tramp_links *fentry = &tlinks[BPF_TRAMP_FENTRY];
|
|
|
|
struct bpf_tramp_links *fexit = &tlinks[BPF_TRAMP_FEXIT];
|
|
|
|
struct bpf_tramp_links *fmod_ret = &tlinks[BPF_TRAMP_MODIFY_RETURN];
|
2024-01-22 19:32:07 -07:00
|
|
|
bool is_struct_ops = flags & BPF_TRAMP_F_INDIRECT;
|
2023-02-15 06:52:05 -07:00
|
|
|
void *orig_call = func_addr;
|
|
|
|
bool save_ret;
|
|
|
|
u32 insn;
|
|
|
|
|
2023-07-21 03:06:27 -07:00
|
|
|
/* Two types of generated trampoline stack layout:
|
|
|
|
*
|
|
|
|
* 1. trampoline called from function entry
|
|
|
|
* --------------------------------------
|
|
|
|
* FP + 8 [ RA to parent func ] return address to parent
|
|
|
|
* function
|
|
|
|
* FP + 0 [ FP of parent func ] frame pointer of parent
|
|
|
|
* function
|
|
|
|
* FP - 8 [ T0 to traced func ] return address of traced
|
|
|
|
* function
|
|
|
|
* FP - 16 [ FP of traced func ] frame pointer of traced
|
|
|
|
* function
|
|
|
|
* --------------------------------------
|
2023-02-15 06:52:05 -07:00
|
|
|
*
|
2023-07-21 03:06:27 -07:00
|
|
|
* 2. trampoline called directly
|
|
|
|
* --------------------------------------
|
|
|
|
* FP - 8 [ RA to caller func ] return address to caller
|
2023-02-15 06:52:05 -07:00
|
|
|
* function
|
2023-07-21 03:06:27 -07:00
|
|
|
* FP - 16 [ FP of caller func ] frame pointer of caller
|
2023-02-15 06:52:05 -07:00
|
|
|
* function
|
2023-07-21 03:06:27 -07:00
|
|
|
* --------------------------------------
|
2023-02-15 06:52:05 -07:00
|
|
|
*
|
|
|
|
* FP - retval_off [ return value ] BPF_TRAMP_F_CALL_ORIG or
|
|
|
|
* BPF_TRAMP_F_RET_FENTRY_RET
|
|
|
|
* [ argN ]
|
|
|
|
* [ ... ]
|
|
|
|
* FP - args_off [ arg1 ]
|
|
|
|
*
|
|
|
|
* FP - nregs_off [ regs count ]
|
|
|
|
*
|
|
|
|
* FP - ip_off [ traced func ] BPF_TRAMP_F_IP_ARG
|
|
|
|
*
|
|
|
|
* FP - run_ctx_off [ bpf_tramp_run_ctx ]
|
|
|
|
*
|
|
|
|
* FP - sreg_off [ callee saved reg ]
|
|
|
|
*
|
|
|
|
* [ pads ] pads for 16 bytes alignment
|
|
|
|
*/
|
|
|
|
|
|
|
|
if (flags & (BPF_TRAMP_F_ORIG_STACK | BPF_TRAMP_F_SHARE_IPMODIFY))
|
|
|
|
return -ENOTSUPP;
|
|
|
|
|
|
|
|
/* extra regiters for struct arguments */
|
|
|
|
for (i = 0; i < m->nr_args; i++)
|
|
|
|
if (m->arg_flags[i] & BTF_FMODEL_STRUCT_ARG)
|
|
|
|
nregs += round_up(m->arg_size[i], 8) / 8 - 1;
|
|
|
|
|
|
|
|
/* 8 arguments passed by registers */
|
|
|
|
if (nregs > 8)
|
|
|
|
return -ENOTSUPP;
|
|
|
|
|
2023-07-21 03:06:27 -07:00
|
|
|
/* room of trampoline frame to store return address and frame pointer */
|
|
|
|
stack_size += 16;
|
2023-02-15 06:52:05 -07:00
|
|
|
|
|
|
|
save_ret = flags & (BPF_TRAMP_F_CALL_ORIG | BPF_TRAMP_F_RET_FENTRY_RET);
|
|
|
|
if (save_ret) {
|
2023-10-04 05:07:06 -07:00
|
|
|
stack_size += 16; /* Save both A5 (BPF R0) and A0 */
|
2023-02-15 06:52:05 -07:00
|
|
|
retval_off = stack_size;
|
|
|
|
}
|
|
|
|
|
|
|
|
stack_size += nregs * 8;
|
|
|
|
args_off = stack_size;
|
|
|
|
|
|
|
|
stack_size += 8;
|
|
|
|
nregs_off = stack_size;
|
|
|
|
|
|
|
|
if (flags & BPF_TRAMP_F_IP_ARG) {
|
|
|
|
stack_size += 8;
|
|
|
|
ip_off = stack_size;
|
|
|
|
}
|
|
|
|
|
|
|
|
stack_size += round_up(sizeof(struct bpf_tramp_run_ctx), 8);
|
|
|
|
run_ctx_off = stack_size;
|
|
|
|
|
|
|
|
stack_size += 8;
|
|
|
|
sreg_off = stack_size;
|
|
|
|
|
|
|
|
stack_size = round_up(stack_size, 16);
|
|
|
|
|
2024-01-22 19:32:07 -07:00
|
|
|
if (!is_struct_ops) {
|
2023-07-21 03:06:27 -07:00
|
|
|
/* For the trampoline called from function entry,
|
|
|
|
* the frame of traced function and the frame of
|
|
|
|
* trampoline need to be considered.
|
|
|
|
*/
|
|
|
|
emit_addi(RV_REG_SP, RV_REG_SP, -16, ctx);
|
|
|
|
emit_sd(RV_REG_SP, 8, RV_REG_RA, ctx);
|
|
|
|
emit_sd(RV_REG_SP, 0, RV_REG_FP, ctx);
|
|
|
|
emit_addi(RV_REG_FP, RV_REG_SP, 16, ctx);
|
|
|
|
|
|
|
|
emit_addi(RV_REG_SP, RV_REG_SP, -stack_size, ctx);
|
|
|
|
emit_sd(RV_REG_SP, stack_size - 8, RV_REG_T0, ctx);
|
|
|
|
emit_sd(RV_REG_SP, stack_size - 16, RV_REG_FP, ctx);
|
|
|
|
emit_addi(RV_REG_FP, RV_REG_SP, stack_size, ctx);
|
|
|
|
} else {
|
2024-03-03 10:02:07 -07:00
|
|
|
/* emit kcfi hash */
|
|
|
|
emit_kcfi(cfi_get_func_hash(func_addr), ctx);
|
2023-07-21 03:06:27 -07:00
|
|
|
/* For the trampoline called directly, just handle
|
|
|
|
* the frame of trampoline.
|
|
|
|
*/
|
|
|
|
emit_addi(RV_REG_SP, RV_REG_SP, -stack_size, ctx);
|
|
|
|
emit_sd(RV_REG_SP, stack_size - 8, RV_REG_RA, ctx);
|
|
|
|
emit_sd(RV_REG_SP, stack_size - 16, RV_REG_FP, ctx);
|
|
|
|
emit_addi(RV_REG_FP, RV_REG_SP, stack_size, ctx);
|
|
|
|
}
|
2023-02-15 06:52:05 -07:00
|
|
|
|
|
|
|
/* callee saved register S1 to pass start time */
|
|
|
|
emit_sd(RV_REG_FP, -sreg_off, RV_REG_S1, ctx);
|
|
|
|
|
|
|
|
/* store ip address of the traced function */
|
|
|
|
if (flags & BPF_TRAMP_F_IP_ARG) {
|
|
|
|
emit_imm(RV_REG_T1, (const s64)func_addr, ctx);
|
|
|
|
emit_sd(RV_REG_FP, -ip_off, RV_REG_T1, ctx);
|
|
|
|
}
|
|
|
|
|
|
|
|
emit_li(RV_REG_T1, nregs, ctx);
|
|
|
|
emit_sd(RV_REG_FP, -nregs_off, RV_REG_T1, ctx);
|
|
|
|
|
|
|
|
store_args(nregs, args_off, ctx);
|
|
|
|
|
|
|
|
/* skip to actual body of traced function */
|
|
|
|
if (flags & BPF_TRAMP_F_SKIP_FRAME)
|
2023-07-21 03:06:27 -07:00
|
|
|
orig_call += RV_FENTRY_NINSNS * 4;
|
2023-02-15 06:52:05 -07:00
|
|
|
|
|
|
|
if (flags & BPF_TRAMP_F_CALL_ORIG) {
|
|
|
|
emit_imm(RV_REG_A0, (const s64)im, ctx);
|
|
|
|
ret = emit_call((const u64)__bpf_tramp_enter, true, ctx);
|
|
|
|
if (ret)
|
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
|
|
|
for (i = 0; i < fentry->nr_links; i++) {
|
|
|
|
ret = invoke_bpf_prog(fentry->links[i], args_off, retval_off, run_ctx_off,
|
|
|
|
flags & BPF_TRAMP_F_RET_FENTRY_RET, ctx);
|
|
|
|
if (ret)
|
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (fmod_ret->nr_links) {
|
|
|
|
branches_off = kcalloc(fmod_ret->nr_links, sizeof(int), GFP_KERNEL);
|
|
|
|
if (!branches_off)
|
|
|
|
return -ENOMEM;
|
|
|
|
|
|
|
|
/* cleanup to avoid garbage return value confusion */
|
|
|
|
emit_sd(RV_REG_FP, -retval_off, RV_REG_ZERO, ctx);
|
|
|
|
for (i = 0; i < fmod_ret->nr_links; i++) {
|
|
|
|
ret = invoke_bpf_prog(fmod_ret->links[i], args_off, retval_off,
|
|
|
|
run_ctx_off, true, ctx);
|
|
|
|
if (ret)
|
|
|
|
goto out;
|
|
|
|
emit_ld(RV_REG_T1, -retval_off, RV_REG_FP, ctx);
|
|
|
|
branches_off[i] = ctx->ninsns;
|
|
|
|
/* nop reserved for conditional jump */
|
|
|
|
emit(rv_nop(), ctx);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
if (flags & BPF_TRAMP_F_CALL_ORIG) {
|
|
|
|
restore_args(nregs, args_off, ctx);
|
|
|
|
ret = emit_call((const u64)orig_call, true, ctx);
|
|
|
|
if (ret)
|
|
|
|
goto out;
|
|
|
|
emit_sd(RV_REG_FP, -retval_off, RV_REG_A0, ctx);
|
2023-10-04 05:07:06 -07:00
|
|
|
emit_sd(RV_REG_FP, -(retval_off - 8), regmap[BPF_REG_0], ctx);
|
2023-02-15 06:52:05 -07:00
|
|
|
im->ip_after_call = ctx->insns + ctx->ninsns;
|
|
|
|
/* 2 nops reserved for auipc+jalr pair */
|
|
|
|
emit(rv_nop(), ctx);
|
|
|
|
emit(rv_nop(), ctx);
|
|
|
|
}
|
|
|
|
|
|
|
|
/* update branches saved in invoke_bpf_mod_ret with bnez */
|
|
|
|
for (i = 0; ctx->insns && i < fmod_ret->nr_links; i++) {
|
|
|
|
offset = ninsns_rvoff(ctx->ninsns - branches_off[i]);
|
|
|
|
insn = rv_bne(RV_REG_T1, RV_REG_ZERO, offset >> 1);
|
|
|
|
*(u32 *)(ctx->insns + branches_off[i]) = insn;
|
|
|
|
}
|
|
|
|
|
|
|
|
for (i = 0; i < fexit->nr_links; i++) {
|
|
|
|
ret = invoke_bpf_prog(fexit->links[i], args_off, retval_off,
|
|
|
|
run_ctx_off, false, ctx);
|
|
|
|
if (ret)
|
|
|
|
goto out;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (flags & BPF_TRAMP_F_CALL_ORIG) {
|
|
|
|
im->ip_epilogue = ctx->insns + ctx->ninsns;
|
|
|
|
emit_imm(RV_REG_A0, (const s64)im, ctx);
|
|
|
|
ret = emit_call((const u64)__bpf_tramp_exit, true, ctx);
|
|
|
|
if (ret)
|
|
|
|
goto out;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (flags & BPF_TRAMP_F_RESTORE_REGS)
|
|
|
|
restore_args(nregs, args_off, ctx);
|
|
|
|
|
2023-10-04 05:07:06 -07:00
|
|
|
if (save_ret) {
|
2023-02-15 06:52:05 -07:00
|
|
|
emit_ld(RV_REG_A0, -retval_off, RV_REG_FP, ctx);
|
2023-10-04 05:07:06 -07:00
|
|
|
emit_ld(regmap[BPF_REG_0], -(retval_off - 8), RV_REG_FP, ctx);
|
|
|
|
}
|
2023-02-15 06:52:05 -07:00
|
|
|
|
|
|
|
emit_ld(RV_REG_S1, -sreg_off, RV_REG_FP, ctx);
|
|
|
|
|
2024-01-22 19:32:07 -07:00
|
|
|
if (!is_struct_ops) {
|
2023-07-21 03:06:27 -07:00
|
|
|
/* trampoline called from function entry */
|
|
|
|
emit_ld(RV_REG_T0, stack_size - 8, RV_REG_SP, ctx);
|
|
|
|
emit_ld(RV_REG_FP, stack_size - 16, RV_REG_SP, ctx);
|
|
|
|
emit_addi(RV_REG_SP, RV_REG_SP, stack_size, ctx);
|
2023-02-15 06:52:05 -07:00
|
|
|
|
2023-07-21 03:06:27 -07:00
|
|
|
emit_ld(RV_REG_RA, 8, RV_REG_SP, ctx);
|
|
|
|
emit_ld(RV_REG_FP, 0, RV_REG_SP, ctx);
|
|
|
|
emit_addi(RV_REG_SP, RV_REG_SP, 16, ctx);
|
2023-02-15 06:52:05 -07:00
|
|
|
|
2023-07-21 03:06:27 -07:00
|
|
|
if (flags & BPF_TRAMP_F_SKIP_FRAME)
|
|
|
|
/* return to parent function */
|
|
|
|
emit_jalr(RV_REG_ZERO, RV_REG_RA, 0, ctx);
|
|
|
|
else
|
|
|
|
/* return to traced function */
|
|
|
|
emit_jalr(RV_REG_ZERO, RV_REG_T0, 0, ctx);
|
|
|
|
} else {
|
|
|
|
/* trampoline called directly */
|
|
|
|
emit_ld(RV_REG_RA, stack_size - 8, RV_REG_SP, ctx);
|
|
|
|
emit_ld(RV_REG_FP, stack_size - 16, RV_REG_SP, ctx);
|
|
|
|
emit_addi(RV_REG_SP, RV_REG_SP, stack_size, ctx);
|
|
|
|
|
|
|
|
emit_jalr(RV_REG_ZERO, RV_REG_RA, 0, ctx);
|
|
|
|
}
|
2023-02-15 06:52:05 -07:00
|
|
|
|
|
|
|
ret = ctx->ninsns;
|
|
|
|
out:
|
|
|
|
kfree(branches_off);
|
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
2023-12-06 15:40:52 -07:00
|
|
|
int arch_bpf_trampoline_size(const struct btf_func_model *m, u32 flags,
|
|
|
|
struct bpf_tramp_links *tlinks, void *func_addr)
|
|
|
|
{
|
|
|
|
struct bpf_tramp_image im;
|
|
|
|
struct rv_jit_context ctx;
|
|
|
|
int ret;
|
|
|
|
|
|
|
|
ctx.ninsns = 0;
|
|
|
|
ctx.insns = NULL;
|
|
|
|
ctx.ro_insns = NULL;
|
|
|
|
ret = __arch_prepare_bpf_trampoline(&im, m, tlinks, func_addr, flags, &ctx);
|
|
|
|
|
|
|
|
return ret < 0 ? ret : ninsns_rvoff(ctx.ninsns);
|
|
|
|
}
|
|
|
|
|
2023-02-15 06:52:05 -07:00
|
|
|
int arch_prepare_bpf_trampoline(struct bpf_tramp_image *im, void *image,
|
|
|
|
void *image_end, const struct btf_func_model *m,
|
|
|
|
u32 flags, struct bpf_tramp_links *tlinks,
|
|
|
|
void *func_addr)
|
|
|
|
{
|
|
|
|
int ret;
|
|
|
|
struct rv_jit_context ctx;
|
|
|
|
|
|
|
|
ctx.ninsns = 0;
|
2023-08-31 06:12:29 -07:00
|
|
|
/*
|
|
|
|
* The bpf_int_jit_compile() uses a RW buffer (ctx.insns) to write the
|
|
|
|
* JITed instructions and later copies it to a RX region (ctx.ro_insns).
|
|
|
|
* It also uses ctx.ro_insns to calculate offsets for jumps etc. As the
|
|
|
|
* trampoline image uses the same memory area for writing and execution,
|
|
|
|
* both ctx.insns and ctx.ro_insns can be set to image.
|
|
|
|
*/
|
2023-02-15 06:52:05 -07:00
|
|
|
ctx.insns = image;
|
2023-08-31 06:12:29 -07:00
|
|
|
ctx.ro_insns = image;
|
2023-02-15 06:52:05 -07:00
|
|
|
ret = __arch_prepare_bpf_trampoline(im, m, tlinks, func_addr, flags, &ctx);
|
|
|
|
if (ret < 0)
|
|
|
|
return ret;
|
|
|
|
|
|
|
|
bpf_flush_icache(ctx.insns, ctx.insns + ctx.ninsns);
|
|
|
|
|
|
|
|
return ninsns_rvoff(ret);
|
|
|
|
}
|
|
|
|
|
2020-03-04 22:02:04 -07:00
|
|
|
int bpf_jit_emit_insn(const struct bpf_insn *insn, struct rv_jit_context *ctx,
|
|
|
|
bool extra_pass)
|
bpf, riscv: add BPF JIT for RV64G
This commit adds a BPF JIT for RV64G.
The JIT is a two-pass JIT, and has a dynamic prolog/epilogue (similar
to the MIPS64 BPF JIT) instead of static ones (e.g. x86_64).
At the moment the RISC-V Linux port does not support
CONFIG_HAVE_KPROBES, which means that CONFIG_BPF_EVENTS is not
supported. Thus, no tests involving BPF_PROG_TYPE_TRACEPOINT,
BPF_PROG_TYPE_PERF_EVENT, BPF_PROG_TYPE_KPROBE and
BPF_PROG_TYPE_RAW_TRACEPOINT passes.
The implementation does not support "far branching" (>4KiB).
Test results:
# modprobe test_bpf
test_bpf: Summary: 378 PASSED, 0 FAILED, [366/366 JIT'ed]
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier
...
Summary: 761 PASSED, 507 SKIPPED, 2 FAILED
Note that "test_verifier" was run with one build with
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y and one without, otherwise
many of the the tests that require unaligned access were skipped.
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
0
No CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
59
The two failing test_verifier tests are:
"ld_abs: vlan + abs, test 1"
"ld_abs: jump around ld_abs"
This is due to that "far branching" involved in those tests.
All tests where done on QEMU (QEMU emulator version 3.1.50
(v3.1.0-688-g8ae951fbc106)).
Signed-off-by: Björn Töpel <bjorn.topel@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2019-02-05 05:41:22 -07:00
|
|
|
{
|
|
|
|
bool is64 = BPF_CLASS(insn->code) == BPF_ALU64 ||
|
|
|
|
BPF_CLASS(insn->code) == BPF_JMP;
|
2020-04-06 15:16:04 -07:00
|
|
|
int s, e, rvoff, ret, i = insn - ctx->prog->insnsi;
|
2019-05-24 15:25:27 -07:00
|
|
|
struct bpf_prog_aux *aux = ctx->prog->aux;
|
bpf, riscv: add BPF JIT for RV64G
This commit adds a BPF JIT for RV64G.
The JIT is a two-pass JIT, and has a dynamic prolog/epilogue (similar
to the MIPS64 BPF JIT) instead of static ones (e.g. x86_64).
At the moment the RISC-V Linux port does not support
CONFIG_HAVE_KPROBES, which means that CONFIG_BPF_EVENTS is not
supported. Thus, no tests involving BPF_PROG_TYPE_TRACEPOINT,
BPF_PROG_TYPE_PERF_EVENT, BPF_PROG_TYPE_KPROBE and
BPF_PROG_TYPE_RAW_TRACEPOINT passes.
The implementation does not support "far branching" (>4KiB).
Test results:
# modprobe test_bpf
test_bpf: Summary: 378 PASSED, 0 FAILED, [366/366 JIT'ed]
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier
...
Summary: 761 PASSED, 507 SKIPPED, 2 FAILED
Note that "test_verifier" was run with one build with
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y and one without, otherwise
many of the the tests that require unaligned access were skipped.
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
0
No CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
59
The two failing test_verifier tests are:
"ld_abs: vlan + abs, test 1"
"ld_abs: jump around ld_abs"
This is due to that "far branching" involved in those tests.
All tests where done on QEMU (QEMU emulator version 3.1.50
(v3.1.0-688-g8ae951fbc106)).
Signed-off-by: Björn Töpel <bjorn.topel@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2019-02-05 05:41:22 -07:00
|
|
|
u8 rd = -1, rs = -1, code = insn->code;
|
|
|
|
s16 off = insn->off;
|
|
|
|
s32 imm = insn->imm;
|
|
|
|
|
|
|
|
init_regs(&rd, &rs, insn, ctx);
|
|
|
|
|
|
|
|
switch (code) {
|
|
|
|
/* dst = src */
|
|
|
|
case BPF_ALU | BPF_MOV | BPF_X:
|
|
|
|
case BPF_ALU64 | BPF_MOV | BPF_X:
|
2024-04-04 04:42:03 -07:00
|
|
|
if (insn_is_cast_user(insn)) {
|
|
|
|
emit_mv(RV_REG_T1, rs, ctx);
|
|
|
|
emit_zextw(RV_REG_T1, RV_REG_T1, ctx);
|
|
|
|
emit_imm(rd, (ctx->user_vm_start >> 32) << 32, ctx);
|
|
|
|
emit(rv_beq(RV_REG_T1, RV_REG_ZERO, 4), ctx);
|
|
|
|
emit_or(RV_REG_T1, rd, RV_REG_T1, ctx);
|
|
|
|
emit_mv(rd, RV_REG_T1, ctx);
|
|
|
|
break;
|
2024-05-02 08:18:51 -07:00
|
|
|
} else if (insn_is_mov_percpu_addr(insn)) {
|
|
|
|
if (rd != rs)
|
|
|
|
emit_mv(rd, rs, ctx);
|
|
|
|
#ifdef CONFIG_SMP
|
|
|
|
/* Load current CPU number in T1 */
|
|
|
|
emit_ld(RV_REG_T1, offsetof(struct thread_info, cpu),
|
|
|
|
RV_REG_TP, ctx);
|
|
|
|
/* << 3 because offsets are 8 bytes */
|
|
|
|
emit_slli(RV_REG_T1, RV_REG_T1, 3, ctx);
|
|
|
|
/* Load address of __per_cpu_offset array in T2 */
|
|
|
|
emit_addr(RV_REG_T2, (u64)&__per_cpu_offset, extra_pass, ctx);
|
|
|
|
/* Add offset of current CPU to __per_cpu_offset */
|
|
|
|
emit_add(RV_REG_T1, RV_REG_T2, RV_REG_T1, ctx);
|
|
|
|
/* Load __per_cpu_offset[cpu] in T1 */
|
|
|
|
emit_ld(RV_REG_T1, 0, RV_REG_T1, ctx);
|
|
|
|
/* Add the offset to Rd */
|
|
|
|
emit_add(rd, rd, RV_REG_T1, ctx);
|
|
|
|
#endif
|
2024-04-04 04:42:03 -07:00
|
|
|
}
|
2019-05-24 15:25:27 -07:00
|
|
|
if (imm == 1) {
|
|
|
|
/* Special mov32 for zext */
|
2024-01-15 06:12:31 -07:00
|
|
|
emit_zextw(rd, rd, ctx);
|
2019-05-24 15:25:27 -07:00
|
|
|
break;
|
|
|
|
}
|
2023-08-24 02:49:57 -07:00
|
|
|
switch (insn->off) {
|
|
|
|
case 0:
|
|
|
|
emit_mv(rd, rs, ctx);
|
|
|
|
break;
|
|
|
|
case 8:
|
2024-01-15 06:12:34 -07:00
|
|
|
emit_sextb(rd, rs, ctx);
|
|
|
|
break;
|
2023-08-24 02:49:57 -07:00
|
|
|
case 16:
|
2024-01-15 06:12:34 -07:00
|
|
|
emit_sexth(rd, rs, ctx);
|
2023-08-24 02:49:57 -07:00
|
|
|
break;
|
|
|
|
case 32:
|
2024-01-15 06:12:30 -07:00
|
|
|
emit_sextw(rd, rs, ctx);
|
2023-08-24 02:49:57 -07:00
|
|
|
break;
|
|
|
|
}
|
2019-05-24 15:25:27 -07:00
|
|
|
if (!is64 && !aux->verifier_zext)
|
2024-01-15 06:12:31 -07:00
|
|
|
emit_zextw(rd, rd, ctx);
|
bpf, riscv: add BPF JIT for RV64G
This commit adds a BPF JIT for RV64G.
The JIT is a two-pass JIT, and has a dynamic prolog/epilogue (similar
to the MIPS64 BPF JIT) instead of static ones (e.g. x86_64).
At the moment the RISC-V Linux port does not support
CONFIG_HAVE_KPROBES, which means that CONFIG_BPF_EVENTS is not
supported. Thus, no tests involving BPF_PROG_TYPE_TRACEPOINT,
BPF_PROG_TYPE_PERF_EVENT, BPF_PROG_TYPE_KPROBE and
BPF_PROG_TYPE_RAW_TRACEPOINT passes.
The implementation does not support "far branching" (>4KiB).
Test results:
# modprobe test_bpf
test_bpf: Summary: 378 PASSED, 0 FAILED, [366/366 JIT'ed]
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier
...
Summary: 761 PASSED, 507 SKIPPED, 2 FAILED
Note that "test_verifier" was run with one build with
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y and one without, otherwise
many of the the tests that require unaligned access were skipped.
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
0
No CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
59
The two failing test_verifier tests are:
"ld_abs: vlan + abs, test 1"
"ld_abs: jump around ld_abs"
This is due to that "far branching" involved in those tests.
All tests where done on QEMU (QEMU emulator version 3.1.50
(v3.1.0-688-g8ae951fbc106)).
Signed-off-by: Björn Töpel <bjorn.topel@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2019-02-05 05:41:22 -07:00
|
|
|
break;
|
|
|
|
|
|
|
|
/* dst = dst OP src */
|
|
|
|
case BPF_ALU | BPF_ADD | BPF_X:
|
|
|
|
case BPF_ALU64 | BPF_ADD | BPF_X:
|
bpf, riscv: Use compressed instructions in the rv64 JIT
This patch uses the RVC support and encodings from bpf_jit.h to optimize
the rv64 jit.
The optimizations work by replacing emit(rv_X(...)) with a call to a
helper function emit_X, which will emit a compressed version of the
instruction when possible, and when RVC is enabled.
The JIT continues to pass all tests in lib/test_bpf.c, and introduces
no new failures to test_verifier; both with and without RVC being enabled.
Most changes are straightforward replacements of emit(rv_X(...), ctx)
with emit_X(..., ctx), with the following exceptions bearing mention;
* Change emit_imm to sign-extend the value in "lower", since the
checks for RVC (and the instructions themselves) treat the value as
signed. Otherwise, small negative immediates will not be recognized as
encodable using an RVC instruction. For example, without this change,
emit_imm(rd, -1, ctx) would cause lower to become 4095, which is not a
6b int even though a "c.li rd, -1" instruction suffices.
* For {BPF_MOV,BPF_ADD} BPF_X, drop using addiw,addw in the 32-bit
cases since the values are zero-extended into the upper 32 bits in
the following instructions anyways, and the addition commutes with
zero-extension. (BPF_SUB BPF_X must still use subw since subtraction
does not commute with zero-extension.)
This patch avoids optimizing branches and jumps to use RVC instructions
since surrounding code often makes assumptions about the sizes of
emitted instructions. Optimizing these will require changing these
functions (e.g., emit_branch) to dynamically compute jump offsets.
The following are examples of the JITed code for the verifier selftest
"direct packet read test#3 for CGROUP_SKB OK", without and with RVC
enabled, respectively. The former uses 178 bytes, and the latter uses 112,
for a ~37% reduction in code size for this example.
Without RVC:
0: 02000813 addi a6,zero,32
4: fd010113 addi sp,sp,-48
8: 02813423 sd s0,40(sp)
c: 02913023 sd s1,32(sp)
10: 01213c23 sd s2,24(sp)
14: 01313823 sd s3,16(sp)
18: 01413423 sd s4,8(sp)
1c: 03010413 addi s0,sp,48
20: 03056683 lwu a3,48(a0)
24: 02069693 slli a3,a3,0x20
28: 0206d693 srli a3,a3,0x20
2c: 03456703 lwu a4,52(a0)
30: 02071713 slli a4,a4,0x20
34: 02075713 srli a4,a4,0x20
38: 03856483 lwu s1,56(a0)
3c: 02049493 slli s1,s1,0x20
40: 0204d493 srli s1,s1,0x20
44: 03c56903 lwu s2,60(a0)
48: 02091913 slli s2,s2,0x20
4c: 02095913 srli s2,s2,0x20
50: 04056983 lwu s3,64(a0)
54: 02099993 slli s3,s3,0x20
58: 0209d993 srli s3,s3,0x20
5c: 09056a03 lwu s4,144(a0)
60: 020a1a13 slli s4,s4,0x20
64: 020a5a13 srli s4,s4,0x20
68: 00900313 addi t1,zero,9
6c: 006a7463 bgeu s4,t1,0x74
70: 00000a13 addi s4,zero,0
74: 02d52823 sw a3,48(a0)
78: 02e52a23 sw a4,52(a0)
7c: 02952c23 sw s1,56(a0)
80: 03252e23 sw s2,60(a0)
84: 05352023 sw s3,64(a0)
88: 00000793 addi a5,zero,0
8c: 02813403 ld s0,40(sp)
90: 02013483 ld s1,32(sp)
94: 01813903 ld s2,24(sp)
98: 01013983 ld s3,16(sp)
9c: 00813a03 ld s4,8(sp)
a0: 03010113 addi sp,sp,48
a4: 00078513 addi a0,a5,0
a8: 00008067 jalr zero,0(ra)
With RVC:
0: 02000813 addi a6,zero,32
4: 7179 c.addi16sp sp,-48
6: f422 c.sdsp s0,40(sp)
8: f026 c.sdsp s1,32(sp)
a: ec4a c.sdsp s2,24(sp)
c: e84e c.sdsp s3,16(sp)
e: e452 c.sdsp s4,8(sp)
10: 1800 c.addi4spn s0,sp,48
12: 03056683 lwu a3,48(a0)
16: 1682 c.slli a3,0x20
18: 9281 c.srli a3,0x20
1a: 03456703 lwu a4,52(a0)
1e: 1702 c.slli a4,0x20
20: 9301 c.srli a4,0x20
22: 03856483 lwu s1,56(a0)
26: 1482 c.slli s1,0x20
28: 9081 c.srli s1,0x20
2a: 03c56903 lwu s2,60(a0)
2e: 1902 c.slli s2,0x20
30: 02095913 srli s2,s2,0x20
34: 04056983 lwu s3,64(a0)
38: 1982 c.slli s3,0x20
3a: 0209d993 srli s3,s3,0x20
3e: 09056a03 lwu s4,144(a0)
42: 1a02 c.slli s4,0x20
44: 020a5a13 srli s4,s4,0x20
48: 4325 c.li t1,9
4a: 006a7363 bgeu s4,t1,0x50
4e: 4a01 c.li s4,0
50: d914 c.sw a3,48(a0)
52: d958 c.sw a4,52(a0)
54: dd04 c.sw s1,56(a0)
56: 03252e23 sw s2,60(a0)
5a: 05352023 sw s3,64(a0)
5e: 4781 c.li a5,0
60: 7422 c.ldsp s0,40(sp)
62: 7482 c.ldsp s1,32(sp)
64: 6962 c.ldsp s2,24(sp)
66: 69c2 c.ldsp s3,16(sp)
68: 6a22 c.ldsp s4,8(sp)
6a: 6145 c.addi16sp sp,48
6c: 853e c.mv a0,a5
6e: 8082 c.jr ra
Signed-off-by: Luke Nelson <luke.r.nels@gmail.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Cc: Björn Töpel <bjorn.topel@gmail.com>
Link: https://lore.kernel.org/bpf/20200721025241.8077-4-luke.r.nels@gmail.com
2020-07-20 19:52:40 -07:00
|
|
|
emit_add(rd, rd, rs, ctx);
|
2019-07-04 17:18:02 -07:00
|
|
|
if (!is64 && !aux->verifier_zext)
|
2024-01-15 06:12:31 -07:00
|
|
|
emit_zextw(rd, rd, ctx);
|
bpf, riscv: add BPF JIT for RV64G
This commit adds a BPF JIT for RV64G.
The JIT is a two-pass JIT, and has a dynamic prolog/epilogue (similar
to the MIPS64 BPF JIT) instead of static ones (e.g. x86_64).
At the moment the RISC-V Linux port does not support
CONFIG_HAVE_KPROBES, which means that CONFIG_BPF_EVENTS is not
supported. Thus, no tests involving BPF_PROG_TYPE_TRACEPOINT,
BPF_PROG_TYPE_PERF_EVENT, BPF_PROG_TYPE_KPROBE and
BPF_PROG_TYPE_RAW_TRACEPOINT passes.
The implementation does not support "far branching" (>4KiB).
Test results:
# modprobe test_bpf
test_bpf: Summary: 378 PASSED, 0 FAILED, [366/366 JIT'ed]
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier
...
Summary: 761 PASSED, 507 SKIPPED, 2 FAILED
Note that "test_verifier" was run with one build with
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y and one without, otherwise
many of the the tests that require unaligned access were skipped.
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
0
No CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
59
The two failing test_verifier tests are:
"ld_abs: vlan + abs, test 1"
"ld_abs: jump around ld_abs"
This is due to that "far branching" involved in those tests.
All tests where done on QEMU (QEMU emulator version 3.1.50
(v3.1.0-688-g8ae951fbc106)).
Signed-off-by: Björn Töpel <bjorn.topel@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2019-02-05 05:41:22 -07:00
|
|
|
break;
|
|
|
|
case BPF_ALU | BPF_SUB | BPF_X:
|
|
|
|
case BPF_ALU64 | BPF_SUB | BPF_X:
|
bpf, riscv: Use compressed instructions in the rv64 JIT
This patch uses the RVC support and encodings from bpf_jit.h to optimize
the rv64 jit.
The optimizations work by replacing emit(rv_X(...)) with a call to a
helper function emit_X, which will emit a compressed version of the
instruction when possible, and when RVC is enabled.
The JIT continues to pass all tests in lib/test_bpf.c, and introduces
no new failures to test_verifier; both with and without RVC being enabled.
Most changes are straightforward replacements of emit(rv_X(...), ctx)
with emit_X(..., ctx), with the following exceptions bearing mention;
* Change emit_imm to sign-extend the value in "lower", since the
checks for RVC (and the instructions themselves) treat the value as
signed. Otherwise, small negative immediates will not be recognized as
encodable using an RVC instruction. For example, without this change,
emit_imm(rd, -1, ctx) would cause lower to become 4095, which is not a
6b int even though a "c.li rd, -1" instruction suffices.
* For {BPF_MOV,BPF_ADD} BPF_X, drop using addiw,addw in the 32-bit
cases since the values are zero-extended into the upper 32 bits in
the following instructions anyways, and the addition commutes with
zero-extension. (BPF_SUB BPF_X must still use subw since subtraction
does not commute with zero-extension.)
This patch avoids optimizing branches and jumps to use RVC instructions
since surrounding code often makes assumptions about the sizes of
emitted instructions. Optimizing these will require changing these
functions (e.g., emit_branch) to dynamically compute jump offsets.
The following are examples of the JITed code for the verifier selftest
"direct packet read test#3 for CGROUP_SKB OK", without and with RVC
enabled, respectively. The former uses 178 bytes, and the latter uses 112,
for a ~37% reduction in code size for this example.
Without RVC:
0: 02000813 addi a6,zero,32
4: fd010113 addi sp,sp,-48
8: 02813423 sd s0,40(sp)
c: 02913023 sd s1,32(sp)
10: 01213c23 sd s2,24(sp)
14: 01313823 sd s3,16(sp)
18: 01413423 sd s4,8(sp)
1c: 03010413 addi s0,sp,48
20: 03056683 lwu a3,48(a0)
24: 02069693 slli a3,a3,0x20
28: 0206d693 srli a3,a3,0x20
2c: 03456703 lwu a4,52(a0)
30: 02071713 slli a4,a4,0x20
34: 02075713 srli a4,a4,0x20
38: 03856483 lwu s1,56(a0)
3c: 02049493 slli s1,s1,0x20
40: 0204d493 srli s1,s1,0x20
44: 03c56903 lwu s2,60(a0)
48: 02091913 slli s2,s2,0x20
4c: 02095913 srli s2,s2,0x20
50: 04056983 lwu s3,64(a0)
54: 02099993 slli s3,s3,0x20
58: 0209d993 srli s3,s3,0x20
5c: 09056a03 lwu s4,144(a0)
60: 020a1a13 slli s4,s4,0x20
64: 020a5a13 srli s4,s4,0x20
68: 00900313 addi t1,zero,9
6c: 006a7463 bgeu s4,t1,0x74
70: 00000a13 addi s4,zero,0
74: 02d52823 sw a3,48(a0)
78: 02e52a23 sw a4,52(a0)
7c: 02952c23 sw s1,56(a0)
80: 03252e23 sw s2,60(a0)
84: 05352023 sw s3,64(a0)
88: 00000793 addi a5,zero,0
8c: 02813403 ld s0,40(sp)
90: 02013483 ld s1,32(sp)
94: 01813903 ld s2,24(sp)
98: 01013983 ld s3,16(sp)
9c: 00813a03 ld s4,8(sp)
a0: 03010113 addi sp,sp,48
a4: 00078513 addi a0,a5,0
a8: 00008067 jalr zero,0(ra)
With RVC:
0: 02000813 addi a6,zero,32
4: 7179 c.addi16sp sp,-48
6: f422 c.sdsp s0,40(sp)
8: f026 c.sdsp s1,32(sp)
a: ec4a c.sdsp s2,24(sp)
c: e84e c.sdsp s3,16(sp)
e: e452 c.sdsp s4,8(sp)
10: 1800 c.addi4spn s0,sp,48
12: 03056683 lwu a3,48(a0)
16: 1682 c.slli a3,0x20
18: 9281 c.srli a3,0x20
1a: 03456703 lwu a4,52(a0)
1e: 1702 c.slli a4,0x20
20: 9301 c.srli a4,0x20
22: 03856483 lwu s1,56(a0)
26: 1482 c.slli s1,0x20
28: 9081 c.srli s1,0x20
2a: 03c56903 lwu s2,60(a0)
2e: 1902 c.slli s2,0x20
30: 02095913 srli s2,s2,0x20
34: 04056983 lwu s3,64(a0)
38: 1982 c.slli s3,0x20
3a: 0209d993 srli s3,s3,0x20
3e: 09056a03 lwu s4,144(a0)
42: 1a02 c.slli s4,0x20
44: 020a5a13 srli s4,s4,0x20
48: 4325 c.li t1,9
4a: 006a7363 bgeu s4,t1,0x50
4e: 4a01 c.li s4,0
50: d914 c.sw a3,48(a0)
52: d958 c.sw a4,52(a0)
54: dd04 c.sw s1,56(a0)
56: 03252e23 sw s2,60(a0)
5a: 05352023 sw s3,64(a0)
5e: 4781 c.li a5,0
60: 7422 c.ldsp s0,40(sp)
62: 7482 c.ldsp s1,32(sp)
64: 6962 c.ldsp s2,24(sp)
66: 69c2 c.ldsp s3,16(sp)
68: 6a22 c.ldsp s4,8(sp)
6a: 6145 c.addi16sp sp,48
6c: 853e c.mv a0,a5
6e: 8082 c.jr ra
Signed-off-by: Luke Nelson <luke.r.nels@gmail.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Cc: Björn Töpel <bjorn.topel@gmail.com>
Link: https://lore.kernel.org/bpf/20200721025241.8077-4-luke.r.nels@gmail.com
2020-07-20 19:52:40 -07:00
|
|
|
if (is64)
|
|
|
|
emit_sub(rd, rd, rs, ctx);
|
|
|
|
else
|
|
|
|
emit_subw(rd, rd, rs, ctx);
|
|
|
|
|
2019-07-04 17:18:02 -07:00
|
|
|
if (!is64 && !aux->verifier_zext)
|
2024-01-15 06:12:31 -07:00
|
|
|
emit_zextw(rd, rd, ctx);
|
bpf, riscv: add BPF JIT for RV64G
This commit adds a BPF JIT for RV64G.
The JIT is a two-pass JIT, and has a dynamic prolog/epilogue (similar
to the MIPS64 BPF JIT) instead of static ones (e.g. x86_64).
At the moment the RISC-V Linux port does not support
CONFIG_HAVE_KPROBES, which means that CONFIG_BPF_EVENTS is not
supported. Thus, no tests involving BPF_PROG_TYPE_TRACEPOINT,
BPF_PROG_TYPE_PERF_EVENT, BPF_PROG_TYPE_KPROBE and
BPF_PROG_TYPE_RAW_TRACEPOINT passes.
The implementation does not support "far branching" (>4KiB).
Test results:
# modprobe test_bpf
test_bpf: Summary: 378 PASSED, 0 FAILED, [366/366 JIT'ed]
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier
...
Summary: 761 PASSED, 507 SKIPPED, 2 FAILED
Note that "test_verifier" was run with one build with
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y and one without, otherwise
many of the the tests that require unaligned access were skipped.
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
0
No CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
59
The two failing test_verifier tests are:
"ld_abs: vlan + abs, test 1"
"ld_abs: jump around ld_abs"
This is due to that "far branching" involved in those tests.
All tests where done on QEMU (QEMU emulator version 3.1.50
(v3.1.0-688-g8ae951fbc106)).
Signed-off-by: Björn Töpel <bjorn.topel@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2019-02-05 05:41:22 -07:00
|
|
|
break;
|
|
|
|
case BPF_ALU | BPF_AND | BPF_X:
|
|
|
|
case BPF_ALU64 | BPF_AND | BPF_X:
|
bpf, riscv: Use compressed instructions in the rv64 JIT
This patch uses the RVC support and encodings from bpf_jit.h to optimize
the rv64 jit.
The optimizations work by replacing emit(rv_X(...)) with a call to a
helper function emit_X, which will emit a compressed version of the
instruction when possible, and when RVC is enabled.
The JIT continues to pass all tests in lib/test_bpf.c, and introduces
no new failures to test_verifier; both with and without RVC being enabled.
Most changes are straightforward replacements of emit(rv_X(...), ctx)
with emit_X(..., ctx), with the following exceptions bearing mention;
* Change emit_imm to sign-extend the value in "lower", since the
checks for RVC (and the instructions themselves) treat the value as
signed. Otherwise, small negative immediates will not be recognized as
encodable using an RVC instruction. For example, without this change,
emit_imm(rd, -1, ctx) would cause lower to become 4095, which is not a
6b int even though a "c.li rd, -1" instruction suffices.
* For {BPF_MOV,BPF_ADD} BPF_X, drop using addiw,addw in the 32-bit
cases since the values are zero-extended into the upper 32 bits in
the following instructions anyways, and the addition commutes with
zero-extension. (BPF_SUB BPF_X must still use subw since subtraction
does not commute with zero-extension.)
This patch avoids optimizing branches and jumps to use RVC instructions
since surrounding code often makes assumptions about the sizes of
emitted instructions. Optimizing these will require changing these
functions (e.g., emit_branch) to dynamically compute jump offsets.
The following are examples of the JITed code for the verifier selftest
"direct packet read test#3 for CGROUP_SKB OK", without and with RVC
enabled, respectively. The former uses 178 bytes, and the latter uses 112,
for a ~37% reduction in code size for this example.
Without RVC:
0: 02000813 addi a6,zero,32
4: fd010113 addi sp,sp,-48
8: 02813423 sd s0,40(sp)
c: 02913023 sd s1,32(sp)
10: 01213c23 sd s2,24(sp)
14: 01313823 sd s3,16(sp)
18: 01413423 sd s4,8(sp)
1c: 03010413 addi s0,sp,48
20: 03056683 lwu a3,48(a0)
24: 02069693 slli a3,a3,0x20
28: 0206d693 srli a3,a3,0x20
2c: 03456703 lwu a4,52(a0)
30: 02071713 slli a4,a4,0x20
34: 02075713 srli a4,a4,0x20
38: 03856483 lwu s1,56(a0)
3c: 02049493 slli s1,s1,0x20
40: 0204d493 srli s1,s1,0x20
44: 03c56903 lwu s2,60(a0)
48: 02091913 slli s2,s2,0x20
4c: 02095913 srli s2,s2,0x20
50: 04056983 lwu s3,64(a0)
54: 02099993 slli s3,s3,0x20
58: 0209d993 srli s3,s3,0x20
5c: 09056a03 lwu s4,144(a0)
60: 020a1a13 slli s4,s4,0x20
64: 020a5a13 srli s4,s4,0x20
68: 00900313 addi t1,zero,9
6c: 006a7463 bgeu s4,t1,0x74
70: 00000a13 addi s4,zero,0
74: 02d52823 sw a3,48(a0)
78: 02e52a23 sw a4,52(a0)
7c: 02952c23 sw s1,56(a0)
80: 03252e23 sw s2,60(a0)
84: 05352023 sw s3,64(a0)
88: 00000793 addi a5,zero,0
8c: 02813403 ld s0,40(sp)
90: 02013483 ld s1,32(sp)
94: 01813903 ld s2,24(sp)
98: 01013983 ld s3,16(sp)
9c: 00813a03 ld s4,8(sp)
a0: 03010113 addi sp,sp,48
a4: 00078513 addi a0,a5,0
a8: 00008067 jalr zero,0(ra)
With RVC:
0: 02000813 addi a6,zero,32
4: 7179 c.addi16sp sp,-48
6: f422 c.sdsp s0,40(sp)
8: f026 c.sdsp s1,32(sp)
a: ec4a c.sdsp s2,24(sp)
c: e84e c.sdsp s3,16(sp)
e: e452 c.sdsp s4,8(sp)
10: 1800 c.addi4spn s0,sp,48
12: 03056683 lwu a3,48(a0)
16: 1682 c.slli a3,0x20
18: 9281 c.srli a3,0x20
1a: 03456703 lwu a4,52(a0)
1e: 1702 c.slli a4,0x20
20: 9301 c.srli a4,0x20
22: 03856483 lwu s1,56(a0)
26: 1482 c.slli s1,0x20
28: 9081 c.srli s1,0x20
2a: 03c56903 lwu s2,60(a0)
2e: 1902 c.slli s2,0x20
30: 02095913 srli s2,s2,0x20
34: 04056983 lwu s3,64(a0)
38: 1982 c.slli s3,0x20
3a: 0209d993 srli s3,s3,0x20
3e: 09056a03 lwu s4,144(a0)
42: 1a02 c.slli s4,0x20
44: 020a5a13 srli s4,s4,0x20
48: 4325 c.li t1,9
4a: 006a7363 bgeu s4,t1,0x50
4e: 4a01 c.li s4,0
50: d914 c.sw a3,48(a0)
52: d958 c.sw a4,52(a0)
54: dd04 c.sw s1,56(a0)
56: 03252e23 sw s2,60(a0)
5a: 05352023 sw s3,64(a0)
5e: 4781 c.li a5,0
60: 7422 c.ldsp s0,40(sp)
62: 7482 c.ldsp s1,32(sp)
64: 6962 c.ldsp s2,24(sp)
66: 69c2 c.ldsp s3,16(sp)
68: 6a22 c.ldsp s4,8(sp)
6a: 6145 c.addi16sp sp,48
6c: 853e c.mv a0,a5
6e: 8082 c.jr ra
Signed-off-by: Luke Nelson <luke.r.nels@gmail.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Cc: Björn Töpel <bjorn.topel@gmail.com>
Link: https://lore.kernel.org/bpf/20200721025241.8077-4-luke.r.nels@gmail.com
2020-07-20 19:52:40 -07:00
|
|
|
emit_and(rd, rd, rs, ctx);
|
2019-07-04 17:18:02 -07:00
|
|
|
if (!is64 && !aux->verifier_zext)
|
2024-01-15 06:12:31 -07:00
|
|
|
emit_zextw(rd, rd, ctx);
|
bpf, riscv: add BPF JIT for RV64G
This commit adds a BPF JIT for RV64G.
The JIT is a two-pass JIT, and has a dynamic prolog/epilogue (similar
to the MIPS64 BPF JIT) instead of static ones (e.g. x86_64).
At the moment the RISC-V Linux port does not support
CONFIG_HAVE_KPROBES, which means that CONFIG_BPF_EVENTS is not
supported. Thus, no tests involving BPF_PROG_TYPE_TRACEPOINT,
BPF_PROG_TYPE_PERF_EVENT, BPF_PROG_TYPE_KPROBE and
BPF_PROG_TYPE_RAW_TRACEPOINT passes.
The implementation does not support "far branching" (>4KiB).
Test results:
# modprobe test_bpf
test_bpf: Summary: 378 PASSED, 0 FAILED, [366/366 JIT'ed]
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier
...
Summary: 761 PASSED, 507 SKIPPED, 2 FAILED
Note that "test_verifier" was run with one build with
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y and one without, otherwise
many of the the tests that require unaligned access were skipped.
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
0
No CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
59
The two failing test_verifier tests are:
"ld_abs: vlan + abs, test 1"
"ld_abs: jump around ld_abs"
This is due to that "far branching" involved in those tests.
All tests where done on QEMU (QEMU emulator version 3.1.50
(v3.1.0-688-g8ae951fbc106)).
Signed-off-by: Björn Töpel <bjorn.topel@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2019-02-05 05:41:22 -07:00
|
|
|
break;
|
|
|
|
case BPF_ALU | BPF_OR | BPF_X:
|
|
|
|
case BPF_ALU64 | BPF_OR | BPF_X:
|
bpf, riscv: Use compressed instructions in the rv64 JIT
This patch uses the RVC support and encodings from bpf_jit.h to optimize
the rv64 jit.
The optimizations work by replacing emit(rv_X(...)) with a call to a
helper function emit_X, which will emit a compressed version of the
instruction when possible, and when RVC is enabled.
The JIT continues to pass all tests in lib/test_bpf.c, and introduces
no new failures to test_verifier; both with and without RVC being enabled.
Most changes are straightforward replacements of emit(rv_X(...), ctx)
with emit_X(..., ctx), with the following exceptions bearing mention;
* Change emit_imm to sign-extend the value in "lower", since the
checks for RVC (and the instructions themselves) treat the value as
signed. Otherwise, small negative immediates will not be recognized as
encodable using an RVC instruction. For example, without this change,
emit_imm(rd, -1, ctx) would cause lower to become 4095, which is not a
6b int even though a "c.li rd, -1" instruction suffices.
* For {BPF_MOV,BPF_ADD} BPF_X, drop using addiw,addw in the 32-bit
cases since the values are zero-extended into the upper 32 bits in
the following instructions anyways, and the addition commutes with
zero-extension. (BPF_SUB BPF_X must still use subw since subtraction
does not commute with zero-extension.)
This patch avoids optimizing branches and jumps to use RVC instructions
since surrounding code often makes assumptions about the sizes of
emitted instructions. Optimizing these will require changing these
functions (e.g., emit_branch) to dynamically compute jump offsets.
The following are examples of the JITed code for the verifier selftest
"direct packet read test#3 for CGROUP_SKB OK", without and with RVC
enabled, respectively. The former uses 178 bytes, and the latter uses 112,
for a ~37% reduction in code size for this example.
Without RVC:
0: 02000813 addi a6,zero,32
4: fd010113 addi sp,sp,-48
8: 02813423 sd s0,40(sp)
c: 02913023 sd s1,32(sp)
10: 01213c23 sd s2,24(sp)
14: 01313823 sd s3,16(sp)
18: 01413423 sd s4,8(sp)
1c: 03010413 addi s0,sp,48
20: 03056683 lwu a3,48(a0)
24: 02069693 slli a3,a3,0x20
28: 0206d693 srli a3,a3,0x20
2c: 03456703 lwu a4,52(a0)
30: 02071713 slli a4,a4,0x20
34: 02075713 srli a4,a4,0x20
38: 03856483 lwu s1,56(a0)
3c: 02049493 slli s1,s1,0x20
40: 0204d493 srli s1,s1,0x20
44: 03c56903 lwu s2,60(a0)
48: 02091913 slli s2,s2,0x20
4c: 02095913 srli s2,s2,0x20
50: 04056983 lwu s3,64(a0)
54: 02099993 slli s3,s3,0x20
58: 0209d993 srli s3,s3,0x20
5c: 09056a03 lwu s4,144(a0)
60: 020a1a13 slli s4,s4,0x20
64: 020a5a13 srli s4,s4,0x20
68: 00900313 addi t1,zero,9
6c: 006a7463 bgeu s4,t1,0x74
70: 00000a13 addi s4,zero,0
74: 02d52823 sw a3,48(a0)
78: 02e52a23 sw a4,52(a0)
7c: 02952c23 sw s1,56(a0)
80: 03252e23 sw s2,60(a0)
84: 05352023 sw s3,64(a0)
88: 00000793 addi a5,zero,0
8c: 02813403 ld s0,40(sp)
90: 02013483 ld s1,32(sp)
94: 01813903 ld s2,24(sp)
98: 01013983 ld s3,16(sp)
9c: 00813a03 ld s4,8(sp)
a0: 03010113 addi sp,sp,48
a4: 00078513 addi a0,a5,0
a8: 00008067 jalr zero,0(ra)
With RVC:
0: 02000813 addi a6,zero,32
4: 7179 c.addi16sp sp,-48
6: f422 c.sdsp s0,40(sp)
8: f026 c.sdsp s1,32(sp)
a: ec4a c.sdsp s2,24(sp)
c: e84e c.sdsp s3,16(sp)
e: e452 c.sdsp s4,8(sp)
10: 1800 c.addi4spn s0,sp,48
12: 03056683 lwu a3,48(a0)
16: 1682 c.slli a3,0x20
18: 9281 c.srli a3,0x20
1a: 03456703 lwu a4,52(a0)
1e: 1702 c.slli a4,0x20
20: 9301 c.srli a4,0x20
22: 03856483 lwu s1,56(a0)
26: 1482 c.slli s1,0x20
28: 9081 c.srli s1,0x20
2a: 03c56903 lwu s2,60(a0)
2e: 1902 c.slli s2,0x20
30: 02095913 srli s2,s2,0x20
34: 04056983 lwu s3,64(a0)
38: 1982 c.slli s3,0x20
3a: 0209d993 srli s3,s3,0x20
3e: 09056a03 lwu s4,144(a0)
42: 1a02 c.slli s4,0x20
44: 020a5a13 srli s4,s4,0x20
48: 4325 c.li t1,9
4a: 006a7363 bgeu s4,t1,0x50
4e: 4a01 c.li s4,0
50: d914 c.sw a3,48(a0)
52: d958 c.sw a4,52(a0)
54: dd04 c.sw s1,56(a0)
56: 03252e23 sw s2,60(a0)
5a: 05352023 sw s3,64(a0)
5e: 4781 c.li a5,0
60: 7422 c.ldsp s0,40(sp)
62: 7482 c.ldsp s1,32(sp)
64: 6962 c.ldsp s2,24(sp)
66: 69c2 c.ldsp s3,16(sp)
68: 6a22 c.ldsp s4,8(sp)
6a: 6145 c.addi16sp sp,48
6c: 853e c.mv a0,a5
6e: 8082 c.jr ra
Signed-off-by: Luke Nelson <luke.r.nels@gmail.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Cc: Björn Töpel <bjorn.topel@gmail.com>
Link: https://lore.kernel.org/bpf/20200721025241.8077-4-luke.r.nels@gmail.com
2020-07-20 19:52:40 -07:00
|
|
|
emit_or(rd, rd, rs, ctx);
|
2019-07-04 17:18:02 -07:00
|
|
|
if (!is64 && !aux->verifier_zext)
|
2024-01-15 06:12:31 -07:00
|
|
|
emit_zextw(rd, rd, ctx);
|
bpf, riscv: add BPF JIT for RV64G
This commit adds a BPF JIT for RV64G.
The JIT is a two-pass JIT, and has a dynamic prolog/epilogue (similar
to the MIPS64 BPF JIT) instead of static ones (e.g. x86_64).
At the moment the RISC-V Linux port does not support
CONFIG_HAVE_KPROBES, which means that CONFIG_BPF_EVENTS is not
supported. Thus, no tests involving BPF_PROG_TYPE_TRACEPOINT,
BPF_PROG_TYPE_PERF_EVENT, BPF_PROG_TYPE_KPROBE and
BPF_PROG_TYPE_RAW_TRACEPOINT passes.
The implementation does not support "far branching" (>4KiB).
Test results:
# modprobe test_bpf
test_bpf: Summary: 378 PASSED, 0 FAILED, [366/366 JIT'ed]
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier
...
Summary: 761 PASSED, 507 SKIPPED, 2 FAILED
Note that "test_verifier" was run with one build with
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y and one without, otherwise
many of the the tests that require unaligned access were skipped.
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
0
No CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
59
The two failing test_verifier tests are:
"ld_abs: vlan + abs, test 1"
"ld_abs: jump around ld_abs"
This is due to that "far branching" involved in those tests.
All tests where done on QEMU (QEMU emulator version 3.1.50
(v3.1.0-688-g8ae951fbc106)).
Signed-off-by: Björn Töpel <bjorn.topel@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2019-02-05 05:41:22 -07:00
|
|
|
break;
|
|
|
|
case BPF_ALU | BPF_XOR | BPF_X:
|
|
|
|
case BPF_ALU64 | BPF_XOR | BPF_X:
|
bpf, riscv: Use compressed instructions in the rv64 JIT
This patch uses the RVC support and encodings from bpf_jit.h to optimize
the rv64 jit.
The optimizations work by replacing emit(rv_X(...)) with a call to a
helper function emit_X, which will emit a compressed version of the
instruction when possible, and when RVC is enabled.
The JIT continues to pass all tests in lib/test_bpf.c, and introduces
no new failures to test_verifier; both with and without RVC being enabled.
Most changes are straightforward replacements of emit(rv_X(...), ctx)
with emit_X(..., ctx), with the following exceptions bearing mention;
* Change emit_imm to sign-extend the value in "lower", since the
checks for RVC (and the instructions themselves) treat the value as
signed. Otherwise, small negative immediates will not be recognized as
encodable using an RVC instruction. For example, without this change,
emit_imm(rd, -1, ctx) would cause lower to become 4095, which is not a
6b int even though a "c.li rd, -1" instruction suffices.
* For {BPF_MOV,BPF_ADD} BPF_X, drop using addiw,addw in the 32-bit
cases since the values are zero-extended into the upper 32 bits in
the following instructions anyways, and the addition commutes with
zero-extension. (BPF_SUB BPF_X must still use subw since subtraction
does not commute with zero-extension.)
This patch avoids optimizing branches and jumps to use RVC instructions
since surrounding code often makes assumptions about the sizes of
emitted instructions. Optimizing these will require changing these
functions (e.g., emit_branch) to dynamically compute jump offsets.
The following are examples of the JITed code for the verifier selftest
"direct packet read test#3 for CGROUP_SKB OK", without and with RVC
enabled, respectively. The former uses 178 bytes, and the latter uses 112,
for a ~37% reduction in code size for this example.
Without RVC:
0: 02000813 addi a6,zero,32
4: fd010113 addi sp,sp,-48
8: 02813423 sd s0,40(sp)
c: 02913023 sd s1,32(sp)
10: 01213c23 sd s2,24(sp)
14: 01313823 sd s3,16(sp)
18: 01413423 sd s4,8(sp)
1c: 03010413 addi s0,sp,48
20: 03056683 lwu a3,48(a0)
24: 02069693 slli a3,a3,0x20
28: 0206d693 srli a3,a3,0x20
2c: 03456703 lwu a4,52(a0)
30: 02071713 slli a4,a4,0x20
34: 02075713 srli a4,a4,0x20
38: 03856483 lwu s1,56(a0)
3c: 02049493 slli s1,s1,0x20
40: 0204d493 srli s1,s1,0x20
44: 03c56903 lwu s2,60(a0)
48: 02091913 slli s2,s2,0x20
4c: 02095913 srli s2,s2,0x20
50: 04056983 lwu s3,64(a0)
54: 02099993 slli s3,s3,0x20
58: 0209d993 srli s3,s3,0x20
5c: 09056a03 lwu s4,144(a0)
60: 020a1a13 slli s4,s4,0x20
64: 020a5a13 srli s4,s4,0x20
68: 00900313 addi t1,zero,9
6c: 006a7463 bgeu s4,t1,0x74
70: 00000a13 addi s4,zero,0
74: 02d52823 sw a3,48(a0)
78: 02e52a23 sw a4,52(a0)
7c: 02952c23 sw s1,56(a0)
80: 03252e23 sw s2,60(a0)
84: 05352023 sw s3,64(a0)
88: 00000793 addi a5,zero,0
8c: 02813403 ld s0,40(sp)
90: 02013483 ld s1,32(sp)
94: 01813903 ld s2,24(sp)
98: 01013983 ld s3,16(sp)
9c: 00813a03 ld s4,8(sp)
a0: 03010113 addi sp,sp,48
a4: 00078513 addi a0,a5,0
a8: 00008067 jalr zero,0(ra)
With RVC:
0: 02000813 addi a6,zero,32
4: 7179 c.addi16sp sp,-48
6: f422 c.sdsp s0,40(sp)
8: f026 c.sdsp s1,32(sp)
a: ec4a c.sdsp s2,24(sp)
c: e84e c.sdsp s3,16(sp)
e: e452 c.sdsp s4,8(sp)
10: 1800 c.addi4spn s0,sp,48
12: 03056683 lwu a3,48(a0)
16: 1682 c.slli a3,0x20
18: 9281 c.srli a3,0x20
1a: 03456703 lwu a4,52(a0)
1e: 1702 c.slli a4,0x20
20: 9301 c.srli a4,0x20
22: 03856483 lwu s1,56(a0)
26: 1482 c.slli s1,0x20
28: 9081 c.srli s1,0x20
2a: 03c56903 lwu s2,60(a0)
2e: 1902 c.slli s2,0x20
30: 02095913 srli s2,s2,0x20
34: 04056983 lwu s3,64(a0)
38: 1982 c.slli s3,0x20
3a: 0209d993 srli s3,s3,0x20
3e: 09056a03 lwu s4,144(a0)
42: 1a02 c.slli s4,0x20
44: 020a5a13 srli s4,s4,0x20
48: 4325 c.li t1,9
4a: 006a7363 bgeu s4,t1,0x50
4e: 4a01 c.li s4,0
50: d914 c.sw a3,48(a0)
52: d958 c.sw a4,52(a0)
54: dd04 c.sw s1,56(a0)
56: 03252e23 sw s2,60(a0)
5a: 05352023 sw s3,64(a0)
5e: 4781 c.li a5,0
60: 7422 c.ldsp s0,40(sp)
62: 7482 c.ldsp s1,32(sp)
64: 6962 c.ldsp s2,24(sp)
66: 69c2 c.ldsp s3,16(sp)
68: 6a22 c.ldsp s4,8(sp)
6a: 6145 c.addi16sp sp,48
6c: 853e c.mv a0,a5
6e: 8082 c.jr ra
Signed-off-by: Luke Nelson <luke.r.nels@gmail.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Cc: Björn Töpel <bjorn.topel@gmail.com>
Link: https://lore.kernel.org/bpf/20200721025241.8077-4-luke.r.nels@gmail.com
2020-07-20 19:52:40 -07:00
|
|
|
emit_xor(rd, rd, rs, ctx);
|
2019-07-04 17:18:02 -07:00
|
|
|
if (!is64 && !aux->verifier_zext)
|
2024-01-15 06:12:31 -07:00
|
|
|
emit_zextw(rd, rd, ctx);
|
bpf, riscv: add BPF JIT for RV64G
This commit adds a BPF JIT for RV64G.
The JIT is a two-pass JIT, and has a dynamic prolog/epilogue (similar
to the MIPS64 BPF JIT) instead of static ones (e.g. x86_64).
At the moment the RISC-V Linux port does not support
CONFIG_HAVE_KPROBES, which means that CONFIG_BPF_EVENTS is not
supported. Thus, no tests involving BPF_PROG_TYPE_TRACEPOINT,
BPF_PROG_TYPE_PERF_EVENT, BPF_PROG_TYPE_KPROBE and
BPF_PROG_TYPE_RAW_TRACEPOINT passes.
The implementation does not support "far branching" (>4KiB).
Test results:
# modprobe test_bpf
test_bpf: Summary: 378 PASSED, 0 FAILED, [366/366 JIT'ed]
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier
...
Summary: 761 PASSED, 507 SKIPPED, 2 FAILED
Note that "test_verifier" was run with one build with
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y and one without, otherwise
many of the the tests that require unaligned access were skipped.
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
0
No CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
59
The two failing test_verifier tests are:
"ld_abs: vlan + abs, test 1"
"ld_abs: jump around ld_abs"
This is due to that "far branching" involved in those tests.
All tests where done on QEMU (QEMU emulator version 3.1.50
(v3.1.0-688-g8ae951fbc106)).
Signed-off-by: Björn Töpel <bjorn.topel@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2019-02-05 05:41:22 -07:00
|
|
|
break;
|
|
|
|
case BPF_ALU | BPF_MUL | BPF_X:
|
|
|
|
case BPF_ALU64 | BPF_MUL | BPF_X:
|
|
|
|
emit(is64 ? rv_mul(rd, rd, rs) : rv_mulw(rd, rd, rs), ctx);
|
2019-05-24 15:25:27 -07:00
|
|
|
if (!is64 && !aux->verifier_zext)
|
2024-01-15 06:12:31 -07:00
|
|
|
emit_zextw(rd, rd, ctx);
|
bpf, riscv: add BPF JIT for RV64G
This commit adds a BPF JIT for RV64G.
The JIT is a two-pass JIT, and has a dynamic prolog/epilogue (similar
to the MIPS64 BPF JIT) instead of static ones (e.g. x86_64).
At the moment the RISC-V Linux port does not support
CONFIG_HAVE_KPROBES, which means that CONFIG_BPF_EVENTS is not
supported. Thus, no tests involving BPF_PROG_TYPE_TRACEPOINT,
BPF_PROG_TYPE_PERF_EVENT, BPF_PROG_TYPE_KPROBE and
BPF_PROG_TYPE_RAW_TRACEPOINT passes.
The implementation does not support "far branching" (>4KiB).
Test results:
# modprobe test_bpf
test_bpf: Summary: 378 PASSED, 0 FAILED, [366/366 JIT'ed]
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier
...
Summary: 761 PASSED, 507 SKIPPED, 2 FAILED
Note that "test_verifier" was run with one build with
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y and one without, otherwise
many of the the tests that require unaligned access were skipped.
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
0
No CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
59
The two failing test_verifier tests are:
"ld_abs: vlan + abs, test 1"
"ld_abs: jump around ld_abs"
This is due to that "far branching" involved in those tests.
All tests where done on QEMU (QEMU emulator version 3.1.50
(v3.1.0-688-g8ae951fbc106)).
Signed-off-by: Björn Töpel <bjorn.topel@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2019-02-05 05:41:22 -07:00
|
|
|
break;
|
|
|
|
case BPF_ALU | BPF_DIV | BPF_X:
|
|
|
|
case BPF_ALU64 | BPF_DIV | BPF_X:
|
2023-08-24 02:49:59 -07:00
|
|
|
if (off)
|
|
|
|
emit(is64 ? rv_div(rd, rd, rs) : rv_divw(rd, rd, rs), ctx);
|
|
|
|
else
|
|
|
|
emit(is64 ? rv_divu(rd, rd, rs) : rv_divuw(rd, rd, rs), ctx);
|
2019-05-24 15:25:27 -07:00
|
|
|
if (!is64 && !aux->verifier_zext)
|
2024-01-15 06:12:31 -07:00
|
|
|
emit_zextw(rd, rd, ctx);
|
bpf, riscv: add BPF JIT for RV64G
This commit adds a BPF JIT for RV64G.
The JIT is a two-pass JIT, and has a dynamic prolog/epilogue (similar
to the MIPS64 BPF JIT) instead of static ones (e.g. x86_64).
At the moment the RISC-V Linux port does not support
CONFIG_HAVE_KPROBES, which means that CONFIG_BPF_EVENTS is not
supported. Thus, no tests involving BPF_PROG_TYPE_TRACEPOINT,
BPF_PROG_TYPE_PERF_EVENT, BPF_PROG_TYPE_KPROBE and
BPF_PROG_TYPE_RAW_TRACEPOINT passes.
The implementation does not support "far branching" (>4KiB).
Test results:
# modprobe test_bpf
test_bpf: Summary: 378 PASSED, 0 FAILED, [366/366 JIT'ed]
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier
...
Summary: 761 PASSED, 507 SKIPPED, 2 FAILED
Note that "test_verifier" was run with one build with
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y and one without, otherwise
many of the the tests that require unaligned access were skipped.
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
0
No CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
59
The two failing test_verifier tests are:
"ld_abs: vlan + abs, test 1"
"ld_abs: jump around ld_abs"
This is due to that "far branching" involved in those tests.
All tests where done on QEMU (QEMU emulator version 3.1.50
(v3.1.0-688-g8ae951fbc106)).
Signed-off-by: Björn Töpel <bjorn.topel@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2019-02-05 05:41:22 -07:00
|
|
|
break;
|
|
|
|
case BPF_ALU | BPF_MOD | BPF_X:
|
|
|
|
case BPF_ALU64 | BPF_MOD | BPF_X:
|
2023-08-24 02:49:59 -07:00
|
|
|
if (off)
|
|
|
|
emit(is64 ? rv_rem(rd, rd, rs) : rv_remw(rd, rd, rs), ctx);
|
|
|
|
else
|
|
|
|
emit(is64 ? rv_remu(rd, rd, rs) : rv_remuw(rd, rd, rs), ctx);
|
2019-05-24 15:25:27 -07:00
|
|
|
if (!is64 && !aux->verifier_zext)
|
2024-01-15 06:12:31 -07:00
|
|
|
emit_zextw(rd, rd, ctx);
|
bpf, riscv: add BPF JIT for RV64G
This commit adds a BPF JIT for RV64G.
The JIT is a two-pass JIT, and has a dynamic prolog/epilogue (similar
to the MIPS64 BPF JIT) instead of static ones (e.g. x86_64).
At the moment the RISC-V Linux port does not support
CONFIG_HAVE_KPROBES, which means that CONFIG_BPF_EVENTS is not
supported. Thus, no tests involving BPF_PROG_TYPE_TRACEPOINT,
BPF_PROG_TYPE_PERF_EVENT, BPF_PROG_TYPE_KPROBE and
BPF_PROG_TYPE_RAW_TRACEPOINT passes.
The implementation does not support "far branching" (>4KiB).
Test results:
# modprobe test_bpf
test_bpf: Summary: 378 PASSED, 0 FAILED, [366/366 JIT'ed]
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier
...
Summary: 761 PASSED, 507 SKIPPED, 2 FAILED
Note that "test_verifier" was run with one build with
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y and one without, otherwise
many of the the tests that require unaligned access were skipped.
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
0
No CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
59
The two failing test_verifier tests are:
"ld_abs: vlan + abs, test 1"
"ld_abs: jump around ld_abs"
This is due to that "far branching" involved in those tests.
All tests where done on QEMU (QEMU emulator version 3.1.50
(v3.1.0-688-g8ae951fbc106)).
Signed-off-by: Björn Töpel <bjorn.topel@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2019-02-05 05:41:22 -07:00
|
|
|
break;
|
|
|
|
case BPF_ALU | BPF_LSH | BPF_X:
|
|
|
|
case BPF_ALU64 | BPF_LSH | BPF_X:
|
|
|
|
emit(is64 ? rv_sll(rd, rd, rs) : rv_sllw(rd, rd, rs), ctx);
|
2020-05-05 17:03:17 -07:00
|
|
|
if (!is64 && !aux->verifier_zext)
|
2024-01-15 06:12:31 -07:00
|
|
|
emit_zextw(rd, rd, ctx);
|
bpf, riscv: add BPF JIT for RV64G
This commit adds a BPF JIT for RV64G.
The JIT is a two-pass JIT, and has a dynamic prolog/epilogue (similar
to the MIPS64 BPF JIT) instead of static ones (e.g. x86_64).
At the moment the RISC-V Linux port does not support
CONFIG_HAVE_KPROBES, which means that CONFIG_BPF_EVENTS is not
supported. Thus, no tests involving BPF_PROG_TYPE_TRACEPOINT,
BPF_PROG_TYPE_PERF_EVENT, BPF_PROG_TYPE_KPROBE and
BPF_PROG_TYPE_RAW_TRACEPOINT passes.
The implementation does not support "far branching" (>4KiB).
Test results:
# modprobe test_bpf
test_bpf: Summary: 378 PASSED, 0 FAILED, [366/366 JIT'ed]
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier
...
Summary: 761 PASSED, 507 SKIPPED, 2 FAILED
Note that "test_verifier" was run with one build with
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y and one without, otherwise
many of the the tests that require unaligned access were skipped.
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
0
No CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
59
The two failing test_verifier tests are:
"ld_abs: vlan + abs, test 1"
"ld_abs: jump around ld_abs"
This is due to that "far branching" involved in those tests.
All tests where done on QEMU (QEMU emulator version 3.1.50
(v3.1.0-688-g8ae951fbc106)).
Signed-off-by: Björn Töpel <bjorn.topel@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2019-02-05 05:41:22 -07:00
|
|
|
break;
|
|
|
|
case BPF_ALU | BPF_RSH | BPF_X:
|
|
|
|
case BPF_ALU64 | BPF_RSH | BPF_X:
|
|
|
|
emit(is64 ? rv_srl(rd, rd, rs) : rv_srlw(rd, rd, rs), ctx);
|
2019-07-04 17:18:02 -07:00
|
|
|
if (!is64 && !aux->verifier_zext)
|
2024-01-15 06:12:31 -07:00
|
|
|
emit_zextw(rd, rd, ctx);
|
bpf, riscv: add BPF JIT for RV64G
This commit adds a BPF JIT for RV64G.
The JIT is a two-pass JIT, and has a dynamic prolog/epilogue (similar
to the MIPS64 BPF JIT) instead of static ones (e.g. x86_64).
At the moment the RISC-V Linux port does not support
CONFIG_HAVE_KPROBES, which means that CONFIG_BPF_EVENTS is not
supported. Thus, no tests involving BPF_PROG_TYPE_TRACEPOINT,
BPF_PROG_TYPE_PERF_EVENT, BPF_PROG_TYPE_KPROBE and
BPF_PROG_TYPE_RAW_TRACEPOINT passes.
The implementation does not support "far branching" (>4KiB).
Test results:
# modprobe test_bpf
test_bpf: Summary: 378 PASSED, 0 FAILED, [366/366 JIT'ed]
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier
...
Summary: 761 PASSED, 507 SKIPPED, 2 FAILED
Note that "test_verifier" was run with one build with
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y and one without, otherwise
many of the the tests that require unaligned access were skipped.
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
0
No CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
59
The two failing test_verifier tests are:
"ld_abs: vlan + abs, test 1"
"ld_abs: jump around ld_abs"
This is due to that "far branching" involved in those tests.
All tests where done on QEMU (QEMU emulator version 3.1.50
(v3.1.0-688-g8ae951fbc106)).
Signed-off-by: Björn Töpel <bjorn.topel@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2019-02-05 05:41:22 -07:00
|
|
|
break;
|
|
|
|
case BPF_ALU | BPF_ARSH | BPF_X:
|
|
|
|
case BPF_ALU64 | BPF_ARSH | BPF_X:
|
|
|
|
emit(is64 ? rv_sra(rd, rd, rs) : rv_sraw(rd, rd, rs), ctx);
|
2019-07-04 17:18:02 -07:00
|
|
|
if (!is64 && !aux->verifier_zext)
|
2024-01-15 06:12:31 -07:00
|
|
|
emit_zextw(rd, rd, ctx);
|
bpf, riscv: add BPF JIT for RV64G
This commit adds a BPF JIT for RV64G.
The JIT is a two-pass JIT, and has a dynamic prolog/epilogue (similar
to the MIPS64 BPF JIT) instead of static ones (e.g. x86_64).
At the moment the RISC-V Linux port does not support
CONFIG_HAVE_KPROBES, which means that CONFIG_BPF_EVENTS is not
supported. Thus, no tests involving BPF_PROG_TYPE_TRACEPOINT,
BPF_PROG_TYPE_PERF_EVENT, BPF_PROG_TYPE_KPROBE and
BPF_PROG_TYPE_RAW_TRACEPOINT passes.
The implementation does not support "far branching" (>4KiB).
Test results:
# modprobe test_bpf
test_bpf: Summary: 378 PASSED, 0 FAILED, [366/366 JIT'ed]
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier
...
Summary: 761 PASSED, 507 SKIPPED, 2 FAILED
Note that "test_verifier" was run with one build with
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y and one without, otherwise
many of the the tests that require unaligned access were skipped.
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
0
No CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
59
The two failing test_verifier tests are:
"ld_abs: vlan + abs, test 1"
"ld_abs: jump around ld_abs"
This is due to that "far branching" involved in those tests.
All tests where done on QEMU (QEMU emulator version 3.1.50
(v3.1.0-688-g8ae951fbc106)).
Signed-off-by: Björn Töpel <bjorn.topel@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2019-02-05 05:41:22 -07:00
|
|
|
break;
|
|
|
|
|
|
|
|
/* dst = -dst */
|
|
|
|
case BPF_ALU | BPF_NEG:
|
|
|
|
case BPF_ALU64 | BPF_NEG:
|
bpf, riscv: Use compressed instructions in the rv64 JIT
This patch uses the RVC support and encodings from bpf_jit.h to optimize
the rv64 jit.
The optimizations work by replacing emit(rv_X(...)) with a call to a
helper function emit_X, which will emit a compressed version of the
instruction when possible, and when RVC is enabled.
The JIT continues to pass all tests in lib/test_bpf.c, and introduces
no new failures to test_verifier; both with and without RVC being enabled.
Most changes are straightforward replacements of emit(rv_X(...), ctx)
with emit_X(..., ctx), with the following exceptions bearing mention;
* Change emit_imm to sign-extend the value in "lower", since the
checks for RVC (and the instructions themselves) treat the value as
signed. Otherwise, small negative immediates will not be recognized as
encodable using an RVC instruction. For example, without this change,
emit_imm(rd, -1, ctx) would cause lower to become 4095, which is not a
6b int even though a "c.li rd, -1" instruction suffices.
* For {BPF_MOV,BPF_ADD} BPF_X, drop using addiw,addw in the 32-bit
cases since the values are zero-extended into the upper 32 bits in
the following instructions anyways, and the addition commutes with
zero-extension. (BPF_SUB BPF_X must still use subw since subtraction
does not commute with zero-extension.)
This patch avoids optimizing branches and jumps to use RVC instructions
since surrounding code often makes assumptions about the sizes of
emitted instructions. Optimizing these will require changing these
functions (e.g., emit_branch) to dynamically compute jump offsets.
The following are examples of the JITed code for the verifier selftest
"direct packet read test#3 for CGROUP_SKB OK", without and with RVC
enabled, respectively. The former uses 178 bytes, and the latter uses 112,
for a ~37% reduction in code size for this example.
Without RVC:
0: 02000813 addi a6,zero,32
4: fd010113 addi sp,sp,-48
8: 02813423 sd s0,40(sp)
c: 02913023 sd s1,32(sp)
10: 01213c23 sd s2,24(sp)
14: 01313823 sd s3,16(sp)
18: 01413423 sd s4,8(sp)
1c: 03010413 addi s0,sp,48
20: 03056683 lwu a3,48(a0)
24: 02069693 slli a3,a3,0x20
28: 0206d693 srli a3,a3,0x20
2c: 03456703 lwu a4,52(a0)
30: 02071713 slli a4,a4,0x20
34: 02075713 srli a4,a4,0x20
38: 03856483 lwu s1,56(a0)
3c: 02049493 slli s1,s1,0x20
40: 0204d493 srli s1,s1,0x20
44: 03c56903 lwu s2,60(a0)
48: 02091913 slli s2,s2,0x20
4c: 02095913 srli s2,s2,0x20
50: 04056983 lwu s3,64(a0)
54: 02099993 slli s3,s3,0x20
58: 0209d993 srli s3,s3,0x20
5c: 09056a03 lwu s4,144(a0)
60: 020a1a13 slli s4,s4,0x20
64: 020a5a13 srli s4,s4,0x20
68: 00900313 addi t1,zero,9
6c: 006a7463 bgeu s4,t1,0x74
70: 00000a13 addi s4,zero,0
74: 02d52823 sw a3,48(a0)
78: 02e52a23 sw a4,52(a0)
7c: 02952c23 sw s1,56(a0)
80: 03252e23 sw s2,60(a0)
84: 05352023 sw s3,64(a0)
88: 00000793 addi a5,zero,0
8c: 02813403 ld s0,40(sp)
90: 02013483 ld s1,32(sp)
94: 01813903 ld s2,24(sp)
98: 01013983 ld s3,16(sp)
9c: 00813a03 ld s4,8(sp)
a0: 03010113 addi sp,sp,48
a4: 00078513 addi a0,a5,0
a8: 00008067 jalr zero,0(ra)
With RVC:
0: 02000813 addi a6,zero,32
4: 7179 c.addi16sp sp,-48
6: f422 c.sdsp s0,40(sp)
8: f026 c.sdsp s1,32(sp)
a: ec4a c.sdsp s2,24(sp)
c: e84e c.sdsp s3,16(sp)
e: e452 c.sdsp s4,8(sp)
10: 1800 c.addi4spn s0,sp,48
12: 03056683 lwu a3,48(a0)
16: 1682 c.slli a3,0x20
18: 9281 c.srli a3,0x20
1a: 03456703 lwu a4,52(a0)
1e: 1702 c.slli a4,0x20
20: 9301 c.srli a4,0x20
22: 03856483 lwu s1,56(a0)
26: 1482 c.slli s1,0x20
28: 9081 c.srli s1,0x20
2a: 03c56903 lwu s2,60(a0)
2e: 1902 c.slli s2,0x20
30: 02095913 srli s2,s2,0x20
34: 04056983 lwu s3,64(a0)
38: 1982 c.slli s3,0x20
3a: 0209d993 srli s3,s3,0x20
3e: 09056a03 lwu s4,144(a0)
42: 1a02 c.slli s4,0x20
44: 020a5a13 srli s4,s4,0x20
48: 4325 c.li t1,9
4a: 006a7363 bgeu s4,t1,0x50
4e: 4a01 c.li s4,0
50: d914 c.sw a3,48(a0)
52: d958 c.sw a4,52(a0)
54: dd04 c.sw s1,56(a0)
56: 03252e23 sw s2,60(a0)
5a: 05352023 sw s3,64(a0)
5e: 4781 c.li a5,0
60: 7422 c.ldsp s0,40(sp)
62: 7482 c.ldsp s1,32(sp)
64: 6962 c.ldsp s2,24(sp)
66: 69c2 c.ldsp s3,16(sp)
68: 6a22 c.ldsp s4,8(sp)
6a: 6145 c.addi16sp sp,48
6c: 853e c.mv a0,a5
6e: 8082 c.jr ra
Signed-off-by: Luke Nelson <luke.r.nels@gmail.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Cc: Björn Töpel <bjorn.topel@gmail.com>
Link: https://lore.kernel.org/bpf/20200721025241.8077-4-luke.r.nels@gmail.com
2020-07-20 19:52:40 -07:00
|
|
|
emit_sub(rd, RV_REG_ZERO, rd, ctx);
|
2019-07-04 17:18:02 -07:00
|
|
|
if (!is64 && !aux->verifier_zext)
|
2024-01-15 06:12:31 -07:00
|
|
|
emit_zextw(rd, rd, ctx);
|
bpf, riscv: add BPF JIT for RV64G
This commit adds a BPF JIT for RV64G.
The JIT is a two-pass JIT, and has a dynamic prolog/epilogue (similar
to the MIPS64 BPF JIT) instead of static ones (e.g. x86_64).
At the moment the RISC-V Linux port does not support
CONFIG_HAVE_KPROBES, which means that CONFIG_BPF_EVENTS is not
supported. Thus, no tests involving BPF_PROG_TYPE_TRACEPOINT,
BPF_PROG_TYPE_PERF_EVENT, BPF_PROG_TYPE_KPROBE and
BPF_PROG_TYPE_RAW_TRACEPOINT passes.
The implementation does not support "far branching" (>4KiB).
Test results:
# modprobe test_bpf
test_bpf: Summary: 378 PASSED, 0 FAILED, [366/366 JIT'ed]
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier
...
Summary: 761 PASSED, 507 SKIPPED, 2 FAILED
Note that "test_verifier" was run with one build with
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y and one without, otherwise
many of the the tests that require unaligned access were skipped.
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
0
No CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
59
The two failing test_verifier tests are:
"ld_abs: vlan + abs, test 1"
"ld_abs: jump around ld_abs"
This is due to that "far branching" involved in those tests.
All tests where done on QEMU (QEMU emulator version 3.1.50
(v3.1.0-688-g8ae951fbc106)).
Signed-off-by: Björn Töpel <bjorn.topel@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2019-02-05 05:41:22 -07:00
|
|
|
break;
|
|
|
|
|
|
|
|
/* dst = BSWAP##imm(dst) */
|
|
|
|
case BPF_ALU | BPF_END | BPF_FROM_LE:
|
2020-05-05 17:03:18 -07:00
|
|
|
switch (imm) {
|
|
|
|
case 16:
|
2024-01-15 06:12:35 -07:00
|
|
|
emit_zexth(rd, rd, ctx);
|
2020-05-05 17:03:18 -07:00
|
|
|
break;
|
|
|
|
case 32:
|
|
|
|
if (!aux->verifier_zext)
|
2024-01-15 06:12:31 -07:00
|
|
|
emit_zextw(rd, rd, ctx);
|
2020-05-05 17:03:18 -07:00
|
|
|
break;
|
|
|
|
case 64:
|
|
|
|
/* Do nothing */
|
|
|
|
break;
|
|
|
|
}
|
bpf, riscv: add BPF JIT for RV64G
This commit adds a BPF JIT for RV64G.
The JIT is a two-pass JIT, and has a dynamic prolog/epilogue (similar
to the MIPS64 BPF JIT) instead of static ones (e.g. x86_64).
At the moment the RISC-V Linux port does not support
CONFIG_HAVE_KPROBES, which means that CONFIG_BPF_EVENTS is not
supported. Thus, no tests involving BPF_PROG_TYPE_TRACEPOINT,
BPF_PROG_TYPE_PERF_EVENT, BPF_PROG_TYPE_KPROBE and
BPF_PROG_TYPE_RAW_TRACEPOINT passes.
The implementation does not support "far branching" (>4KiB).
Test results:
# modprobe test_bpf
test_bpf: Summary: 378 PASSED, 0 FAILED, [366/366 JIT'ed]
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier
...
Summary: 761 PASSED, 507 SKIPPED, 2 FAILED
Note that "test_verifier" was run with one build with
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y and one without, otherwise
many of the the tests that require unaligned access were skipped.
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
0
No CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
59
The two failing test_verifier tests are:
"ld_abs: vlan + abs, test 1"
"ld_abs: jump around ld_abs"
This is due to that "far branching" involved in those tests.
All tests where done on QEMU (QEMU emulator version 3.1.50
(v3.1.0-688-g8ae951fbc106)).
Signed-off-by: Björn Töpel <bjorn.topel@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2019-02-05 05:41:22 -07:00
|
|
|
break;
|
|
|
|
case BPF_ALU | BPF_END | BPF_FROM_BE:
|
2023-08-24 02:50:00 -07:00
|
|
|
case BPF_ALU64 | BPF_END | BPF_FROM_LE:
|
2024-01-15 06:12:35 -07:00
|
|
|
emit_bswap(rd, imm, ctx);
|
bpf, riscv: add BPF JIT for RV64G
This commit adds a BPF JIT for RV64G.
The JIT is a two-pass JIT, and has a dynamic prolog/epilogue (similar
to the MIPS64 BPF JIT) instead of static ones (e.g. x86_64).
At the moment the RISC-V Linux port does not support
CONFIG_HAVE_KPROBES, which means that CONFIG_BPF_EVENTS is not
supported. Thus, no tests involving BPF_PROG_TYPE_TRACEPOINT,
BPF_PROG_TYPE_PERF_EVENT, BPF_PROG_TYPE_KPROBE and
BPF_PROG_TYPE_RAW_TRACEPOINT passes.
The implementation does not support "far branching" (>4KiB).
Test results:
# modprobe test_bpf
test_bpf: Summary: 378 PASSED, 0 FAILED, [366/366 JIT'ed]
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier
...
Summary: 761 PASSED, 507 SKIPPED, 2 FAILED
Note that "test_verifier" was run with one build with
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y and one without, otherwise
many of the the tests that require unaligned access were skipped.
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
0
No CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
59
The two failing test_verifier tests are:
"ld_abs: vlan + abs, test 1"
"ld_abs: jump around ld_abs"
This is due to that "far branching" involved in those tests.
All tests where done on QEMU (QEMU emulator version 3.1.50
(v3.1.0-688-g8ae951fbc106)).
Signed-off-by: Björn Töpel <bjorn.topel@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2019-02-05 05:41:22 -07:00
|
|
|
break;
|
|
|
|
|
|
|
|
/* dst = imm */
|
|
|
|
case BPF_ALU | BPF_MOV | BPF_K:
|
|
|
|
case BPF_ALU64 | BPF_MOV | BPF_K:
|
|
|
|
emit_imm(rd, imm, ctx);
|
2019-05-24 15:25:27 -07:00
|
|
|
if (!is64 && !aux->verifier_zext)
|
2024-01-15 06:12:31 -07:00
|
|
|
emit_zextw(rd, rd, ctx);
|
bpf, riscv: add BPF JIT for RV64G
This commit adds a BPF JIT for RV64G.
The JIT is a two-pass JIT, and has a dynamic prolog/epilogue (similar
to the MIPS64 BPF JIT) instead of static ones (e.g. x86_64).
At the moment the RISC-V Linux port does not support
CONFIG_HAVE_KPROBES, which means that CONFIG_BPF_EVENTS is not
supported. Thus, no tests involving BPF_PROG_TYPE_TRACEPOINT,
BPF_PROG_TYPE_PERF_EVENT, BPF_PROG_TYPE_KPROBE and
BPF_PROG_TYPE_RAW_TRACEPOINT passes.
The implementation does not support "far branching" (>4KiB).
Test results:
# modprobe test_bpf
test_bpf: Summary: 378 PASSED, 0 FAILED, [366/366 JIT'ed]
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier
...
Summary: 761 PASSED, 507 SKIPPED, 2 FAILED
Note that "test_verifier" was run with one build with
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y and one without, otherwise
many of the the tests that require unaligned access were skipped.
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
0
No CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
59
The two failing test_verifier tests are:
"ld_abs: vlan + abs, test 1"
"ld_abs: jump around ld_abs"
This is due to that "far branching" involved in those tests.
All tests where done on QEMU (QEMU emulator version 3.1.50
(v3.1.0-688-g8ae951fbc106)).
Signed-off-by: Björn Töpel <bjorn.topel@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2019-02-05 05:41:22 -07:00
|
|
|
break;
|
|
|
|
|
|
|
|
/* dst = dst OP imm */
|
|
|
|
case BPF_ALU | BPF_ADD | BPF_K:
|
|
|
|
case BPF_ALU64 | BPF_ADD | BPF_K:
|
|
|
|
if (is_12b_int(imm)) {
|
bpf, riscv: Use compressed instructions in the rv64 JIT
This patch uses the RVC support and encodings from bpf_jit.h to optimize
the rv64 jit.
The optimizations work by replacing emit(rv_X(...)) with a call to a
helper function emit_X, which will emit a compressed version of the
instruction when possible, and when RVC is enabled.
The JIT continues to pass all tests in lib/test_bpf.c, and introduces
no new failures to test_verifier; both with and without RVC being enabled.
Most changes are straightforward replacements of emit(rv_X(...), ctx)
with emit_X(..., ctx), with the following exceptions bearing mention;
* Change emit_imm to sign-extend the value in "lower", since the
checks for RVC (and the instructions themselves) treat the value as
signed. Otherwise, small negative immediates will not be recognized as
encodable using an RVC instruction. For example, without this change,
emit_imm(rd, -1, ctx) would cause lower to become 4095, which is not a
6b int even though a "c.li rd, -1" instruction suffices.
* For {BPF_MOV,BPF_ADD} BPF_X, drop using addiw,addw in the 32-bit
cases since the values are zero-extended into the upper 32 bits in
the following instructions anyways, and the addition commutes with
zero-extension. (BPF_SUB BPF_X must still use subw since subtraction
does not commute with zero-extension.)
This patch avoids optimizing branches and jumps to use RVC instructions
since surrounding code often makes assumptions about the sizes of
emitted instructions. Optimizing these will require changing these
functions (e.g., emit_branch) to dynamically compute jump offsets.
The following are examples of the JITed code for the verifier selftest
"direct packet read test#3 for CGROUP_SKB OK", without and with RVC
enabled, respectively. The former uses 178 bytes, and the latter uses 112,
for a ~37% reduction in code size for this example.
Without RVC:
0: 02000813 addi a6,zero,32
4: fd010113 addi sp,sp,-48
8: 02813423 sd s0,40(sp)
c: 02913023 sd s1,32(sp)
10: 01213c23 sd s2,24(sp)
14: 01313823 sd s3,16(sp)
18: 01413423 sd s4,8(sp)
1c: 03010413 addi s0,sp,48
20: 03056683 lwu a3,48(a0)
24: 02069693 slli a3,a3,0x20
28: 0206d693 srli a3,a3,0x20
2c: 03456703 lwu a4,52(a0)
30: 02071713 slli a4,a4,0x20
34: 02075713 srli a4,a4,0x20
38: 03856483 lwu s1,56(a0)
3c: 02049493 slli s1,s1,0x20
40: 0204d493 srli s1,s1,0x20
44: 03c56903 lwu s2,60(a0)
48: 02091913 slli s2,s2,0x20
4c: 02095913 srli s2,s2,0x20
50: 04056983 lwu s3,64(a0)
54: 02099993 slli s3,s3,0x20
58: 0209d993 srli s3,s3,0x20
5c: 09056a03 lwu s4,144(a0)
60: 020a1a13 slli s4,s4,0x20
64: 020a5a13 srli s4,s4,0x20
68: 00900313 addi t1,zero,9
6c: 006a7463 bgeu s4,t1,0x74
70: 00000a13 addi s4,zero,0
74: 02d52823 sw a3,48(a0)
78: 02e52a23 sw a4,52(a0)
7c: 02952c23 sw s1,56(a0)
80: 03252e23 sw s2,60(a0)
84: 05352023 sw s3,64(a0)
88: 00000793 addi a5,zero,0
8c: 02813403 ld s0,40(sp)
90: 02013483 ld s1,32(sp)
94: 01813903 ld s2,24(sp)
98: 01013983 ld s3,16(sp)
9c: 00813a03 ld s4,8(sp)
a0: 03010113 addi sp,sp,48
a4: 00078513 addi a0,a5,0
a8: 00008067 jalr zero,0(ra)
With RVC:
0: 02000813 addi a6,zero,32
4: 7179 c.addi16sp sp,-48
6: f422 c.sdsp s0,40(sp)
8: f026 c.sdsp s1,32(sp)
a: ec4a c.sdsp s2,24(sp)
c: e84e c.sdsp s3,16(sp)
e: e452 c.sdsp s4,8(sp)
10: 1800 c.addi4spn s0,sp,48
12: 03056683 lwu a3,48(a0)
16: 1682 c.slli a3,0x20
18: 9281 c.srli a3,0x20
1a: 03456703 lwu a4,52(a0)
1e: 1702 c.slli a4,0x20
20: 9301 c.srli a4,0x20
22: 03856483 lwu s1,56(a0)
26: 1482 c.slli s1,0x20
28: 9081 c.srli s1,0x20
2a: 03c56903 lwu s2,60(a0)
2e: 1902 c.slli s2,0x20
30: 02095913 srli s2,s2,0x20
34: 04056983 lwu s3,64(a0)
38: 1982 c.slli s3,0x20
3a: 0209d993 srli s3,s3,0x20
3e: 09056a03 lwu s4,144(a0)
42: 1a02 c.slli s4,0x20
44: 020a5a13 srli s4,s4,0x20
48: 4325 c.li t1,9
4a: 006a7363 bgeu s4,t1,0x50
4e: 4a01 c.li s4,0
50: d914 c.sw a3,48(a0)
52: d958 c.sw a4,52(a0)
54: dd04 c.sw s1,56(a0)
56: 03252e23 sw s2,60(a0)
5a: 05352023 sw s3,64(a0)
5e: 4781 c.li a5,0
60: 7422 c.ldsp s0,40(sp)
62: 7482 c.ldsp s1,32(sp)
64: 6962 c.ldsp s2,24(sp)
66: 69c2 c.ldsp s3,16(sp)
68: 6a22 c.ldsp s4,8(sp)
6a: 6145 c.addi16sp sp,48
6c: 853e c.mv a0,a5
6e: 8082 c.jr ra
Signed-off-by: Luke Nelson <luke.r.nels@gmail.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Cc: Björn Töpel <bjorn.topel@gmail.com>
Link: https://lore.kernel.org/bpf/20200721025241.8077-4-luke.r.nels@gmail.com
2020-07-20 19:52:40 -07:00
|
|
|
emit_addi(rd, rd, imm, ctx);
|
bpf, riscv: add BPF JIT for RV64G
This commit adds a BPF JIT for RV64G.
The JIT is a two-pass JIT, and has a dynamic prolog/epilogue (similar
to the MIPS64 BPF JIT) instead of static ones (e.g. x86_64).
At the moment the RISC-V Linux port does not support
CONFIG_HAVE_KPROBES, which means that CONFIG_BPF_EVENTS is not
supported. Thus, no tests involving BPF_PROG_TYPE_TRACEPOINT,
BPF_PROG_TYPE_PERF_EVENT, BPF_PROG_TYPE_KPROBE and
BPF_PROG_TYPE_RAW_TRACEPOINT passes.
The implementation does not support "far branching" (>4KiB).
Test results:
# modprobe test_bpf
test_bpf: Summary: 378 PASSED, 0 FAILED, [366/366 JIT'ed]
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier
...
Summary: 761 PASSED, 507 SKIPPED, 2 FAILED
Note that "test_verifier" was run with one build with
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y and one without, otherwise
many of the the tests that require unaligned access were skipped.
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
0
No CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
59
The two failing test_verifier tests are:
"ld_abs: vlan + abs, test 1"
"ld_abs: jump around ld_abs"
This is due to that "far branching" involved in those tests.
All tests where done on QEMU (QEMU emulator version 3.1.50
(v3.1.0-688-g8ae951fbc106)).
Signed-off-by: Björn Töpel <bjorn.topel@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2019-02-05 05:41:22 -07:00
|
|
|
} else {
|
|
|
|
emit_imm(RV_REG_T1, imm, ctx);
|
bpf, riscv: Use compressed instructions in the rv64 JIT
This patch uses the RVC support and encodings from bpf_jit.h to optimize
the rv64 jit.
The optimizations work by replacing emit(rv_X(...)) with a call to a
helper function emit_X, which will emit a compressed version of the
instruction when possible, and when RVC is enabled.
The JIT continues to pass all tests in lib/test_bpf.c, and introduces
no new failures to test_verifier; both with and without RVC being enabled.
Most changes are straightforward replacements of emit(rv_X(...), ctx)
with emit_X(..., ctx), with the following exceptions bearing mention;
* Change emit_imm to sign-extend the value in "lower", since the
checks for RVC (and the instructions themselves) treat the value as
signed. Otherwise, small negative immediates will not be recognized as
encodable using an RVC instruction. For example, without this change,
emit_imm(rd, -1, ctx) would cause lower to become 4095, which is not a
6b int even though a "c.li rd, -1" instruction suffices.
* For {BPF_MOV,BPF_ADD} BPF_X, drop using addiw,addw in the 32-bit
cases since the values are zero-extended into the upper 32 bits in
the following instructions anyways, and the addition commutes with
zero-extension. (BPF_SUB BPF_X must still use subw since subtraction
does not commute with zero-extension.)
This patch avoids optimizing branches and jumps to use RVC instructions
since surrounding code often makes assumptions about the sizes of
emitted instructions. Optimizing these will require changing these
functions (e.g., emit_branch) to dynamically compute jump offsets.
The following are examples of the JITed code for the verifier selftest
"direct packet read test#3 for CGROUP_SKB OK", without and with RVC
enabled, respectively. The former uses 178 bytes, and the latter uses 112,
for a ~37% reduction in code size for this example.
Without RVC:
0: 02000813 addi a6,zero,32
4: fd010113 addi sp,sp,-48
8: 02813423 sd s0,40(sp)
c: 02913023 sd s1,32(sp)
10: 01213c23 sd s2,24(sp)
14: 01313823 sd s3,16(sp)
18: 01413423 sd s4,8(sp)
1c: 03010413 addi s0,sp,48
20: 03056683 lwu a3,48(a0)
24: 02069693 slli a3,a3,0x20
28: 0206d693 srli a3,a3,0x20
2c: 03456703 lwu a4,52(a0)
30: 02071713 slli a4,a4,0x20
34: 02075713 srli a4,a4,0x20
38: 03856483 lwu s1,56(a0)
3c: 02049493 slli s1,s1,0x20
40: 0204d493 srli s1,s1,0x20
44: 03c56903 lwu s2,60(a0)
48: 02091913 slli s2,s2,0x20
4c: 02095913 srli s2,s2,0x20
50: 04056983 lwu s3,64(a0)
54: 02099993 slli s3,s3,0x20
58: 0209d993 srli s3,s3,0x20
5c: 09056a03 lwu s4,144(a0)
60: 020a1a13 slli s4,s4,0x20
64: 020a5a13 srli s4,s4,0x20
68: 00900313 addi t1,zero,9
6c: 006a7463 bgeu s4,t1,0x74
70: 00000a13 addi s4,zero,0
74: 02d52823 sw a3,48(a0)
78: 02e52a23 sw a4,52(a0)
7c: 02952c23 sw s1,56(a0)
80: 03252e23 sw s2,60(a0)
84: 05352023 sw s3,64(a0)
88: 00000793 addi a5,zero,0
8c: 02813403 ld s0,40(sp)
90: 02013483 ld s1,32(sp)
94: 01813903 ld s2,24(sp)
98: 01013983 ld s3,16(sp)
9c: 00813a03 ld s4,8(sp)
a0: 03010113 addi sp,sp,48
a4: 00078513 addi a0,a5,0
a8: 00008067 jalr zero,0(ra)
With RVC:
0: 02000813 addi a6,zero,32
4: 7179 c.addi16sp sp,-48
6: f422 c.sdsp s0,40(sp)
8: f026 c.sdsp s1,32(sp)
a: ec4a c.sdsp s2,24(sp)
c: e84e c.sdsp s3,16(sp)
e: e452 c.sdsp s4,8(sp)
10: 1800 c.addi4spn s0,sp,48
12: 03056683 lwu a3,48(a0)
16: 1682 c.slli a3,0x20
18: 9281 c.srli a3,0x20
1a: 03456703 lwu a4,52(a0)
1e: 1702 c.slli a4,0x20
20: 9301 c.srli a4,0x20
22: 03856483 lwu s1,56(a0)
26: 1482 c.slli s1,0x20
28: 9081 c.srli s1,0x20
2a: 03c56903 lwu s2,60(a0)
2e: 1902 c.slli s2,0x20
30: 02095913 srli s2,s2,0x20
34: 04056983 lwu s3,64(a0)
38: 1982 c.slli s3,0x20
3a: 0209d993 srli s3,s3,0x20
3e: 09056a03 lwu s4,144(a0)
42: 1a02 c.slli s4,0x20
44: 020a5a13 srli s4,s4,0x20
48: 4325 c.li t1,9
4a: 006a7363 bgeu s4,t1,0x50
4e: 4a01 c.li s4,0
50: d914 c.sw a3,48(a0)
52: d958 c.sw a4,52(a0)
54: dd04 c.sw s1,56(a0)
56: 03252e23 sw s2,60(a0)
5a: 05352023 sw s3,64(a0)
5e: 4781 c.li a5,0
60: 7422 c.ldsp s0,40(sp)
62: 7482 c.ldsp s1,32(sp)
64: 6962 c.ldsp s2,24(sp)
66: 69c2 c.ldsp s3,16(sp)
68: 6a22 c.ldsp s4,8(sp)
6a: 6145 c.addi16sp sp,48
6c: 853e c.mv a0,a5
6e: 8082 c.jr ra
Signed-off-by: Luke Nelson <luke.r.nels@gmail.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Cc: Björn Töpel <bjorn.topel@gmail.com>
Link: https://lore.kernel.org/bpf/20200721025241.8077-4-luke.r.nels@gmail.com
2020-07-20 19:52:40 -07:00
|
|
|
emit_add(rd, rd, RV_REG_T1, ctx);
|
bpf, riscv: add BPF JIT for RV64G
This commit adds a BPF JIT for RV64G.
The JIT is a two-pass JIT, and has a dynamic prolog/epilogue (similar
to the MIPS64 BPF JIT) instead of static ones (e.g. x86_64).
At the moment the RISC-V Linux port does not support
CONFIG_HAVE_KPROBES, which means that CONFIG_BPF_EVENTS is not
supported. Thus, no tests involving BPF_PROG_TYPE_TRACEPOINT,
BPF_PROG_TYPE_PERF_EVENT, BPF_PROG_TYPE_KPROBE and
BPF_PROG_TYPE_RAW_TRACEPOINT passes.
The implementation does not support "far branching" (>4KiB).
Test results:
# modprobe test_bpf
test_bpf: Summary: 378 PASSED, 0 FAILED, [366/366 JIT'ed]
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier
...
Summary: 761 PASSED, 507 SKIPPED, 2 FAILED
Note that "test_verifier" was run with one build with
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y and one without, otherwise
many of the the tests that require unaligned access were skipped.
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
0
No CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
59
The two failing test_verifier tests are:
"ld_abs: vlan + abs, test 1"
"ld_abs: jump around ld_abs"
This is due to that "far branching" involved in those tests.
All tests where done on QEMU (QEMU emulator version 3.1.50
(v3.1.0-688-g8ae951fbc106)).
Signed-off-by: Björn Töpel <bjorn.topel@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2019-02-05 05:41:22 -07:00
|
|
|
}
|
2019-05-24 15:25:27 -07:00
|
|
|
if (!is64 && !aux->verifier_zext)
|
2024-01-15 06:12:31 -07:00
|
|
|
emit_zextw(rd, rd, ctx);
|
bpf, riscv: add BPF JIT for RV64G
This commit adds a BPF JIT for RV64G.
The JIT is a two-pass JIT, and has a dynamic prolog/epilogue (similar
to the MIPS64 BPF JIT) instead of static ones (e.g. x86_64).
At the moment the RISC-V Linux port does not support
CONFIG_HAVE_KPROBES, which means that CONFIG_BPF_EVENTS is not
supported. Thus, no tests involving BPF_PROG_TYPE_TRACEPOINT,
BPF_PROG_TYPE_PERF_EVENT, BPF_PROG_TYPE_KPROBE and
BPF_PROG_TYPE_RAW_TRACEPOINT passes.
The implementation does not support "far branching" (>4KiB).
Test results:
# modprobe test_bpf
test_bpf: Summary: 378 PASSED, 0 FAILED, [366/366 JIT'ed]
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier
...
Summary: 761 PASSED, 507 SKIPPED, 2 FAILED
Note that "test_verifier" was run with one build with
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y and one without, otherwise
many of the the tests that require unaligned access were skipped.
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
0
No CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
59
The two failing test_verifier tests are:
"ld_abs: vlan + abs, test 1"
"ld_abs: jump around ld_abs"
This is due to that "far branching" involved in those tests.
All tests where done on QEMU (QEMU emulator version 3.1.50
(v3.1.0-688-g8ae951fbc106)).
Signed-off-by: Björn Töpel <bjorn.topel@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2019-02-05 05:41:22 -07:00
|
|
|
break;
|
|
|
|
case BPF_ALU | BPF_SUB | BPF_K:
|
|
|
|
case BPF_ALU64 | BPF_SUB | BPF_K:
|
|
|
|
if (is_12b_int(-imm)) {
|
bpf, riscv: Use compressed instructions in the rv64 JIT
This patch uses the RVC support and encodings from bpf_jit.h to optimize
the rv64 jit.
The optimizations work by replacing emit(rv_X(...)) with a call to a
helper function emit_X, which will emit a compressed version of the
instruction when possible, and when RVC is enabled.
The JIT continues to pass all tests in lib/test_bpf.c, and introduces
no new failures to test_verifier; both with and without RVC being enabled.
Most changes are straightforward replacements of emit(rv_X(...), ctx)
with emit_X(..., ctx), with the following exceptions bearing mention;
* Change emit_imm to sign-extend the value in "lower", since the
checks for RVC (and the instructions themselves) treat the value as
signed. Otherwise, small negative immediates will not be recognized as
encodable using an RVC instruction. For example, without this change,
emit_imm(rd, -1, ctx) would cause lower to become 4095, which is not a
6b int even though a "c.li rd, -1" instruction suffices.
* For {BPF_MOV,BPF_ADD} BPF_X, drop using addiw,addw in the 32-bit
cases since the values are zero-extended into the upper 32 bits in
the following instructions anyways, and the addition commutes with
zero-extension. (BPF_SUB BPF_X must still use subw since subtraction
does not commute with zero-extension.)
This patch avoids optimizing branches and jumps to use RVC instructions
since surrounding code often makes assumptions about the sizes of
emitted instructions. Optimizing these will require changing these
functions (e.g., emit_branch) to dynamically compute jump offsets.
The following are examples of the JITed code for the verifier selftest
"direct packet read test#3 for CGROUP_SKB OK", without and with RVC
enabled, respectively. The former uses 178 bytes, and the latter uses 112,
for a ~37% reduction in code size for this example.
Without RVC:
0: 02000813 addi a6,zero,32
4: fd010113 addi sp,sp,-48
8: 02813423 sd s0,40(sp)
c: 02913023 sd s1,32(sp)
10: 01213c23 sd s2,24(sp)
14: 01313823 sd s3,16(sp)
18: 01413423 sd s4,8(sp)
1c: 03010413 addi s0,sp,48
20: 03056683 lwu a3,48(a0)
24: 02069693 slli a3,a3,0x20
28: 0206d693 srli a3,a3,0x20
2c: 03456703 lwu a4,52(a0)
30: 02071713 slli a4,a4,0x20
34: 02075713 srli a4,a4,0x20
38: 03856483 lwu s1,56(a0)
3c: 02049493 slli s1,s1,0x20
40: 0204d493 srli s1,s1,0x20
44: 03c56903 lwu s2,60(a0)
48: 02091913 slli s2,s2,0x20
4c: 02095913 srli s2,s2,0x20
50: 04056983 lwu s3,64(a0)
54: 02099993 slli s3,s3,0x20
58: 0209d993 srli s3,s3,0x20
5c: 09056a03 lwu s4,144(a0)
60: 020a1a13 slli s4,s4,0x20
64: 020a5a13 srli s4,s4,0x20
68: 00900313 addi t1,zero,9
6c: 006a7463 bgeu s4,t1,0x74
70: 00000a13 addi s4,zero,0
74: 02d52823 sw a3,48(a0)
78: 02e52a23 sw a4,52(a0)
7c: 02952c23 sw s1,56(a0)
80: 03252e23 sw s2,60(a0)
84: 05352023 sw s3,64(a0)
88: 00000793 addi a5,zero,0
8c: 02813403 ld s0,40(sp)
90: 02013483 ld s1,32(sp)
94: 01813903 ld s2,24(sp)
98: 01013983 ld s3,16(sp)
9c: 00813a03 ld s4,8(sp)
a0: 03010113 addi sp,sp,48
a4: 00078513 addi a0,a5,0
a8: 00008067 jalr zero,0(ra)
With RVC:
0: 02000813 addi a6,zero,32
4: 7179 c.addi16sp sp,-48
6: f422 c.sdsp s0,40(sp)
8: f026 c.sdsp s1,32(sp)
a: ec4a c.sdsp s2,24(sp)
c: e84e c.sdsp s3,16(sp)
e: e452 c.sdsp s4,8(sp)
10: 1800 c.addi4spn s0,sp,48
12: 03056683 lwu a3,48(a0)
16: 1682 c.slli a3,0x20
18: 9281 c.srli a3,0x20
1a: 03456703 lwu a4,52(a0)
1e: 1702 c.slli a4,0x20
20: 9301 c.srli a4,0x20
22: 03856483 lwu s1,56(a0)
26: 1482 c.slli s1,0x20
28: 9081 c.srli s1,0x20
2a: 03c56903 lwu s2,60(a0)
2e: 1902 c.slli s2,0x20
30: 02095913 srli s2,s2,0x20
34: 04056983 lwu s3,64(a0)
38: 1982 c.slli s3,0x20
3a: 0209d993 srli s3,s3,0x20
3e: 09056a03 lwu s4,144(a0)
42: 1a02 c.slli s4,0x20
44: 020a5a13 srli s4,s4,0x20
48: 4325 c.li t1,9
4a: 006a7363 bgeu s4,t1,0x50
4e: 4a01 c.li s4,0
50: d914 c.sw a3,48(a0)
52: d958 c.sw a4,52(a0)
54: dd04 c.sw s1,56(a0)
56: 03252e23 sw s2,60(a0)
5a: 05352023 sw s3,64(a0)
5e: 4781 c.li a5,0
60: 7422 c.ldsp s0,40(sp)
62: 7482 c.ldsp s1,32(sp)
64: 6962 c.ldsp s2,24(sp)
66: 69c2 c.ldsp s3,16(sp)
68: 6a22 c.ldsp s4,8(sp)
6a: 6145 c.addi16sp sp,48
6c: 853e c.mv a0,a5
6e: 8082 c.jr ra
Signed-off-by: Luke Nelson <luke.r.nels@gmail.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Cc: Björn Töpel <bjorn.topel@gmail.com>
Link: https://lore.kernel.org/bpf/20200721025241.8077-4-luke.r.nels@gmail.com
2020-07-20 19:52:40 -07:00
|
|
|
emit_addi(rd, rd, -imm, ctx);
|
bpf, riscv: add BPF JIT for RV64G
This commit adds a BPF JIT for RV64G.
The JIT is a two-pass JIT, and has a dynamic prolog/epilogue (similar
to the MIPS64 BPF JIT) instead of static ones (e.g. x86_64).
At the moment the RISC-V Linux port does not support
CONFIG_HAVE_KPROBES, which means that CONFIG_BPF_EVENTS is not
supported. Thus, no tests involving BPF_PROG_TYPE_TRACEPOINT,
BPF_PROG_TYPE_PERF_EVENT, BPF_PROG_TYPE_KPROBE and
BPF_PROG_TYPE_RAW_TRACEPOINT passes.
The implementation does not support "far branching" (>4KiB).
Test results:
# modprobe test_bpf
test_bpf: Summary: 378 PASSED, 0 FAILED, [366/366 JIT'ed]
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier
...
Summary: 761 PASSED, 507 SKIPPED, 2 FAILED
Note that "test_verifier" was run with one build with
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y and one without, otherwise
many of the the tests that require unaligned access were skipped.
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
0
No CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
59
The two failing test_verifier tests are:
"ld_abs: vlan + abs, test 1"
"ld_abs: jump around ld_abs"
This is due to that "far branching" involved in those tests.
All tests where done on QEMU (QEMU emulator version 3.1.50
(v3.1.0-688-g8ae951fbc106)).
Signed-off-by: Björn Töpel <bjorn.topel@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2019-02-05 05:41:22 -07:00
|
|
|
} else {
|
|
|
|
emit_imm(RV_REG_T1, imm, ctx);
|
bpf, riscv: Use compressed instructions in the rv64 JIT
This patch uses the RVC support and encodings from bpf_jit.h to optimize
the rv64 jit.
The optimizations work by replacing emit(rv_X(...)) with a call to a
helper function emit_X, which will emit a compressed version of the
instruction when possible, and when RVC is enabled.
The JIT continues to pass all tests in lib/test_bpf.c, and introduces
no new failures to test_verifier; both with and without RVC being enabled.
Most changes are straightforward replacements of emit(rv_X(...), ctx)
with emit_X(..., ctx), with the following exceptions bearing mention;
* Change emit_imm to sign-extend the value in "lower", since the
checks for RVC (and the instructions themselves) treat the value as
signed. Otherwise, small negative immediates will not be recognized as
encodable using an RVC instruction. For example, without this change,
emit_imm(rd, -1, ctx) would cause lower to become 4095, which is not a
6b int even though a "c.li rd, -1" instruction suffices.
* For {BPF_MOV,BPF_ADD} BPF_X, drop using addiw,addw in the 32-bit
cases since the values are zero-extended into the upper 32 bits in
the following instructions anyways, and the addition commutes with
zero-extension. (BPF_SUB BPF_X must still use subw since subtraction
does not commute with zero-extension.)
This patch avoids optimizing branches and jumps to use RVC instructions
since surrounding code often makes assumptions about the sizes of
emitted instructions. Optimizing these will require changing these
functions (e.g., emit_branch) to dynamically compute jump offsets.
The following are examples of the JITed code for the verifier selftest
"direct packet read test#3 for CGROUP_SKB OK", without and with RVC
enabled, respectively. The former uses 178 bytes, and the latter uses 112,
for a ~37% reduction in code size for this example.
Without RVC:
0: 02000813 addi a6,zero,32
4: fd010113 addi sp,sp,-48
8: 02813423 sd s0,40(sp)
c: 02913023 sd s1,32(sp)
10: 01213c23 sd s2,24(sp)
14: 01313823 sd s3,16(sp)
18: 01413423 sd s4,8(sp)
1c: 03010413 addi s0,sp,48
20: 03056683 lwu a3,48(a0)
24: 02069693 slli a3,a3,0x20
28: 0206d693 srli a3,a3,0x20
2c: 03456703 lwu a4,52(a0)
30: 02071713 slli a4,a4,0x20
34: 02075713 srli a4,a4,0x20
38: 03856483 lwu s1,56(a0)
3c: 02049493 slli s1,s1,0x20
40: 0204d493 srli s1,s1,0x20
44: 03c56903 lwu s2,60(a0)
48: 02091913 slli s2,s2,0x20
4c: 02095913 srli s2,s2,0x20
50: 04056983 lwu s3,64(a0)
54: 02099993 slli s3,s3,0x20
58: 0209d993 srli s3,s3,0x20
5c: 09056a03 lwu s4,144(a0)
60: 020a1a13 slli s4,s4,0x20
64: 020a5a13 srli s4,s4,0x20
68: 00900313 addi t1,zero,9
6c: 006a7463 bgeu s4,t1,0x74
70: 00000a13 addi s4,zero,0
74: 02d52823 sw a3,48(a0)
78: 02e52a23 sw a4,52(a0)
7c: 02952c23 sw s1,56(a0)
80: 03252e23 sw s2,60(a0)
84: 05352023 sw s3,64(a0)
88: 00000793 addi a5,zero,0
8c: 02813403 ld s0,40(sp)
90: 02013483 ld s1,32(sp)
94: 01813903 ld s2,24(sp)
98: 01013983 ld s3,16(sp)
9c: 00813a03 ld s4,8(sp)
a0: 03010113 addi sp,sp,48
a4: 00078513 addi a0,a5,0
a8: 00008067 jalr zero,0(ra)
With RVC:
0: 02000813 addi a6,zero,32
4: 7179 c.addi16sp sp,-48
6: f422 c.sdsp s0,40(sp)
8: f026 c.sdsp s1,32(sp)
a: ec4a c.sdsp s2,24(sp)
c: e84e c.sdsp s3,16(sp)
e: e452 c.sdsp s4,8(sp)
10: 1800 c.addi4spn s0,sp,48
12: 03056683 lwu a3,48(a0)
16: 1682 c.slli a3,0x20
18: 9281 c.srli a3,0x20
1a: 03456703 lwu a4,52(a0)
1e: 1702 c.slli a4,0x20
20: 9301 c.srli a4,0x20
22: 03856483 lwu s1,56(a0)
26: 1482 c.slli s1,0x20
28: 9081 c.srli s1,0x20
2a: 03c56903 lwu s2,60(a0)
2e: 1902 c.slli s2,0x20
30: 02095913 srli s2,s2,0x20
34: 04056983 lwu s3,64(a0)
38: 1982 c.slli s3,0x20
3a: 0209d993 srli s3,s3,0x20
3e: 09056a03 lwu s4,144(a0)
42: 1a02 c.slli s4,0x20
44: 020a5a13 srli s4,s4,0x20
48: 4325 c.li t1,9
4a: 006a7363 bgeu s4,t1,0x50
4e: 4a01 c.li s4,0
50: d914 c.sw a3,48(a0)
52: d958 c.sw a4,52(a0)
54: dd04 c.sw s1,56(a0)
56: 03252e23 sw s2,60(a0)
5a: 05352023 sw s3,64(a0)
5e: 4781 c.li a5,0
60: 7422 c.ldsp s0,40(sp)
62: 7482 c.ldsp s1,32(sp)
64: 6962 c.ldsp s2,24(sp)
66: 69c2 c.ldsp s3,16(sp)
68: 6a22 c.ldsp s4,8(sp)
6a: 6145 c.addi16sp sp,48
6c: 853e c.mv a0,a5
6e: 8082 c.jr ra
Signed-off-by: Luke Nelson <luke.r.nels@gmail.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Cc: Björn Töpel <bjorn.topel@gmail.com>
Link: https://lore.kernel.org/bpf/20200721025241.8077-4-luke.r.nels@gmail.com
2020-07-20 19:52:40 -07:00
|
|
|
emit_sub(rd, rd, RV_REG_T1, ctx);
|
bpf, riscv: add BPF JIT for RV64G
This commit adds a BPF JIT for RV64G.
The JIT is a two-pass JIT, and has a dynamic prolog/epilogue (similar
to the MIPS64 BPF JIT) instead of static ones (e.g. x86_64).
At the moment the RISC-V Linux port does not support
CONFIG_HAVE_KPROBES, which means that CONFIG_BPF_EVENTS is not
supported. Thus, no tests involving BPF_PROG_TYPE_TRACEPOINT,
BPF_PROG_TYPE_PERF_EVENT, BPF_PROG_TYPE_KPROBE and
BPF_PROG_TYPE_RAW_TRACEPOINT passes.
The implementation does not support "far branching" (>4KiB).
Test results:
# modprobe test_bpf
test_bpf: Summary: 378 PASSED, 0 FAILED, [366/366 JIT'ed]
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier
...
Summary: 761 PASSED, 507 SKIPPED, 2 FAILED
Note that "test_verifier" was run with one build with
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y and one without, otherwise
many of the the tests that require unaligned access were skipped.
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
0
No CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
59
The two failing test_verifier tests are:
"ld_abs: vlan + abs, test 1"
"ld_abs: jump around ld_abs"
This is due to that "far branching" involved in those tests.
All tests where done on QEMU (QEMU emulator version 3.1.50
(v3.1.0-688-g8ae951fbc106)).
Signed-off-by: Björn Töpel <bjorn.topel@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2019-02-05 05:41:22 -07:00
|
|
|
}
|
2019-05-24 15:25:27 -07:00
|
|
|
if (!is64 && !aux->verifier_zext)
|
2024-01-15 06:12:31 -07:00
|
|
|
emit_zextw(rd, rd, ctx);
|
bpf, riscv: add BPF JIT for RV64G
This commit adds a BPF JIT for RV64G.
The JIT is a two-pass JIT, and has a dynamic prolog/epilogue (similar
to the MIPS64 BPF JIT) instead of static ones (e.g. x86_64).
At the moment the RISC-V Linux port does not support
CONFIG_HAVE_KPROBES, which means that CONFIG_BPF_EVENTS is not
supported. Thus, no tests involving BPF_PROG_TYPE_TRACEPOINT,
BPF_PROG_TYPE_PERF_EVENT, BPF_PROG_TYPE_KPROBE and
BPF_PROG_TYPE_RAW_TRACEPOINT passes.
The implementation does not support "far branching" (>4KiB).
Test results:
# modprobe test_bpf
test_bpf: Summary: 378 PASSED, 0 FAILED, [366/366 JIT'ed]
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier
...
Summary: 761 PASSED, 507 SKIPPED, 2 FAILED
Note that "test_verifier" was run with one build with
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y and one without, otherwise
many of the the tests that require unaligned access were skipped.
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
0
No CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
59
The two failing test_verifier tests are:
"ld_abs: vlan + abs, test 1"
"ld_abs: jump around ld_abs"
This is due to that "far branching" involved in those tests.
All tests where done on QEMU (QEMU emulator version 3.1.50
(v3.1.0-688-g8ae951fbc106)).
Signed-off-by: Björn Töpel <bjorn.topel@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2019-02-05 05:41:22 -07:00
|
|
|
break;
|
|
|
|
case BPF_ALU | BPF_AND | BPF_K:
|
|
|
|
case BPF_ALU64 | BPF_AND | BPF_K:
|
|
|
|
if (is_12b_int(imm)) {
|
bpf, riscv: Use compressed instructions in the rv64 JIT
This patch uses the RVC support and encodings from bpf_jit.h to optimize
the rv64 jit.
The optimizations work by replacing emit(rv_X(...)) with a call to a
helper function emit_X, which will emit a compressed version of the
instruction when possible, and when RVC is enabled.
The JIT continues to pass all tests in lib/test_bpf.c, and introduces
no new failures to test_verifier; both with and without RVC being enabled.
Most changes are straightforward replacements of emit(rv_X(...), ctx)
with emit_X(..., ctx), with the following exceptions bearing mention;
* Change emit_imm to sign-extend the value in "lower", since the
checks for RVC (and the instructions themselves) treat the value as
signed. Otherwise, small negative immediates will not be recognized as
encodable using an RVC instruction. For example, without this change,
emit_imm(rd, -1, ctx) would cause lower to become 4095, which is not a
6b int even though a "c.li rd, -1" instruction suffices.
* For {BPF_MOV,BPF_ADD} BPF_X, drop using addiw,addw in the 32-bit
cases since the values are zero-extended into the upper 32 bits in
the following instructions anyways, and the addition commutes with
zero-extension. (BPF_SUB BPF_X must still use subw since subtraction
does not commute with zero-extension.)
This patch avoids optimizing branches and jumps to use RVC instructions
since surrounding code often makes assumptions about the sizes of
emitted instructions. Optimizing these will require changing these
functions (e.g., emit_branch) to dynamically compute jump offsets.
The following are examples of the JITed code for the verifier selftest
"direct packet read test#3 for CGROUP_SKB OK", without and with RVC
enabled, respectively. The former uses 178 bytes, and the latter uses 112,
for a ~37% reduction in code size for this example.
Without RVC:
0: 02000813 addi a6,zero,32
4: fd010113 addi sp,sp,-48
8: 02813423 sd s0,40(sp)
c: 02913023 sd s1,32(sp)
10: 01213c23 sd s2,24(sp)
14: 01313823 sd s3,16(sp)
18: 01413423 sd s4,8(sp)
1c: 03010413 addi s0,sp,48
20: 03056683 lwu a3,48(a0)
24: 02069693 slli a3,a3,0x20
28: 0206d693 srli a3,a3,0x20
2c: 03456703 lwu a4,52(a0)
30: 02071713 slli a4,a4,0x20
34: 02075713 srli a4,a4,0x20
38: 03856483 lwu s1,56(a0)
3c: 02049493 slli s1,s1,0x20
40: 0204d493 srli s1,s1,0x20
44: 03c56903 lwu s2,60(a0)
48: 02091913 slli s2,s2,0x20
4c: 02095913 srli s2,s2,0x20
50: 04056983 lwu s3,64(a0)
54: 02099993 slli s3,s3,0x20
58: 0209d993 srli s3,s3,0x20
5c: 09056a03 lwu s4,144(a0)
60: 020a1a13 slli s4,s4,0x20
64: 020a5a13 srli s4,s4,0x20
68: 00900313 addi t1,zero,9
6c: 006a7463 bgeu s4,t1,0x74
70: 00000a13 addi s4,zero,0
74: 02d52823 sw a3,48(a0)
78: 02e52a23 sw a4,52(a0)
7c: 02952c23 sw s1,56(a0)
80: 03252e23 sw s2,60(a0)
84: 05352023 sw s3,64(a0)
88: 00000793 addi a5,zero,0
8c: 02813403 ld s0,40(sp)
90: 02013483 ld s1,32(sp)
94: 01813903 ld s2,24(sp)
98: 01013983 ld s3,16(sp)
9c: 00813a03 ld s4,8(sp)
a0: 03010113 addi sp,sp,48
a4: 00078513 addi a0,a5,0
a8: 00008067 jalr zero,0(ra)
With RVC:
0: 02000813 addi a6,zero,32
4: 7179 c.addi16sp sp,-48
6: f422 c.sdsp s0,40(sp)
8: f026 c.sdsp s1,32(sp)
a: ec4a c.sdsp s2,24(sp)
c: e84e c.sdsp s3,16(sp)
e: e452 c.sdsp s4,8(sp)
10: 1800 c.addi4spn s0,sp,48
12: 03056683 lwu a3,48(a0)
16: 1682 c.slli a3,0x20
18: 9281 c.srli a3,0x20
1a: 03456703 lwu a4,52(a0)
1e: 1702 c.slli a4,0x20
20: 9301 c.srli a4,0x20
22: 03856483 lwu s1,56(a0)
26: 1482 c.slli s1,0x20
28: 9081 c.srli s1,0x20
2a: 03c56903 lwu s2,60(a0)
2e: 1902 c.slli s2,0x20
30: 02095913 srli s2,s2,0x20
34: 04056983 lwu s3,64(a0)
38: 1982 c.slli s3,0x20
3a: 0209d993 srli s3,s3,0x20
3e: 09056a03 lwu s4,144(a0)
42: 1a02 c.slli s4,0x20
44: 020a5a13 srli s4,s4,0x20
48: 4325 c.li t1,9
4a: 006a7363 bgeu s4,t1,0x50
4e: 4a01 c.li s4,0
50: d914 c.sw a3,48(a0)
52: d958 c.sw a4,52(a0)
54: dd04 c.sw s1,56(a0)
56: 03252e23 sw s2,60(a0)
5a: 05352023 sw s3,64(a0)
5e: 4781 c.li a5,0
60: 7422 c.ldsp s0,40(sp)
62: 7482 c.ldsp s1,32(sp)
64: 6962 c.ldsp s2,24(sp)
66: 69c2 c.ldsp s3,16(sp)
68: 6a22 c.ldsp s4,8(sp)
6a: 6145 c.addi16sp sp,48
6c: 853e c.mv a0,a5
6e: 8082 c.jr ra
Signed-off-by: Luke Nelson <luke.r.nels@gmail.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Cc: Björn Töpel <bjorn.topel@gmail.com>
Link: https://lore.kernel.org/bpf/20200721025241.8077-4-luke.r.nels@gmail.com
2020-07-20 19:52:40 -07:00
|
|
|
emit_andi(rd, rd, imm, ctx);
|
bpf, riscv: add BPF JIT for RV64G
This commit adds a BPF JIT for RV64G.
The JIT is a two-pass JIT, and has a dynamic prolog/epilogue (similar
to the MIPS64 BPF JIT) instead of static ones (e.g. x86_64).
At the moment the RISC-V Linux port does not support
CONFIG_HAVE_KPROBES, which means that CONFIG_BPF_EVENTS is not
supported. Thus, no tests involving BPF_PROG_TYPE_TRACEPOINT,
BPF_PROG_TYPE_PERF_EVENT, BPF_PROG_TYPE_KPROBE and
BPF_PROG_TYPE_RAW_TRACEPOINT passes.
The implementation does not support "far branching" (>4KiB).
Test results:
# modprobe test_bpf
test_bpf: Summary: 378 PASSED, 0 FAILED, [366/366 JIT'ed]
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier
...
Summary: 761 PASSED, 507 SKIPPED, 2 FAILED
Note that "test_verifier" was run with one build with
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y and one without, otherwise
many of the the tests that require unaligned access were skipped.
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
0
No CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
59
The two failing test_verifier tests are:
"ld_abs: vlan + abs, test 1"
"ld_abs: jump around ld_abs"
This is due to that "far branching" involved in those tests.
All tests where done on QEMU (QEMU emulator version 3.1.50
(v3.1.0-688-g8ae951fbc106)).
Signed-off-by: Björn Töpel <bjorn.topel@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2019-02-05 05:41:22 -07:00
|
|
|
} else {
|
|
|
|
emit_imm(RV_REG_T1, imm, ctx);
|
bpf, riscv: Use compressed instructions in the rv64 JIT
This patch uses the RVC support and encodings from bpf_jit.h to optimize
the rv64 jit.
The optimizations work by replacing emit(rv_X(...)) with a call to a
helper function emit_X, which will emit a compressed version of the
instruction when possible, and when RVC is enabled.
The JIT continues to pass all tests in lib/test_bpf.c, and introduces
no new failures to test_verifier; both with and without RVC being enabled.
Most changes are straightforward replacements of emit(rv_X(...), ctx)
with emit_X(..., ctx), with the following exceptions bearing mention;
* Change emit_imm to sign-extend the value in "lower", since the
checks for RVC (and the instructions themselves) treat the value as
signed. Otherwise, small negative immediates will not be recognized as
encodable using an RVC instruction. For example, without this change,
emit_imm(rd, -1, ctx) would cause lower to become 4095, which is not a
6b int even though a "c.li rd, -1" instruction suffices.
* For {BPF_MOV,BPF_ADD} BPF_X, drop using addiw,addw in the 32-bit
cases since the values are zero-extended into the upper 32 bits in
the following instructions anyways, and the addition commutes with
zero-extension. (BPF_SUB BPF_X must still use subw since subtraction
does not commute with zero-extension.)
This patch avoids optimizing branches and jumps to use RVC instructions
since surrounding code often makes assumptions about the sizes of
emitted instructions. Optimizing these will require changing these
functions (e.g., emit_branch) to dynamically compute jump offsets.
The following are examples of the JITed code for the verifier selftest
"direct packet read test#3 for CGROUP_SKB OK", without and with RVC
enabled, respectively. The former uses 178 bytes, and the latter uses 112,
for a ~37% reduction in code size for this example.
Without RVC:
0: 02000813 addi a6,zero,32
4: fd010113 addi sp,sp,-48
8: 02813423 sd s0,40(sp)
c: 02913023 sd s1,32(sp)
10: 01213c23 sd s2,24(sp)
14: 01313823 sd s3,16(sp)
18: 01413423 sd s4,8(sp)
1c: 03010413 addi s0,sp,48
20: 03056683 lwu a3,48(a0)
24: 02069693 slli a3,a3,0x20
28: 0206d693 srli a3,a3,0x20
2c: 03456703 lwu a4,52(a0)
30: 02071713 slli a4,a4,0x20
34: 02075713 srli a4,a4,0x20
38: 03856483 lwu s1,56(a0)
3c: 02049493 slli s1,s1,0x20
40: 0204d493 srli s1,s1,0x20
44: 03c56903 lwu s2,60(a0)
48: 02091913 slli s2,s2,0x20
4c: 02095913 srli s2,s2,0x20
50: 04056983 lwu s3,64(a0)
54: 02099993 slli s3,s3,0x20
58: 0209d993 srli s3,s3,0x20
5c: 09056a03 lwu s4,144(a0)
60: 020a1a13 slli s4,s4,0x20
64: 020a5a13 srli s4,s4,0x20
68: 00900313 addi t1,zero,9
6c: 006a7463 bgeu s4,t1,0x74
70: 00000a13 addi s4,zero,0
74: 02d52823 sw a3,48(a0)
78: 02e52a23 sw a4,52(a0)
7c: 02952c23 sw s1,56(a0)
80: 03252e23 sw s2,60(a0)
84: 05352023 sw s3,64(a0)
88: 00000793 addi a5,zero,0
8c: 02813403 ld s0,40(sp)
90: 02013483 ld s1,32(sp)
94: 01813903 ld s2,24(sp)
98: 01013983 ld s3,16(sp)
9c: 00813a03 ld s4,8(sp)
a0: 03010113 addi sp,sp,48
a4: 00078513 addi a0,a5,0
a8: 00008067 jalr zero,0(ra)
With RVC:
0: 02000813 addi a6,zero,32
4: 7179 c.addi16sp sp,-48
6: f422 c.sdsp s0,40(sp)
8: f026 c.sdsp s1,32(sp)
a: ec4a c.sdsp s2,24(sp)
c: e84e c.sdsp s3,16(sp)
e: e452 c.sdsp s4,8(sp)
10: 1800 c.addi4spn s0,sp,48
12: 03056683 lwu a3,48(a0)
16: 1682 c.slli a3,0x20
18: 9281 c.srli a3,0x20
1a: 03456703 lwu a4,52(a0)
1e: 1702 c.slli a4,0x20
20: 9301 c.srli a4,0x20
22: 03856483 lwu s1,56(a0)
26: 1482 c.slli s1,0x20
28: 9081 c.srli s1,0x20
2a: 03c56903 lwu s2,60(a0)
2e: 1902 c.slli s2,0x20
30: 02095913 srli s2,s2,0x20
34: 04056983 lwu s3,64(a0)
38: 1982 c.slli s3,0x20
3a: 0209d993 srli s3,s3,0x20
3e: 09056a03 lwu s4,144(a0)
42: 1a02 c.slli s4,0x20
44: 020a5a13 srli s4,s4,0x20
48: 4325 c.li t1,9
4a: 006a7363 bgeu s4,t1,0x50
4e: 4a01 c.li s4,0
50: d914 c.sw a3,48(a0)
52: d958 c.sw a4,52(a0)
54: dd04 c.sw s1,56(a0)
56: 03252e23 sw s2,60(a0)
5a: 05352023 sw s3,64(a0)
5e: 4781 c.li a5,0
60: 7422 c.ldsp s0,40(sp)
62: 7482 c.ldsp s1,32(sp)
64: 6962 c.ldsp s2,24(sp)
66: 69c2 c.ldsp s3,16(sp)
68: 6a22 c.ldsp s4,8(sp)
6a: 6145 c.addi16sp sp,48
6c: 853e c.mv a0,a5
6e: 8082 c.jr ra
Signed-off-by: Luke Nelson <luke.r.nels@gmail.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Cc: Björn Töpel <bjorn.topel@gmail.com>
Link: https://lore.kernel.org/bpf/20200721025241.8077-4-luke.r.nels@gmail.com
2020-07-20 19:52:40 -07:00
|
|
|
emit_and(rd, rd, RV_REG_T1, ctx);
|
bpf, riscv: add BPF JIT for RV64G
This commit adds a BPF JIT for RV64G.
The JIT is a two-pass JIT, and has a dynamic prolog/epilogue (similar
to the MIPS64 BPF JIT) instead of static ones (e.g. x86_64).
At the moment the RISC-V Linux port does not support
CONFIG_HAVE_KPROBES, which means that CONFIG_BPF_EVENTS is not
supported. Thus, no tests involving BPF_PROG_TYPE_TRACEPOINT,
BPF_PROG_TYPE_PERF_EVENT, BPF_PROG_TYPE_KPROBE and
BPF_PROG_TYPE_RAW_TRACEPOINT passes.
The implementation does not support "far branching" (>4KiB).
Test results:
# modprobe test_bpf
test_bpf: Summary: 378 PASSED, 0 FAILED, [366/366 JIT'ed]
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier
...
Summary: 761 PASSED, 507 SKIPPED, 2 FAILED
Note that "test_verifier" was run with one build with
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y and one without, otherwise
many of the the tests that require unaligned access were skipped.
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
0
No CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
59
The two failing test_verifier tests are:
"ld_abs: vlan + abs, test 1"
"ld_abs: jump around ld_abs"
This is due to that "far branching" involved in those tests.
All tests where done on QEMU (QEMU emulator version 3.1.50
(v3.1.0-688-g8ae951fbc106)).
Signed-off-by: Björn Töpel <bjorn.topel@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2019-02-05 05:41:22 -07:00
|
|
|
}
|
2019-05-24 15:25:27 -07:00
|
|
|
if (!is64 && !aux->verifier_zext)
|
2024-01-15 06:12:31 -07:00
|
|
|
emit_zextw(rd, rd, ctx);
|
bpf, riscv: add BPF JIT for RV64G
This commit adds a BPF JIT for RV64G.
The JIT is a two-pass JIT, and has a dynamic prolog/epilogue (similar
to the MIPS64 BPF JIT) instead of static ones (e.g. x86_64).
At the moment the RISC-V Linux port does not support
CONFIG_HAVE_KPROBES, which means that CONFIG_BPF_EVENTS is not
supported. Thus, no tests involving BPF_PROG_TYPE_TRACEPOINT,
BPF_PROG_TYPE_PERF_EVENT, BPF_PROG_TYPE_KPROBE and
BPF_PROG_TYPE_RAW_TRACEPOINT passes.
The implementation does not support "far branching" (>4KiB).
Test results:
# modprobe test_bpf
test_bpf: Summary: 378 PASSED, 0 FAILED, [366/366 JIT'ed]
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier
...
Summary: 761 PASSED, 507 SKIPPED, 2 FAILED
Note that "test_verifier" was run with one build with
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y and one without, otherwise
many of the the tests that require unaligned access were skipped.
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
0
No CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
59
The two failing test_verifier tests are:
"ld_abs: vlan + abs, test 1"
"ld_abs: jump around ld_abs"
This is due to that "far branching" involved in those tests.
All tests where done on QEMU (QEMU emulator version 3.1.50
(v3.1.0-688-g8ae951fbc106)).
Signed-off-by: Björn Töpel <bjorn.topel@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2019-02-05 05:41:22 -07:00
|
|
|
break;
|
|
|
|
case BPF_ALU | BPF_OR | BPF_K:
|
|
|
|
case BPF_ALU64 | BPF_OR | BPF_K:
|
|
|
|
if (is_12b_int(imm)) {
|
|
|
|
emit(rv_ori(rd, rd, imm), ctx);
|
|
|
|
} else {
|
|
|
|
emit_imm(RV_REG_T1, imm, ctx);
|
bpf, riscv: Use compressed instructions in the rv64 JIT
This patch uses the RVC support and encodings from bpf_jit.h to optimize
the rv64 jit.
The optimizations work by replacing emit(rv_X(...)) with a call to a
helper function emit_X, which will emit a compressed version of the
instruction when possible, and when RVC is enabled.
The JIT continues to pass all tests in lib/test_bpf.c, and introduces
no new failures to test_verifier; both with and without RVC being enabled.
Most changes are straightforward replacements of emit(rv_X(...), ctx)
with emit_X(..., ctx), with the following exceptions bearing mention;
* Change emit_imm to sign-extend the value in "lower", since the
checks for RVC (and the instructions themselves) treat the value as
signed. Otherwise, small negative immediates will not be recognized as
encodable using an RVC instruction. For example, without this change,
emit_imm(rd, -1, ctx) would cause lower to become 4095, which is not a
6b int even though a "c.li rd, -1" instruction suffices.
* For {BPF_MOV,BPF_ADD} BPF_X, drop using addiw,addw in the 32-bit
cases since the values are zero-extended into the upper 32 bits in
the following instructions anyways, and the addition commutes with
zero-extension. (BPF_SUB BPF_X must still use subw since subtraction
does not commute with zero-extension.)
This patch avoids optimizing branches and jumps to use RVC instructions
since surrounding code often makes assumptions about the sizes of
emitted instructions. Optimizing these will require changing these
functions (e.g., emit_branch) to dynamically compute jump offsets.
The following are examples of the JITed code for the verifier selftest
"direct packet read test#3 for CGROUP_SKB OK", without and with RVC
enabled, respectively. The former uses 178 bytes, and the latter uses 112,
for a ~37% reduction in code size for this example.
Without RVC:
0: 02000813 addi a6,zero,32
4: fd010113 addi sp,sp,-48
8: 02813423 sd s0,40(sp)
c: 02913023 sd s1,32(sp)
10: 01213c23 sd s2,24(sp)
14: 01313823 sd s3,16(sp)
18: 01413423 sd s4,8(sp)
1c: 03010413 addi s0,sp,48
20: 03056683 lwu a3,48(a0)
24: 02069693 slli a3,a3,0x20
28: 0206d693 srli a3,a3,0x20
2c: 03456703 lwu a4,52(a0)
30: 02071713 slli a4,a4,0x20
34: 02075713 srli a4,a4,0x20
38: 03856483 lwu s1,56(a0)
3c: 02049493 slli s1,s1,0x20
40: 0204d493 srli s1,s1,0x20
44: 03c56903 lwu s2,60(a0)
48: 02091913 slli s2,s2,0x20
4c: 02095913 srli s2,s2,0x20
50: 04056983 lwu s3,64(a0)
54: 02099993 slli s3,s3,0x20
58: 0209d993 srli s3,s3,0x20
5c: 09056a03 lwu s4,144(a0)
60: 020a1a13 slli s4,s4,0x20
64: 020a5a13 srli s4,s4,0x20
68: 00900313 addi t1,zero,9
6c: 006a7463 bgeu s4,t1,0x74
70: 00000a13 addi s4,zero,0
74: 02d52823 sw a3,48(a0)
78: 02e52a23 sw a4,52(a0)
7c: 02952c23 sw s1,56(a0)
80: 03252e23 sw s2,60(a0)
84: 05352023 sw s3,64(a0)
88: 00000793 addi a5,zero,0
8c: 02813403 ld s0,40(sp)
90: 02013483 ld s1,32(sp)
94: 01813903 ld s2,24(sp)
98: 01013983 ld s3,16(sp)
9c: 00813a03 ld s4,8(sp)
a0: 03010113 addi sp,sp,48
a4: 00078513 addi a0,a5,0
a8: 00008067 jalr zero,0(ra)
With RVC:
0: 02000813 addi a6,zero,32
4: 7179 c.addi16sp sp,-48
6: f422 c.sdsp s0,40(sp)
8: f026 c.sdsp s1,32(sp)
a: ec4a c.sdsp s2,24(sp)
c: e84e c.sdsp s3,16(sp)
e: e452 c.sdsp s4,8(sp)
10: 1800 c.addi4spn s0,sp,48
12: 03056683 lwu a3,48(a0)
16: 1682 c.slli a3,0x20
18: 9281 c.srli a3,0x20
1a: 03456703 lwu a4,52(a0)
1e: 1702 c.slli a4,0x20
20: 9301 c.srli a4,0x20
22: 03856483 lwu s1,56(a0)
26: 1482 c.slli s1,0x20
28: 9081 c.srli s1,0x20
2a: 03c56903 lwu s2,60(a0)
2e: 1902 c.slli s2,0x20
30: 02095913 srli s2,s2,0x20
34: 04056983 lwu s3,64(a0)
38: 1982 c.slli s3,0x20
3a: 0209d993 srli s3,s3,0x20
3e: 09056a03 lwu s4,144(a0)
42: 1a02 c.slli s4,0x20
44: 020a5a13 srli s4,s4,0x20
48: 4325 c.li t1,9
4a: 006a7363 bgeu s4,t1,0x50
4e: 4a01 c.li s4,0
50: d914 c.sw a3,48(a0)
52: d958 c.sw a4,52(a0)
54: dd04 c.sw s1,56(a0)
56: 03252e23 sw s2,60(a0)
5a: 05352023 sw s3,64(a0)
5e: 4781 c.li a5,0
60: 7422 c.ldsp s0,40(sp)
62: 7482 c.ldsp s1,32(sp)
64: 6962 c.ldsp s2,24(sp)
66: 69c2 c.ldsp s3,16(sp)
68: 6a22 c.ldsp s4,8(sp)
6a: 6145 c.addi16sp sp,48
6c: 853e c.mv a0,a5
6e: 8082 c.jr ra
Signed-off-by: Luke Nelson <luke.r.nels@gmail.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Cc: Björn Töpel <bjorn.topel@gmail.com>
Link: https://lore.kernel.org/bpf/20200721025241.8077-4-luke.r.nels@gmail.com
2020-07-20 19:52:40 -07:00
|
|
|
emit_or(rd, rd, RV_REG_T1, ctx);
|
bpf, riscv: add BPF JIT for RV64G
This commit adds a BPF JIT for RV64G.
The JIT is a two-pass JIT, and has a dynamic prolog/epilogue (similar
to the MIPS64 BPF JIT) instead of static ones (e.g. x86_64).
At the moment the RISC-V Linux port does not support
CONFIG_HAVE_KPROBES, which means that CONFIG_BPF_EVENTS is not
supported. Thus, no tests involving BPF_PROG_TYPE_TRACEPOINT,
BPF_PROG_TYPE_PERF_EVENT, BPF_PROG_TYPE_KPROBE and
BPF_PROG_TYPE_RAW_TRACEPOINT passes.
The implementation does not support "far branching" (>4KiB).
Test results:
# modprobe test_bpf
test_bpf: Summary: 378 PASSED, 0 FAILED, [366/366 JIT'ed]
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier
...
Summary: 761 PASSED, 507 SKIPPED, 2 FAILED
Note that "test_verifier" was run with one build with
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y and one without, otherwise
many of the the tests that require unaligned access were skipped.
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
0
No CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
59
The two failing test_verifier tests are:
"ld_abs: vlan + abs, test 1"
"ld_abs: jump around ld_abs"
This is due to that "far branching" involved in those tests.
All tests where done on QEMU (QEMU emulator version 3.1.50
(v3.1.0-688-g8ae951fbc106)).
Signed-off-by: Björn Töpel <bjorn.topel@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2019-02-05 05:41:22 -07:00
|
|
|
}
|
2019-05-24 15:25:27 -07:00
|
|
|
if (!is64 && !aux->verifier_zext)
|
2024-01-15 06:12:31 -07:00
|
|
|
emit_zextw(rd, rd, ctx);
|
bpf, riscv: add BPF JIT for RV64G
This commit adds a BPF JIT for RV64G.
The JIT is a two-pass JIT, and has a dynamic prolog/epilogue (similar
to the MIPS64 BPF JIT) instead of static ones (e.g. x86_64).
At the moment the RISC-V Linux port does not support
CONFIG_HAVE_KPROBES, which means that CONFIG_BPF_EVENTS is not
supported. Thus, no tests involving BPF_PROG_TYPE_TRACEPOINT,
BPF_PROG_TYPE_PERF_EVENT, BPF_PROG_TYPE_KPROBE and
BPF_PROG_TYPE_RAW_TRACEPOINT passes.
The implementation does not support "far branching" (>4KiB).
Test results:
# modprobe test_bpf
test_bpf: Summary: 378 PASSED, 0 FAILED, [366/366 JIT'ed]
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier
...
Summary: 761 PASSED, 507 SKIPPED, 2 FAILED
Note that "test_verifier" was run with one build with
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y and one without, otherwise
many of the the tests that require unaligned access were skipped.
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
0
No CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
59
The two failing test_verifier tests are:
"ld_abs: vlan + abs, test 1"
"ld_abs: jump around ld_abs"
This is due to that "far branching" involved in those tests.
All tests where done on QEMU (QEMU emulator version 3.1.50
(v3.1.0-688-g8ae951fbc106)).
Signed-off-by: Björn Töpel <bjorn.topel@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2019-02-05 05:41:22 -07:00
|
|
|
break;
|
|
|
|
case BPF_ALU | BPF_XOR | BPF_K:
|
|
|
|
case BPF_ALU64 | BPF_XOR | BPF_K:
|
|
|
|
if (is_12b_int(imm)) {
|
|
|
|
emit(rv_xori(rd, rd, imm), ctx);
|
|
|
|
} else {
|
|
|
|
emit_imm(RV_REG_T1, imm, ctx);
|
bpf, riscv: Use compressed instructions in the rv64 JIT
This patch uses the RVC support and encodings from bpf_jit.h to optimize
the rv64 jit.
The optimizations work by replacing emit(rv_X(...)) with a call to a
helper function emit_X, which will emit a compressed version of the
instruction when possible, and when RVC is enabled.
The JIT continues to pass all tests in lib/test_bpf.c, and introduces
no new failures to test_verifier; both with and without RVC being enabled.
Most changes are straightforward replacements of emit(rv_X(...), ctx)
with emit_X(..., ctx), with the following exceptions bearing mention;
* Change emit_imm to sign-extend the value in "lower", since the
checks for RVC (and the instructions themselves) treat the value as
signed. Otherwise, small negative immediates will not be recognized as
encodable using an RVC instruction. For example, without this change,
emit_imm(rd, -1, ctx) would cause lower to become 4095, which is not a
6b int even though a "c.li rd, -1" instruction suffices.
* For {BPF_MOV,BPF_ADD} BPF_X, drop using addiw,addw in the 32-bit
cases since the values are zero-extended into the upper 32 bits in
the following instructions anyways, and the addition commutes with
zero-extension. (BPF_SUB BPF_X must still use subw since subtraction
does not commute with zero-extension.)
This patch avoids optimizing branches and jumps to use RVC instructions
since surrounding code often makes assumptions about the sizes of
emitted instructions. Optimizing these will require changing these
functions (e.g., emit_branch) to dynamically compute jump offsets.
The following are examples of the JITed code for the verifier selftest
"direct packet read test#3 for CGROUP_SKB OK", without and with RVC
enabled, respectively. The former uses 178 bytes, and the latter uses 112,
for a ~37% reduction in code size for this example.
Without RVC:
0: 02000813 addi a6,zero,32
4: fd010113 addi sp,sp,-48
8: 02813423 sd s0,40(sp)
c: 02913023 sd s1,32(sp)
10: 01213c23 sd s2,24(sp)
14: 01313823 sd s3,16(sp)
18: 01413423 sd s4,8(sp)
1c: 03010413 addi s0,sp,48
20: 03056683 lwu a3,48(a0)
24: 02069693 slli a3,a3,0x20
28: 0206d693 srli a3,a3,0x20
2c: 03456703 lwu a4,52(a0)
30: 02071713 slli a4,a4,0x20
34: 02075713 srli a4,a4,0x20
38: 03856483 lwu s1,56(a0)
3c: 02049493 slli s1,s1,0x20
40: 0204d493 srli s1,s1,0x20
44: 03c56903 lwu s2,60(a0)
48: 02091913 slli s2,s2,0x20
4c: 02095913 srli s2,s2,0x20
50: 04056983 lwu s3,64(a0)
54: 02099993 slli s3,s3,0x20
58: 0209d993 srli s3,s3,0x20
5c: 09056a03 lwu s4,144(a0)
60: 020a1a13 slli s4,s4,0x20
64: 020a5a13 srli s4,s4,0x20
68: 00900313 addi t1,zero,9
6c: 006a7463 bgeu s4,t1,0x74
70: 00000a13 addi s4,zero,0
74: 02d52823 sw a3,48(a0)
78: 02e52a23 sw a4,52(a0)
7c: 02952c23 sw s1,56(a0)
80: 03252e23 sw s2,60(a0)
84: 05352023 sw s3,64(a0)
88: 00000793 addi a5,zero,0
8c: 02813403 ld s0,40(sp)
90: 02013483 ld s1,32(sp)
94: 01813903 ld s2,24(sp)
98: 01013983 ld s3,16(sp)
9c: 00813a03 ld s4,8(sp)
a0: 03010113 addi sp,sp,48
a4: 00078513 addi a0,a5,0
a8: 00008067 jalr zero,0(ra)
With RVC:
0: 02000813 addi a6,zero,32
4: 7179 c.addi16sp sp,-48
6: f422 c.sdsp s0,40(sp)
8: f026 c.sdsp s1,32(sp)
a: ec4a c.sdsp s2,24(sp)
c: e84e c.sdsp s3,16(sp)
e: e452 c.sdsp s4,8(sp)
10: 1800 c.addi4spn s0,sp,48
12: 03056683 lwu a3,48(a0)
16: 1682 c.slli a3,0x20
18: 9281 c.srli a3,0x20
1a: 03456703 lwu a4,52(a0)
1e: 1702 c.slli a4,0x20
20: 9301 c.srli a4,0x20
22: 03856483 lwu s1,56(a0)
26: 1482 c.slli s1,0x20
28: 9081 c.srli s1,0x20
2a: 03c56903 lwu s2,60(a0)
2e: 1902 c.slli s2,0x20
30: 02095913 srli s2,s2,0x20
34: 04056983 lwu s3,64(a0)
38: 1982 c.slli s3,0x20
3a: 0209d993 srli s3,s3,0x20
3e: 09056a03 lwu s4,144(a0)
42: 1a02 c.slli s4,0x20
44: 020a5a13 srli s4,s4,0x20
48: 4325 c.li t1,9
4a: 006a7363 bgeu s4,t1,0x50
4e: 4a01 c.li s4,0
50: d914 c.sw a3,48(a0)
52: d958 c.sw a4,52(a0)
54: dd04 c.sw s1,56(a0)
56: 03252e23 sw s2,60(a0)
5a: 05352023 sw s3,64(a0)
5e: 4781 c.li a5,0
60: 7422 c.ldsp s0,40(sp)
62: 7482 c.ldsp s1,32(sp)
64: 6962 c.ldsp s2,24(sp)
66: 69c2 c.ldsp s3,16(sp)
68: 6a22 c.ldsp s4,8(sp)
6a: 6145 c.addi16sp sp,48
6c: 853e c.mv a0,a5
6e: 8082 c.jr ra
Signed-off-by: Luke Nelson <luke.r.nels@gmail.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Cc: Björn Töpel <bjorn.topel@gmail.com>
Link: https://lore.kernel.org/bpf/20200721025241.8077-4-luke.r.nels@gmail.com
2020-07-20 19:52:40 -07:00
|
|
|
emit_xor(rd, rd, RV_REG_T1, ctx);
|
bpf, riscv: add BPF JIT for RV64G
This commit adds a BPF JIT for RV64G.
The JIT is a two-pass JIT, and has a dynamic prolog/epilogue (similar
to the MIPS64 BPF JIT) instead of static ones (e.g. x86_64).
At the moment the RISC-V Linux port does not support
CONFIG_HAVE_KPROBES, which means that CONFIG_BPF_EVENTS is not
supported. Thus, no tests involving BPF_PROG_TYPE_TRACEPOINT,
BPF_PROG_TYPE_PERF_EVENT, BPF_PROG_TYPE_KPROBE and
BPF_PROG_TYPE_RAW_TRACEPOINT passes.
The implementation does not support "far branching" (>4KiB).
Test results:
# modprobe test_bpf
test_bpf: Summary: 378 PASSED, 0 FAILED, [366/366 JIT'ed]
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier
...
Summary: 761 PASSED, 507 SKIPPED, 2 FAILED
Note that "test_verifier" was run with one build with
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y and one without, otherwise
many of the the tests that require unaligned access were skipped.
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
0
No CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
59
The two failing test_verifier tests are:
"ld_abs: vlan + abs, test 1"
"ld_abs: jump around ld_abs"
This is due to that "far branching" involved in those tests.
All tests where done on QEMU (QEMU emulator version 3.1.50
(v3.1.0-688-g8ae951fbc106)).
Signed-off-by: Björn Töpel <bjorn.topel@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2019-02-05 05:41:22 -07:00
|
|
|
}
|
2019-05-24 15:25:27 -07:00
|
|
|
if (!is64 && !aux->verifier_zext)
|
2024-01-15 06:12:31 -07:00
|
|
|
emit_zextw(rd, rd, ctx);
|
bpf, riscv: add BPF JIT for RV64G
This commit adds a BPF JIT for RV64G.
The JIT is a two-pass JIT, and has a dynamic prolog/epilogue (similar
to the MIPS64 BPF JIT) instead of static ones (e.g. x86_64).
At the moment the RISC-V Linux port does not support
CONFIG_HAVE_KPROBES, which means that CONFIG_BPF_EVENTS is not
supported. Thus, no tests involving BPF_PROG_TYPE_TRACEPOINT,
BPF_PROG_TYPE_PERF_EVENT, BPF_PROG_TYPE_KPROBE and
BPF_PROG_TYPE_RAW_TRACEPOINT passes.
The implementation does not support "far branching" (>4KiB).
Test results:
# modprobe test_bpf
test_bpf: Summary: 378 PASSED, 0 FAILED, [366/366 JIT'ed]
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier
...
Summary: 761 PASSED, 507 SKIPPED, 2 FAILED
Note that "test_verifier" was run with one build with
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y and one without, otherwise
many of the the tests that require unaligned access were skipped.
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
0
No CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
59
The two failing test_verifier tests are:
"ld_abs: vlan + abs, test 1"
"ld_abs: jump around ld_abs"
This is due to that "far branching" involved in those tests.
All tests where done on QEMU (QEMU emulator version 3.1.50
(v3.1.0-688-g8ae951fbc106)).
Signed-off-by: Björn Töpel <bjorn.topel@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2019-02-05 05:41:22 -07:00
|
|
|
break;
|
|
|
|
case BPF_ALU | BPF_MUL | BPF_K:
|
|
|
|
case BPF_ALU64 | BPF_MUL | BPF_K:
|
|
|
|
emit_imm(RV_REG_T1, imm, ctx);
|
|
|
|
emit(is64 ? rv_mul(rd, rd, RV_REG_T1) :
|
|
|
|
rv_mulw(rd, rd, RV_REG_T1), ctx);
|
2019-05-24 15:25:27 -07:00
|
|
|
if (!is64 && !aux->verifier_zext)
|
2024-01-15 06:12:31 -07:00
|
|
|
emit_zextw(rd, rd, ctx);
|
bpf, riscv: add BPF JIT for RV64G
This commit adds a BPF JIT for RV64G.
The JIT is a two-pass JIT, and has a dynamic prolog/epilogue (similar
to the MIPS64 BPF JIT) instead of static ones (e.g. x86_64).
At the moment the RISC-V Linux port does not support
CONFIG_HAVE_KPROBES, which means that CONFIG_BPF_EVENTS is not
supported. Thus, no tests involving BPF_PROG_TYPE_TRACEPOINT,
BPF_PROG_TYPE_PERF_EVENT, BPF_PROG_TYPE_KPROBE and
BPF_PROG_TYPE_RAW_TRACEPOINT passes.
The implementation does not support "far branching" (>4KiB).
Test results:
# modprobe test_bpf
test_bpf: Summary: 378 PASSED, 0 FAILED, [366/366 JIT'ed]
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier
...
Summary: 761 PASSED, 507 SKIPPED, 2 FAILED
Note that "test_verifier" was run with one build with
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y and one without, otherwise
many of the the tests that require unaligned access were skipped.
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
0
No CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
59
The two failing test_verifier tests are:
"ld_abs: vlan + abs, test 1"
"ld_abs: jump around ld_abs"
This is due to that "far branching" involved in those tests.
All tests where done on QEMU (QEMU emulator version 3.1.50
(v3.1.0-688-g8ae951fbc106)).
Signed-off-by: Björn Töpel <bjorn.topel@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2019-02-05 05:41:22 -07:00
|
|
|
break;
|
|
|
|
case BPF_ALU | BPF_DIV | BPF_K:
|
|
|
|
case BPF_ALU64 | BPF_DIV | BPF_K:
|
|
|
|
emit_imm(RV_REG_T1, imm, ctx);
|
2023-08-24 02:49:59 -07:00
|
|
|
if (off)
|
|
|
|
emit(is64 ? rv_div(rd, rd, RV_REG_T1) :
|
|
|
|
rv_divw(rd, rd, RV_REG_T1), ctx);
|
|
|
|
else
|
|
|
|
emit(is64 ? rv_divu(rd, rd, RV_REG_T1) :
|
|
|
|
rv_divuw(rd, rd, RV_REG_T1), ctx);
|
2019-05-24 15:25:27 -07:00
|
|
|
if (!is64 && !aux->verifier_zext)
|
2024-01-15 06:12:31 -07:00
|
|
|
emit_zextw(rd, rd, ctx);
|
bpf, riscv: add BPF JIT for RV64G
This commit adds a BPF JIT for RV64G.
The JIT is a two-pass JIT, and has a dynamic prolog/epilogue (similar
to the MIPS64 BPF JIT) instead of static ones (e.g. x86_64).
At the moment the RISC-V Linux port does not support
CONFIG_HAVE_KPROBES, which means that CONFIG_BPF_EVENTS is not
supported. Thus, no tests involving BPF_PROG_TYPE_TRACEPOINT,
BPF_PROG_TYPE_PERF_EVENT, BPF_PROG_TYPE_KPROBE and
BPF_PROG_TYPE_RAW_TRACEPOINT passes.
The implementation does not support "far branching" (>4KiB).
Test results:
# modprobe test_bpf
test_bpf: Summary: 378 PASSED, 0 FAILED, [366/366 JIT'ed]
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier
...
Summary: 761 PASSED, 507 SKIPPED, 2 FAILED
Note that "test_verifier" was run with one build with
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y and one without, otherwise
many of the the tests that require unaligned access were skipped.
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
0
No CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
59
The two failing test_verifier tests are:
"ld_abs: vlan + abs, test 1"
"ld_abs: jump around ld_abs"
This is due to that "far branching" involved in those tests.
All tests where done on QEMU (QEMU emulator version 3.1.50
(v3.1.0-688-g8ae951fbc106)).
Signed-off-by: Björn Töpel <bjorn.topel@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2019-02-05 05:41:22 -07:00
|
|
|
break;
|
|
|
|
case BPF_ALU | BPF_MOD | BPF_K:
|
|
|
|
case BPF_ALU64 | BPF_MOD | BPF_K:
|
|
|
|
emit_imm(RV_REG_T1, imm, ctx);
|
2023-08-24 02:49:59 -07:00
|
|
|
if (off)
|
|
|
|
emit(is64 ? rv_rem(rd, rd, RV_REG_T1) :
|
|
|
|
rv_remw(rd, rd, RV_REG_T1), ctx);
|
|
|
|
else
|
|
|
|
emit(is64 ? rv_remu(rd, rd, RV_REG_T1) :
|
|
|
|
rv_remuw(rd, rd, RV_REG_T1), ctx);
|
2019-05-24 15:25:27 -07:00
|
|
|
if (!is64 && !aux->verifier_zext)
|
2024-01-15 06:12:31 -07:00
|
|
|
emit_zextw(rd, rd, ctx);
|
bpf, riscv: add BPF JIT for RV64G
This commit adds a BPF JIT for RV64G.
The JIT is a two-pass JIT, and has a dynamic prolog/epilogue (similar
to the MIPS64 BPF JIT) instead of static ones (e.g. x86_64).
At the moment the RISC-V Linux port does not support
CONFIG_HAVE_KPROBES, which means that CONFIG_BPF_EVENTS is not
supported. Thus, no tests involving BPF_PROG_TYPE_TRACEPOINT,
BPF_PROG_TYPE_PERF_EVENT, BPF_PROG_TYPE_KPROBE and
BPF_PROG_TYPE_RAW_TRACEPOINT passes.
The implementation does not support "far branching" (>4KiB).
Test results:
# modprobe test_bpf
test_bpf: Summary: 378 PASSED, 0 FAILED, [366/366 JIT'ed]
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier
...
Summary: 761 PASSED, 507 SKIPPED, 2 FAILED
Note that "test_verifier" was run with one build with
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y and one without, otherwise
many of the the tests that require unaligned access were skipped.
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
0
No CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
59
The two failing test_verifier tests are:
"ld_abs: vlan + abs, test 1"
"ld_abs: jump around ld_abs"
This is due to that "far branching" involved in those tests.
All tests where done on QEMU (QEMU emulator version 3.1.50
(v3.1.0-688-g8ae951fbc106)).
Signed-off-by: Björn Töpel <bjorn.topel@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2019-02-05 05:41:22 -07:00
|
|
|
break;
|
|
|
|
case BPF_ALU | BPF_LSH | BPF_K:
|
|
|
|
case BPF_ALU64 | BPF_LSH | BPF_K:
|
bpf, riscv: Use compressed instructions in the rv64 JIT
This patch uses the RVC support and encodings from bpf_jit.h to optimize
the rv64 jit.
The optimizations work by replacing emit(rv_X(...)) with a call to a
helper function emit_X, which will emit a compressed version of the
instruction when possible, and when RVC is enabled.
The JIT continues to pass all tests in lib/test_bpf.c, and introduces
no new failures to test_verifier; both with and without RVC being enabled.
Most changes are straightforward replacements of emit(rv_X(...), ctx)
with emit_X(..., ctx), with the following exceptions bearing mention;
* Change emit_imm to sign-extend the value in "lower", since the
checks for RVC (and the instructions themselves) treat the value as
signed. Otherwise, small negative immediates will not be recognized as
encodable using an RVC instruction. For example, without this change,
emit_imm(rd, -1, ctx) would cause lower to become 4095, which is not a
6b int even though a "c.li rd, -1" instruction suffices.
* For {BPF_MOV,BPF_ADD} BPF_X, drop using addiw,addw in the 32-bit
cases since the values are zero-extended into the upper 32 bits in
the following instructions anyways, and the addition commutes with
zero-extension. (BPF_SUB BPF_X must still use subw since subtraction
does not commute with zero-extension.)
This patch avoids optimizing branches and jumps to use RVC instructions
since surrounding code often makes assumptions about the sizes of
emitted instructions. Optimizing these will require changing these
functions (e.g., emit_branch) to dynamically compute jump offsets.
The following are examples of the JITed code for the verifier selftest
"direct packet read test#3 for CGROUP_SKB OK", without and with RVC
enabled, respectively. The former uses 178 bytes, and the latter uses 112,
for a ~37% reduction in code size for this example.
Without RVC:
0: 02000813 addi a6,zero,32
4: fd010113 addi sp,sp,-48
8: 02813423 sd s0,40(sp)
c: 02913023 sd s1,32(sp)
10: 01213c23 sd s2,24(sp)
14: 01313823 sd s3,16(sp)
18: 01413423 sd s4,8(sp)
1c: 03010413 addi s0,sp,48
20: 03056683 lwu a3,48(a0)
24: 02069693 slli a3,a3,0x20
28: 0206d693 srli a3,a3,0x20
2c: 03456703 lwu a4,52(a0)
30: 02071713 slli a4,a4,0x20
34: 02075713 srli a4,a4,0x20
38: 03856483 lwu s1,56(a0)
3c: 02049493 slli s1,s1,0x20
40: 0204d493 srli s1,s1,0x20
44: 03c56903 lwu s2,60(a0)
48: 02091913 slli s2,s2,0x20
4c: 02095913 srli s2,s2,0x20
50: 04056983 lwu s3,64(a0)
54: 02099993 slli s3,s3,0x20
58: 0209d993 srli s3,s3,0x20
5c: 09056a03 lwu s4,144(a0)
60: 020a1a13 slli s4,s4,0x20
64: 020a5a13 srli s4,s4,0x20
68: 00900313 addi t1,zero,9
6c: 006a7463 bgeu s4,t1,0x74
70: 00000a13 addi s4,zero,0
74: 02d52823 sw a3,48(a0)
78: 02e52a23 sw a4,52(a0)
7c: 02952c23 sw s1,56(a0)
80: 03252e23 sw s2,60(a0)
84: 05352023 sw s3,64(a0)
88: 00000793 addi a5,zero,0
8c: 02813403 ld s0,40(sp)
90: 02013483 ld s1,32(sp)
94: 01813903 ld s2,24(sp)
98: 01013983 ld s3,16(sp)
9c: 00813a03 ld s4,8(sp)
a0: 03010113 addi sp,sp,48
a4: 00078513 addi a0,a5,0
a8: 00008067 jalr zero,0(ra)
With RVC:
0: 02000813 addi a6,zero,32
4: 7179 c.addi16sp sp,-48
6: f422 c.sdsp s0,40(sp)
8: f026 c.sdsp s1,32(sp)
a: ec4a c.sdsp s2,24(sp)
c: e84e c.sdsp s3,16(sp)
e: e452 c.sdsp s4,8(sp)
10: 1800 c.addi4spn s0,sp,48
12: 03056683 lwu a3,48(a0)
16: 1682 c.slli a3,0x20
18: 9281 c.srli a3,0x20
1a: 03456703 lwu a4,52(a0)
1e: 1702 c.slli a4,0x20
20: 9301 c.srli a4,0x20
22: 03856483 lwu s1,56(a0)
26: 1482 c.slli s1,0x20
28: 9081 c.srli s1,0x20
2a: 03c56903 lwu s2,60(a0)
2e: 1902 c.slli s2,0x20
30: 02095913 srli s2,s2,0x20
34: 04056983 lwu s3,64(a0)
38: 1982 c.slli s3,0x20
3a: 0209d993 srli s3,s3,0x20
3e: 09056a03 lwu s4,144(a0)
42: 1a02 c.slli s4,0x20
44: 020a5a13 srli s4,s4,0x20
48: 4325 c.li t1,9
4a: 006a7363 bgeu s4,t1,0x50
4e: 4a01 c.li s4,0
50: d914 c.sw a3,48(a0)
52: d958 c.sw a4,52(a0)
54: dd04 c.sw s1,56(a0)
56: 03252e23 sw s2,60(a0)
5a: 05352023 sw s3,64(a0)
5e: 4781 c.li a5,0
60: 7422 c.ldsp s0,40(sp)
62: 7482 c.ldsp s1,32(sp)
64: 6962 c.ldsp s2,24(sp)
66: 69c2 c.ldsp s3,16(sp)
68: 6a22 c.ldsp s4,8(sp)
6a: 6145 c.addi16sp sp,48
6c: 853e c.mv a0,a5
6e: 8082 c.jr ra
Signed-off-by: Luke Nelson <luke.r.nels@gmail.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Cc: Björn Töpel <bjorn.topel@gmail.com>
Link: https://lore.kernel.org/bpf/20200721025241.8077-4-luke.r.nels@gmail.com
2020-07-20 19:52:40 -07:00
|
|
|
emit_slli(rd, rd, imm, ctx);
|
|
|
|
|
2020-05-05 17:03:17 -07:00
|
|
|
if (!is64 && !aux->verifier_zext)
|
2024-01-15 06:12:31 -07:00
|
|
|
emit_zextw(rd, rd, ctx);
|
bpf, riscv: add BPF JIT for RV64G
This commit adds a BPF JIT for RV64G.
The JIT is a two-pass JIT, and has a dynamic prolog/epilogue (similar
to the MIPS64 BPF JIT) instead of static ones (e.g. x86_64).
At the moment the RISC-V Linux port does not support
CONFIG_HAVE_KPROBES, which means that CONFIG_BPF_EVENTS is not
supported. Thus, no tests involving BPF_PROG_TYPE_TRACEPOINT,
BPF_PROG_TYPE_PERF_EVENT, BPF_PROG_TYPE_KPROBE and
BPF_PROG_TYPE_RAW_TRACEPOINT passes.
The implementation does not support "far branching" (>4KiB).
Test results:
# modprobe test_bpf
test_bpf: Summary: 378 PASSED, 0 FAILED, [366/366 JIT'ed]
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier
...
Summary: 761 PASSED, 507 SKIPPED, 2 FAILED
Note that "test_verifier" was run with one build with
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y and one without, otherwise
many of the the tests that require unaligned access were skipped.
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
0
No CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
59
The two failing test_verifier tests are:
"ld_abs: vlan + abs, test 1"
"ld_abs: jump around ld_abs"
This is due to that "far branching" involved in those tests.
All tests where done on QEMU (QEMU emulator version 3.1.50
(v3.1.0-688-g8ae951fbc106)).
Signed-off-by: Björn Töpel <bjorn.topel@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2019-02-05 05:41:22 -07:00
|
|
|
break;
|
|
|
|
case BPF_ALU | BPF_RSH | BPF_K:
|
|
|
|
case BPF_ALU64 | BPF_RSH | BPF_K:
|
bpf, riscv: Use compressed instructions in the rv64 JIT
This patch uses the RVC support and encodings from bpf_jit.h to optimize
the rv64 jit.
The optimizations work by replacing emit(rv_X(...)) with a call to a
helper function emit_X, which will emit a compressed version of the
instruction when possible, and when RVC is enabled.
The JIT continues to pass all tests in lib/test_bpf.c, and introduces
no new failures to test_verifier; both with and without RVC being enabled.
Most changes are straightforward replacements of emit(rv_X(...), ctx)
with emit_X(..., ctx), with the following exceptions bearing mention;
* Change emit_imm to sign-extend the value in "lower", since the
checks for RVC (and the instructions themselves) treat the value as
signed. Otherwise, small negative immediates will not be recognized as
encodable using an RVC instruction. For example, without this change,
emit_imm(rd, -1, ctx) would cause lower to become 4095, which is not a
6b int even though a "c.li rd, -1" instruction suffices.
* For {BPF_MOV,BPF_ADD} BPF_X, drop using addiw,addw in the 32-bit
cases since the values are zero-extended into the upper 32 bits in
the following instructions anyways, and the addition commutes with
zero-extension. (BPF_SUB BPF_X must still use subw since subtraction
does not commute with zero-extension.)
This patch avoids optimizing branches and jumps to use RVC instructions
since surrounding code often makes assumptions about the sizes of
emitted instructions. Optimizing these will require changing these
functions (e.g., emit_branch) to dynamically compute jump offsets.
The following are examples of the JITed code for the verifier selftest
"direct packet read test#3 for CGROUP_SKB OK", without and with RVC
enabled, respectively. The former uses 178 bytes, and the latter uses 112,
for a ~37% reduction in code size for this example.
Without RVC:
0: 02000813 addi a6,zero,32
4: fd010113 addi sp,sp,-48
8: 02813423 sd s0,40(sp)
c: 02913023 sd s1,32(sp)
10: 01213c23 sd s2,24(sp)
14: 01313823 sd s3,16(sp)
18: 01413423 sd s4,8(sp)
1c: 03010413 addi s0,sp,48
20: 03056683 lwu a3,48(a0)
24: 02069693 slli a3,a3,0x20
28: 0206d693 srli a3,a3,0x20
2c: 03456703 lwu a4,52(a0)
30: 02071713 slli a4,a4,0x20
34: 02075713 srli a4,a4,0x20
38: 03856483 lwu s1,56(a0)
3c: 02049493 slli s1,s1,0x20
40: 0204d493 srli s1,s1,0x20
44: 03c56903 lwu s2,60(a0)
48: 02091913 slli s2,s2,0x20
4c: 02095913 srli s2,s2,0x20
50: 04056983 lwu s3,64(a0)
54: 02099993 slli s3,s3,0x20
58: 0209d993 srli s3,s3,0x20
5c: 09056a03 lwu s4,144(a0)
60: 020a1a13 slli s4,s4,0x20
64: 020a5a13 srli s4,s4,0x20
68: 00900313 addi t1,zero,9
6c: 006a7463 bgeu s4,t1,0x74
70: 00000a13 addi s4,zero,0
74: 02d52823 sw a3,48(a0)
78: 02e52a23 sw a4,52(a0)
7c: 02952c23 sw s1,56(a0)
80: 03252e23 sw s2,60(a0)
84: 05352023 sw s3,64(a0)
88: 00000793 addi a5,zero,0
8c: 02813403 ld s0,40(sp)
90: 02013483 ld s1,32(sp)
94: 01813903 ld s2,24(sp)
98: 01013983 ld s3,16(sp)
9c: 00813a03 ld s4,8(sp)
a0: 03010113 addi sp,sp,48
a4: 00078513 addi a0,a5,0
a8: 00008067 jalr zero,0(ra)
With RVC:
0: 02000813 addi a6,zero,32
4: 7179 c.addi16sp sp,-48
6: f422 c.sdsp s0,40(sp)
8: f026 c.sdsp s1,32(sp)
a: ec4a c.sdsp s2,24(sp)
c: e84e c.sdsp s3,16(sp)
e: e452 c.sdsp s4,8(sp)
10: 1800 c.addi4spn s0,sp,48
12: 03056683 lwu a3,48(a0)
16: 1682 c.slli a3,0x20
18: 9281 c.srli a3,0x20
1a: 03456703 lwu a4,52(a0)
1e: 1702 c.slli a4,0x20
20: 9301 c.srli a4,0x20
22: 03856483 lwu s1,56(a0)
26: 1482 c.slli s1,0x20
28: 9081 c.srli s1,0x20
2a: 03c56903 lwu s2,60(a0)
2e: 1902 c.slli s2,0x20
30: 02095913 srli s2,s2,0x20
34: 04056983 lwu s3,64(a0)
38: 1982 c.slli s3,0x20
3a: 0209d993 srli s3,s3,0x20
3e: 09056a03 lwu s4,144(a0)
42: 1a02 c.slli s4,0x20
44: 020a5a13 srli s4,s4,0x20
48: 4325 c.li t1,9
4a: 006a7363 bgeu s4,t1,0x50
4e: 4a01 c.li s4,0
50: d914 c.sw a3,48(a0)
52: d958 c.sw a4,52(a0)
54: dd04 c.sw s1,56(a0)
56: 03252e23 sw s2,60(a0)
5a: 05352023 sw s3,64(a0)
5e: 4781 c.li a5,0
60: 7422 c.ldsp s0,40(sp)
62: 7482 c.ldsp s1,32(sp)
64: 6962 c.ldsp s2,24(sp)
66: 69c2 c.ldsp s3,16(sp)
68: 6a22 c.ldsp s4,8(sp)
6a: 6145 c.addi16sp sp,48
6c: 853e c.mv a0,a5
6e: 8082 c.jr ra
Signed-off-by: Luke Nelson <luke.r.nels@gmail.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Cc: Björn Töpel <bjorn.topel@gmail.com>
Link: https://lore.kernel.org/bpf/20200721025241.8077-4-luke.r.nels@gmail.com
2020-07-20 19:52:40 -07:00
|
|
|
if (is64)
|
|
|
|
emit_srli(rd, rd, imm, ctx);
|
|
|
|
else
|
|
|
|
emit(rv_srliw(rd, rd, imm), ctx);
|
|
|
|
|
2020-05-05 17:03:17 -07:00
|
|
|
if (!is64 && !aux->verifier_zext)
|
2024-01-15 06:12:31 -07:00
|
|
|
emit_zextw(rd, rd, ctx);
|
bpf, riscv: add BPF JIT for RV64G
This commit adds a BPF JIT for RV64G.
The JIT is a two-pass JIT, and has a dynamic prolog/epilogue (similar
to the MIPS64 BPF JIT) instead of static ones (e.g. x86_64).
At the moment the RISC-V Linux port does not support
CONFIG_HAVE_KPROBES, which means that CONFIG_BPF_EVENTS is not
supported. Thus, no tests involving BPF_PROG_TYPE_TRACEPOINT,
BPF_PROG_TYPE_PERF_EVENT, BPF_PROG_TYPE_KPROBE and
BPF_PROG_TYPE_RAW_TRACEPOINT passes.
The implementation does not support "far branching" (>4KiB).
Test results:
# modprobe test_bpf
test_bpf: Summary: 378 PASSED, 0 FAILED, [366/366 JIT'ed]
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier
...
Summary: 761 PASSED, 507 SKIPPED, 2 FAILED
Note that "test_verifier" was run with one build with
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y and one without, otherwise
many of the the tests that require unaligned access were skipped.
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
0
No CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
59
The two failing test_verifier tests are:
"ld_abs: vlan + abs, test 1"
"ld_abs: jump around ld_abs"
This is due to that "far branching" involved in those tests.
All tests where done on QEMU (QEMU emulator version 3.1.50
(v3.1.0-688-g8ae951fbc106)).
Signed-off-by: Björn Töpel <bjorn.topel@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2019-02-05 05:41:22 -07:00
|
|
|
break;
|
|
|
|
case BPF_ALU | BPF_ARSH | BPF_K:
|
|
|
|
case BPF_ALU64 | BPF_ARSH | BPF_K:
|
bpf, riscv: Use compressed instructions in the rv64 JIT
This patch uses the RVC support and encodings from bpf_jit.h to optimize
the rv64 jit.
The optimizations work by replacing emit(rv_X(...)) with a call to a
helper function emit_X, which will emit a compressed version of the
instruction when possible, and when RVC is enabled.
The JIT continues to pass all tests in lib/test_bpf.c, and introduces
no new failures to test_verifier; both with and without RVC being enabled.
Most changes are straightforward replacements of emit(rv_X(...), ctx)
with emit_X(..., ctx), with the following exceptions bearing mention;
* Change emit_imm to sign-extend the value in "lower", since the
checks for RVC (and the instructions themselves) treat the value as
signed. Otherwise, small negative immediates will not be recognized as
encodable using an RVC instruction. For example, without this change,
emit_imm(rd, -1, ctx) would cause lower to become 4095, which is not a
6b int even though a "c.li rd, -1" instruction suffices.
* For {BPF_MOV,BPF_ADD} BPF_X, drop using addiw,addw in the 32-bit
cases since the values are zero-extended into the upper 32 bits in
the following instructions anyways, and the addition commutes with
zero-extension. (BPF_SUB BPF_X must still use subw since subtraction
does not commute with zero-extension.)
This patch avoids optimizing branches and jumps to use RVC instructions
since surrounding code often makes assumptions about the sizes of
emitted instructions. Optimizing these will require changing these
functions (e.g., emit_branch) to dynamically compute jump offsets.
The following are examples of the JITed code for the verifier selftest
"direct packet read test#3 for CGROUP_SKB OK", without and with RVC
enabled, respectively. The former uses 178 bytes, and the latter uses 112,
for a ~37% reduction in code size for this example.
Without RVC:
0: 02000813 addi a6,zero,32
4: fd010113 addi sp,sp,-48
8: 02813423 sd s0,40(sp)
c: 02913023 sd s1,32(sp)
10: 01213c23 sd s2,24(sp)
14: 01313823 sd s3,16(sp)
18: 01413423 sd s4,8(sp)
1c: 03010413 addi s0,sp,48
20: 03056683 lwu a3,48(a0)
24: 02069693 slli a3,a3,0x20
28: 0206d693 srli a3,a3,0x20
2c: 03456703 lwu a4,52(a0)
30: 02071713 slli a4,a4,0x20
34: 02075713 srli a4,a4,0x20
38: 03856483 lwu s1,56(a0)
3c: 02049493 slli s1,s1,0x20
40: 0204d493 srli s1,s1,0x20
44: 03c56903 lwu s2,60(a0)
48: 02091913 slli s2,s2,0x20
4c: 02095913 srli s2,s2,0x20
50: 04056983 lwu s3,64(a0)
54: 02099993 slli s3,s3,0x20
58: 0209d993 srli s3,s3,0x20
5c: 09056a03 lwu s4,144(a0)
60: 020a1a13 slli s4,s4,0x20
64: 020a5a13 srli s4,s4,0x20
68: 00900313 addi t1,zero,9
6c: 006a7463 bgeu s4,t1,0x74
70: 00000a13 addi s4,zero,0
74: 02d52823 sw a3,48(a0)
78: 02e52a23 sw a4,52(a0)
7c: 02952c23 sw s1,56(a0)
80: 03252e23 sw s2,60(a0)
84: 05352023 sw s3,64(a0)
88: 00000793 addi a5,zero,0
8c: 02813403 ld s0,40(sp)
90: 02013483 ld s1,32(sp)
94: 01813903 ld s2,24(sp)
98: 01013983 ld s3,16(sp)
9c: 00813a03 ld s4,8(sp)
a0: 03010113 addi sp,sp,48
a4: 00078513 addi a0,a5,0
a8: 00008067 jalr zero,0(ra)
With RVC:
0: 02000813 addi a6,zero,32
4: 7179 c.addi16sp sp,-48
6: f422 c.sdsp s0,40(sp)
8: f026 c.sdsp s1,32(sp)
a: ec4a c.sdsp s2,24(sp)
c: e84e c.sdsp s3,16(sp)
e: e452 c.sdsp s4,8(sp)
10: 1800 c.addi4spn s0,sp,48
12: 03056683 lwu a3,48(a0)
16: 1682 c.slli a3,0x20
18: 9281 c.srli a3,0x20
1a: 03456703 lwu a4,52(a0)
1e: 1702 c.slli a4,0x20
20: 9301 c.srli a4,0x20
22: 03856483 lwu s1,56(a0)
26: 1482 c.slli s1,0x20
28: 9081 c.srli s1,0x20
2a: 03c56903 lwu s2,60(a0)
2e: 1902 c.slli s2,0x20
30: 02095913 srli s2,s2,0x20
34: 04056983 lwu s3,64(a0)
38: 1982 c.slli s3,0x20
3a: 0209d993 srli s3,s3,0x20
3e: 09056a03 lwu s4,144(a0)
42: 1a02 c.slli s4,0x20
44: 020a5a13 srli s4,s4,0x20
48: 4325 c.li t1,9
4a: 006a7363 bgeu s4,t1,0x50
4e: 4a01 c.li s4,0
50: d914 c.sw a3,48(a0)
52: d958 c.sw a4,52(a0)
54: dd04 c.sw s1,56(a0)
56: 03252e23 sw s2,60(a0)
5a: 05352023 sw s3,64(a0)
5e: 4781 c.li a5,0
60: 7422 c.ldsp s0,40(sp)
62: 7482 c.ldsp s1,32(sp)
64: 6962 c.ldsp s2,24(sp)
66: 69c2 c.ldsp s3,16(sp)
68: 6a22 c.ldsp s4,8(sp)
6a: 6145 c.addi16sp sp,48
6c: 853e c.mv a0,a5
6e: 8082 c.jr ra
Signed-off-by: Luke Nelson <luke.r.nels@gmail.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Cc: Björn Töpel <bjorn.topel@gmail.com>
Link: https://lore.kernel.org/bpf/20200721025241.8077-4-luke.r.nels@gmail.com
2020-07-20 19:52:40 -07:00
|
|
|
if (is64)
|
|
|
|
emit_srai(rd, rd, imm, ctx);
|
|
|
|
else
|
|
|
|
emit(rv_sraiw(rd, rd, imm), ctx);
|
|
|
|
|
2020-05-05 17:03:17 -07:00
|
|
|
if (!is64 && !aux->verifier_zext)
|
2024-01-15 06:12:31 -07:00
|
|
|
emit_zextw(rd, rd, ctx);
|
bpf, riscv: add BPF JIT for RV64G
This commit adds a BPF JIT for RV64G.
The JIT is a two-pass JIT, and has a dynamic prolog/epilogue (similar
to the MIPS64 BPF JIT) instead of static ones (e.g. x86_64).
At the moment the RISC-V Linux port does not support
CONFIG_HAVE_KPROBES, which means that CONFIG_BPF_EVENTS is not
supported. Thus, no tests involving BPF_PROG_TYPE_TRACEPOINT,
BPF_PROG_TYPE_PERF_EVENT, BPF_PROG_TYPE_KPROBE and
BPF_PROG_TYPE_RAW_TRACEPOINT passes.
The implementation does not support "far branching" (>4KiB).
Test results:
# modprobe test_bpf
test_bpf: Summary: 378 PASSED, 0 FAILED, [366/366 JIT'ed]
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier
...
Summary: 761 PASSED, 507 SKIPPED, 2 FAILED
Note that "test_verifier" was run with one build with
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y and one without, otherwise
many of the the tests that require unaligned access were skipped.
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
0
No CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
59
The two failing test_verifier tests are:
"ld_abs: vlan + abs, test 1"
"ld_abs: jump around ld_abs"
This is due to that "far branching" involved in those tests.
All tests where done on QEMU (QEMU emulator version 3.1.50
(v3.1.0-688-g8ae951fbc106)).
Signed-off-by: Björn Töpel <bjorn.topel@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2019-02-05 05:41:22 -07:00
|
|
|
break;
|
|
|
|
|
|
|
|
/* JUMP off */
|
|
|
|
case BPF_JMP | BPF_JA:
|
2023-08-24 02:49:58 -07:00
|
|
|
case BPF_JMP32 | BPF_JA:
|
|
|
|
if (BPF_CLASS(code) == BPF_JMP)
|
|
|
|
rvoff = rv_offset(i, off, ctx);
|
|
|
|
else
|
|
|
|
rvoff = rv_offset(i, imm, ctx);
|
2023-02-15 06:52:03 -07:00
|
|
|
ret = emit_jump_and_link(RV_REG_ZERO, rvoff, true, ctx);
|
2020-04-06 15:16:04 -07:00
|
|
|
if (ret)
|
|
|
|
return ret;
|
bpf, riscv: add BPF JIT for RV64G
This commit adds a BPF JIT for RV64G.
The JIT is a two-pass JIT, and has a dynamic prolog/epilogue (similar
to the MIPS64 BPF JIT) instead of static ones (e.g. x86_64).
At the moment the RISC-V Linux port does not support
CONFIG_HAVE_KPROBES, which means that CONFIG_BPF_EVENTS is not
supported. Thus, no tests involving BPF_PROG_TYPE_TRACEPOINT,
BPF_PROG_TYPE_PERF_EVENT, BPF_PROG_TYPE_KPROBE and
BPF_PROG_TYPE_RAW_TRACEPOINT passes.
The implementation does not support "far branching" (>4KiB).
Test results:
# modprobe test_bpf
test_bpf: Summary: 378 PASSED, 0 FAILED, [366/366 JIT'ed]
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier
...
Summary: 761 PASSED, 507 SKIPPED, 2 FAILED
Note that "test_verifier" was run with one build with
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y and one without, otherwise
many of the the tests that require unaligned access were skipped.
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
0
No CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
59
The two failing test_verifier tests are:
"ld_abs: vlan + abs, test 1"
"ld_abs: jump around ld_abs"
This is due to that "far branching" involved in those tests.
All tests where done on QEMU (QEMU emulator version 3.1.50
(v3.1.0-688-g8ae951fbc106)).
Signed-off-by: Björn Töpel <bjorn.topel@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2019-02-05 05:41:22 -07:00
|
|
|
break;
|
|
|
|
|
|
|
|
/* IF (dst COND src) JUMP off */
|
|
|
|
case BPF_JMP | BPF_JEQ | BPF_X:
|
|
|
|
case BPF_JMP32 | BPF_JEQ | BPF_X:
|
|
|
|
case BPF_JMP | BPF_JGT | BPF_X:
|
|
|
|
case BPF_JMP32 | BPF_JGT | BPF_X:
|
|
|
|
case BPF_JMP | BPF_JLT | BPF_X:
|
|
|
|
case BPF_JMP32 | BPF_JLT | BPF_X:
|
|
|
|
case BPF_JMP | BPF_JGE | BPF_X:
|
|
|
|
case BPF_JMP32 | BPF_JGE | BPF_X:
|
|
|
|
case BPF_JMP | BPF_JLE | BPF_X:
|
|
|
|
case BPF_JMP32 | BPF_JLE | BPF_X:
|
|
|
|
case BPF_JMP | BPF_JNE | BPF_X:
|
|
|
|
case BPF_JMP32 | BPF_JNE | BPF_X:
|
|
|
|
case BPF_JMP | BPF_JSGT | BPF_X:
|
|
|
|
case BPF_JMP32 | BPF_JSGT | BPF_X:
|
|
|
|
case BPF_JMP | BPF_JSLT | BPF_X:
|
|
|
|
case BPF_JMP32 | BPF_JSLT | BPF_X:
|
|
|
|
case BPF_JMP | BPF_JSGE | BPF_X:
|
|
|
|
case BPF_JMP32 | BPF_JSGE | BPF_X:
|
|
|
|
case BPF_JMP | BPF_JSLE | BPF_X:
|
|
|
|
case BPF_JMP32 | BPF_JSLE | BPF_X:
|
|
|
|
case BPF_JMP | BPF_JSET | BPF_X:
|
|
|
|
case BPF_JMP32 | BPF_JSET | BPF_X:
|
2019-12-16 02:13:36 -07:00
|
|
|
rvoff = rv_offset(i, off, ctx);
|
|
|
|
if (!is64) {
|
|
|
|
s = ctx->ninsns;
|
2024-01-15 06:12:32 -07:00
|
|
|
if (is_signed_bpf_cond(BPF_OP(code))) {
|
|
|
|
emit_sextw_alt(&rs, RV_REG_T1, ctx);
|
|
|
|
emit_sextw_alt(&rd, RV_REG_T2, ctx);
|
|
|
|
} else {
|
|
|
|
emit_zextw_alt(&rs, RV_REG_T1, ctx);
|
|
|
|
emit_zextw_alt(&rd, RV_REG_T2, ctx);
|
|
|
|
}
|
2019-12-16 02:13:36 -07:00
|
|
|
e = ctx->ninsns;
|
|
|
|
|
|
|
|
/* Adjust for extra insns */
|
2020-07-20 19:52:38 -07:00
|
|
|
rvoff -= ninsns_rvoff(e - s);
|
2019-12-16 02:13:36 -07:00
|
|
|
}
|
|
|
|
|
|
|
|
if (BPF_OP(code) == BPF_JSET) {
|
|
|
|
/* Adjust for and */
|
|
|
|
rvoff -= 4;
|
bpf, riscv: Use compressed instructions in the rv64 JIT
This patch uses the RVC support and encodings from bpf_jit.h to optimize
the rv64 jit.
The optimizations work by replacing emit(rv_X(...)) with a call to a
helper function emit_X, which will emit a compressed version of the
instruction when possible, and when RVC is enabled.
The JIT continues to pass all tests in lib/test_bpf.c, and introduces
no new failures to test_verifier; both with and without RVC being enabled.
Most changes are straightforward replacements of emit(rv_X(...), ctx)
with emit_X(..., ctx), with the following exceptions bearing mention;
* Change emit_imm to sign-extend the value in "lower", since the
checks for RVC (and the instructions themselves) treat the value as
signed. Otherwise, small negative immediates will not be recognized as
encodable using an RVC instruction. For example, without this change,
emit_imm(rd, -1, ctx) would cause lower to become 4095, which is not a
6b int even though a "c.li rd, -1" instruction suffices.
* For {BPF_MOV,BPF_ADD} BPF_X, drop using addiw,addw in the 32-bit
cases since the values are zero-extended into the upper 32 bits in
the following instructions anyways, and the addition commutes with
zero-extension. (BPF_SUB BPF_X must still use subw since subtraction
does not commute with zero-extension.)
This patch avoids optimizing branches and jumps to use RVC instructions
since surrounding code often makes assumptions about the sizes of
emitted instructions. Optimizing these will require changing these
functions (e.g., emit_branch) to dynamically compute jump offsets.
The following are examples of the JITed code for the verifier selftest
"direct packet read test#3 for CGROUP_SKB OK", without and with RVC
enabled, respectively. The former uses 178 bytes, and the latter uses 112,
for a ~37% reduction in code size for this example.
Without RVC:
0: 02000813 addi a6,zero,32
4: fd010113 addi sp,sp,-48
8: 02813423 sd s0,40(sp)
c: 02913023 sd s1,32(sp)
10: 01213c23 sd s2,24(sp)
14: 01313823 sd s3,16(sp)
18: 01413423 sd s4,8(sp)
1c: 03010413 addi s0,sp,48
20: 03056683 lwu a3,48(a0)
24: 02069693 slli a3,a3,0x20
28: 0206d693 srli a3,a3,0x20
2c: 03456703 lwu a4,52(a0)
30: 02071713 slli a4,a4,0x20
34: 02075713 srli a4,a4,0x20
38: 03856483 lwu s1,56(a0)
3c: 02049493 slli s1,s1,0x20
40: 0204d493 srli s1,s1,0x20
44: 03c56903 lwu s2,60(a0)
48: 02091913 slli s2,s2,0x20
4c: 02095913 srli s2,s2,0x20
50: 04056983 lwu s3,64(a0)
54: 02099993 slli s3,s3,0x20
58: 0209d993 srli s3,s3,0x20
5c: 09056a03 lwu s4,144(a0)
60: 020a1a13 slli s4,s4,0x20
64: 020a5a13 srli s4,s4,0x20
68: 00900313 addi t1,zero,9
6c: 006a7463 bgeu s4,t1,0x74
70: 00000a13 addi s4,zero,0
74: 02d52823 sw a3,48(a0)
78: 02e52a23 sw a4,52(a0)
7c: 02952c23 sw s1,56(a0)
80: 03252e23 sw s2,60(a0)
84: 05352023 sw s3,64(a0)
88: 00000793 addi a5,zero,0
8c: 02813403 ld s0,40(sp)
90: 02013483 ld s1,32(sp)
94: 01813903 ld s2,24(sp)
98: 01013983 ld s3,16(sp)
9c: 00813a03 ld s4,8(sp)
a0: 03010113 addi sp,sp,48
a4: 00078513 addi a0,a5,0
a8: 00008067 jalr zero,0(ra)
With RVC:
0: 02000813 addi a6,zero,32
4: 7179 c.addi16sp sp,-48
6: f422 c.sdsp s0,40(sp)
8: f026 c.sdsp s1,32(sp)
a: ec4a c.sdsp s2,24(sp)
c: e84e c.sdsp s3,16(sp)
e: e452 c.sdsp s4,8(sp)
10: 1800 c.addi4spn s0,sp,48
12: 03056683 lwu a3,48(a0)
16: 1682 c.slli a3,0x20
18: 9281 c.srli a3,0x20
1a: 03456703 lwu a4,52(a0)
1e: 1702 c.slli a4,0x20
20: 9301 c.srli a4,0x20
22: 03856483 lwu s1,56(a0)
26: 1482 c.slli s1,0x20
28: 9081 c.srli s1,0x20
2a: 03c56903 lwu s2,60(a0)
2e: 1902 c.slli s2,0x20
30: 02095913 srli s2,s2,0x20
34: 04056983 lwu s3,64(a0)
38: 1982 c.slli s3,0x20
3a: 0209d993 srli s3,s3,0x20
3e: 09056a03 lwu s4,144(a0)
42: 1a02 c.slli s4,0x20
44: 020a5a13 srli s4,s4,0x20
48: 4325 c.li t1,9
4a: 006a7363 bgeu s4,t1,0x50
4e: 4a01 c.li s4,0
50: d914 c.sw a3,48(a0)
52: d958 c.sw a4,52(a0)
54: dd04 c.sw s1,56(a0)
56: 03252e23 sw s2,60(a0)
5a: 05352023 sw s3,64(a0)
5e: 4781 c.li a5,0
60: 7422 c.ldsp s0,40(sp)
62: 7482 c.ldsp s1,32(sp)
64: 6962 c.ldsp s2,24(sp)
66: 69c2 c.ldsp s3,16(sp)
68: 6a22 c.ldsp s4,8(sp)
6a: 6145 c.addi16sp sp,48
6c: 853e c.mv a0,a5
6e: 8082 c.jr ra
Signed-off-by: Luke Nelson <luke.r.nels@gmail.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Cc: Björn Töpel <bjorn.topel@gmail.com>
Link: https://lore.kernel.org/bpf/20200721025241.8077-4-luke.r.nels@gmail.com
2020-07-20 19:52:40 -07:00
|
|
|
emit_and(RV_REG_T1, rd, rs, ctx);
|
2024-01-15 06:12:32 -07:00
|
|
|
emit_branch(BPF_JNE, RV_REG_T1, RV_REG_ZERO, rvoff, ctx);
|
2019-12-16 02:13:36 -07:00
|
|
|
} else {
|
|
|
|
emit_branch(BPF_OP(code), rd, rs, rvoff, ctx);
|
|
|
|
}
|
bpf, riscv: add BPF JIT for RV64G
This commit adds a BPF JIT for RV64G.
The JIT is a two-pass JIT, and has a dynamic prolog/epilogue (similar
to the MIPS64 BPF JIT) instead of static ones (e.g. x86_64).
At the moment the RISC-V Linux port does not support
CONFIG_HAVE_KPROBES, which means that CONFIG_BPF_EVENTS is not
supported. Thus, no tests involving BPF_PROG_TYPE_TRACEPOINT,
BPF_PROG_TYPE_PERF_EVENT, BPF_PROG_TYPE_KPROBE and
BPF_PROG_TYPE_RAW_TRACEPOINT passes.
The implementation does not support "far branching" (>4KiB).
Test results:
# modprobe test_bpf
test_bpf: Summary: 378 PASSED, 0 FAILED, [366/366 JIT'ed]
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier
...
Summary: 761 PASSED, 507 SKIPPED, 2 FAILED
Note that "test_verifier" was run with one build with
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y and one without, otherwise
many of the the tests that require unaligned access were skipped.
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
0
No CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
59
The two failing test_verifier tests are:
"ld_abs: vlan + abs, test 1"
"ld_abs: jump around ld_abs"
This is due to that "far branching" involved in those tests.
All tests where done on QEMU (QEMU emulator version 3.1.50
(v3.1.0-688-g8ae951fbc106)).
Signed-off-by: Björn Töpel <bjorn.topel@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2019-02-05 05:41:22 -07:00
|
|
|
break;
|
|
|
|
|
|
|
|
/* IF (dst COND imm) JUMP off */
|
|
|
|
case BPF_JMP | BPF_JEQ | BPF_K:
|
|
|
|
case BPF_JMP32 | BPF_JEQ | BPF_K:
|
|
|
|
case BPF_JMP | BPF_JGT | BPF_K:
|
|
|
|
case BPF_JMP32 | BPF_JGT | BPF_K:
|
|
|
|
case BPF_JMP | BPF_JLT | BPF_K:
|
|
|
|
case BPF_JMP32 | BPF_JLT | BPF_K:
|
|
|
|
case BPF_JMP | BPF_JGE | BPF_K:
|
|
|
|
case BPF_JMP32 | BPF_JGE | BPF_K:
|
|
|
|
case BPF_JMP | BPF_JLE | BPF_K:
|
|
|
|
case BPF_JMP32 | BPF_JLE | BPF_K:
|
|
|
|
case BPF_JMP | BPF_JNE | BPF_K:
|
|
|
|
case BPF_JMP32 | BPF_JNE | BPF_K:
|
|
|
|
case BPF_JMP | BPF_JSGT | BPF_K:
|
|
|
|
case BPF_JMP32 | BPF_JSGT | BPF_K:
|
|
|
|
case BPF_JMP | BPF_JSLT | BPF_K:
|
|
|
|
case BPF_JMP32 | BPF_JSLT | BPF_K:
|
|
|
|
case BPF_JMP | BPF_JSGE | BPF_K:
|
|
|
|
case BPF_JMP32 | BPF_JSGE | BPF_K:
|
|
|
|
case BPF_JMP | BPF_JSLE | BPF_K:
|
|
|
|
case BPF_JMP32 | BPF_JSLE | BPF_K:
|
2019-12-16 02:13:36 -07:00
|
|
|
rvoff = rv_offset(i, off, ctx);
|
|
|
|
s = ctx->ninsns;
|
2024-01-15 06:12:32 -07:00
|
|
|
if (imm)
|
2020-05-05 17:03:19 -07:00
|
|
|
emit_imm(RV_REG_T1, imm, ctx);
|
2024-01-15 06:12:32 -07:00
|
|
|
rs = imm ? RV_REG_T1 : RV_REG_ZERO;
|
2019-12-16 02:13:36 -07:00
|
|
|
if (!is64) {
|
2024-01-15 06:12:32 -07:00
|
|
|
if (is_signed_bpf_cond(BPF_OP(code))) {
|
|
|
|
emit_sextw_alt(&rd, RV_REG_T2, ctx);
|
|
|
|
/* rs has been sign extended */
|
|
|
|
} else {
|
|
|
|
emit_zextw_alt(&rd, RV_REG_T2, ctx);
|
|
|
|
if (imm)
|
|
|
|
emit_zextw(rs, rs, ctx);
|
|
|
|
}
|
2019-12-16 02:13:36 -07:00
|
|
|
}
|
|
|
|
e = ctx->ninsns;
|
|
|
|
|
|
|
|
/* Adjust for extra insns */
|
2020-07-20 19:52:38 -07:00
|
|
|
rvoff -= ninsns_rvoff(e - s);
|
bpf, riscv: Optimize BPF_JSET BPF_K using andi on RV64
This patch optimizes BPF_JSET BPF_K by using a RISC-V andi instruction
when the BPF immediate fits in 12 bits, instead of first loading the
immediate to a temporary register.
Examples of generated code with and without this optimization:
BPF_JMP_IMM(BPF_JSET, R1, 2, 1) without optimization:
20: li t1,2
24: and t1,a0,t1
28: bnez t1,0x30
BPF_JMP_IMM(BPF_JSET, R1, 2, 1) with optimization:
20: andi t1,a0,2
24: bnez t1,0x2c
BPF_JMP32_IMM(BPF_JSET, R1, 2, 1) without optimization:
20: li t1,2
24: mv t2,a0
28: slli t2,t2,0x20
2c: srli t2,t2,0x20
30: slli t1,t1,0x20
34: srli t1,t1,0x20
38: and t1,t2,t1
3c: bnez t1,0x44
BPF_JMP32_IMM(BPF_JSET, R1, 2, 1) with optimization:
20: andi t1,a0,2
24: bnez t1,0x2c
In these examples, because the upper 32 bits of the sign-extended
immediate are 0, BPF_JMP BPF_JSET and BPF_JMP32 BPF_JSET are equivalent
and therefore the JIT produces identical code for them.
Co-developed-by: Xi Wang <xi.wang@gmail.com>
Signed-off-by: Xi Wang <xi.wang@gmail.com>
Signed-off-by: Luke Nelson <luke.r.nels@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Reviewed-by: Björn Töpel <bjorn.topel@gmail.com>
Acked-by: Björn Töpel <bjorn.topel@gmail.com>
Link: https://lore.kernel.org/bpf/20200506000320.28965-5-luke.r.nels@gmail.com
2020-05-05 17:03:20 -07:00
|
|
|
emit_branch(BPF_OP(code), rd, rs, rvoff, ctx);
|
|
|
|
break;
|
2019-12-16 02:13:36 -07:00
|
|
|
|
bpf, riscv: Optimize BPF_JSET BPF_K using andi on RV64
This patch optimizes BPF_JSET BPF_K by using a RISC-V andi instruction
when the BPF immediate fits in 12 bits, instead of first loading the
immediate to a temporary register.
Examples of generated code with and without this optimization:
BPF_JMP_IMM(BPF_JSET, R1, 2, 1) without optimization:
20: li t1,2
24: and t1,a0,t1
28: bnez t1,0x30
BPF_JMP_IMM(BPF_JSET, R1, 2, 1) with optimization:
20: andi t1,a0,2
24: bnez t1,0x2c
BPF_JMP32_IMM(BPF_JSET, R1, 2, 1) without optimization:
20: li t1,2
24: mv t2,a0
28: slli t2,t2,0x20
2c: srli t2,t2,0x20
30: slli t1,t1,0x20
34: srli t1,t1,0x20
38: and t1,t2,t1
3c: bnez t1,0x44
BPF_JMP32_IMM(BPF_JSET, R1, 2, 1) with optimization:
20: andi t1,a0,2
24: bnez t1,0x2c
In these examples, because the upper 32 bits of the sign-extended
immediate are 0, BPF_JMP BPF_JSET and BPF_JMP32 BPF_JSET are equivalent
and therefore the JIT produces identical code for them.
Co-developed-by: Xi Wang <xi.wang@gmail.com>
Signed-off-by: Xi Wang <xi.wang@gmail.com>
Signed-off-by: Luke Nelson <luke.r.nels@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Reviewed-by: Björn Töpel <bjorn.topel@gmail.com>
Acked-by: Björn Töpel <bjorn.topel@gmail.com>
Link: https://lore.kernel.org/bpf/20200506000320.28965-5-luke.r.nels@gmail.com
2020-05-05 17:03:20 -07:00
|
|
|
case BPF_JMP | BPF_JSET | BPF_K:
|
|
|
|
case BPF_JMP32 | BPF_JSET | BPF_K:
|
|
|
|
rvoff = rv_offset(i, off, ctx);
|
|
|
|
s = ctx->ninsns;
|
|
|
|
if (is_12b_int(imm)) {
|
bpf, riscv: Use compressed instructions in the rv64 JIT
This patch uses the RVC support and encodings from bpf_jit.h to optimize
the rv64 jit.
The optimizations work by replacing emit(rv_X(...)) with a call to a
helper function emit_X, which will emit a compressed version of the
instruction when possible, and when RVC is enabled.
The JIT continues to pass all tests in lib/test_bpf.c, and introduces
no new failures to test_verifier; both with and without RVC being enabled.
Most changes are straightforward replacements of emit(rv_X(...), ctx)
with emit_X(..., ctx), with the following exceptions bearing mention;
* Change emit_imm to sign-extend the value in "lower", since the
checks for RVC (and the instructions themselves) treat the value as
signed. Otherwise, small negative immediates will not be recognized as
encodable using an RVC instruction. For example, without this change,
emit_imm(rd, -1, ctx) would cause lower to become 4095, which is not a
6b int even though a "c.li rd, -1" instruction suffices.
* For {BPF_MOV,BPF_ADD} BPF_X, drop using addiw,addw in the 32-bit
cases since the values are zero-extended into the upper 32 bits in
the following instructions anyways, and the addition commutes with
zero-extension. (BPF_SUB BPF_X must still use subw since subtraction
does not commute with zero-extension.)
This patch avoids optimizing branches and jumps to use RVC instructions
since surrounding code often makes assumptions about the sizes of
emitted instructions. Optimizing these will require changing these
functions (e.g., emit_branch) to dynamically compute jump offsets.
The following are examples of the JITed code for the verifier selftest
"direct packet read test#3 for CGROUP_SKB OK", without and with RVC
enabled, respectively. The former uses 178 bytes, and the latter uses 112,
for a ~37% reduction in code size for this example.
Without RVC:
0: 02000813 addi a6,zero,32
4: fd010113 addi sp,sp,-48
8: 02813423 sd s0,40(sp)
c: 02913023 sd s1,32(sp)
10: 01213c23 sd s2,24(sp)
14: 01313823 sd s3,16(sp)
18: 01413423 sd s4,8(sp)
1c: 03010413 addi s0,sp,48
20: 03056683 lwu a3,48(a0)
24: 02069693 slli a3,a3,0x20
28: 0206d693 srli a3,a3,0x20
2c: 03456703 lwu a4,52(a0)
30: 02071713 slli a4,a4,0x20
34: 02075713 srli a4,a4,0x20
38: 03856483 lwu s1,56(a0)
3c: 02049493 slli s1,s1,0x20
40: 0204d493 srli s1,s1,0x20
44: 03c56903 lwu s2,60(a0)
48: 02091913 slli s2,s2,0x20
4c: 02095913 srli s2,s2,0x20
50: 04056983 lwu s3,64(a0)
54: 02099993 slli s3,s3,0x20
58: 0209d993 srli s3,s3,0x20
5c: 09056a03 lwu s4,144(a0)
60: 020a1a13 slli s4,s4,0x20
64: 020a5a13 srli s4,s4,0x20
68: 00900313 addi t1,zero,9
6c: 006a7463 bgeu s4,t1,0x74
70: 00000a13 addi s4,zero,0
74: 02d52823 sw a3,48(a0)
78: 02e52a23 sw a4,52(a0)
7c: 02952c23 sw s1,56(a0)
80: 03252e23 sw s2,60(a0)
84: 05352023 sw s3,64(a0)
88: 00000793 addi a5,zero,0
8c: 02813403 ld s0,40(sp)
90: 02013483 ld s1,32(sp)
94: 01813903 ld s2,24(sp)
98: 01013983 ld s3,16(sp)
9c: 00813a03 ld s4,8(sp)
a0: 03010113 addi sp,sp,48
a4: 00078513 addi a0,a5,0
a8: 00008067 jalr zero,0(ra)
With RVC:
0: 02000813 addi a6,zero,32
4: 7179 c.addi16sp sp,-48
6: f422 c.sdsp s0,40(sp)
8: f026 c.sdsp s1,32(sp)
a: ec4a c.sdsp s2,24(sp)
c: e84e c.sdsp s3,16(sp)
e: e452 c.sdsp s4,8(sp)
10: 1800 c.addi4spn s0,sp,48
12: 03056683 lwu a3,48(a0)
16: 1682 c.slli a3,0x20
18: 9281 c.srli a3,0x20
1a: 03456703 lwu a4,52(a0)
1e: 1702 c.slli a4,0x20
20: 9301 c.srli a4,0x20
22: 03856483 lwu s1,56(a0)
26: 1482 c.slli s1,0x20
28: 9081 c.srli s1,0x20
2a: 03c56903 lwu s2,60(a0)
2e: 1902 c.slli s2,0x20
30: 02095913 srli s2,s2,0x20
34: 04056983 lwu s3,64(a0)
38: 1982 c.slli s3,0x20
3a: 0209d993 srli s3,s3,0x20
3e: 09056a03 lwu s4,144(a0)
42: 1a02 c.slli s4,0x20
44: 020a5a13 srli s4,s4,0x20
48: 4325 c.li t1,9
4a: 006a7363 bgeu s4,t1,0x50
4e: 4a01 c.li s4,0
50: d914 c.sw a3,48(a0)
52: d958 c.sw a4,52(a0)
54: dd04 c.sw s1,56(a0)
56: 03252e23 sw s2,60(a0)
5a: 05352023 sw s3,64(a0)
5e: 4781 c.li a5,0
60: 7422 c.ldsp s0,40(sp)
62: 7482 c.ldsp s1,32(sp)
64: 6962 c.ldsp s2,24(sp)
66: 69c2 c.ldsp s3,16(sp)
68: 6a22 c.ldsp s4,8(sp)
6a: 6145 c.addi16sp sp,48
6c: 853e c.mv a0,a5
6e: 8082 c.jr ra
Signed-off-by: Luke Nelson <luke.r.nels@gmail.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Cc: Björn Töpel <bjorn.topel@gmail.com>
Link: https://lore.kernel.org/bpf/20200721025241.8077-4-luke.r.nels@gmail.com
2020-07-20 19:52:40 -07:00
|
|
|
emit_andi(RV_REG_T1, rd, imm, ctx);
|
2019-12-16 02:13:36 -07:00
|
|
|
} else {
|
bpf, riscv: Optimize BPF_JSET BPF_K using andi on RV64
This patch optimizes BPF_JSET BPF_K by using a RISC-V andi instruction
when the BPF immediate fits in 12 bits, instead of first loading the
immediate to a temporary register.
Examples of generated code with and without this optimization:
BPF_JMP_IMM(BPF_JSET, R1, 2, 1) without optimization:
20: li t1,2
24: and t1,a0,t1
28: bnez t1,0x30
BPF_JMP_IMM(BPF_JSET, R1, 2, 1) with optimization:
20: andi t1,a0,2
24: bnez t1,0x2c
BPF_JMP32_IMM(BPF_JSET, R1, 2, 1) without optimization:
20: li t1,2
24: mv t2,a0
28: slli t2,t2,0x20
2c: srli t2,t2,0x20
30: slli t1,t1,0x20
34: srli t1,t1,0x20
38: and t1,t2,t1
3c: bnez t1,0x44
BPF_JMP32_IMM(BPF_JSET, R1, 2, 1) with optimization:
20: andi t1,a0,2
24: bnez t1,0x2c
In these examples, because the upper 32 bits of the sign-extended
immediate are 0, BPF_JMP BPF_JSET and BPF_JMP32 BPF_JSET are equivalent
and therefore the JIT produces identical code for them.
Co-developed-by: Xi Wang <xi.wang@gmail.com>
Signed-off-by: Xi Wang <xi.wang@gmail.com>
Signed-off-by: Luke Nelson <luke.r.nels@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Reviewed-by: Björn Töpel <bjorn.topel@gmail.com>
Acked-by: Björn Töpel <bjorn.topel@gmail.com>
Link: https://lore.kernel.org/bpf/20200506000320.28965-5-luke.r.nels@gmail.com
2020-05-05 17:03:20 -07:00
|
|
|
emit_imm(RV_REG_T1, imm, ctx);
|
bpf, riscv: Use compressed instructions in the rv64 JIT
This patch uses the RVC support and encodings from bpf_jit.h to optimize
the rv64 jit.
The optimizations work by replacing emit(rv_X(...)) with a call to a
helper function emit_X, which will emit a compressed version of the
instruction when possible, and when RVC is enabled.
The JIT continues to pass all tests in lib/test_bpf.c, and introduces
no new failures to test_verifier; both with and without RVC being enabled.
Most changes are straightforward replacements of emit(rv_X(...), ctx)
with emit_X(..., ctx), with the following exceptions bearing mention;
* Change emit_imm to sign-extend the value in "lower", since the
checks for RVC (and the instructions themselves) treat the value as
signed. Otherwise, small negative immediates will not be recognized as
encodable using an RVC instruction. For example, without this change,
emit_imm(rd, -1, ctx) would cause lower to become 4095, which is not a
6b int even though a "c.li rd, -1" instruction suffices.
* For {BPF_MOV,BPF_ADD} BPF_X, drop using addiw,addw in the 32-bit
cases since the values are zero-extended into the upper 32 bits in
the following instructions anyways, and the addition commutes with
zero-extension. (BPF_SUB BPF_X must still use subw since subtraction
does not commute with zero-extension.)
This patch avoids optimizing branches and jumps to use RVC instructions
since surrounding code often makes assumptions about the sizes of
emitted instructions. Optimizing these will require changing these
functions (e.g., emit_branch) to dynamically compute jump offsets.
The following are examples of the JITed code for the verifier selftest
"direct packet read test#3 for CGROUP_SKB OK", without and with RVC
enabled, respectively. The former uses 178 bytes, and the latter uses 112,
for a ~37% reduction in code size for this example.
Without RVC:
0: 02000813 addi a6,zero,32
4: fd010113 addi sp,sp,-48
8: 02813423 sd s0,40(sp)
c: 02913023 sd s1,32(sp)
10: 01213c23 sd s2,24(sp)
14: 01313823 sd s3,16(sp)
18: 01413423 sd s4,8(sp)
1c: 03010413 addi s0,sp,48
20: 03056683 lwu a3,48(a0)
24: 02069693 slli a3,a3,0x20
28: 0206d693 srli a3,a3,0x20
2c: 03456703 lwu a4,52(a0)
30: 02071713 slli a4,a4,0x20
34: 02075713 srli a4,a4,0x20
38: 03856483 lwu s1,56(a0)
3c: 02049493 slli s1,s1,0x20
40: 0204d493 srli s1,s1,0x20
44: 03c56903 lwu s2,60(a0)
48: 02091913 slli s2,s2,0x20
4c: 02095913 srli s2,s2,0x20
50: 04056983 lwu s3,64(a0)
54: 02099993 slli s3,s3,0x20
58: 0209d993 srli s3,s3,0x20
5c: 09056a03 lwu s4,144(a0)
60: 020a1a13 slli s4,s4,0x20
64: 020a5a13 srli s4,s4,0x20
68: 00900313 addi t1,zero,9
6c: 006a7463 bgeu s4,t1,0x74
70: 00000a13 addi s4,zero,0
74: 02d52823 sw a3,48(a0)
78: 02e52a23 sw a4,52(a0)
7c: 02952c23 sw s1,56(a0)
80: 03252e23 sw s2,60(a0)
84: 05352023 sw s3,64(a0)
88: 00000793 addi a5,zero,0
8c: 02813403 ld s0,40(sp)
90: 02013483 ld s1,32(sp)
94: 01813903 ld s2,24(sp)
98: 01013983 ld s3,16(sp)
9c: 00813a03 ld s4,8(sp)
a0: 03010113 addi sp,sp,48
a4: 00078513 addi a0,a5,0
a8: 00008067 jalr zero,0(ra)
With RVC:
0: 02000813 addi a6,zero,32
4: 7179 c.addi16sp sp,-48
6: f422 c.sdsp s0,40(sp)
8: f026 c.sdsp s1,32(sp)
a: ec4a c.sdsp s2,24(sp)
c: e84e c.sdsp s3,16(sp)
e: e452 c.sdsp s4,8(sp)
10: 1800 c.addi4spn s0,sp,48
12: 03056683 lwu a3,48(a0)
16: 1682 c.slli a3,0x20
18: 9281 c.srli a3,0x20
1a: 03456703 lwu a4,52(a0)
1e: 1702 c.slli a4,0x20
20: 9301 c.srli a4,0x20
22: 03856483 lwu s1,56(a0)
26: 1482 c.slli s1,0x20
28: 9081 c.srli s1,0x20
2a: 03c56903 lwu s2,60(a0)
2e: 1902 c.slli s2,0x20
30: 02095913 srli s2,s2,0x20
34: 04056983 lwu s3,64(a0)
38: 1982 c.slli s3,0x20
3a: 0209d993 srli s3,s3,0x20
3e: 09056a03 lwu s4,144(a0)
42: 1a02 c.slli s4,0x20
44: 020a5a13 srli s4,s4,0x20
48: 4325 c.li t1,9
4a: 006a7363 bgeu s4,t1,0x50
4e: 4a01 c.li s4,0
50: d914 c.sw a3,48(a0)
52: d958 c.sw a4,52(a0)
54: dd04 c.sw s1,56(a0)
56: 03252e23 sw s2,60(a0)
5a: 05352023 sw s3,64(a0)
5e: 4781 c.li a5,0
60: 7422 c.ldsp s0,40(sp)
62: 7482 c.ldsp s1,32(sp)
64: 6962 c.ldsp s2,24(sp)
66: 69c2 c.ldsp s3,16(sp)
68: 6a22 c.ldsp s4,8(sp)
6a: 6145 c.addi16sp sp,48
6c: 853e c.mv a0,a5
6e: 8082 c.jr ra
Signed-off-by: Luke Nelson <luke.r.nels@gmail.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Cc: Björn Töpel <bjorn.topel@gmail.com>
Link: https://lore.kernel.org/bpf/20200721025241.8077-4-luke.r.nels@gmail.com
2020-07-20 19:52:40 -07:00
|
|
|
emit_and(RV_REG_T1, rd, RV_REG_T1, ctx);
|
2019-12-16 02:13:36 -07:00
|
|
|
}
|
bpf, riscv: Optimize BPF_JSET BPF_K using andi on RV64
This patch optimizes BPF_JSET BPF_K by using a RISC-V andi instruction
when the BPF immediate fits in 12 bits, instead of first loading the
immediate to a temporary register.
Examples of generated code with and without this optimization:
BPF_JMP_IMM(BPF_JSET, R1, 2, 1) without optimization:
20: li t1,2
24: and t1,a0,t1
28: bnez t1,0x30
BPF_JMP_IMM(BPF_JSET, R1, 2, 1) with optimization:
20: andi t1,a0,2
24: bnez t1,0x2c
BPF_JMP32_IMM(BPF_JSET, R1, 2, 1) without optimization:
20: li t1,2
24: mv t2,a0
28: slli t2,t2,0x20
2c: srli t2,t2,0x20
30: slli t1,t1,0x20
34: srli t1,t1,0x20
38: and t1,t2,t1
3c: bnez t1,0x44
BPF_JMP32_IMM(BPF_JSET, R1, 2, 1) with optimization:
20: andi t1,a0,2
24: bnez t1,0x2c
In these examples, because the upper 32 bits of the sign-extended
immediate are 0, BPF_JMP BPF_JSET and BPF_JMP32 BPF_JSET are equivalent
and therefore the JIT produces identical code for them.
Co-developed-by: Xi Wang <xi.wang@gmail.com>
Signed-off-by: Xi Wang <xi.wang@gmail.com>
Signed-off-by: Luke Nelson <luke.r.nels@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Reviewed-by: Björn Töpel <bjorn.topel@gmail.com>
Acked-by: Björn Töpel <bjorn.topel@gmail.com>
Link: https://lore.kernel.org/bpf/20200506000320.28965-5-luke.r.nels@gmail.com
2020-05-05 17:03:20 -07:00
|
|
|
/* For jset32, we should clear the upper 32 bits of t1, but
|
|
|
|
* sign-extension is sufficient here and saves one instruction,
|
|
|
|
* as t1 is used only in comparison against zero.
|
|
|
|
*/
|
|
|
|
if (!is64 && imm < 0)
|
2024-01-15 06:12:30 -07:00
|
|
|
emit_sextw(RV_REG_T1, RV_REG_T1, ctx);
|
bpf, riscv: Optimize BPF_JSET BPF_K using andi on RV64
This patch optimizes BPF_JSET BPF_K by using a RISC-V andi instruction
when the BPF immediate fits in 12 bits, instead of first loading the
immediate to a temporary register.
Examples of generated code with and without this optimization:
BPF_JMP_IMM(BPF_JSET, R1, 2, 1) without optimization:
20: li t1,2
24: and t1,a0,t1
28: bnez t1,0x30
BPF_JMP_IMM(BPF_JSET, R1, 2, 1) with optimization:
20: andi t1,a0,2
24: bnez t1,0x2c
BPF_JMP32_IMM(BPF_JSET, R1, 2, 1) without optimization:
20: li t1,2
24: mv t2,a0
28: slli t2,t2,0x20
2c: srli t2,t2,0x20
30: slli t1,t1,0x20
34: srli t1,t1,0x20
38: and t1,t2,t1
3c: bnez t1,0x44
BPF_JMP32_IMM(BPF_JSET, R1, 2, 1) with optimization:
20: andi t1,a0,2
24: bnez t1,0x2c
In these examples, because the upper 32 bits of the sign-extended
immediate are 0, BPF_JMP BPF_JSET and BPF_JMP32 BPF_JSET are equivalent
and therefore the JIT produces identical code for them.
Co-developed-by: Xi Wang <xi.wang@gmail.com>
Signed-off-by: Xi Wang <xi.wang@gmail.com>
Signed-off-by: Luke Nelson <luke.r.nels@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Reviewed-by: Björn Töpel <bjorn.topel@gmail.com>
Acked-by: Björn Töpel <bjorn.topel@gmail.com>
Link: https://lore.kernel.org/bpf/20200506000320.28965-5-luke.r.nels@gmail.com
2020-05-05 17:03:20 -07:00
|
|
|
e = ctx->ninsns;
|
2020-07-20 19:52:38 -07:00
|
|
|
rvoff -= ninsns_rvoff(e - s);
|
bpf, riscv: Optimize BPF_JSET BPF_K using andi on RV64
This patch optimizes BPF_JSET BPF_K by using a RISC-V andi instruction
when the BPF immediate fits in 12 bits, instead of first loading the
immediate to a temporary register.
Examples of generated code with and without this optimization:
BPF_JMP_IMM(BPF_JSET, R1, 2, 1) without optimization:
20: li t1,2
24: and t1,a0,t1
28: bnez t1,0x30
BPF_JMP_IMM(BPF_JSET, R1, 2, 1) with optimization:
20: andi t1,a0,2
24: bnez t1,0x2c
BPF_JMP32_IMM(BPF_JSET, R1, 2, 1) without optimization:
20: li t1,2
24: mv t2,a0
28: slli t2,t2,0x20
2c: srli t2,t2,0x20
30: slli t1,t1,0x20
34: srli t1,t1,0x20
38: and t1,t2,t1
3c: bnez t1,0x44
BPF_JMP32_IMM(BPF_JSET, R1, 2, 1) with optimization:
20: andi t1,a0,2
24: bnez t1,0x2c
In these examples, because the upper 32 bits of the sign-extended
immediate are 0, BPF_JMP BPF_JSET and BPF_JMP32 BPF_JSET are equivalent
and therefore the JIT produces identical code for them.
Co-developed-by: Xi Wang <xi.wang@gmail.com>
Signed-off-by: Xi Wang <xi.wang@gmail.com>
Signed-off-by: Luke Nelson <luke.r.nels@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Reviewed-by: Björn Töpel <bjorn.topel@gmail.com>
Acked-by: Björn Töpel <bjorn.topel@gmail.com>
Link: https://lore.kernel.org/bpf/20200506000320.28965-5-luke.r.nels@gmail.com
2020-05-05 17:03:20 -07:00
|
|
|
emit_branch(BPF_JNE, RV_REG_T1, RV_REG_ZERO, rvoff, ctx);
|
bpf, riscv: add BPF JIT for RV64G
This commit adds a BPF JIT for RV64G.
The JIT is a two-pass JIT, and has a dynamic prolog/epilogue (similar
to the MIPS64 BPF JIT) instead of static ones (e.g. x86_64).
At the moment the RISC-V Linux port does not support
CONFIG_HAVE_KPROBES, which means that CONFIG_BPF_EVENTS is not
supported. Thus, no tests involving BPF_PROG_TYPE_TRACEPOINT,
BPF_PROG_TYPE_PERF_EVENT, BPF_PROG_TYPE_KPROBE and
BPF_PROG_TYPE_RAW_TRACEPOINT passes.
The implementation does not support "far branching" (>4KiB).
Test results:
# modprobe test_bpf
test_bpf: Summary: 378 PASSED, 0 FAILED, [366/366 JIT'ed]
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier
...
Summary: 761 PASSED, 507 SKIPPED, 2 FAILED
Note that "test_verifier" was run with one build with
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y and one without, otherwise
many of the the tests that require unaligned access were skipped.
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
0
No CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
59
The two failing test_verifier tests are:
"ld_abs: vlan + abs, test 1"
"ld_abs: jump around ld_abs"
This is due to that "far branching" involved in those tests.
All tests where done on QEMU (QEMU emulator version 3.1.50
(v3.1.0-688-g8ae951fbc106)).
Signed-off-by: Björn Töpel <bjorn.topel@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2019-02-05 05:41:22 -07:00
|
|
|
break;
|
|
|
|
|
|
|
|
/* function call */
|
|
|
|
case BPF_JMP | BPF_CALL:
|
|
|
|
{
|
2023-02-15 06:52:03 -07:00
|
|
|
bool fixed_addr;
|
bpf, riscv: add BPF JIT for RV64G
This commit adds a BPF JIT for RV64G.
The JIT is a two-pass JIT, and has a dynamic prolog/epilogue (similar
to the MIPS64 BPF JIT) instead of static ones (e.g. x86_64).
At the moment the RISC-V Linux port does not support
CONFIG_HAVE_KPROBES, which means that CONFIG_BPF_EVENTS is not
supported. Thus, no tests involving BPF_PROG_TYPE_TRACEPOINT,
BPF_PROG_TYPE_PERF_EVENT, BPF_PROG_TYPE_KPROBE and
BPF_PROG_TYPE_RAW_TRACEPOINT passes.
The implementation does not support "far branching" (>4KiB).
Test results:
# modprobe test_bpf
test_bpf: Summary: 378 PASSED, 0 FAILED, [366/366 JIT'ed]
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier
...
Summary: 761 PASSED, 507 SKIPPED, 2 FAILED
Note that "test_verifier" was run with one build with
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y and one without, otherwise
many of the the tests that require unaligned access were skipped.
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
0
No CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
59
The two failing test_verifier tests are:
"ld_abs: vlan + abs, test 1"
"ld_abs: jump around ld_abs"
This is due to that "far branching" involved in those tests.
All tests where done on QEMU (QEMU emulator version 3.1.50
(v3.1.0-688-g8ae951fbc106)).
Signed-off-by: Björn Töpel <bjorn.topel@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2019-02-05 05:41:22 -07:00
|
|
|
u64 addr;
|
|
|
|
|
2024-05-02 08:18:52 -07:00
|
|
|
/* Inline calls to bpf_get_smp_processor_id()
|
|
|
|
*
|
|
|
|
* RV_REG_TP holds the address of the current CPU's task_struct and thread_info is
|
|
|
|
* at offset 0 in task_struct.
|
|
|
|
* Load cpu from thread_info:
|
|
|
|
* Set R0 to ((struct thread_info *)(RV_REG_TP))->cpu
|
|
|
|
*
|
|
|
|
* This replicates the implementation of raw_smp_processor_id() on RISCV
|
|
|
|
*/
|
|
|
|
if (insn->src_reg == 0 && insn->imm == BPF_FUNC_get_smp_processor_id) {
|
|
|
|
/* Load current CPU number in R0 */
|
|
|
|
emit_ld(bpf_to_rv_reg(BPF_REG_0, ctx), offsetof(struct thread_info, cpu),
|
|
|
|
RV_REG_TP, ctx);
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
|
bpf, riscv: add BPF JIT for RV64G
This commit adds a BPF JIT for RV64G.
The JIT is a two-pass JIT, and has a dynamic prolog/epilogue (similar
to the MIPS64 BPF JIT) instead of static ones (e.g. x86_64).
At the moment the RISC-V Linux port does not support
CONFIG_HAVE_KPROBES, which means that CONFIG_BPF_EVENTS is not
supported. Thus, no tests involving BPF_PROG_TYPE_TRACEPOINT,
BPF_PROG_TYPE_PERF_EVENT, BPF_PROG_TYPE_KPROBE and
BPF_PROG_TYPE_RAW_TRACEPOINT passes.
The implementation does not support "far branching" (>4KiB).
Test results:
# modprobe test_bpf
test_bpf: Summary: 378 PASSED, 0 FAILED, [366/366 JIT'ed]
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier
...
Summary: 761 PASSED, 507 SKIPPED, 2 FAILED
Note that "test_verifier" was run with one build with
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y and one without, otherwise
many of the the tests that require unaligned access were skipped.
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
0
No CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
59
The two failing test_verifier tests are:
"ld_abs: vlan + abs, test 1"
"ld_abs: jump around ld_abs"
This is due to that "far branching" involved in those tests.
All tests where done on QEMU (QEMU emulator version 3.1.50
(v3.1.0-688-g8ae951fbc106)).
Signed-off-by: Björn Töpel <bjorn.topel@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2019-02-05 05:41:22 -07:00
|
|
|
mark_call(ctx);
|
2023-02-15 06:52:03 -07:00
|
|
|
ret = bpf_jit_get_func_addr(ctx->prog, insn, extra_pass,
|
|
|
|
&addr, &fixed_addr);
|
bpf, riscv: add BPF JIT for RV64G
This commit adds a BPF JIT for RV64G.
The JIT is a two-pass JIT, and has a dynamic prolog/epilogue (similar
to the MIPS64 BPF JIT) instead of static ones (e.g. x86_64).
At the moment the RISC-V Linux port does not support
CONFIG_HAVE_KPROBES, which means that CONFIG_BPF_EVENTS is not
supported. Thus, no tests involving BPF_PROG_TYPE_TRACEPOINT,
BPF_PROG_TYPE_PERF_EVENT, BPF_PROG_TYPE_KPROBE and
BPF_PROG_TYPE_RAW_TRACEPOINT passes.
The implementation does not support "far branching" (>4KiB).
Test results:
# modprobe test_bpf
test_bpf: Summary: 378 PASSED, 0 FAILED, [366/366 JIT'ed]
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier
...
Summary: 761 PASSED, 507 SKIPPED, 2 FAILED
Note that "test_verifier" was run with one build with
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y and one without, otherwise
many of the the tests that require unaligned access were skipped.
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
0
No CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
59
The two failing test_verifier tests are:
"ld_abs: vlan + abs, test 1"
"ld_abs: jump around ld_abs"
This is due to that "far branching" involved in those tests.
All tests where done on QEMU (QEMU emulator version 3.1.50
(v3.1.0-688-g8ae951fbc106)).
Signed-off-by: Björn Töpel <bjorn.topel@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2019-02-05 05:41:22 -07:00
|
|
|
if (ret < 0)
|
|
|
|
return ret;
|
2023-02-15 06:52:03 -07:00
|
|
|
|
riscv, bpf: Fix kfunc parameters incompatibility between bpf and riscv abi
We encountered a failing case when running selftest in no_alu32 mode:
The failure case is `kfunc_call/kfunc_call_test4` and its source code is
like bellow:
```
long bpf_kfunc_call_test4(signed char a, short b, int c, long d) __ksym;
int kfunc_call_test4(struct __sk_buff *skb)
{
...
tmp = bpf_kfunc_call_test4(-3, -30, -200, -1000);
...
}
```
And its corresponding asm code is:
```
0: r1 = -3
1: r2 = -30
2: r3 = 0xffffff38 # opcode: 18 03 00 00 38 ff ff ff 00 00 00 00 00 00 00 00
4: r4 = -1000
5: call bpf_kfunc_call_test4
```
insn 2 is parsed to ld_imm64 insn to emit 0x00000000ffffff38 imm, and
converted to int type and then send to bpf_kfunc_call_test4. But since
it is zero-extended in the bpf calling convention, riscv jit will
directly treat it as an unsigned 32-bit int value, and then fails with
the message "actual 4294966063 != expected -1234".
The reason is the incompatibility between bpf and riscv abi, that is,
bpf will do zero-extension on uint, but riscv64 requires sign-extension
on int or uint. We can solve this problem by sign extending the 32-bit
parameters in kfunc.
The issue is related to [0], and thanks to Yonghong and Alexei.
Link: https://github.com/llvm/llvm-project/pull/84874 [0]
Fixes: d40c3847b485 ("riscv, bpf: Add kfunc support for RV64")
Signed-off-by: Pu Lehui <pulehui@huawei.com>
Tested-by: Puranjay Mohan <puranjay12@gmail.com>
Reviewed-by: Puranjay Mohan <puranjay12@gmail.com>
Link: https://lore.kernel.org/r/20240324103306.2202954-1-pulehui@huaweicloud.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2024-03-24 03:33:06 -07:00
|
|
|
if (insn->src_reg == BPF_PSEUDO_KFUNC_CALL) {
|
|
|
|
const struct btf_func_model *fm;
|
|
|
|
int idx;
|
|
|
|
|
|
|
|
fm = bpf_jit_find_kfunc_model(ctx->prog, insn);
|
|
|
|
if (!fm)
|
|
|
|
return -EINVAL;
|
|
|
|
|
|
|
|
for (idx = 0; idx < fm->nr_args; idx++) {
|
|
|
|
u8 reg = bpf_to_rv_reg(BPF_REG_1 + idx, ctx);
|
|
|
|
|
|
|
|
if (fm->arg_size[idx] == sizeof(int))
|
|
|
|
emit_sextw(reg, reg, ctx);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2023-02-15 06:52:03 -07:00
|
|
|
ret = emit_call(addr, fixed_addr, ctx);
|
2019-12-16 02:13:41 -07:00
|
|
|
if (ret)
|
|
|
|
return ret;
|
2023-02-15 06:52:03 -07:00
|
|
|
|
riscv, bpf: Sign-extend return values
The RISC-V architecture does not expose sub-registers, and hold all
32-bit values in a sign-extended format [1] [2]:
| The compiler and calling convention maintain an invariant that all
| 32-bit values are held in a sign-extended format in 64-bit
| registers. Even 32-bit unsigned integers extend bit 31 into bits
| 63 through 32. Consequently, conversion between unsigned and
| signed 32-bit integers is a no-op, as is conversion from a signed
| 32-bit integer to a signed 64-bit integer.
While BPF, on the other hand, exposes sub-registers, and use
zero-extension (similar to arm64/x86).
This has led to some subtle bugs, where a BPF JITted program has not
sign-extended the a0 register (return value in RISC-V land), passed
the return value up the kernel, e.g.:
| int from_bpf(void);
|
| long foo(void)
| {
| return from_bpf();
| }
Here, a0 would be 0xffff_ffff, instead of the expected
0xffff_ffff_ffff_ffff.
Internally, the RISC-V JIT uses a5 as a dedicated register for BPF
return values.
Keep a5 zero-extended, but explicitly sign-extend a0 (which is used
outside BPF land). Now that a0 (RISC-V ABI) and a5 (BPF ABI) differs,
a0 is only moved to a5 for non-BPF native calls (BPF_PSEUDO_CALL).
Fixes: 2353ecc6f91f ("bpf, riscv: add BPF JIT for RV64G")
Signed-off-by: Björn Töpel <bjorn@rivosinc.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Link: https://github.com/riscv/riscv-isa-manual/releases/download/riscv-isa-release-056b6ff-2023-10-02/unpriv-isa-asciidoc.pdf # [2]
Link: https://github.com/riscv-non-isa/riscv-elf-psabi-doc/releases/download/draft-20230929-e5c800e661a53efe3c2678d71a306323b60eb13b/riscv-abi.pdf # [2]
Link: https://lore.kernel.org/bpf/20231004120706.52848-2-bjorn@kernel.org
2023-10-04 05:07:05 -07:00
|
|
|
if (insn->src_reg != BPF_PSEUDO_CALL)
|
|
|
|
emit_mv(bpf_to_rv_reg(BPF_REG_0, ctx), RV_REG_A0, ctx);
|
bpf, riscv: add BPF JIT for RV64G
This commit adds a BPF JIT for RV64G.
The JIT is a two-pass JIT, and has a dynamic prolog/epilogue (similar
to the MIPS64 BPF JIT) instead of static ones (e.g. x86_64).
At the moment the RISC-V Linux port does not support
CONFIG_HAVE_KPROBES, which means that CONFIG_BPF_EVENTS is not
supported. Thus, no tests involving BPF_PROG_TYPE_TRACEPOINT,
BPF_PROG_TYPE_PERF_EVENT, BPF_PROG_TYPE_KPROBE and
BPF_PROG_TYPE_RAW_TRACEPOINT passes.
The implementation does not support "far branching" (>4KiB).
Test results:
# modprobe test_bpf
test_bpf: Summary: 378 PASSED, 0 FAILED, [366/366 JIT'ed]
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier
...
Summary: 761 PASSED, 507 SKIPPED, 2 FAILED
Note that "test_verifier" was run with one build with
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y and one without, otherwise
many of the the tests that require unaligned access were skipped.
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
0
No CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
59
The two failing test_verifier tests are:
"ld_abs: vlan + abs, test 1"
"ld_abs: jump around ld_abs"
This is due to that "far branching" involved in those tests.
All tests where done on QEMU (QEMU emulator version 3.1.50
(v3.1.0-688-g8ae951fbc106)).
Signed-off-by: Björn Töpel <bjorn.topel@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2019-02-05 05:41:22 -07:00
|
|
|
break;
|
|
|
|
}
|
|
|
|
/* tail call */
|
|
|
|
case BPF_JMP | BPF_TAIL_CALL:
|
|
|
|
if (emit_bpf_tail_call(i, ctx))
|
|
|
|
return -1;
|
|
|
|
break;
|
|
|
|
|
|
|
|
/* function return */
|
|
|
|
case BPF_JMP | BPF_EXIT:
|
|
|
|
if (i == ctx->prog->len - 1)
|
|
|
|
break;
|
|
|
|
|
|
|
|
rvoff = epilogue_offset(ctx);
|
2023-02-15 06:52:03 -07:00
|
|
|
ret = emit_jump_and_link(RV_REG_ZERO, rvoff, true, ctx);
|
2020-04-06 15:16:04 -07:00
|
|
|
if (ret)
|
|
|
|
return ret;
|
bpf, riscv: add BPF JIT for RV64G
This commit adds a BPF JIT for RV64G.
The JIT is a two-pass JIT, and has a dynamic prolog/epilogue (similar
to the MIPS64 BPF JIT) instead of static ones (e.g. x86_64).
At the moment the RISC-V Linux port does not support
CONFIG_HAVE_KPROBES, which means that CONFIG_BPF_EVENTS is not
supported. Thus, no tests involving BPF_PROG_TYPE_TRACEPOINT,
BPF_PROG_TYPE_PERF_EVENT, BPF_PROG_TYPE_KPROBE and
BPF_PROG_TYPE_RAW_TRACEPOINT passes.
The implementation does not support "far branching" (>4KiB).
Test results:
# modprobe test_bpf
test_bpf: Summary: 378 PASSED, 0 FAILED, [366/366 JIT'ed]
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier
...
Summary: 761 PASSED, 507 SKIPPED, 2 FAILED
Note that "test_verifier" was run with one build with
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y and one without, otherwise
many of the the tests that require unaligned access were skipped.
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
0
No CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
59
The two failing test_verifier tests are:
"ld_abs: vlan + abs, test 1"
"ld_abs: jump around ld_abs"
This is due to that "far branching" involved in those tests.
All tests where done on QEMU (QEMU emulator version 3.1.50
(v3.1.0-688-g8ae951fbc106)).
Signed-off-by: Björn Töpel <bjorn.topel@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2019-02-05 05:41:22 -07:00
|
|
|
break;
|
|
|
|
|
|
|
|
/* dst = imm64 */
|
|
|
|
case BPF_LD | BPF_IMM | BPF_DW:
|
|
|
|
{
|
|
|
|
struct bpf_insn insn1 = insn[1];
|
|
|
|
u64 imm64;
|
|
|
|
|
|
|
|
imm64 = (u64)insn1.imm << 32 | (u32)imm;
|
2022-12-06 02:14:10 -07:00
|
|
|
if (bpf_pseudo_func(insn)) {
|
|
|
|
/* fixed-length insns for extra jit pass */
|
|
|
|
ret = emit_addr(rd, imm64, extra_pass, ctx);
|
|
|
|
if (ret)
|
|
|
|
return ret;
|
|
|
|
} else {
|
|
|
|
emit_imm(rd, imm64, ctx);
|
|
|
|
}
|
|
|
|
|
bpf, riscv: add BPF JIT for RV64G
This commit adds a BPF JIT for RV64G.
The JIT is a two-pass JIT, and has a dynamic prolog/epilogue (similar
to the MIPS64 BPF JIT) instead of static ones (e.g. x86_64).
At the moment the RISC-V Linux port does not support
CONFIG_HAVE_KPROBES, which means that CONFIG_BPF_EVENTS is not
supported. Thus, no tests involving BPF_PROG_TYPE_TRACEPOINT,
BPF_PROG_TYPE_PERF_EVENT, BPF_PROG_TYPE_KPROBE and
BPF_PROG_TYPE_RAW_TRACEPOINT passes.
The implementation does not support "far branching" (>4KiB).
Test results:
# modprobe test_bpf
test_bpf: Summary: 378 PASSED, 0 FAILED, [366/366 JIT'ed]
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier
...
Summary: 761 PASSED, 507 SKIPPED, 2 FAILED
Note that "test_verifier" was run with one build with
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y and one without, otherwise
many of the the tests that require unaligned access were skipped.
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
0
No CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
59
The two failing test_verifier tests are:
"ld_abs: vlan + abs, test 1"
"ld_abs: jump around ld_abs"
This is due to that "far branching" involved in those tests.
All tests where done on QEMU (QEMU emulator version 3.1.50
(v3.1.0-688-g8ae951fbc106)).
Signed-off-by: Björn Töpel <bjorn.topel@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2019-02-05 05:41:22 -07:00
|
|
|
return 1;
|
|
|
|
}
|
|
|
|
|
2023-08-24 02:49:56 -07:00
|
|
|
/* LDX: dst = *(unsigned size *)(src + off) */
|
bpf, riscv: add BPF JIT for RV64G
This commit adds a BPF JIT for RV64G.
The JIT is a two-pass JIT, and has a dynamic prolog/epilogue (similar
to the MIPS64 BPF JIT) instead of static ones (e.g. x86_64).
At the moment the RISC-V Linux port does not support
CONFIG_HAVE_KPROBES, which means that CONFIG_BPF_EVENTS is not
supported. Thus, no tests involving BPF_PROG_TYPE_TRACEPOINT,
BPF_PROG_TYPE_PERF_EVENT, BPF_PROG_TYPE_KPROBE and
BPF_PROG_TYPE_RAW_TRACEPOINT passes.
The implementation does not support "far branching" (>4KiB).
Test results:
# modprobe test_bpf
test_bpf: Summary: 378 PASSED, 0 FAILED, [366/366 JIT'ed]
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier
...
Summary: 761 PASSED, 507 SKIPPED, 2 FAILED
Note that "test_verifier" was run with one build with
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y and one without, otherwise
many of the the tests that require unaligned access were skipped.
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
0
No CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
59
The two failing test_verifier tests are:
"ld_abs: vlan + abs, test 1"
"ld_abs: jump around ld_abs"
This is due to that "far branching" involved in those tests.
All tests where done on QEMU (QEMU emulator version 3.1.50
(v3.1.0-688-g8ae951fbc106)).
Signed-off-by: Björn Töpel <bjorn.topel@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2019-02-05 05:41:22 -07:00
|
|
|
case BPF_LDX | BPF_MEM | BPF_B:
|
|
|
|
case BPF_LDX | BPF_MEM | BPF_H:
|
|
|
|
case BPF_LDX | BPF_MEM | BPF_W:
|
|
|
|
case BPF_LDX | BPF_MEM | BPF_DW:
|
2021-10-27 04:18:22 -07:00
|
|
|
case BPF_LDX | BPF_PROBE_MEM | BPF_B:
|
|
|
|
case BPF_LDX | BPF_PROBE_MEM | BPF_H:
|
|
|
|
case BPF_LDX | BPF_PROBE_MEM | BPF_W:
|
|
|
|
case BPF_LDX | BPF_PROBE_MEM | BPF_DW:
|
2023-08-24 02:49:56 -07:00
|
|
|
/* LDSX: dst = *(signed size *)(src + off) */
|
|
|
|
case BPF_LDX | BPF_MEMSX | BPF_B:
|
|
|
|
case BPF_LDX | BPF_MEMSX | BPF_H:
|
|
|
|
case BPF_LDX | BPF_MEMSX | BPF_W:
|
|
|
|
case BPF_LDX | BPF_PROBE_MEMSX | BPF_B:
|
|
|
|
case BPF_LDX | BPF_PROBE_MEMSX | BPF_H:
|
|
|
|
case BPF_LDX | BPF_PROBE_MEMSX | BPF_W:
|
2024-04-04 04:42:02 -07:00
|
|
|
/* LDX | PROBE_MEM32: dst = *(unsigned size *)(src + RV_REG_ARENA + off) */
|
|
|
|
case BPF_LDX | BPF_PROBE_MEM32 | BPF_B:
|
|
|
|
case BPF_LDX | BPF_PROBE_MEM32 | BPF_H:
|
|
|
|
case BPF_LDX | BPF_PROBE_MEM32 | BPF_W:
|
|
|
|
case BPF_LDX | BPF_PROBE_MEM32 | BPF_DW:
|
2021-10-27 04:18:22 -07:00
|
|
|
{
|
|
|
|
int insn_len, insns_start;
|
2023-08-24 02:49:56 -07:00
|
|
|
bool sign_ext;
|
|
|
|
|
|
|
|
sign_ext = BPF_MODE(insn->code) == BPF_MEMSX ||
|
|
|
|
BPF_MODE(insn->code) == BPF_PROBE_MEMSX;
|
2021-10-27 04:18:22 -07:00
|
|
|
|
2024-04-04 04:42:02 -07:00
|
|
|
if (BPF_MODE(insn->code) == BPF_PROBE_MEM32) {
|
|
|
|
emit_add(RV_REG_T2, rs, RV_REG_ARENA, ctx);
|
|
|
|
rs = RV_REG_T2;
|
|
|
|
}
|
|
|
|
|
2021-10-27 04:18:22 -07:00
|
|
|
switch (BPF_SIZE(code)) {
|
|
|
|
case BPF_B:
|
|
|
|
if (is_12b_int(off)) {
|
|
|
|
insns_start = ctx->ninsns;
|
2023-08-24 02:49:56 -07:00
|
|
|
if (sign_ext)
|
|
|
|
emit(rv_lb(rd, off, rs), ctx);
|
|
|
|
else
|
|
|
|
emit(rv_lbu(rd, off, rs), ctx);
|
2021-10-27 04:18:22 -07:00
|
|
|
insn_len = ctx->ninsns - insns_start;
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
|
|
|
|
emit_imm(RV_REG_T1, off, ctx);
|
|
|
|
emit_add(RV_REG_T1, RV_REG_T1, rs, ctx);
|
|
|
|
insns_start = ctx->ninsns;
|
2023-08-24 02:49:56 -07:00
|
|
|
if (sign_ext)
|
|
|
|
emit(rv_lb(rd, 0, RV_REG_T1), ctx);
|
|
|
|
else
|
|
|
|
emit(rv_lbu(rd, 0, RV_REG_T1), ctx);
|
2021-10-27 04:18:22 -07:00
|
|
|
insn_len = ctx->ninsns - insns_start;
|
|
|
|
break;
|
|
|
|
case BPF_H:
|
|
|
|
if (is_12b_int(off)) {
|
|
|
|
insns_start = ctx->ninsns;
|
2023-08-24 02:49:56 -07:00
|
|
|
if (sign_ext)
|
|
|
|
emit(rv_lh(rd, off, rs), ctx);
|
|
|
|
else
|
|
|
|
emit(rv_lhu(rd, off, rs), ctx);
|
2021-10-27 04:18:22 -07:00
|
|
|
insn_len = ctx->ninsns - insns_start;
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
|
|
|
|
emit_imm(RV_REG_T1, off, ctx);
|
|
|
|
emit_add(RV_REG_T1, RV_REG_T1, rs, ctx);
|
|
|
|
insns_start = ctx->ninsns;
|
2023-08-24 02:49:56 -07:00
|
|
|
if (sign_ext)
|
|
|
|
emit(rv_lh(rd, 0, RV_REG_T1), ctx);
|
|
|
|
else
|
|
|
|
emit(rv_lhu(rd, 0, RV_REG_T1), ctx);
|
2021-10-27 04:18:22 -07:00
|
|
|
insn_len = ctx->ninsns - insns_start;
|
|
|
|
break;
|
|
|
|
case BPF_W:
|
|
|
|
if (is_12b_int(off)) {
|
|
|
|
insns_start = ctx->ninsns;
|
2023-08-24 02:49:56 -07:00
|
|
|
if (sign_ext)
|
|
|
|
emit(rv_lw(rd, off, rs), ctx);
|
|
|
|
else
|
|
|
|
emit(rv_lwu(rd, off, rs), ctx);
|
2021-10-27 04:18:22 -07:00
|
|
|
insn_len = ctx->ninsns - insns_start;
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
|
|
|
|
emit_imm(RV_REG_T1, off, ctx);
|
|
|
|
emit_add(RV_REG_T1, RV_REG_T1, rs, ctx);
|
|
|
|
insns_start = ctx->ninsns;
|
2023-08-24 02:49:56 -07:00
|
|
|
if (sign_ext)
|
|
|
|
emit(rv_lw(rd, 0, RV_REG_T1), ctx);
|
|
|
|
else
|
|
|
|
emit(rv_lwu(rd, 0, RV_REG_T1), ctx);
|
2021-10-27 04:18:22 -07:00
|
|
|
insn_len = ctx->ninsns - insns_start;
|
|
|
|
break;
|
|
|
|
case BPF_DW:
|
|
|
|
if (is_12b_int(off)) {
|
|
|
|
insns_start = ctx->ninsns;
|
|
|
|
emit_ld(rd, off, rs, ctx);
|
|
|
|
insn_len = ctx->ninsns - insns_start;
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
|
|
|
|
emit_imm(RV_REG_T1, off, ctx);
|
|
|
|
emit_add(RV_REG_T1, RV_REG_T1, rs, ctx);
|
|
|
|
insns_start = ctx->ninsns;
|
|
|
|
emit_ld(rd, 0, RV_REG_T1, ctx);
|
|
|
|
insn_len = ctx->ninsns - insns_start;
|
bpf, riscv: add BPF JIT for RV64G
This commit adds a BPF JIT for RV64G.
The JIT is a two-pass JIT, and has a dynamic prolog/epilogue (similar
to the MIPS64 BPF JIT) instead of static ones (e.g. x86_64).
At the moment the RISC-V Linux port does not support
CONFIG_HAVE_KPROBES, which means that CONFIG_BPF_EVENTS is not
supported. Thus, no tests involving BPF_PROG_TYPE_TRACEPOINT,
BPF_PROG_TYPE_PERF_EVENT, BPF_PROG_TYPE_KPROBE and
BPF_PROG_TYPE_RAW_TRACEPOINT passes.
The implementation does not support "far branching" (>4KiB).
Test results:
# modprobe test_bpf
test_bpf: Summary: 378 PASSED, 0 FAILED, [366/366 JIT'ed]
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier
...
Summary: 761 PASSED, 507 SKIPPED, 2 FAILED
Note that "test_verifier" was run with one build with
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y and one without, otherwise
many of the the tests that require unaligned access were skipped.
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
0
No CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
59
The two failing test_verifier tests are:
"ld_abs: vlan + abs, test 1"
"ld_abs: jump around ld_abs"
This is due to that "far branching" involved in those tests.
All tests where done on QEMU (QEMU emulator version 3.1.50
(v3.1.0-688-g8ae951fbc106)).
Signed-off-by: Björn Töpel <bjorn.topel@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2019-02-05 05:41:22 -07:00
|
|
|
break;
|
|
|
|
}
|
|
|
|
|
2021-10-27 04:18:22 -07:00
|
|
|
ret = add_exception_handler(insn, ctx, rd, insn_len);
|
|
|
|
if (ret)
|
|
|
|
return ret;
|
2023-08-24 02:49:55 -07:00
|
|
|
|
|
|
|
if (BPF_SIZE(code) != BPF_DW && insn_is_zext(&insn[1]))
|
|
|
|
return 1;
|
bpf, riscv: add BPF JIT for RV64G
This commit adds a BPF JIT for RV64G.
The JIT is a two-pass JIT, and has a dynamic prolog/epilogue (similar
to the MIPS64 BPF JIT) instead of static ones (e.g. x86_64).
At the moment the RISC-V Linux port does not support
CONFIG_HAVE_KPROBES, which means that CONFIG_BPF_EVENTS is not
supported. Thus, no tests involving BPF_PROG_TYPE_TRACEPOINT,
BPF_PROG_TYPE_PERF_EVENT, BPF_PROG_TYPE_KPROBE and
BPF_PROG_TYPE_RAW_TRACEPOINT passes.
The implementation does not support "far branching" (>4KiB).
Test results:
# modprobe test_bpf
test_bpf: Summary: 378 PASSED, 0 FAILED, [366/366 JIT'ed]
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier
...
Summary: 761 PASSED, 507 SKIPPED, 2 FAILED
Note that "test_verifier" was run with one build with
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y and one without, otherwise
many of the the tests that require unaligned access were skipped.
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
0
No CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
59
The two failing test_verifier tests are:
"ld_abs: vlan + abs, test 1"
"ld_abs: jump around ld_abs"
This is due to that "far branching" involved in those tests.
All tests where done on QEMU (QEMU emulator version 3.1.50
(v3.1.0-688-g8ae951fbc106)).
Signed-off-by: Björn Töpel <bjorn.topel@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2019-02-05 05:41:22 -07:00
|
|
|
break;
|
2021-10-27 04:18:22 -07:00
|
|
|
}
|
2021-07-13 01:18:31 -07:00
|
|
|
/* speculation barrier */
|
|
|
|
case BPF_ST | BPF_NOSPEC:
|
|
|
|
break;
|
|
|
|
|
bpf, riscv: add BPF JIT for RV64G
This commit adds a BPF JIT for RV64G.
The JIT is a two-pass JIT, and has a dynamic prolog/epilogue (similar
to the MIPS64 BPF JIT) instead of static ones (e.g. x86_64).
At the moment the RISC-V Linux port does not support
CONFIG_HAVE_KPROBES, which means that CONFIG_BPF_EVENTS is not
supported. Thus, no tests involving BPF_PROG_TYPE_TRACEPOINT,
BPF_PROG_TYPE_PERF_EVENT, BPF_PROG_TYPE_KPROBE and
BPF_PROG_TYPE_RAW_TRACEPOINT passes.
The implementation does not support "far branching" (>4KiB).
Test results:
# modprobe test_bpf
test_bpf: Summary: 378 PASSED, 0 FAILED, [366/366 JIT'ed]
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier
...
Summary: 761 PASSED, 507 SKIPPED, 2 FAILED
Note that "test_verifier" was run with one build with
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y and one without, otherwise
many of the the tests that require unaligned access were skipped.
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
0
No CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
59
The two failing test_verifier tests are:
"ld_abs: vlan + abs, test 1"
"ld_abs: jump around ld_abs"
This is due to that "far branching" involved in those tests.
All tests where done on QEMU (QEMU emulator version 3.1.50
(v3.1.0-688-g8ae951fbc106)).
Signed-off-by: Björn Töpel <bjorn.topel@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2019-02-05 05:41:22 -07:00
|
|
|
/* ST: *(size *)(dst + off) = imm */
|
|
|
|
case BPF_ST | BPF_MEM | BPF_B:
|
|
|
|
emit_imm(RV_REG_T1, imm, ctx);
|
|
|
|
if (is_12b_int(off)) {
|
|
|
|
emit(rv_sb(rd, off, RV_REG_T1), ctx);
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
|
|
|
|
emit_imm(RV_REG_T2, off, ctx);
|
bpf, riscv: Use compressed instructions in the rv64 JIT
This patch uses the RVC support and encodings from bpf_jit.h to optimize
the rv64 jit.
The optimizations work by replacing emit(rv_X(...)) with a call to a
helper function emit_X, which will emit a compressed version of the
instruction when possible, and when RVC is enabled.
The JIT continues to pass all tests in lib/test_bpf.c, and introduces
no new failures to test_verifier; both with and without RVC being enabled.
Most changes are straightforward replacements of emit(rv_X(...), ctx)
with emit_X(..., ctx), with the following exceptions bearing mention;
* Change emit_imm to sign-extend the value in "lower", since the
checks for RVC (and the instructions themselves) treat the value as
signed. Otherwise, small negative immediates will not be recognized as
encodable using an RVC instruction. For example, without this change,
emit_imm(rd, -1, ctx) would cause lower to become 4095, which is not a
6b int even though a "c.li rd, -1" instruction suffices.
* For {BPF_MOV,BPF_ADD} BPF_X, drop using addiw,addw in the 32-bit
cases since the values are zero-extended into the upper 32 bits in
the following instructions anyways, and the addition commutes with
zero-extension. (BPF_SUB BPF_X must still use subw since subtraction
does not commute with zero-extension.)
This patch avoids optimizing branches and jumps to use RVC instructions
since surrounding code often makes assumptions about the sizes of
emitted instructions. Optimizing these will require changing these
functions (e.g., emit_branch) to dynamically compute jump offsets.
The following are examples of the JITed code for the verifier selftest
"direct packet read test#3 for CGROUP_SKB OK", without and with RVC
enabled, respectively. The former uses 178 bytes, and the latter uses 112,
for a ~37% reduction in code size for this example.
Without RVC:
0: 02000813 addi a6,zero,32
4: fd010113 addi sp,sp,-48
8: 02813423 sd s0,40(sp)
c: 02913023 sd s1,32(sp)
10: 01213c23 sd s2,24(sp)
14: 01313823 sd s3,16(sp)
18: 01413423 sd s4,8(sp)
1c: 03010413 addi s0,sp,48
20: 03056683 lwu a3,48(a0)
24: 02069693 slli a3,a3,0x20
28: 0206d693 srli a3,a3,0x20
2c: 03456703 lwu a4,52(a0)
30: 02071713 slli a4,a4,0x20
34: 02075713 srli a4,a4,0x20
38: 03856483 lwu s1,56(a0)
3c: 02049493 slli s1,s1,0x20
40: 0204d493 srli s1,s1,0x20
44: 03c56903 lwu s2,60(a0)
48: 02091913 slli s2,s2,0x20
4c: 02095913 srli s2,s2,0x20
50: 04056983 lwu s3,64(a0)
54: 02099993 slli s3,s3,0x20
58: 0209d993 srli s3,s3,0x20
5c: 09056a03 lwu s4,144(a0)
60: 020a1a13 slli s4,s4,0x20
64: 020a5a13 srli s4,s4,0x20
68: 00900313 addi t1,zero,9
6c: 006a7463 bgeu s4,t1,0x74
70: 00000a13 addi s4,zero,0
74: 02d52823 sw a3,48(a0)
78: 02e52a23 sw a4,52(a0)
7c: 02952c23 sw s1,56(a0)
80: 03252e23 sw s2,60(a0)
84: 05352023 sw s3,64(a0)
88: 00000793 addi a5,zero,0
8c: 02813403 ld s0,40(sp)
90: 02013483 ld s1,32(sp)
94: 01813903 ld s2,24(sp)
98: 01013983 ld s3,16(sp)
9c: 00813a03 ld s4,8(sp)
a0: 03010113 addi sp,sp,48
a4: 00078513 addi a0,a5,0
a8: 00008067 jalr zero,0(ra)
With RVC:
0: 02000813 addi a6,zero,32
4: 7179 c.addi16sp sp,-48
6: f422 c.sdsp s0,40(sp)
8: f026 c.sdsp s1,32(sp)
a: ec4a c.sdsp s2,24(sp)
c: e84e c.sdsp s3,16(sp)
e: e452 c.sdsp s4,8(sp)
10: 1800 c.addi4spn s0,sp,48
12: 03056683 lwu a3,48(a0)
16: 1682 c.slli a3,0x20
18: 9281 c.srli a3,0x20
1a: 03456703 lwu a4,52(a0)
1e: 1702 c.slli a4,0x20
20: 9301 c.srli a4,0x20
22: 03856483 lwu s1,56(a0)
26: 1482 c.slli s1,0x20
28: 9081 c.srli s1,0x20
2a: 03c56903 lwu s2,60(a0)
2e: 1902 c.slli s2,0x20
30: 02095913 srli s2,s2,0x20
34: 04056983 lwu s3,64(a0)
38: 1982 c.slli s3,0x20
3a: 0209d993 srli s3,s3,0x20
3e: 09056a03 lwu s4,144(a0)
42: 1a02 c.slli s4,0x20
44: 020a5a13 srli s4,s4,0x20
48: 4325 c.li t1,9
4a: 006a7363 bgeu s4,t1,0x50
4e: 4a01 c.li s4,0
50: d914 c.sw a3,48(a0)
52: d958 c.sw a4,52(a0)
54: dd04 c.sw s1,56(a0)
56: 03252e23 sw s2,60(a0)
5a: 05352023 sw s3,64(a0)
5e: 4781 c.li a5,0
60: 7422 c.ldsp s0,40(sp)
62: 7482 c.ldsp s1,32(sp)
64: 6962 c.ldsp s2,24(sp)
66: 69c2 c.ldsp s3,16(sp)
68: 6a22 c.ldsp s4,8(sp)
6a: 6145 c.addi16sp sp,48
6c: 853e c.mv a0,a5
6e: 8082 c.jr ra
Signed-off-by: Luke Nelson <luke.r.nels@gmail.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Cc: Björn Töpel <bjorn.topel@gmail.com>
Link: https://lore.kernel.org/bpf/20200721025241.8077-4-luke.r.nels@gmail.com
2020-07-20 19:52:40 -07:00
|
|
|
emit_add(RV_REG_T2, RV_REG_T2, rd, ctx);
|
bpf, riscv: add BPF JIT for RV64G
This commit adds a BPF JIT for RV64G.
The JIT is a two-pass JIT, and has a dynamic prolog/epilogue (similar
to the MIPS64 BPF JIT) instead of static ones (e.g. x86_64).
At the moment the RISC-V Linux port does not support
CONFIG_HAVE_KPROBES, which means that CONFIG_BPF_EVENTS is not
supported. Thus, no tests involving BPF_PROG_TYPE_TRACEPOINT,
BPF_PROG_TYPE_PERF_EVENT, BPF_PROG_TYPE_KPROBE and
BPF_PROG_TYPE_RAW_TRACEPOINT passes.
The implementation does not support "far branching" (>4KiB).
Test results:
# modprobe test_bpf
test_bpf: Summary: 378 PASSED, 0 FAILED, [366/366 JIT'ed]
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier
...
Summary: 761 PASSED, 507 SKIPPED, 2 FAILED
Note that "test_verifier" was run with one build with
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y and one without, otherwise
many of the the tests that require unaligned access were skipped.
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
0
No CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
59
The two failing test_verifier tests are:
"ld_abs: vlan + abs, test 1"
"ld_abs: jump around ld_abs"
This is due to that "far branching" involved in those tests.
All tests where done on QEMU (QEMU emulator version 3.1.50
(v3.1.0-688-g8ae951fbc106)).
Signed-off-by: Björn Töpel <bjorn.topel@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2019-02-05 05:41:22 -07:00
|
|
|
emit(rv_sb(RV_REG_T2, 0, RV_REG_T1), ctx);
|
|
|
|
break;
|
|
|
|
|
|
|
|
case BPF_ST | BPF_MEM | BPF_H:
|
|
|
|
emit_imm(RV_REG_T1, imm, ctx);
|
|
|
|
if (is_12b_int(off)) {
|
|
|
|
emit(rv_sh(rd, off, RV_REG_T1), ctx);
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
|
|
|
|
emit_imm(RV_REG_T2, off, ctx);
|
bpf, riscv: Use compressed instructions in the rv64 JIT
This patch uses the RVC support and encodings from bpf_jit.h to optimize
the rv64 jit.
The optimizations work by replacing emit(rv_X(...)) with a call to a
helper function emit_X, which will emit a compressed version of the
instruction when possible, and when RVC is enabled.
The JIT continues to pass all tests in lib/test_bpf.c, and introduces
no new failures to test_verifier; both with and without RVC being enabled.
Most changes are straightforward replacements of emit(rv_X(...), ctx)
with emit_X(..., ctx), with the following exceptions bearing mention;
* Change emit_imm to sign-extend the value in "lower", since the
checks for RVC (and the instructions themselves) treat the value as
signed. Otherwise, small negative immediates will not be recognized as
encodable using an RVC instruction. For example, without this change,
emit_imm(rd, -1, ctx) would cause lower to become 4095, which is not a
6b int even though a "c.li rd, -1" instruction suffices.
* For {BPF_MOV,BPF_ADD} BPF_X, drop using addiw,addw in the 32-bit
cases since the values are zero-extended into the upper 32 bits in
the following instructions anyways, and the addition commutes with
zero-extension. (BPF_SUB BPF_X must still use subw since subtraction
does not commute with zero-extension.)
This patch avoids optimizing branches and jumps to use RVC instructions
since surrounding code often makes assumptions about the sizes of
emitted instructions. Optimizing these will require changing these
functions (e.g., emit_branch) to dynamically compute jump offsets.
The following are examples of the JITed code for the verifier selftest
"direct packet read test#3 for CGROUP_SKB OK", without and with RVC
enabled, respectively. The former uses 178 bytes, and the latter uses 112,
for a ~37% reduction in code size for this example.
Without RVC:
0: 02000813 addi a6,zero,32
4: fd010113 addi sp,sp,-48
8: 02813423 sd s0,40(sp)
c: 02913023 sd s1,32(sp)
10: 01213c23 sd s2,24(sp)
14: 01313823 sd s3,16(sp)
18: 01413423 sd s4,8(sp)
1c: 03010413 addi s0,sp,48
20: 03056683 lwu a3,48(a0)
24: 02069693 slli a3,a3,0x20
28: 0206d693 srli a3,a3,0x20
2c: 03456703 lwu a4,52(a0)
30: 02071713 slli a4,a4,0x20
34: 02075713 srli a4,a4,0x20
38: 03856483 lwu s1,56(a0)
3c: 02049493 slli s1,s1,0x20
40: 0204d493 srli s1,s1,0x20
44: 03c56903 lwu s2,60(a0)
48: 02091913 slli s2,s2,0x20
4c: 02095913 srli s2,s2,0x20
50: 04056983 lwu s3,64(a0)
54: 02099993 slli s3,s3,0x20
58: 0209d993 srli s3,s3,0x20
5c: 09056a03 lwu s4,144(a0)
60: 020a1a13 slli s4,s4,0x20
64: 020a5a13 srli s4,s4,0x20
68: 00900313 addi t1,zero,9
6c: 006a7463 bgeu s4,t1,0x74
70: 00000a13 addi s4,zero,0
74: 02d52823 sw a3,48(a0)
78: 02e52a23 sw a4,52(a0)
7c: 02952c23 sw s1,56(a0)
80: 03252e23 sw s2,60(a0)
84: 05352023 sw s3,64(a0)
88: 00000793 addi a5,zero,0
8c: 02813403 ld s0,40(sp)
90: 02013483 ld s1,32(sp)
94: 01813903 ld s2,24(sp)
98: 01013983 ld s3,16(sp)
9c: 00813a03 ld s4,8(sp)
a0: 03010113 addi sp,sp,48
a4: 00078513 addi a0,a5,0
a8: 00008067 jalr zero,0(ra)
With RVC:
0: 02000813 addi a6,zero,32
4: 7179 c.addi16sp sp,-48
6: f422 c.sdsp s0,40(sp)
8: f026 c.sdsp s1,32(sp)
a: ec4a c.sdsp s2,24(sp)
c: e84e c.sdsp s3,16(sp)
e: e452 c.sdsp s4,8(sp)
10: 1800 c.addi4spn s0,sp,48
12: 03056683 lwu a3,48(a0)
16: 1682 c.slli a3,0x20
18: 9281 c.srli a3,0x20
1a: 03456703 lwu a4,52(a0)
1e: 1702 c.slli a4,0x20
20: 9301 c.srli a4,0x20
22: 03856483 lwu s1,56(a0)
26: 1482 c.slli s1,0x20
28: 9081 c.srli s1,0x20
2a: 03c56903 lwu s2,60(a0)
2e: 1902 c.slli s2,0x20
30: 02095913 srli s2,s2,0x20
34: 04056983 lwu s3,64(a0)
38: 1982 c.slli s3,0x20
3a: 0209d993 srli s3,s3,0x20
3e: 09056a03 lwu s4,144(a0)
42: 1a02 c.slli s4,0x20
44: 020a5a13 srli s4,s4,0x20
48: 4325 c.li t1,9
4a: 006a7363 bgeu s4,t1,0x50
4e: 4a01 c.li s4,0
50: d914 c.sw a3,48(a0)
52: d958 c.sw a4,52(a0)
54: dd04 c.sw s1,56(a0)
56: 03252e23 sw s2,60(a0)
5a: 05352023 sw s3,64(a0)
5e: 4781 c.li a5,0
60: 7422 c.ldsp s0,40(sp)
62: 7482 c.ldsp s1,32(sp)
64: 6962 c.ldsp s2,24(sp)
66: 69c2 c.ldsp s3,16(sp)
68: 6a22 c.ldsp s4,8(sp)
6a: 6145 c.addi16sp sp,48
6c: 853e c.mv a0,a5
6e: 8082 c.jr ra
Signed-off-by: Luke Nelson <luke.r.nels@gmail.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Cc: Björn Töpel <bjorn.topel@gmail.com>
Link: https://lore.kernel.org/bpf/20200721025241.8077-4-luke.r.nels@gmail.com
2020-07-20 19:52:40 -07:00
|
|
|
emit_add(RV_REG_T2, RV_REG_T2, rd, ctx);
|
bpf, riscv: add BPF JIT for RV64G
This commit adds a BPF JIT for RV64G.
The JIT is a two-pass JIT, and has a dynamic prolog/epilogue (similar
to the MIPS64 BPF JIT) instead of static ones (e.g. x86_64).
At the moment the RISC-V Linux port does not support
CONFIG_HAVE_KPROBES, which means that CONFIG_BPF_EVENTS is not
supported. Thus, no tests involving BPF_PROG_TYPE_TRACEPOINT,
BPF_PROG_TYPE_PERF_EVENT, BPF_PROG_TYPE_KPROBE and
BPF_PROG_TYPE_RAW_TRACEPOINT passes.
The implementation does not support "far branching" (>4KiB).
Test results:
# modprobe test_bpf
test_bpf: Summary: 378 PASSED, 0 FAILED, [366/366 JIT'ed]
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier
...
Summary: 761 PASSED, 507 SKIPPED, 2 FAILED
Note that "test_verifier" was run with one build with
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y and one without, otherwise
many of the the tests that require unaligned access were skipped.
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
0
No CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
59
The two failing test_verifier tests are:
"ld_abs: vlan + abs, test 1"
"ld_abs: jump around ld_abs"
This is due to that "far branching" involved in those tests.
All tests where done on QEMU (QEMU emulator version 3.1.50
(v3.1.0-688-g8ae951fbc106)).
Signed-off-by: Björn Töpel <bjorn.topel@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2019-02-05 05:41:22 -07:00
|
|
|
emit(rv_sh(RV_REG_T2, 0, RV_REG_T1), ctx);
|
|
|
|
break;
|
|
|
|
case BPF_ST | BPF_MEM | BPF_W:
|
|
|
|
emit_imm(RV_REG_T1, imm, ctx);
|
|
|
|
if (is_12b_int(off)) {
|
bpf, riscv: Use compressed instructions in the rv64 JIT
This patch uses the RVC support and encodings from bpf_jit.h to optimize
the rv64 jit.
The optimizations work by replacing emit(rv_X(...)) with a call to a
helper function emit_X, which will emit a compressed version of the
instruction when possible, and when RVC is enabled.
The JIT continues to pass all tests in lib/test_bpf.c, and introduces
no new failures to test_verifier; both with and without RVC being enabled.
Most changes are straightforward replacements of emit(rv_X(...), ctx)
with emit_X(..., ctx), with the following exceptions bearing mention;
* Change emit_imm to sign-extend the value in "lower", since the
checks for RVC (and the instructions themselves) treat the value as
signed. Otherwise, small negative immediates will not be recognized as
encodable using an RVC instruction. For example, without this change,
emit_imm(rd, -1, ctx) would cause lower to become 4095, which is not a
6b int even though a "c.li rd, -1" instruction suffices.
* For {BPF_MOV,BPF_ADD} BPF_X, drop using addiw,addw in the 32-bit
cases since the values are zero-extended into the upper 32 bits in
the following instructions anyways, and the addition commutes with
zero-extension. (BPF_SUB BPF_X must still use subw since subtraction
does not commute with zero-extension.)
This patch avoids optimizing branches and jumps to use RVC instructions
since surrounding code often makes assumptions about the sizes of
emitted instructions. Optimizing these will require changing these
functions (e.g., emit_branch) to dynamically compute jump offsets.
The following are examples of the JITed code for the verifier selftest
"direct packet read test#3 for CGROUP_SKB OK", without and with RVC
enabled, respectively. The former uses 178 bytes, and the latter uses 112,
for a ~37% reduction in code size for this example.
Without RVC:
0: 02000813 addi a6,zero,32
4: fd010113 addi sp,sp,-48
8: 02813423 sd s0,40(sp)
c: 02913023 sd s1,32(sp)
10: 01213c23 sd s2,24(sp)
14: 01313823 sd s3,16(sp)
18: 01413423 sd s4,8(sp)
1c: 03010413 addi s0,sp,48
20: 03056683 lwu a3,48(a0)
24: 02069693 slli a3,a3,0x20
28: 0206d693 srli a3,a3,0x20
2c: 03456703 lwu a4,52(a0)
30: 02071713 slli a4,a4,0x20
34: 02075713 srli a4,a4,0x20
38: 03856483 lwu s1,56(a0)
3c: 02049493 slli s1,s1,0x20
40: 0204d493 srli s1,s1,0x20
44: 03c56903 lwu s2,60(a0)
48: 02091913 slli s2,s2,0x20
4c: 02095913 srli s2,s2,0x20
50: 04056983 lwu s3,64(a0)
54: 02099993 slli s3,s3,0x20
58: 0209d993 srli s3,s3,0x20
5c: 09056a03 lwu s4,144(a0)
60: 020a1a13 slli s4,s4,0x20
64: 020a5a13 srli s4,s4,0x20
68: 00900313 addi t1,zero,9
6c: 006a7463 bgeu s4,t1,0x74
70: 00000a13 addi s4,zero,0
74: 02d52823 sw a3,48(a0)
78: 02e52a23 sw a4,52(a0)
7c: 02952c23 sw s1,56(a0)
80: 03252e23 sw s2,60(a0)
84: 05352023 sw s3,64(a0)
88: 00000793 addi a5,zero,0
8c: 02813403 ld s0,40(sp)
90: 02013483 ld s1,32(sp)
94: 01813903 ld s2,24(sp)
98: 01013983 ld s3,16(sp)
9c: 00813a03 ld s4,8(sp)
a0: 03010113 addi sp,sp,48
a4: 00078513 addi a0,a5,0
a8: 00008067 jalr zero,0(ra)
With RVC:
0: 02000813 addi a6,zero,32
4: 7179 c.addi16sp sp,-48
6: f422 c.sdsp s0,40(sp)
8: f026 c.sdsp s1,32(sp)
a: ec4a c.sdsp s2,24(sp)
c: e84e c.sdsp s3,16(sp)
e: e452 c.sdsp s4,8(sp)
10: 1800 c.addi4spn s0,sp,48
12: 03056683 lwu a3,48(a0)
16: 1682 c.slli a3,0x20
18: 9281 c.srli a3,0x20
1a: 03456703 lwu a4,52(a0)
1e: 1702 c.slli a4,0x20
20: 9301 c.srli a4,0x20
22: 03856483 lwu s1,56(a0)
26: 1482 c.slli s1,0x20
28: 9081 c.srli s1,0x20
2a: 03c56903 lwu s2,60(a0)
2e: 1902 c.slli s2,0x20
30: 02095913 srli s2,s2,0x20
34: 04056983 lwu s3,64(a0)
38: 1982 c.slli s3,0x20
3a: 0209d993 srli s3,s3,0x20
3e: 09056a03 lwu s4,144(a0)
42: 1a02 c.slli s4,0x20
44: 020a5a13 srli s4,s4,0x20
48: 4325 c.li t1,9
4a: 006a7363 bgeu s4,t1,0x50
4e: 4a01 c.li s4,0
50: d914 c.sw a3,48(a0)
52: d958 c.sw a4,52(a0)
54: dd04 c.sw s1,56(a0)
56: 03252e23 sw s2,60(a0)
5a: 05352023 sw s3,64(a0)
5e: 4781 c.li a5,0
60: 7422 c.ldsp s0,40(sp)
62: 7482 c.ldsp s1,32(sp)
64: 6962 c.ldsp s2,24(sp)
66: 69c2 c.ldsp s3,16(sp)
68: 6a22 c.ldsp s4,8(sp)
6a: 6145 c.addi16sp sp,48
6c: 853e c.mv a0,a5
6e: 8082 c.jr ra
Signed-off-by: Luke Nelson <luke.r.nels@gmail.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Cc: Björn Töpel <bjorn.topel@gmail.com>
Link: https://lore.kernel.org/bpf/20200721025241.8077-4-luke.r.nels@gmail.com
2020-07-20 19:52:40 -07:00
|
|
|
emit_sw(rd, off, RV_REG_T1, ctx);
|
bpf, riscv: add BPF JIT for RV64G
This commit adds a BPF JIT for RV64G.
The JIT is a two-pass JIT, and has a dynamic prolog/epilogue (similar
to the MIPS64 BPF JIT) instead of static ones (e.g. x86_64).
At the moment the RISC-V Linux port does not support
CONFIG_HAVE_KPROBES, which means that CONFIG_BPF_EVENTS is not
supported. Thus, no tests involving BPF_PROG_TYPE_TRACEPOINT,
BPF_PROG_TYPE_PERF_EVENT, BPF_PROG_TYPE_KPROBE and
BPF_PROG_TYPE_RAW_TRACEPOINT passes.
The implementation does not support "far branching" (>4KiB).
Test results:
# modprobe test_bpf
test_bpf: Summary: 378 PASSED, 0 FAILED, [366/366 JIT'ed]
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier
...
Summary: 761 PASSED, 507 SKIPPED, 2 FAILED
Note that "test_verifier" was run with one build with
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y and one without, otherwise
many of the the tests that require unaligned access were skipped.
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
0
No CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
59
The two failing test_verifier tests are:
"ld_abs: vlan + abs, test 1"
"ld_abs: jump around ld_abs"
This is due to that "far branching" involved in those tests.
All tests where done on QEMU (QEMU emulator version 3.1.50
(v3.1.0-688-g8ae951fbc106)).
Signed-off-by: Björn Töpel <bjorn.topel@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2019-02-05 05:41:22 -07:00
|
|
|
break;
|
|
|
|
}
|
|
|
|
|
|
|
|
emit_imm(RV_REG_T2, off, ctx);
|
bpf, riscv: Use compressed instructions in the rv64 JIT
This patch uses the RVC support and encodings from bpf_jit.h to optimize
the rv64 jit.
The optimizations work by replacing emit(rv_X(...)) with a call to a
helper function emit_X, which will emit a compressed version of the
instruction when possible, and when RVC is enabled.
The JIT continues to pass all tests in lib/test_bpf.c, and introduces
no new failures to test_verifier; both with and without RVC being enabled.
Most changes are straightforward replacements of emit(rv_X(...), ctx)
with emit_X(..., ctx), with the following exceptions bearing mention;
* Change emit_imm to sign-extend the value in "lower", since the
checks for RVC (and the instructions themselves) treat the value as
signed. Otherwise, small negative immediates will not be recognized as
encodable using an RVC instruction. For example, without this change,
emit_imm(rd, -1, ctx) would cause lower to become 4095, which is not a
6b int even though a "c.li rd, -1" instruction suffices.
* For {BPF_MOV,BPF_ADD} BPF_X, drop using addiw,addw in the 32-bit
cases since the values are zero-extended into the upper 32 bits in
the following instructions anyways, and the addition commutes with
zero-extension. (BPF_SUB BPF_X must still use subw since subtraction
does not commute with zero-extension.)
This patch avoids optimizing branches and jumps to use RVC instructions
since surrounding code often makes assumptions about the sizes of
emitted instructions. Optimizing these will require changing these
functions (e.g., emit_branch) to dynamically compute jump offsets.
The following are examples of the JITed code for the verifier selftest
"direct packet read test#3 for CGROUP_SKB OK", without and with RVC
enabled, respectively. The former uses 178 bytes, and the latter uses 112,
for a ~37% reduction in code size for this example.
Without RVC:
0: 02000813 addi a6,zero,32
4: fd010113 addi sp,sp,-48
8: 02813423 sd s0,40(sp)
c: 02913023 sd s1,32(sp)
10: 01213c23 sd s2,24(sp)
14: 01313823 sd s3,16(sp)
18: 01413423 sd s4,8(sp)
1c: 03010413 addi s0,sp,48
20: 03056683 lwu a3,48(a0)
24: 02069693 slli a3,a3,0x20
28: 0206d693 srli a3,a3,0x20
2c: 03456703 lwu a4,52(a0)
30: 02071713 slli a4,a4,0x20
34: 02075713 srli a4,a4,0x20
38: 03856483 lwu s1,56(a0)
3c: 02049493 slli s1,s1,0x20
40: 0204d493 srli s1,s1,0x20
44: 03c56903 lwu s2,60(a0)
48: 02091913 slli s2,s2,0x20
4c: 02095913 srli s2,s2,0x20
50: 04056983 lwu s3,64(a0)
54: 02099993 slli s3,s3,0x20
58: 0209d993 srli s3,s3,0x20
5c: 09056a03 lwu s4,144(a0)
60: 020a1a13 slli s4,s4,0x20
64: 020a5a13 srli s4,s4,0x20
68: 00900313 addi t1,zero,9
6c: 006a7463 bgeu s4,t1,0x74
70: 00000a13 addi s4,zero,0
74: 02d52823 sw a3,48(a0)
78: 02e52a23 sw a4,52(a0)
7c: 02952c23 sw s1,56(a0)
80: 03252e23 sw s2,60(a0)
84: 05352023 sw s3,64(a0)
88: 00000793 addi a5,zero,0
8c: 02813403 ld s0,40(sp)
90: 02013483 ld s1,32(sp)
94: 01813903 ld s2,24(sp)
98: 01013983 ld s3,16(sp)
9c: 00813a03 ld s4,8(sp)
a0: 03010113 addi sp,sp,48
a4: 00078513 addi a0,a5,0
a8: 00008067 jalr zero,0(ra)
With RVC:
0: 02000813 addi a6,zero,32
4: 7179 c.addi16sp sp,-48
6: f422 c.sdsp s0,40(sp)
8: f026 c.sdsp s1,32(sp)
a: ec4a c.sdsp s2,24(sp)
c: e84e c.sdsp s3,16(sp)
e: e452 c.sdsp s4,8(sp)
10: 1800 c.addi4spn s0,sp,48
12: 03056683 lwu a3,48(a0)
16: 1682 c.slli a3,0x20
18: 9281 c.srli a3,0x20
1a: 03456703 lwu a4,52(a0)
1e: 1702 c.slli a4,0x20
20: 9301 c.srli a4,0x20
22: 03856483 lwu s1,56(a0)
26: 1482 c.slli s1,0x20
28: 9081 c.srli s1,0x20
2a: 03c56903 lwu s2,60(a0)
2e: 1902 c.slli s2,0x20
30: 02095913 srli s2,s2,0x20
34: 04056983 lwu s3,64(a0)
38: 1982 c.slli s3,0x20
3a: 0209d993 srli s3,s3,0x20
3e: 09056a03 lwu s4,144(a0)
42: 1a02 c.slli s4,0x20
44: 020a5a13 srli s4,s4,0x20
48: 4325 c.li t1,9
4a: 006a7363 bgeu s4,t1,0x50
4e: 4a01 c.li s4,0
50: d914 c.sw a3,48(a0)
52: d958 c.sw a4,52(a0)
54: dd04 c.sw s1,56(a0)
56: 03252e23 sw s2,60(a0)
5a: 05352023 sw s3,64(a0)
5e: 4781 c.li a5,0
60: 7422 c.ldsp s0,40(sp)
62: 7482 c.ldsp s1,32(sp)
64: 6962 c.ldsp s2,24(sp)
66: 69c2 c.ldsp s3,16(sp)
68: 6a22 c.ldsp s4,8(sp)
6a: 6145 c.addi16sp sp,48
6c: 853e c.mv a0,a5
6e: 8082 c.jr ra
Signed-off-by: Luke Nelson <luke.r.nels@gmail.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Cc: Björn Töpel <bjorn.topel@gmail.com>
Link: https://lore.kernel.org/bpf/20200721025241.8077-4-luke.r.nels@gmail.com
2020-07-20 19:52:40 -07:00
|
|
|
emit_add(RV_REG_T2, RV_REG_T2, rd, ctx);
|
|
|
|
emit_sw(RV_REG_T2, 0, RV_REG_T1, ctx);
|
bpf, riscv: add BPF JIT for RV64G
This commit adds a BPF JIT for RV64G.
The JIT is a two-pass JIT, and has a dynamic prolog/epilogue (similar
to the MIPS64 BPF JIT) instead of static ones (e.g. x86_64).
At the moment the RISC-V Linux port does not support
CONFIG_HAVE_KPROBES, which means that CONFIG_BPF_EVENTS is not
supported. Thus, no tests involving BPF_PROG_TYPE_TRACEPOINT,
BPF_PROG_TYPE_PERF_EVENT, BPF_PROG_TYPE_KPROBE and
BPF_PROG_TYPE_RAW_TRACEPOINT passes.
The implementation does not support "far branching" (>4KiB).
Test results:
# modprobe test_bpf
test_bpf: Summary: 378 PASSED, 0 FAILED, [366/366 JIT'ed]
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier
...
Summary: 761 PASSED, 507 SKIPPED, 2 FAILED
Note that "test_verifier" was run with one build with
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y and one without, otherwise
many of the the tests that require unaligned access were skipped.
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
0
No CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
59
The two failing test_verifier tests are:
"ld_abs: vlan + abs, test 1"
"ld_abs: jump around ld_abs"
This is due to that "far branching" involved in those tests.
All tests where done on QEMU (QEMU emulator version 3.1.50
(v3.1.0-688-g8ae951fbc106)).
Signed-off-by: Björn Töpel <bjorn.topel@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2019-02-05 05:41:22 -07:00
|
|
|
break;
|
|
|
|
case BPF_ST | BPF_MEM | BPF_DW:
|
|
|
|
emit_imm(RV_REG_T1, imm, ctx);
|
|
|
|
if (is_12b_int(off)) {
|
bpf, riscv: Use compressed instructions in the rv64 JIT
This patch uses the RVC support and encodings from bpf_jit.h to optimize
the rv64 jit.
The optimizations work by replacing emit(rv_X(...)) with a call to a
helper function emit_X, which will emit a compressed version of the
instruction when possible, and when RVC is enabled.
The JIT continues to pass all tests in lib/test_bpf.c, and introduces
no new failures to test_verifier; both with and without RVC being enabled.
Most changes are straightforward replacements of emit(rv_X(...), ctx)
with emit_X(..., ctx), with the following exceptions bearing mention;
* Change emit_imm to sign-extend the value in "lower", since the
checks for RVC (and the instructions themselves) treat the value as
signed. Otherwise, small negative immediates will not be recognized as
encodable using an RVC instruction. For example, without this change,
emit_imm(rd, -1, ctx) would cause lower to become 4095, which is not a
6b int even though a "c.li rd, -1" instruction suffices.
* For {BPF_MOV,BPF_ADD} BPF_X, drop using addiw,addw in the 32-bit
cases since the values are zero-extended into the upper 32 bits in
the following instructions anyways, and the addition commutes with
zero-extension. (BPF_SUB BPF_X must still use subw since subtraction
does not commute with zero-extension.)
This patch avoids optimizing branches and jumps to use RVC instructions
since surrounding code often makes assumptions about the sizes of
emitted instructions. Optimizing these will require changing these
functions (e.g., emit_branch) to dynamically compute jump offsets.
The following are examples of the JITed code for the verifier selftest
"direct packet read test#3 for CGROUP_SKB OK", without and with RVC
enabled, respectively. The former uses 178 bytes, and the latter uses 112,
for a ~37% reduction in code size for this example.
Without RVC:
0: 02000813 addi a6,zero,32
4: fd010113 addi sp,sp,-48
8: 02813423 sd s0,40(sp)
c: 02913023 sd s1,32(sp)
10: 01213c23 sd s2,24(sp)
14: 01313823 sd s3,16(sp)
18: 01413423 sd s4,8(sp)
1c: 03010413 addi s0,sp,48
20: 03056683 lwu a3,48(a0)
24: 02069693 slli a3,a3,0x20
28: 0206d693 srli a3,a3,0x20
2c: 03456703 lwu a4,52(a0)
30: 02071713 slli a4,a4,0x20
34: 02075713 srli a4,a4,0x20
38: 03856483 lwu s1,56(a0)
3c: 02049493 slli s1,s1,0x20
40: 0204d493 srli s1,s1,0x20
44: 03c56903 lwu s2,60(a0)
48: 02091913 slli s2,s2,0x20
4c: 02095913 srli s2,s2,0x20
50: 04056983 lwu s3,64(a0)
54: 02099993 slli s3,s3,0x20
58: 0209d993 srli s3,s3,0x20
5c: 09056a03 lwu s4,144(a0)
60: 020a1a13 slli s4,s4,0x20
64: 020a5a13 srli s4,s4,0x20
68: 00900313 addi t1,zero,9
6c: 006a7463 bgeu s4,t1,0x74
70: 00000a13 addi s4,zero,0
74: 02d52823 sw a3,48(a0)
78: 02e52a23 sw a4,52(a0)
7c: 02952c23 sw s1,56(a0)
80: 03252e23 sw s2,60(a0)
84: 05352023 sw s3,64(a0)
88: 00000793 addi a5,zero,0
8c: 02813403 ld s0,40(sp)
90: 02013483 ld s1,32(sp)
94: 01813903 ld s2,24(sp)
98: 01013983 ld s3,16(sp)
9c: 00813a03 ld s4,8(sp)
a0: 03010113 addi sp,sp,48
a4: 00078513 addi a0,a5,0
a8: 00008067 jalr zero,0(ra)
With RVC:
0: 02000813 addi a6,zero,32
4: 7179 c.addi16sp sp,-48
6: f422 c.sdsp s0,40(sp)
8: f026 c.sdsp s1,32(sp)
a: ec4a c.sdsp s2,24(sp)
c: e84e c.sdsp s3,16(sp)
e: e452 c.sdsp s4,8(sp)
10: 1800 c.addi4spn s0,sp,48
12: 03056683 lwu a3,48(a0)
16: 1682 c.slli a3,0x20
18: 9281 c.srli a3,0x20
1a: 03456703 lwu a4,52(a0)
1e: 1702 c.slli a4,0x20
20: 9301 c.srli a4,0x20
22: 03856483 lwu s1,56(a0)
26: 1482 c.slli s1,0x20
28: 9081 c.srli s1,0x20
2a: 03c56903 lwu s2,60(a0)
2e: 1902 c.slli s2,0x20
30: 02095913 srli s2,s2,0x20
34: 04056983 lwu s3,64(a0)
38: 1982 c.slli s3,0x20
3a: 0209d993 srli s3,s3,0x20
3e: 09056a03 lwu s4,144(a0)
42: 1a02 c.slli s4,0x20
44: 020a5a13 srli s4,s4,0x20
48: 4325 c.li t1,9
4a: 006a7363 bgeu s4,t1,0x50
4e: 4a01 c.li s4,0
50: d914 c.sw a3,48(a0)
52: d958 c.sw a4,52(a0)
54: dd04 c.sw s1,56(a0)
56: 03252e23 sw s2,60(a0)
5a: 05352023 sw s3,64(a0)
5e: 4781 c.li a5,0
60: 7422 c.ldsp s0,40(sp)
62: 7482 c.ldsp s1,32(sp)
64: 6962 c.ldsp s2,24(sp)
66: 69c2 c.ldsp s3,16(sp)
68: 6a22 c.ldsp s4,8(sp)
6a: 6145 c.addi16sp sp,48
6c: 853e c.mv a0,a5
6e: 8082 c.jr ra
Signed-off-by: Luke Nelson <luke.r.nels@gmail.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Cc: Björn Töpel <bjorn.topel@gmail.com>
Link: https://lore.kernel.org/bpf/20200721025241.8077-4-luke.r.nels@gmail.com
2020-07-20 19:52:40 -07:00
|
|
|
emit_sd(rd, off, RV_REG_T1, ctx);
|
bpf, riscv: add BPF JIT for RV64G
This commit adds a BPF JIT for RV64G.
The JIT is a two-pass JIT, and has a dynamic prolog/epilogue (similar
to the MIPS64 BPF JIT) instead of static ones (e.g. x86_64).
At the moment the RISC-V Linux port does not support
CONFIG_HAVE_KPROBES, which means that CONFIG_BPF_EVENTS is not
supported. Thus, no tests involving BPF_PROG_TYPE_TRACEPOINT,
BPF_PROG_TYPE_PERF_EVENT, BPF_PROG_TYPE_KPROBE and
BPF_PROG_TYPE_RAW_TRACEPOINT passes.
The implementation does not support "far branching" (>4KiB).
Test results:
# modprobe test_bpf
test_bpf: Summary: 378 PASSED, 0 FAILED, [366/366 JIT'ed]
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier
...
Summary: 761 PASSED, 507 SKIPPED, 2 FAILED
Note that "test_verifier" was run with one build with
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y and one without, otherwise
many of the the tests that require unaligned access were skipped.
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
0
No CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
59
The two failing test_verifier tests are:
"ld_abs: vlan + abs, test 1"
"ld_abs: jump around ld_abs"
This is due to that "far branching" involved in those tests.
All tests where done on QEMU (QEMU emulator version 3.1.50
(v3.1.0-688-g8ae951fbc106)).
Signed-off-by: Björn Töpel <bjorn.topel@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2019-02-05 05:41:22 -07:00
|
|
|
break;
|
|
|
|
}
|
|
|
|
|
|
|
|
emit_imm(RV_REG_T2, off, ctx);
|
bpf, riscv: Use compressed instructions in the rv64 JIT
This patch uses the RVC support and encodings from bpf_jit.h to optimize
the rv64 jit.
The optimizations work by replacing emit(rv_X(...)) with a call to a
helper function emit_X, which will emit a compressed version of the
instruction when possible, and when RVC is enabled.
The JIT continues to pass all tests in lib/test_bpf.c, and introduces
no new failures to test_verifier; both with and without RVC being enabled.
Most changes are straightforward replacements of emit(rv_X(...), ctx)
with emit_X(..., ctx), with the following exceptions bearing mention;
* Change emit_imm to sign-extend the value in "lower", since the
checks for RVC (and the instructions themselves) treat the value as
signed. Otherwise, small negative immediates will not be recognized as
encodable using an RVC instruction. For example, without this change,
emit_imm(rd, -1, ctx) would cause lower to become 4095, which is not a
6b int even though a "c.li rd, -1" instruction suffices.
* For {BPF_MOV,BPF_ADD} BPF_X, drop using addiw,addw in the 32-bit
cases since the values are zero-extended into the upper 32 bits in
the following instructions anyways, and the addition commutes with
zero-extension. (BPF_SUB BPF_X must still use subw since subtraction
does not commute with zero-extension.)
This patch avoids optimizing branches and jumps to use RVC instructions
since surrounding code often makes assumptions about the sizes of
emitted instructions. Optimizing these will require changing these
functions (e.g., emit_branch) to dynamically compute jump offsets.
The following are examples of the JITed code for the verifier selftest
"direct packet read test#3 for CGROUP_SKB OK", without and with RVC
enabled, respectively. The former uses 178 bytes, and the latter uses 112,
for a ~37% reduction in code size for this example.
Without RVC:
0: 02000813 addi a6,zero,32
4: fd010113 addi sp,sp,-48
8: 02813423 sd s0,40(sp)
c: 02913023 sd s1,32(sp)
10: 01213c23 sd s2,24(sp)
14: 01313823 sd s3,16(sp)
18: 01413423 sd s4,8(sp)
1c: 03010413 addi s0,sp,48
20: 03056683 lwu a3,48(a0)
24: 02069693 slli a3,a3,0x20
28: 0206d693 srli a3,a3,0x20
2c: 03456703 lwu a4,52(a0)
30: 02071713 slli a4,a4,0x20
34: 02075713 srli a4,a4,0x20
38: 03856483 lwu s1,56(a0)
3c: 02049493 slli s1,s1,0x20
40: 0204d493 srli s1,s1,0x20
44: 03c56903 lwu s2,60(a0)
48: 02091913 slli s2,s2,0x20
4c: 02095913 srli s2,s2,0x20
50: 04056983 lwu s3,64(a0)
54: 02099993 slli s3,s3,0x20
58: 0209d993 srli s3,s3,0x20
5c: 09056a03 lwu s4,144(a0)
60: 020a1a13 slli s4,s4,0x20
64: 020a5a13 srli s4,s4,0x20
68: 00900313 addi t1,zero,9
6c: 006a7463 bgeu s4,t1,0x74
70: 00000a13 addi s4,zero,0
74: 02d52823 sw a3,48(a0)
78: 02e52a23 sw a4,52(a0)
7c: 02952c23 sw s1,56(a0)
80: 03252e23 sw s2,60(a0)
84: 05352023 sw s3,64(a0)
88: 00000793 addi a5,zero,0
8c: 02813403 ld s0,40(sp)
90: 02013483 ld s1,32(sp)
94: 01813903 ld s2,24(sp)
98: 01013983 ld s3,16(sp)
9c: 00813a03 ld s4,8(sp)
a0: 03010113 addi sp,sp,48
a4: 00078513 addi a0,a5,0
a8: 00008067 jalr zero,0(ra)
With RVC:
0: 02000813 addi a6,zero,32
4: 7179 c.addi16sp sp,-48
6: f422 c.sdsp s0,40(sp)
8: f026 c.sdsp s1,32(sp)
a: ec4a c.sdsp s2,24(sp)
c: e84e c.sdsp s3,16(sp)
e: e452 c.sdsp s4,8(sp)
10: 1800 c.addi4spn s0,sp,48
12: 03056683 lwu a3,48(a0)
16: 1682 c.slli a3,0x20
18: 9281 c.srli a3,0x20
1a: 03456703 lwu a4,52(a0)
1e: 1702 c.slli a4,0x20
20: 9301 c.srli a4,0x20
22: 03856483 lwu s1,56(a0)
26: 1482 c.slli s1,0x20
28: 9081 c.srli s1,0x20
2a: 03c56903 lwu s2,60(a0)
2e: 1902 c.slli s2,0x20
30: 02095913 srli s2,s2,0x20
34: 04056983 lwu s3,64(a0)
38: 1982 c.slli s3,0x20
3a: 0209d993 srli s3,s3,0x20
3e: 09056a03 lwu s4,144(a0)
42: 1a02 c.slli s4,0x20
44: 020a5a13 srli s4,s4,0x20
48: 4325 c.li t1,9
4a: 006a7363 bgeu s4,t1,0x50
4e: 4a01 c.li s4,0
50: d914 c.sw a3,48(a0)
52: d958 c.sw a4,52(a0)
54: dd04 c.sw s1,56(a0)
56: 03252e23 sw s2,60(a0)
5a: 05352023 sw s3,64(a0)
5e: 4781 c.li a5,0
60: 7422 c.ldsp s0,40(sp)
62: 7482 c.ldsp s1,32(sp)
64: 6962 c.ldsp s2,24(sp)
66: 69c2 c.ldsp s3,16(sp)
68: 6a22 c.ldsp s4,8(sp)
6a: 6145 c.addi16sp sp,48
6c: 853e c.mv a0,a5
6e: 8082 c.jr ra
Signed-off-by: Luke Nelson <luke.r.nels@gmail.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Cc: Björn Töpel <bjorn.topel@gmail.com>
Link: https://lore.kernel.org/bpf/20200721025241.8077-4-luke.r.nels@gmail.com
2020-07-20 19:52:40 -07:00
|
|
|
emit_add(RV_REG_T2, RV_REG_T2, rd, ctx);
|
|
|
|
emit_sd(RV_REG_T2, 0, RV_REG_T1, ctx);
|
bpf, riscv: add BPF JIT for RV64G
This commit adds a BPF JIT for RV64G.
The JIT is a two-pass JIT, and has a dynamic prolog/epilogue (similar
to the MIPS64 BPF JIT) instead of static ones (e.g. x86_64).
At the moment the RISC-V Linux port does not support
CONFIG_HAVE_KPROBES, which means that CONFIG_BPF_EVENTS is not
supported. Thus, no tests involving BPF_PROG_TYPE_TRACEPOINT,
BPF_PROG_TYPE_PERF_EVENT, BPF_PROG_TYPE_KPROBE and
BPF_PROG_TYPE_RAW_TRACEPOINT passes.
The implementation does not support "far branching" (>4KiB).
Test results:
# modprobe test_bpf
test_bpf: Summary: 378 PASSED, 0 FAILED, [366/366 JIT'ed]
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier
...
Summary: 761 PASSED, 507 SKIPPED, 2 FAILED
Note that "test_verifier" was run with one build with
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y and one without, otherwise
many of the the tests that require unaligned access were skipped.
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
0
No CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
59
The two failing test_verifier tests are:
"ld_abs: vlan + abs, test 1"
"ld_abs: jump around ld_abs"
This is due to that "far branching" involved in those tests.
All tests where done on QEMU (QEMU emulator version 3.1.50
(v3.1.0-688-g8ae951fbc106)).
Signed-off-by: Björn Töpel <bjorn.topel@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2019-02-05 05:41:22 -07:00
|
|
|
break;
|
|
|
|
|
2024-04-04 04:42:02 -07:00
|
|
|
case BPF_ST | BPF_PROBE_MEM32 | BPF_B:
|
|
|
|
case BPF_ST | BPF_PROBE_MEM32 | BPF_H:
|
|
|
|
case BPF_ST | BPF_PROBE_MEM32 | BPF_W:
|
|
|
|
case BPF_ST | BPF_PROBE_MEM32 | BPF_DW:
|
|
|
|
{
|
|
|
|
int insn_len, insns_start;
|
|
|
|
|
|
|
|
emit_add(RV_REG_T3, rd, RV_REG_ARENA, ctx);
|
|
|
|
rd = RV_REG_T3;
|
|
|
|
|
|
|
|
/* Load imm to a register then store it */
|
|
|
|
emit_imm(RV_REG_T1, imm, ctx);
|
|
|
|
|
|
|
|
switch (BPF_SIZE(code)) {
|
|
|
|
case BPF_B:
|
|
|
|
if (is_12b_int(off)) {
|
|
|
|
insns_start = ctx->ninsns;
|
|
|
|
emit(rv_sb(rd, off, RV_REG_T1), ctx);
|
|
|
|
insn_len = ctx->ninsns - insns_start;
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
|
|
|
|
emit_imm(RV_REG_T2, off, ctx);
|
|
|
|
emit_add(RV_REG_T2, RV_REG_T2, rd, ctx);
|
|
|
|
insns_start = ctx->ninsns;
|
|
|
|
emit(rv_sb(RV_REG_T2, 0, RV_REG_T1), ctx);
|
|
|
|
insn_len = ctx->ninsns - insns_start;
|
|
|
|
break;
|
|
|
|
case BPF_H:
|
|
|
|
if (is_12b_int(off)) {
|
|
|
|
insns_start = ctx->ninsns;
|
|
|
|
emit(rv_sh(rd, off, RV_REG_T1), ctx);
|
|
|
|
insn_len = ctx->ninsns - insns_start;
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
|
|
|
|
emit_imm(RV_REG_T2, off, ctx);
|
|
|
|
emit_add(RV_REG_T2, RV_REG_T2, rd, ctx);
|
|
|
|
insns_start = ctx->ninsns;
|
|
|
|
emit(rv_sh(RV_REG_T2, 0, RV_REG_T1), ctx);
|
|
|
|
insn_len = ctx->ninsns - insns_start;
|
|
|
|
break;
|
|
|
|
case BPF_W:
|
|
|
|
if (is_12b_int(off)) {
|
|
|
|
insns_start = ctx->ninsns;
|
|
|
|
emit_sw(rd, off, RV_REG_T1, ctx);
|
|
|
|
insn_len = ctx->ninsns - insns_start;
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
|
|
|
|
emit_imm(RV_REG_T2, off, ctx);
|
|
|
|
emit_add(RV_REG_T2, RV_REG_T2, rd, ctx);
|
|
|
|
insns_start = ctx->ninsns;
|
|
|
|
emit_sw(RV_REG_T2, 0, RV_REG_T1, ctx);
|
|
|
|
insn_len = ctx->ninsns - insns_start;
|
|
|
|
break;
|
|
|
|
case BPF_DW:
|
|
|
|
if (is_12b_int(off)) {
|
|
|
|
insns_start = ctx->ninsns;
|
|
|
|
emit_sd(rd, off, RV_REG_T1, ctx);
|
|
|
|
insn_len = ctx->ninsns - insns_start;
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
|
|
|
|
emit_imm(RV_REG_T2, off, ctx);
|
|
|
|
emit_add(RV_REG_T2, RV_REG_T2, rd, ctx);
|
|
|
|
insns_start = ctx->ninsns;
|
|
|
|
emit_sd(RV_REG_T2, 0, RV_REG_T1, ctx);
|
|
|
|
insn_len = ctx->ninsns - insns_start;
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
|
|
|
|
ret = add_exception_handler(insn, ctx, REG_DONT_CLEAR_MARKER,
|
|
|
|
insn_len);
|
|
|
|
if (ret)
|
|
|
|
return ret;
|
|
|
|
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
|
bpf, riscv: add BPF JIT for RV64G
This commit adds a BPF JIT for RV64G.
The JIT is a two-pass JIT, and has a dynamic prolog/epilogue (similar
to the MIPS64 BPF JIT) instead of static ones (e.g. x86_64).
At the moment the RISC-V Linux port does not support
CONFIG_HAVE_KPROBES, which means that CONFIG_BPF_EVENTS is not
supported. Thus, no tests involving BPF_PROG_TYPE_TRACEPOINT,
BPF_PROG_TYPE_PERF_EVENT, BPF_PROG_TYPE_KPROBE and
BPF_PROG_TYPE_RAW_TRACEPOINT passes.
The implementation does not support "far branching" (>4KiB).
Test results:
# modprobe test_bpf
test_bpf: Summary: 378 PASSED, 0 FAILED, [366/366 JIT'ed]
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier
...
Summary: 761 PASSED, 507 SKIPPED, 2 FAILED
Note that "test_verifier" was run with one build with
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y and one without, otherwise
many of the the tests that require unaligned access were skipped.
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
0
No CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
59
The two failing test_verifier tests are:
"ld_abs: vlan + abs, test 1"
"ld_abs: jump around ld_abs"
This is due to that "far branching" involved in those tests.
All tests where done on QEMU (QEMU emulator version 3.1.50
(v3.1.0-688-g8ae951fbc106)).
Signed-off-by: Björn Töpel <bjorn.topel@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2019-02-05 05:41:22 -07:00
|
|
|
/* STX: *(size *)(dst + off) = src */
|
|
|
|
case BPF_STX | BPF_MEM | BPF_B:
|
|
|
|
if (is_12b_int(off)) {
|
|
|
|
emit(rv_sb(rd, off, rs), ctx);
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
|
|
|
|
emit_imm(RV_REG_T1, off, ctx);
|
bpf, riscv: Use compressed instructions in the rv64 JIT
This patch uses the RVC support and encodings from bpf_jit.h to optimize
the rv64 jit.
The optimizations work by replacing emit(rv_X(...)) with a call to a
helper function emit_X, which will emit a compressed version of the
instruction when possible, and when RVC is enabled.
The JIT continues to pass all tests in lib/test_bpf.c, and introduces
no new failures to test_verifier; both with and without RVC being enabled.
Most changes are straightforward replacements of emit(rv_X(...), ctx)
with emit_X(..., ctx), with the following exceptions bearing mention;
* Change emit_imm to sign-extend the value in "lower", since the
checks for RVC (and the instructions themselves) treat the value as
signed. Otherwise, small negative immediates will not be recognized as
encodable using an RVC instruction. For example, without this change,
emit_imm(rd, -1, ctx) would cause lower to become 4095, which is not a
6b int even though a "c.li rd, -1" instruction suffices.
* For {BPF_MOV,BPF_ADD} BPF_X, drop using addiw,addw in the 32-bit
cases since the values are zero-extended into the upper 32 bits in
the following instructions anyways, and the addition commutes with
zero-extension. (BPF_SUB BPF_X must still use subw since subtraction
does not commute with zero-extension.)
This patch avoids optimizing branches and jumps to use RVC instructions
since surrounding code often makes assumptions about the sizes of
emitted instructions. Optimizing these will require changing these
functions (e.g., emit_branch) to dynamically compute jump offsets.
The following are examples of the JITed code for the verifier selftest
"direct packet read test#3 for CGROUP_SKB OK", without and with RVC
enabled, respectively. The former uses 178 bytes, and the latter uses 112,
for a ~37% reduction in code size for this example.
Without RVC:
0: 02000813 addi a6,zero,32
4: fd010113 addi sp,sp,-48
8: 02813423 sd s0,40(sp)
c: 02913023 sd s1,32(sp)
10: 01213c23 sd s2,24(sp)
14: 01313823 sd s3,16(sp)
18: 01413423 sd s4,8(sp)
1c: 03010413 addi s0,sp,48
20: 03056683 lwu a3,48(a0)
24: 02069693 slli a3,a3,0x20
28: 0206d693 srli a3,a3,0x20
2c: 03456703 lwu a4,52(a0)
30: 02071713 slli a4,a4,0x20
34: 02075713 srli a4,a4,0x20
38: 03856483 lwu s1,56(a0)
3c: 02049493 slli s1,s1,0x20
40: 0204d493 srli s1,s1,0x20
44: 03c56903 lwu s2,60(a0)
48: 02091913 slli s2,s2,0x20
4c: 02095913 srli s2,s2,0x20
50: 04056983 lwu s3,64(a0)
54: 02099993 slli s3,s3,0x20
58: 0209d993 srli s3,s3,0x20
5c: 09056a03 lwu s4,144(a0)
60: 020a1a13 slli s4,s4,0x20
64: 020a5a13 srli s4,s4,0x20
68: 00900313 addi t1,zero,9
6c: 006a7463 bgeu s4,t1,0x74
70: 00000a13 addi s4,zero,0
74: 02d52823 sw a3,48(a0)
78: 02e52a23 sw a4,52(a0)
7c: 02952c23 sw s1,56(a0)
80: 03252e23 sw s2,60(a0)
84: 05352023 sw s3,64(a0)
88: 00000793 addi a5,zero,0
8c: 02813403 ld s0,40(sp)
90: 02013483 ld s1,32(sp)
94: 01813903 ld s2,24(sp)
98: 01013983 ld s3,16(sp)
9c: 00813a03 ld s4,8(sp)
a0: 03010113 addi sp,sp,48
a4: 00078513 addi a0,a5,0
a8: 00008067 jalr zero,0(ra)
With RVC:
0: 02000813 addi a6,zero,32
4: 7179 c.addi16sp sp,-48
6: f422 c.sdsp s0,40(sp)
8: f026 c.sdsp s1,32(sp)
a: ec4a c.sdsp s2,24(sp)
c: e84e c.sdsp s3,16(sp)
e: e452 c.sdsp s4,8(sp)
10: 1800 c.addi4spn s0,sp,48
12: 03056683 lwu a3,48(a0)
16: 1682 c.slli a3,0x20
18: 9281 c.srli a3,0x20
1a: 03456703 lwu a4,52(a0)
1e: 1702 c.slli a4,0x20
20: 9301 c.srli a4,0x20
22: 03856483 lwu s1,56(a0)
26: 1482 c.slli s1,0x20
28: 9081 c.srli s1,0x20
2a: 03c56903 lwu s2,60(a0)
2e: 1902 c.slli s2,0x20
30: 02095913 srli s2,s2,0x20
34: 04056983 lwu s3,64(a0)
38: 1982 c.slli s3,0x20
3a: 0209d993 srli s3,s3,0x20
3e: 09056a03 lwu s4,144(a0)
42: 1a02 c.slli s4,0x20
44: 020a5a13 srli s4,s4,0x20
48: 4325 c.li t1,9
4a: 006a7363 bgeu s4,t1,0x50
4e: 4a01 c.li s4,0
50: d914 c.sw a3,48(a0)
52: d958 c.sw a4,52(a0)
54: dd04 c.sw s1,56(a0)
56: 03252e23 sw s2,60(a0)
5a: 05352023 sw s3,64(a0)
5e: 4781 c.li a5,0
60: 7422 c.ldsp s0,40(sp)
62: 7482 c.ldsp s1,32(sp)
64: 6962 c.ldsp s2,24(sp)
66: 69c2 c.ldsp s3,16(sp)
68: 6a22 c.ldsp s4,8(sp)
6a: 6145 c.addi16sp sp,48
6c: 853e c.mv a0,a5
6e: 8082 c.jr ra
Signed-off-by: Luke Nelson <luke.r.nels@gmail.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Cc: Björn Töpel <bjorn.topel@gmail.com>
Link: https://lore.kernel.org/bpf/20200721025241.8077-4-luke.r.nels@gmail.com
2020-07-20 19:52:40 -07:00
|
|
|
emit_add(RV_REG_T1, RV_REG_T1, rd, ctx);
|
bpf, riscv: add BPF JIT for RV64G
This commit adds a BPF JIT for RV64G.
The JIT is a two-pass JIT, and has a dynamic prolog/epilogue (similar
to the MIPS64 BPF JIT) instead of static ones (e.g. x86_64).
At the moment the RISC-V Linux port does not support
CONFIG_HAVE_KPROBES, which means that CONFIG_BPF_EVENTS is not
supported. Thus, no tests involving BPF_PROG_TYPE_TRACEPOINT,
BPF_PROG_TYPE_PERF_EVENT, BPF_PROG_TYPE_KPROBE and
BPF_PROG_TYPE_RAW_TRACEPOINT passes.
The implementation does not support "far branching" (>4KiB).
Test results:
# modprobe test_bpf
test_bpf: Summary: 378 PASSED, 0 FAILED, [366/366 JIT'ed]
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier
...
Summary: 761 PASSED, 507 SKIPPED, 2 FAILED
Note that "test_verifier" was run with one build with
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y and one without, otherwise
many of the the tests that require unaligned access were skipped.
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
0
No CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
59
The two failing test_verifier tests are:
"ld_abs: vlan + abs, test 1"
"ld_abs: jump around ld_abs"
This is due to that "far branching" involved in those tests.
All tests where done on QEMU (QEMU emulator version 3.1.50
(v3.1.0-688-g8ae951fbc106)).
Signed-off-by: Björn Töpel <bjorn.topel@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2019-02-05 05:41:22 -07:00
|
|
|
emit(rv_sb(RV_REG_T1, 0, rs), ctx);
|
|
|
|
break;
|
|
|
|
case BPF_STX | BPF_MEM | BPF_H:
|
|
|
|
if (is_12b_int(off)) {
|
|
|
|
emit(rv_sh(rd, off, rs), ctx);
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
|
|
|
|
emit_imm(RV_REG_T1, off, ctx);
|
bpf, riscv: Use compressed instructions in the rv64 JIT
This patch uses the RVC support and encodings from bpf_jit.h to optimize
the rv64 jit.
The optimizations work by replacing emit(rv_X(...)) with a call to a
helper function emit_X, which will emit a compressed version of the
instruction when possible, and when RVC is enabled.
The JIT continues to pass all tests in lib/test_bpf.c, and introduces
no new failures to test_verifier; both with and without RVC being enabled.
Most changes are straightforward replacements of emit(rv_X(...), ctx)
with emit_X(..., ctx), with the following exceptions bearing mention;
* Change emit_imm to sign-extend the value in "lower", since the
checks for RVC (and the instructions themselves) treat the value as
signed. Otherwise, small negative immediates will not be recognized as
encodable using an RVC instruction. For example, without this change,
emit_imm(rd, -1, ctx) would cause lower to become 4095, which is not a
6b int even though a "c.li rd, -1" instruction suffices.
* For {BPF_MOV,BPF_ADD} BPF_X, drop using addiw,addw in the 32-bit
cases since the values are zero-extended into the upper 32 bits in
the following instructions anyways, and the addition commutes with
zero-extension. (BPF_SUB BPF_X must still use subw since subtraction
does not commute with zero-extension.)
This patch avoids optimizing branches and jumps to use RVC instructions
since surrounding code often makes assumptions about the sizes of
emitted instructions. Optimizing these will require changing these
functions (e.g., emit_branch) to dynamically compute jump offsets.
The following are examples of the JITed code for the verifier selftest
"direct packet read test#3 for CGROUP_SKB OK", without and with RVC
enabled, respectively. The former uses 178 bytes, and the latter uses 112,
for a ~37% reduction in code size for this example.
Without RVC:
0: 02000813 addi a6,zero,32
4: fd010113 addi sp,sp,-48
8: 02813423 sd s0,40(sp)
c: 02913023 sd s1,32(sp)
10: 01213c23 sd s2,24(sp)
14: 01313823 sd s3,16(sp)
18: 01413423 sd s4,8(sp)
1c: 03010413 addi s0,sp,48
20: 03056683 lwu a3,48(a0)
24: 02069693 slli a3,a3,0x20
28: 0206d693 srli a3,a3,0x20
2c: 03456703 lwu a4,52(a0)
30: 02071713 slli a4,a4,0x20
34: 02075713 srli a4,a4,0x20
38: 03856483 lwu s1,56(a0)
3c: 02049493 slli s1,s1,0x20
40: 0204d493 srli s1,s1,0x20
44: 03c56903 lwu s2,60(a0)
48: 02091913 slli s2,s2,0x20
4c: 02095913 srli s2,s2,0x20
50: 04056983 lwu s3,64(a0)
54: 02099993 slli s3,s3,0x20
58: 0209d993 srli s3,s3,0x20
5c: 09056a03 lwu s4,144(a0)
60: 020a1a13 slli s4,s4,0x20
64: 020a5a13 srli s4,s4,0x20
68: 00900313 addi t1,zero,9
6c: 006a7463 bgeu s4,t1,0x74
70: 00000a13 addi s4,zero,0
74: 02d52823 sw a3,48(a0)
78: 02e52a23 sw a4,52(a0)
7c: 02952c23 sw s1,56(a0)
80: 03252e23 sw s2,60(a0)
84: 05352023 sw s3,64(a0)
88: 00000793 addi a5,zero,0
8c: 02813403 ld s0,40(sp)
90: 02013483 ld s1,32(sp)
94: 01813903 ld s2,24(sp)
98: 01013983 ld s3,16(sp)
9c: 00813a03 ld s4,8(sp)
a0: 03010113 addi sp,sp,48
a4: 00078513 addi a0,a5,0
a8: 00008067 jalr zero,0(ra)
With RVC:
0: 02000813 addi a6,zero,32
4: 7179 c.addi16sp sp,-48
6: f422 c.sdsp s0,40(sp)
8: f026 c.sdsp s1,32(sp)
a: ec4a c.sdsp s2,24(sp)
c: e84e c.sdsp s3,16(sp)
e: e452 c.sdsp s4,8(sp)
10: 1800 c.addi4spn s0,sp,48
12: 03056683 lwu a3,48(a0)
16: 1682 c.slli a3,0x20
18: 9281 c.srli a3,0x20
1a: 03456703 lwu a4,52(a0)
1e: 1702 c.slli a4,0x20
20: 9301 c.srli a4,0x20
22: 03856483 lwu s1,56(a0)
26: 1482 c.slli s1,0x20
28: 9081 c.srli s1,0x20
2a: 03c56903 lwu s2,60(a0)
2e: 1902 c.slli s2,0x20
30: 02095913 srli s2,s2,0x20
34: 04056983 lwu s3,64(a0)
38: 1982 c.slli s3,0x20
3a: 0209d993 srli s3,s3,0x20
3e: 09056a03 lwu s4,144(a0)
42: 1a02 c.slli s4,0x20
44: 020a5a13 srli s4,s4,0x20
48: 4325 c.li t1,9
4a: 006a7363 bgeu s4,t1,0x50
4e: 4a01 c.li s4,0
50: d914 c.sw a3,48(a0)
52: d958 c.sw a4,52(a0)
54: dd04 c.sw s1,56(a0)
56: 03252e23 sw s2,60(a0)
5a: 05352023 sw s3,64(a0)
5e: 4781 c.li a5,0
60: 7422 c.ldsp s0,40(sp)
62: 7482 c.ldsp s1,32(sp)
64: 6962 c.ldsp s2,24(sp)
66: 69c2 c.ldsp s3,16(sp)
68: 6a22 c.ldsp s4,8(sp)
6a: 6145 c.addi16sp sp,48
6c: 853e c.mv a0,a5
6e: 8082 c.jr ra
Signed-off-by: Luke Nelson <luke.r.nels@gmail.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Cc: Björn Töpel <bjorn.topel@gmail.com>
Link: https://lore.kernel.org/bpf/20200721025241.8077-4-luke.r.nels@gmail.com
2020-07-20 19:52:40 -07:00
|
|
|
emit_add(RV_REG_T1, RV_REG_T1, rd, ctx);
|
bpf, riscv: add BPF JIT for RV64G
This commit adds a BPF JIT for RV64G.
The JIT is a two-pass JIT, and has a dynamic prolog/epilogue (similar
to the MIPS64 BPF JIT) instead of static ones (e.g. x86_64).
At the moment the RISC-V Linux port does not support
CONFIG_HAVE_KPROBES, which means that CONFIG_BPF_EVENTS is not
supported. Thus, no tests involving BPF_PROG_TYPE_TRACEPOINT,
BPF_PROG_TYPE_PERF_EVENT, BPF_PROG_TYPE_KPROBE and
BPF_PROG_TYPE_RAW_TRACEPOINT passes.
The implementation does not support "far branching" (>4KiB).
Test results:
# modprobe test_bpf
test_bpf: Summary: 378 PASSED, 0 FAILED, [366/366 JIT'ed]
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier
...
Summary: 761 PASSED, 507 SKIPPED, 2 FAILED
Note that "test_verifier" was run with one build with
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y and one without, otherwise
many of the the tests that require unaligned access were skipped.
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
0
No CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
59
The two failing test_verifier tests are:
"ld_abs: vlan + abs, test 1"
"ld_abs: jump around ld_abs"
This is due to that "far branching" involved in those tests.
All tests where done on QEMU (QEMU emulator version 3.1.50
(v3.1.0-688-g8ae951fbc106)).
Signed-off-by: Björn Töpel <bjorn.topel@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2019-02-05 05:41:22 -07:00
|
|
|
emit(rv_sh(RV_REG_T1, 0, rs), ctx);
|
|
|
|
break;
|
|
|
|
case BPF_STX | BPF_MEM | BPF_W:
|
|
|
|
if (is_12b_int(off)) {
|
bpf, riscv: Use compressed instructions in the rv64 JIT
This patch uses the RVC support and encodings from bpf_jit.h to optimize
the rv64 jit.
The optimizations work by replacing emit(rv_X(...)) with a call to a
helper function emit_X, which will emit a compressed version of the
instruction when possible, and when RVC is enabled.
The JIT continues to pass all tests in lib/test_bpf.c, and introduces
no new failures to test_verifier; both with and without RVC being enabled.
Most changes are straightforward replacements of emit(rv_X(...), ctx)
with emit_X(..., ctx), with the following exceptions bearing mention;
* Change emit_imm to sign-extend the value in "lower", since the
checks for RVC (and the instructions themselves) treat the value as
signed. Otherwise, small negative immediates will not be recognized as
encodable using an RVC instruction. For example, without this change,
emit_imm(rd, -1, ctx) would cause lower to become 4095, which is not a
6b int even though a "c.li rd, -1" instruction suffices.
* For {BPF_MOV,BPF_ADD} BPF_X, drop using addiw,addw in the 32-bit
cases since the values are zero-extended into the upper 32 bits in
the following instructions anyways, and the addition commutes with
zero-extension. (BPF_SUB BPF_X must still use subw since subtraction
does not commute with zero-extension.)
This patch avoids optimizing branches and jumps to use RVC instructions
since surrounding code often makes assumptions about the sizes of
emitted instructions. Optimizing these will require changing these
functions (e.g., emit_branch) to dynamically compute jump offsets.
The following are examples of the JITed code for the verifier selftest
"direct packet read test#3 for CGROUP_SKB OK", without and with RVC
enabled, respectively. The former uses 178 bytes, and the latter uses 112,
for a ~37% reduction in code size for this example.
Without RVC:
0: 02000813 addi a6,zero,32
4: fd010113 addi sp,sp,-48
8: 02813423 sd s0,40(sp)
c: 02913023 sd s1,32(sp)
10: 01213c23 sd s2,24(sp)
14: 01313823 sd s3,16(sp)
18: 01413423 sd s4,8(sp)
1c: 03010413 addi s0,sp,48
20: 03056683 lwu a3,48(a0)
24: 02069693 slli a3,a3,0x20
28: 0206d693 srli a3,a3,0x20
2c: 03456703 lwu a4,52(a0)
30: 02071713 slli a4,a4,0x20
34: 02075713 srli a4,a4,0x20
38: 03856483 lwu s1,56(a0)
3c: 02049493 slli s1,s1,0x20
40: 0204d493 srli s1,s1,0x20
44: 03c56903 lwu s2,60(a0)
48: 02091913 slli s2,s2,0x20
4c: 02095913 srli s2,s2,0x20
50: 04056983 lwu s3,64(a0)
54: 02099993 slli s3,s3,0x20
58: 0209d993 srli s3,s3,0x20
5c: 09056a03 lwu s4,144(a0)
60: 020a1a13 slli s4,s4,0x20
64: 020a5a13 srli s4,s4,0x20
68: 00900313 addi t1,zero,9
6c: 006a7463 bgeu s4,t1,0x74
70: 00000a13 addi s4,zero,0
74: 02d52823 sw a3,48(a0)
78: 02e52a23 sw a4,52(a0)
7c: 02952c23 sw s1,56(a0)
80: 03252e23 sw s2,60(a0)
84: 05352023 sw s3,64(a0)
88: 00000793 addi a5,zero,0
8c: 02813403 ld s0,40(sp)
90: 02013483 ld s1,32(sp)
94: 01813903 ld s2,24(sp)
98: 01013983 ld s3,16(sp)
9c: 00813a03 ld s4,8(sp)
a0: 03010113 addi sp,sp,48
a4: 00078513 addi a0,a5,0
a8: 00008067 jalr zero,0(ra)
With RVC:
0: 02000813 addi a6,zero,32
4: 7179 c.addi16sp sp,-48
6: f422 c.sdsp s0,40(sp)
8: f026 c.sdsp s1,32(sp)
a: ec4a c.sdsp s2,24(sp)
c: e84e c.sdsp s3,16(sp)
e: e452 c.sdsp s4,8(sp)
10: 1800 c.addi4spn s0,sp,48
12: 03056683 lwu a3,48(a0)
16: 1682 c.slli a3,0x20
18: 9281 c.srli a3,0x20
1a: 03456703 lwu a4,52(a0)
1e: 1702 c.slli a4,0x20
20: 9301 c.srli a4,0x20
22: 03856483 lwu s1,56(a0)
26: 1482 c.slli s1,0x20
28: 9081 c.srli s1,0x20
2a: 03c56903 lwu s2,60(a0)
2e: 1902 c.slli s2,0x20
30: 02095913 srli s2,s2,0x20
34: 04056983 lwu s3,64(a0)
38: 1982 c.slli s3,0x20
3a: 0209d993 srli s3,s3,0x20
3e: 09056a03 lwu s4,144(a0)
42: 1a02 c.slli s4,0x20
44: 020a5a13 srli s4,s4,0x20
48: 4325 c.li t1,9
4a: 006a7363 bgeu s4,t1,0x50
4e: 4a01 c.li s4,0
50: d914 c.sw a3,48(a0)
52: d958 c.sw a4,52(a0)
54: dd04 c.sw s1,56(a0)
56: 03252e23 sw s2,60(a0)
5a: 05352023 sw s3,64(a0)
5e: 4781 c.li a5,0
60: 7422 c.ldsp s0,40(sp)
62: 7482 c.ldsp s1,32(sp)
64: 6962 c.ldsp s2,24(sp)
66: 69c2 c.ldsp s3,16(sp)
68: 6a22 c.ldsp s4,8(sp)
6a: 6145 c.addi16sp sp,48
6c: 853e c.mv a0,a5
6e: 8082 c.jr ra
Signed-off-by: Luke Nelson <luke.r.nels@gmail.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Cc: Björn Töpel <bjorn.topel@gmail.com>
Link: https://lore.kernel.org/bpf/20200721025241.8077-4-luke.r.nels@gmail.com
2020-07-20 19:52:40 -07:00
|
|
|
emit_sw(rd, off, rs, ctx);
|
bpf, riscv: add BPF JIT for RV64G
This commit adds a BPF JIT for RV64G.
The JIT is a two-pass JIT, and has a dynamic prolog/epilogue (similar
to the MIPS64 BPF JIT) instead of static ones (e.g. x86_64).
At the moment the RISC-V Linux port does not support
CONFIG_HAVE_KPROBES, which means that CONFIG_BPF_EVENTS is not
supported. Thus, no tests involving BPF_PROG_TYPE_TRACEPOINT,
BPF_PROG_TYPE_PERF_EVENT, BPF_PROG_TYPE_KPROBE and
BPF_PROG_TYPE_RAW_TRACEPOINT passes.
The implementation does not support "far branching" (>4KiB).
Test results:
# modprobe test_bpf
test_bpf: Summary: 378 PASSED, 0 FAILED, [366/366 JIT'ed]
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier
...
Summary: 761 PASSED, 507 SKIPPED, 2 FAILED
Note that "test_verifier" was run with one build with
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y and one without, otherwise
many of the the tests that require unaligned access were skipped.
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
0
No CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
59
The two failing test_verifier tests are:
"ld_abs: vlan + abs, test 1"
"ld_abs: jump around ld_abs"
This is due to that "far branching" involved in those tests.
All tests where done on QEMU (QEMU emulator version 3.1.50
(v3.1.0-688-g8ae951fbc106)).
Signed-off-by: Björn Töpel <bjorn.topel@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2019-02-05 05:41:22 -07:00
|
|
|
break;
|
|
|
|
}
|
|
|
|
|
|
|
|
emit_imm(RV_REG_T1, off, ctx);
|
bpf, riscv: Use compressed instructions in the rv64 JIT
This patch uses the RVC support and encodings from bpf_jit.h to optimize
the rv64 jit.
The optimizations work by replacing emit(rv_X(...)) with a call to a
helper function emit_X, which will emit a compressed version of the
instruction when possible, and when RVC is enabled.
The JIT continues to pass all tests in lib/test_bpf.c, and introduces
no new failures to test_verifier; both with and without RVC being enabled.
Most changes are straightforward replacements of emit(rv_X(...), ctx)
with emit_X(..., ctx), with the following exceptions bearing mention;
* Change emit_imm to sign-extend the value in "lower", since the
checks for RVC (and the instructions themselves) treat the value as
signed. Otherwise, small negative immediates will not be recognized as
encodable using an RVC instruction. For example, without this change,
emit_imm(rd, -1, ctx) would cause lower to become 4095, which is not a
6b int even though a "c.li rd, -1" instruction suffices.
* For {BPF_MOV,BPF_ADD} BPF_X, drop using addiw,addw in the 32-bit
cases since the values are zero-extended into the upper 32 bits in
the following instructions anyways, and the addition commutes with
zero-extension. (BPF_SUB BPF_X must still use subw since subtraction
does not commute with zero-extension.)
This patch avoids optimizing branches and jumps to use RVC instructions
since surrounding code often makes assumptions about the sizes of
emitted instructions. Optimizing these will require changing these
functions (e.g., emit_branch) to dynamically compute jump offsets.
The following are examples of the JITed code for the verifier selftest
"direct packet read test#3 for CGROUP_SKB OK", without and with RVC
enabled, respectively. The former uses 178 bytes, and the latter uses 112,
for a ~37% reduction in code size for this example.
Without RVC:
0: 02000813 addi a6,zero,32
4: fd010113 addi sp,sp,-48
8: 02813423 sd s0,40(sp)
c: 02913023 sd s1,32(sp)
10: 01213c23 sd s2,24(sp)
14: 01313823 sd s3,16(sp)
18: 01413423 sd s4,8(sp)
1c: 03010413 addi s0,sp,48
20: 03056683 lwu a3,48(a0)
24: 02069693 slli a3,a3,0x20
28: 0206d693 srli a3,a3,0x20
2c: 03456703 lwu a4,52(a0)
30: 02071713 slli a4,a4,0x20
34: 02075713 srli a4,a4,0x20
38: 03856483 lwu s1,56(a0)
3c: 02049493 slli s1,s1,0x20
40: 0204d493 srli s1,s1,0x20
44: 03c56903 lwu s2,60(a0)
48: 02091913 slli s2,s2,0x20
4c: 02095913 srli s2,s2,0x20
50: 04056983 lwu s3,64(a0)
54: 02099993 slli s3,s3,0x20
58: 0209d993 srli s3,s3,0x20
5c: 09056a03 lwu s4,144(a0)
60: 020a1a13 slli s4,s4,0x20
64: 020a5a13 srli s4,s4,0x20
68: 00900313 addi t1,zero,9
6c: 006a7463 bgeu s4,t1,0x74
70: 00000a13 addi s4,zero,0
74: 02d52823 sw a3,48(a0)
78: 02e52a23 sw a4,52(a0)
7c: 02952c23 sw s1,56(a0)
80: 03252e23 sw s2,60(a0)
84: 05352023 sw s3,64(a0)
88: 00000793 addi a5,zero,0
8c: 02813403 ld s0,40(sp)
90: 02013483 ld s1,32(sp)
94: 01813903 ld s2,24(sp)
98: 01013983 ld s3,16(sp)
9c: 00813a03 ld s4,8(sp)
a0: 03010113 addi sp,sp,48
a4: 00078513 addi a0,a5,0
a8: 00008067 jalr zero,0(ra)
With RVC:
0: 02000813 addi a6,zero,32
4: 7179 c.addi16sp sp,-48
6: f422 c.sdsp s0,40(sp)
8: f026 c.sdsp s1,32(sp)
a: ec4a c.sdsp s2,24(sp)
c: e84e c.sdsp s3,16(sp)
e: e452 c.sdsp s4,8(sp)
10: 1800 c.addi4spn s0,sp,48
12: 03056683 lwu a3,48(a0)
16: 1682 c.slli a3,0x20
18: 9281 c.srli a3,0x20
1a: 03456703 lwu a4,52(a0)
1e: 1702 c.slli a4,0x20
20: 9301 c.srli a4,0x20
22: 03856483 lwu s1,56(a0)
26: 1482 c.slli s1,0x20
28: 9081 c.srli s1,0x20
2a: 03c56903 lwu s2,60(a0)
2e: 1902 c.slli s2,0x20
30: 02095913 srli s2,s2,0x20
34: 04056983 lwu s3,64(a0)
38: 1982 c.slli s3,0x20
3a: 0209d993 srli s3,s3,0x20
3e: 09056a03 lwu s4,144(a0)
42: 1a02 c.slli s4,0x20
44: 020a5a13 srli s4,s4,0x20
48: 4325 c.li t1,9
4a: 006a7363 bgeu s4,t1,0x50
4e: 4a01 c.li s4,0
50: d914 c.sw a3,48(a0)
52: d958 c.sw a4,52(a0)
54: dd04 c.sw s1,56(a0)
56: 03252e23 sw s2,60(a0)
5a: 05352023 sw s3,64(a0)
5e: 4781 c.li a5,0
60: 7422 c.ldsp s0,40(sp)
62: 7482 c.ldsp s1,32(sp)
64: 6962 c.ldsp s2,24(sp)
66: 69c2 c.ldsp s3,16(sp)
68: 6a22 c.ldsp s4,8(sp)
6a: 6145 c.addi16sp sp,48
6c: 853e c.mv a0,a5
6e: 8082 c.jr ra
Signed-off-by: Luke Nelson <luke.r.nels@gmail.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Cc: Björn Töpel <bjorn.topel@gmail.com>
Link: https://lore.kernel.org/bpf/20200721025241.8077-4-luke.r.nels@gmail.com
2020-07-20 19:52:40 -07:00
|
|
|
emit_add(RV_REG_T1, RV_REG_T1, rd, ctx);
|
|
|
|
emit_sw(RV_REG_T1, 0, rs, ctx);
|
bpf, riscv: add BPF JIT for RV64G
This commit adds a BPF JIT for RV64G.
The JIT is a two-pass JIT, and has a dynamic prolog/epilogue (similar
to the MIPS64 BPF JIT) instead of static ones (e.g. x86_64).
At the moment the RISC-V Linux port does not support
CONFIG_HAVE_KPROBES, which means that CONFIG_BPF_EVENTS is not
supported. Thus, no tests involving BPF_PROG_TYPE_TRACEPOINT,
BPF_PROG_TYPE_PERF_EVENT, BPF_PROG_TYPE_KPROBE and
BPF_PROG_TYPE_RAW_TRACEPOINT passes.
The implementation does not support "far branching" (>4KiB).
Test results:
# modprobe test_bpf
test_bpf: Summary: 378 PASSED, 0 FAILED, [366/366 JIT'ed]
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier
...
Summary: 761 PASSED, 507 SKIPPED, 2 FAILED
Note that "test_verifier" was run with one build with
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y and one without, otherwise
many of the the tests that require unaligned access were skipped.
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
0
No CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
59
The two failing test_verifier tests are:
"ld_abs: vlan + abs, test 1"
"ld_abs: jump around ld_abs"
This is due to that "far branching" involved in those tests.
All tests where done on QEMU (QEMU emulator version 3.1.50
(v3.1.0-688-g8ae951fbc106)).
Signed-off-by: Björn Töpel <bjorn.topel@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2019-02-05 05:41:22 -07:00
|
|
|
break;
|
|
|
|
case BPF_STX | BPF_MEM | BPF_DW:
|
|
|
|
if (is_12b_int(off)) {
|
bpf, riscv: Use compressed instructions in the rv64 JIT
This patch uses the RVC support and encodings from bpf_jit.h to optimize
the rv64 jit.
The optimizations work by replacing emit(rv_X(...)) with a call to a
helper function emit_X, which will emit a compressed version of the
instruction when possible, and when RVC is enabled.
The JIT continues to pass all tests in lib/test_bpf.c, and introduces
no new failures to test_verifier; both with and without RVC being enabled.
Most changes are straightforward replacements of emit(rv_X(...), ctx)
with emit_X(..., ctx), with the following exceptions bearing mention;
* Change emit_imm to sign-extend the value in "lower", since the
checks for RVC (and the instructions themselves) treat the value as
signed. Otherwise, small negative immediates will not be recognized as
encodable using an RVC instruction. For example, without this change,
emit_imm(rd, -1, ctx) would cause lower to become 4095, which is not a
6b int even though a "c.li rd, -1" instruction suffices.
* For {BPF_MOV,BPF_ADD} BPF_X, drop using addiw,addw in the 32-bit
cases since the values are zero-extended into the upper 32 bits in
the following instructions anyways, and the addition commutes with
zero-extension. (BPF_SUB BPF_X must still use subw since subtraction
does not commute with zero-extension.)
This patch avoids optimizing branches and jumps to use RVC instructions
since surrounding code often makes assumptions about the sizes of
emitted instructions. Optimizing these will require changing these
functions (e.g., emit_branch) to dynamically compute jump offsets.
The following are examples of the JITed code for the verifier selftest
"direct packet read test#3 for CGROUP_SKB OK", without and with RVC
enabled, respectively. The former uses 178 bytes, and the latter uses 112,
for a ~37% reduction in code size for this example.
Without RVC:
0: 02000813 addi a6,zero,32
4: fd010113 addi sp,sp,-48
8: 02813423 sd s0,40(sp)
c: 02913023 sd s1,32(sp)
10: 01213c23 sd s2,24(sp)
14: 01313823 sd s3,16(sp)
18: 01413423 sd s4,8(sp)
1c: 03010413 addi s0,sp,48
20: 03056683 lwu a3,48(a0)
24: 02069693 slli a3,a3,0x20
28: 0206d693 srli a3,a3,0x20
2c: 03456703 lwu a4,52(a0)
30: 02071713 slli a4,a4,0x20
34: 02075713 srli a4,a4,0x20
38: 03856483 lwu s1,56(a0)
3c: 02049493 slli s1,s1,0x20
40: 0204d493 srli s1,s1,0x20
44: 03c56903 lwu s2,60(a0)
48: 02091913 slli s2,s2,0x20
4c: 02095913 srli s2,s2,0x20
50: 04056983 lwu s3,64(a0)
54: 02099993 slli s3,s3,0x20
58: 0209d993 srli s3,s3,0x20
5c: 09056a03 lwu s4,144(a0)
60: 020a1a13 slli s4,s4,0x20
64: 020a5a13 srli s4,s4,0x20
68: 00900313 addi t1,zero,9
6c: 006a7463 bgeu s4,t1,0x74
70: 00000a13 addi s4,zero,0
74: 02d52823 sw a3,48(a0)
78: 02e52a23 sw a4,52(a0)
7c: 02952c23 sw s1,56(a0)
80: 03252e23 sw s2,60(a0)
84: 05352023 sw s3,64(a0)
88: 00000793 addi a5,zero,0
8c: 02813403 ld s0,40(sp)
90: 02013483 ld s1,32(sp)
94: 01813903 ld s2,24(sp)
98: 01013983 ld s3,16(sp)
9c: 00813a03 ld s4,8(sp)
a0: 03010113 addi sp,sp,48
a4: 00078513 addi a0,a5,0
a8: 00008067 jalr zero,0(ra)
With RVC:
0: 02000813 addi a6,zero,32
4: 7179 c.addi16sp sp,-48
6: f422 c.sdsp s0,40(sp)
8: f026 c.sdsp s1,32(sp)
a: ec4a c.sdsp s2,24(sp)
c: e84e c.sdsp s3,16(sp)
e: e452 c.sdsp s4,8(sp)
10: 1800 c.addi4spn s0,sp,48
12: 03056683 lwu a3,48(a0)
16: 1682 c.slli a3,0x20
18: 9281 c.srli a3,0x20
1a: 03456703 lwu a4,52(a0)
1e: 1702 c.slli a4,0x20
20: 9301 c.srli a4,0x20
22: 03856483 lwu s1,56(a0)
26: 1482 c.slli s1,0x20
28: 9081 c.srli s1,0x20
2a: 03c56903 lwu s2,60(a0)
2e: 1902 c.slli s2,0x20
30: 02095913 srli s2,s2,0x20
34: 04056983 lwu s3,64(a0)
38: 1982 c.slli s3,0x20
3a: 0209d993 srli s3,s3,0x20
3e: 09056a03 lwu s4,144(a0)
42: 1a02 c.slli s4,0x20
44: 020a5a13 srli s4,s4,0x20
48: 4325 c.li t1,9
4a: 006a7363 bgeu s4,t1,0x50
4e: 4a01 c.li s4,0
50: d914 c.sw a3,48(a0)
52: d958 c.sw a4,52(a0)
54: dd04 c.sw s1,56(a0)
56: 03252e23 sw s2,60(a0)
5a: 05352023 sw s3,64(a0)
5e: 4781 c.li a5,0
60: 7422 c.ldsp s0,40(sp)
62: 7482 c.ldsp s1,32(sp)
64: 6962 c.ldsp s2,24(sp)
66: 69c2 c.ldsp s3,16(sp)
68: 6a22 c.ldsp s4,8(sp)
6a: 6145 c.addi16sp sp,48
6c: 853e c.mv a0,a5
6e: 8082 c.jr ra
Signed-off-by: Luke Nelson <luke.r.nels@gmail.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Cc: Björn Töpel <bjorn.topel@gmail.com>
Link: https://lore.kernel.org/bpf/20200721025241.8077-4-luke.r.nels@gmail.com
2020-07-20 19:52:40 -07:00
|
|
|
emit_sd(rd, off, rs, ctx);
|
bpf, riscv: add BPF JIT for RV64G
This commit adds a BPF JIT for RV64G.
The JIT is a two-pass JIT, and has a dynamic prolog/epilogue (similar
to the MIPS64 BPF JIT) instead of static ones (e.g. x86_64).
At the moment the RISC-V Linux port does not support
CONFIG_HAVE_KPROBES, which means that CONFIG_BPF_EVENTS is not
supported. Thus, no tests involving BPF_PROG_TYPE_TRACEPOINT,
BPF_PROG_TYPE_PERF_EVENT, BPF_PROG_TYPE_KPROBE and
BPF_PROG_TYPE_RAW_TRACEPOINT passes.
The implementation does not support "far branching" (>4KiB).
Test results:
# modprobe test_bpf
test_bpf: Summary: 378 PASSED, 0 FAILED, [366/366 JIT'ed]
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier
...
Summary: 761 PASSED, 507 SKIPPED, 2 FAILED
Note that "test_verifier" was run with one build with
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y and one without, otherwise
many of the the tests that require unaligned access were skipped.
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
0
No CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
59
The two failing test_verifier tests are:
"ld_abs: vlan + abs, test 1"
"ld_abs: jump around ld_abs"
This is due to that "far branching" involved in those tests.
All tests where done on QEMU (QEMU emulator version 3.1.50
(v3.1.0-688-g8ae951fbc106)).
Signed-off-by: Björn Töpel <bjorn.topel@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2019-02-05 05:41:22 -07:00
|
|
|
break;
|
|
|
|
}
|
|
|
|
|
|
|
|
emit_imm(RV_REG_T1, off, ctx);
|
bpf, riscv: Use compressed instructions in the rv64 JIT
This patch uses the RVC support and encodings from bpf_jit.h to optimize
the rv64 jit.
The optimizations work by replacing emit(rv_X(...)) with a call to a
helper function emit_X, which will emit a compressed version of the
instruction when possible, and when RVC is enabled.
The JIT continues to pass all tests in lib/test_bpf.c, and introduces
no new failures to test_verifier; both with and without RVC being enabled.
Most changes are straightforward replacements of emit(rv_X(...), ctx)
with emit_X(..., ctx), with the following exceptions bearing mention;
* Change emit_imm to sign-extend the value in "lower", since the
checks for RVC (and the instructions themselves) treat the value as
signed. Otherwise, small negative immediates will not be recognized as
encodable using an RVC instruction. For example, without this change,
emit_imm(rd, -1, ctx) would cause lower to become 4095, which is not a
6b int even though a "c.li rd, -1" instruction suffices.
* For {BPF_MOV,BPF_ADD} BPF_X, drop using addiw,addw in the 32-bit
cases since the values are zero-extended into the upper 32 bits in
the following instructions anyways, and the addition commutes with
zero-extension. (BPF_SUB BPF_X must still use subw since subtraction
does not commute with zero-extension.)
This patch avoids optimizing branches and jumps to use RVC instructions
since surrounding code often makes assumptions about the sizes of
emitted instructions. Optimizing these will require changing these
functions (e.g., emit_branch) to dynamically compute jump offsets.
The following are examples of the JITed code for the verifier selftest
"direct packet read test#3 for CGROUP_SKB OK", without and with RVC
enabled, respectively. The former uses 178 bytes, and the latter uses 112,
for a ~37% reduction in code size for this example.
Without RVC:
0: 02000813 addi a6,zero,32
4: fd010113 addi sp,sp,-48
8: 02813423 sd s0,40(sp)
c: 02913023 sd s1,32(sp)
10: 01213c23 sd s2,24(sp)
14: 01313823 sd s3,16(sp)
18: 01413423 sd s4,8(sp)
1c: 03010413 addi s0,sp,48
20: 03056683 lwu a3,48(a0)
24: 02069693 slli a3,a3,0x20
28: 0206d693 srli a3,a3,0x20
2c: 03456703 lwu a4,52(a0)
30: 02071713 slli a4,a4,0x20
34: 02075713 srli a4,a4,0x20
38: 03856483 lwu s1,56(a0)
3c: 02049493 slli s1,s1,0x20
40: 0204d493 srli s1,s1,0x20
44: 03c56903 lwu s2,60(a0)
48: 02091913 slli s2,s2,0x20
4c: 02095913 srli s2,s2,0x20
50: 04056983 lwu s3,64(a0)
54: 02099993 slli s3,s3,0x20
58: 0209d993 srli s3,s3,0x20
5c: 09056a03 lwu s4,144(a0)
60: 020a1a13 slli s4,s4,0x20
64: 020a5a13 srli s4,s4,0x20
68: 00900313 addi t1,zero,9
6c: 006a7463 bgeu s4,t1,0x74
70: 00000a13 addi s4,zero,0
74: 02d52823 sw a3,48(a0)
78: 02e52a23 sw a4,52(a0)
7c: 02952c23 sw s1,56(a0)
80: 03252e23 sw s2,60(a0)
84: 05352023 sw s3,64(a0)
88: 00000793 addi a5,zero,0
8c: 02813403 ld s0,40(sp)
90: 02013483 ld s1,32(sp)
94: 01813903 ld s2,24(sp)
98: 01013983 ld s3,16(sp)
9c: 00813a03 ld s4,8(sp)
a0: 03010113 addi sp,sp,48
a4: 00078513 addi a0,a5,0
a8: 00008067 jalr zero,0(ra)
With RVC:
0: 02000813 addi a6,zero,32
4: 7179 c.addi16sp sp,-48
6: f422 c.sdsp s0,40(sp)
8: f026 c.sdsp s1,32(sp)
a: ec4a c.sdsp s2,24(sp)
c: e84e c.sdsp s3,16(sp)
e: e452 c.sdsp s4,8(sp)
10: 1800 c.addi4spn s0,sp,48
12: 03056683 lwu a3,48(a0)
16: 1682 c.slli a3,0x20
18: 9281 c.srli a3,0x20
1a: 03456703 lwu a4,52(a0)
1e: 1702 c.slli a4,0x20
20: 9301 c.srli a4,0x20
22: 03856483 lwu s1,56(a0)
26: 1482 c.slli s1,0x20
28: 9081 c.srli s1,0x20
2a: 03c56903 lwu s2,60(a0)
2e: 1902 c.slli s2,0x20
30: 02095913 srli s2,s2,0x20
34: 04056983 lwu s3,64(a0)
38: 1982 c.slli s3,0x20
3a: 0209d993 srli s3,s3,0x20
3e: 09056a03 lwu s4,144(a0)
42: 1a02 c.slli s4,0x20
44: 020a5a13 srli s4,s4,0x20
48: 4325 c.li t1,9
4a: 006a7363 bgeu s4,t1,0x50
4e: 4a01 c.li s4,0
50: d914 c.sw a3,48(a0)
52: d958 c.sw a4,52(a0)
54: dd04 c.sw s1,56(a0)
56: 03252e23 sw s2,60(a0)
5a: 05352023 sw s3,64(a0)
5e: 4781 c.li a5,0
60: 7422 c.ldsp s0,40(sp)
62: 7482 c.ldsp s1,32(sp)
64: 6962 c.ldsp s2,24(sp)
66: 69c2 c.ldsp s3,16(sp)
68: 6a22 c.ldsp s4,8(sp)
6a: 6145 c.addi16sp sp,48
6c: 853e c.mv a0,a5
6e: 8082 c.jr ra
Signed-off-by: Luke Nelson <luke.r.nels@gmail.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Cc: Björn Töpel <bjorn.topel@gmail.com>
Link: https://lore.kernel.org/bpf/20200721025241.8077-4-luke.r.nels@gmail.com
2020-07-20 19:52:40 -07:00
|
|
|
emit_add(RV_REG_T1, RV_REG_T1, rd, ctx);
|
|
|
|
emit_sd(RV_REG_T1, 0, rs, ctx);
|
bpf, riscv: add BPF JIT for RV64G
This commit adds a BPF JIT for RV64G.
The JIT is a two-pass JIT, and has a dynamic prolog/epilogue (similar
to the MIPS64 BPF JIT) instead of static ones (e.g. x86_64).
At the moment the RISC-V Linux port does not support
CONFIG_HAVE_KPROBES, which means that CONFIG_BPF_EVENTS is not
supported. Thus, no tests involving BPF_PROG_TYPE_TRACEPOINT,
BPF_PROG_TYPE_PERF_EVENT, BPF_PROG_TYPE_KPROBE and
BPF_PROG_TYPE_RAW_TRACEPOINT passes.
The implementation does not support "far branching" (>4KiB).
Test results:
# modprobe test_bpf
test_bpf: Summary: 378 PASSED, 0 FAILED, [366/366 JIT'ed]
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier
...
Summary: 761 PASSED, 507 SKIPPED, 2 FAILED
Note that "test_verifier" was run with one build with
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y and one without, otherwise
many of the the tests that require unaligned access were skipped.
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
0
No CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
59
The two failing test_verifier tests are:
"ld_abs: vlan + abs, test 1"
"ld_abs: jump around ld_abs"
This is due to that "far branching" involved in those tests.
All tests where done on QEMU (QEMU emulator version 3.1.50
(v3.1.0-688-g8ae951fbc106)).
Signed-off-by: Björn Töpel <bjorn.topel@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2019-02-05 05:41:22 -07:00
|
|
|
break;
|
2021-01-14 11:17:44 -07:00
|
|
|
case BPF_STX | BPF_ATOMIC | BPF_W:
|
|
|
|
case BPF_STX | BPF_ATOMIC | BPF_DW:
|
2022-04-10 03:12:46 -07:00
|
|
|
emit_atomic(rd, rs, off, imm,
|
|
|
|
BPF_SIZE(code) == BPF_DW, ctx);
|
bpf, riscv: add BPF JIT for RV64G
This commit adds a BPF JIT for RV64G.
The JIT is a two-pass JIT, and has a dynamic prolog/epilogue (similar
to the MIPS64 BPF JIT) instead of static ones (e.g. x86_64).
At the moment the RISC-V Linux port does not support
CONFIG_HAVE_KPROBES, which means that CONFIG_BPF_EVENTS is not
supported. Thus, no tests involving BPF_PROG_TYPE_TRACEPOINT,
BPF_PROG_TYPE_PERF_EVENT, BPF_PROG_TYPE_KPROBE and
BPF_PROG_TYPE_RAW_TRACEPOINT passes.
The implementation does not support "far branching" (>4KiB).
Test results:
# modprobe test_bpf
test_bpf: Summary: 378 PASSED, 0 FAILED, [366/366 JIT'ed]
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier
...
Summary: 761 PASSED, 507 SKIPPED, 2 FAILED
Note that "test_verifier" was run with one build with
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y and one without, otherwise
many of the the tests that require unaligned access were skipped.
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
0
No CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
59
The two failing test_verifier tests are:
"ld_abs: vlan + abs, test 1"
"ld_abs: jump around ld_abs"
This is due to that "far branching" involved in those tests.
All tests where done on QEMU (QEMU emulator version 3.1.50
(v3.1.0-688-g8ae951fbc106)).
Signed-off-by: Björn Töpel <bjorn.topel@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2019-02-05 05:41:22 -07:00
|
|
|
break;
|
2024-04-04 04:42:02 -07:00
|
|
|
|
|
|
|
case BPF_STX | BPF_PROBE_MEM32 | BPF_B:
|
|
|
|
case BPF_STX | BPF_PROBE_MEM32 | BPF_H:
|
|
|
|
case BPF_STX | BPF_PROBE_MEM32 | BPF_W:
|
|
|
|
case BPF_STX | BPF_PROBE_MEM32 | BPF_DW:
|
|
|
|
{
|
|
|
|
int insn_len, insns_start;
|
|
|
|
|
|
|
|
emit_add(RV_REG_T2, rd, RV_REG_ARENA, ctx);
|
|
|
|
rd = RV_REG_T2;
|
|
|
|
|
|
|
|
switch (BPF_SIZE(code)) {
|
|
|
|
case BPF_B:
|
|
|
|
if (is_12b_int(off)) {
|
|
|
|
insns_start = ctx->ninsns;
|
|
|
|
emit(rv_sb(rd, off, rs), ctx);
|
|
|
|
insn_len = ctx->ninsns - insns_start;
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
|
|
|
|
emit_imm(RV_REG_T1, off, ctx);
|
|
|
|
emit_add(RV_REG_T1, RV_REG_T1, rd, ctx);
|
|
|
|
insns_start = ctx->ninsns;
|
|
|
|
emit(rv_sb(RV_REG_T1, 0, rs), ctx);
|
|
|
|
insn_len = ctx->ninsns - insns_start;
|
|
|
|
break;
|
|
|
|
case BPF_H:
|
|
|
|
if (is_12b_int(off)) {
|
|
|
|
insns_start = ctx->ninsns;
|
|
|
|
emit(rv_sh(rd, off, rs), ctx);
|
|
|
|
insn_len = ctx->ninsns - insns_start;
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
|
|
|
|
emit_imm(RV_REG_T1, off, ctx);
|
|
|
|
emit_add(RV_REG_T1, RV_REG_T1, rd, ctx);
|
|
|
|
insns_start = ctx->ninsns;
|
|
|
|
emit(rv_sh(RV_REG_T1, 0, rs), ctx);
|
|
|
|
insn_len = ctx->ninsns - insns_start;
|
|
|
|
break;
|
|
|
|
case BPF_W:
|
|
|
|
if (is_12b_int(off)) {
|
|
|
|
insns_start = ctx->ninsns;
|
|
|
|
emit_sw(rd, off, rs, ctx);
|
|
|
|
insn_len = ctx->ninsns - insns_start;
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
|
|
|
|
emit_imm(RV_REG_T1, off, ctx);
|
|
|
|
emit_add(RV_REG_T1, RV_REG_T1, rd, ctx);
|
|
|
|
insns_start = ctx->ninsns;
|
|
|
|
emit_sw(RV_REG_T1, 0, rs, ctx);
|
|
|
|
insn_len = ctx->ninsns - insns_start;
|
|
|
|
break;
|
|
|
|
case BPF_DW:
|
|
|
|
if (is_12b_int(off)) {
|
|
|
|
insns_start = ctx->ninsns;
|
|
|
|
emit_sd(rd, off, rs, ctx);
|
|
|
|
insn_len = ctx->ninsns - insns_start;
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
|
|
|
|
emit_imm(RV_REG_T1, off, ctx);
|
|
|
|
emit_add(RV_REG_T1, RV_REG_T1, rd, ctx);
|
|
|
|
insns_start = ctx->ninsns;
|
|
|
|
emit_sd(RV_REG_T1, 0, rs, ctx);
|
|
|
|
insn_len = ctx->ninsns - insns_start;
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
|
|
|
|
ret = add_exception_handler(insn, ctx, REG_DONT_CLEAR_MARKER,
|
|
|
|
insn_len);
|
|
|
|
if (ret)
|
|
|
|
return ret;
|
|
|
|
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
|
bpf, riscv: add BPF JIT for RV64G
This commit adds a BPF JIT for RV64G.
The JIT is a two-pass JIT, and has a dynamic prolog/epilogue (similar
to the MIPS64 BPF JIT) instead of static ones (e.g. x86_64).
At the moment the RISC-V Linux port does not support
CONFIG_HAVE_KPROBES, which means that CONFIG_BPF_EVENTS is not
supported. Thus, no tests involving BPF_PROG_TYPE_TRACEPOINT,
BPF_PROG_TYPE_PERF_EVENT, BPF_PROG_TYPE_KPROBE and
BPF_PROG_TYPE_RAW_TRACEPOINT passes.
The implementation does not support "far branching" (>4KiB).
Test results:
# modprobe test_bpf
test_bpf: Summary: 378 PASSED, 0 FAILED, [366/366 JIT'ed]
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier
...
Summary: 761 PASSED, 507 SKIPPED, 2 FAILED
Note that "test_verifier" was run with one build with
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y and one without, otherwise
many of the the tests that require unaligned access were skipped.
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
0
No CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
59
The two failing test_verifier tests are:
"ld_abs: vlan + abs, test 1"
"ld_abs: jump around ld_abs"
This is due to that "far branching" involved in those tests.
All tests where done on QEMU (QEMU emulator version 3.1.50
(v3.1.0-688-g8ae951fbc106)).
Signed-off-by: Björn Töpel <bjorn.topel@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2019-02-05 05:41:22 -07:00
|
|
|
default:
|
|
|
|
pr_err("bpf-jit: unknown opcode %02x\n", code);
|
|
|
|
return -EINVAL;
|
|
|
|
}
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2024-03-03 10:02:07 -07:00
|
|
|
void bpf_jit_build_prologue(struct rv_jit_context *ctx, bool is_subprog)
|
bpf, riscv: add BPF JIT for RV64G
This commit adds a BPF JIT for RV64G.
The JIT is a two-pass JIT, and has a dynamic prolog/epilogue (similar
to the MIPS64 BPF JIT) instead of static ones (e.g. x86_64).
At the moment the RISC-V Linux port does not support
CONFIG_HAVE_KPROBES, which means that CONFIG_BPF_EVENTS is not
supported. Thus, no tests involving BPF_PROG_TYPE_TRACEPOINT,
BPF_PROG_TYPE_PERF_EVENT, BPF_PROG_TYPE_KPROBE and
BPF_PROG_TYPE_RAW_TRACEPOINT passes.
The implementation does not support "far branching" (>4KiB).
Test results:
# modprobe test_bpf
test_bpf: Summary: 378 PASSED, 0 FAILED, [366/366 JIT'ed]
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier
...
Summary: 761 PASSED, 507 SKIPPED, 2 FAILED
Note that "test_verifier" was run with one build with
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y and one without, otherwise
many of the the tests that require unaligned access were skipped.
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
0
No CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
59
The two failing test_verifier tests are:
"ld_abs: vlan + abs, test 1"
"ld_abs: jump around ld_abs"
This is due to that "far branching" involved in those tests.
All tests where done on QEMU (QEMU emulator version 3.1.50
(v3.1.0-688-g8ae951fbc106)).
Signed-off-by: Björn Töpel <bjorn.topel@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2019-02-05 05:41:22 -07:00
|
|
|
{
|
2023-02-15 06:52:04 -07:00
|
|
|
int i, stack_adjust = 0, store_offset, bpf_stack_adjust;
|
bpf, riscv: add BPF JIT for RV64G
This commit adds a BPF JIT for RV64G.
The JIT is a two-pass JIT, and has a dynamic prolog/epilogue (similar
to the MIPS64 BPF JIT) instead of static ones (e.g. x86_64).
At the moment the RISC-V Linux port does not support
CONFIG_HAVE_KPROBES, which means that CONFIG_BPF_EVENTS is not
supported. Thus, no tests involving BPF_PROG_TYPE_TRACEPOINT,
BPF_PROG_TYPE_PERF_EVENT, BPF_PROG_TYPE_KPROBE and
BPF_PROG_TYPE_RAW_TRACEPOINT passes.
The implementation does not support "far branching" (>4KiB).
Test results:
# modprobe test_bpf
test_bpf: Summary: 378 PASSED, 0 FAILED, [366/366 JIT'ed]
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier
...
Summary: 761 PASSED, 507 SKIPPED, 2 FAILED
Note that "test_verifier" was run with one build with
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y and one without, otherwise
many of the the tests that require unaligned access were skipped.
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
0
No CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
59
The two failing test_verifier tests are:
"ld_abs: vlan + abs, test 1"
"ld_abs: jump around ld_abs"
This is due to that "far branching" involved in those tests.
All tests where done on QEMU (QEMU emulator version 3.1.50
(v3.1.0-688-g8ae951fbc106)).
Signed-off-by: Björn Töpel <bjorn.topel@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2019-02-05 05:41:22 -07:00
|
|
|
|
2019-12-16 02:13:35 -07:00
|
|
|
bpf_stack_adjust = round_up(ctx->prog->aux->stack_depth, 16);
|
|
|
|
if (bpf_stack_adjust)
|
|
|
|
mark_fp(ctx);
|
|
|
|
|
bpf, riscv: add BPF JIT for RV64G
This commit adds a BPF JIT for RV64G.
The JIT is a two-pass JIT, and has a dynamic prolog/epilogue (similar
to the MIPS64 BPF JIT) instead of static ones (e.g. x86_64).
At the moment the RISC-V Linux port does not support
CONFIG_HAVE_KPROBES, which means that CONFIG_BPF_EVENTS is not
supported. Thus, no tests involving BPF_PROG_TYPE_TRACEPOINT,
BPF_PROG_TYPE_PERF_EVENT, BPF_PROG_TYPE_KPROBE and
BPF_PROG_TYPE_RAW_TRACEPOINT passes.
The implementation does not support "far branching" (>4KiB).
Test results:
# modprobe test_bpf
test_bpf: Summary: 378 PASSED, 0 FAILED, [366/366 JIT'ed]
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier
...
Summary: 761 PASSED, 507 SKIPPED, 2 FAILED
Note that "test_verifier" was run with one build with
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y and one without, otherwise
many of the the tests that require unaligned access were skipped.
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
0
No CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
59
The two failing test_verifier tests are:
"ld_abs: vlan + abs, test 1"
"ld_abs: jump around ld_abs"
This is due to that "far branching" involved in those tests.
All tests where done on QEMU (QEMU emulator version 3.1.50
(v3.1.0-688-g8ae951fbc106)).
Signed-off-by: Björn Töpel <bjorn.topel@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2019-02-05 05:41:22 -07:00
|
|
|
if (seen_reg(RV_REG_RA, ctx))
|
|
|
|
stack_adjust += 8;
|
|
|
|
stack_adjust += 8; /* RV_REG_FP */
|
|
|
|
if (seen_reg(RV_REG_S1, ctx))
|
|
|
|
stack_adjust += 8;
|
|
|
|
if (seen_reg(RV_REG_S2, ctx))
|
|
|
|
stack_adjust += 8;
|
|
|
|
if (seen_reg(RV_REG_S3, ctx))
|
|
|
|
stack_adjust += 8;
|
|
|
|
if (seen_reg(RV_REG_S4, ctx))
|
|
|
|
stack_adjust += 8;
|
|
|
|
if (seen_reg(RV_REG_S5, ctx))
|
|
|
|
stack_adjust += 8;
|
|
|
|
if (seen_reg(RV_REG_S6, ctx))
|
|
|
|
stack_adjust += 8;
|
2024-04-04 04:42:02 -07:00
|
|
|
if (ctx->arena_vm_start)
|
|
|
|
stack_adjust += 8;
|
bpf, riscv: add BPF JIT for RV64G
This commit adds a BPF JIT for RV64G.
The JIT is a two-pass JIT, and has a dynamic prolog/epilogue (similar
to the MIPS64 BPF JIT) instead of static ones (e.g. x86_64).
At the moment the RISC-V Linux port does not support
CONFIG_HAVE_KPROBES, which means that CONFIG_BPF_EVENTS is not
supported. Thus, no tests involving BPF_PROG_TYPE_TRACEPOINT,
BPF_PROG_TYPE_PERF_EVENT, BPF_PROG_TYPE_KPROBE and
BPF_PROG_TYPE_RAW_TRACEPOINT passes.
The implementation does not support "far branching" (>4KiB).
Test results:
# modprobe test_bpf
test_bpf: Summary: 378 PASSED, 0 FAILED, [366/366 JIT'ed]
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier
...
Summary: 761 PASSED, 507 SKIPPED, 2 FAILED
Note that "test_verifier" was run with one build with
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y and one without, otherwise
many of the the tests that require unaligned access were skipped.
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
0
No CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
59
The two failing test_verifier tests are:
"ld_abs: vlan + abs, test 1"
"ld_abs: jump around ld_abs"
This is due to that "far branching" involved in those tests.
All tests where done on QEMU (QEMU emulator version 3.1.50
(v3.1.0-688-g8ae951fbc106)).
Signed-off-by: Björn Töpel <bjorn.topel@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2019-02-05 05:41:22 -07:00
|
|
|
|
|
|
|
stack_adjust = round_up(stack_adjust, 16);
|
|
|
|
stack_adjust += bpf_stack_adjust;
|
|
|
|
|
|
|
|
store_offset = stack_adjust - 8;
|
|
|
|
|
2024-03-03 10:02:07 -07:00
|
|
|
/* emit kcfi type preamble immediately before the first insn */
|
|
|
|
emit_kcfi(is_subprog ? cfi_bpf_subprog_hash : cfi_bpf_hash, ctx);
|
|
|
|
|
2023-07-21 03:06:27 -07:00
|
|
|
/* nops reserved for auipc+jalr pair */
|
|
|
|
for (i = 0; i < RV_FENTRY_NINSNS; i++)
|
2023-02-15 06:52:04 -07:00
|
|
|
emit(rv_nop(), ctx);
|
|
|
|
|
bpf, riscv: add BPF JIT for RV64G
This commit adds a BPF JIT for RV64G.
The JIT is a two-pass JIT, and has a dynamic prolog/epilogue (similar
to the MIPS64 BPF JIT) instead of static ones (e.g. x86_64).
At the moment the RISC-V Linux port does not support
CONFIG_HAVE_KPROBES, which means that CONFIG_BPF_EVENTS is not
supported. Thus, no tests involving BPF_PROG_TYPE_TRACEPOINT,
BPF_PROG_TYPE_PERF_EVENT, BPF_PROG_TYPE_KPROBE and
BPF_PROG_TYPE_RAW_TRACEPOINT passes.
The implementation does not support "far branching" (>4KiB).
Test results:
# modprobe test_bpf
test_bpf: Summary: 378 PASSED, 0 FAILED, [366/366 JIT'ed]
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier
...
Summary: 761 PASSED, 507 SKIPPED, 2 FAILED
Note that "test_verifier" was run with one build with
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y and one without, otherwise
many of the the tests that require unaligned access were skipped.
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
0
No CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
59
The two failing test_verifier tests are:
"ld_abs: vlan + abs, test 1"
"ld_abs: jump around ld_abs"
This is due to that "far branching" involved in those tests.
All tests where done on QEMU (QEMU emulator version 3.1.50
(v3.1.0-688-g8ae951fbc106)).
Signed-off-by: Björn Töpel <bjorn.topel@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2019-02-05 05:41:22 -07:00
|
|
|
/* First instruction is always setting the tail-call-counter
|
|
|
|
* (TCC) register. This instruction is skipped for tail calls.
|
bpf, riscv: Use compressed instructions in the rv64 JIT
This patch uses the RVC support and encodings from bpf_jit.h to optimize
the rv64 jit.
The optimizations work by replacing emit(rv_X(...)) with a call to a
helper function emit_X, which will emit a compressed version of the
instruction when possible, and when RVC is enabled.
The JIT continues to pass all tests in lib/test_bpf.c, and introduces
no new failures to test_verifier; both with and without RVC being enabled.
Most changes are straightforward replacements of emit(rv_X(...), ctx)
with emit_X(..., ctx), with the following exceptions bearing mention;
* Change emit_imm to sign-extend the value in "lower", since the
checks for RVC (and the instructions themselves) treat the value as
signed. Otherwise, small negative immediates will not be recognized as
encodable using an RVC instruction. For example, without this change,
emit_imm(rd, -1, ctx) would cause lower to become 4095, which is not a
6b int even though a "c.li rd, -1" instruction suffices.
* For {BPF_MOV,BPF_ADD} BPF_X, drop using addiw,addw in the 32-bit
cases since the values are zero-extended into the upper 32 bits in
the following instructions anyways, and the addition commutes with
zero-extension. (BPF_SUB BPF_X must still use subw since subtraction
does not commute with zero-extension.)
This patch avoids optimizing branches and jumps to use RVC instructions
since surrounding code often makes assumptions about the sizes of
emitted instructions. Optimizing these will require changing these
functions (e.g., emit_branch) to dynamically compute jump offsets.
The following are examples of the JITed code for the verifier selftest
"direct packet read test#3 for CGROUP_SKB OK", without and with RVC
enabled, respectively. The former uses 178 bytes, and the latter uses 112,
for a ~37% reduction in code size for this example.
Without RVC:
0: 02000813 addi a6,zero,32
4: fd010113 addi sp,sp,-48
8: 02813423 sd s0,40(sp)
c: 02913023 sd s1,32(sp)
10: 01213c23 sd s2,24(sp)
14: 01313823 sd s3,16(sp)
18: 01413423 sd s4,8(sp)
1c: 03010413 addi s0,sp,48
20: 03056683 lwu a3,48(a0)
24: 02069693 slli a3,a3,0x20
28: 0206d693 srli a3,a3,0x20
2c: 03456703 lwu a4,52(a0)
30: 02071713 slli a4,a4,0x20
34: 02075713 srli a4,a4,0x20
38: 03856483 lwu s1,56(a0)
3c: 02049493 slli s1,s1,0x20
40: 0204d493 srli s1,s1,0x20
44: 03c56903 lwu s2,60(a0)
48: 02091913 slli s2,s2,0x20
4c: 02095913 srli s2,s2,0x20
50: 04056983 lwu s3,64(a0)
54: 02099993 slli s3,s3,0x20
58: 0209d993 srli s3,s3,0x20
5c: 09056a03 lwu s4,144(a0)
60: 020a1a13 slli s4,s4,0x20
64: 020a5a13 srli s4,s4,0x20
68: 00900313 addi t1,zero,9
6c: 006a7463 bgeu s4,t1,0x74
70: 00000a13 addi s4,zero,0
74: 02d52823 sw a3,48(a0)
78: 02e52a23 sw a4,52(a0)
7c: 02952c23 sw s1,56(a0)
80: 03252e23 sw s2,60(a0)
84: 05352023 sw s3,64(a0)
88: 00000793 addi a5,zero,0
8c: 02813403 ld s0,40(sp)
90: 02013483 ld s1,32(sp)
94: 01813903 ld s2,24(sp)
98: 01013983 ld s3,16(sp)
9c: 00813a03 ld s4,8(sp)
a0: 03010113 addi sp,sp,48
a4: 00078513 addi a0,a5,0
a8: 00008067 jalr zero,0(ra)
With RVC:
0: 02000813 addi a6,zero,32
4: 7179 c.addi16sp sp,-48
6: f422 c.sdsp s0,40(sp)
8: f026 c.sdsp s1,32(sp)
a: ec4a c.sdsp s2,24(sp)
c: e84e c.sdsp s3,16(sp)
e: e452 c.sdsp s4,8(sp)
10: 1800 c.addi4spn s0,sp,48
12: 03056683 lwu a3,48(a0)
16: 1682 c.slli a3,0x20
18: 9281 c.srli a3,0x20
1a: 03456703 lwu a4,52(a0)
1e: 1702 c.slli a4,0x20
20: 9301 c.srli a4,0x20
22: 03856483 lwu s1,56(a0)
26: 1482 c.slli s1,0x20
28: 9081 c.srli s1,0x20
2a: 03c56903 lwu s2,60(a0)
2e: 1902 c.slli s2,0x20
30: 02095913 srli s2,s2,0x20
34: 04056983 lwu s3,64(a0)
38: 1982 c.slli s3,0x20
3a: 0209d993 srli s3,s3,0x20
3e: 09056a03 lwu s4,144(a0)
42: 1a02 c.slli s4,0x20
44: 020a5a13 srli s4,s4,0x20
48: 4325 c.li t1,9
4a: 006a7363 bgeu s4,t1,0x50
4e: 4a01 c.li s4,0
50: d914 c.sw a3,48(a0)
52: d958 c.sw a4,52(a0)
54: dd04 c.sw s1,56(a0)
56: 03252e23 sw s2,60(a0)
5a: 05352023 sw s3,64(a0)
5e: 4781 c.li a5,0
60: 7422 c.ldsp s0,40(sp)
62: 7482 c.ldsp s1,32(sp)
64: 6962 c.ldsp s2,24(sp)
66: 69c2 c.ldsp s3,16(sp)
68: 6a22 c.ldsp s4,8(sp)
6a: 6145 c.addi16sp sp,48
6c: 853e c.mv a0,a5
6e: 8082 c.jr ra
Signed-off-by: Luke Nelson <luke.r.nels@gmail.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Cc: Björn Töpel <bjorn.topel@gmail.com>
Link: https://lore.kernel.org/bpf/20200721025241.8077-4-luke.r.nels@gmail.com
2020-07-20 19:52:40 -07:00
|
|
|
* Force using a 4-byte (non-compressed) instruction.
|
bpf, riscv: add BPF JIT for RV64G
This commit adds a BPF JIT for RV64G.
The JIT is a two-pass JIT, and has a dynamic prolog/epilogue (similar
to the MIPS64 BPF JIT) instead of static ones (e.g. x86_64).
At the moment the RISC-V Linux port does not support
CONFIG_HAVE_KPROBES, which means that CONFIG_BPF_EVENTS is not
supported. Thus, no tests involving BPF_PROG_TYPE_TRACEPOINT,
BPF_PROG_TYPE_PERF_EVENT, BPF_PROG_TYPE_KPROBE and
BPF_PROG_TYPE_RAW_TRACEPOINT passes.
The implementation does not support "far branching" (>4KiB).
Test results:
# modprobe test_bpf
test_bpf: Summary: 378 PASSED, 0 FAILED, [366/366 JIT'ed]
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier
...
Summary: 761 PASSED, 507 SKIPPED, 2 FAILED
Note that "test_verifier" was run with one build with
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y and one without, otherwise
many of the the tests that require unaligned access were skipped.
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
0
No CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
59
The two failing test_verifier tests are:
"ld_abs: vlan + abs, test 1"
"ld_abs: jump around ld_abs"
This is due to that "far branching" involved in those tests.
All tests where done on QEMU (QEMU emulator version 3.1.50
(v3.1.0-688-g8ae951fbc106)).
Signed-off-by: Björn Töpel <bjorn.topel@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2019-02-05 05:41:22 -07:00
|
|
|
*/
|
|
|
|
emit(rv_addi(RV_REG_TCC, RV_REG_ZERO, MAX_TAIL_CALL_CNT), ctx);
|
|
|
|
|
bpf, riscv: Use compressed instructions in the rv64 JIT
This patch uses the RVC support and encodings from bpf_jit.h to optimize
the rv64 jit.
The optimizations work by replacing emit(rv_X(...)) with a call to a
helper function emit_X, which will emit a compressed version of the
instruction when possible, and when RVC is enabled.
The JIT continues to pass all tests in lib/test_bpf.c, and introduces
no new failures to test_verifier; both with and without RVC being enabled.
Most changes are straightforward replacements of emit(rv_X(...), ctx)
with emit_X(..., ctx), with the following exceptions bearing mention;
* Change emit_imm to sign-extend the value in "lower", since the
checks for RVC (and the instructions themselves) treat the value as
signed. Otherwise, small negative immediates will not be recognized as
encodable using an RVC instruction. For example, without this change,
emit_imm(rd, -1, ctx) would cause lower to become 4095, which is not a
6b int even though a "c.li rd, -1" instruction suffices.
* For {BPF_MOV,BPF_ADD} BPF_X, drop using addiw,addw in the 32-bit
cases since the values are zero-extended into the upper 32 bits in
the following instructions anyways, and the addition commutes with
zero-extension. (BPF_SUB BPF_X must still use subw since subtraction
does not commute with zero-extension.)
This patch avoids optimizing branches and jumps to use RVC instructions
since surrounding code often makes assumptions about the sizes of
emitted instructions. Optimizing these will require changing these
functions (e.g., emit_branch) to dynamically compute jump offsets.
The following are examples of the JITed code for the verifier selftest
"direct packet read test#3 for CGROUP_SKB OK", without and with RVC
enabled, respectively. The former uses 178 bytes, and the latter uses 112,
for a ~37% reduction in code size for this example.
Without RVC:
0: 02000813 addi a6,zero,32
4: fd010113 addi sp,sp,-48
8: 02813423 sd s0,40(sp)
c: 02913023 sd s1,32(sp)
10: 01213c23 sd s2,24(sp)
14: 01313823 sd s3,16(sp)
18: 01413423 sd s4,8(sp)
1c: 03010413 addi s0,sp,48
20: 03056683 lwu a3,48(a0)
24: 02069693 slli a3,a3,0x20
28: 0206d693 srli a3,a3,0x20
2c: 03456703 lwu a4,52(a0)
30: 02071713 slli a4,a4,0x20
34: 02075713 srli a4,a4,0x20
38: 03856483 lwu s1,56(a0)
3c: 02049493 slli s1,s1,0x20
40: 0204d493 srli s1,s1,0x20
44: 03c56903 lwu s2,60(a0)
48: 02091913 slli s2,s2,0x20
4c: 02095913 srli s2,s2,0x20
50: 04056983 lwu s3,64(a0)
54: 02099993 slli s3,s3,0x20
58: 0209d993 srli s3,s3,0x20
5c: 09056a03 lwu s4,144(a0)
60: 020a1a13 slli s4,s4,0x20
64: 020a5a13 srli s4,s4,0x20
68: 00900313 addi t1,zero,9
6c: 006a7463 bgeu s4,t1,0x74
70: 00000a13 addi s4,zero,0
74: 02d52823 sw a3,48(a0)
78: 02e52a23 sw a4,52(a0)
7c: 02952c23 sw s1,56(a0)
80: 03252e23 sw s2,60(a0)
84: 05352023 sw s3,64(a0)
88: 00000793 addi a5,zero,0
8c: 02813403 ld s0,40(sp)
90: 02013483 ld s1,32(sp)
94: 01813903 ld s2,24(sp)
98: 01013983 ld s3,16(sp)
9c: 00813a03 ld s4,8(sp)
a0: 03010113 addi sp,sp,48
a4: 00078513 addi a0,a5,0
a8: 00008067 jalr zero,0(ra)
With RVC:
0: 02000813 addi a6,zero,32
4: 7179 c.addi16sp sp,-48
6: f422 c.sdsp s0,40(sp)
8: f026 c.sdsp s1,32(sp)
a: ec4a c.sdsp s2,24(sp)
c: e84e c.sdsp s3,16(sp)
e: e452 c.sdsp s4,8(sp)
10: 1800 c.addi4spn s0,sp,48
12: 03056683 lwu a3,48(a0)
16: 1682 c.slli a3,0x20
18: 9281 c.srli a3,0x20
1a: 03456703 lwu a4,52(a0)
1e: 1702 c.slli a4,0x20
20: 9301 c.srli a4,0x20
22: 03856483 lwu s1,56(a0)
26: 1482 c.slli s1,0x20
28: 9081 c.srli s1,0x20
2a: 03c56903 lwu s2,60(a0)
2e: 1902 c.slli s2,0x20
30: 02095913 srli s2,s2,0x20
34: 04056983 lwu s3,64(a0)
38: 1982 c.slli s3,0x20
3a: 0209d993 srli s3,s3,0x20
3e: 09056a03 lwu s4,144(a0)
42: 1a02 c.slli s4,0x20
44: 020a5a13 srli s4,s4,0x20
48: 4325 c.li t1,9
4a: 006a7363 bgeu s4,t1,0x50
4e: 4a01 c.li s4,0
50: d914 c.sw a3,48(a0)
52: d958 c.sw a4,52(a0)
54: dd04 c.sw s1,56(a0)
56: 03252e23 sw s2,60(a0)
5a: 05352023 sw s3,64(a0)
5e: 4781 c.li a5,0
60: 7422 c.ldsp s0,40(sp)
62: 7482 c.ldsp s1,32(sp)
64: 6962 c.ldsp s2,24(sp)
66: 69c2 c.ldsp s3,16(sp)
68: 6a22 c.ldsp s4,8(sp)
6a: 6145 c.addi16sp sp,48
6c: 853e c.mv a0,a5
6e: 8082 c.jr ra
Signed-off-by: Luke Nelson <luke.r.nels@gmail.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Cc: Björn Töpel <bjorn.topel@gmail.com>
Link: https://lore.kernel.org/bpf/20200721025241.8077-4-luke.r.nels@gmail.com
2020-07-20 19:52:40 -07:00
|
|
|
emit_addi(RV_REG_SP, RV_REG_SP, -stack_adjust, ctx);
|
bpf, riscv: add BPF JIT for RV64G
This commit adds a BPF JIT for RV64G.
The JIT is a two-pass JIT, and has a dynamic prolog/epilogue (similar
to the MIPS64 BPF JIT) instead of static ones (e.g. x86_64).
At the moment the RISC-V Linux port does not support
CONFIG_HAVE_KPROBES, which means that CONFIG_BPF_EVENTS is not
supported. Thus, no tests involving BPF_PROG_TYPE_TRACEPOINT,
BPF_PROG_TYPE_PERF_EVENT, BPF_PROG_TYPE_KPROBE and
BPF_PROG_TYPE_RAW_TRACEPOINT passes.
The implementation does not support "far branching" (>4KiB).
Test results:
# modprobe test_bpf
test_bpf: Summary: 378 PASSED, 0 FAILED, [366/366 JIT'ed]
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier
...
Summary: 761 PASSED, 507 SKIPPED, 2 FAILED
Note that "test_verifier" was run with one build with
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y and one without, otherwise
many of the the tests that require unaligned access were skipped.
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
0
No CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
59
The two failing test_verifier tests are:
"ld_abs: vlan + abs, test 1"
"ld_abs: jump around ld_abs"
This is due to that "far branching" involved in those tests.
All tests where done on QEMU (QEMU emulator version 3.1.50
(v3.1.0-688-g8ae951fbc106)).
Signed-off-by: Björn Töpel <bjorn.topel@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2019-02-05 05:41:22 -07:00
|
|
|
|
|
|
|
if (seen_reg(RV_REG_RA, ctx)) {
|
bpf, riscv: Use compressed instructions in the rv64 JIT
This patch uses the RVC support and encodings from bpf_jit.h to optimize
the rv64 jit.
The optimizations work by replacing emit(rv_X(...)) with a call to a
helper function emit_X, which will emit a compressed version of the
instruction when possible, and when RVC is enabled.
The JIT continues to pass all tests in lib/test_bpf.c, and introduces
no new failures to test_verifier; both with and without RVC being enabled.
Most changes are straightforward replacements of emit(rv_X(...), ctx)
with emit_X(..., ctx), with the following exceptions bearing mention;
* Change emit_imm to sign-extend the value in "lower", since the
checks for RVC (and the instructions themselves) treat the value as
signed. Otherwise, small negative immediates will not be recognized as
encodable using an RVC instruction. For example, without this change,
emit_imm(rd, -1, ctx) would cause lower to become 4095, which is not a
6b int even though a "c.li rd, -1" instruction suffices.
* For {BPF_MOV,BPF_ADD} BPF_X, drop using addiw,addw in the 32-bit
cases since the values are zero-extended into the upper 32 bits in
the following instructions anyways, and the addition commutes with
zero-extension. (BPF_SUB BPF_X must still use subw since subtraction
does not commute with zero-extension.)
This patch avoids optimizing branches and jumps to use RVC instructions
since surrounding code often makes assumptions about the sizes of
emitted instructions. Optimizing these will require changing these
functions (e.g., emit_branch) to dynamically compute jump offsets.
The following are examples of the JITed code for the verifier selftest
"direct packet read test#3 for CGROUP_SKB OK", without and with RVC
enabled, respectively. The former uses 178 bytes, and the latter uses 112,
for a ~37% reduction in code size for this example.
Without RVC:
0: 02000813 addi a6,zero,32
4: fd010113 addi sp,sp,-48
8: 02813423 sd s0,40(sp)
c: 02913023 sd s1,32(sp)
10: 01213c23 sd s2,24(sp)
14: 01313823 sd s3,16(sp)
18: 01413423 sd s4,8(sp)
1c: 03010413 addi s0,sp,48
20: 03056683 lwu a3,48(a0)
24: 02069693 slli a3,a3,0x20
28: 0206d693 srli a3,a3,0x20
2c: 03456703 lwu a4,52(a0)
30: 02071713 slli a4,a4,0x20
34: 02075713 srli a4,a4,0x20
38: 03856483 lwu s1,56(a0)
3c: 02049493 slli s1,s1,0x20
40: 0204d493 srli s1,s1,0x20
44: 03c56903 lwu s2,60(a0)
48: 02091913 slli s2,s2,0x20
4c: 02095913 srli s2,s2,0x20
50: 04056983 lwu s3,64(a0)
54: 02099993 slli s3,s3,0x20
58: 0209d993 srli s3,s3,0x20
5c: 09056a03 lwu s4,144(a0)
60: 020a1a13 slli s4,s4,0x20
64: 020a5a13 srli s4,s4,0x20
68: 00900313 addi t1,zero,9
6c: 006a7463 bgeu s4,t1,0x74
70: 00000a13 addi s4,zero,0
74: 02d52823 sw a3,48(a0)
78: 02e52a23 sw a4,52(a0)
7c: 02952c23 sw s1,56(a0)
80: 03252e23 sw s2,60(a0)
84: 05352023 sw s3,64(a0)
88: 00000793 addi a5,zero,0
8c: 02813403 ld s0,40(sp)
90: 02013483 ld s1,32(sp)
94: 01813903 ld s2,24(sp)
98: 01013983 ld s3,16(sp)
9c: 00813a03 ld s4,8(sp)
a0: 03010113 addi sp,sp,48
a4: 00078513 addi a0,a5,0
a8: 00008067 jalr zero,0(ra)
With RVC:
0: 02000813 addi a6,zero,32
4: 7179 c.addi16sp sp,-48
6: f422 c.sdsp s0,40(sp)
8: f026 c.sdsp s1,32(sp)
a: ec4a c.sdsp s2,24(sp)
c: e84e c.sdsp s3,16(sp)
e: e452 c.sdsp s4,8(sp)
10: 1800 c.addi4spn s0,sp,48
12: 03056683 lwu a3,48(a0)
16: 1682 c.slli a3,0x20
18: 9281 c.srli a3,0x20
1a: 03456703 lwu a4,52(a0)
1e: 1702 c.slli a4,0x20
20: 9301 c.srli a4,0x20
22: 03856483 lwu s1,56(a0)
26: 1482 c.slli s1,0x20
28: 9081 c.srli s1,0x20
2a: 03c56903 lwu s2,60(a0)
2e: 1902 c.slli s2,0x20
30: 02095913 srli s2,s2,0x20
34: 04056983 lwu s3,64(a0)
38: 1982 c.slli s3,0x20
3a: 0209d993 srli s3,s3,0x20
3e: 09056a03 lwu s4,144(a0)
42: 1a02 c.slli s4,0x20
44: 020a5a13 srli s4,s4,0x20
48: 4325 c.li t1,9
4a: 006a7363 bgeu s4,t1,0x50
4e: 4a01 c.li s4,0
50: d914 c.sw a3,48(a0)
52: d958 c.sw a4,52(a0)
54: dd04 c.sw s1,56(a0)
56: 03252e23 sw s2,60(a0)
5a: 05352023 sw s3,64(a0)
5e: 4781 c.li a5,0
60: 7422 c.ldsp s0,40(sp)
62: 7482 c.ldsp s1,32(sp)
64: 6962 c.ldsp s2,24(sp)
66: 69c2 c.ldsp s3,16(sp)
68: 6a22 c.ldsp s4,8(sp)
6a: 6145 c.addi16sp sp,48
6c: 853e c.mv a0,a5
6e: 8082 c.jr ra
Signed-off-by: Luke Nelson <luke.r.nels@gmail.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Cc: Björn Töpel <bjorn.topel@gmail.com>
Link: https://lore.kernel.org/bpf/20200721025241.8077-4-luke.r.nels@gmail.com
2020-07-20 19:52:40 -07:00
|
|
|
emit_sd(RV_REG_SP, store_offset, RV_REG_RA, ctx);
|
bpf, riscv: add BPF JIT for RV64G
This commit adds a BPF JIT for RV64G.
The JIT is a two-pass JIT, and has a dynamic prolog/epilogue (similar
to the MIPS64 BPF JIT) instead of static ones (e.g. x86_64).
At the moment the RISC-V Linux port does not support
CONFIG_HAVE_KPROBES, which means that CONFIG_BPF_EVENTS is not
supported. Thus, no tests involving BPF_PROG_TYPE_TRACEPOINT,
BPF_PROG_TYPE_PERF_EVENT, BPF_PROG_TYPE_KPROBE and
BPF_PROG_TYPE_RAW_TRACEPOINT passes.
The implementation does not support "far branching" (>4KiB).
Test results:
# modprobe test_bpf
test_bpf: Summary: 378 PASSED, 0 FAILED, [366/366 JIT'ed]
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier
...
Summary: 761 PASSED, 507 SKIPPED, 2 FAILED
Note that "test_verifier" was run with one build with
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y and one without, otherwise
many of the the tests that require unaligned access were skipped.
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
0
No CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
59
The two failing test_verifier tests are:
"ld_abs: vlan + abs, test 1"
"ld_abs: jump around ld_abs"
This is due to that "far branching" involved in those tests.
All tests where done on QEMU (QEMU emulator version 3.1.50
(v3.1.0-688-g8ae951fbc106)).
Signed-off-by: Björn Töpel <bjorn.topel@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2019-02-05 05:41:22 -07:00
|
|
|
store_offset -= 8;
|
|
|
|
}
|
bpf, riscv: Use compressed instructions in the rv64 JIT
This patch uses the RVC support and encodings from bpf_jit.h to optimize
the rv64 jit.
The optimizations work by replacing emit(rv_X(...)) with a call to a
helper function emit_X, which will emit a compressed version of the
instruction when possible, and when RVC is enabled.
The JIT continues to pass all tests in lib/test_bpf.c, and introduces
no new failures to test_verifier; both with and without RVC being enabled.
Most changes are straightforward replacements of emit(rv_X(...), ctx)
with emit_X(..., ctx), with the following exceptions bearing mention;
* Change emit_imm to sign-extend the value in "lower", since the
checks for RVC (and the instructions themselves) treat the value as
signed. Otherwise, small negative immediates will not be recognized as
encodable using an RVC instruction. For example, without this change,
emit_imm(rd, -1, ctx) would cause lower to become 4095, which is not a
6b int even though a "c.li rd, -1" instruction suffices.
* For {BPF_MOV,BPF_ADD} BPF_X, drop using addiw,addw in the 32-bit
cases since the values are zero-extended into the upper 32 bits in
the following instructions anyways, and the addition commutes with
zero-extension. (BPF_SUB BPF_X must still use subw since subtraction
does not commute with zero-extension.)
This patch avoids optimizing branches and jumps to use RVC instructions
since surrounding code often makes assumptions about the sizes of
emitted instructions. Optimizing these will require changing these
functions (e.g., emit_branch) to dynamically compute jump offsets.
The following are examples of the JITed code for the verifier selftest
"direct packet read test#3 for CGROUP_SKB OK", without and with RVC
enabled, respectively. The former uses 178 bytes, and the latter uses 112,
for a ~37% reduction in code size for this example.
Without RVC:
0: 02000813 addi a6,zero,32
4: fd010113 addi sp,sp,-48
8: 02813423 sd s0,40(sp)
c: 02913023 sd s1,32(sp)
10: 01213c23 sd s2,24(sp)
14: 01313823 sd s3,16(sp)
18: 01413423 sd s4,8(sp)
1c: 03010413 addi s0,sp,48
20: 03056683 lwu a3,48(a0)
24: 02069693 slli a3,a3,0x20
28: 0206d693 srli a3,a3,0x20
2c: 03456703 lwu a4,52(a0)
30: 02071713 slli a4,a4,0x20
34: 02075713 srli a4,a4,0x20
38: 03856483 lwu s1,56(a0)
3c: 02049493 slli s1,s1,0x20
40: 0204d493 srli s1,s1,0x20
44: 03c56903 lwu s2,60(a0)
48: 02091913 slli s2,s2,0x20
4c: 02095913 srli s2,s2,0x20
50: 04056983 lwu s3,64(a0)
54: 02099993 slli s3,s3,0x20
58: 0209d993 srli s3,s3,0x20
5c: 09056a03 lwu s4,144(a0)
60: 020a1a13 slli s4,s4,0x20
64: 020a5a13 srli s4,s4,0x20
68: 00900313 addi t1,zero,9
6c: 006a7463 bgeu s4,t1,0x74
70: 00000a13 addi s4,zero,0
74: 02d52823 sw a3,48(a0)
78: 02e52a23 sw a4,52(a0)
7c: 02952c23 sw s1,56(a0)
80: 03252e23 sw s2,60(a0)
84: 05352023 sw s3,64(a0)
88: 00000793 addi a5,zero,0
8c: 02813403 ld s0,40(sp)
90: 02013483 ld s1,32(sp)
94: 01813903 ld s2,24(sp)
98: 01013983 ld s3,16(sp)
9c: 00813a03 ld s4,8(sp)
a0: 03010113 addi sp,sp,48
a4: 00078513 addi a0,a5,0
a8: 00008067 jalr zero,0(ra)
With RVC:
0: 02000813 addi a6,zero,32
4: 7179 c.addi16sp sp,-48
6: f422 c.sdsp s0,40(sp)
8: f026 c.sdsp s1,32(sp)
a: ec4a c.sdsp s2,24(sp)
c: e84e c.sdsp s3,16(sp)
e: e452 c.sdsp s4,8(sp)
10: 1800 c.addi4spn s0,sp,48
12: 03056683 lwu a3,48(a0)
16: 1682 c.slli a3,0x20
18: 9281 c.srli a3,0x20
1a: 03456703 lwu a4,52(a0)
1e: 1702 c.slli a4,0x20
20: 9301 c.srli a4,0x20
22: 03856483 lwu s1,56(a0)
26: 1482 c.slli s1,0x20
28: 9081 c.srli s1,0x20
2a: 03c56903 lwu s2,60(a0)
2e: 1902 c.slli s2,0x20
30: 02095913 srli s2,s2,0x20
34: 04056983 lwu s3,64(a0)
38: 1982 c.slli s3,0x20
3a: 0209d993 srli s3,s3,0x20
3e: 09056a03 lwu s4,144(a0)
42: 1a02 c.slli s4,0x20
44: 020a5a13 srli s4,s4,0x20
48: 4325 c.li t1,9
4a: 006a7363 bgeu s4,t1,0x50
4e: 4a01 c.li s4,0
50: d914 c.sw a3,48(a0)
52: d958 c.sw a4,52(a0)
54: dd04 c.sw s1,56(a0)
56: 03252e23 sw s2,60(a0)
5a: 05352023 sw s3,64(a0)
5e: 4781 c.li a5,0
60: 7422 c.ldsp s0,40(sp)
62: 7482 c.ldsp s1,32(sp)
64: 6962 c.ldsp s2,24(sp)
66: 69c2 c.ldsp s3,16(sp)
68: 6a22 c.ldsp s4,8(sp)
6a: 6145 c.addi16sp sp,48
6c: 853e c.mv a0,a5
6e: 8082 c.jr ra
Signed-off-by: Luke Nelson <luke.r.nels@gmail.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Cc: Björn Töpel <bjorn.topel@gmail.com>
Link: https://lore.kernel.org/bpf/20200721025241.8077-4-luke.r.nels@gmail.com
2020-07-20 19:52:40 -07:00
|
|
|
emit_sd(RV_REG_SP, store_offset, RV_REG_FP, ctx);
|
bpf, riscv: add BPF JIT for RV64G
This commit adds a BPF JIT for RV64G.
The JIT is a two-pass JIT, and has a dynamic prolog/epilogue (similar
to the MIPS64 BPF JIT) instead of static ones (e.g. x86_64).
At the moment the RISC-V Linux port does not support
CONFIG_HAVE_KPROBES, which means that CONFIG_BPF_EVENTS is not
supported. Thus, no tests involving BPF_PROG_TYPE_TRACEPOINT,
BPF_PROG_TYPE_PERF_EVENT, BPF_PROG_TYPE_KPROBE and
BPF_PROG_TYPE_RAW_TRACEPOINT passes.
The implementation does not support "far branching" (>4KiB).
Test results:
# modprobe test_bpf
test_bpf: Summary: 378 PASSED, 0 FAILED, [366/366 JIT'ed]
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier
...
Summary: 761 PASSED, 507 SKIPPED, 2 FAILED
Note that "test_verifier" was run with one build with
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y and one without, otherwise
many of the the tests that require unaligned access were skipped.
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
0
No CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
59
The two failing test_verifier tests are:
"ld_abs: vlan + abs, test 1"
"ld_abs: jump around ld_abs"
This is due to that "far branching" involved in those tests.
All tests where done on QEMU (QEMU emulator version 3.1.50
(v3.1.0-688-g8ae951fbc106)).
Signed-off-by: Björn Töpel <bjorn.topel@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2019-02-05 05:41:22 -07:00
|
|
|
store_offset -= 8;
|
|
|
|
if (seen_reg(RV_REG_S1, ctx)) {
|
bpf, riscv: Use compressed instructions in the rv64 JIT
This patch uses the RVC support and encodings from bpf_jit.h to optimize
the rv64 jit.
The optimizations work by replacing emit(rv_X(...)) with a call to a
helper function emit_X, which will emit a compressed version of the
instruction when possible, and when RVC is enabled.
The JIT continues to pass all tests in lib/test_bpf.c, and introduces
no new failures to test_verifier; both with and without RVC being enabled.
Most changes are straightforward replacements of emit(rv_X(...), ctx)
with emit_X(..., ctx), with the following exceptions bearing mention;
* Change emit_imm to sign-extend the value in "lower", since the
checks for RVC (and the instructions themselves) treat the value as
signed. Otherwise, small negative immediates will not be recognized as
encodable using an RVC instruction. For example, without this change,
emit_imm(rd, -1, ctx) would cause lower to become 4095, which is not a
6b int even though a "c.li rd, -1" instruction suffices.
* For {BPF_MOV,BPF_ADD} BPF_X, drop using addiw,addw in the 32-bit
cases since the values are zero-extended into the upper 32 bits in
the following instructions anyways, and the addition commutes with
zero-extension. (BPF_SUB BPF_X must still use subw since subtraction
does not commute with zero-extension.)
This patch avoids optimizing branches and jumps to use RVC instructions
since surrounding code often makes assumptions about the sizes of
emitted instructions. Optimizing these will require changing these
functions (e.g., emit_branch) to dynamically compute jump offsets.
The following are examples of the JITed code for the verifier selftest
"direct packet read test#3 for CGROUP_SKB OK", without and with RVC
enabled, respectively. The former uses 178 bytes, and the latter uses 112,
for a ~37% reduction in code size for this example.
Without RVC:
0: 02000813 addi a6,zero,32
4: fd010113 addi sp,sp,-48
8: 02813423 sd s0,40(sp)
c: 02913023 sd s1,32(sp)
10: 01213c23 sd s2,24(sp)
14: 01313823 sd s3,16(sp)
18: 01413423 sd s4,8(sp)
1c: 03010413 addi s0,sp,48
20: 03056683 lwu a3,48(a0)
24: 02069693 slli a3,a3,0x20
28: 0206d693 srli a3,a3,0x20
2c: 03456703 lwu a4,52(a0)
30: 02071713 slli a4,a4,0x20
34: 02075713 srli a4,a4,0x20
38: 03856483 lwu s1,56(a0)
3c: 02049493 slli s1,s1,0x20
40: 0204d493 srli s1,s1,0x20
44: 03c56903 lwu s2,60(a0)
48: 02091913 slli s2,s2,0x20
4c: 02095913 srli s2,s2,0x20
50: 04056983 lwu s3,64(a0)
54: 02099993 slli s3,s3,0x20
58: 0209d993 srli s3,s3,0x20
5c: 09056a03 lwu s4,144(a0)
60: 020a1a13 slli s4,s4,0x20
64: 020a5a13 srli s4,s4,0x20
68: 00900313 addi t1,zero,9
6c: 006a7463 bgeu s4,t1,0x74
70: 00000a13 addi s4,zero,0
74: 02d52823 sw a3,48(a0)
78: 02e52a23 sw a4,52(a0)
7c: 02952c23 sw s1,56(a0)
80: 03252e23 sw s2,60(a0)
84: 05352023 sw s3,64(a0)
88: 00000793 addi a5,zero,0
8c: 02813403 ld s0,40(sp)
90: 02013483 ld s1,32(sp)
94: 01813903 ld s2,24(sp)
98: 01013983 ld s3,16(sp)
9c: 00813a03 ld s4,8(sp)
a0: 03010113 addi sp,sp,48
a4: 00078513 addi a0,a5,0
a8: 00008067 jalr zero,0(ra)
With RVC:
0: 02000813 addi a6,zero,32
4: 7179 c.addi16sp sp,-48
6: f422 c.sdsp s0,40(sp)
8: f026 c.sdsp s1,32(sp)
a: ec4a c.sdsp s2,24(sp)
c: e84e c.sdsp s3,16(sp)
e: e452 c.sdsp s4,8(sp)
10: 1800 c.addi4spn s0,sp,48
12: 03056683 lwu a3,48(a0)
16: 1682 c.slli a3,0x20
18: 9281 c.srli a3,0x20
1a: 03456703 lwu a4,52(a0)
1e: 1702 c.slli a4,0x20
20: 9301 c.srli a4,0x20
22: 03856483 lwu s1,56(a0)
26: 1482 c.slli s1,0x20
28: 9081 c.srli s1,0x20
2a: 03c56903 lwu s2,60(a0)
2e: 1902 c.slli s2,0x20
30: 02095913 srli s2,s2,0x20
34: 04056983 lwu s3,64(a0)
38: 1982 c.slli s3,0x20
3a: 0209d993 srli s3,s3,0x20
3e: 09056a03 lwu s4,144(a0)
42: 1a02 c.slli s4,0x20
44: 020a5a13 srli s4,s4,0x20
48: 4325 c.li t1,9
4a: 006a7363 bgeu s4,t1,0x50
4e: 4a01 c.li s4,0
50: d914 c.sw a3,48(a0)
52: d958 c.sw a4,52(a0)
54: dd04 c.sw s1,56(a0)
56: 03252e23 sw s2,60(a0)
5a: 05352023 sw s3,64(a0)
5e: 4781 c.li a5,0
60: 7422 c.ldsp s0,40(sp)
62: 7482 c.ldsp s1,32(sp)
64: 6962 c.ldsp s2,24(sp)
66: 69c2 c.ldsp s3,16(sp)
68: 6a22 c.ldsp s4,8(sp)
6a: 6145 c.addi16sp sp,48
6c: 853e c.mv a0,a5
6e: 8082 c.jr ra
Signed-off-by: Luke Nelson <luke.r.nels@gmail.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Cc: Björn Töpel <bjorn.topel@gmail.com>
Link: https://lore.kernel.org/bpf/20200721025241.8077-4-luke.r.nels@gmail.com
2020-07-20 19:52:40 -07:00
|
|
|
emit_sd(RV_REG_SP, store_offset, RV_REG_S1, ctx);
|
bpf, riscv: add BPF JIT for RV64G
This commit adds a BPF JIT for RV64G.
The JIT is a two-pass JIT, and has a dynamic prolog/epilogue (similar
to the MIPS64 BPF JIT) instead of static ones (e.g. x86_64).
At the moment the RISC-V Linux port does not support
CONFIG_HAVE_KPROBES, which means that CONFIG_BPF_EVENTS is not
supported. Thus, no tests involving BPF_PROG_TYPE_TRACEPOINT,
BPF_PROG_TYPE_PERF_EVENT, BPF_PROG_TYPE_KPROBE and
BPF_PROG_TYPE_RAW_TRACEPOINT passes.
The implementation does not support "far branching" (>4KiB).
Test results:
# modprobe test_bpf
test_bpf: Summary: 378 PASSED, 0 FAILED, [366/366 JIT'ed]
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier
...
Summary: 761 PASSED, 507 SKIPPED, 2 FAILED
Note that "test_verifier" was run with one build with
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y and one without, otherwise
many of the the tests that require unaligned access were skipped.
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
0
No CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
59
The two failing test_verifier tests are:
"ld_abs: vlan + abs, test 1"
"ld_abs: jump around ld_abs"
This is due to that "far branching" involved in those tests.
All tests where done on QEMU (QEMU emulator version 3.1.50
(v3.1.0-688-g8ae951fbc106)).
Signed-off-by: Björn Töpel <bjorn.topel@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2019-02-05 05:41:22 -07:00
|
|
|
store_offset -= 8;
|
|
|
|
}
|
|
|
|
if (seen_reg(RV_REG_S2, ctx)) {
|
bpf, riscv: Use compressed instructions in the rv64 JIT
This patch uses the RVC support and encodings from bpf_jit.h to optimize
the rv64 jit.
The optimizations work by replacing emit(rv_X(...)) with a call to a
helper function emit_X, which will emit a compressed version of the
instruction when possible, and when RVC is enabled.
The JIT continues to pass all tests in lib/test_bpf.c, and introduces
no new failures to test_verifier; both with and without RVC being enabled.
Most changes are straightforward replacements of emit(rv_X(...), ctx)
with emit_X(..., ctx), with the following exceptions bearing mention;
* Change emit_imm to sign-extend the value in "lower", since the
checks for RVC (and the instructions themselves) treat the value as
signed. Otherwise, small negative immediates will not be recognized as
encodable using an RVC instruction. For example, without this change,
emit_imm(rd, -1, ctx) would cause lower to become 4095, which is not a
6b int even though a "c.li rd, -1" instruction suffices.
* For {BPF_MOV,BPF_ADD} BPF_X, drop using addiw,addw in the 32-bit
cases since the values are zero-extended into the upper 32 bits in
the following instructions anyways, and the addition commutes with
zero-extension. (BPF_SUB BPF_X must still use subw since subtraction
does not commute with zero-extension.)
This patch avoids optimizing branches and jumps to use RVC instructions
since surrounding code often makes assumptions about the sizes of
emitted instructions. Optimizing these will require changing these
functions (e.g., emit_branch) to dynamically compute jump offsets.
The following are examples of the JITed code for the verifier selftest
"direct packet read test#3 for CGROUP_SKB OK", without and with RVC
enabled, respectively. The former uses 178 bytes, and the latter uses 112,
for a ~37% reduction in code size for this example.
Without RVC:
0: 02000813 addi a6,zero,32
4: fd010113 addi sp,sp,-48
8: 02813423 sd s0,40(sp)
c: 02913023 sd s1,32(sp)
10: 01213c23 sd s2,24(sp)
14: 01313823 sd s3,16(sp)
18: 01413423 sd s4,8(sp)
1c: 03010413 addi s0,sp,48
20: 03056683 lwu a3,48(a0)
24: 02069693 slli a3,a3,0x20
28: 0206d693 srli a3,a3,0x20
2c: 03456703 lwu a4,52(a0)
30: 02071713 slli a4,a4,0x20
34: 02075713 srli a4,a4,0x20
38: 03856483 lwu s1,56(a0)
3c: 02049493 slli s1,s1,0x20
40: 0204d493 srli s1,s1,0x20
44: 03c56903 lwu s2,60(a0)
48: 02091913 slli s2,s2,0x20
4c: 02095913 srli s2,s2,0x20
50: 04056983 lwu s3,64(a0)
54: 02099993 slli s3,s3,0x20
58: 0209d993 srli s3,s3,0x20
5c: 09056a03 lwu s4,144(a0)
60: 020a1a13 slli s4,s4,0x20
64: 020a5a13 srli s4,s4,0x20
68: 00900313 addi t1,zero,9
6c: 006a7463 bgeu s4,t1,0x74
70: 00000a13 addi s4,zero,0
74: 02d52823 sw a3,48(a0)
78: 02e52a23 sw a4,52(a0)
7c: 02952c23 sw s1,56(a0)
80: 03252e23 sw s2,60(a0)
84: 05352023 sw s3,64(a0)
88: 00000793 addi a5,zero,0
8c: 02813403 ld s0,40(sp)
90: 02013483 ld s1,32(sp)
94: 01813903 ld s2,24(sp)
98: 01013983 ld s3,16(sp)
9c: 00813a03 ld s4,8(sp)
a0: 03010113 addi sp,sp,48
a4: 00078513 addi a0,a5,0
a8: 00008067 jalr zero,0(ra)
With RVC:
0: 02000813 addi a6,zero,32
4: 7179 c.addi16sp sp,-48
6: f422 c.sdsp s0,40(sp)
8: f026 c.sdsp s1,32(sp)
a: ec4a c.sdsp s2,24(sp)
c: e84e c.sdsp s3,16(sp)
e: e452 c.sdsp s4,8(sp)
10: 1800 c.addi4spn s0,sp,48
12: 03056683 lwu a3,48(a0)
16: 1682 c.slli a3,0x20
18: 9281 c.srli a3,0x20
1a: 03456703 lwu a4,52(a0)
1e: 1702 c.slli a4,0x20
20: 9301 c.srli a4,0x20
22: 03856483 lwu s1,56(a0)
26: 1482 c.slli s1,0x20
28: 9081 c.srli s1,0x20
2a: 03c56903 lwu s2,60(a0)
2e: 1902 c.slli s2,0x20
30: 02095913 srli s2,s2,0x20
34: 04056983 lwu s3,64(a0)
38: 1982 c.slli s3,0x20
3a: 0209d993 srli s3,s3,0x20
3e: 09056a03 lwu s4,144(a0)
42: 1a02 c.slli s4,0x20
44: 020a5a13 srli s4,s4,0x20
48: 4325 c.li t1,9
4a: 006a7363 bgeu s4,t1,0x50
4e: 4a01 c.li s4,0
50: d914 c.sw a3,48(a0)
52: d958 c.sw a4,52(a0)
54: dd04 c.sw s1,56(a0)
56: 03252e23 sw s2,60(a0)
5a: 05352023 sw s3,64(a0)
5e: 4781 c.li a5,0
60: 7422 c.ldsp s0,40(sp)
62: 7482 c.ldsp s1,32(sp)
64: 6962 c.ldsp s2,24(sp)
66: 69c2 c.ldsp s3,16(sp)
68: 6a22 c.ldsp s4,8(sp)
6a: 6145 c.addi16sp sp,48
6c: 853e c.mv a0,a5
6e: 8082 c.jr ra
Signed-off-by: Luke Nelson <luke.r.nels@gmail.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Cc: Björn Töpel <bjorn.topel@gmail.com>
Link: https://lore.kernel.org/bpf/20200721025241.8077-4-luke.r.nels@gmail.com
2020-07-20 19:52:40 -07:00
|
|
|
emit_sd(RV_REG_SP, store_offset, RV_REG_S2, ctx);
|
bpf, riscv: add BPF JIT for RV64G
This commit adds a BPF JIT for RV64G.
The JIT is a two-pass JIT, and has a dynamic prolog/epilogue (similar
to the MIPS64 BPF JIT) instead of static ones (e.g. x86_64).
At the moment the RISC-V Linux port does not support
CONFIG_HAVE_KPROBES, which means that CONFIG_BPF_EVENTS is not
supported. Thus, no tests involving BPF_PROG_TYPE_TRACEPOINT,
BPF_PROG_TYPE_PERF_EVENT, BPF_PROG_TYPE_KPROBE and
BPF_PROG_TYPE_RAW_TRACEPOINT passes.
The implementation does not support "far branching" (>4KiB).
Test results:
# modprobe test_bpf
test_bpf: Summary: 378 PASSED, 0 FAILED, [366/366 JIT'ed]
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier
...
Summary: 761 PASSED, 507 SKIPPED, 2 FAILED
Note that "test_verifier" was run with one build with
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y and one without, otherwise
many of the the tests that require unaligned access were skipped.
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
0
No CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
59
The two failing test_verifier tests are:
"ld_abs: vlan + abs, test 1"
"ld_abs: jump around ld_abs"
This is due to that "far branching" involved in those tests.
All tests where done on QEMU (QEMU emulator version 3.1.50
(v3.1.0-688-g8ae951fbc106)).
Signed-off-by: Björn Töpel <bjorn.topel@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2019-02-05 05:41:22 -07:00
|
|
|
store_offset -= 8;
|
|
|
|
}
|
|
|
|
if (seen_reg(RV_REG_S3, ctx)) {
|
bpf, riscv: Use compressed instructions in the rv64 JIT
This patch uses the RVC support and encodings from bpf_jit.h to optimize
the rv64 jit.
The optimizations work by replacing emit(rv_X(...)) with a call to a
helper function emit_X, which will emit a compressed version of the
instruction when possible, and when RVC is enabled.
The JIT continues to pass all tests in lib/test_bpf.c, and introduces
no new failures to test_verifier; both with and without RVC being enabled.
Most changes are straightforward replacements of emit(rv_X(...), ctx)
with emit_X(..., ctx), with the following exceptions bearing mention;
* Change emit_imm to sign-extend the value in "lower", since the
checks for RVC (and the instructions themselves) treat the value as
signed. Otherwise, small negative immediates will not be recognized as
encodable using an RVC instruction. For example, without this change,
emit_imm(rd, -1, ctx) would cause lower to become 4095, which is not a
6b int even though a "c.li rd, -1" instruction suffices.
* For {BPF_MOV,BPF_ADD} BPF_X, drop using addiw,addw in the 32-bit
cases since the values are zero-extended into the upper 32 bits in
the following instructions anyways, and the addition commutes with
zero-extension. (BPF_SUB BPF_X must still use subw since subtraction
does not commute with zero-extension.)
This patch avoids optimizing branches and jumps to use RVC instructions
since surrounding code often makes assumptions about the sizes of
emitted instructions. Optimizing these will require changing these
functions (e.g., emit_branch) to dynamically compute jump offsets.
The following are examples of the JITed code for the verifier selftest
"direct packet read test#3 for CGROUP_SKB OK", without and with RVC
enabled, respectively. The former uses 178 bytes, and the latter uses 112,
for a ~37% reduction in code size for this example.
Without RVC:
0: 02000813 addi a6,zero,32
4: fd010113 addi sp,sp,-48
8: 02813423 sd s0,40(sp)
c: 02913023 sd s1,32(sp)
10: 01213c23 sd s2,24(sp)
14: 01313823 sd s3,16(sp)
18: 01413423 sd s4,8(sp)
1c: 03010413 addi s0,sp,48
20: 03056683 lwu a3,48(a0)
24: 02069693 slli a3,a3,0x20
28: 0206d693 srli a3,a3,0x20
2c: 03456703 lwu a4,52(a0)
30: 02071713 slli a4,a4,0x20
34: 02075713 srli a4,a4,0x20
38: 03856483 lwu s1,56(a0)
3c: 02049493 slli s1,s1,0x20
40: 0204d493 srli s1,s1,0x20
44: 03c56903 lwu s2,60(a0)
48: 02091913 slli s2,s2,0x20
4c: 02095913 srli s2,s2,0x20
50: 04056983 lwu s3,64(a0)
54: 02099993 slli s3,s3,0x20
58: 0209d993 srli s3,s3,0x20
5c: 09056a03 lwu s4,144(a0)
60: 020a1a13 slli s4,s4,0x20
64: 020a5a13 srli s4,s4,0x20
68: 00900313 addi t1,zero,9
6c: 006a7463 bgeu s4,t1,0x74
70: 00000a13 addi s4,zero,0
74: 02d52823 sw a3,48(a0)
78: 02e52a23 sw a4,52(a0)
7c: 02952c23 sw s1,56(a0)
80: 03252e23 sw s2,60(a0)
84: 05352023 sw s3,64(a0)
88: 00000793 addi a5,zero,0
8c: 02813403 ld s0,40(sp)
90: 02013483 ld s1,32(sp)
94: 01813903 ld s2,24(sp)
98: 01013983 ld s3,16(sp)
9c: 00813a03 ld s4,8(sp)
a0: 03010113 addi sp,sp,48
a4: 00078513 addi a0,a5,0
a8: 00008067 jalr zero,0(ra)
With RVC:
0: 02000813 addi a6,zero,32
4: 7179 c.addi16sp sp,-48
6: f422 c.sdsp s0,40(sp)
8: f026 c.sdsp s1,32(sp)
a: ec4a c.sdsp s2,24(sp)
c: e84e c.sdsp s3,16(sp)
e: e452 c.sdsp s4,8(sp)
10: 1800 c.addi4spn s0,sp,48
12: 03056683 lwu a3,48(a0)
16: 1682 c.slli a3,0x20
18: 9281 c.srli a3,0x20
1a: 03456703 lwu a4,52(a0)
1e: 1702 c.slli a4,0x20
20: 9301 c.srli a4,0x20
22: 03856483 lwu s1,56(a0)
26: 1482 c.slli s1,0x20
28: 9081 c.srli s1,0x20
2a: 03c56903 lwu s2,60(a0)
2e: 1902 c.slli s2,0x20
30: 02095913 srli s2,s2,0x20
34: 04056983 lwu s3,64(a0)
38: 1982 c.slli s3,0x20
3a: 0209d993 srli s3,s3,0x20
3e: 09056a03 lwu s4,144(a0)
42: 1a02 c.slli s4,0x20
44: 020a5a13 srli s4,s4,0x20
48: 4325 c.li t1,9
4a: 006a7363 bgeu s4,t1,0x50
4e: 4a01 c.li s4,0
50: d914 c.sw a3,48(a0)
52: d958 c.sw a4,52(a0)
54: dd04 c.sw s1,56(a0)
56: 03252e23 sw s2,60(a0)
5a: 05352023 sw s3,64(a0)
5e: 4781 c.li a5,0
60: 7422 c.ldsp s0,40(sp)
62: 7482 c.ldsp s1,32(sp)
64: 6962 c.ldsp s2,24(sp)
66: 69c2 c.ldsp s3,16(sp)
68: 6a22 c.ldsp s4,8(sp)
6a: 6145 c.addi16sp sp,48
6c: 853e c.mv a0,a5
6e: 8082 c.jr ra
Signed-off-by: Luke Nelson <luke.r.nels@gmail.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Cc: Björn Töpel <bjorn.topel@gmail.com>
Link: https://lore.kernel.org/bpf/20200721025241.8077-4-luke.r.nels@gmail.com
2020-07-20 19:52:40 -07:00
|
|
|
emit_sd(RV_REG_SP, store_offset, RV_REG_S3, ctx);
|
bpf, riscv: add BPF JIT for RV64G
This commit adds a BPF JIT for RV64G.
The JIT is a two-pass JIT, and has a dynamic prolog/epilogue (similar
to the MIPS64 BPF JIT) instead of static ones (e.g. x86_64).
At the moment the RISC-V Linux port does not support
CONFIG_HAVE_KPROBES, which means that CONFIG_BPF_EVENTS is not
supported. Thus, no tests involving BPF_PROG_TYPE_TRACEPOINT,
BPF_PROG_TYPE_PERF_EVENT, BPF_PROG_TYPE_KPROBE and
BPF_PROG_TYPE_RAW_TRACEPOINT passes.
The implementation does not support "far branching" (>4KiB).
Test results:
# modprobe test_bpf
test_bpf: Summary: 378 PASSED, 0 FAILED, [366/366 JIT'ed]
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier
...
Summary: 761 PASSED, 507 SKIPPED, 2 FAILED
Note that "test_verifier" was run with one build with
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y and one without, otherwise
many of the the tests that require unaligned access were skipped.
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
0
No CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
59
The two failing test_verifier tests are:
"ld_abs: vlan + abs, test 1"
"ld_abs: jump around ld_abs"
This is due to that "far branching" involved in those tests.
All tests where done on QEMU (QEMU emulator version 3.1.50
(v3.1.0-688-g8ae951fbc106)).
Signed-off-by: Björn Töpel <bjorn.topel@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2019-02-05 05:41:22 -07:00
|
|
|
store_offset -= 8;
|
|
|
|
}
|
|
|
|
if (seen_reg(RV_REG_S4, ctx)) {
|
bpf, riscv: Use compressed instructions in the rv64 JIT
This patch uses the RVC support and encodings from bpf_jit.h to optimize
the rv64 jit.
The optimizations work by replacing emit(rv_X(...)) with a call to a
helper function emit_X, which will emit a compressed version of the
instruction when possible, and when RVC is enabled.
The JIT continues to pass all tests in lib/test_bpf.c, and introduces
no new failures to test_verifier; both with and without RVC being enabled.
Most changes are straightforward replacements of emit(rv_X(...), ctx)
with emit_X(..., ctx), with the following exceptions bearing mention;
* Change emit_imm to sign-extend the value in "lower", since the
checks for RVC (and the instructions themselves) treat the value as
signed. Otherwise, small negative immediates will not be recognized as
encodable using an RVC instruction. For example, without this change,
emit_imm(rd, -1, ctx) would cause lower to become 4095, which is not a
6b int even though a "c.li rd, -1" instruction suffices.
* For {BPF_MOV,BPF_ADD} BPF_X, drop using addiw,addw in the 32-bit
cases since the values are zero-extended into the upper 32 bits in
the following instructions anyways, and the addition commutes with
zero-extension. (BPF_SUB BPF_X must still use subw since subtraction
does not commute with zero-extension.)
This patch avoids optimizing branches and jumps to use RVC instructions
since surrounding code often makes assumptions about the sizes of
emitted instructions. Optimizing these will require changing these
functions (e.g., emit_branch) to dynamically compute jump offsets.
The following are examples of the JITed code for the verifier selftest
"direct packet read test#3 for CGROUP_SKB OK", without and with RVC
enabled, respectively. The former uses 178 bytes, and the latter uses 112,
for a ~37% reduction in code size for this example.
Without RVC:
0: 02000813 addi a6,zero,32
4: fd010113 addi sp,sp,-48
8: 02813423 sd s0,40(sp)
c: 02913023 sd s1,32(sp)
10: 01213c23 sd s2,24(sp)
14: 01313823 sd s3,16(sp)
18: 01413423 sd s4,8(sp)
1c: 03010413 addi s0,sp,48
20: 03056683 lwu a3,48(a0)
24: 02069693 slli a3,a3,0x20
28: 0206d693 srli a3,a3,0x20
2c: 03456703 lwu a4,52(a0)
30: 02071713 slli a4,a4,0x20
34: 02075713 srli a4,a4,0x20
38: 03856483 lwu s1,56(a0)
3c: 02049493 slli s1,s1,0x20
40: 0204d493 srli s1,s1,0x20
44: 03c56903 lwu s2,60(a0)
48: 02091913 slli s2,s2,0x20
4c: 02095913 srli s2,s2,0x20
50: 04056983 lwu s3,64(a0)
54: 02099993 slli s3,s3,0x20
58: 0209d993 srli s3,s3,0x20
5c: 09056a03 lwu s4,144(a0)
60: 020a1a13 slli s4,s4,0x20
64: 020a5a13 srli s4,s4,0x20
68: 00900313 addi t1,zero,9
6c: 006a7463 bgeu s4,t1,0x74
70: 00000a13 addi s4,zero,0
74: 02d52823 sw a3,48(a0)
78: 02e52a23 sw a4,52(a0)
7c: 02952c23 sw s1,56(a0)
80: 03252e23 sw s2,60(a0)
84: 05352023 sw s3,64(a0)
88: 00000793 addi a5,zero,0
8c: 02813403 ld s0,40(sp)
90: 02013483 ld s1,32(sp)
94: 01813903 ld s2,24(sp)
98: 01013983 ld s3,16(sp)
9c: 00813a03 ld s4,8(sp)
a0: 03010113 addi sp,sp,48
a4: 00078513 addi a0,a5,0
a8: 00008067 jalr zero,0(ra)
With RVC:
0: 02000813 addi a6,zero,32
4: 7179 c.addi16sp sp,-48
6: f422 c.sdsp s0,40(sp)
8: f026 c.sdsp s1,32(sp)
a: ec4a c.sdsp s2,24(sp)
c: e84e c.sdsp s3,16(sp)
e: e452 c.sdsp s4,8(sp)
10: 1800 c.addi4spn s0,sp,48
12: 03056683 lwu a3,48(a0)
16: 1682 c.slli a3,0x20
18: 9281 c.srli a3,0x20
1a: 03456703 lwu a4,52(a0)
1e: 1702 c.slli a4,0x20
20: 9301 c.srli a4,0x20
22: 03856483 lwu s1,56(a0)
26: 1482 c.slli s1,0x20
28: 9081 c.srli s1,0x20
2a: 03c56903 lwu s2,60(a0)
2e: 1902 c.slli s2,0x20
30: 02095913 srli s2,s2,0x20
34: 04056983 lwu s3,64(a0)
38: 1982 c.slli s3,0x20
3a: 0209d993 srli s3,s3,0x20
3e: 09056a03 lwu s4,144(a0)
42: 1a02 c.slli s4,0x20
44: 020a5a13 srli s4,s4,0x20
48: 4325 c.li t1,9
4a: 006a7363 bgeu s4,t1,0x50
4e: 4a01 c.li s4,0
50: d914 c.sw a3,48(a0)
52: d958 c.sw a4,52(a0)
54: dd04 c.sw s1,56(a0)
56: 03252e23 sw s2,60(a0)
5a: 05352023 sw s3,64(a0)
5e: 4781 c.li a5,0
60: 7422 c.ldsp s0,40(sp)
62: 7482 c.ldsp s1,32(sp)
64: 6962 c.ldsp s2,24(sp)
66: 69c2 c.ldsp s3,16(sp)
68: 6a22 c.ldsp s4,8(sp)
6a: 6145 c.addi16sp sp,48
6c: 853e c.mv a0,a5
6e: 8082 c.jr ra
Signed-off-by: Luke Nelson <luke.r.nels@gmail.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Cc: Björn Töpel <bjorn.topel@gmail.com>
Link: https://lore.kernel.org/bpf/20200721025241.8077-4-luke.r.nels@gmail.com
2020-07-20 19:52:40 -07:00
|
|
|
emit_sd(RV_REG_SP, store_offset, RV_REG_S4, ctx);
|
bpf, riscv: add BPF JIT for RV64G
This commit adds a BPF JIT for RV64G.
The JIT is a two-pass JIT, and has a dynamic prolog/epilogue (similar
to the MIPS64 BPF JIT) instead of static ones (e.g. x86_64).
At the moment the RISC-V Linux port does not support
CONFIG_HAVE_KPROBES, which means that CONFIG_BPF_EVENTS is not
supported. Thus, no tests involving BPF_PROG_TYPE_TRACEPOINT,
BPF_PROG_TYPE_PERF_EVENT, BPF_PROG_TYPE_KPROBE and
BPF_PROG_TYPE_RAW_TRACEPOINT passes.
The implementation does not support "far branching" (>4KiB).
Test results:
# modprobe test_bpf
test_bpf: Summary: 378 PASSED, 0 FAILED, [366/366 JIT'ed]
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier
...
Summary: 761 PASSED, 507 SKIPPED, 2 FAILED
Note that "test_verifier" was run with one build with
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y and one without, otherwise
many of the the tests that require unaligned access were skipped.
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
0
No CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
59
The two failing test_verifier tests are:
"ld_abs: vlan + abs, test 1"
"ld_abs: jump around ld_abs"
This is due to that "far branching" involved in those tests.
All tests where done on QEMU (QEMU emulator version 3.1.50
(v3.1.0-688-g8ae951fbc106)).
Signed-off-by: Björn Töpel <bjorn.topel@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2019-02-05 05:41:22 -07:00
|
|
|
store_offset -= 8;
|
|
|
|
}
|
|
|
|
if (seen_reg(RV_REG_S5, ctx)) {
|
bpf, riscv: Use compressed instructions in the rv64 JIT
This patch uses the RVC support and encodings from bpf_jit.h to optimize
the rv64 jit.
The optimizations work by replacing emit(rv_X(...)) with a call to a
helper function emit_X, which will emit a compressed version of the
instruction when possible, and when RVC is enabled.
The JIT continues to pass all tests in lib/test_bpf.c, and introduces
no new failures to test_verifier; both with and without RVC being enabled.
Most changes are straightforward replacements of emit(rv_X(...), ctx)
with emit_X(..., ctx), with the following exceptions bearing mention;
* Change emit_imm to sign-extend the value in "lower", since the
checks for RVC (and the instructions themselves) treat the value as
signed. Otherwise, small negative immediates will not be recognized as
encodable using an RVC instruction. For example, without this change,
emit_imm(rd, -1, ctx) would cause lower to become 4095, which is not a
6b int even though a "c.li rd, -1" instruction suffices.
* For {BPF_MOV,BPF_ADD} BPF_X, drop using addiw,addw in the 32-bit
cases since the values are zero-extended into the upper 32 bits in
the following instructions anyways, and the addition commutes with
zero-extension. (BPF_SUB BPF_X must still use subw since subtraction
does not commute with zero-extension.)
This patch avoids optimizing branches and jumps to use RVC instructions
since surrounding code often makes assumptions about the sizes of
emitted instructions. Optimizing these will require changing these
functions (e.g., emit_branch) to dynamically compute jump offsets.
The following are examples of the JITed code for the verifier selftest
"direct packet read test#3 for CGROUP_SKB OK", without and with RVC
enabled, respectively. The former uses 178 bytes, and the latter uses 112,
for a ~37% reduction in code size for this example.
Without RVC:
0: 02000813 addi a6,zero,32
4: fd010113 addi sp,sp,-48
8: 02813423 sd s0,40(sp)
c: 02913023 sd s1,32(sp)
10: 01213c23 sd s2,24(sp)
14: 01313823 sd s3,16(sp)
18: 01413423 sd s4,8(sp)
1c: 03010413 addi s0,sp,48
20: 03056683 lwu a3,48(a0)
24: 02069693 slli a3,a3,0x20
28: 0206d693 srli a3,a3,0x20
2c: 03456703 lwu a4,52(a0)
30: 02071713 slli a4,a4,0x20
34: 02075713 srli a4,a4,0x20
38: 03856483 lwu s1,56(a0)
3c: 02049493 slli s1,s1,0x20
40: 0204d493 srli s1,s1,0x20
44: 03c56903 lwu s2,60(a0)
48: 02091913 slli s2,s2,0x20
4c: 02095913 srli s2,s2,0x20
50: 04056983 lwu s3,64(a0)
54: 02099993 slli s3,s3,0x20
58: 0209d993 srli s3,s3,0x20
5c: 09056a03 lwu s4,144(a0)
60: 020a1a13 slli s4,s4,0x20
64: 020a5a13 srli s4,s4,0x20
68: 00900313 addi t1,zero,9
6c: 006a7463 bgeu s4,t1,0x74
70: 00000a13 addi s4,zero,0
74: 02d52823 sw a3,48(a0)
78: 02e52a23 sw a4,52(a0)
7c: 02952c23 sw s1,56(a0)
80: 03252e23 sw s2,60(a0)
84: 05352023 sw s3,64(a0)
88: 00000793 addi a5,zero,0
8c: 02813403 ld s0,40(sp)
90: 02013483 ld s1,32(sp)
94: 01813903 ld s2,24(sp)
98: 01013983 ld s3,16(sp)
9c: 00813a03 ld s4,8(sp)
a0: 03010113 addi sp,sp,48
a4: 00078513 addi a0,a5,0
a8: 00008067 jalr zero,0(ra)
With RVC:
0: 02000813 addi a6,zero,32
4: 7179 c.addi16sp sp,-48
6: f422 c.sdsp s0,40(sp)
8: f026 c.sdsp s1,32(sp)
a: ec4a c.sdsp s2,24(sp)
c: e84e c.sdsp s3,16(sp)
e: e452 c.sdsp s4,8(sp)
10: 1800 c.addi4spn s0,sp,48
12: 03056683 lwu a3,48(a0)
16: 1682 c.slli a3,0x20
18: 9281 c.srli a3,0x20
1a: 03456703 lwu a4,52(a0)
1e: 1702 c.slli a4,0x20
20: 9301 c.srli a4,0x20
22: 03856483 lwu s1,56(a0)
26: 1482 c.slli s1,0x20
28: 9081 c.srli s1,0x20
2a: 03c56903 lwu s2,60(a0)
2e: 1902 c.slli s2,0x20
30: 02095913 srli s2,s2,0x20
34: 04056983 lwu s3,64(a0)
38: 1982 c.slli s3,0x20
3a: 0209d993 srli s3,s3,0x20
3e: 09056a03 lwu s4,144(a0)
42: 1a02 c.slli s4,0x20
44: 020a5a13 srli s4,s4,0x20
48: 4325 c.li t1,9
4a: 006a7363 bgeu s4,t1,0x50
4e: 4a01 c.li s4,0
50: d914 c.sw a3,48(a0)
52: d958 c.sw a4,52(a0)
54: dd04 c.sw s1,56(a0)
56: 03252e23 sw s2,60(a0)
5a: 05352023 sw s3,64(a0)
5e: 4781 c.li a5,0
60: 7422 c.ldsp s0,40(sp)
62: 7482 c.ldsp s1,32(sp)
64: 6962 c.ldsp s2,24(sp)
66: 69c2 c.ldsp s3,16(sp)
68: 6a22 c.ldsp s4,8(sp)
6a: 6145 c.addi16sp sp,48
6c: 853e c.mv a0,a5
6e: 8082 c.jr ra
Signed-off-by: Luke Nelson <luke.r.nels@gmail.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Cc: Björn Töpel <bjorn.topel@gmail.com>
Link: https://lore.kernel.org/bpf/20200721025241.8077-4-luke.r.nels@gmail.com
2020-07-20 19:52:40 -07:00
|
|
|
emit_sd(RV_REG_SP, store_offset, RV_REG_S5, ctx);
|
bpf, riscv: add BPF JIT for RV64G
This commit adds a BPF JIT for RV64G.
The JIT is a two-pass JIT, and has a dynamic prolog/epilogue (similar
to the MIPS64 BPF JIT) instead of static ones (e.g. x86_64).
At the moment the RISC-V Linux port does not support
CONFIG_HAVE_KPROBES, which means that CONFIG_BPF_EVENTS is not
supported. Thus, no tests involving BPF_PROG_TYPE_TRACEPOINT,
BPF_PROG_TYPE_PERF_EVENT, BPF_PROG_TYPE_KPROBE and
BPF_PROG_TYPE_RAW_TRACEPOINT passes.
The implementation does not support "far branching" (>4KiB).
Test results:
# modprobe test_bpf
test_bpf: Summary: 378 PASSED, 0 FAILED, [366/366 JIT'ed]
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier
...
Summary: 761 PASSED, 507 SKIPPED, 2 FAILED
Note that "test_verifier" was run with one build with
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y and one without, otherwise
many of the the tests that require unaligned access were skipped.
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
0
No CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
59
The two failing test_verifier tests are:
"ld_abs: vlan + abs, test 1"
"ld_abs: jump around ld_abs"
This is due to that "far branching" involved in those tests.
All tests where done on QEMU (QEMU emulator version 3.1.50
(v3.1.0-688-g8ae951fbc106)).
Signed-off-by: Björn Töpel <bjorn.topel@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2019-02-05 05:41:22 -07:00
|
|
|
store_offset -= 8;
|
|
|
|
}
|
|
|
|
if (seen_reg(RV_REG_S6, ctx)) {
|
bpf, riscv: Use compressed instructions in the rv64 JIT
This patch uses the RVC support and encodings from bpf_jit.h to optimize
the rv64 jit.
The optimizations work by replacing emit(rv_X(...)) with a call to a
helper function emit_X, which will emit a compressed version of the
instruction when possible, and when RVC is enabled.
The JIT continues to pass all tests in lib/test_bpf.c, and introduces
no new failures to test_verifier; both with and without RVC being enabled.
Most changes are straightforward replacements of emit(rv_X(...), ctx)
with emit_X(..., ctx), with the following exceptions bearing mention;
* Change emit_imm to sign-extend the value in "lower", since the
checks for RVC (and the instructions themselves) treat the value as
signed. Otherwise, small negative immediates will not be recognized as
encodable using an RVC instruction. For example, without this change,
emit_imm(rd, -1, ctx) would cause lower to become 4095, which is not a
6b int even though a "c.li rd, -1" instruction suffices.
* For {BPF_MOV,BPF_ADD} BPF_X, drop using addiw,addw in the 32-bit
cases since the values are zero-extended into the upper 32 bits in
the following instructions anyways, and the addition commutes with
zero-extension. (BPF_SUB BPF_X must still use subw since subtraction
does not commute with zero-extension.)
This patch avoids optimizing branches and jumps to use RVC instructions
since surrounding code often makes assumptions about the sizes of
emitted instructions. Optimizing these will require changing these
functions (e.g., emit_branch) to dynamically compute jump offsets.
The following are examples of the JITed code for the verifier selftest
"direct packet read test#3 for CGROUP_SKB OK", without and with RVC
enabled, respectively. The former uses 178 bytes, and the latter uses 112,
for a ~37% reduction in code size for this example.
Without RVC:
0: 02000813 addi a6,zero,32
4: fd010113 addi sp,sp,-48
8: 02813423 sd s0,40(sp)
c: 02913023 sd s1,32(sp)
10: 01213c23 sd s2,24(sp)
14: 01313823 sd s3,16(sp)
18: 01413423 sd s4,8(sp)
1c: 03010413 addi s0,sp,48
20: 03056683 lwu a3,48(a0)
24: 02069693 slli a3,a3,0x20
28: 0206d693 srli a3,a3,0x20
2c: 03456703 lwu a4,52(a0)
30: 02071713 slli a4,a4,0x20
34: 02075713 srli a4,a4,0x20
38: 03856483 lwu s1,56(a0)
3c: 02049493 slli s1,s1,0x20
40: 0204d493 srli s1,s1,0x20
44: 03c56903 lwu s2,60(a0)
48: 02091913 slli s2,s2,0x20
4c: 02095913 srli s2,s2,0x20
50: 04056983 lwu s3,64(a0)
54: 02099993 slli s3,s3,0x20
58: 0209d993 srli s3,s3,0x20
5c: 09056a03 lwu s4,144(a0)
60: 020a1a13 slli s4,s4,0x20
64: 020a5a13 srli s4,s4,0x20
68: 00900313 addi t1,zero,9
6c: 006a7463 bgeu s4,t1,0x74
70: 00000a13 addi s4,zero,0
74: 02d52823 sw a3,48(a0)
78: 02e52a23 sw a4,52(a0)
7c: 02952c23 sw s1,56(a0)
80: 03252e23 sw s2,60(a0)
84: 05352023 sw s3,64(a0)
88: 00000793 addi a5,zero,0
8c: 02813403 ld s0,40(sp)
90: 02013483 ld s1,32(sp)
94: 01813903 ld s2,24(sp)
98: 01013983 ld s3,16(sp)
9c: 00813a03 ld s4,8(sp)
a0: 03010113 addi sp,sp,48
a4: 00078513 addi a0,a5,0
a8: 00008067 jalr zero,0(ra)
With RVC:
0: 02000813 addi a6,zero,32
4: 7179 c.addi16sp sp,-48
6: f422 c.sdsp s0,40(sp)
8: f026 c.sdsp s1,32(sp)
a: ec4a c.sdsp s2,24(sp)
c: e84e c.sdsp s3,16(sp)
e: e452 c.sdsp s4,8(sp)
10: 1800 c.addi4spn s0,sp,48
12: 03056683 lwu a3,48(a0)
16: 1682 c.slli a3,0x20
18: 9281 c.srli a3,0x20
1a: 03456703 lwu a4,52(a0)
1e: 1702 c.slli a4,0x20
20: 9301 c.srli a4,0x20
22: 03856483 lwu s1,56(a0)
26: 1482 c.slli s1,0x20
28: 9081 c.srli s1,0x20
2a: 03c56903 lwu s2,60(a0)
2e: 1902 c.slli s2,0x20
30: 02095913 srli s2,s2,0x20
34: 04056983 lwu s3,64(a0)
38: 1982 c.slli s3,0x20
3a: 0209d993 srli s3,s3,0x20
3e: 09056a03 lwu s4,144(a0)
42: 1a02 c.slli s4,0x20
44: 020a5a13 srli s4,s4,0x20
48: 4325 c.li t1,9
4a: 006a7363 bgeu s4,t1,0x50
4e: 4a01 c.li s4,0
50: d914 c.sw a3,48(a0)
52: d958 c.sw a4,52(a0)
54: dd04 c.sw s1,56(a0)
56: 03252e23 sw s2,60(a0)
5a: 05352023 sw s3,64(a0)
5e: 4781 c.li a5,0
60: 7422 c.ldsp s0,40(sp)
62: 7482 c.ldsp s1,32(sp)
64: 6962 c.ldsp s2,24(sp)
66: 69c2 c.ldsp s3,16(sp)
68: 6a22 c.ldsp s4,8(sp)
6a: 6145 c.addi16sp sp,48
6c: 853e c.mv a0,a5
6e: 8082 c.jr ra
Signed-off-by: Luke Nelson <luke.r.nels@gmail.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Cc: Björn Töpel <bjorn.topel@gmail.com>
Link: https://lore.kernel.org/bpf/20200721025241.8077-4-luke.r.nels@gmail.com
2020-07-20 19:52:40 -07:00
|
|
|
emit_sd(RV_REG_SP, store_offset, RV_REG_S6, ctx);
|
bpf, riscv: add BPF JIT for RV64G
This commit adds a BPF JIT for RV64G.
The JIT is a two-pass JIT, and has a dynamic prolog/epilogue (similar
to the MIPS64 BPF JIT) instead of static ones (e.g. x86_64).
At the moment the RISC-V Linux port does not support
CONFIG_HAVE_KPROBES, which means that CONFIG_BPF_EVENTS is not
supported. Thus, no tests involving BPF_PROG_TYPE_TRACEPOINT,
BPF_PROG_TYPE_PERF_EVENT, BPF_PROG_TYPE_KPROBE and
BPF_PROG_TYPE_RAW_TRACEPOINT passes.
The implementation does not support "far branching" (>4KiB).
Test results:
# modprobe test_bpf
test_bpf: Summary: 378 PASSED, 0 FAILED, [366/366 JIT'ed]
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier
...
Summary: 761 PASSED, 507 SKIPPED, 2 FAILED
Note that "test_verifier" was run with one build with
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y and one without, otherwise
many of the the tests that require unaligned access were skipped.
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
0
No CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
59
The two failing test_verifier tests are:
"ld_abs: vlan + abs, test 1"
"ld_abs: jump around ld_abs"
This is due to that "far branching" involved in those tests.
All tests where done on QEMU (QEMU emulator version 3.1.50
(v3.1.0-688-g8ae951fbc106)).
Signed-off-by: Björn Töpel <bjorn.topel@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2019-02-05 05:41:22 -07:00
|
|
|
store_offset -= 8;
|
|
|
|
}
|
2024-04-04 04:42:02 -07:00
|
|
|
if (ctx->arena_vm_start) {
|
|
|
|
emit_sd(RV_REG_SP, store_offset, RV_REG_ARENA, ctx);
|
|
|
|
store_offset -= 8;
|
|
|
|
}
|
bpf, riscv: add BPF JIT for RV64G
This commit adds a BPF JIT for RV64G.
The JIT is a two-pass JIT, and has a dynamic prolog/epilogue (similar
to the MIPS64 BPF JIT) instead of static ones (e.g. x86_64).
At the moment the RISC-V Linux port does not support
CONFIG_HAVE_KPROBES, which means that CONFIG_BPF_EVENTS is not
supported. Thus, no tests involving BPF_PROG_TYPE_TRACEPOINT,
BPF_PROG_TYPE_PERF_EVENT, BPF_PROG_TYPE_KPROBE and
BPF_PROG_TYPE_RAW_TRACEPOINT passes.
The implementation does not support "far branching" (>4KiB).
Test results:
# modprobe test_bpf
test_bpf: Summary: 378 PASSED, 0 FAILED, [366/366 JIT'ed]
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier
...
Summary: 761 PASSED, 507 SKIPPED, 2 FAILED
Note that "test_verifier" was run with one build with
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y and one without, otherwise
many of the the tests that require unaligned access were skipped.
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
0
No CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
59
The two failing test_verifier tests are:
"ld_abs: vlan + abs, test 1"
"ld_abs: jump around ld_abs"
This is due to that "far branching" involved in those tests.
All tests where done on QEMU (QEMU emulator version 3.1.50
(v3.1.0-688-g8ae951fbc106)).
Signed-off-by: Björn Töpel <bjorn.topel@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2019-02-05 05:41:22 -07:00
|
|
|
|
bpf, riscv: Use compressed instructions in the rv64 JIT
This patch uses the RVC support and encodings from bpf_jit.h to optimize
the rv64 jit.
The optimizations work by replacing emit(rv_X(...)) with a call to a
helper function emit_X, which will emit a compressed version of the
instruction when possible, and when RVC is enabled.
The JIT continues to pass all tests in lib/test_bpf.c, and introduces
no new failures to test_verifier; both with and without RVC being enabled.
Most changes are straightforward replacements of emit(rv_X(...), ctx)
with emit_X(..., ctx), with the following exceptions bearing mention;
* Change emit_imm to sign-extend the value in "lower", since the
checks for RVC (and the instructions themselves) treat the value as
signed. Otherwise, small negative immediates will not be recognized as
encodable using an RVC instruction. For example, without this change,
emit_imm(rd, -1, ctx) would cause lower to become 4095, which is not a
6b int even though a "c.li rd, -1" instruction suffices.
* For {BPF_MOV,BPF_ADD} BPF_X, drop using addiw,addw in the 32-bit
cases since the values are zero-extended into the upper 32 bits in
the following instructions anyways, and the addition commutes with
zero-extension. (BPF_SUB BPF_X must still use subw since subtraction
does not commute with zero-extension.)
This patch avoids optimizing branches and jumps to use RVC instructions
since surrounding code often makes assumptions about the sizes of
emitted instructions. Optimizing these will require changing these
functions (e.g., emit_branch) to dynamically compute jump offsets.
The following are examples of the JITed code for the verifier selftest
"direct packet read test#3 for CGROUP_SKB OK", without and with RVC
enabled, respectively. The former uses 178 bytes, and the latter uses 112,
for a ~37% reduction in code size for this example.
Without RVC:
0: 02000813 addi a6,zero,32
4: fd010113 addi sp,sp,-48
8: 02813423 sd s0,40(sp)
c: 02913023 sd s1,32(sp)
10: 01213c23 sd s2,24(sp)
14: 01313823 sd s3,16(sp)
18: 01413423 sd s4,8(sp)
1c: 03010413 addi s0,sp,48
20: 03056683 lwu a3,48(a0)
24: 02069693 slli a3,a3,0x20
28: 0206d693 srli a3,a3,0x20
2c: 03456703 lwu a4,52(a0)
30: 02071713 slli a4,a4,0x20
34: 02075713 srli a4,a4,0x20
38: 03856483 lwu s1,56(a0)
3c: 02049493 slli s1,s1,0x20
40: 0204d493 srli s1,s1,0x20
44: 03c56903 lwu s2,60(a0)
48: 02091913 slli s2,s2,0x20
4c: 02095913 srli s2,s2,0x20
50: 04056983 lwu s3,64(a0)
54: 02099993 slli s3,s3,0x20
58: 0209d993 srli s3,s3,0x20
5c: 09056a03 lwu s4,144(a0)
60: 020a1a13 slli s4,s4,0x20
64: 020a5a13 srli s4,s4,0x20
68: 00900313 addi t1,zero,9
6c: 006a7463 bgeu s4,t1,0x74
70: 00000a13 addi s4,zero,0
74: 02d52823 sw a3,48(a0)
78: 02e52a23 sw a4,52(a0)
7c: 02952c23 sw s1,56(a0)
80: 03252e23 sw s2,60(a0)
84: 05352023 sw s3,64(a0)
88: 00000793 addi a5,zero,0
8c: 02813403 ld s0,40(sp)
90: 02013483 ld s1,32(sp)
94: 01813903 ld s2,24(sp)
98: 01013983 ld s3,16(sp)
9c: 00813a03 ld s4,8(sp)
a0: 03010113 addi sp,sp,48
a4: 00078513 addi a0,a5,0
a8: 00008067 jalr zero,0(ra)
With RVC:
0: 02000813 addi a6,zero,32
4: 7179 c.addi16sp sp,-48
6: f422 c.sdsp s0,40(sp)
8: f026 c.sdsp s1,32(sp)
a: ec4a c.sdsp s2,24(sp)
c: e84e c.sdsp s3,16(sp)
e: e452 c.sdsp s4,8(sp)
10: 1800 c.addi4spn s0,sp,48
12: 03056683 lwu a3,48(a0)
16: 1682 c.slli a3,0x20
18: 9281 c.srli a3,0x20
1a: 03456703 lwu a4,52(a0)
1e: 1702 c.slli a4,0x20
20: 9301 c.srli a4,0x20
22: 03856483 lwu s1,56(a0)
26: 1482 c.slli s1,0x20
28: 9081 c.srli s1,0x20
2a: 03c56903 lwu s2,60(a0)
2e: 1902 c.slli s2,0x20
30: 02095913 srli s2,s2,0x20
34: 04056983 lwu s3,64(a0)
38: 1982 c.slli s3,0x20
3a: 0209d993 srli s3,s3,0x20
3e: 09056a03 lwu s4,144(a0)
42: 1a02 c.slli s4,0x20
44: 020a5a13 srli s4,s4,0x20
48: 4325 c.li t1,9
4a: 006a7363 bgeu s4,t1,0x50
4e: 4a01 c.li s4,0
50: d914 c.sw a3,48(a0)
52: d958 c.sw a4,52(a0)
54: dd04 c.sw s1,56(a0)
56: 03252e23 sw s2,60(a0)
5a: 05352023 sw s3,64(a0)
5e: 4781 c.li a5,0
60: 7422 c.ldsp s0,40(sp)
62: 7482 c.ldsp s1,32(sp)
64: 6962 c.ldsp s2,24(sp)
66: 69c2 c.ldsp s3,16(sp)
68: 6a22 c.ldsp s4,8(sp)
6a: 6145 c.addi16sp sp,48
6c: 853e c.mv a0,a5
6e: 8082 c.jr ra
Signed-off-by: Luke Nelson <luke.r.nels@gmail.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Cc: Björn Töpel <bjorn.topel@gmail.com>
Link: https://lore.kernel.org/bpf/20200721025241.8077-4-luke.r.nels@gmail.com
2020-07-20 19:52:40 -07:00
|
|
|
emit_addi(RV_REG_FP, RV_REG_SP, stack_adjust, ctx);
|
bpf, riscv: add BPF JIT for RV64G
This commit adds a BPF JIT for RV64G.
The JIT is a two-pass JIT, and has a dynamic prolog/epilogue (similar
to the MIPS64 BPF JIT) instead of static ones (e.g. x86_64).
At the moment the RISC-V Linux port does not support
CONFIG_HAVE_KPROBES, which means that CONFIG_BPF_EVENTS is not
supported. Thus, no tests involving BPF_PROG_TYPE_TRACEPOINT,
BPF_PROG_TYPE_PERF_EVENT, BPF_PROG_TYPE_KPROBE and
BPF_PROG_TYPE_RAW_TRACEPOINT passes.
The implementation does not support "far branching" (>4KiB).
Test results:
# modprobe test_bpf
test_bpf: Summary: 378 PASSED, 0 FAILED, [366/366 JIT'ed]
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier
...
Summary: 761 PASSED, 507 SKIPPED, 2 FAILED
Note that "test_verifier" was run with one build with
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y and one without, otherwise
many of the the tests that require unaligned access were skipped.
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
0
No CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
59
The two failing test_verifier tests are:
"ld_abs: vlan + abs, test 1"
"ld_abs: jump around ld_abs"
This is due to that "far branching" involved in those tests.
All tests where done on QEMU (QEMU emulator version 3.1.50
(v3.1.0-688-g8ae951fbc106)).
Signed-off-by: Björn Töpel <bjorn.topel@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2019-02-05 05:41:22 -07:00
|
|
|
|
|
|
|
if (bpf_stack_adjust)
|
bpf, riscv: Use compressed instructions in the rv64 JIT
This patch uses the RVC support and encodings from bpf_jit.h to optimize
the rv64 jit.
The optimizations work by replacing emit(rv_X(...)) with a call to a
helper function emit_X, which will emit a compressed version of the
instruction when possible, and when RVC is enabled.
The JIT continues to pass all tests in lib/test_bpf.c, and introduces
no new failures to test_verifier; both with and without RVC being enabled.
Most changes are straightforward replacements of emit(rv_X(...), ctx)
with emit_X(..., ctx), with the following exceptions bearing mention;
* Change emit_imm to sign-extend the value in "lower", since the
checks for RVC (and the instructions themselves) treat the value as
signed. Otherwise, small negative immediates will not be recognized as
encodable using an RVC instruction. For example, without this change,
emit_imm(rd, -1, ctx) would cause lower to become 4095, which is not a
6b int even though a "c.li rd, -1" instruction suffices.
* For {BPF_MOV,BPF_ADD} BPF_X, drop using addiw,addw in the 32-bit
cases since the values are zero-extended into the upper 32 bits in
the following instructions anyways, and the addition commutes with
zero-extension. (BPF_SUB BPF_X must still use subw since subtraction
does not commute with zero-extension.)
This patch avoids optimizing branches and jumps to use RVC instructions
since surrounding code often makes assumptions about the sizes of
emitted instructions. Optimizing these will require changing these
functions (e.g., emit_branch) to dynamically compute jump offsets.
The following are examples of the JITed code for the verifier selftest
"direct packet read test#3 for CGROUP_SKB OK", without and with RVC
enabled, respectively. The former uses 178 bytes, and the latter uses 112,
for a ~37% reduction in code size for this example.
Without RVC:
0: 02000813 addi a6,zero,32
4: fd010113 addi sp,sp,-48
8: 02813423 sd s0,40(sp)
c: 02913023 sd s1,32(sp)
10: 01213c23 sd s2,24(sp)
14: 01313823 sd s3,16(sp)
18: 01413423 sd s4,8(sp)
1c: 03010413 addi s0,sp,48
20: 03056683 lwu a3,48(a0)
24: 02069693 slli a3,a3,0x20
28: 0206d693 srli a3,a3,0x20
2c: 03456703 lwu a4,52(a0)
30: 02071713 slli a4,a4,0x20
34: 02075713 srli a4,a4,0x20
38: 03856483 lwu s1,56(a0)
3c: 02049493 slli s1,s1,0x20
40: 0204d493 srli s1,s1,0x20
44: 03c56903 lwu s2,60(a0)
48: 02091913 slli s2,s2,0x20
4c: 02095913 srli s2,s2,0x20
50: 04056983 lwu s3,64(a0)
54: 02099993 slli s3,s3,0x20
58: 0209d993 srli s3,s3,0x20
5c: 09056a03 lwu s4,144(a0)
60: 020a1a13 slli s4,s4,0x20
64: 020a5a13 srli s4,s4,0x20
68: 00900313 addi t1,zero,9
6c: 006a7463 bgeu s4,t1,0x74
70: 00000a13 addi s4,zero,0
74: 02d52823 sw a3,48(a0)
78: 02e52a23 sw a4,52(a0)
7c: 02952c23 sw s1,56(a0)
80: 03252e23 sw s2,60(a0)
84: 05352023 sw s3,64(a0)
88: 00000793 addi a5,zero,0
8c: 02813403 ld s0,40(sp)
90: 02013483 ld s1,32(sp)
94: 01813903 ld s2,24(sp)
98: 01013983 ld s3,16(sp)
9c: 00813a03 ld s4,8(sp)
a0: 03010113 addi sp,sp,48
a4: 00078513 addi a0,a5,0
a8: 00008067 jalr zero,0(ra)
With RVC:
0: 02000813 addi a6,zero,32
4: 7179 c.addi16sp sp,-48
6: f422 c.sdsp s0,40(sp)
8: f026 c.sdsp s1,32(sp)
a: ec4a c.sdsp s2,24(sp)
c: e84e c.sdsp s3,16(sp)
e: e452 c.sdsp s4,8(sp)
10: 1800 c.addi4spn s0,sp,48
12: 03056683 lwu a3,48(a0)
16: 1682 c.slli a3,0x20
18: 9281 c.srli a3,0x20
1a: 03456703 lwu a4,52(a0)
1e: 1702 c.slli a4,0x20
20: 9301 c.srli a4,0x20
22: 03856483 lwu s1,56(a0)
26: 1482 c.slli s1,0x20
28: 9081 c.srli s1,0x20
2a: 03c56903 lwu s2,60(a0)
2e: 1902 c.slli s2,0x20
30: 02095913 srli s2,s2,0x20
34: 04056983 lwu s3,64(a0)
38: 1982 c.slli s3,0x20
3a: 0209d993 srli s3,s3,0x20
3e: 09056a03 lwu s4,144(a0)
42: 1a02 c.slli s4,0x20
44: 020a5a13 srli s4,s4,0x20
48: 4325 c.li t1,9
4a: 006a7363 bgeu s4,t1,0x50
4e: 4a01 c.li s4,0
50: d914 c.sw a3,48(a0)
52: d958 c.sw a4,52(a0)
54: dd04 c.sw s1,56(a0)
56: 03252e23 sw s2,60(a0)
5a: 05352023 sw s3,64(a0)
5e: 4781 c.li a5,0
60: 7422 c.ldsp s0,40(sp)
62: 7482 c.ldsp s1,32(sp)
64: 6962 c.ldsp s2,24(sp)
66: 69c2 c.ldsp s3,16(sp)
68: 6a22 c.ldsp s4,8(sp)
6a: 6145 c.addi16sp sp,48
6c: 853e c.mv a0,a5
6e: 8082 c.jr ra
Signed-off-by: Luke Nelson <luke.r.nels@gmail.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Cc: Björn Töpel <bjorn.topel@gmail.com>
Link: https://lore.kernel.org/bpf/20200721025241.8077-4-luke.r.nels@gmail.com
2020-07-20 19:52:40 -07:00
|
|
|
emit_addi(RV_REG_S5, RV_REG_SP, bpf_stack_adjust, ctx);
|
bpf, riscv: add BPF JIT for RV64G
This commit adds a BPF JIT for RV64G.
The JIT is a two-pass JIT, and has a dynamic prolog/epilogue (similar
to the MIPS64 BPF JIT) instead of static ones (e.g. x86_64).
At the moment the RISC-V Linux port does not support
CONFIG_HAVE_KPROBES, which means that CONFIG_BPF_EVENTS is not
supported. Thus, no tests involving BPF_PROG_TYPE_TRACEPOINT,
BPF_PROG_TYPE_PERF_EVENT, BPF_PROG_TYPE_KPROBE and
BPF_PROG_TYPE_RAW_TRACEPOINT passes.
The implementation does not support "far branching" (>4KiB).
Test results:
# modprobe test_bpf
test_bpf: Summary: 378 PASSED, 0 FAILED, [366/366 JIT'ed]
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier
...
Summary: 761 PASSED, 507 SKIPPED, 2 FAILED
Note that "test_verifier" was run with one build with
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y and one without, otherwise
many of the the tests that require unaligned access were skipped.
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
0
No CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
59
The two failing test_verifier tests are:
"ld_abs: vlan + abs, test 1"
"ld_abs: jump around ld_abs"
This is due to that "far branching" involved in those tests.
All tests where done on QEMU (QEMU emulator version 3.1.50
(v3.1.0-688-g8ae951fbc106)).
Signed-off-by: Björn Töpel <bjorn.topel@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2019-02-05 05:41:22 -07:00
|
|
|
|
|
|
|
/* Program contains calls and tail calls, so RV_REG_TCC need
|
|
|
|
* to be saved across calls.
|
|
|
|
*/
|
|
|
|
if (seen_tail_call(ctx) && seen_call(ctx))
|
bpf, riscv: Use compressed instructions in the rv64 JIT
This patch uses the RVC support and encodings from bpf_jit.h to optimize
the rv64 jit.
The optimizations work by replacing emit(rv_X(...)) with a call to a
helper function emit_X, which will emit a compressed version of the
instruction when possible, and when RVC is enabled.
The JIT continues to pass all tests in lib/test_bpf.c, and introduces
no new failures to test_verifier; both with and without RVC being enabled.
Most changes are straightforward replacements of emit(rv_X(...), ctx)
with emit_X(..., ctx), with the following exceptions bearing mention;
* Change emit_imm to sign-extend the value in "lower", since the
checks for RVC (and the instructions themselves) treat the value as
signed. Otherwise, small negative immediates will not be recognized as
encodable using an RVC instruction. For example, without this change,
emit_imm(rd, -1, ctx) would cause lower to become 4095, which is not a
6b int even though a "c.li rd, -1" instruction suffices.
* For {BPF_MOV,BPF_ADD} BPF_X, drop using addiw,addw in the 32-bit
cases since the values are zero-extended into the upper 32 bits in
the following instructions anyways, and the addition commutes with
zero-extension. (BPF_SUB BPF_X must still use subw since subtraction
does not commute with zero-extension.)
This patch avoids optimizing branches and jumps to use RVC instructions
since surrounding code often makes assumptions about the sizes of
emitted instructions. Optimizing these will require changing these
functions (e.g., emit_branch) to dynamically compute jump offsets.
The following are examples of the JITed code for the verifier selftest
"direct packet read test#3 for CGROUP_SKB OK", without and with RVC
enabled, respectively. The former uses 178 bytes, and the latter uses 112,
for a ~37% reduction in code size for this example.
Without RVC:
0: 02000813 addi a6,zero,32
4: fd010113 addi sp,sp,-48
8: 02813423 sd s0,40(sp)
c: 02913023 sd s1,32(sp)
10: 01213c23 sd s2,24(sp)
14: 01313823 sd s3,16(sp)
18: 01413423 sd s4,8(sp)
1c: 03010413 addi s0,sp,48
20: 03056683 lwu a3,48(a0)
24: 02069693 slli a3,a3,0x20
28: 0206d693 srli a3,a3,0x20
2c: 03456703 lwu a4,52(a0)
30: 02071713 slli a4,a4,0x20
34: 02075713 srli a4,a4,0x20
38: 03856483 lwu s1,56(a0)
3c: 02049493 slli s1,s1,0x20
40: 0204d493 srli s1,s1,0x20
44: 03c56903 lwu s2,60(a0)
48: 02091913 slli s2,s2,0x20
4c: 02095913 srli s2,s2,0x20
50: 04056983 lwu s3,64(a0)
54: 02099993 slli s3,s3,0x20
58: 0209d993 srli s3,s3,0x20
5c: 09056a03 lwu s4,144(a0)
60: 020a1a13 slli s4,s4,0x20
64: 020a5a13 srli s4,s4,0x20
68: 00900313 addi t1,zero,9
6c: 006a7463 bgeu s4,t1,0x74
70: 00000a13 addi s4,zero,0
74: 02d52823 sw a3,48(a0)
78: 02e52a23 sw a4,52(a0)
7c: 02952c23 sw s1,56(a0)
80: 03252e23 sw s2,60(a0)
84: 05352023 sw s3,64(a0)
88: 00000793 addi a5,zero,0
8c: 02813403 ld s0,40(sp)
90: 02013483 ld s1,32(sp)
94: 01813903 ld s2,24(sp)
98: 01013983 ld s3,16(sp)
9c: 00813a03 ld s4,8(sp)
a0: 03010113 addi sp,sp,48
a4: 00078513 addi a0,a5,0
a8: 00008067 jalr zero,0(ra)
With RVC:
0: 02000813 addi a6,zero,32
4: 7179 c.addi16sp sp,-48
6: f422 c.sdsp s0,40(sp)
8: f026 c.sdsp s1,32(sp)
a: ec4a c.sdsp s2,24(sp)
c: e84e c.sdsp s3,16(sp)
e: e452 c.sdsp s4,8(sp)
10: 1800 c.addi4spn s0,sp,48
12: 03056683 lwu a3,48(a0)
16: 1682 c.slli a3,0x20
18: 9281 c.srli a3,0x20
1a: 03456703 lwu a4,52(a0)
1e: 1702 c.slli a4,0x20
20: 9301 c.srli a4,0x20
22: 03856483 lwu s1,56(a0)
26: 1482 c.slli s1,0x20
28: 9081 c.srli s1,0x20
2a: 03c56903 lwu s2,60(a0)
2e: 1902 c.slli s2,0x20
30: 02095913 srli s2,s2,0x20
34: 04056983 lwu s3,64(a0)
38: 1982 c.slli s3,0x20
3a: 0209d993 srli s3,s3,0x20
3e: 09056a03 lwu s4,144(a0)
42: 1a02 c.slli s4,0x20
44: 020a5a13 srli s4,s4,0x20
48: 4325 c.li t1,9
4a: 006a7363 bgeu s4,t1,0x50
4e: 4a01 c.li s4,0
50: d914 c.sw a3,48(a0)
52: d958 c.sw a4,52(a0)
54: dd04 c.sw s1,56(a0)
56: 03252e23 sw s2,60(a0)
5a: 05352023 sw s3,64(a0)
5e: 4781 c.li a5,0
60: 7422 c.ldsp s0,40(sp)
62: 7482 c.ldsp s1,32(sp)
64: 6962 c.ldsp s2,24(sp)
66: 69c2 c.ldsp s3,16(sp)
68: 6a22 c.ldsp s4,8(sp)
6a: 6145 c.addi16sp sp,48
6c: 853e c.mv a0,a5
6e: 8082 c.jr ra
Signed-off-by: Luke Nelson <luke.r.nels@gmail.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Cc: Björn Töpel <bjorn.topel@gmail.com>
Link: https://lore.kernel.org/bpf/20200721025241.8077-4-luke.r.nels@gmail.com
2020-07-20 19:52:40 -07:00
|
|
|
emit_mv(RV_REG_TCC_SAVED, RV_REG_TCC, ctx);
|
bpf, riscv: add BPF JIT for RV64G
This commit adds a BPF JIT for RV64G.
The JIT is a two-pass JIT, and has a dynamic prolog/epilogue (similar
to the MIPS64 BPF JIT) instead of static ones (e.g. x86_64).
At the moment the RISC-V Linux port does not support
CONFIG_HAVE_KPROBES, which means that CONFIG_BPF_EVENTS is not
supported. Thus, no tests involving BPF_PROG_TYPE_TRACEPOINT,
BPF_PROG_TYPE_PERF_EVENT, BPF_PROG_TYPE_KPROBE and
BPF_PROG_TYPE_RAW_TRACEPOINT passes.
The implementation does not support "far branching" (>4KiB).
Test results:
# modprobe test_bpf
test_bpf: Summary: 378 PASSED, 0 FAILED, [366/366 JIT'ed]
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier
...
Summary: 761 PASSED, 507 SKIPPED, 2 FAILED
Note that "test_verifier" was run with one build with
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y and one without, otherwise
many of the the tests that require unaligned access were skipped.
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
0
No CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
59
The two failing test_verifier tests are:
"ld_abs: vlan + abs, test 1"
"ld_abs: jump around ld_abs"
This is due to that "far branching" involved in those tests.
All tests where done on QEMU (QEMU emulator version 3.1.50
(v3.1.0-688-g8ae951fbc106)).
Signed-off-by: Björn Töpel <bjorn.topel@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2019-02-05 05:41:22 -07:00
|
|
|
|
|
|
|
ctx->stack_size = stack_adjust;
|
2024-04-04 04:42:02 -07:00
|
|
|
|
|
|
|
if (ctx->arena_vm_start)
|
|
|
|
emit_imm(RV_REG_ARENA, ctx->arena_vm_start, ctx);
|
bpf, riscv: add BPF JIT for RV64G
This commit adds a BPF JIT for RV64G.
The JIT is a two-pass JIT, and has a dynamic prolog/epilogue (similar
to the MIPS64 BPF JIT) instead of static ones (e.g. x86_64).
At the moment the RISC-V Linux port does not support
CONFIG_HAVE_KPROBES, which means that CONFIG_BPF_EVENTS is not
supported. Thus, no tests involving BPF_PROG_TYPE_TRACEPOINT,
BPF_PROG_TYPE_PERF_EVENT, BPF_PROG_TYPE_KPROBE and
BPF_PROG_TYPE_RAW_TRACEPOINT passes.
The implementation does not support "far branching" (>4KiB).
Test results:
# modprobe test_bpf
test_bpf: Summary: 378 PASSED, 0 FAILED, [366/366 JIT'ed]
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier
...
Summary: 761 PASSED, 507 SKIPPED, 2 FAILED
Note that "test_verifier" was run with one build with
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y and one without, otherwise
many of the the tests that require unaligned access were skipped.
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
0
No CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
59
The two failing test_verifier tests are:
"ld_abs: vlan + abs, test 1"
"ld_abs: jump around ld_abs"
This is due to that "far branching" involved in those tests.
All tests where done on QEMU (QEMU emulator version 3.1.50
(v3.1.0-688-g8ae951fbc106)).
Signed-off-by: Björn Töpel <bjorn.topel@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2019-02-05 05:41:22 -07:00
|
|
|
}
|
|
|
|
|
2020-03-04 22:02:04 -07:00
|
|
|
void bpf_jit_build_epilogue(struct rv_jit_context *ctx)
|
bpf, riscv: add BPF JIT for RV64G
This commit adds a BPF JIT for RV64G.
The JIT is a two-pass JIT, and has a dynamic prolog/epilogue (similar
to the MIPS64 BPF JIT) instead of static ones (e.g. x86_64).
At the moment the RISC-V Linux port does not support
CONFIG_HAVE_KPROBES, which means that CONFIG_BPF_EVENTS is not
supported. Thus, no tests involving BPF_PROG_TYPE_TRACEPOINT,
BPF_PROG_TYPE_PERF_EVENT, BPF_PROG_TYPE_KPROBE and
BPF_PROG_TYPE_RAW_TRACEPOINT passes.
The implementation does not support "far branching" (>4KiB).
Test results:
# modprobe test_bpf
test_bpf: Summary: 378 PASSED, 0 FAILED, [366/366 JIT'ed]
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier
...
Summary: 761 PASSED, 507 SKIPPED, 2 FAILED
Note that "test_verifier" was run with one build with
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y and one without, otherwise
many of the the tests that require unaligned access were skipped.
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
0
No CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
59
The two failing test_verifier tests are:
"ld_abs: vlan + abs, test 1"
"ld_abs: jump around ld_abs"
This is due to that "far branching" involved in those tests.
All tests where done on QEMU (QEMU emulator version 3.1.50
(v3.1.0-688-g8ae951fbc106)).
Signed-off-by: Björn Töpel <bjorn.topel@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2019-02-05 05:41:22 -07:00
|
|
|
{
|
2019-12-16 02:13:39 -07:00
|
|
|
__build_epilogue(false, ctx);
|
bpf, riscv: add BPF JIT for RV64G
This commit adds a BPF JIT for RV64G.
The JIT is a two-pass JIT, and has a dynamic prolog/epilogue (similar
to the MIPS64 BPF JIT) instead of static ones (e.g. x86_64).
At the moment the RISC-V Linux port does not support
CONFIG_HAVE_KPROBES, which means that CONFIG_BPF_EVENTS is not
supported. Thus, no tests involving BPF_PROG_TYPE_TRACEPOINT,
BPF_PROG_TYPE_PERF_EVENT, BPF_PROG_TYPE_KPROBE and
BPF_PROG_TYPE_RAW_TRACEPOINT passes.
The implementation does not support "far branching" (>4KiB).
Test results:
# modprobe test_bpf
test_bpf: Summary: 378 PASSED, 0 FAILED, [366/366 JIT'ed]
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier
...
Summary: 761 PASSED, 507 SKIPPED, 2 FAILED
Note that "test_verifier" was run with one build with
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y and one without, otherwise
many of the the tests that require unaligned access were skipped.
CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
0
No CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS:
# echo 1 > /proc/sys/kernel/unprivileged_bpf_disabled
# ./test_verifier | grep -c 'NOTE.*unknown align'
59
The two failing test_verifier tests are:
"ld_abs: vlan + abs, test 1"
"ld_abs: jump around ld_abs"
This is due to that "far branching" involved in those tests.
All tests where done on QEMU (QEMU emulator version 3.1.50
(v3.1.0-688-g8ae951fbc106)).
Signed-off-by: Björn Töpel <bjorn.topel@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2019-02-05 05:41:22 -07:00
|
|
|
}
|
2023-02-21 07:06:56 -07:00
|
|
|
|
|
|
|
bool bpf_jit_supports_kfunc_call(void)
|
|
|
|
{
|
|
|
|
return true;
|
|
|
|
}
|
2024-01-30 05:46:58 -07:00
|
|
|
|
|
|
|
bool bpf_jit_supports_ptr_xchg(void)
|
|
|
|
{
|
|
|
|
return true;
|
|
|
|
}
|
2024-04-04 04:42:03 -07:00
|
|
|
|
|
|
|
bool bpf_jit_supports_arena(void)
|
|
|
|
{
|
|
|
|
return true;
|
|
|
|
}
|
2024-05-02 08:18:51 -07:00
|
|
|
|
|
|
|
bool bpf_jit_supports_percpu_insn(void)
|
|
|
|
{
|
|
|
|
return true;
|
|
|
|
}
|
2024-05-02 08:18:52 -07:00
|
|
|
|
|
|
|
bool bpf_jit_inlines_helper_call(s32 imm)
|
|
|
|
{
|
|
|
|
switch (imm) {
|
|
|
|
case BPF_FUNC_get_smp_processor_id:
|
|
|
|
return true;
|
|
|
|
default:
|
|
|
|
return false;
|
|
|
|
}
|
|
|
|
}
|