Skip to contentSkip to conversation

iommu/io-pgtable-arm: Make allocations NUMA-aware

PATCH New Robin Murphy 3 participants 3 messages May 22, 2018 – May 29, 2018

Subsystems (2)
lore
Share & tools
lore

Robin Murphy

author reviewer
#0 lore raw

PATCH v2 1/1 iommu/io-pgtable-arm: Make allocations NUMA-aware

We would generally expect pagetables to be read by the IOMMU more than written by the CPU, so in NUMA systems it makes sense to locate them close to the former and avoid cross-node pagetable walks if at all possible. As it turns out, we already have a handle on the IOMMU device for the sake of coherency management, so it's trivial to grab the appropriate NUMA node when allocating new pagetable pages.

Note that we drop the semantics of alloc_pages_exact(), but that's fine since they have never been necessary: the only time we're allocating more than one page is for stage 2 top-level concatenation, but since that is based on the number of IPA bits, the size is always some exact power of two anyway.

---

v2: Retain equivalent highmem check

 drivers/iommu/io-pgtable-arm.c | 13 +++++++++----
 1 file changed, 9 insertions(+), 4 deletions(-)
drivers/iommu/io-pgtable-arm.c +9 -4
index 39c2a056da21..5194755bba29 100644
@@ -231,12 +231,17 @@ static void *__arm_lpae_alloc_pages(size_t size, gfp_t gfp,
 				    struct io_pgtable_cfg *cfg)
 {
 	struct device *dev = cfg->iommu_dev;
+	int order = get_order(size);
+	struct page *p;
 	dma_addr_t dma;
-	void *pages = alloc_pages_exact(size, gfp | __GFP_ZERO);
+	void *pages;
 
-	if (!pages)
+	VM_BUG_ON((gfp & __GFP_HIGHMEM));
+	p = alloc_pages_node(dev_to_node(dev), gfp | __GFP_ZERO, order);
+	if (!p)
 		return NULL;
 
+	pages = page_address(p);
 	if (!(cfg->quirks & IO_PGTABLE_QUIRK_NO_DMA)) {
 		dma = dma_map_single(dev, pages, size, DMA_TO_DEVICE);
 		if (dma_mapping_error(dev, dma))
@@ -256,7 +261,7 @@ static void *__arm_lpae_alloc_pages(size_t size, gfp_t gfp,
 	dev_err(dev, "Cannot accommodate DMA translation for IOMMU page tables\n");
 	dma_unmap_single(dev, dma, size, DMA_TO_DEVICE);
 out_free:
-	free_pages_exact(pages, size);
+	__free_pages(p, order);
 	return NULL;
 }
 
@@ -266,7 +271,7 @@ static void __arm_lpae_free_pages(void *pages, size_t size,
 	if (!(cfg->quirks & IO_PGTABLE_QUIRK_NO_DMA))
 		dma_unmap_single(cfg->iommu_dev, __arm_lpae_dma_addr(pages),
 				 size, DMA_TO_DEVICE);
-	free_pages_exact(pages, size);
+	free_pages((unsigned long)pages, get_order(size));
 }
 
 static void __arm_lpae_sync_pte(arm_lpae_iopte *ptep,
Signature · 1 line
2.17.0.dirty
#1 lore raw

PATCH v2 1/1 iommu/io-pgtable-arm: Make allocations NUMA-aware

Quoted 13 lines · Robin Murphy

We would generally expect pagetables to be read by the IOMMU more than written by the CPU, so in NUMA systems it makes sense to locate them close to the former and avoid cross-node pagetable walks if at all possible. As it turns out, we already have a handle on the IOMMU device for the sake of coherency management, so it's trivial to grab the appropriate NUMA node when allocating new pagetable pages.

Note that we drop the semantics of alloc_pages_exact(), but that's fine since they have never been necessary: the only time we're allocating more than one page is for stage 2 top-level concatenation, but since that is based on the number of IPA bits, the size is always some exact power of two anyway.

Quoted 8 lines

---

v2: Retain equivalent highmem check

 drivers/iommu/io-pgtable-arm.c | 13 +++++++++----
 1 file changed, 9 insertions(+), 4 deletions(-)

Thanks, Robin.

Joerg -- please you can pick this up for 4.18? I don't have any other SMMU patches pending, so it doesn't seem worth putting together a pull request just for this.

Cheers,

Will

Quoted 45 lines
drivers/iommu/io-pgtable-arm.c +9 -4
index 39c2a056da21..5194755bba29 100644
@@ -231,12 +231,17 @@ static void *__arm_lpae_alloc_pages(size_t size, gfp_t gfp,
 				    struct io_pgtable_cfg *cfg)
 {
 	struct device *dev = cfg->iommu_dev;
+	int order = get_order(size);
+	struct page *p;
 	dma_addr_t dma;
-	void *pages = alloc_pages_exact(size, gfp | __GFP_ZERO);
+	void *pages;
 
-	if (!pages)
+	VM_BUG_ON((gfp & __GFP_HIGHMEM));
+	p = alloc_pages_node(dev_to_node(dev), gfp | __GFP_ZERO, order);
+	if (!p)
 		return NULL;
 
+	pages = page_address(p);
 	if (!(cfg->quirks & IO_PGTABLE_QUIRK_NO_DMA)) {
 		dma = dma_map_single(dev, pages, size, DMA_TO_DEVICE);
 		if (dma_mapping_error(dev, dma))
@@ -256,7 +261,7 @@ static void *__arm_lpae_alloc_pages(size_t size, gfp_t gfp,
 	dev_err(dev, "Cannot accommodate DMA translation for IOMMU page tables\n");
 	dma_unmap_single(dev, dma, size, DMA_TO_DEVICE);
 out_free:
-	free_pages_exact(pages, size);
+	__free_pages(p, order);
 	return NULL;
 }
 
@@ -266,7 +271,7 @@ static void __arm_lpae_free_pages(void *pages, size_t size,
 	if (!(cfg->quirks & IO_PGTABLE_QUIRK_NO_DMA))
 		dma_unmap_single(cfg->iommu_dev, __arm_lpae_dma_addr(pages),
 				 size, DMA_TO_DEVICE);
-	free_pages_exact(pages, size);
+	free_pages((unsigned long)pages, get_order(size));
 }
 
 static void __arm_lpae_sync_pte(arm_lpae_iopte *ptep,
Signature · 1 line
2.17.0.dirty