The goal for this series is to avoid device private memory TLB invalidations when migrating a range of addresses from system memory to device private memory and some of those pages have already been migrated. The approach taken is to introduce a new mmu notifier invalidation event type and use that in the device driver to skip invalidation callbacks from migrate_vma_setup(). The device driver is also then expected to handle device MMU invalidations as part of the migrate_vma_setup(), migrate_vma_pages(), migrate_vma_finalize() process. Note that this is opt-in. A device driver can simply invalidate its MMU in the mmu notifier callback and not handle MMU invalidations in the migration sequence.
This series is based on linux-5.8.0-rc4 and the patches I sent for ("mm/hmm/nouveau: add PMD system memory mapping") https://lore.kernel.org/linux-mm/20200701225352.9649-1-rcampbell@nvidia.com There are no logical dependencies, but there would be merge conflicts which could be resolved if this were to be applied before the other series.
Also, this replaces the need for the following two patches I sent: ("mm: fix migrate_vma_setup() src_owner and normal pages") https://lore.kernel.org/linux-mm/20200622222008.9971-1-rcampbell@nvidia.com ("nouveau: fix mixed normal and device private page migration") https://lore.kernel.org/lkml/20200622233854.10889-3-rcampbell@nvidia.com
Ralph Campbell (5): nouveau: fix storing invalid ptes mm/migrate: add a direction parameter to migrate_vma mm/notifier: add migration invalidation type nouveau/svm: use the new migration invalidation mm/hmm/test: use the new migration invalidation
arch/powerpc/kvm/book3s_hv_uvmem.c | 2 ++ drivers/gpu/drm/nouveau/nouveau_dmem.c | 13 ++++++-- drivers/gpu/drm/nouveau/nouveau_svm.c | 10 +++++- drivers/gpu/drm/nouveau/nouveau_svm.h | 1 + .../drm/nouveau/nvkm/subdev/mmu/vmmgp100.c | 13 +++++--- include/linux/migrate.h | 12 +++++-- include/linux/mmu_notifier.h | 7 ++++ lib/test_hmm.c | 33 +++++++++++-------- mm/migrate.c | 13 ++++++-- 9 files changed, 77 insertions(+), 27 deletions(-)
When migrating a range of system memory to device private memory, some of the pages in the address range may not be migrating. In this case, the non migrating pages won't have a new GPU MMU entry to store but the nvif_object_ioctl() NVIF_VMM_V0_PFNMAP method doesn't check the input and stores a bad valid GPU page table entry. Fix this by skipping the invalid input PTEs when updating the GPU page tables.
Signed-off-by: Ralph Campbell rcampbell@nvidia.com --- drivers/gpu/drm/nouveau/nvkm/subdev/mmu/vmmgp100.c | 13 +++++++++---- 1 file changed, 9 insertions(+), 4 deletions(-)
diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/vmmgp100.c b/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/vmmgp100.c index ed37fddd063f..7eabe9fe0d2b 100644 --- a/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/vmmgp100.c +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/mmu/vmmgp100.c @@ -79,8 +79,12 @@ gp100_vmm_pgt_pfn(struct nvkm_vmm *vmm, struct nvkm_mmu_pt *pt, dma_addr_t addr;
nvkm_kmap(pt->memory); - while (ptes--) { + for (; ptes; ptes--, map->pfn++) { u64 data = 0; + + if (!(*map->pfn & NVKM_VMM_PFN_V)) + continue; + if (!(*map->pfn & NVKM_VMM_PFN_W)) data |= BIT_ULL(6); /* RO. */
@@ -100,7 +104,6 @@ gp100_vmm_pgt_pfn(struct nvkm_vmm *vmm, struct nvkm_mmu_pt *pt, }
VMM_WO064(pt, vmm, ptei++ * 8, data); - map->pfn++; } nvkm_done(pt->memory); } @@ -310,9 +313,12 @@ gp100_vmm_pd0_pfn(struct nvkm_vmm *vmm, struct nvkm_mmu_pt *pt, dma_addr_t addr;
nvkm_kmap(pt->memory); - while (ptes--) { + for (; ptes; ptes--, map->pfn++) { u64 data = 0;
+ if (!(*map->pfn & NVKM_VMM_PFN_V)) + continue; + if (!(*map->pfn & NVKM_VMM_PFN_W)) data |= BIT_ULL(6); /* RO. */
@@ -332,7 +338,6 @@ gp100_vmm_pd0_pfn(struct nvkm_vmm *vmm, struct nvkm_mmu_pt *pt, }
VMM_WO064(pt, vmm, ptei++ * 16, data); - map->pfn++; } nvkm_done(pt->memory); }
The src_owner field in struct migrate_vma is being used for two purposes, it implies the direction of the migration and it identifies device private pages owned by the caller. Split this into separate parameters so the src_owner field can be used just to identify device private pages owned by the caller of migrate_vma_setup().
Signed-off-by: Ralph Campbell rcampbell@nvidia.com --- arch/powerpc/kvm/book3s_hv_uvmem.c | 2 ++ drivers/gpu/drm/nouveau/nouveau_dmem.c | 2 ++ include/linux/migrate.h | 12 +++++++++--- lib/test_hmm.c | 2 ++ mm/migrate.c | 5 +++-- 5 files changed, 18 insertions(+), 5 deletions(-)
diff --git a/arch/powerpc/kvm/book3s_hv_uvmem.c b/arch/powerpc/kvm/book3s_hv_uvmem.c index 09d8119024db..acbf14cd2d72 100644 --- a/arch/powerpc/kvm/book3s_hv_uvmem.c +++ b/arch/powerpc/kvm/book3s_hv_uvmem.c @@ -400,6 +400,7 @@ kvmppc_svm_page_in(struct vm_area_struct *vma, unsigned long start, mig.end = end; mig.src = &src_pfn; mig.dst = &dst_pfn; + mig.dir = MIGRATE_VMA_FROM_SYSTEM;
/* * We come here with mmap_lock write lock held just for @@ -578,6 +579,7 @@ kvmppc_svm_page_out(struct vm_area_struct *vma, unsigned long start, mig.src = &src_pfn; mig.dst = &dst_pfn; mig.src_owner = &kvmppc_uvmem_pgmap; + mig.dir = MIGRATE_VMA_FROM_DEVICE_PRIVATE;
mutex_lock(&kvm->arch.uvmem_lock); /* The requested page is already paged-out, nothing to do */ diff --git a/drivers/gpu/drm/nouveau/nouveau_dmem.c b/drivers/gpu/drm/nouveau/nouveau_dmem.c index cc9993837508..762e0168bc8d 100644 --- a/drivers/gpu/drm/nouveau/nouveau_dmem.c +++ b/drivers/gpu/drm/nouveau/nouveau_dmem.c @@ -183,6 +183,7 @@ static vm_fault_t nouveau_dmem_migrate_to_ram(struct vm_fault *vmf) .src = &src, .dst = &dst, .src_owner = drm->dev, + .dir = MIGRATE_VMA_FROM_DEVICE_PRIVATE, };
/* @@ -615,6 +616,7 @@ nouveau_dmem_migrate_vma(struct nouveau_drm *drm, struct migrate_vma args = { .vma = vma, .start = start, + .dir = MIGRATE_VMA_FROM_SYSTEM, }; unsigned long i; u64 *pfns; diff --git a/include/linux/migrate.h b/include/linux/migrate.h index 3e546cbf03dd..620f2235d7d4 100644 --- a/include/linux/migrate.h +++ b/include/linux/migrate.h @@ -180,6 +180,11 @@ static inline unsigned long migrate_pfn(unsigned long pfn) return (pfn << MIGRATE_PFN_SHIFT) | MIGRATE_PFN_VALID; }
+enum migrate_vma_direction { + MIGRATE_VMA_FROM_SYSTEM, + MIGRATE_VMA_FROM_DEVICE_PRIVATE, +}; + struct migrate_vma { struct vm_area_struct *vma; /* @@ -199,11 +204,12 @@ struct migrate_vma {
/* * Set to the owner value also stored in page->pgmap->owner for - * migrating out of device private memory. If set only device - * private pages with this owner are migrated. If not set - * device private pages are not migrated at all. + * migrating device private memory. The direction also needs to + * be set to MIGRATE_VMA_FROM_DEVICE_PRIVATE. */ void *src_owner; + + enum migrate_vma_direction dir; };
int migrate_vma_setup(struct migrate_vma *args); diff --git a/lib/test_hmm.c b/lib/test_hmm.c index 9aa577afc269..1bd60cfb5a25 100644 --- a/lib/test_hmm.c +++ b/lib/test_hmm.c @@ -703,6 +703,7 @@ static int dmirror_migrate(struct dmirror *dmirror, args.start = addr; args.end = next; args.src_owner = NULL; + args.dir = MIGRATE_VMA_FROM_SYSTEM; ret = migrate_vma_setup(&args); if (ret) goto out; @@ -1054,6 +1055,7 @@ static vm_fault_t dmirror_devmem_fault(struct vm_fault *vmf) args.src = &src_pfns; args.dst = &dst_pfns; args.src_owner = dmirror->mdevice; + args.dir = MIGRATE_VMA_FROM_DEVICE_PRIVATE;
if (migrate_vma_setup(&args)) return VM_FAULT_SIGBUS; diff --git a/mm/migrate.c b/mm/migrate.c index f37729673558..2bbc5c4c672e 100644 --- a/mm/migrate.c +++ b/mm/migrate.c @@ -2287,7 +2287,8 @@ static int migrate_vma_collect_pmd(pmd_t *pmdp, goto next;
page = device_private_entry_to_page(entry); - if (page->pgmap->owner != migrate->src_owner) + if (migrate->dir != MIGRATE_VMA_FROM_DEVICE_PRIVATE || + page->pgmap->owner != migrate->src_owner) goto next;
mpfn = migrate_pfn(page_to_pfn(page)) | @@ -2295,7 +2296,7 @@ static int migrate_vma_collect_pmd(pmd_t *pmdp, if (is_write_device_private_entry(entry)) mpfn |= MIGRATE_PFN_WRITE; } else { - if (migrate->src_owner) + if (migrate->dir != MIGRATE_VMA_FROM_SYSTEM) goto next; pfn = pte_pfn(pte); if (is_zero_pfn(pfn)) {
On Mon, Jul 06, 2020 at 03:23:44PM -0700, Ralph Campbell wrote:
The src_owner field in struct migrate_vma is being used for two purposes, it implies the direction of the migration and it identifies device private pages owned by the caller. Split this into separate parameters so the src_owner field can be used just to identify device private pages owned by the caller of migrate_vma_setup().
Signed-off-by: Ralph Campbell rcampbell@nvidia.com
arch/powerpc/kvm/book3s_hv_uvmem.c | 2 ++ drivers/gpu/drm/nouveau/nouveau_dmem.c | 2 ++ include/linux/migrate.h | 12 +++++++++--- lib/test_hmm.c | 2 ++ mm/migrate.c | 5 +++-- 5 files changed, 18 insertions(+), 5 deletions(-)
diff --git a/arch/powerpc/kvm/book3s_hv_uvmem.c b/arch/powerpc/kvm/book3s_hv_uvmem.c index 09d8119024db..acbf14cd2d72 100644 --- a/arch/powerpc/kvm/book3s_hv_uvmem.c +++ b/arch/powerpc/kvm/book3s_hv_uvmem.c @@ -400,6 +400,7 @@ kvmppc_svm_page_in(struct vm_area_struct *vma, unsigned long start, mig.end = end; mig.src = &src_pfn; mig.dst = &dst_pfn;
- mig.dir = MIGRATE_VMA_FROM_SYSTEM;
/* * We come here with mmap_lock write lock held just for @@ -578,6 +579,7 @@ kvmppc_svm_page_out(struct vm_area_struct *vma, unsigned long start, mig.src = &src_pfn; mig.dst = &dst_pfn; mig.src_owner = &kvmppc_uvmem_pgmap;
- mig.dir = MIGRATE_VMA_FROM_DEVICE_PRIVATE;
Reviewed-by: Bharata B Rao bharata@linux.ibm.com
for the above kvmppc change.
Currently migrate_vma_setup() calls mmu_notifier_invalidate_range_start() which flushes all device private page mappings whether or not a page is being migrated to/from device private memory. In order to not disrupt device mappings that are not being migrated, shift the responsibility for clearing device private mappings to the device driver and leave CPU page table unmapping handled by migrate_vma_setup(). To support this, the caller of migrate_vma_setup() should always set struct migrate_vma::src_owner to a non NULL value that matches the device private page->pgmap->owner. This value is then passed to the struct mmu_notifier_range with a new event type which the driver's invalidation function can use to avoid device MMU invalidations.
Signed-off-by: Ralph Campbell rcampbell@nvidia.com --- include/linux/mmu_notifier.h | 7 +++++++ mm/migrate.c | 8 +++++++- 2 files changed, 14 insertions(+), 1 deletion(-)
diff --git a/include/linux/mmu_notifier.h b/include/linux/mmu_notifier.h index fc68f3570e19..bd0b34dbe4de 100644 --- a/include/linux/mmu_notifier.h +++ b/include/linux/mmu_notifier.h @@ -38,6 +38,10 @@ struct mmu_interval_notifier; * * @MMU_NOTIFY_RELEASE: used during mmu_interval_notifier invalidate to signal * that the mm refcount is zero and the range is no longer accessible. + * + * @MMU_NOTIFY_MIGRATE: used during migrate_vma_collect() invalidate to signal + * a device driver to possibly ignore the invalidation if the src_own + * field matches. */ enum mmu_notifier_event { MMU_NOTIFY_UNMAP = 0, @@ -46,6 +50,7 @@ enum mmu_notifier_event { MMU_NOTIFY_PROTECTION_PAGE, MMU_NOTIFY_SOFT_DIRTY, MMU_NOTIFY_RELEASE, + MMU_NOTIFY_MIGRATE, };
#define MMU_NOTIFIER_RANGE_BLOCKABLE (1 << 0) @@ -264,6 +269,7 @@ struct mmu_notifier_range { unsigned long end; unsigned flags; enum mmu_notifier_event event; + void *data; };
static inline int mm_has_notifiers(struct mm_struct *mm) @@ -513,6 +519,7 @@ static inline void mmu_notifier_range_init(struct mmu_notifier_range *range, range->start = start; range->end = end; range->flags = flags; + range->data = NULL; }
#define ptep_clear_flush_young_notify(__vma, __address, __ptep) \ diff --git a/mm/migrate.c b/mm/migrate.c index 2bbc5c4c672e..62270e6727b0 100644 --- a/mm/migrate.c +++ b/mm/migrate.c @@ -2391,8 +2391,14 @@ static void migrate_vma_collect(struct migrate_vma *migrate) { struct mmu_notifier_range range;
- mmu_notifier_range_init(&range, MMU_NOTIFY_CLEAR, 0, NULL, + /* + * Note that the src_owner is passed to the mmu notifier callback so + * that the registered device driver can skip invalidating device + * private page mappings that won't be migrated. + */ + mmu_notifier_range_init(&range, MMU_NOTIFY_MIGRATE, 0, migrate->vma, migrate->vma->vm_mm, migrate->start, migrate->end); + range.data = migrate->src_owner; mmu_notifier_invalidate_range_start(&range);
walk_page_range(migrate->vma->vm_mm, migrate->start, migrate->end,
On Mon, Jul 06, 2020 at 03:23:45PM -0700, Ralph Campbell wrote:
Currently migrate_vma_setup() calls mmu_notifier_invalidate_range_start() which flushes all device private page mappings whether or not a page is being migrated to/from device private memory. In order to not disrupt device mappings that are not being migrated, shift the responsibility for clearing device private mappings to the device driver and leave CPU page table unmapping handled by migrate_vma_setup(). To support this, the caller of migrate_vma_setup() should always set struct migrate_vma::src_owner to a non NULL value that matches the device private page->pgmap->owner. This value is then passed to the struct mmu_notifier_range with a new event type which the driver's invalidation function can use to avoid device MMU invalidations.
Signed-off-by: Ralph Campbell rcampbell@nvidia.com include/linux/mmu_notifier.h | 7 +++++++ mm/migrate.c | 8 +++++++- 2 files changed, 14 insertions(+), 1 deletion(-)
diff --git a/include/linux/mmu_notifier.h b/include/linux/mmu_notifier.h index fc68f3570e19..bd0b34dbe4de 100644 +++ b/include/linux/mmu_notifier.h @@ -38,6 +38,10 @@ struct mmu_interval_notifier;
- @MMU_NOTIFY_RELEASE: used during mmu_interval_notifier invalidate to signal
- that the mm refcount is zero and the range is no longer accessible.
- @MMU_NOTIFY_MIGRATE: used during migrate_vma_collect() invalidate to signal
- a device driver to possibly ignore the invalidation if the src_own
*/
- field matches.
enum mmu_notifier_event { MMU_NOTIFY_UNMAP = 0, @@ -46,6 +50,7 @@ enum mmu_notifier_event { MMU_NOTIFY_PROTECTION_PAGE, MMU_NOTIFY_SOFT_DIRTY, MMU_NOTIFY_RELEASE,
- MMU_NOTIFY_MIGRATE,
}; #define MMU_NOTIFIER_RANGE_BLOCKABLE (1 << 0) @@ -264,6 +269,7 @@ struct mmu_notifier_range { unsigned long end; unsigned flags; enum mmu_notifier_event event;
- void *data;
};
This generic member usually ends up a bit ugly, can we do a tagged union instead?
union { void *migrate_pgmap_owner; };
and probably drop the union until we actually need two things here.
Jason
On 7/10/20 12:39 PM, Jason Gunthorpe wrote:
On Mon, Jul 06, 2020 at 03:23:45PM -0700, Ralph Campbell wrote:
Currently migrate_vma_setup() calls mmu_notifier_invalidate_range_start() which flushes all device private page mappings whether or not a page is being migrated to/from device private memory. In order to not disrupt device mappings that are not being migrated, shift the responsibility for clearing device private mappings to the device driver and leave CPU page table unmapping handled by migrate_vma_setup(). To support this, the caller of migrate_vma_setup() should always set struct migrate_vma::src_owner to a non NULL value that matches the device private page->pgmap->owner. This value is then passed to the struct mmu_notifier_range with a new event type which the driver's invalidation function can use to avoid device MMU invalidations.
Signed-off-by: Ralph Campbell rcampbell@nvidia.com include/linux/mmu_notifier.h | 7 +++++++ mm/migrate.c | 8 +++++++- 2 files changed, 14 insertions(+), 1 deletion(-)
diff --git a/include/linux/mmu_notifier.h b/include/linux/mmu_notifier.h index fc68f3570e19..bd0b34dbe4de 100644 +++ b/include/linux/mmu_notifier.h @@ -38,6 +38,10 @@ struct mmu_interval_notifier;
- @MMU_NOTIFY_RELEASE: used during mmu_interval_notifier invalidate to signal
- that the mm refcount is zero and the range is no longer accessible.
- @MMU_NOTIFY_MIGRATE: used during migrate_vma_collect() invalidate to signal
- a device driver to possibly ignore the invalidation if the src_own
*/ enum mmu_notifier_event { MMU_NOTIFY_UNMAP = 0,
- field matches.
@@ -46,6 +50,7 @@ enum mmu_notifier_event { MMU_NOTIFY_PROTECTION_PAGE, MMU_NOTIFY_SOFT_DIRTY, MMU_NOTIFY_RELEASE,
- MMU_NOTIFY_MIGRATE, };
#define MMU_NOTIFIER_RANGE_BLOCKABLE (1 << 0) @@ -264,6 +269,7 @@ struct mmu_notifier_range { unsigned long end; unsigned flags; enum mmu_notifier_event event;
- void *data; };
This generic member usually ends up a bit ugly, can we do a tagged union instead?
union { void *migrate_pgmap_owner; };
and probably drop the union until we actually need two things here.
Jason
OK, I'll send a v2 with this change.
Use the new MMU_NOTIFY_MIGRATE event to skip GPU MMU invalidations of device private memory and handle the invalidation in the driver as part of migrating device private memory.
Signed-off-by: Ralph Campbell rcampbell@nvidia.com --- drivers/gpu/drm/nouveau/nouveau_dmem.c | 11 ++++++++--- drivers/gpu/drm/nouveau/nouveau_svm.c | 10 +++++++++- drivers/gpu/drm/nouveau/nouveau_svm.h | 1 + 3 files changed, 18 insertions(+), 4 deletions(-)
diff --git a/drivers/gpu/drm/nouveau/nouveau_dmem.c b/drivers/gpu/drm/nouveau/nouveau_dmem.c index 762e0168bc8d..da793e301c0c 100644 --- a/drivers/gpu/drm/nouveau/nouveau_dmem.c +++ b/drivers/gpu/drm/nouveau/nouveau_dmem.c @@ -154,6 +154,8 @@ static vm_fault_t nouveau_dmem_fault_copy_one(struct nouveau_drm *drm, if (dma_mapping_error(dev, *dma_addr)) goto error_free_page;
+ nouveau_svmm_invalidate(spage->zone_device_data, args->start, + args->end); if (drm->dmem->migrate.copy_func(drm, 1, NOUVEAU_APER_HOST, *dma_addr, NOUVEAU_APER_VRAM, nouveau_dmem_page_addr(spage))) goto error_dma_unmap; @@ -531,7 +533,8 @@ nouveau_dmem_init(struct nouveau_drm *drm) }
static unsigned long nouveau_dmem_migrate_copy_one(struct nouveau_drm *drm, - unsigned long src, dma_addr_t *dma_addr, u64 *pfn) + struct nouveau_svmm *svmm, unsigned long src, + dma_addr_t *dma_addr, u64 *pfn) { struct device *dev = drm->dev->dev; struct page *dpage, *spage; @@ -561,6 +564,7 @@ static unsigned long nouveau_dmem_migrate_copy_one(struct nouveau_drm *drm, goto out_free_page; }
+ dpage->zone_device_data = svmm; *pfn = NVIF_VMM_PFNMAP_V0_V | NVIF_VMM_PFNMAP_V0_VRAM | ((paddr >> PAGE_SHIFT) << NVIF_VMM_PFNMAP_V0_ADDR_SHIFT); if (src & MIGRATE_PFN_WRITE) @@ -584,8 +588,8 @@ static void nouveau_dmem_migrate_chunk(struct nouveau_drm *drm, unsigned long addr = args->start, nr_dma = 0, i;
for (i = 0; addr < args->end; i++) { - args->dst[i] = nouveau_dmem_migrate_copy_one(drm, args->src[i], - dma_addrs + nr_dma, pfns + i); + args->dst[i] = nouveau_dmem_migrate_copy_one(drm, svmm, + args->src[i], dma_addrs + nr_dma, pfns + i); if (!dma_mapping_error(drm->dev->dev, dma_addrs[nr_dma])) nr_dma++; addr += PAGE_SIZE; @@ -616,6 +620,7 @@ nouveau_dmem_migrate_vma(struct nouveau_drm *drm, struct migrate_vma args = { .vma = vma, .start = start, + .src_owner = drm->dev, .dir = MIGRATE_VMA_FROM_SYSTEM, }; unsigned long i; diff --git a/drivers/gpu/drm/nouveau/nouveau_svm.c b/drivers/gpu/drm/nouveau/nouveau_svm.c index 9dd296d3c378..ccfd4304d737 100644 --- a/drivers/gpu/drm/nouveau/nouveau_svm.c +++ b/drivers/gpu/drm/nouveau/nouveau_svm.c @@ -246,7 +246,7 @@ nouveau_svmm_join(struct nouveau_svmm *svmm, u64 inst) }
/* Invalidate SVMM address-range on GPU. */ -static void +void nouveau_svmm_invalidate(struct nouveau_svmm *svmm, u64 start, u64 limit) { if (limit > start) { @@ -279,6 +279,14 @@ nouveau_svmm_invalidate_range_start(struct mmu_notifier *mn, if (unlikely(!svmm->vmm)) goto out;
+ /* + * Ignore invalidation callbacks for device private pages since + * the invalidation is handled as part of the migration process. + */ + if (update->event == MMU_NOTIFY_MIGRATE && + update->data == svmm->vmm->cli->drm->dev) + goto out; + if (limit > svmm->unmanaged.start && start < svmm->unmanaged.limit) { if (start < svmm->unmanaged.start) { nouveau_svmm_invalidate(svmm, start, diff --git a/drivers/gpu/drm/nouveau/nouveau_svm.h b/drivers/gpu/drm/nouveau/nouveau_svm.h index f0fcd1b72e8b..bb2d56e50e0c 100644 --- a/drivers/gpu/drm/nouveau/nouveau_svm.h +++ b/drivers/gpu/drm/nouveau/nouveau_svm.h @@ -19,6 +19,7 @@ int nouveau_svmm_join(struct nouveau_svmm *, u64 inst); void nouveau_svmm_part(struct nouveau_svmm *, u64 inst); int nouveau_svmm_bind(struct drm_device *, void *, struct drm_file *);
+void nouveau_svmm_invalidate(struct nouveau_svmm *svmm, u64 start, u64 limit); u64 *nouveau_pfns_alloc(unsigned long npages); void nouveau_pfns_free(u64 *pfns); void nouveau_pfns_map(struct nouveau_svmm *svmm, struct mm_struct *mm,
Use the new MMU_NOTIFY_MIGRATE event to skip MMU invalidations of device private memory and handle the invalidation in the driver as part of migrating device private memory.
Signed-off-by: Ralph Campbell rcampbell@nvidia.com --- lib/test_hmm.c | 31 ++++++++++++++++++------------- 1 file changed, 18 insertions(+), 13 deletions(-)
diff --git a/lib/test_hmm.c b/lib/test_hmm.c index 1bd60cfb5a25..a170e5f7fb2e 100644 --- a/lib/test_hmm.c +++ b/lib/test_hmm.c @@ -214,6 +214,14 @@ static bool dmirror_interval_invalidate(struct mmu_interval_notifier *mni, { struct dmirror *dmirror = container_of(mni, struct dmirror, notifier);
+ /* + * Ignore invalidation callbacks for device private pages since + * the invalidation is handled as part of the migration process. + */ + if (range->event == MMU_NOTIFY_MIGRATE && + range->data == dmirror->mdevice) + return true; + if (mmu_notifier_range_blockable(range)) mutex_lock(&dmirror->mutex); else if (!mutex_trylock(&dmirror->mutex)) @@ -702,7 +710,7 @@ static int dmirror_migrate(struct dmirror *dmirror, args.dst = dst_pfns; args.start = addr; args.end = next; - args.src_owner = NULL; + args.src_owner = dmirror->mdevice; args.dir = MIGRATE_VMA_FROM_SYSTEM; ret = migrate_vma_setup(&args); if (ret) @@ -992,7 +1000,7 @@ static void dmirror_devmem_free(struct page *page) }
static vm_fault_t dmirror_devmem_fault_alloc_and_copy(struct migrate_vma *args, - struct dmirror_device *mdevice) + struct dmirror *dmirror) { const unsigned long *src = args->src; unsigned long *dst = args->dst; @@ -1014,6 +1022,7 @@ static vm_fault_t dmirror_devmem_fault_alloc_and_copy(struct migrate_vma *args, continue;
lock_page(dpage); + xa_erase(&dmirror->pt, addr >> PAGE_SHIFT); copy_highpage(dpage, spage); *dst = migrate_pfn(page_to_pfn(dpage)) | MIGRATE_PFN_LOCKED; if (*src & MIGRATE_PFN_WRITE) @@ -1022,15 +1031,6 @@ static vm_fault_t dmirror_devmem_fault_alloc_and_copy(struct migrate_vma *args, return 0; }
-static void dmirror_devmem_fault_finalize_and_map(struct migrate_vma *args, - struct dmirror *dmirror) -{ - /* Invalidate the device's page table mapping. */ - mutex_lock(&dmirror->mutex); - dmirror_do_update(dmirror, args->start, args->end); - mutex_unlock(&dmirror->mutex); -} - static vm_fault_t dmirror_devmem_fault(struct vm_fault *vmf) { struct migrate_vma args; @@ -1060,11 +1060,16 @@ static vm_fault_t dmirror_devmem_fault(struct vm_fault *vmf) if (migrate_vma_setup(&args)) return VM_FAULT_SIGBUS;
- ret = dmirror_devmem_fault_alloc_and_copy(&args, dmirror->mdevice); + ret = dmirror_devmem_fault_alloc_and_copy(&args, dmirror); if (ret) return ret; migrate_vma_pages(&args); - dmirror_devmem_fault_finalize_and_map(&args, dmirror); + /* + * No device finalize step is needed since + * dmirror_devmem_fault_alloc_and_copy() will have already + * invalidated the device page table. We could reinstate device MMU + * entries for pages that didn't migrate but that should be rare. + */ migrate_vma_finalize(&args); return 0; }
On Mon, Jul 06, 2020 at 03:23:42PM -0700, Ralph Campbell wrote:
The goal for this series is to avoid device private memory TLB invalidations when migrating a range of addresses from system memory to device private memory and some of those pages have already been migrated. The approach taken is to introduce a new mmu notifier invalidation event type and use that in the device driver to skip invalidation callbacks from migrate_vma_setup(). The device driver is also then expected to handle device MMU invalidations as part of the migrate_vma_setup(), migrate_vma_pages(), migrate_vma_finalize() process. Note that this is opt-in. A device driver can simply invalidate its MMU in the mmu notifier callback and not handle MMU invalidations in the migration sequence.
In the kvmppc secure guest usecase,
1. We ensure that we don't issue migrate_vma() calls for pages that have already been migrated to the device side (which is actually secure memory for us that is managed by Ultravisor firmware)
2. The page table mappings on the device side (secure memory) are managed transparent to the kernel by the Ultravisor firmware.
Hence I assume that no specific action would be required by the kvmppc usecase due to this patchset. In fact, we never registered for this mmu notifier events.
Regards, Bharata.
linux-kselftest-mirror@lists.linaro.org
-
Bharata B Rao -
Jason Gunthorpe -
Ralph Campbell