1
linux/arch/cris/kernel/process.c
Nick Piggin 5bfb5d690f [PATCH] sched: disable preempt in idle tasks
Run idle threads with preempt disabled.

Also corrected a bugs in arm26's cpu_idle (make it actually call schedule()).
How did it ever work before?

Might fix the CPU hotplugging hang which Nigel Cunningham noted.

We think the bug hits if the idle thread is preempted after checking
need_resched() and before going to sleep, then the CPU offlined.

After calling stop_machine_run, the CPU eventually returns from preemption and
into the idle thread and goes to sleep.  The CPU will continue executing
previous idle and have no chance to call play_dead.

By disabling preemption until we are ready to explicitly schedule, this bug is
fixed and the idle threads generally become more robust.

From: alexs <ashepard@u.washington.edu>

  PPC build fix

From: Yoichi Yuasa <yuasa@hh.iij4u.or.jp>

  MIPS build fix

Signed-off-by: Nick Piggin <npiggin@suse.de>
Signed-off-by: Yoichi Yuasa <yuasa@hh.iij4u.or.jp>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-11-09 07:56:33 -08:00

293 lines
7.7 KiB
C

/* $Id: process.c,v 1.21 2005/03/04 08:16:17 starvik Exp $
*
* linux/arch/cris/kernel/process.c
*
* Copyright (C) 1995 Linus Torvalds
* Copyright (C) 2000-2002 Axis Communications AB
*
* Authors: Bjorn Wesen (bjornw@axis.com)
*
* $Log: process.c,v $
* Revision 1.21 2005/03/04 08:16:17 starvik
* Merge of Linux 2.6.11.
*
* Revision 1.20 2005/01/18 05:57:22 starvik
* Renamed hlt_counter to cris_hlt_counter and made it global.
*
* Revision 1.19 2004/10/19 13:07:43 starvik
* Merge of Linux 2.6.9
*
* Revision 1.18 2004/08/16 12:37:23 starvik
* Merge of Linux 2.6.8
*
* Revision 1.17 2004/04/05 13:53:48 starvik
* Merge of Linux 2.6.5
*
* Revision 1.16 2003/10/27 08:04:33 starvik
* Merge of Linux 2.6.0-test9
*
* Revision 1.15 2003/09/11 07:29:52 starvik
* Merge of Linux 2.6.0-test5
*
* Revision 1.14 2003/06/10 10:21:12 johana
* Moved thread_saved_pc() from arch/cris/kernel/process.c to
* subarch specific process.c. arch-v32 has an erp, no irp.
*
* Revision 1.13 2003/04/09 05:20:47 starvik
* Merge of Linux 2.5.67
*
* Revision 1.12 2002/12/11 15:41:11 starvik
* Extracted v10 (ETRAX 100LX) specific stuff to arch/cris/arch-v10/kernel
*
* Revision 1.11 2002/12/10 09:00:10 starvik
* Merge of Linux 2.5.51
*
* Revision 1.10 2002/11/27 08:42:34 starvik
* Argument to user_regs() is thread_info*
*
* Revision 1.9 2002/11/26 09:44:21 starvik
* New threads exits through ret_from_fork (necessary for preemptive scheduling)
*
* Revision 1.8 2002/11/19 14:35:24 starvik
* Changes from linux 2.4
* Changed struct initializer syntax to the currently prefered notation
*
* Revision 1.7 2002/11/18 07:39:42 starvik
* thread_saved_pc moved here from processor.h
*
* Revision 1.6 2002/11/14 06:51:27 starvik
* Made cpu_idle more similar with other archs
* init_task_union -> init_thread_union
* Updated for new interrupt macros
* sys_clone and do_fork have a new argument, user_tid
*
* Revision 1.5 2002/11/05 06:45:11 starvik
* Merge of Linux 2.5.45
*
* Revision 1.4 2002/02/05 15:37:44 bjornw
* Need init_task.h
*
* Revision 1.3 2002/01/21 15:22:49 bjornw
* current->counter is gone
*
* Revision 1.22 2001/11/13 09:40:43 orjanf
* Added dump_fpu (needed for core dumps).
*
* Revision 1.21 2001/11/12 18:26:21 pkj
* Fixed compiler warnings.
*
* Revision 1.20 2001/10/03 08:21:39 jonashg
* cause_of_death does not exist if CONFIG_SVINTO_SIM is defined.
*
* Revision 1.19 2001/09/26 11:52:54 bjornw
* INIT_MMAP is gone in 2.4.10
*
* Revision 1.18 2001/08/21 21:43:51 hp
* Move last watchdog fix inside #ifdef CONFIG_ETRAX_WATCHDOG
*
* Revision 1.17 2001/08/21 13:48:01 jonashg
* Added fix by HP to avoid oops when doing a hard_reset_now.
*
* Revision 1.16 2001/06/21 02:00:40 hp
* * entry.S: Include asm/unistd.h.
* (_sys_call_table): Use section .rodata, not .data.
* (_kernel_thread): Move from...
* * process.c: ... here.
* * entryoffsets.c (VAL): Break out from...
* (OF): Use VAL.
* (LCLONE_VM): New asmified value from CLONE_VM.
*
* Revision 1.15 2001/06/20 16:31:57 hp
* Add comments to describe empty functions according to review.
*
* Revision 1.14 2001/05/29 11:27:59 markusl
* Fixed so that hard_reset_now will do reset even if watchdog wasn't enabled
*
* Revision 1.13 2001/03/20 19:44:06 bjornw
* Use the 7th syscall argument for regs instead of current_regs
*
*/
/*
* This file handles the architecture-dependent parts of process handling..
*/
#include <asm/atomic.h>
#include <asm/pgtable.h>
#include <asm/uaccess.h>
#include <asm/irq.h>
#include <linux/module.h>
#include <linux/spinlock.h>
#include <linux/fs_struct.h>
#include <linux/init_task.h>
#include <linux/sched.h>
#include <linux/fs.h>
#include <linux/user.h>
#include <linux/elfcore.h>
#include <linux/mqueue.h>
#include <linux/reboot.h>
//#define DEBUG
/*
* Initial task structure. Make this a per-architecture thing,
* because different architectures tend to have different
* alignment requirements and potentially different initial
* setup.
*/
static struct fs_struct init_fs = INIT_FS;
static struct files_struct init_files = INIT_FILES;
static struct signal_struct init_signals = INIT_SIGNALS(init_signals);
static struct sighand_struct init_sighand = INIT_SIGHAND(init_sighand);
struct mm_struct init_mm = INIT_MM(init_mm);
EXPORT_SYMBOL(init_mm);
/*
* Initial thread structure.
*
* We need to make sure that this is 8192-byte aligned due to the
* way process stacks are handled. This is done by having a special
* "init_task" linker map entry..
*/
union thread_union init_thread_union
__attribute__((__section__(".data.init_task"))) =
{ INIT_THREAD_INFO(init_task) };
/*
* Initial task structure.
*
* All other task structs will be allocated on slabs in fork.c
*/
struct task_struct init_task = INIT_TASK(init_task);
EXPORT_SYMBOL(init_task);
/*
* The hlt_counter, disable_hlt and enable_hlt is just here as a hook if
* there would ever be a halt sequence (for power save when idle) with
* some largish delay when halting or resuming *and* a driver that can't
* afford that delay. The hlt_counter would then be checked before
* executing the halt sequence, and the driver marks the unhaltable
* region by enable_hlt/disable_hlt.
*/
int cris_hlt_counter=0;
void disable_hlt(void)
{
cris_hlt_counter++;
}
EXPORT_SYMBOL(disable_hlt);
void enable_hlt(void)
{
cris_hlt_counter--;
}
EXPORT_SYMBOL(enable_hlt);
/*
* The following aren't currently used.
*/
void (*pm_idle)(void);
extern void default_idle(void);
/*
* The idle thread. There's no useful work to be
* done, so just try to conserve power and have a
* low exit latency (ie sit in a loop waiting for
* somebody to say that they'd like to reschedule)
*/
void cpu_idle (void)
{
/* endless idle loop with no priority at all */
while (1) {
while (!need_resched()) {
void (*idle)(void);
/*
* Mark this as an RCU critical section so that
* synchronize_kernel() in the unload path waits
* for our completion.
*/
idle = pm_idle;
if (!idle)
idle = default_idle;
idle();
}
preempt_enable_no_resched();
schedule();
preempt_disable();
}
}
void hard_reset_now (void);
void machine_restart(char *cmd)
{
hard_reset_now();
}
/*
* Similar to machine_power_off, but don't shut off power. Add code
* here to freeze the system for e.g. post-mortem debug purpose when
* possible. This halt has nothing to do with the idle halt.
*/
void machine_halt(void)
{
}
/* If or when software power-off is implemented, add code here. */
void machine_power_off(void)
{
}
/*
* When a process does an "exec", machine state like FPU and debug
* registers need to be reset. This is a hook function for that.
* Currently we don't have any such state to reset, so this is empty.
*/
void flush_thread(void)
{
}
/*
* fill in the user structure for a core dump..
*/
void dump_thread(struct pt_regs * regs, struct user * dump)
{
#if 0
int i;
/* changed the size calculations - should hopefully work better. lbt */
dump->magic = CMAGIC;
dump->start_code = 0;
dump->start_stack = regs->esp & ~(PAGE_SIZE - 1);
dump->u_tsize = ((unsigned long) current->mm->end_code) >> PAGE_SHIFT;
dump->u_dsize = ((unsigned long) (current->mm->brk + (PAGE_SIZE-1))) >> PAGE_SHIFT;
dump->u_dsize -= dump->u_tsize;
dump->u_ssize = 0;
for (i = 0; i < 8; i++)
dump->u_debugreg[i] = current->debugreg[i];
if (dump->start_stack < TASK_SIZE)
dump->u_ssize = ((unsigned long) (TASK_SIZE - dump->start_stack)) >> PAGE_SHIFT;
dump->regs = *regs;
dump->u_fpvalid = dump_fpu (regs, &dump->i387);
#endif
}
/* Fill in the fpu structure for a core dump. */
int dump_fpu(struct pt_regs *regs, elf_fpregset_t *fpu)
{
return 0;
}