]> git.proxmox.com Git - mirror_ubuntu-hirsute-kernel.git/commitdiff
drm/amdgpu: remove sorting of CS BOs
authorChristian König <christian.koenig@amd.com>
Fri, 15 Apr 2016 15:19:17 +0000 (17:19 +0200)
committerAlex Deucher <alexander.deucher@amd.com>
Thu, 5 May 2016 00:23:12 +0000 (20:23 -0400)
Not needed any more.

Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c

index d7e0b0b9a1bcb8965af3042fc57ee74d882bbc0d..2ebba295d0e4fabdd499ad2da860e75141d94a6a 100644 (file)
@@ -24,7 +24,6 @@
  * Authors:
  *    Jerome Glisse <glisse@freedesktop.org>
  */
-#include <linux/list_sort.h>
 #include <linux/pagemap.h>
 #include <drm/drmP.h>
 #include <drm/amdgpu_drm.h>
@@ -527,16 +526,6 @@ static int amdgpu_cs_sync_rings(struct amdgpu_cs_parser *p)
        return 0;
 }
 
-static int cmp_size_smaller_first(void *priv, struct list_head *a,
-                                 struct list_head *b)
-{
-       struct amdgpu_bo_list_entry *la = list_entry(a, struct amdgpu_bo_list_entry, tv.head);
-       struct amdgpu_bo_list_entry *lb = list_entry(b, struct amdgpu_bo_list_entry, tv.head);
-
-       /* Sort A before B if A is smaller. */
-       return (int)la->robj->tbo.num_pages - (int)lb->robj->tbo.num_pages;
-}
-
 /**
  * cs_parser_fini() - clean parser states
  * @parser:    parser structure holding parsing context.
@@ -553,18 +542,6 @@ static void amdgpu_cs_parser_fini(struct amdgpu_cs_parser *parser, int error, bo
        if (!error) {
                amdgpu_vm_move_pt_bos_in_lru(parser->adev, &fpriv->vm);
 
-               /* Sort the buffer list from the smallest to largest buffer,
-                * which affects the order of buffers in the LRU list.
-                * This assures that the smallest buffers are added first
-                * to the LRU list, so they are likely to be later evicted
-                * first, instead of large buffers whose eviction is more
-                * expensive.
-                *
-                * This slightly lowers the number of bytes moved by TTM
-                * per frame under memory pressure.
-                */
-               list_sort(NULL, &parser->validated, cmp_size_smaller_first);
-
                ttm_eu_fence_buffer_objects(&parser->ticket,
                                            &parser->validated,
                                            parser->fence);