[PATCH] iommu/iova: using separate rcache for SAC and DAC

[brookxu.cn]

From: Chunguang Xu <chunguang.xu@xxxxxxxxxx>

While iommu_dma_forcedac disable, for PCI device kernel
try SAC first, if failed then try DAC. Since now rcache
does not distinguish SAC and DAC, if all PFNs contained
in cpu loaded cache is larger than SAC max PFN, but the
SAC address space is sufficient, as cpu loaded cached is
not empty, kernel will iova_alloc () to alloc IOVA. For
PCI device, kernel alloc SAC most, loaded cache may
invalid for SAC alloc for a long time, kernel will enter
alloc_iova() slow path frequencely, as result performance
is degrade. To circumvent this problem, SAC and DAC maybe
better to use separate caches.

Signed-off-by: Chunguang Xu <chunguang.xu@xxxxxxxxxx>
---
 drivers/iommu/iova.c | 28 +++++++++++++++++++++-------
 1 file changed, 21 insertions(+), 7 deletions(-)

diff --git a/drivers/iommu/iova.c b/drivers/iommu/iova.c
index 47d1983dfa2a..d5775719a143 100644
--- a/drivers/iommu/iova.c
+++ b/drivers/iommu/iova.c
@@ -16,6 +16,7 @@
 #define IOVA_ANCHOR	~0UL
 
 #define IOVA_RANGE_CACHE_MAX_SIZE 6	/* log of max cached IOVA range size (in pages) */
+#define IOVA_RANGE_CACHE_ARRAY_SIZE (2 * IOVA_RANGE_CACHE_MAX_SIZE)
 
 static bool iova_rcache_insert(struct iova_domain *iovad,
 			       unsigned long pfn,
@@ -723,13 +724,13 @@ int iova_domain_init_rcaches(struct iova_domain *iovad)
 	unsigned int cpu;
 	int i, ret;
 
-	iovad->rcaches = kcalloc(IOVA_RANGE_CACHE_MAX_SIZE,
+	iovad->rcaches = kcalloc(IOVA_RANGE_CACHE_ARRAY_SIZE,
 				 sizeof(struct iova_rcache),
 				 GFP_KERNEL);
 	if (!iovad->rcaches)
 		return -ENOMEM;
 
-	for (i = 0; i < IOVA_RANGE_CACHE_MAX_SIZE; ++i) {
+	for (i = 0; i < IOVA_RANGE_CACHE_ARRAY_SIZE; ++i) {
 		struct iova_cpu_rcache *cpu_rcache;
 		struct iova_rcache *rcache;
 
@@ -825,11 +826,15 @@ static bool iova_rcache_insert(struct iova_domain *iovad, unsigned long pfn,
 			       unsigned long size)
 {
 	unsigned int log_size = order_base_2(size);
+	unsigned int index;
 
 	if (log_size >= IOVA_RANGE_CACHE_MAX_SIZE)
 		return false;
 
-	return __iova_rcache_insert(iovad, &iovad->rcaches[log_size], pfn);
+	if (pfn > DMA_BIT_MASK(32))
+		index = log_size + IOVA_RANGE_CACHE_MAX_SIZE;
+
+	return __iova_rcache_insert(iovad, &iovad->rcaches[index], pfn);
 }
 
 /*
@@ -881,11 +886,20 @@ static unsigned long iova_rcache_get(struct iova_domain *iovad,
 				     unsigned long limit_pfn)
 {
 	unsigned int log_size = order_base_2(size);
+	unsigned long iova_pfn;
+	unsigned int index;
 
 	if (log_size >= IOVA_RANGE_CACHE_MAX_SIZE || !iovad->rcaches)
 		return 0;
 
-	return __iova_rcache_get(&iovad->rcaches[log_size], limit_pfn - size);
+	iova_pfn = __iova_rcache_get(&iovad->rcaches[log_size], limit_pfn - size);
+
+	if (!iova_pfn && limit_pfn > DMA_BIT_MASK(32)) {
+		index = log_size + IOVA_RANGE_CACHE_MAX_SIZE;
+		iova_pfn = __iova_rcache_get(&iovad->rcaches[index], limit_pfn - size);
+	}
+
+	return iova_pfn
 }
 
 /*
@@ -898,7 +912,7 @@ static void free_iova_rcaches(struct iova_domain *iovad)
 	unsigned int cpu;
 	int i, j;
 
-	for (i = 0; i < IOVA_RANGE_CACHE_MAX_SIZE; ++i) {
+	for (i = 0; i < IOVA_RANGE_CACHE_ARRAY_SIZE; ++i) {
 		rcache = &iovad->rcaches[i];
 		if (!rcache->cpu_rcaches)
 			break;
@@ -926,7 +940,7 @@ static void free_cpu_cached_iovas(unsigned int cpu, struct iova_domain *iovad)
 	unsigned long flags;
 	int i;
 
-	for (i = 0; i < IOVA_RANGE_CACHE_MAX_SIZE; ++i) {
+	for (i = 0; i < IOVA_RANGE_CACHE_ARRAY_SIZE; ++i) {
 		rcache = &iovad->rcaches[i];
 		cpu_rcache = per_cpu_ptr(rcache->cpu_rcaches, cpu);
 		spin_lock_irqsave(&cpu_rcache->lock, flags);
@@ -945,7 +959,7 @@ static void free_global_cached_iovas(struct iova_domain *iovad)
 	unsigned long flags;
 	int i, j;
 
-	for (i = 0; i < IOVA_RANGE_CACHE_MAX_SIZE; ++i) {
+	for (i = 0; i < IOVA_RANGE_CACHE_ARRAY_SIZE; ++i) {
 		rcache = &iovad->rcaches[i];
 		spin_lock_irqsave(&rcache->lock, flags);
 		for (j = 0; j < rcache->depot_size; ++j) {
-- 
2.31.1