1
linux/arch/powerpc/include/asm/iommu.h

330 lines
10 KiB
C
Raw Permalink Normal View History

/* SPDX-License-Identifier: GPL-2.0-or-later */
/*
* Copyright (C) 2001 Mike Corrigan & Dave Engebretsen, IBM Corporation
* Rewrite, cleanup:
* Copyright (C) 2004 Olof Johansson <olof@lixom.net>, IBM Corporation
*/
#ifndef _ASM_IOMMU_H
#define _ASM_IOMMU_H
#ifdef __KERNEL__
#include <linux/compiler.h>
#include <linux/spinlock.h>
#include <linux/device.h>
#include <linux/dma-map-ops.h>
#include <linux/bitops.h>
#include <asm/machdep.h>
#include <asm/types.h>
#include <asm/pci-bridge.h>
#include <asm/asm-const.h>
#define IOMMU_PAGE_SHIFT_4K 12
#define IOMMU_PAGE_SIZE_4K (ASM_CONST(1) << IOMMU_PAGE_SHIFT_4K)
#define IOMMU_PAGE_MASK_4K (~((1 << IOMMU_PAGE_SHIFT_4K) - 1))
#define IOMMU_PAGE_ALIGN_4K(addr) ALIGN(addr, IOMMU_PAGE_SIZE_4K)
#define IOMMU_PAGE_SIZE(tblptr) (ASM_CONST(1) << (tblptr)->it_page_shift)
#define IOMMU_PAGE_MASK(tblptr) (~((1 << (tblptr)->it_page_shift) - 1))
#define IOMMU_PAGE_ALIGN(addr, tblptr) ALIGN(addr, IOMMU_PAGE_SIZE(tblptr))
#define DIRECT64_PROPNAME "linux,direct64-ddr-window-info"
#define DMA64_PROPNAME "linux,dma64-ddr-window-info"
powerpc/pseries/iommu: Split Dynamic DMA Window to be used in Hybrid mode Dynamic DMA Window (DDW) supports TCEs that are backed by 2MB page size. In most configurations, DDW is big enough to pre-map all of LPAR memory for IO. Pre-mapping of memory for DMA results in improvements in IO performance. Persistent memory, vPMEM, can be assigned to an LPAR as well. vPMEM is not contiguous with LPAR memory and usually is assigned at high memory addresses. This makes is not possible to pre-map both vPMEM and LPAR memory in the same DDW. For a dedicated adapter this limitation is not an issue. Dedicated adapters can have both Default DMA window, which is backed by 4K page size and a DDW backed by 2MB page size TCEs. In this scenario, LPAR memory is pre-mapped in the DDW. Any DMA going to the vPMEM is routed via dynamically allocated TCEs in the default window. The issue arises with SR-IOV adapters. There is only one DMA window - either Default or DDW. If an LPAR has vPMEM assigned, memory is not pre-mapped in the DDW since TCEs needs to be allocated for vPMEM as well. In this case, DDW is created and TCEs are dynamically allocated for both vPMEM and LPAR memory. Today, DDW is only used in single mode - direct mapped TCEs or dynamically mapped TCEs. This enhancement breaks a single DDW in 2 regions - 1. First region to pre-map LPAR memory 2. Second region to dynamically allocate TCEs for IO to vPMEM The DDW is split only if it is big enough to pre-map complete LPAR memory and still have some space left to dynamically map vPMEM. Maximum size possible DDW is created as permitted by the Hypervisor. Signed-off-by: Gaurav Batra <gbatra@linux.ibm.com> Reviewed-by: Brian King <brking@linux.vnet.ibm.com> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://msgid.link/20240514014608.35537-1-gbatra@linux.ibm.com
2024-05-13 18:46:08 -07:00
#define MIN_DDW_VPMEM_DMA_WINDOW SZ_2G
/* Boot time flags */
extern int iommu_is_off;
extern int iommu_force_on;
powerpc/iommu: Move tce_xxx callbacks from ppc_md to iommu_table This adds a iommu_table_ops struct and puts pointer to it into the iommu_table struct. This moves tce_build/tce_free/tce_get/tce_flush callbacks from ppc_md to the new struct where they really belong to. This adds the requirement for @it_ops to be initialized before calling iommu_init_table() to make sure that we do not leave any IOMMU table with iommu_table_ops uninitialized. This is not a parameter of iommu_init_table() though as there will be cases when iommu_init_table() will not be called on TCE tables, for example - VFIO. This does s/tce_build/set/, s/tce_free/clear/ and removes "tce_" redundant prefixes. This removes tce_xxx_rm handlers from ppc_md but does not add them to iommu_table_ops as this will be done later if we decide to support TCE hypercalls in real mode. This removes _vm callbacks as only virtual mode is supported by now so this also removes @rm parameter. For pSeries, this always uses tce_buildmulti_pSeriesLP/ tce_buildmulti_pSeriesLP. This changes multi callback to fall back to tce_build_pSeriesLP/tce_free_pSeriesLP if FW_FEATURE_MULTITCE is not present. The reason for this is we still have to support "multitce=off" boot parameter in disable_multitce() and we do not want to walk through all IOMMU tables in the system and replace "multi" callbacks with single ones. For powernv, this defines _ops per PHB type which are P5IOC2/IODA1/IODA2. This makes the callbacks for them public. Later patches will extend callbacks for IODA1/2. No change in behaviour is expected. Signed-off-by: Alexey Kardashevskiy <aik@ozlabs.ru> Reviewed-by: David Gibson <david@gibson.dropbear.id.au> Reviewed-by: Gavin Shan <gwshan@linux.vnet.ibm.com> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
2015-06-04 23:35:06 -07:00
struct iommu_table_ops {
powerpc/iommu/powernv: Release replaced TCE At the moment writing new TCE value to the IOMMU table fails with EBUSY if there is a valid entry already. However PAPR specification allows the guest to write new TCE value without clearing it first. Another problem this patch is addressing is the use of pool locks for external IOMMU users such as VFIO. The pool locks are to protect DMA page allocator rather than entries and since the host kernel does not control what pages are in use, there is no point in pool locks and exchange()+put_page(oldtce) is sufficient to avoid possible races. This adds an exchange() callback to iommu_table_ops which does the same thing as set() plus it returns replaced TCE and DMA direction so the caller can release the pages afterwards. The exchange() receives a physical address unlike set() which receives linear mapping address; and returns a physical address as the clear() does. This implements exchange() for P5IOC2/IODA/IODA2. This adds a requirement for a platform to have exchange() implemented in order to support VFIO. This replaces iommu_tce_build() and iommu_clear_tce() with a single iommu_tce_xchg(). This makes sure that TCE permission bits are not set in TCE passed to IOMMU API as those are to be calculated by platform code from DMA direction. This moves SetPageDirty() to the IOMMU code to make it work for both VFIO ioctl interface in in-kernel TCE acceleration (when it becomes available later). Signed-off-by: Alexey Kardashevskiy <aik@ozlabs.ru> [aw: for the vfio related changes] Acked-by: Alex Williamson <alex.williamson@redhat.com> Reviewed-by: David Gibson <david@gibson.dropbear.id.au> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
2015-06-04 23:35:15 -07:00
/*
* When called with direction==DMA_NONE, it is equal to clear().
* uaddr is a linear map address.
*/
powerpc/iommu: Move tce_xxx callbacks from ppc_md to iommu_table This adds a iommu_table_ops struct and puts pointer to it into the iommu_table struct. This moves tce_build/tce_free/tce_get/tce_flush callbacks from ppc_md to the new struct where they really belong to. This adds the requirement for @it_ops to be initialized before calling iommu_init_table() to make sure that we do not leave any IOMMU table with iommu_table_ops uninitialized. This is not a parameter of iommu_init_table() though as there will be cases when iommu_init_table() will not be called on TCE tables, for example - VFIO. This does s/tce_build/set/, s/tce_free/clear/ and removes "tce_" redundant prefixes. This removes tce_xxx_rm handlers from ppc_md but does not add them to iommu_table_ops as this will be done later if we decide to support TCE hypercalls in real mode. This removes _vm callbacks as only virtual mode is supported by now so this also removes @rm parameter. For pSeries, this always uses tce_buildmulti_pSeriesLP/ tce_buildmulti_pSeriesLP. This changes multi callback to fall back to tce_build_pSeriesLP/tce_free_pSeriesLP if FW_FEATURE_MULTITCE is not present. The reason for this is we still have to support "multitce=off" boot parameter in disable_multitce() and we do not want to walk through all IOMMU tables in the system and replace "multi" callbacks with single ones. For powernv, this defines _ops per PHB type which are P5IOC2/IODA1/IODA2. This makes the callbacks for them public. Later patches will extend callbacks for IODA1/2. No change in behaviour is expected. Signed-off-by: Alexey Kardashevskiy <aik@ozlabs.ru> Reviewed-by: David Gibson <david@gibson.dropbear.id.au> Reviewed-by: Gavin Shan <gwshan@linux.vnet.ibm.com> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
2015-06-04 23:35:06 -07:00
int (*set)(struct iommu_table *tbl,
long index, long npages,
unsigned long uaddr,
enum dma_data_direction direction,
dma-mapping: use unsigned long for dma_attrs The dma-mapping core and the implementations do not change the DMA attributes passed by pointer. Thus the pointer can point to const data. However the attributes do not have to be a bitfield. Instead unsigned long will do fine: 1. This is just simpler. Both in terms of reading the code and setting attributes. Instead of initializing local attributes on the stack and passing pointer to it to dma_set_attr(), just set the bits. 2. It brings safeness and checking for const correctness because the attributes are passed by value. Semantic patches for this change (at least most of them): virtual patch virtual context @r@ identifier f, attrs; @@ f(..., - struct dma_attrs *attrs + unsigned long attrs , ...) { ... } @@ identifier r.f; @@ f(..., - NULL + 0 ) and // Options: --all-includes virtual patch virtual context @r@ identifier f, attrs; type t; @@ t f(..., struct dma_attrs *attrs); @@ identifier r.f; @@ f(..., - NULL + 0 ) Link: http://lkml.kernel.org/r/1468399300-5399-2-git-send-email-k.kozlowski@samsung.com Signed-off-by: Krzysztof Kozlowski <k.kozlowski@samsung.com> Acked-by: Vineet Gupta <vgupta@synopsys.com> Acked-by: Robin Murphy <robin.murphy@arm.com> Acked-by: Hans-Christian Noren Egtvedt <egtvedt@samfundet.no> Acked-by: Mark Salter <msalter@redhat.com> [c6x] Acked-by: Jesper Nilsson <jesper.nilsson@axis.com> [cris] Acked-by: Daniel Vetter <daniel.vetter@ffwll.ch> [drm] Reviewed-by: Bart Van Assche <bart.vanassche@sandisk.com> Acked-by: Joerg Roedel <jroedel@suse.de> [iommu] Acked-by: Fabien Dessenne <fabien.dessenne@st.com> [bdisp] Reviewed-by: Marek Szyprowski <m.szyprowski@samsung.com> [vb2-core] Acked-by: David Vrabel <david.vrabel@citrix.com> [xen] Acked-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com> [xen swiotlb] Acked-by: Joerg Roedel <jroedel@suse.de> [iommu] Acked-by: Richard Kuo <rkuo@codeaurora.org> [hexagon] Acked-by: Geert Uytterhoeven <geert@linux-m68k.org> [m68k] Acked-by: Gerald Schaefer <gerald.schaefer@de.ibm.com> [s390] Acked-by: Bjorn Andersson <bjorn.andersson@linaro.org> Acked-by: Hans-Christian Noren Egtvedt <egtvedt@samfundet.no> [avr32] Acked-by: Vineet Gupta <vgupta@synopsys.com> [arc] Acked-by: Robin Murphy <robin.murphy@arm.com> [arm64 and dma-iommu] Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-08-03 13:46:00 -07:00
unsigned long attrs);
powerpc/iommu/powernv: Release replaced TCE At the moment writing new TCE value to the IOMMU table fails with EBUSY if there is a valid entry already. However PAPR specification allows the guest to write new TCE value without clearing it first. Another problem this patch is addressing is the use of pool locks for external IOMMU users such as VFIO. The pool locks are to protect DMA page allocator rather than entries and since the host kernel does not control what pages are in use, there is no point in pool locks and exchange()+put_page(oldtce) is sufficient to avoid possible races. This adds an exchange() callback to iommu_table_ops which does the same thing as set() plus it returns replaced TCE and DMA direction so the caller can release the pages afterwards. The exchange() receives a physical address unlike set() which receives linear mapping address; and returns a physical address as the clear() does. This implements exchange() for P5IOC2/IODA/IODA2. This adds a requirement for a platform to have exchange() implemented in order to support VFIO. This replaces iommu_tce_build() and iommu_clear_tce() with a single iommu_tce_xchg(). This makes sure that TCE permission bits are not set in TCE passed to IOMMU API as those are to be calculated by platform code from DMA direction. This moves SetPageDirty() to the IOMMU code to make it work for both VFIO ioctl interface in in-kernel TCE acceleration (when it becomes available later). Signed-off-by: Alexey Kardashevskiy <aik@ozlabs.ru> [aw: for the vfio related changes] Acked-by: Alex Williamson <alex.williamson@redhat.com> Reviewed-by: David Gibson <david@gibson.dropbear.id.au> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
2015-06-04 23:35:15 -07:00
#ifdef CONFIG_IOMMU_API
/*
* Exchanges existing TCE with new TCE plus direction bits;
* returns old TCE and DMA direction mask.
* @tce is a physical address.
*/
int (*xchg_no_kill)(struct iommu_table *tbl,
powerpc/iommu/powernv: Release replaced TCE At the moment writing new TCE value to the IOMMU table fails with EBUSY if there is a valid entry already. However PAPR specification allows the guest to write new TCE value without clearing it first. Another problem this patch is addressing is the use of pool locks for external IOMMU users such as VFIO. The pool locks are to protect DMA page allocator rather than entries and since the host kernel does not control what pages are in use, there is no point in pool locks and exchange()+put_page(oldtce) is sufficient to avoid possible races. This adds an exchange() callback to iommu_table_ops which does the same thing as set() plus it returns replaced TCE and DMA direction so the caller can release the pages afterwards. The exchange() receives a physical address unlike set() which receives linear mapping address; and returns a physical address as the clear() does. This implements exchange() for P5IOC2/IODA/IODA2. This adds a requirement for a platform to have exchange() implemented in order to support VFIO. This replaces iommu_tce_build() and iommu_clear_tce() with a single iommu_tce_xchg(). This makes sure that TCE permission bits are not set in TCE passed to IOMMU API as those are to be calculated by platform code from DMA direction. This moves SetPageDirty() to the IOMMU code to make it work for both VFIO ioctl interface in in-kernel TCE acceleration (when it becomes available later). Signed-off-by: Alexey Kardashevskiy <aik@ozlabs.ru> [aw: for the vfio related changes] Acked-by: Alex Williamson <alex.williamson@redhat.com> Reviewed-by: David Gibson <david@gibson.dropbear.id.au> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
2015-06-04 23:35:15 -07:00
long index,
unsigned long *hpa,
KVM: PPC: Book3s: Retire H_PUT_TCE/etc real mode handlers LoPAPR defines guest visible IOMMU with hypercalls to use it - H_PUT_TCE/etc. Implemented first on POWER7 where hypercalls would trap in the KVM in the real mode (with MMU off). The problem with the real mode is some memory is not available and some API usage crashed the host but enabling MMU was an expensive operation. The problems with the real mode handlers are: 1. Occasionally these cannot complete the request so the code is copied+modified to work in the virtual mode, very little is shared; 2. The real mode handlers have to be linked into vmlinux to work; 3. An exception in real mode immediately reboots the machine. If the small DMA window is used, the real mode handlers bring better performance. However since POWER8, there has always been a bigger DMA window which VMs use to map the entire VM memory to avoid calling H_PUT_TCE. Such 1:1 mapping happens once and uses H_PUT_TCE_INDIRECT (a bulk version of H_PUT_TCE) which virtual mode handler is even closer to its real mode version. On POWER9 hypercalls trap straight to the virtual mode so the real mode handlers never execute on POWER9 and later CPUs. So with the current use of the DMA windows and MMU improvements in POWER9 and later, there is no point in duplicating the code. The 32bit passed through devices may slow down but we do not have many of these in practice. For example, with this applied, a 1Gbit ethernet adapter still demostrates above 800Mbit/s of actual throughput. This removes the real mode handlers from KVM and related code from the powernv platform. This updates the list of implemented hcalls in KVM-HV as the realmode handlers are removed. This changes ABI - kvmppc_h_get_tce() moves to the KVM module and kvmppc_find_table() is static now. Signed-off-by: Alexey Kardashevskiy <aik@ozlabs.ru> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/20220506053755.3820702-1-aik@ozlabs.ru
2022-05-05 22:37:55 -07:00
enum dma_data_direction *direction);
void (*tce_kill)(struct iommu_table *tbl,
unsigned long index,
KVM: PPC: Book3s: Retire H_PUT_TCE/etc real mode handlers LoPAPR defines guest visible IOMMU with hypercalls to use it - H_PUT_TCE/etc. Implemented first on POWER7 where hypercalls would trap in the KVM in the real mode (with MMU off). The problem with the real mode is some memory is not available and some API usage crashed the host but enabling MMU was an expensive operation. The problems with the real mode handlers are: 1. Occasionally these cannot complete the request so the code is copied+modified to work in the virtual mode, very little is shared; 2. The real mode handlers have to be linked into vmlinux to work; 3. An exception in real mode immediately reboots the machine. If the small DMA window is used, the real mode handlers bring better performance. However since POWER8, there has always been a bigger DMA window which VMs use to map the entire VM memory to avoid calling H_PUT_TCE. Such 1:1 mapping happens once and uses H_PUT_TCE_INDIRECT (a bulk version of H_PUT_TCE) which virtual mode handler is even closer to its real mode version. On POWER9 hypercalls trap straight to the virtual mode so the real mode handlers never execute on POWER9 and later CPUs. So with the current use of the DMA windows and MMU improvements in POWER9 and later, there is no point in duplicating the code. The 32bit passed through devices may slow down but we do not have many of these in practice. For example, with this applied, a 1Gbit ethernet adapter still demostrates above 800Mbit/s of actual throughput. This removes the real mode handlers from KVM and related code from the powernv platform. This updates the list of implemented hcalls in KVM-HV as the realmode handlers are removed. This changes ABI - kvmppc_h_get_tce() moves to the KVM module and kvmppc_find_table() is static now. Signed-off-by: Alexey Kardashevskiy <aik@ozlabs.ru> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/20220506053755.3820702-1-aik@ozlabs.ru
2022-05-05 22:37:55 -07:00
unsigned long pages);
powerpc/powernv/ioda: Allocate indirect TCE levels on demand At the moment we allocate the entire TCE table, twice (hardware part and userspace translation cache). This normally works as we normally have contigous memory and the guest will map entire RAM for 64bit DMA. However if we have sparse RAM (one example is a memory device), then we will allocate TCEs which will never be used as the guest only maps actual memory for DMA. If it is a single level TCE table, there is nothing we can really do but if it a multilevel table, we can skip allocating TCEs we know we won't need. This adds ability to allocate only first level, saving memory. This changes iommu_table::free() to avoid allocating of an extra level; iommu_table::set() will do this when needed. This adds @alloc parameter to iommu_table::exchange() to tell the callback if it can allocate an extra level; the flag is set to "false" for the realmode KVM handlers of H_PUT_TCE hcalls and the callback returns H_TOO_HARD. This still requires the entire table to be counted in mm::locked_vm. To be conservative, this only does on-demand allocation when the usespace cache table is requested which is the case of VFIO. The example math for a system replicating a powernv setup with NVLink2 in a guest: 16GB RAM mapped at 0x0 128GB GPU RAM window (16GB of actual RAM) mapped at 0x244000000000 the table to cover that all with 64K pages takes: (((0x244000000000 + 0x2000000000) >> 16)*8)>>20 = 4556MB If we allocate only necessary TCE levels, we will only need: (((0x400000000 + 0x400000000) >> 16)*8)>>20 = 4MB (plus some for indirect levels). Signed-off-by: Alexey Kardashevskiy <aik@ozlabs.ru> Reviewed-by: David Gibson <david@gibson.dropbear.id.au> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
2018-07-03 23:13:49 -07:00
__be64 *(*useraddrptr)(struct iommu_table *tbl, long index, bool alloc);
powerpc/iommu/powernv: Release replaced TCE At the moment writing new TCE value to the IOMMU table fails with EBUSY if there is a valid entry already. However PAPR specification allows the guest to write new TCE value without clearing it first. Another problem this patch is addressing is the use of pool locks for external IOMMU users such as VFIO. The pool locks are to protect DMA page allocator rather than entries and since the host kernel does not control what pages are in use, there is no point in pool locks and exchange()+put_page(oldtce) is sufficient to avoid possible races. This adds an exchange() callback to iommu_table_ops which does the same thing as set() plus it returns replaced TCE and DMA direction so the caller can release the pages afterwards. The exchange() receives a physical address unlike set() which receives linear mapping address; and returns a physical address as the clear() does. This implements exchange() for P5IOC2/IODA/IODA2. This adds a requirement for a platform to have exchange() implemented in order to support VFIO. This replaces iommu_tce_build() and iommu_clear_tce() with a single iommu_tce_xchg(). This makes sure that TCE permission bits are not set in TCE passed to IOMMU API as those are to be calculated by platform code from DMA direction. This moves SetPageDirty() to the IOMMU code to make it work for both VFIO ioctl interface in in-kernel TCE acceleration (when it becomes available later). Signed-off-by: Alexey Kardashevskiy <aik@ozlabs.ru> [aw: for the vfio related changes] Acked-by: Alex Williamson <alex.williamson@redhat.com> Reviewed-by: David Gibson <david@gibson.dropbear.id.au> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
2015-06-04 23:35:15 -07:00
#endif
powerpc/iommu: Move tce_xxx callbacks from ppc_md to iommu_table This adds a iommu_table_ops struct and puts pointer to it into the iommu_table struct. This moves tce_build/tce_free/tce_get/tce_flush callbacks from ppc_md to the new struct where they really belong to. This adds the requirement for @it_ops to be initialized before calling iommu_init_table() to make sure that we do not leave any IOMMU table with iommu_table_ops uninitialized. This is not a parameter of iommu_init_table() though as there will be cases when iommu_init_table() will not be called on TCE tables, for example - VFIO. This does s/tce_build/set/, s/tce_free/clear/ and removes "tce_" redundant prefixes. This removes tce_xxx_rm handlers from ppc_md but does not add them to iommu_table_ops as this will be done later if we decide to support TCE hypercalls in real mode. This removes _vm callbacks as only virtual mode is supported by now so this also removes @rm parameter. For pSeries, this always uses tce_buildmulti_pSeriesLP/ tce_buildmulti_pSeriesLP. This changes multi callback to fall back to tce_build_pSeriesLP/tce_free_pSeriesLP if FW_FEATURE_MULTITCE is not present. The reason for this is we still have to support "multitce=off" boot parameter in disable_multitce() and we do not want to walk through all IOMMU tables in the system and replace "multi" callbacks with single ones. For powernv, this defines _ops per PHB type which are P5IOC2/IODA1/IODA2. This makes the callbacks for them public. Later patches will extend callbacks for IODA1/2. No change in behaviour is expected. Signed-off-by: Alexey Kardashevskiy <aik@ozlabs.ru> Reviewed-by: David Gibson <david@gibson.dropbear.id.au> Reviewed-by: Gavin Shan <gwshan@linux.vnet.ibm.com> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
2015-06-04 23:35:06 -07:00
void (*clear)(struct iommu_table *tbl,
long index, long npages);
powerpc/iommu/powernv: Release replaced TCE At the moment writing new TCE value to the IOMMU table fails with EBUSY if there is a valid entry already. However PAPR specification allows the guest to write new TCE value without clearing it first. Another problem this patch is addressing is the use of pool locks for external IOMMU users such as VFIO. The pool locks are to protect DMA page allocator rather than entries and since the host kernel does not control what pages are in use, there is no point in pool locks and exchange()+put_page(oldtce) is sufficient to avoid possible races. This adds an exchange() callback to iommu_table_ops which does the same thing as set() plus it returns replaced TCE and DMA direction so the caller can release the pages afterwards. The exchange() receives a physical address unlike set() which receives linear mapping address; and returns a physical address as the clear() does. This implements exchange() for P5IOC2/IODA/IODA2. This adds a requirement for a platform to have exchange() implemented in order to support VFIO. This replaces iommu_tce_build() and iommu_clear_tce() with a single iommu_tce_xchg(). This makes sure that TCE permission bits are not set in TCE passed to IOMMU API as those are to be calculated by platform code from DMA direction. This moves SetPageDirty() to the IOMMU code to make it work for both VFIO ioctl interface in in-kernel TCE acceleration (when it becomes available later). Signed-off-by: Alexey Kardashevskiy <aik@ozlabs.ru> [aw: for the vfio related changes] Acked-by: Alex Williamson <alex.williamson@redhat.com> Reviewed-by: David Gibson <david@gibson.dropbear.id.au> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
2015-06-04 23:35:15 -07:00
/* get() returns a physical address */
powerpc/iommu: Move tce_xxx callbacks from ppc_md to iommu_table This adds a iommu_table_ops struct and puts pointer to it into the iommu_table struct. This moves tce_build/tce_free/tce_get/tce_flush callbacks from ppc_md to the new struct where they really belong to. This adds the requirement for @it_ops to be initialized before calling iommu_init_table() to make sure that we do not leave any IOMMU table with iommu_table_ops uninitialized. This is not a parameter of iommu_init_table() though as there will be cases when iommu_init_table() will not be called on TCE tables, for example - VFIO. This does s/tce_build/set/, s/tce_free/clear/ and removes "tce_" redundant prefixes. This removes tce_xxx_rm handlers from ppc_md but does not add them to iommu_table_ops as this will be done later if we decide to support TCE hypercalls in real mode. This removes _vm callbacks as only virtual mode is supported by now so this also removes @rm parameter. For pSeries, this always uses tce_buildmulti_pSeriesLP/ tce_buildmulti_pSeriesLP. This changes multi callback to fall back to tce_build_pSeriesLP/tce_free_pSeriesLP if FW_FEATURE_MULTITCE is not present. The reason for this is we still have to support "multitce=off" boot parameter in disable_multitce() and we do not want to walk through all IOMMU tables in the system and replace "multi" callbacks with single ones. For powernv, this defines _ops per PHB type which are P5IOC2/IODA1/IODA2. This makes the callbacks for them public. Later patches will extend callbacks for IODA1/2. No change in behaviour is expected. Signed-off-by: Alexey Kardashevskiy <aik@ozlabs.ru> Reviewed-by: David Gibson <david@gibson.dropbear.id.au> Reviewed-by: Gavin Shan <gwshan@linux.vnet.ibm.com> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
2015-06-04 23:35:06 -07:00
unsigned long (*get)(struct iommu_table *tbl, long index);
void (*flush)(struct iommu_table *tbl);
void (*free)(struct iommu_table *tbl);
powerpc/iommu: Move tce_xxx callbacks from ppc_md to iommu_table This adds a iommu_table_ops struct and puts pointer to it into the iommu_table struct. This moves tce_build/tce_free/tce_get/tce_flush callbacks from ppc_md to the new struct where they really belong to. This adds the requirement for @it_ops to be initialized before calling iommu_init_table() to make sure that we do not leave any IOMMU table with iommu_table_ops uninitialized. This is not a parameter of iommu_init_table() though as there will be cases when iommu_init_table() will not be called on TCE tables, for example - VFIO. This does s/tce_build/set/, s/tce_free/clear/ and removes "tce_" redundant prefixes. This removes tce_xxx_rm handlers from ppc_md but does not add them to iommu_table_ops as this will be done later if we decide to support TCE hypercalls in real mode. This removes _vm callbacks as only virtual mode is supported by now so this also removes @rm parameter. For pSeries, this always uses tce_buildmulti_pSeriesLP/ tce_buildmulti_pSeriesLP. This changes multi callback to fall back to tce_build_pSeriesLP/tce_free_pSeriesLP if FW_FEATURE_MULTITCE is not present. The reason for this is we still have to support "multitce=off" boot parameter in disable_multitce() and we do not want to walk through all IOMMU tables in the system and replace "multi" callbacks with single ones. For powernv, this defines _ops per PHB type which are P5IOC2/IODA1/IODA2. This makes the callbacks for them public. Later patches will extend callbacks for IODA1/2. No change in behaviour is expected. Signed-off-by: Alexey Kardashevskiy <aik@ozlabs.ru> Reviewed-by: David Gibson <david@gibson.dropbear.id.au> Reviewed-by: Gavin Shan <gwshan@linux.vnet.ibm.com> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
2015-06-04 23:35:06 -07:00
};
/* These are used by VIO */
extern struct iommu_table_ops iommu_table_lpar_multi_ops;
extern struct iommu_table_ops iommu_table_pseries_ops;
/*
* IOMAP_MAX_ORDER defines the largest contiguous block
* of dma space we can get. IOMAP_MAX_ORDER = 13
* allows up to 2**12 pages (4096 * 4096) = 16 MB
*/
#define IOMAP_MAX_ORDER 13
#define IOMMU_POOL_HASHBITS 2
#define IOMMU_NR_POOLS (1 << IOMMU_POOL_HASHBITS)
struct iommu_pool {
unsigned long start;
unsigned long end;
unsigned long hint;
spinlock_t lock;
} ____cacheline_aligned_in_smp;
struct iommu_table {
unsigned long it_busno; /* Bus number this table belongs to */
unsigned long it_size; /* Size of iommu table in entries */
unsigned long it_indirect_levels;
unsigned long it_level_size;
unsigned long it_allocated_size;
unsigned long it_offset; /* Offset into global table */
unsigned long it_base; /* mapped address of tce table */
unsigned long it_index; /* which iommu table this is */
unsigned long it_type; /* type: PCI or Virtual Bus */
unsigned long it_blocksize; /* Entries in each block (cacheline) */
unsigned long poolsize;
unsigned long nr_pools;
struct iommu_pool large_pool;
struct iommu_pool pools[IOMMU_NR_POOLS];
unsigned long *it_map; /* A simple allocation bitmap for now */
unsigned long it_page_shift;/* table iommu page size */
struct list_head it_group_list;/* List of iommu_table_group_link */
__be64 *it_userspace; /* userspace view of the table */
powerpc/iommu: Move tce_xxx callbacks from ppc_md to iommu_table This adds a iommu_table_ops struct and puts pointer to it into the iommu_table struct. This moves tce_build/tce_free/tce_get/tce_flush callbacks from ppc_md to the new struct where they really belong to. This adds the requirement for @it_ops to be initialized before calling iommu_init_table() to make sure that we do not leave any IOMMU table with iommu_table_ops uninitialized. This is not a parameter of iommu_init_table() though as there will be cases when iommu_init_table() will not be called on TCE tables, for example - VFIO. This does s/tce_build/set/, s/tce_free/clear/ and removes "tce_" redundant prefixes. This removes tce_xxx_rm handlers from ppc_md but does not add them to iommu_table_ops as this will be done later if we decide to support TCE hypercalls in real mode. This removes _vm callbacks as only virtual mode is supported by now so this also removes @rm parameter. For pSeries, this always uses tce_buildmulti_pSeriesLP/ tce_buildmulti_pSeriesLP. This changes multi callback to fall back to tce_build_pSeriesLP/tce_free_pSeriesLP if FW_FEATURE_MULTITCE is not present. The reason for this is we still have to support "multitce=off" boot parameter in disable_multitce() and we do not want to walk through all IOMMU tables in the system and replace "multi" callbacks with single ones. For powernv, this defines _ops per PHB type which are P5IOC2/IODA1/IODA2. This makes the callbacks for them public. Later patches will extend callbacks for IODA1/2. No change in behaviour is expected. Signed-off-by: Alexey Kardashevskiy <aik@ozlabs.ru> Reviewed-by: David Gibson <david@gibson.dropbear.id.au> Reviewed-by: Gavin Shan <gwshan@linux.vnet.ibm.com> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
2015-06-04 23:35:06 -07:00
struct iommu_table_ops *it_ops;
struct kref it_kref;
powerpc/powernv/ioda: Allocate indirect TCE levels on demand At the moment we allocate the entire TCE table, twice (hardware part and userspace translation cache). This normally works as we normally have contigous memory and the guest will map entire RAM for 64bit DMA. However if we have sparse RAM (one example is a memory device), then we will allocate TCEs which will never be used as the guest only maps actual memory for DMA. If it is a single level TCE table, there is nothing we can really do but if it a multilevel table, we can skip allocating TCEs we know we won't need. This adds ability to allocate only first level, saving memory. This changes iommu_table::free() to avoid allocating of an extra level; iommu_table::set() will do this when needed. This adds @alloc parameter to iommu_table::exchange() to tell the callback if it can allocate an extra level; the flag is set to "false" for the realmode KVM handlers of H_PUT_TCE hcalls and the callback returns H_TOO_HARD. This still requires the entire table to be counted in mm::locked_vm. To be conservative, this only does on-demand allocation when the usespace cache table is requested which is the case of VFIO. The example math for a system replicating a powernv setup with NVLink2 in a guest: 16GB RAM mapped at 0x0 128GB GPU RAM window (16GB of actual RAM) mapped at 0x244000000000 the table to cover that all with 64K pages takes: (((0x244000000000 + 0x2000000000) >> 16)*8)>>20 = 4556MB If we allocate only necessary TCE levels, we will only need: (((0x400000000 + 0x400000000) >> 16)*8)>>20 = 4MB (plus some for indirect levels). Signed-off-by: Alexey Kardashevskiy <aik@ozlabs.ru> Reviewed-by: David Gibson <david@gibson.dropbear.id.au> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
2018-07-03 23:13:49 -07:00
int it_nid;
powerpc/powernv/ioda2: Create bigger default window with 64k IOMMU pages At the moment we create a small window only for 32bit devices, the window maps 0..2GB of the PCI space only. For other devices we either use a sketchy bypass or hardware bypass but the former can only work if the amount of RAM is no bigger than the device's DMA mask and the latter requires devices to support at least 59bit DMA. This extends the default DMA window to the maximum size possible to allow a wider DMA mask than just 32bit. The default window size is now limited by the the iommu_table::it_map allocation bitmap which is a contiguous array, 1 bit per an IOMMU page. This increases the default IOMMU page size from hard coded 4K to the system page size to allow wider DMA masks. This increases the level number to not exceed the max order allocation limit per TCE level. By the same time, this keeps minimal levels number as 2 in order to save memory. As the extended window now overlaps the 32bit MMIO region, this adds an area reservation to iommu_init_table(). After this change the default window size is 0x80000000000==1<<43 so devices limited to DMA mask smaller than the amount of system RAM can still use more than just 2GB of memory for DMA. This is an optimization and not a bug fix for DMA API usage. With the on-demand allocation of indirect TCE table levels enabled and 2 levels, the first TCE level size is just 1<<ceil((log2(0x7ffffffffff+1)-16)/2)=16384 TCEs or 2 system pages. Signed-off-by: Alexey Kardashevskiy <aik@ozlabs.ru> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/20190718051139.74787-5-aik@ozlabs.ru
2019-07-17 22:11:39 -07:00
unsigned long it_reserved_start; /* Start of not-DMA-able (MMIO) area */
unsigned long it_reserved_end;
};
#define IOMMU_TABLE_USERSPACE_ENTRY_RO(tbl, entry) \
powerpc/powernv/ioda: Allocate indirect TCE levels on demand At the moment we allocate the entire TCE table, twice (hardware part and userspace translation cache). This normally works as we normally have contigous memory and the guest will map entire RAM for 64bit DMA. However if we have sparse RAM (one example is a memory device), then we will allocate TCEs which will never be used as the guest only maps actual memory for DMA. If it is a single level TCE table, there is nothing we can really do but if it a multilevel table, we can skip allocating TCEs we know we won't need. This adds ability to allocate only first level, saving memory. This changes iommu_table::free() to avoid allocating of an extra level; iommu_table::set() will do this when needed. This adds @alloc parameter to iommu_table::exchange() to tell the callback if it can allocate an extra level; the flag is set to "false" for the realmode KVM handlers of H_PUT_TCE hcalls and the callback returns H_TOO_HARD. This still requires the entire table to be counted in mm::locked_vm. To be conservative, this only does on-demand allocation when the usespace cache table is requested which is the case of VFIO. The example math for a system replicating a powernv setup with NVLink2 in a guest: 16GB RAM mapped at 0x0 128GB GPU RAM window (16GB of actual RAM) mapped at 0x244000000000 the table to cover that all with 64K pages takes: (((0x244000000000 + 0x2000000000) >> 16)*8)>>20 = 4556MB If we allocate only necessary TCE levels, we will only need: (((0x400000000 + 0x400000000) >> 16)*8)>>20 = 4MB (plus some for indirect levels). Signed-off-by: Alexey Kardashevskiy <aik@ozlabs.ru> Reviewed-by: David Gibson <david@gibson.dropbear.id.au> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
2018-07-03 23:13:49 -07:00
((tbl)->it_ops->useraddrptr((tbl), (entry), false))
vfio: powerpc/spapr: Register memory and define IOMMU v2 The existing implementation accounts the whole DMA window in the locked_vm counter. This is going to be worse with multiple containers and huge DMA windows. Also, real-time accounting would requite additional tracking of accounted pages due to the page size difference - IOMMU uses 4K pages and system uses 4K or 64K pages. Another issue is that actual pages pinning/unpinning happens on every DMA map/unmap request. This does not affect the performance much now as we spend way too much time now on switching context between guest/userspace/host but this will start to matter when we add in-kernel DMA map/unmap acceleration. This introduces a new IOMMU type for SPAPR - VFIO_SPAPR_TCE_v2_IOMMU. New IOMMU deprecates VFIO_IOMMU_ENABLE/VFIO_IOMMU_DISABLE and introduces 2 new ioctls to register/unregister DMA memory - VFIO_IOMMU_SPAPR_REGISTER_MEMORY and VFIO_IOMMU_SPAPR_UNREGISTER_MEMORY - which receive user space address and size of a memory region which needs to be pinned/unpinned and counted in locked_vm. New IOMMU splits physical pages pinning and TCE table update into 2 different operations. It requires: 1) guest pages to be registered first 2) consequent map/unmap requests to work only with pre-registered memory. For the default single window case this means that the entire guest (instead of 2GB) needs to be pinned before using VFIO. When a huge DMA window is added, no additional pinning will be required, otherwise it would be guest RAM + 2GB. The new memory registration ioctls are not supported by VFIO_SPAPR_TCE_IOMMU. Dynamic DMA window and in-kernel acceleration will require memory to be preregistered in order to work. The accounting is done per the user process. This advertises v2 SPAPR TCE IOMMU and restricts what the userspace can do with v1 or v2 IOMMUs. In order to support memory pre-registration, we need a way to track the use of every registered memory region and only allow unregistration if a region is not in use anymore. So we need a way to tell from what region the just cleared TCE was from. This adds a userspace view of the TCE table into iommu_table struct. It contains userspace address, one per TCE entry. The table is only allocated when the ownership over an IOMMU group is taken which means it is only used from outside of the powernv code (such as VFIO). As v2 IOMMU supports IODA2 and pre-IODA2 IOMMUs (which do not support DDW API), this creates a default DMA window for IODA2 for consistency. Signed-off-by: Alexey Kardashevskiy <aik@ozlabs.ru> [aw: for the vfio related changes] Acked-by: Alex Williamson <alex.williamson@redhat.com> Reviewed-by: David Gibson <david@gibson.dropbear.id.au> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
2015-06-04 23:35:25 -07:00
#define IOMMU_TABLE_USERSPACE_ENTRY(tbl, entry) \
powerpc/powernv/ioda: Allocate indirect TCE levels on demand At the moment we allocate the entire TCE table, twice (hardware part and userspace translation cache). This normally works as we normally have contigous memory and the guest will map entire RAM for 64bit DMA. However if we have sparse RAM (one example is a memory device), then we will allocate TCEs which will never be used as the guest only maps actual memory for DMA. If it is a single level TCE table, there is nothing we can really do but if it a multilevel table, we can skip allocating TCEs we know we won't need. This adds ability to allocate only first level, saving memory. This changes iommu_table::free() to avoid allocating of an extra level; iommu_table::set() will do this when needed. This adds @alloc parameter to iommu_table::exchange() to tell the callback if it can allocate an extra level; the flag is set to "false" for the realmode KVM handlers of H_PUT_TCE hcalls and the callback returns H_TOO_HARD. This still requires the entire table to be counted in mm::locked_vm. To be conservative, this only does on-demand allocation when the usespace cache table is requested which is the case of VFIO. The example math for a system replicating a powernv setup with NVLink2 in a guest: 16GB RAM mapped at 0x0 128GB GPU RAM window (16GB of actual RAM) mapped at 0x244000000000 the table to cover that all with 64K pages takes: (((0x244000000000 + 0x2000000000) >> 16)*8)>>20 = 4556MB If we allocate only necessary TCE levels, we will only need: (((0x400000000 + 0x400000000) >> 16)*8)>>20 = 4MB (plus some for indirect levels). Signed-off-by: Alexey Kardashevskiy <aik@ozlabs.ru> Reviewed-by: David Gibson <david@gibson.dropbear.id.au> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
2018-07-03 23:13:49 -07:00
((tbl)->it_ops->useraddrptr((tbl), (entry), true))
vfio: powerpc/spapr: Register memory and define IOMMU v2 The existing implementation accounts the whole DMA window in the locked_vm counter. This is going to be worse with multiple containers and huge DMA windows. Also, real-time accounting would requite additional tracking of accounted pages due to the page size difference - IOMMU uses 4K pages and system uses 4K or 64K pages. Another issue is that actual pages pinning/unpinning happens on every DMA map/unmap request. This does not affect the performance much now as we spend way too much time now on switching context between guest/userspace/host but this will start to matter when we add in-kernel DMA map/unmap acceleration. This introduces a new IOMMU type for SPAPR - VFIO_SPAPR_TCE_v2_IOMMU. New IOMMU deprecates VFIO_IOMMU_ENABLE/VFIO_IOMMU_DISABLE and introduces 2 new ioctls to register/unregister DMA memory - VFIO_IOMMU_SPAPR_REGISTER_MEMORY and VFIO_IOMMU_SPAPR_UNREGISTER_MEMORY - which receive user space address and size of a memory region which needs to be pinned/unpinned and counted in locked_vm. New IOMMU splits physical pages pinning and TCE table update into 2 different operations. It requires: 1) guest pages to be registered first 2) consequent map/unmap requests to work only with pre-registered memory. For the default single window case this means that the entire guest (instead of 2GB) needs to be pinned before using VFIO. When a huge DMA window is added, no additional pinning will be required, otherwise it would be guest RAM + 2GB. The new memory registration ioctls are not supported by VFIO_SPAPR_TCE_IOMMU. Dynamic DMA window and in-kernel acceleration will require memory to be preregistered in order to work. The accounting is done per the user process. This advertises v2 SPAPR TCE IOMMU and restricts what the userspace can do with v1 or v2 IOMMUs. In order to support memory pre-registration, we need a way to track the use of every registered memory region and only allow unregistration if a region is not in use anymore. So we need a way to tell from what region the just cleared TCE was from. This adds a userspace view of the TCE table into iommu_table struct. It contains userspace address, one per TCE entry. The table is only allocated when the ownership over an IOMMU group is taken which means it is only used from outside of the powernv code (such as VFIO). As v2 IOMMU supports IODA2 and pre-IODA2 IOMMUs (which do not support DDW API), this creates a default DMA window for IODA2 for consistency. Signed-off-by: Alexey Kardashevskiy <aik@ozlabs.ru> [aw: for the vfio related changes] Acked-by: Alex Williamson <alex.williamson@redhat.com> Reviewed-by: David Gibson <david@gibson.dropbear.id.au> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
2015-06-04 23:35:25 -07:00
/* Pure 2^n version of get_order */
static inline __attribute_const__
int get_iommu_order(unsigned long size, struct iommu_table *tbl)
{
return __ilog2((size - 1) >> tbl->it_page_shift) + 1;
}
struct scatterlist;
#ifdef CONFIG_PPC64
static inline void set_iommu_table_base(struct device *dev,
struct iommu_table *base)
{
dev->archdata.iommu_table_base = base;
}
static inline void *get_iommu_table_base(struct device *dev)
{
return dev->archdata.iommu_table_base;
}
extern int dma_iommu_dma_supported(struct device *dev, u64 mask);
extern struct iommu_table *iommu_tce_table_get(struct iommu_table *tbl);
extern int iommu_tce_table_put(struct iommu_table *tbl);
/* Initializes an iommu_table based in values set in the passed-in
* structure
*/
powerpc/powernv/ioda2: Create bigger default window with 64k IOMMU pages At the moment we create a small window only for 32bit devices, the window maps 0..2GB of the PCI space only. For other devices we either use a sketchy bypass or hardware bypass but the former can only work if the amount of RAM is no bigger than the device's DMA mask and the latter requires devices to support at least 59bit DMA. This extends the default DMA window to the maximum size possible to allow a wider DMA mask than just 32bit. The default window size is now limited by the the iommu_table::it_map allocation bitmap which is a contiguous array, 1 bit per an IOMMU page. This increases the default IOMMU page size from hard coded 4K to the system page size to allow wider DMA masks. This increases the level number to not exceed the max order allocation limit per TCE level. By the same time, this keeps minimal levels number as 2 in order to save memory. As the extended window now overlaps the 32bit MMIO region, this adds an area reservation to iommu_init_table(). After this change the default window size is 0x80000000000==1<<43 so devices limited to DMA mask smaller than the amount of system RAM can still use more than just 2GB of memory for DMA. This is an optimization and not a bug fix for DMA API usage. With the on-demand allocation of indirect TCE table levels enabled and 2 levels, the first TCE level size is just 1<<ceil((log2(0x7ffffffffff+1)-16)/2)=16384 TCEs or 2 system pages. Signed-off-by: Alexey Kardashevskiy <aik@ozlabs.ru> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/20190718051139.74787-5-aik@ozlabs.ru
2019-07-17 22:11:39 -07:00
extern struct iommu_table *iommu_init_table(struct iommu_table *tbl,
int nid, unsigned long res_start, unsigned long res_end);
bool iommu_table_in_use(struct iommu_table *tbl);
extern void iommu_table_reserve_pages(struct iommu_table *tbl,
unsigned long res_start, unsigned long res_end);
extern void iommu_table_clear(struct iommu_table *tbl);
powerpc/powernv/ioda2: Create bigger default window with 64k IOMMU pages At the moment we create a small window only for 32bit devices, the window maps 0..2GB of the PCI space only. For other devices we either use a sketchy bypass or hardware bypass but the former can only work if the amount of RAM is no bigger than the device's DMA mask and the latter requires devices to support at least 59bit DMA. This extends the default DMA window to the maximum size possible to allow a wider DMA mask than just 32bit. The default window size is now limited by the the iommu_table::it_map allocation bitmap which is a contiguous array, 1 bit per an IOMMU page. This increases the default IOMMU page size from hard coded 4K to the system page size to allow wider DMA masks. This increases the level number to not exceed the max order allocation limit per TCE level. By the same time, this keeps minimal levels number as 2 in order to save memory. As the extended window now overlaps the 32bit MMIO region, this adds an area reservation to iommu_init_table(). After this change the default window size is 0x80000000000==1<<43 so devices limited to DMA mask smaller than the amount of system RAM can still use more than just 2GB of memory for DMA. This is an optimization and not a bug fix for DMA API usage. With the on-demand allocation of indirect TCE table levels enabled and 2 levels, the first TCE level size is just 1<<ceil((log2(0x7ffffffffff+1)-16)/2)=16384 TCEs or 2 system pages. Signed-off-by: Alexey Kardashevskiy <aik@ozlabs.ru> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/20190718051139.74787-5-aik@ozlabs.ru
2019-07-17 22:11:39 -07:00
#define IOMMU_TABLE_GROUP_MAX_TABLES 2
powerpc/spapr: vfio: Replace iommu_table with iommu_table_group Modern IBM POWERPC systems support multiple (currently two) TCE tables per IOMMU group (a.k.a. PE). This adds a iommu_table_group container for TCE tables. Right now just one table is supported. This defines iommu_table_group struct which stores pointers to iommu_group and iommu_table(s). This replaces iommu_table with iommu_table_group where iommu_table was used to identify a group: - iommu_register_group(); - iommudata of generic iommu_group; This removes @data from iommu_table as it_table_group provides same access to pnv_ioda_pe. For IODA, instead of embedding iommu_table, the new iommu_table_group keeps pointers to those. The iommu_table structs are allocated dynamically. For P5IOC2, both iommu_table_group and iommu_table are embedded into PE struct. As there is no EEH and SRIOV support for P5IOC2, iommu_free_table() should not be called on iommu_table struct pointers so we can keep it embedded in pnv_phb::p5ioc2. For pSeries, this replaces multiple calls of kzalloc_node() with a new iommu_pseries_alloc_group() helper and stores the table group struct pointer into the pci_dn struct. For release, a iommu_table_free_group() helper is added. This moves iommu_table struct allocation from SR-IOV code to the generic DMA initialization code in pnv_pci_ioda_setup_dma_pe and pnv_pci_ioda2_setup_dma_pe as this is where DMA is actually initialized. This change is here because those lines had to be changed anyway. This should cause no behavioural change. Signed-off-by: Alexey Kardashevskiy <aik@ozlabs.ru> [aw: for the vfio related changes] Acked-by: Alex Williamson <alex.williamson@redhat.com> Reviewed-by: David Gibson <david@gibson.dropbear.id.au> Reviewed-by: Gavin Shan <gwshan@linux.vnet.ibm.com> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
2015-06-04 23:35:08 -07:00
vfio: powerpc/spapr/iommu/powernv/ioda2: Rework IOMMU ownership control This adds tce_iommu_take_ownership() and tce_iommu_release_ownership which call in a loop iommu_take_ownership()/iommu_release_ownership() for every table on the group. As there is just one now, no change in behaviour is expected. At the moment the iommu_table struct has a set_bypass() which enables/ disables DMA bypass on IODA2 PHB. This is exposed to POWERPC IOMMU code which calls this callback when external IOMMU users such as VFIO are about to get over a PHB. The set_bypass() callback is not really an iommu_table function but IOMMU/PE function. This introduces a iommu_table_group_ops struct and adds take_ownership()/release_ownership() callbacks to it which are called when an external user takes/releases control over the IOMMU. This replaces set_bypass() with ownership callbacks as it is not necessarily just bypass enabling, it can be something else/more so let's give it more generic name. The callbacks is implemented for IODA2 only. Other platforms (P5IOC2, IODA1) will use the old iommu_take_ownership/iommu_release_ownership API. The following patches will replace iommu_take_ownership/ iommu_release_ownership calls in IODA2 with full IOMMU table release/ create. As we here and touching bypass control, this removes pnv_pci_ioda2_setup_bypass_pe() as it does not do much more compared to pnv_pci_ioda2_set_bypass. This moves tce_bypass_base initialization to pnv_pci_ioda2_setup_dma_pe. Signed-off-by: Alexey Kardashevskiy <aik@ozlabs.ru> [aw: for the vfio related changes] Acked-by: Alex Williamson <alex.williamson@redhat.com> Reviewed-by: Gavin Shan <gwshan@linux.vnet.ibm.com> Reviewed-by: David Gibson <david@gibson.dropbear.id.au> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
2015-06-04 23:35:10 -07:00
struct iommu_table_group;
struct iommu_table_group_ops {
unsigned long (*get_table_size)(
__u32 page_shift,
__u64 window_size,
__u32 levels);
long (*create_table)(struct iommu_table_group *table_group,
int num,
__u32 page_shift,
__u64 window_size,
__u32 levels,
struct iommu_table **ptbl);
long (*set_window)(struct iommu_table_group *table_group,
int num,
struct iommu_table *tblnew);
long (*unset_window)(struct iommu_table_group *table_group,
int num);
vfio: powerpc/spapr/iommu/powernv/ioda2: Rework IOMMU ownership control This adds tce_iommu_take_ownership() and tce_iommu_release_ownership which call in a loop iommu_take_ownership()/iommu_release_ownership() for every table on the group. As there is just one now, no change in behaviour is expected. At the moment the iommu_table struct has a set_bypass() which enables/ disables DMA bypass on IODA2 PHB. This is exposed to POWERPC IOMMU code which calls this callback when external IOMMU users such as VFIO are about to get over a PHB. The set_bypass() callback is not really an iommu_table function but IOMMU/PE function. This introduces a iommu_table_group_ops struct and adds take_ownership()/release_ownership() callbacks to it which are called when an external user takes/releases control over the IOMMU. This replaces set_bypass() with ownership callbacks as it is not necessarily just bypass enabling, it can be something else/more so let's give it more generic name. The callbacks is implemented for IODA2 only. Other platforms (P5IOC2, IODA1) will use the old iommu_take_ownership/iommu_release_ownership API. The following patches will replace iommu_take_ownership/ iommu_release_ownership calls in IODA2 with full IOMMU table release/ create. As we here and touching bypass control, this removes pnv_pci_ioda2_setup_bypass_pe() as it does not do much more compared to pnv_pci_ioda2_set_bypass. This moves tce_bypass_base initialization to pnv_pci_ioda2_setup_dma_pe. Signed-off-by: Alexey Kardashevskiy <aik@ozlabs.ru> [aw: for the vfio related changes] Acked-by: Alex Williamson <alex.williamson@redhat.com> Reviewed-by: Gavin Shan <gwshan@linux.vnet.ibm.com> Reviewed-by: David Gibson <david@gibson.dropbear.id.au> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
2015-06-04 23:35:10 -07:00
/* Switch ownership from platform code to external user (e.g. VFIO) */
powerpc/iommu: Reimplement the iommu_table_group_ops for pSeries PPC64 IOMMU API defines iommu_table_group_ops which handles DMA windows for PEs, their ownership transfer, create/set/unset the TCE tables for the Dynamic DMA wundows(DDW). VFIOS uses these APIs for support on POWER. The commit 9d67c9433509 ("powerpc/iommu: Add "borrowing" iommu_table_group_ops") implemented partial support for this API with "borrow" mechanism wherein the DMA windows if created already by the host driver, they would be available for VFIO to use. Also, it didn't have the support to control/modify the window size or the IO page size. The current patch implements all the necessary iommu_table_group_ops APIs there by avoiding the "borrrowing". So, just the way it is on the PowerNV platform, with this patch the iommu table group ownership is transferred to the VFIO PPC subdriver, the iommu table, DMA windows creation/deletion all driven through the APIs. The pSeries uses the query-pe-dma-window, create-pe-dma-window and reset-pe-dma-window RTAS calls for DMA window creation, deletion and reset to defaul. The RTAs calls do show some minor differences to the way things are to be handled on the pSeries which are listed below. * On pSeries, the default DMA window size is "fixed" cannot be custom sized as requested by the user. For non-SRIOV VFs, It is fixed at 2GB and for SRIOV VFs, its variable sized based on the capacity assigned to it during the VF assignment to the LPAR. So, for the default DMA window alone the size if requested less than tce32_size, the smaller size is enforced using the iommu table->it_size. * The DMA start address for 32-bit window is 0, and for the 64-bit window in case of PowerNV is hardcoded to TVE select (bit 59) at 512PiB offset. This address is returned at the time of create_table() API call (even before the window is created), the subsequent set_window() call actually opens the DMA window. On pSeries, the DMA start address for 32-bit window is known from the 'ibm,dma-window' DT property. However, the 64-bit window start address is not known until the create-pe-dma RTAS call is made. So, the create_table() which returns the DMA window start address actually opens the DMA window and returns the DMA start address as returned by the Hypervisor for the create-pe-dma RTAS call. * The reset-pe-dma RTAS call resets the DMA windows and restores the default DMA window, however it does not clear the TCE table entries if there are any. In case of ownership transfer from platform domain which used direct mapping, the patch chooses remove-pe-dma instead of reset-pe for the 64-bit window intentionally so that the clear_dma_window() is called. Other than the DMA window management changes mentioned above, the patch also brings back the userspace view for the single level TCE as it existed before commit 090bad39b237a ("powerpc/powernv: Add indirect levels to it_userspace") along with the relavent refactoring. Signed-off-by: Shivaprasad G Bhat <sbhat@linux.ibm.com> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://msgid.link/171923275958.1397.907964437142542242.stgit@linux.ibm.com
2024-06-24 05:39:23 -07:00
long (*take_ownership)(struct iommu_table_group *table_group, struct device *dev);
vfio: powerpc/spapr/iommu/powernv/ioda2: Rework IOMMU ownership control This adds tce_iommu_take_ownership() and tce_iommu_release_ownership which call in a loop iommu_take_ownership()/iommu_release_ownership() for every table on the group. As there is just one now, no change in behaviour is expected. At the moment the iommu_table struct has a set_bypass() which enables/ disables DMA bypass on IODA2 PHB. This is exposed to POWERPC IOMMU code which calls this callback when external IOMMU users such as VFIO are about to get over a PHB. The set_bypass() callback is not really an iommu_table function but IOMMU/PE function. This introduces a iommu_table_group_ops struct and adds take_ownership()/release_ownership() callbacks to it which are called when an external user takes/releases control over the IOMMU. This replaces set_bypass() with ownership callbacks as it is not necessarily just bypass enabling, it can be something else/more so let's give it more generic name. The callbacks is implemented for IODA2 only. Other platforms (P5IOC2, IODA1) will use the old iommu_take_ownership/iommu_release_ownership API. The following patches will replace iommu_take_ownership/ iommu_release_ownership calls in IODA2 with full IOMMU table release/ create. As we here and touching bypass control, this removes pnv_pci_ioda2_setup_bypass_pe() as it does not do much more compared to pnv_pci_ioda2_set_bypass. This moves tce_bypass_base initialization to pnv_pci_ioda2_setup_dma_pe. Signed-off-by: Alexey Kardashevskiy <aik@ozlabs.ru> [aw: for the vfio related changes] Acked-by: Alex Williamson <alex.williamson@redhat.com> Reviewed-by: Gavin Shan <gwshan@linux.vnet.ibm.com> Reviewed-by: David Gibson <david@gibson.dropbear.id.au> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
2015-06-04 23:35:10 -07:00
/* Switch ownership from external user (e.g. VFIO) back to core */
powerpc/iommu: Reimplement the iommu_table_group_ops for pSeries PPC64 IOMMU API defines iommu_table_group_ops which handles DMA windows for PEs, their ownership transfer, create/set/unset the TCE tables for the Dynamic DMA wundows(DDW). VFIOS uses these APIs for support on POWER. The commit 9d67c9433509 ("powerpc/iommu: Add "borrowing" iommu_table_group_ops") implemented partial support for this API with "borrow" mechanism wherein the DMA windows if created already by the host driver, they would be available for VFIO to use. Also, it didn't have the support to control/modify the window size or the IO page size. The current patch implements all the necessary iommu_table_group_ops APIs there by avoiding the "borrrowing". So, just the way it is on the PowerNV platform, with this patch the iommu table group ownership is transferred to the VFIO PPC subdriver, the iommu table, DMA windows creation/deletion all driven through the APIs. The pSeries uses the query-pe-dma-window, create-pe-dma-window and reset-pe-dma-window RTAS calls for DMA window creation, deletion and reset to defaul. The RTAs calls do show some minor differences to the way things are to be handled on the pSeries which are listed below. * On pSeries, the default DMA window size is "fixed" cannot be custom sized as requested by the user. For non-SRIOV VFs, It is fixed at 2GB and for SRIOV VFs, its variable sized based on the capacity assigned to it during the VF assignment to the LPAR. So, for the default DMA window alone the size if requested less than tce32_size, the smaller size is enforced using the iommu table->it_size. * The DMA start address for 32-bit window is 0, and for the 64-bit window in case of PowerNV is hardcoded to TVE select (bit 59) at 512PiB offset. This address is returned at the time of create_table() API call (even before the window is created), the subsequent set_window() call actually opens the DMA window. On pSeries, the DMA start address for 32-bit window is known from the 'ibm,dma-window' DT property. However, the 64-bit window start address is not known until the create-pe-dma RTAS call is made. So, the create_table() which returns the DMA window start address actually opens the DMA window and returns the DMA start address as returned by the Hypervisor for the create-pe-dma RTAS call. * The reset-pe-dma RTAS call resets the DMA windows and restores the default DMA window, however it does not clear the TCE table entries if there are any. In case of ownership transfer from platform domain which used direct mapping, the patch chooses remove-pe-dma instead of reset-pe for the 64-bit window intentionally so that the clear_dma_window() is called. Other than the DMA window management changes mentioned above, the patch also brings back the userspace view for the single level TCE as it existed before commit 090bad39b237a ("powerpc/powernv: Add indirect levels to it_userspace") along with the relavent refactoring. Signed-off-by: Shivaprasad G Bhat <sbhat@linux.ibm.com> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://msgid.link/171923275958.1397.907964437142542242.stgit@linux.ibm.com
2024-06-24 05:39:23 -07:00
void (*release_ownership)(struct iommu_table_group *table_group, struct device *dev);
vfio: powerpc/spapr/iommu/powernv/ioda2: Rework IOMMU ownership control This adds tce_iommu_take_ownership() and tce_iommu_release_ownership which call in a loop iommu_take_ownership()/iommu_release_ownership() for every table on the group. As there is just one now, no change in behaviour is expected. At the moment the iommu_table struct has a set_bypass() which enables/ disables DMA bypass on IODA2 PHB. This is exposed to POWERPC IOMMU code which calls this callback when external IOMMU users such as VFIO are about to get over a PHB. The set_bypass() callback is not really an iommu_table function but IOMMU/PE function. This introduces a iommu_table_group_ops struct and adds take_ownership()/release_ownership() callbacks to it which are called when an external user takes/releases control over the IOMMU. This replaces set_bypass() with ownership callbacks as it is not necessarily just bypass enabling, it can be something else/more so let's give it more generic name. The callbacks is implemented for IODA2 only. Other platforms (P5IOC2, IODA1) will use the old iommu_take_ownership/iommu_release_ownership API. The following patches will replace iommu_take_ownership/ iommu_release_ownership calls in IODA2 with full IOMMU table release/ create. As we here and touching bypass control, this removes pnv_pci_ioda2_setup_bypass_pe() as it does not do much more compared to pnv_pci_ioda2_set_bypass. This moves tce_bypass_base initialization to pnv_pci_ioda2_setup_dma_pe. Signed-off-by: Alexey Kardashevskiy <aik@ozlabs.ru> [aw: for the vfio related changes] Acked-by: Alex Williamson <alex.williamson@redhat.com> Reviewed-by: Gavin Shan <gwshan@linux.vnet.ibm.com> Reviewed-by: David Gibson <david@gibson.dropbear.id.au> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
2015-06-04 23:35:10 -07:00
};
struct iommu_table_group_link {
struct list_head next;
struct rcu_head rcu;
struct iommu_table_group *table_group;
};
powerpc/spapr: vfio: Replace iommu_table with iommu_table_group Modern IBM POWERPC systems support multiple (currently two) TCE tables per IOMMU group (a.k.a. PE). This adds a iommu_table_group container for TCE tables. Right now just one table is supported. This defines iommu_table_group struct which stores pointers to iommu_group and iommu_table(s). This replaces iommu_table with iommu_table_group where iommu_table was used to identify a group: - iommu_register_group(); - iommudata of generic iommu_group; This removes @data from iommu_table as it_table_group provides same access to pnv_ioda_pe. For IODA, instead of embedding iommu_table, the new iommu_table_group keeps pointers to those. The iommu_table structs are allocated dynamically. For P5IOC2, both iommu_table_group and iommu_table are embedded into PE struct. As there is no EEH and SRIOV support for P5IOC2, iommu_free_table() should not be called on iommu_table struct pointers so we can keep it embedded in pnv_phb::p5ioc2. For pSeries, this replaces multiple calls of kzalloc_node() with a new iommu_pseries_alloc_group() helper and stores the table group struct pointer into the pci_dn struct. For release, a iommu_table_free_group() helper is added. This moves iommu_table struct allocation from SR-IOV code to the generic DMA initialization code in pnv_pci_ioda_setup_dma_pe and pnv_pci_ioda2_setup_dma_pe as this is where DMA is actually initialized. This change is here because those lines had to be changed anyway. This should cause no behavioural change. Signed-off-by: Alexey Kardashevskiy <aik@ozlabs.ru> [aw: for the vfio related changes] Acked-by: Alex Williamson <alex.williamson@redhat.com> Reviewed-by: David Gibson <david@gibson.dropbear.id.au> Reviewed-by: Gavin Shan <gwshan@linux.vnet.ibm.com> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
2015-06-04 23:35:08 -07:00
struct iommu_table_group {
/* IOMMU properties */
__u32 tce32_start;
__u32 tce32_size;
__u64 pgsizes; /* Bitmap of supported page sizes */
__u32 max_dynamic_windows_supported;
__u32 max_levels;
powerpc/spapr: vfio: Replace iommu_table with iommu_table_group Modern IBM POWERPC systems support multiple (currently two) TCE tables per IOMMU group (a.k.a. PE). This adds a iommu_table_group container for TCE tables. Right now just one table is supported. This defines iommu_table_group struct which stores pointers to iommu_group and iommu_table(s). This replaces iommu_table with iommu_table_group where iommu_table was used to identify a group: - iommu_register_group(); - iommudata of generic iommu_group; This removes @data from iommu_table as it_table_group provides same access to pnv_ioda_pe. For IODA, instead of embedding iommu_table, the new iommu_table_group keeps pointers to those. The iommu_table structs are allocated dynamically. For P5IOC2, both iommu_table_group and iommu_table are embedded into PE struct. As there is no EEH and SRIOV support for P5IOC2, iommu_free_table() should not be called on iommu_table struct pointers so we can keep it embedded in pnv_phb::p5ioc2. For pSeries, this replaces multiple calls of kzalloc_node() with a new iommu_pseries_alloc_group() helper and stores the table group struct pointer into the pci_dn struct. For release, a iommu_table_free_group() helper is added. This moves iommu_table struct allocation from SR-IOV code to the generic DMA initialization code in pnv_pci_ioda_setup_dma_pe and pnv_pci_ioda2_setup_dma_pe as this is where DMA is actually initialized. This change is here because those lines had to be changed anyway. This should cause no behavioural change. Signed-off-by: Alexey Kardashevskiy <aik@ozlabs.ru> [aw: for the vfio related changes] Acked-by: Alex Williamson <alex.williamson@redhat.com> Reviewed-by: David Gibson <david@gibson.dropbear.id.au> Reviewed-by: Gavin Shan <gwshan@linux.vnet.ibm.com> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
2015-06-04 23:35:08 -07:00
struct iommu_group *group;
struct iommu_table *tables[IOMMU_TABLE_GROUP_MAX_TABLES];
vfio: powerpc/spapr/iommu/powernv/ioda2: Rework IOMMU ownership control This adds tce_iommu_take_ownership() and tce_iommu_release_ownership which call in a loop iommu_take_ownership()/iommu_release_ownership() for every table on the group. As there is just one now, no change in behaviour is expected. At the moment the iommu_table struct has a set_bypass() which enables/ disables DMA bypass on IODA2 PHB. This is exposed to POWERPC IOMMU code which calls this callback when external IOMMU users such as VFIO are about to get over a PHB. The set_bypass() callback is not really an iommu_table function but IOMMU/PE function. This introduces a iommu_table_group_ops struct and adds take_ownership()/release_ownership() callbacks to it which are called when an external user takes/releases control over the IOMMU. This replaces set_bypass() with ownership callbacks as it is not necessarily just bypass enabling, it can be something else/more so let's give it more generic name. The callbacks is implemented for IODA2 only. Other platforms (P5IOC2, IODA1) will use the old iommu_take_ownership/iommu_release_ownership API. The following patches will replace iommu_take_ownership/ iommu_release_ownership calls in IODA2 with full IOMMU table release/ create. As we here and touching bypass control, this removes pnv_pci_ioda2_setup_bypass_pe() as it does not do much more compared to pnv_pci_ioda2_set_bypass. This moves tce_bypass_base initialization to pnv_pci_ioda2_setup_dma_pe. Signed-off-by: Alexey Kardashevskiy <aik@ozlabs.ru> [aw: for the vfio related changes] Acked-by: Alex Williamson <alex.williamson@redhat.com> Reviewed-by: Gavin Shan <gwshan@linux.vnet.ibm.com> Reviewed-by: David Gibson <david@gibson.dropbear.id.au> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
2015-06-04 23:35:10 -07:00
struct iommu_table_group_ops *ops;
powerpc/spapr: vfio: Replace iommu_table with iommu_table_group Modern IBM POWERPC systems support multiple (currently two) TCE tables per IOMMU group (a.k.a. PE). This adds a iommu_table_group container for TCE tables. Right now just one table is supported. This defines iommu_table_group struct which stores pointers to iommu_group and iommu_table(s). This replaces iommu_table with iommu_table_group where iommu_table was used to identify a group: - iommu_register_group(); - iommudata of generic iommu_group; This removes @data from iommu_table as it_table_group provides same access to pnv_ioda_pe. For IODA, instead of embedding iommu_table, the new iommu_table_group keeps pointers to those. The iommu_table structs are allocated dynamically. For P5IOC2, both iommu_table_group and iommu_table are embedded into PE struct. As there is no EEH and SRIOV support for P5IOC2, iommu_free_table() should not be called on iommu_table struct pointers so we can keep it embedded in pnv_phb::p5ioc2. For pSeries, this replaces multiple calls of kzalloc_node() with a new iommu_pseries_alloc_group() helper and stores the table group struct pointer into the pci_dn struct. For release, a iommu_table_free_group() helper is added. This moves iommu_table struct allocation from SR-IOV code to the generic DMA initialization code in pnv_pci_ioda_setup_dma_pe and pnv_pci_ioda2_setup_dma_pe as this is where DMA is actually initialized. This change is here because those lines had to be changed anyway. This should cause no behavioural change. Signed-off-by: Alexey Kardashevskiy <aik@ozlabs.ru> [aw: for the vfio related changes] Acked-by: Alex Williamson <alex.williamson@redhat.com> Reviewed-by: David Gibson <david@gibson.dropbear.id.au> Reviewed-by: Gavin Shan <gwshan@linux.vnet.ibm.com> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
2015-06-04 23:35:08 -07:00
};
#ifdef CONFIG_IOMMU_API
powerpc/spapr: vfio: Replace iommu_table with iommu_table_group Modern IBM POWERPC systems support multiple (currently two) TCE tables per IOMMU group (a.k.a. PE). This adds a iommu_table_group container for TCE tables. Right now just one table is supported. This defines iommu_table_group struct which stores pointers to iommu_group and iommu_table(s). This replaces iommu_table with iommu_table_group where iommu_table was used to identify a group: - iommu_register_group(); - iommudata of generic iommu_group; This removes @data from iommu_table as it_table_group provides same access to pnv_ioda_pe. For IODA, instead of embedding iommu_table, the new iommu_table_group keeps pointers to those. The iommu_table structs are allocated dynamically. For P5IOC2, both iommu_table_group and iommu_table are embedded into PE struct. As there is no EEH and SRIOV support for P5IOC2, iommu_free_table() should not be called on iommu_table struct pointers so we can keep it embedded in pnv_phb::p5ioc2. For pSeries, this replaces multiple calls of kzalloc_node() with a new iommu_pseries_alloc_group() helper and stores the table group struct pointer into the pci_dn struct. For release, a iommu_table_free_group() helper is added. This moves iommu_table struct allocation from SR-IOV code to the generic DMA initialization code in pnv_pci_ioda_setup_dma_pe and pnv_pci_ioda2_setup_dma_pe as this is where DMA is actually initialized. This change is here because those lines had to be changed anyway. This should cause no behavioural change. Signed-off-by: Alexey Kardashevskiy <aik@ozlabs.ru> [aw: for the vfio related changes] Acked-by: Alex Williamson <alex.williamson@redhat.com> Reviewed-by: David Gibson <david@gibson.dropbear.id.au> Reviewed-by: Gavin Shan <gwshan@linux.vnet.ibm.com> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
2015-06-04 23:35:08 -07:00
extern void iommu_register_group(struct iommu_table_group *table_group,
int pci_domain_number, unsigned long pe_num);
powerpc/powernv/pseries: Rework device adding to IOMMU groups The powernv platform registers IOMMU groups and adds devices to them from the pci_controller_ops::setup_bridge() hook except one case when virtual functions (SRIOV VFs) are added from a bus notifier. The pseries platform registers IOMMU groups from the pci_controller_ops::dma_bus_setup() hook and adds devices from the pci_controller_ops::dma_dev_setup() hook. The very same bus notifier used for powernv does not add devices for pseries though as __of_scan_bus() adds devices first, then it does the bus/dev DMA setup. Both platforms use iommu_add_device() which takes a device and expects it to have a valid IOMMU table struct with an iommu_table_group pointer which in turn points the iommu_group struct (which represents an IOMMU group). Although the helper seems easy to use, it relies on some pre-existing device configuration and associated data structures which it does not really need. This simplifies iommu_add_device() to take the table_group pointer directly. Pseries already has a table_group pointer handy and the bus notified is not used anyway. For powernv, this copies the existing bus notifier, makes it work for powernv only which means an easy way of getting to the table_group pointer. This was tested on VFs but should also support physical PCI hotplug. Since iommu_add_device() receives the table_group pointer directly, pseries does not do TCE cache invalidation (the hypervisor does) nor allow multiple groups per a VFIO container (in other words sharing an IOMMU table between partitionable endpoints), this removes iommu_table_group_link from pseries. Signed-off-by: Alexey Kardashevskiy <aik@ozlabs.ru> Reviewed-by: David Gibson <david@gibson.dropbear.id.au> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
2018-12-19 01:52:21 -07:00
extern int iommu_add_device(struct iommu_table_group *table_group,
struct device *dev);
extern long iommu_tce_xchg(struct mm_struct *mm, struct iommu_table *tbl,
unsigned long entry, unsigned long *hpa,
enum dma_data_direction *direction);
extern long iommu_tce_xchg_no_kill(struct mm_struct *mm,
struct iommu_table *tbl,
unsigned long entry, unsigned long *hpa,
enum dma_data_direction *direction);
extern void iommu_tce_kill(struct iommu_table *tbl,
unsigned long entry, unsigned long pages);
int dev_has_iommu_table(struct device *dev, void *data);
powerpc/iommu: Add "borrowing" iommu_table_group_ops PPC64 IOMMU API defines iommu_table_group_ops which handles DMA windows for PEs: control the ownership, create/set/unset a table the hardware for dynamic DMA windows (DDW). VFIO uses the API to implement support on POWER. So far only PowerNV IODA2 (POWER8 and newer machines) implemented this and other cases (POWER7 or nested KVM) did not and instead reused existing iommu_table structs. This means 1) no DDW 2) ownership transfer is done directly in the VFIO SPAPR TCE driver. Soon POWER is going to get its own iommu_ops and ownership control is going to move there. This implements spapr_tce_table_group_ops which borrows iommu_table tables. The upside is that VFIO needs to know less about POWER. The new ops returns the existing table from create_table() and only checks if the same window is already set. This is only going to work if the default DMA window starts table_group.tce32_start and as big as pe->table_group.tce32_size (not the case for IODA2+ PowerNV). This changes iommu_table_group_ops::take_ownership() to return an error if borrowing a table failed. This should not cause any visible change in behavior for PowerNV. pSeries was not that well tested/supported anyway. Signed-off-by: Alexey Kardashevskiy <aik@ozlabs.ru> Signed-off-by: Timothy Pearson <tpearson@raptorengineering.com> Acked-by: Alex Williamson <alex.williamson@redhat.com> [mpe: Fix CONFIG_IOMMU_API=n build (skiroot_defconfig), & formatting] Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://msgid.link/525438831.16998517.1678123820075.JavaMail.zimbra@raptorengineeringinc.com
2023-03-06 10:30:20 -07:00
#else
powerpc/spapr: vfio: Replace iommu_table with iommu_table_group Modern IBM POWERPC systems support multiple (currently two) TCE tables per IOMMU group (a.k.a. PE). This adds a iommu_table_group container for TCE tables. Right now just one table is supported. This defines iommu_table_group struct which stores pointers to iommu_group and iommu_table(s). This replaces iommu_table with iommu_table_group where iommu_table was used to identify a group: - iommu_register_group(); - iommudata of generic iommu_group; This removes @data from iommu_table as it_table_group provides same access to pnv_ioda_pe. For IODA, instead of embedding iommu_table, the new iommu_table_group keeps pointers to those. The iommu_table structs are allocated dynamically. For P5IOC2, both iommu_table_group and iommu_table are embedded into PE struct. As there is no EEH and SRIOV support for P5IOC2, iommu_free_table() should not be called on iommu_table struct pointers so we can keep it embedded in pnv_phb::p5ioc2. For pSeries, this replaces multiple calls of kzalloc_node() with a new iommu_pseries_alloc_group() helper and stores the table group struct pointer into the pci_dn struct. For release, a iommu_table_free_group() helper is added. This moves iommu_table struct allocation from SR-IOV code to the generic DMA initialization code in pnv_pci_ioda_setup_dma_pe and pnv_pci_ioda2_setup_dma_pe as this is where DMA is actually initialized. This change is here because those lines had to be changed anyway. This should cause no behavioural change. Signed-off-by: Alexey Kardashevskiy <aik@ozlabs.ru> [aw: for the vfio related changes] Acked-by: Alex Williamson <alex.williamson@redhat.com> Reviewed-by: David Gibson <david@gibson.dropbear.id.au> Reviewed-by: Gavin Shan <gwshan@linux.vnet.ibm.com> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
2015-06-04 23:35:08 -07:00
static inline void iommu_register_group(struct iommu_table_group *table_group,
int pci_domain_number,
unsigned long pe_num)
{
}
powerpc/powernv/pseries: Rework device adding to IOMMU groups The powernv platform registers IOMMU groups and adds devices to them from the pci_controller_ops::setup_bridge() hook except one case when virtual functions (SRIOV VFs) are added from a bus notifier. The pseries platform registers IOMMU groups from the pci_controller_ops::dma_bus_setup() hook and adds devices from the pci_controller_ops::dma_dev_setup() hook. The very same bus notifier used for powernv does not add devices for pseries though as __of_scan_bus() adds devices first, then it does the bus/dev DMA setup. Both platforms use iommu_add_device() which takes a device and expects it to have a valid IOMMU table struct with an iommu_table_group pointer which in turn points the iommu_group struct (which represents an IOMMU group). Although the helper seems easy to use, it relies on some pre-existing device configuration and associated data structures which it does not really need. This simplifies iommu_add_device() to take the table_group pointer directly. Pseries already has a table_group pointer handy and the bus notified is not used anyway. For powernv, this copies the existing bus notifier, makes it work for powernv only which means an easy way of getting to the table_group pointer. This was tested on VFs but should also support physical PCI hotplug. Since iommu_add_device() receives the table_group pointer directly, pseries does not do TCE cache invalidation (the hypervisor does) nor allow multiple groups per a VFIO container (in other words sharing an IOMMU table between partitionable endpoints), this removes iommu_table_group_link from pseries. Signed-off-by: Alexey Kardashevskiy <aik@ozlabs.ru> Reviewed-by: David Gibson <david@gibson.dropbear.id.au> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
2018-12-19 01:52:21 -07:00
static inline int iommu_add_device(struct iommu_table_group *table_group,
struct device *dev)
{
return 0;
}
static inline int dev_has_iommu_table(struct device *dev, void *data)
{
return 0;
}
#endif /* !CONFIG_IOMMU_API */
u64 dma_iommu_get_required_mask(struct device *dev);
#else
static inline void *get_iommu_table_base(struct device *dev)
{
return NULL;
}
static inline int dma_iommu_dma_supported(struct device *dev, u64 mask)
{
return 0;
}
#endif /* CONFIG_PPC64 */
extern int ppc_iommu_map_sg(struct device *dev, struct iommu_table *tbl,
struct scatterlist *sglist, int nelems,
unsigned long mask,
enum dma_data_direction direction,
dma-mapping: use unsigned long for dma_attrs The dma-mapping core and the implementations do not change the DMA attributes passed by pointer. Thus the pointer can point to const data. However the attributes do not have to be a bitfield. Instead unsigned long will do fine: 1. This is just simpler. Both in terms of reading the code and setting attributes. Instead of initializing local attributes on the stack and passing pointer to it to dma_set_attr(), just set the bits. 2. It brings safeness and checking for const correctness because the attributes are passed by value. Semantic patches for this change (at least most of them): virtual patch virtual context @r@ identifier f, attrs; @@ f(..., - struct dma_attrs *attrs + unsigned long attrs , ...) { ... } @@ identifier r.f; @@ f(..., - NULL + 0 ) and // Options: --all-includes virtual patch virtual context @r@ identifier f, attrs; type t; @@ t f(..., struct dma_attrs *attrs); @@ identifier r.f; @@ f(..., - NULL + 0 ) Link: http://lkml.kernel.org/r/1468399300-5399-2-git-send-email-k.kozlowski@samsung.com Signed-off-by: Krzysztof Kozlowski <k.kozlowski@samsung.com> Acked-by: Vineet Gupta <vgupta@synopsys.com> Acked-by: Robin Murphy <robin.murphy@arm.com> Acked-by: Hans-Christian Noren Egtvedt <egtvedt@samfundet.no> Acked-by: Mark Salter <msalter@redhat.com> [c6x] Acked-by: Jesper Nilsson <jesper.nilsson@axis.com> [cris] Acked-by: Daniel Vetter <daniel.vetter@ffwll.ch> [drm] Reviewed-by: Bart Van Assche <bart.vanassche@sandisk.com> Acked-by: Joerg Roedel <jroedel@suse.de> [iommu] Acked-by: Fabien Dessenne <fabien.dessenne@st.com> [bdisp] Reviewed-by: Marek Szyprowski <m.szyprowski@samsung.com> [vb2-core] Acked-by: David Vrabel <david.vrabel@citrix.com> [xen] Acked-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com> [xen swiotlb] Acked-by: Joerg Roedel <jroedel@suse.de> [iommu] Acked-by: Richard Kuo <rkuo@codeaurora.org> [hexagon] Acked-by: Geert Uytterhoeven <geert@linux-m68k.org> [m68k] Acked-by: Gerald Schaefer <gerald.schaefer@de.ibm.com> [s390] Acked-by: Bjorn Andersson <bjorn.andersson@linaro.org> Acked-by: Hans-Christian Noren Egtvedt <egtvedt@samfundet.no> [avr32] Acked-by: Vineet Gupta <vgupta@synopsys.com> [arc] Acked-by: Robin Murphy <robin.murphy@arm.com> [arm64 and dma-iommu] Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-08-03 13:46:00 -07:00
unsigned long attrs);
extern void ppc_iommu_unmap_sg(struct iommu_table *tbl,
struct scatterlist *sglist,
int nelems,
enum dma_data_direction direction,
dma-mapping: use unsigned long for dma_attrs The dma-mapping core and the implementations do not change the DMA attributes passed by pointer. Thus the pointer can point to const data. However the attributes do not have to be a bitfield. Instead unsigned long will do fine: 1. This is just simpler. Both in terms of reading the code and setting attributes. Instead of initializing local attributes on the stack and passing pointer to it to dma_set_attr(), just set the bits. 2. It brings safeness and checking for const correctness because the attributes are passed by value. Semantic patches for this change (at least most of them): virtual patch virtual context @r@ identifier f, attrs; @@ f(..., - struct dma_attrs *attrs + unsigned long attrs , ...) { ... } @@ identifier r.f; @@ f(..., - NULL + 0 ) and // Options: --all-includes virtual patch virtual context @r@ identifier f, attrs; type t; @@ t f(..., struct dma_attrs *attrs); @@ identifier r.f; @@ f(..., - NULL + 0 ) Link: http://lkml.kernel.org/r/1468399300-5399-2-git-send-email-k.kozlowski@samsung.com Signed-off-by: Krzysztof Kozlowski <k.kozlowski@samsung.com> Acked-by: Vineet Gupta <vgupta@synopsys.com> Acked-by: Robin Murphy <robin.murphy@arm.com> Acked-by: Hans-Christian Noren Egtvedt <egtvedt@samfundet.no> Acked-by: Mark Salter <msalter@redhat.com> [c6x] Acked-by: Jesper Nilsson <jesper.nilsson@axis.com> [cris] Acked-by: Daniel Vetter <daniel.vetter@ffwll.ch> [drm] Reviewed-by: Bart Van Assche <bart.vanassche@sandisk.com> Acked-by: Joerg Roedel <jroedel@suse.de> [iommu] Acked-by: Fabien Dessenne <fabien.dessenne@st.com> [bdisp] Reviewed-by: Marek Szyprowski <m.szyprowski@samsung.com> [vb2-core] Acked-by: David Vrabel <david.vrabel@citrix.com> [xen] Acked-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com> [xen swiotlb] Acked-by: Joerg Roedel <jroedel@suse.de> [iommu] Acked-by: Richard Kuo <rkuo@codeaurora.org> [hexagon] Acked-by: Geert Uytterhoeven <geert@linux-m68k.org> [m68k] Acked-by: Gerald Schaefer <gerald.schaefer@de.ibm.com> [s390] Acked-by: Bjorn Andersson <bjorn.andersson@linaro.org> Acked-by: Hans-Christian Noren Egtvedt <egtvedt@samfundet.no> [avr32] Acked-by: Vineet Gupta <vgupta@synopsys.com> [arc] Acked-by: Robin Murphy <robin.murphy@arm.com> [arm64 and dma-iommu] Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-08-03 13:46:00 -07:00
unsigned long attrs);
extern void *iommu_alloc_coherent(struct device *dev, struct iommu_table *tbl,
size_t size, dma_addr_t *dma_handle,
unsigned long mask, gfp_t flag, int node);
extern void iommu_free_coherent(struct iommu_table *tbl, size_t size,
void *vaddr, dma_addr_t dma_handle);
extern dma_addr_t iommu_map_page(struct device *dev, struct iommu_table *tbl,
struct page *page, unsigned long offset,
size_t size, unsigned long mask,
enum dma_data_direction direction,
dma-mapping: use unsigned long for dma_attrs The dma-mapping core and the implementations do not change the DMA attributes passed by pointer. Thus the pointer can point to const data. However the attributes do not have to be a bitfield. Instead unsigned long will do fine: 1. This is just simpler. Both in terms of reading the code and setting attributes. Instead of initializing local attributes on the stack and passing pointer to it to dma_set_attr(), just set the bits. 2. It brings safeness and checking for const correctness because the attributes are passed by value. Semantic patches for this change (at least most of them): virtual patch virtual context @r@ identifier f, attrs; @@ f(..., - struct dma_attrs *attrs + unsigned long attrs , ...) { ... } @@ identifier r.f; @@ f(..., - NULL + 0 ) and // Options: --all-includes virtual patch virtual context @r@ identifier f, attrs; type t; @@ t f(..., struct dma_attrs *attrs); @@ identifier r.f; @@ f(..., - NULL + 0 ) Link: http://lkml.kernel.org/r/1468399300-5399-2-git-send-email-k.kozlowski@samsung.com Signed-off-by: Krzysztof Kozlowski <k.kozlowski@samsung.com> Acked-by: Vineet Gupta <vgupta@synopsys.com> Acked-by: Robin Murphy <robin.murphy@arm.com> Acked-by: Hans-Christian Noren Egtvedt <egtvedt@samfundet.no> Acked-by: Mark Salter <msalter@redhat.com> [c6x] Acked-by: Jesper Nilsson <jesper.nilsson@axis.com> [cris] Acked-by: Daniel Vetter <daniel.vetter@ffwll.ch> [drm] Reviewed-by: Bart Van Assche <bart.vanassche@sandisk.com> Acked-by: Joerg Roedel <jroedel@suse.de> [iommu] Acked-by: Fabien Dessenne <fabien.dessenne@st.com> [bdisp] Reviewed-by: Marek Szyprowski <m.szyprowski@samsung.com> [vb2-core] Acked-by: David Vrabel <david.vrabel@citrix.com> [xen] Acked-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com> [xen swiotlb] Acked-by: Joerg Roedel <jroedel@suse.de> [iommu] Acked-by: Richard Kuo <rkuo@codeaurora.org> [hexagon] Acked-by: Geert Uytterhoeven <geert@linux-m68k.org> [m68k] Acked-by: Gerald Schaefer <gerald.schaefer@de.ibm.com> [s390] Acked-by: Bjorn Andersson <bjorn.andersson@linaro.org> Acked-by: Hans-Christian Noren Egtvedt <egtvedt@samfundet.no> [avr32] Acked-by: Vineet Gupta <vgupta@synopsys.com> [arc] Acked-by: Robin Murphy <robin.murphy@arm.com> [arm64 and dma-iommu] Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-08-03 13:46:00 -07:00
unsigned long attrs);
extern void iommu_unmap_page(struct iommu_table *tbl, dma_addr_t dma_handle,
size_t size, enum dma_data_direction direction,
dma-mapping: use unsigned long for dma_attrs The dma-mapping core and the implementations do not change the DMA attributes passed by pointer. Thus the pointer can point to const data. However the attributes do not have to be a bitfield. Instead unsigned long will do fine: 1. This is just simpler. Both in terms of reading the code and setting attributes. Instead of initializing local attributes on the stack and passing pointer to it to dma_set_attr(), just set the bits. 2. It brings safeness and checking for const correctness because the attributes are passed by value. Semantic patches for this change (at least most of them): virtual patch virtual context @r@ identifier f, attrs; @@ f(..., - struct dma_attrs *attrs + unsigned long attrs , ...) { ... } @@ identifier r.f; @@ f(..., - NULL + 0 ) and // Options: --all-includes virtual patch virtual context @r@ identifier f, attrs; type t; @@ t f(..., struct dma_attrs *attrs); @@ identifier r.f; @@ f(..., - NULL + 0 ) Link: http://lkml.kernel.org/r/1468399300-5399-2-git-send-email-k.kozlowski@samsung.com Signed-off-by: Krzysztof Kozlowski <k.kozlowski@samsung.com> Acked-by: Vineet Gupta <vgupta@synopsys.com> Acked-by: Robin Murphy <robin.murphy@arm.com> Acked-by: Hans-Christian Noren Egtvedt <egtvedt@samfundet.no> Acked-by: Mark Salter <msalter@redhat.com> [c6x] Acked-by: Jesper Nilsson <jesper.nilsson@axis.com> [cris] Acked-by: Daniel Vetter <daniel.vetter@ffwll.ch> [drm] Reviewed-by: Bart Van Assche <bart.vanassche@sandisk.com> Acked-by: Joerg Roedel <jroedel@suse.de> [iommu] Acked-by: Fabien Dessenne <fabien.dessenne@st.com> [bdisp] Reviewed-by: Marek Szyprowski <m.szyprowski@samsung.com> [vb2-core] Acked-by: David Vrabel <david.vrabel@citrix.com> [xen] Acked-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com> [xen swiotlb] Acked-by: Joerg Roedel <jroedel@suse.de> [iommu] Acked-by: Richard Kuo <rkuo@codeaurora.org> [hexagon] Acked-by: Geert Uytterhoeven <geert@linux-m68k.org> [m68k] Acked-by: Gerald Schaefer <gerald.schaefer@de.ibm.com> [s390] Acked-by: Bjorn Andersson <bjorn.andersson@linaro.org> Acked-by: Hans-Christian Noren Egtvedt <egtvedt@samfundet.no> [avr32] Acked-by: Vineet Gupta <vgupta@synopsys.com> [arc] Acked-by: Robin Murphy <robin.murphy@arm.com> [arm64 and dma-iommu] Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-08-03 13:46:00 -07:00
unsigned long attrs);
void __init iommu_init_early_pSeries(void);
extern void iommu_init_early_dart(struct pci_controller_ops *controller_ops);
extern void iommu_init_early_pasemi(void);
#if defined(CONFIG_PPC64) && defined(CONFIG_PM)
static inline void iommu_restore(void)
{
if (ppc_md.iommu_restore)
ppc_md.iommu_restore();
}
#endif
/* The API to support IOMMU operations for VFIO */
extern int iommu_tce_check_ioba(unsigned long page_shift,
unsigned long offset, unsigned long size,
unsigned long ioba, unsigned long npages);
extern int iommu_tce_check_gpa(unsigned long page_shift,
unsigned long gpa);
#define iommu_tce_clear_param_check(tbl, ioba, tce_value, npages) \
(iommu_tce_check_ioba((tbl)->it_page_shift, \
(tbl)->it_offset, (tbl)->it_size, \
(ioba), (npages)) || (tce_value))
#define iommu_tce_put_param_check(tbl, ioba, gpa) \
(iommu_tce_check_ioba((tbl)->it_page_shift, \
(tbl)->it_offset, (tbl)->it_size, \
(ioba), 1) || \
iommu_tce_check_gpa((tbl)->it_page_shift, (gpa)))
extern void iommu_flush_tce(struct iommu_table *tbl);
extern enum dma_data_direction iommu_tce_direction(unsigned long tce);
extern unsigned long iommu_direction_to_tce_perm(enum dma_data_direction dir);
#ifdef CONFIG_PPC_CELL_NATIVE
extern bool iommu_fixed_is_weak;
#else
#define iommu_fixed_is_weak false
#endif
extern const struct dma_map_ops dma_iommu_ops;
#endif /* __KERNEL__ */
#endif /* _ASM_IOMMU_H */