tegra:video:nvmap: optimize cache_maint operation.

video:tegra:nvmap: Clean whole L1 instead of cleaning by MVA

For large allocations, cleaning each page of the allocation can
take a significant amount of time. If an allocation that nvmap needs
to clean or invalidate out of the cache is significantly larger than
the cache, just flush the entire cache by set/ways.

bug 788967

Reviewed-on: http://git-master/r/19354
(cherry picked from commit c01c12e63b1476501204152356867aeb5091fb80)

tegra:video:nvmap: optimize cache_maint operation.

optimize cache_maint operation for carveout and heap memories.
flush carveout memory allocations on memory free.
Bug 761637

Reviewed-on: http://git-master/r/21205
Conflicts:
	drivers/video/tegra/nvmap/nvmap_dev.c
	drivers/video/tegra/nvmap/nvmap_heap.c
	drivers/video/tegra/nvmap/nvmap_ioctl.c

(cherry picked from commit 731df4df5e895e1d4999359d6d5939fc2095f883)

tegra:video:nvmap: optimize cache flush for system heap pages.

optimize cache flush for pages allocated from system heap.
Bug 788187

Reviewed-on: http://git-master/r/21687
(cherry picked from commit 3f318911ad91410aed53c90494210e2b8f74308b)

Original-Change-Id: Ia7b90ba0b50acfef1b88dd8095219c51733e027f
Reviewed-on: http://git-master/r/23465
Reviewed-by: Kirill Artamonov <kartamonov@nvidia.com>
Tested-by: Kirill Artamonov <kartamonov@nvidia.com>
Reviewed-by: Bharat Nihalani <bnihalani@nvidia.com>

Rebase-Id: R04f618f88ed1d2c7a680d51a8c5113f42de3f667

1 files changed, 16 insertions, 5 deletions

diff --git a/drivers/video/tegra/nvmap/nvmap_handle.c b/drivers/video/tegra/nvmap/nvmap_handle.c
index dc3be30ca2f5..a9150a36cf2a 100644
--- a/drivers/video/tegra/nvmap/nvmap_handle.c
+++ b/drivers/video/tegra/nvmap/nvmap_handle.c
@@ -37,6 +37,7 @@
 
 #include "nvmap.h"
 #include "nvmap_mru.h"
+#include "nvmap_common.h"
 
 #define NVMAP_SECURE_HEAPS	(NVMAP_HEAP_CARVEOUT_IRAM | NVMAP_HEAP_IOVMM)
 #ifdef CONFIG_NVMAP_HIGHMEM_ONLY
@@ -107,7 +108,8 @@ out:
 
 extern void __flush_dcache_page(struct address_space *, struct page *);
 
-static struct page *nvmap_alloc_pages_exact(gfp_t gfp, size_t size)
+static struct page *nvmap_alloc_pages_exact(gfp_t gfp,
+	size_t size, bool flush_inner)
 {
 	struct page *page, *p, *e;
 	unsigned int order;
@@ -127,8 +129,10 @@ static struct page *nvmap_alloc_pages_exact(gfp_t gfp, size_t size)
 		__free_page(p);
 
 	e = page + (size >> PAGE_SHIFT);
-	for (p = page; p < e; p++)
-		__flush_dcache_page(page_mapping(p), p);
+	if (flush_inner) {
+		for (p = page; p < e; p++)
+			__flush_dcache_page(page_mapping(p), p);
+	}
 
 	base = page_to_phys(page);
 	outer_flush_range(base, base + size);
@@ -143,6 +147,7 @@ static int handle_page_alloc(struct nvmap_client *client,
 	pgprot_t prot;
 	unsigned int i = 0;
 	struct page **pages;
+	bool flush_inner = true;
 
 	pages = altalloc(nr_page * sizeof(*pages));
 	if (!pages)
@@ -155,10 +160,14 @@ static int handle_page_alloc(struct nvmap_client *client,
 		contiguous = true;
 #endif
 
+	if (size >= FLUSH_CLEAN_BY_SET_WAY_THRESHOLD) {
+		inner_flush_cache_all();
+		flush_inner = false;
+	}
 	h->pgalloc.area = NULL;
 	if (contiguous) {
 		struct page *page;
-		page = nvmap_alloc_pages_exact(GFP_NVMAP, size);
+		page = nvmap_alloc_pages_exact(GFP_NVMAP, size, flush_inner);
 		if (!page)
 			goto fail;
 
@@ -167,7 +176,8 @@ static int handle_page_alloc(struct nvmap_client *client,
 
 	} else {
 		for (i = 0; i < nr_page; i++) {
-			pages[i] = nvmap_alloc_pages_exact(GFP_NVMAP, PAGE_SIZE);
+			pages[i] = nvmap_alloc_pages_exact(GFP_NVMAP, PAGE_SIZE,
+				flush_inner);
 			if (!pages[i])
 				goto fail;
 		}
@@ -193,6 +203,7 @@ fail:
 	while (i--)
 		__free_page(pages[i]);
 	altfree(pages, nr_page * sizeof(*pages));
+	wmb();
 	return -ENOMEM;
 }