From: Tomasz Nowicki Date: Wed, 20 Sep 2017 08:52:02 +0000 (+0200) Subject: iommu/iova: Make rcache flush optional on IOVA allocation failure X-Git-Url: http://git.maquefel.me/?a=commitdiff_plain;h=538d5b333216c3daa7a5821307164f10af73ec8c;p=linux.git iommu/iova: Make rcache flush optional on IOVA allocation failure Since IOVA allocation failure is not unusual case we need to flush CPUs' rcache in hope we will succeed in next round. However, it is useful to decide whether we need rcache flush step because of two reasons: - Not scalability. On large system with ~100 CPUs iterating and flushing rcache for each CPU becomes serious bottleneck so we may want to defer it. - free_cpu_cached_iovas() does not care about max PFN we are interested in. Thus we may flush our rcaches and still get no new IOVA like in the commonly used scenario: if (dma_limit > DMA_BIT_MASK(32) && dev_is_pci(dev)) iova = alloc_iova_fast(iovad, iova_len, DMA_BIT_MASK(32) >> shift); if (!iova) iova = alloc_iova_fast(iovad, iova_len, dma_limit >> shift); 1. First alloc_iova_fast() call is limited to DMA_BIT_MASK(32) to get PCI devices a SAC address 2. alloc_iova() fails due to full 32-bit space 3. rcaches contain PFNs out of 32-bit space so free_cpu_cached_iovas() throws entries away for nothing and alloc_iova() fails again 4. Next alloc_iova_fast() call cannot take advantage of rcache since we have just defeated caches. In this case we pick the slowest option to proceed. This patch reworks flushed_rcache local flag to be additional function argument instead and control rcache flush step. Also, it updates all users to do the flush as the last chance. Signed-off-by: Tomasz Nowicki Reviewed-by: Robin Murphy Tested-by: Nate Watterson Signed-off-by: Joerg Roedel --- diff --git a/drivers/iommu/amd_iommu.c b/drivers/iommu/amd_iommu.c index 647ab7691aeeb..3d64c844d8b10 100644 --- a/drivers/iommu/amd_iommu.c +++ b/drivers/iommu/amd_iommu.c @@ -1546,10 +1546,11 @@ static unsigned long dma_ops_alloc_iova(struct device *dev, if (dma_mask > DMA_BIT_MASK(32)) pfn = alloc_iova_fast(&dma_dom->iovad, pages, - IOVA_PFN(DMA_BIT_MASK(32))); + IOVA_PFN(DMA_BIT_MASK(32)), false); if (!pfn) - pfn = alloc_iova_fast(&dma_dom->iovad, pages, IOVA_PFN(dma_mask)); + pfn = alloc_iova_fast(&dma_dom->iovad, pages, + IOVA_PFN(dma_mask), true); return (pfn << PAGE_SHIFT); } diff --git a/drivers/iommu/dma-iommu.c b/drivers/iommu/dma-iommu.c index 191be9c80a8af..25914d36c5ace 100644 --- a/drivers/iommu/dma-iommu.c +++ b/drivers/iommu/dma-iommu.c @@ -370,10 +370,12 @@ static dma_addr_t iommu_dma_alloc_iova(struct iommu_domain *domain, /* Try to get PCI devices a SAC address */ if (dma_limit > DMA_BIT_MASK(32) && dev_is_pci(dev)) - iova = alloc_iova_fast(iovad, iova_len, DMA_BIT_MASK(32) >> shift); + iova = alloc_iova_fast(iovad, iova_len, + DMA_BIT_MASK(32) >> shift, false); if (!iova) - iova = alloc_iova_fast(iovad, iova_len, dma_limit >> shift); + iova = alloc_iova_fast(iovad, iova_len, dma_limit >> shift, + true); return (dma_addr_t)iova << shift; } diff --git a/drivers/iommu/intel-iommu.c b/drivers/iommu/intel-iommu.c index ebb48353dd39a..b3914fce82544 100644 --- a/drivers/iommu/intel-iommu.c +++ b/drivers/iommu/intel-iommu.c @@ -3469,11 +3469,12 @@ static unsigned long intel_alloc_iova(struct device *dev, * from higher range */ iova_pfn = alloc_iova_fast(&domain->iovad, nrpages, - IOVA_PFN(DMA_BIT_MASK(32))); + IOVA_PFN(DMA_BIT_MASK(32)), false); if (iova_pfn) return iova_pfn; } - iova_pfn = alloc_iova_fast(&domain->iovad, nrpages, IOVA_PFN(dma_mask)); + iova_pfn = alloc_iova_fast(&domain->iovad, nrpages, + IOVA_PFN(dma_mask), true); if (unlikely(!iova_pfn)) { pr_err("Allocating %ld-page iova for %s failed", nrpages, dev_name(dev)); diff --git a/drivers/iommu/iova.c b/drivers/iommu/iova.c index 3aee64b99df12..84bda3a4dafc7 100644 --- a/drivers/iommu/iova.c +++ b/drivers/iommu/iova.c @@ -395,14 +395,15 @@ EXPORT_SYMBOL_GPL(free_iova); * @iovad: - iova domain in question * @size: - size of page frames to allocate * @limit_pfn: - max limit address + * @flush_rcache: - set to flush rcache on regular allocation failure * This function tries to satisfy an iova allocation from the rcache, - * and falls back to regular allocation on failure. + * and falls back to regular allocation on failure. If regular allocation + * fails too and the flush_rcache flag is set then the rcache will be flushed. */ unsigned long alloc_iova_fast(struct iova_domain *iovad, unsigned long size, - unsigned long limit_pfn) + unsigned long limit_pfn, bool flush_rcache) { - bool flushed_rcache = false; unsigned long iova_pfn; struct iova *new_iova; @@ -415,11 +416,11 @@ retry: if (!new_iova) { unsigned int cpu; - if (flushed_rcache) + if (!flush_rcache) return 0; /* Try replenishing IOVAs by flushing rcache. */ - flushed_rcache = true; + flush_rcache = false; for_each_online_cpu(cpu) free_cpu_cached_iovas(cpu, iovad); goto retry; diff --git a/include/linux/iova.h b/include/linux/iova.h index c696ee81054ed..928442dda565f 100644 --- a/include/linux/iova.h +++ b/include/linux/iova.h @@ -150,7 +150,7 @@ void queue_iova(struct iova_domain *iovad, unsigned long pfn, unsigned long pages, unsigned long data); unsigned long alloc_iova_fast(struct iova_domain *iovad, unsigned long size, - unsigned long limit_pfn); + unsigned long limit_pfn, bool flush_rcache); struct iova *reserve_iova(struct iova_domain *iovad, unsigned long pfn_lo, unsigned long pfn_hi); void copy_reserved_iova(struct iova_domain *from, struct iova_domain *to); @@ -212,7 +212,8 @@ static inline void queue_iova(struct iova_domain *iovad, static inline unsigned long alloc_iova_fast(struct iova_domain *iovad, unsigned long size, - unsigned long limit_pfn) + unsigned long limit_pfn, + bool flush_rcache) { return 0; }