Merge branch 'iommu-memory-accounting' of ssh://gitolite.kernel.org/pub/scm/linux/kernel/git/joro/iommu intoiommufd/for-next

Jason Gunthorpe says:

====================
iommufd follows the same design as KVM and uses memory cgroups to limit
the amount of kernel memory a iommufd file descriptor can pin down. The
various internal data structures already use GFP_KERNEL_ACCOUNT to charge
its own memory.

However, one of the biggest consumers of kernel memory is the IOPTEs
stored under the iommu_domain and these allocations are not tracked.

This series is the first step in fixing it.

The iommu driver contract already includes a 'gfp' argument to the
map_pages op, allowing iommufd to specify GFP_KERNEL_ACCOUNT and then
having the driver allocate the IOPTE tables with that flag will capture a
significant amount of the allocations.

Update the iommu_map() API to pass in the GFP argument, and fix all call
sites. Replace iommu_map_atomic().

Audit the "enterprise" iommu drivers to make sure they do the right thing.
Intel and S390 ignore the GFP argument and always use GFP_ATOMIC. This is
problematic for iommufd anyhow, so fix it. AMD and ARM SMMUv2/3 are
already correct.

A follow up series will be needed to capture the allocations made when the
iommu_domain itself is allocated, which will complete the job.
====================

* 'iommu-memory-accounting' of ssh://gitolite.kernel.org/pub/scm/linux/kernel/git/joro/iommu:
  iommu/s390: Use GFP_KERNEL in sleepable contexts
  iommu/s390: Push the gfp parameter to the kmem_cache_alloc()'s
  iommu/intel: Use GFP_KERNEL in sleepable contexts
  iommu/intel: Support the gfp argument to the map_pages op
  iommu/intel: Add a gfp parameter to alloc_pgtable_page()
  iommufd: Use GFP_KERNEL_ACCOUNT for iommu_map()
  iommu/dma: Use the gfp parameter in __iommu_dma_alloc_noncontiguous()
  iommu: Add a gfp parameter to iommu_map_sg()
  iommu: Remove iommu_map_atomic()
  iommu: Add a gfp parameter to iommu_map()

Link: https://lore.kernel.org/linux-iommu/0-v3-76b587fe28df+6e3-iommu_map_gfp_jgg@nvidia.com
Signed-off-by: Jason Gunthorpe <jgg@nvidia.com>

3 files changed, 23 insertions, 17 deletions

diff --git a/drivers/iommu/intel/iommu.c b/drivers/iommu/intel/iommu.c
index 7cfab5fd5e59..6fd223cf639f 100644
--- a/drivers/iommu/intel/iommu.c
+++ b/drivers/iommu/intel/iommu.c
@@ -362,12 +362,12 @@ static int __init intel_iommu_setup(char *str)
 }
 __setup("intel_iommu=", intel_iommu_setup);
 
-void *alloc_pgtable_page(int node)
+void *alloc_pgtable_page(int node, gfp_t gfp)
 {
 	struct page *page;
 	void *vaddr = NULL;
 
-	page = alloc_pages_node(node, GFP_ATOMIC | __GFP_ZERO, 0);
+	page = alloc_pages_node(node, gfp | __GFP_ZERO, 0);
 	if (page)
 		vaddr = page_address(page);
 	return vaddr;
@@ -612,7 +612,7 @@ struct context_entry *iommu_context_addr(struct intel_iommu *iommu, u8 bus,
 		if (!alloc)
 			return NULL;
 
-		context = alloc_pgtable_page(iommu->node);
+		context = alloc_pgtable_page(iommu->node, GFP_ATOMIC);
 		if (!context)
 			return NULL;
 
@@ -908,7 +908,8 @@ pgtable_walk:
 #endif
 
 static struct dma_pte *pfn_to_dma_pte(struct dmar_domain *domain,
-				      unsigned long pfn, int *target_level)
+				      unsigned long pfn, int *target_level,
+				      gfp_t gfp)
 {
 	struct dma_pte *parent, *pte;
 	int level = agaw_to_level(domain->agaw);
@@ -935,7 +936,7 @@ static struct dma_pte *pfn_to_dma_pte(struct dmar_domain *domain,
 		if (!dma_pte_present(pte)) {
 			uint64_t pteval;
 
-			tmp_page = alloc_pgtable_page(domain->nid);
+			tmp_page = alloc_pgtable_page(domain->nid, gfp);
 
 			if (!tmp_page)
 				return NULL;
@@ -1186,7 +1187,7 @@ static int iommu_alloc_root_entry(struct intel_iommu *iommu)
 {
 	struct root_entry *root;
 
-	root = (struct root_entry *)alloc_pgtable_page(iommu->node);
+	root = (struct root_entry *)alloc_pgtable_page(iommu->node, GFP_ATOMIC);
 	if (!root) {
 		pr_err("Allocating root entry for %s failed\n",
 			iommu->name);
@@ -2150,7 +2151,8 @@ static void switch_to_super_page(struct dmar_domain *domain,
 
 	while (start_pfn <= end_pfn) {
 		if (!pte)
-			pte = pfn_to_dma_pte(domain, start_pfn, &level);
+			pte = pfn_to_dma_pte(domain, start_pfn, &level,
+					     GFP_ATOMIC);
 
 		if (dma_pte_present(pte)) {
 			dma_pte_free_pagetable(domain, start_pfn,
@@ -2172,7 +2174,8 @@ static void switch_to_super_page(struct dmar_domain *domain,
 
 static int
 __domain_mapping(struct dmar_domain *domain, unsigned long iov_pfn,
-		 unsigned long phys_pfn, unsigned long nr_pages, int prot)
+		 unsigned long phys_pfn, unsigned long nr_pages, int prot,
+		 gfp_t gfp)
 {
 	struct dma_pte *first_pte = NULL, *pte = NULL;
 	unsigned int largepage_lvl = 0;
@@ -2202,7 +2205,8 @@ __domain_mapping(struct dmar_domain *domain, unsigned long iov_pfn,
 			largepage_lvl = hardware_largepage_caps(domain, iov_pfn,
 					phys_pfn, nr_pages);
 
-			pte = pfn_to_dma_pte(domain, iov_pfn, &largepage_lvl);
+			pte = pfn_to_dma_pte(domain, iov_pfn, &largepage_lvl,
+					     gfp);
 			if (!pte)
 				return -ENOMEM;
 			first_pte = pte;
@@ -2368,7 +2372,7 @@ static int iommu_domain_identity_map(struct dmar_domain *domain,
 
 	return __domain_mapping(domain, first_vpfn,
 				first_vpfn, last_vpfn - first_vpfn + 1,
-				DMA_PTE_READ|DMA_PTE_WRITE);
+				DMA_PTE_READ|DMA_PTE_WRITE, GFP_KERNEL);
 }
 
 static int md_domain_init(struct dmar_domain *domain, int guest_width);
@@ -2676,7 +2680,7 @@ static int copy_context_table(struct intel_iommu *iommu,
 			if (!old_ce)
 				goto out;
 
-			new_ce = alloc_pgtable_page(iommu->node);
+			new_ce = alloc_pgtable_page(iommu->node, GFP_KERNEL);
 			if (!new_ce)
 				goto out_unmap;
 
@@ -4136,7 +4140,7 @@ static int md_domain_init(struct dmar_domain *domain, int guest_width)
 	domain->max_addr = 0;
 
 	/* always allocate the top pgd */
-	domain->pgd = alloc_pgtable_page(domain->nid);
+	domain->pgd = alloc_pgtable_page(domain->nid, GFP_ATOMIC);
 	if (!domain->pgd)
 		return -ENOMEM;
 	domain_flush_cache(domain, domain->pgd, PAGE_SIZE);
@@ -4298,7 +4302,7 @@ static int intel_iommu_map(struct iommu_domain *domain,
 	   the low bits of hpa would take us onto the next page */
 	size = aligned_nrpages(hpa, size);
 	return __domain_mapping(dmar_domain, iova >> VTD_PAGE_SHIFT,
-				hpa >> VTD_PAGE_SHIFT, size, prot);
+				hpa >> VTD_PAGE_SHIFT, size, prot, gfp);
 }
 
 static int intel_iommu_map_pages(struct iommu_domain *domain,
@@ -4333,7 +4337,8 @@ static size_t intel_iommu_unmap(struct iommu_domain *domain,
 
 	/* Cope with horrid API which requires us to unmap more than the
 	   size argument if it happens to be a large-page mapping. */
-	BUG_ON(!pfn_to_dma_pte(dmar_domain, iova >> VTD_PAGE_SHIFT, &level));
+	BUG_ON(!pfn_to_dma_pte(dmar_domain, iova >> VTD_PAGE_SHIFT, &level,
+			       GFP_ATOMIC));
 
 	if (size < VTD_PAGE_SIZE << level_to_offset_bits(level))
 		size = VTD_PAGE_SIZE << level_to_offset_bits(level);
@@ -4392,7 +4397,8 @@ static phys_addr_t intel_iommu_iova_to_phys(struct iommu_domain *domain,
 	int level = 0;
 	u64 phys = 0;
 
-	pte = pfn_to_dma_pte(dmar_domain, iova >> VTD_PAGE_SHIFT, &level);
+	pte = pfn_to_dma_pte(dmar_domain, iova >> VTD_PAGE_SHIFT, &level,
+			     GFP_ATOMIC);
 	if (pte && dma_pte_present(pte))
 		phys = dma_pte_addr(pte) +
 			(iova & (BIT_MASK(level_to_offset_bits(level) +
diff --git a/drivers/iommu/intel/iommu.h b/drivers/iommu/intel/iommu.h
index 06e61e474856..ca9a035e0110 100644
--- a/drivers/iommu/intel/iommu.h
+++ b/drivers/iommu/intel/iommu.h
@@ -737,7 +737,7 @@ int qi_submit_sync(struct intel_iommu *iommu, struct qi_desc *desc,
 
 extern int dmar_ir_support(void);
 
-void *alloc_pgtable_page(int node);
+void *alloc_pgtable_page(int node, gfp_t gfp);
 void free_pgtable_page(void *vaddr);
 void iommu_flush_write_buffer(struct intel_iommu *iommu);
 struct intel_iommu *device_to_iommu(struct device *dev, u8 *bus, u8 *devfn);
diff --git a/drivers/iommu/intel/pasid.c b/drivers/iommu/intel/pasid.c
index fb3c7020028d..c5bf74e9372d 100644
--- a/drivers/iommu/intel/pasid.c
+++ b/drivers/iommu/intel/pasid.c
@@ -200,7 +200,7 @@ static struct pasid_entry *intel_pasid_get_entry(struct device *dev, u32 pasid)
 retry:
 	entries = get_pasid_table_from_pde(&dir[dir_index]);
 	if (!entries) {
-		entries = alloc_pgtable_page(info->iommu->node);
+		entries = alloc_pgtable_page(info->iommu->node, GFP_ATOMIC);
 		if (!entries)
 			return NULL;