DMA, CMA: support alignment constraint on CMA region

PPC KVM's CMA area management needs alignment constraint on CMA region.
So support it to prepare generalization of CMA area management
functionality.

Additionally, add some comments which tell us why alignment constraint
is needed on CMA region.

Signed-off-by: Joonsoo Kim <[email protected]>
Acked-by: Michal Nazarewicz <[email protected]>
Reviewed-by: Aneesh Kumar K.V <[email protected]>
Cc: Alexander Graf <[email protected]>
Cc: Aneesh Kumar K.V <[email protected]>
Cc: Gleb Natapov <[email protected]>
Acked-by: Marek Szyprowski <[email protected]>
Tested-by: Marek Szyprowski <[email protected]>
Cc: Minchan Kim <[email protected]>
Cc: Paolo Bonzini <[email protected]>
Cc: Zhang Yanfei <[email protected]>
Cc: Benjamin Herrenschmidt <[email protected]>
Cc: Paul Mackerras <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>

1 files changed, 18 insertions, 8 deletions

diff --git a/drivers/base/dma-contiguous.c b/drivers/base/dma-contiguous.c
index 9021762227a7..5f62c284072c 100644
--- a/drivers/base/dma-contiguous.c
+++ b/drivers/base/dma-contiguous.c
@@ -32,6 +32,7 @@
 #include <linux/swap.h>
 #include <linux/mm_types.h>
 #include <linux/dma-contiguous.h>
+#include <linux/log2.h>
 
 struct cma {
 	unsigned long	base_pfn;
@@ -215,17 +216,16 @@ core_initcall(cma_init_reserved_areas);
 
 static int __init __dma_contiguous_reserve_area(phys_addr_t size,
 				phys_addr_t base, phys_addr_t limit,
+				phys_addr_t alignment,
 				struct cma **res_cma, bool fixed)
 {
 	struct cma *cma = &cma_areas[cma_area_count];
-	phys_addr_t alignment;
 	int ret = 0;
 
-	pr_debug("%s(size %lx, base %08lx, limit %08lx)\n", __func__,
-		 (unsigned long)size, (unsigned long)base,
-		 (unsigned long)limit);
+	pr_debug("%s(size %lx, base %08lx, limit %08lx alignment %08lx)\n",
+		__func__, (unsigned long)size, (unsigned long)base,
+		(unsigned long)limit, (unsigned long)alignment);
 
-	/* Sanity checks */
 	if (cma_area_count == ARRAY_SIZE(cma_areas)) {
 		pr_err("Not enough slots for CMA reserved regions!\n");
 		return -ENOSPC;
@@ -234,8 +234,17 @@ static int __init __dma_contiguous_reserve_area(phys_addr_t size,
 	if (!size)
 		return -EINVAL;
 
-	/* Sanitise input arguments */
-	alignment = PAGE_SIZE << max(MAX_ORDER - 1, pageblock_order);
+	if (alignment && !is_power_of_2(alignment))
+		return -EINVAL;
+
+	/*
+	 * Sanitise input arguments.
+	 * Pages both ends in CMA area could be merged into adjacent unmovable
+	 * migratetype page by page allocator's buddy algorithm. In the case,
+	 * you couldn't get a contiguous memory, which is not what we want.
+	 */
+	alignment = max(alignment,
+		(phys_addr_t)PAGE_SIZE << max(MAX_ORDER - 1, pageblock_order));
 	base = ALIGN(base, alignment);
 	size = ALIGN(size, alignment);
 	limit &= ~(alignment - 1);
@@ -299,7 +308,8 @@ int __init dma_contiguous_reserve_area(phys_addr_t size, phys_addr_t base,
 {
 	int ret;
 
-	ret = __dma_contiguous_reserve_area(size, base, limit, res_cma, fixed);
+	ret = __dma_contiguous_reserve_area(size, base, limit, 0,
+						res_cma, fixed);
 	if (ret)
 		return ret;