Merge tag 'for-linus-hmm' of git://git.kernel.org/pub/scm/linux/kernel/git/rdma/rdma

Pull hmm updates from Jason Gunthorpe:
 "This series adds a selftest for hmm_range_fault() and several of the
  DEVICE_PRIVATE migration related actions, and another simplification
  for hmm_range_fault()'s API.

   - Simplify hmm_range_fault() with a simpler return code, no
     HMM_PFN_SPECIAL, and no customizable output PFN format

   - Add a selftest for hmm_range_fault() and DEVICE_PRIVATE related
     functionality"

* tag 'for-linus-hmm' of git://git.kernel.org/pub/scm/linux/kernel/git/rdma/rdma:
  MAINTAINERS: add HMM selftests
  mm/hmm/test: add selftests for HMM
  mm/hmm/test: add selftest driver for HMM
  mm/hmm: remove the customizable pfn format from hmm_range_fault
  mm/hmm: remove HMM_PFN_SPECIAL
  drm/amdgpu: remove dead code after hmm_range_fault()
  mm/hmm: make hmm_range_fault return 0 or -1

4 files changed, 82 insertions, 98 deletions

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
index 6309ff72bd78..b1c62da527c5 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
@@ -766,18 +766,6 @@ struct amdgpu_ttm_tt {
 };
 
 #ifdef CONFIG_DRM_AMDGPU_USERPTR
-/* flags used by HMM internal, not related to CPU/GPU PTE flags */
-static const uint64_t hmm_range_flags[HMM_PFN_FLAG_MAX] = {
-	(1 << 0), /* HMM_PFN_VALID */
-	(1 << 1), /* HMM_PFN_WRITE */
-};
-
-static const uint64_t hmm_range_values[HMM_PFN_VALUE_MAX] = {
-	0xfffffffffffffffeUL, /* HMM_PFN_ERROR */
-	0, /* HMM_PFN_NONE */
-	0xfffffffffffffffcUL /* HMM_PFN_SPECIAL */
-};
-
 /**
  * amdgpu_ttm_tt_get_user_pages - get device accessible pages that back user
  * memory and start HMM tracking CPU page table update
@@ -816,18 +804,15 @@ int amdgpu_ttm_tt_get_user_pages(struct amdgpu_bo *bo, struct page **pages)
 		goto out;
 	}
 	range->notifier = &bo->notifier;
-	range->flags = hmm_range_flags;
-	range->values = hmm_range_values;
-	range->pfn_shift = PAGE_SHIFT;
 	range->start = bo->notifier.interval_tree.start;
 	range->end = bo->notifier.interval_tree.last + 1;
-	range->default_flags = hmm_range_flags[HMM_PFN_VALID];
+	range->default_flags = HMM_PFN_REQ_FAULT;
 	if (!amdgpu_ttm_tt_is_readonly(ttm))
-		range->default_flags |= range->flags[HMM_PFN_WRITE];
+		range->default_flags |= HMM_PFN_REQ_WRITE;
 
-	range->pfns = kvmalloc_array(ttm->num_pages, sizeof(*range->pfns),
-				     GFP_KERNEL);
-	if (unlikely(!range->pfns)) {
+	range->hmm_pfns = kvmalloc_array(ttm->num_pages,
+					 sizeof(*range->hmm_pfns), GFP_KERNEL);
+	if (unlikely(!range->hmm_pfns)) {
 		r = -ENOMEM;
 		goto out_free_ranges;
 	}
@@ -852,27 +837,23 @@ retry:
 	down_read(&mm->mmap_sem);
 	r = hmm_range_fault(range);
 	up_read(&mm->mmap_sem);
-	if (unlikely(r <= 0)) {
+	if (unlikely(r)) {
 		/*
 		 * FIXME: This timeout should encompass the retry from
 		 * mmu_interval_read_retry() as well.
 		 */
-		if ((r == 0 || r == -EBUSY) && !time_after(jiffies, timeout))
+		if (r == -EBUSY && !time_after(jiffies, timeout))
 			goto retry;
 		goto out_free_pfns;
 	}
 
-	for (i = 0; i < ttm->num_pages; i++) {
-		/* FIXME: The pages cannot be touched outside the notifier_lock */
-		pages[i] = hmm_device_entry_to_page(range, range->pfns[i]);
-		if (unlikely(!pages[i])) {
-			pr_err("Page fault failed for pfn[%lu] = 0x%llx\n",
-			       i, range->pfns[i]);
-			r = -ENOMEM;
-
-			goto out_free_pfns;
-		}
-	}
+	/*
+	 * Due to default_flags, all pages are HMM_PFN_VALID or
+	 * hmm_range_fault() fails. FIXME: The pages cannot be touched outside
+	 * the notifier_lock, and mmu_interval_read_retry() must be done first.
+	 */
+	for (i = 0; i < ttm->num_pages; i++)
+		pages[i] = hmm_pfn_to_page(range->hmm_pfns[i]);
 
 	gtt->range = range;
 	mmput(mm);
@@ -882,7 +863,7 @@ retry:
 out_unlock:
 	up_read(&mm->mmap_sem);
 out_free_pfns:
-	kvfree(range->pfns);
+	kvfree(range->hmm_pfns);
 out_free_ranges:
 	kfree(range);
 out:
@@ -907,7 +888,7 @@ bool amdgpu_ttm_tt_get_user_pages_done(struct ttm_tt *ttm)
 	DRM_DEBUG_DRIVER("user_pages_done 0x%llx pages 0x%lx\n",
 		gtt->userptr, ttm->num_pages);
 
-	WARN_ONCE(!gtt->range || !gtt->range->pfns,
+	WARN_ONCE(!gtt->range || !gtt->range->hmm_pfns,
 		"No user pages to check\n");
 
 	if (gtt->range) {
@@ -917,7 +898,7 @@ bool amdgpu_ttm_tt_get_user_pages_done(struct ttm_tt *ttm)
 		 */
 		r = mmu_interval_read_retry(gtt->range->notifier,
 					 gtt->range->notifier_seq);
-		kvfree(gtt->range->pfns);
+		kvfree(gtt->range->hmm_pfns);
 		kfree(gtt->range);
 		gtt->range = NULL;
 	}
@@ -1008,8 +989,7 @@ static void amdgpu_ttm_tt_unpin_userptr(struct ttm_tt *ttm)
 
 		for (i = 0; i < ttm->num_pages; i++) {
 			if (ttm->pages[i] !=
-				hmm_device_entry_to_page(gtt->range,
-					      gtt->range->pfns[i]))
+			    hmm_pfn_to_page(gtt->range->hmm_pfns[i]))
 				break;
 		}
 
diff --git a/drivers/gpu/drm/nouveau/nouveau_dmem.c b/drivers/gpu/drm/nouveau/nouveau_dmem.c
index ad89e09a0be3..3364904eccff 100644
--- a/drivers/gpu/drm/nouveau/nouveau_dmem.c
+++ b/drivers/gpu/drm/nouveau/nouveau_dmem.c
@@ -85,7 +85,7 @@ static inline struct nouveau_dmem *page_to_dmem(struct page *page)
 	return container_of(page->pgmap, struct nouveau_dmem, pagemap);
 }
 
-static unsigned long nouveau_dmem_page_addr(struct page *page)
+unsigned long nouveau_dmem_page_addr(struct page *page)
 {
 	struct nouveau_dmem_chunk *chunk = page->zone_device_data;
 	unsigned long idx = page_to_pfn(page) - chunk->pfn_first;
@@ -671,28 +671,3 @@ out_free_src:
 out:
 	return ret;
 }
-
-void
-nouveau_dmem_convert_pfn(struct nouveau_drm *drm,
-			 struct hmm_range *range)
-{
-	unsigned long i, npages;
-
-	npages = (range->end - range->start) >> PAGE_SHIFT;
-	for (i = 0; i < npages; ++i) {
-		struct page *page;
-		uint64_t addr;
-
-		page = hmm_device_entry_to_page(range, range->pfns[i]);
-		if (page == NULL)
-			continue;
-
-		if (!is_device_private_page(page))
-			continue;
-
-		addr = nouveau_dmem_page_addr(page);
-		range->pfns[i] &= ((1UL << range->pfn_shift) - 1);
-		range->pfns[i] |= (addr >> PAGE_SHIFT) << range->pfn_shift;
-		range->pfns[i] |= NVIF_VMM_PFNMAP_V0_VRAM;
-	}
-}
diff --git a/drivers/gpu/drm/nouveau/nouveau_dmem.h b/drivers/gpu/drm/nouveau/nouveau_dmem.h
index 92394be5d649..db3b59b210af 100644
--- a/drivers/gpu/drm/nouveau/nouveau_dmem.h
+++ b/drivers/gpu/drm/nouveau/nouveau_dmem.h
@@ -37,9 +37,8 @@ int nouveau_dmem_migrate_vma(struct nouveau_drm *drm,
 			     struct vm_area_struct *vma,
 			     unsigned long start,
 			     unsigned long end);
+unsigned long nouveau_dmem_page_addr(struct page *page);
 
-void nouveau_dmem_convert_pfn(struct nouveau_drm *drm,
-			      struct hmm_range *range);
 #else /* IS_ENABLED(CONFIG_DRM_NOUVEAU_SVM) */
 static inline void nouveau_dmem_init(struct nouveau_drm *drm) {}
 static inline void nouveau_dmem_fini(struct nouveau_drm *drm) {}
diff --git a/drivers/gpu/drm/nouveau/nouveau_svm.c b/drivers/gpu/drm/nouveau/nouveau_svm.c
index 645fedd77e21..407e34a5c0ab 100644
--- a/drivers/gpu/drm/nouveau/nouveau_svm.c
+++ b/drivers/gpu/drm/nouveau/nouveau_svm.c
@@ -369,19 +369,6 @@ out_free:
 	return ret;
 }
 
-static const u64
-nouveau_svm_pfn_flags[HMM_PFN_FLAG_MAX] = {
-	[HMM_PFN_VALID         ] = NVIF_VMM_PFNMAP_V0_V,
-	[HMM_PFN_WRITE         ] = NVIF_VMM_PFNMAP_V0_W,
-};
-
-static const u64
-nouveau_svm_pfn_values[HMM_PFN_VALUE_MAX] = {
-	[HMM_PFN_ERROR  ] = ~NVIF_VMM_PFNMAP_V0_V,
-	[HMM_PFN_NONE   ] =  NVIF_VMM_PFNMAP_V0_NONE,
-	[HMM_PFN_SPECIAL] = ~NVIF_VMM_PFNMAP_V0_V,
-};
-
 /* Issue fault replay for GPU to retry accesses that faulted previously. */
 static void
 nouveau_svm_fault_replay(struct nouveau_svm *svm)
@@ -519,9 +506,45 @@ static const struct mmu_interval_notifier_ops nouveau_svm_mni_ops = {
 	.invalidate = nouveau_svm_range_invalidate,
 };
 
+static void nouveau_hmm_convert_pfn(struct nouveau_drm *drm,
+				    struct hmm_range *range, u64 *ioctl_addr)
+{
+	unsigned long i, npages;
+
+	/*
+	 * The ioctl_addr prepared here is passed through nvif_object_ioctl()
+	 * to an eventual DMA map in something like gp100_vmm_pgt_pfn()
+	 *
+	 * This is all just encoding the internal hmm representation into a
+	 * different nouveau internal representation.
+	 */
+	npages = (range->end - range->start) >> PAGE_SHIFT;
+	for (i = 0; i < npages; ++i) {
+		struct page *page;
+
+		if (!(range->hmm_pfns[i] & HMM_PFN_VALID)) {
+			ioctl_addr[i] = 0;
+			continue;
+		}
+
+		page = hmm_pfn_to_page(range->hmm_pfns[i]);
+		if (is_device_private_page(page))
+			ioctl_addr[i] = nouveau_dmem_page_addr(page) |
+					NVIF_VMM_PFNMAP_V0_V |
+					NVIF_VMM_PFNMAP_V0_VRAM;
+		else
+			ioctl_addr[i] = page_to_phys(page) |
+					NVIF_VMM_PFNMAP_V0_V |
+					NVIF_VMM_PFNMAP_V0_HOST;
+		if (range->hmm_pfns[i] & HMM_PFN_WRITE)
+			ioctl_addr[i] |= NVIF_VMM_PFNMAP_V0_W;
+	}
+}
+
 static int nouveau_range_fault(struct nouveau_svmm *svmm,
 			       struct nouveau_drm *drm, void *data, u32 size,
-			       u64 *pfns, struct svm_notifier *notifier)
+			       unsigned long hmm_pfns[], u64 *ioctl_addr,
+			       struct svm_notifier *notifier)
 {
 	unsigned long timeout =
 		jiffies + msecs_to_jiffies(HMM_RANGE_DEFAULT_TIMEOUT);
@@ -530,26 +553,27 @@ static int nouveau_range_fault(struct nouveau_svmm *svmm,
 		.notifier = &notifier->notifier,
 		.start = notifier->notifier.interval_tree.start,
 		.end = notifier->notifier.interval_tree.last + 1,
-		.pfns = pfns,
-		.flags = nouveau_svm_pfn_flags,
-		.values = nouveau_svm_pfn_values,
-		.pfn_shift = NVIF_VMM_PFNMAP_V0_ADDR_SHIFT,
+		.pfn_flags_mask = HMM_PFN_REQ_FAULT | HMM_PFN_REQ_WRITE,
+		.hmm_pfns = hmm_pfns,
 	};
 	struct mm_struct *mm = notifier->notifier.mm;
-	long ret;
+	int ret;
 
 	while (true) {
 		if (time_after(jiffies, timeout))
 			return -EBUSY;
 
 		range.notifier_seq = mmu_interval_read_begin(range.notifier);
-		range.default_flags = 0;
-		range.pfn_flags_mask = -1UL;
 		down_read(&mm->mmap_sem);
 		ret = hmm_range_fault(&range);
 		up_read(&mm->mmap_sem);
-		if (ret <= 0) {
-			if (ret == 0 || ret == -EBUSY)
+		if (ret) {
+			/*
+			 * FIXME: the input PFN_REQ flags are destroyed on
+			 * -EBUSY, we need to regenerate them, also for the
+			 * other continue below
+			 */
+			if (ret == -EBUSY)
 				continue;
 			return ret;
 		}
@@ -563,7 +587,7 @@ static int nouveau_range_fault(struct nouveau_svmm *svmm,
 		break;
 	}
 
-	nouveau_dmem_convert_pfn(drm, &range);
+	nouveau_hmm_convert_pfn(drm, &range, ioctl_addr);
 
 	svmm->vmm->vmm.object.client->super = true;
 	ret = nvif_object_ioctl(&svmm->vmm->vmm.object, data, size, NULL);
@@ -590,6 +614,7 @@ nouveau_svm_fault(struct nvif_notify *notify)
 		} i;
 		u64 phys[16];
 	} args;
+	unsigned long hmm_pfns[ARRAY_SIZE(args.phys)];
 	struct vm_area_struct *vma;
 	u64 inst, start, limit;
 	int fi, fn, pi, fill;
@@ -705,12 +730,17 @@ nouveau_svm_fault(struct nvif_notify *notify)
 			 * access flags.
 			 *XXX: atomic?
 			 */
-			if (buffer->fault[fn]->access != 0 /* READ. */ &&
-			    buffer->fault[fn]->access != 3 /* PREFETCH. */) {
-				args.phys[pi++] = NVIF_VMM_PFNMAP_V0_V |
-						  NVIF_VMM_PFNMAP_V0_W;
-			} else {
-				args.phys[pi++] = NVIF_VMM_PFNMAP_V0_V;
+			switch (buffer->fault[fn]->access) {
+			case 0: /* READ. */
+				hmm_pfns[pi++] = HMM_PFN_REQ_FAULT;
+				break;
+			case 3: /* PREFETCH. */
+				hmm_pfns[pi++] = 0;
+				break;
+			default:
+				hmm_pfns[pi++] = HMM_PFN_REQ_FAULT |
+						 HMM_PFN_REQ_WRITE;
+				break;
 			}
 			args.i.p.size = pi << PAGE_SHIFT;
 
@@ -738,7 +768,7 @@ nouveau_svm_fault(struct nvif_notify *notify)
 			fill = (buffer->fault[fn    ]->addr -
 				buffer->fault[fn - 1]->addr) >> PAGE_SHIFT;
 			while (--fill)
-				args.phys[pi++] = NVIF_VMM_PFNMAP_V0_NONE;
+				hmm_pfns[pi++] = 0;
 		}
 
 		SVMM_DBG(svmm, "wndw %016llx-%016llx covering %d fault(s)",
@@ -754,7 +784,7 @@ nouveau_svm_fault(struct nvif_notify *notify)
 			ret = nouveau_range_fault(
 				svmm, svm->drm, &args,
 				sizeof(args.i) + pi * sizeof(args.phys[0]),
-				args.phys, &notifier);
+				hmm_pfns, args.phys, &notifier);
 			mmu_interval_notifier_remove(&notifier.notifier);
 		}
 		mmput(mm);