current/SOURCES/0001-i965-Don-t-do-the-temporary-and-blit-copy-for-INVALI.patch

From 3a6271890cd40c1e37870802f9f3231b3170d27f Mon Sep 17 00:00:00 2001
From: Eric Anholt <eric@anholt.net>
Date: Tue, 24 Dec 2013 15:11:29 -0800
Subject: [PATCH] i965: Don't do the temporary-and-blit-copy for
 INVALIDATE_RANGE maps.

We definitely want to fall through to the unsynchronized map case, instead
of wasting bandwidth on a copy.  Prevents a -43.2407% +/- 1.06113% (n=49)
performance regression on aa10perf when teaching glamor to provide the
GL_INVALIDATE_RANGE_BIT information.

This is a performance fix, which I usually wouldn't cherry-pick to stable.
But this was really was just a bug in the code, its presence would
discourage developers from giving us the best information they can, and I
think we've got fairly high confidence in the unsynchronized map path
already.

Cc: 10.0 9.2 <mesa-stable@lists.freedesktop.org>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
(cherry picked from commit f46563fe1c8a5560e4de0adf03e3d8770b7fc734)
---
 src/mesa/drivers/dri/i965/intel_buffer_objects.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/src/mesa/drivers/dri/i965/intel_buffer_objects.c b/src/mesa/drivers/dri/i965/intel_buffer_objects.c
index 779198f..a87ebf7 100644
--- a/src/mesa/drivers/dri/i965/intel_buffer_objects.c
+++ b/src/mesa/drivers/dri/i965/intel_buffer_objects.c
@@ -357,7 +357,8 @@ intel_bufferobj_map_range(struct gl_context * ctx,
     * BO, and we'll copy what they put in there out at unmap or
     * FlushRange time.
     */
-   if ((access & GL_MAP_INVALIDATE_RANGE_BIT) &&
+   if (!(access & GL_MAP_UNSYNCHRONIZED_BIT) &&
+       (access & GL_MAP_INVALIDATE_RANGE_BIT) &&
        drm_intel_bo_busy(intel_obj->buffer)) {
       if (access & GL_MAP_FLUSH_EXPLICIT_BIT) {
         intel_obj->range_map_buffer = malloc(length);
-- 
1.8.4.5

1	From 3a6271890cd40c1e37870802f9f3231b3170d27f Mon Sep 17 00:00:00 2001
2	From: Eric Anholt <eric@anholt.net>
3	Date: Tue, 24 Dec 2013 15:11:29 -0800
4	Subject: [PATCH] i965: Don't do the temporary-and-blit-copy for
5	INVALIDATE_RANGE maps.
6
7	We definitely want to fall through to the unsynchronized map case, instead
8	of wasting bandwidth on a copy. Prevents a -43.2407% +/- 1.06113% (n=49)
9	performance regression on aa10perf when teaching glamor to provide the
10	GL_INVALIDATE_RANGE_BIT information.
11
12	This is a performance fix, which I usually wouldn't cherry-pick to stable.
13	But this was really was just a bug in the code, its presence would
14	discourage developers from giving us the best information they can, and I
15	think we've got fairly high confidence in the unsynchronized map path
16	already.
17
18	Cc: 10.0 9.2 <mesa-stable@lists.freedesktop.org>
19	Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
20	(cherry picked from commit f46563fe1c8a5560e4de0adf03e3d8770b7fc734)
21	---
22	src/mesa/drivers/dri/i965/intel_buffer_objects.c \| 3 ++-
23	1 file changed, 2 insertions(+), 1 deletion(-)
24
25	diff --git a/src/mesa/drivers/dri/i965/intel_buffer_objects.c b/src/mesa/drivers/dri/i965/intel_buffer_objects.c
26	index 779198f..a87ebf7 100644
27	--- a/src/mesa/drivers/dri/i965/intel_buffer_objects.c
28	+++ b/src/mesa/drivers/dri/i965/intel_buffer_objects.c
29	@@ -357,7 +357,8 @@ intel_bufferobj_map_range(struct gl_context * ctx,
30	* BO, and we'll copy what they put in there out at unmap or
31	* FlushRange time.
32	*/
33	- if ((access & GL_MAP_INVALIDATE_RANGE_BIT) &&
34	+ if (!(access & GL_MAP_UNSYNCHRONIZED_BIT) &&
35	+ (access & GL_MAP_INVALIDATE_RANGE_BIT) &&
36	drm_intel_bo_busy(intel_obj->buffer)) {
37	if (access & GL_MAP_FLUSH_EXPLICIT_BIT) {
38	intel_obj->range_map_buffer = malloc(length);
39	--
40	1.8.4.5
41