[Mesa-dev] [PATCH][RFC] mesa/main: Clamp rgba with streamed sse

2014-10-31 Thread Juha-Pekka Heikkila
Signed-off-by: Juha-Pekka Heikkila --- src/mesa/main/colormac.h | 20 +++ src/mesa/main/pixeltransfer.c | 59 --- 2 files changed, 64 insertions(+), 15 deletions(-) diff --git a/src/mesa/main/colormac.h b/src/mesa/main/colormac.h index c8a

[Mesa-dev] [PATCH][RFC] Clamp rgba floats with sse

2014-10-31 Thread Juha-Pekka Heikkila
I was motivated to send this to list after seeing Timothy Arceri's add SSE optimization for glDrawElements patch. This patch is not very nice with its ifdefs which are needed to get gcc to optimize loops when possible and still avoid sse code where at the minimum sse2 is not present. What this ca

Re: [Mesa-dev] [PATCH][RFC] mesa/main: Clamp rgba with streamed sse

2014-10-31 Thread Jose Fonseca
On 31/10/14 10:13, Juha-Pekka Heikkila wrote: defined(__SSE2__) && defined(__GNUC__) Instead of duplicate this expression everywhere lets create a "HAVE_SSE2_INTRIN" define. Not only this expression is complex, it will become even more when we updated it for MSVC. See src/gallium/include

[Mesa-dev] [demos PATCH] glxinfo/wglinfo: reverse order of the gl_versions[] array

2014-10-31 Thread Brian Paul
By going from newest to oldest we can simplify the glxinfo/wglinfo loops over the possible GL versions. --- src/wgl/wglinfo.c |2 +- src/xdemos/glinfo_common.h | 40 +--- src/xdemos/glxinfo.c |2 +- 3 files changed, 23 insertions(+), 21

[Mesa-dev] [Bugzilla] Your account mesa-dev@lists.freedesktop.org is being impersonated

2014-10-31 Thread bugzilla-daemon
ajax at nwnk dot net has used the 'sudo' feature to access Bugzilla using your account. ajax at nwnk dot net provided the following reason for doing this: fixing email settings If you feel that this action was inappropriate, please contact sitewrangl...@lists.freedesktop.org. Fo

Re: [Mesa-dev] [PATCH][RFC] mesa/main: Clamp rgba with streamed sse

2014-10-31 Thread Matt Turner
On Fri, Oct 31, 2014 at 4:12 AM, Jose Fonseca wrote: > On 31/10/14 10:13, Juha-Pekka Heikkila wrote: >> >> defined(__SSE2__) && defined(__GNUC__) > > > Instead of duplicate this expression everywhere lets create a > "HAVE_SSE2_INTRIN" define. Not only this expression is complex, it will > becom

Re: [Mesa-dev] [Bugzilla] Your account mesa-dev@lists.freedesktop.org is being impersonated

2014-10-31 Thread Marek Olšák
Hi everyone, Just so you know what this is all about: I requested that the "New: " prefix be removed from the first emailed message from bug reports. If you consider it annoying too, I recommend going to bugs.freedesktop.org, Preferences -> General Preferences tab -> at the bottom "Add 'New:' to

Re: [Mesa-dev] [PATCH][RFC] mesa/main: Clamp rgba with streamed sse

2014-10-31 Thread Matt Turner
On Fri, Oct 31, 2014 at 3:13 AM, Juha-Pekka Heikkila wrote: > Signed-off-by: Juha-Pekka Heikkila > --- > src/mesa/main/colormac.h | 20 +++ > src/mesa/main/pixeltransfer.c | 59 > --- > 2 files changed, 64 insertions(+), 15 deletions(-) >

Re: [Mesa-dev] [PATCH][RFC] mesa/main: Clamp rgba with streamed sse

2014-10-31 Thread Jose Fonseca
On 31/10/14 17:01, Matt Turner wrote: On Fri, Oct 31, 2014 at 4:12 AM, Jose Fonseca wrote: On 31/10/14 10:13, Juha-Pekka Heikkila wrote: defined(__SSE2__) && defined(__GNUC__) Instead of duplicate this expression everywhere lets create a "HAVE_SSE2_INTRIN" define. Not only this express

[Mesa-dev] [PATCH 2/2] glsl: Remove now useless dot optimization on basis vect

2014-10-31 Thread Matt Turner
The optimization in commit d056863b covers these cases, which were the first optimizations I added to the GLSL compiler. --- src/glsl/ir.cpp| 48 -- src/glsl/ir.h | 24 +++ src/glsl/opt_algebraic.cpp | 23

[Mesa-dev] [PATCH 1/2] glsl: Emit mul instead of dot if only one component left.

2014-10-31 Thread Matt Turner
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=85683 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=85691 --- src/glsl/opt_algebraic.cpp | 5 - 1 file changed, 4 insertions(+), 1 deletion(-) diff --git a/src/glsl/opt_algebraic.cpp b/src/glsl/opt_algebraic.cpp index c8d1ba1..a2

[Mesa-dev] [Bug 85712] glClear can hang when the window being rendered is concurrently closed

2014-10-31 Thread bugzilla-daemon
https://bugs.freedesktop.org/show_bug.cgi?id=85712 Bug ID: 85712 Summary: glClear can hang when the window being rendered is concurrently closed Product: Mesa Version: unspecified Hardware: All OS: All

[Mesa-dev] [Bug 85712] glClear can hang when the window being rendered is concurrently closed

2014-10-31 Thread bugzilla-daemon
https://bugs.freedesktop.org/show_bug.cgi?id=85712 --- Comment #1 from Steven Stewart-Gallus --- Created attachment 108741 --> https://bugs.freedesktop.org/attachment.cgi?id=108741&action=edit glxinfo output -- You are receiving this mail because: You are the assignee for the bug. ___

[Mesa-dev] [Bug 85712] glClear can hang when the window being rendered is concurrently closed

2014-10-31 Thread bugzilla-daemon
https://bugs.freedesktop.org/show_bug.cgi?id=85712 --- Comment #2 from Steven Stewart-Gallus --- I suspect the problem can occur on machines other than my Intel machine but I can't confirm. The problem occurs when one closes a window while it is being rendered to with glClear. The xcb_wait_for_re

Re: [Mesa-dev] [PATCH][RFC] mesa/main: Clamp rgba with streamed sse

2014-10-31 Thread Roland Scheidegger
Am 31.10.2014 um 18:17 schrieb Matt Turner: > On Fri, Oct 31, 2014 at 3:13 AM, Juha-Pekka Heikkila > wrote: >> Signed-off-by: Juha-Pekka Heikkila >> --- >> src/mesa/main/colormac.h | 20 +++ >> src/mesa/main/pixeltransfer.c | 59 >> --- >>

[Mesa-dev] [PATCH 3/3] i965/fs: Remove force uncompressed stack.

2014-10-31 Thread Matt Turner
Last use was in shader_time. --- src/mesa/drivers/dri/i965/brw_fs.cpp | 17 - src/mesa/drivers/dri/i965/brw_fs.h | 5 - src/mesa/drivers/dri/i965/brw_fs_visitor.cpp | 5 - 3 files changed, 27 deletions(-) diff --git a/src/mesa/drivers/dri/i965/brw_fs.cp

[Mesa-dev] [PATCH 2/3] i965/fs: Use execution size of 1 for some shader_time operations.

2014-10-31 Thread Matt Turner
The ADDs depended on dispatch_width, which really isn't what we wanted. --- src/mesa/drivers/dri/i965/brw_fs.cpp | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/src/mesa/drivers/dri/i965/brw_fs.cpp b/src/mesa/drivers/dri/i965/brw_fs.cpp index b5b9b0f..e6b284b 100644 --- a/src/

[Mesa-dev] [PATCH 1/3] i965/fs: Use mov(4) instructions to read timestamp.

2014-10-31 Thread Matt Turner
We only want fields 0-2. --- src/mesa/drivers/dri/i965/brw_fs.cpp | 9 - 1 file changed, 4 insertions(+), 5 deletions(-) diff --git a/src/mesa/drivers/dri/i965/brw_fs.cpp b/src/mesa/drivers/dri/i965/brw_fs.cpp index aa1d8d2..b5b9b0f 100644 --- a/src/mesa/drivers/dri/i965/brw_fs.cpp +++ b

[Mesa-dev] [Bug 85712] glClear can hang when the window being rendered is concurrently closed

2014-10-31 Thread bugzilla-daemon
https://bugs.freedesktop.org/show_bug.cgi?id=85712 --- Comment #3 from Steven Stewart-Gallus --- Created attachment 108743 --> https://bugs.freedesktop.org/attachment.cgi?id=108743&action=edit A very hacky test case One can use this test case and kill the child to reproduce the bug. -- You a

Re: [Mesa-dev] [PATCH 4/7] util: simplify temp register selection in u_pstipple.c

2014-10-31 Thread Charmaine Lee
The series looks good. Just a minor comment below. Reviewed-by: Charmaine Lee >From: mesa-dev on behalf of Brian >Paul >Sent: Thursday, October 30, 2014 8:04 PM >To: mesa-dev@lists.freedesktop.org >Subject: [Mesa-dev] [PATCH 4/7] util: simplify temp register selection in > u_pstipple.c

Re: [Mesa-dev] [PATCH 4/7] util: simplify temp register selection in u_pstipple.c

2014-10-31 Thread Brian Paul
On 10/31/2014 02:52 PM, Charmaine Lee wrote: The series looks good. Just a minor comment below. Reviewed-by: Charmaine Lee From: mesa-dev on behalf of Brian Paul Sent: Thursday, October 30, 2014 8:04 PM To: mesa-dev@lists.freedesktop.org Subject: [Mesa-dev] [PATCH 4/7] util: simplify temp

[Mesa-dev] [PATCH] gallium/docs: fix NRM, NRM4 docs

2014-10-31 Thread Brian Paul
Need to do a sqrt(). FWIW, the html that Sphinx 1.1.3 generates for the math expressions looks completely broken. --- src/gallium/docs/source/tgsi.rst | 36 1 file changed, 24 insertions(+), 12 deletions(-) diff --git a/src/gallium/docs/source/tgsi.rst b/sr

Re: [Mesa-dev] [PATCH 3/3] i965/fs: Remove force uncompressed stack.

2014-10-31 Thread Jason Ekstrand
Yes, please! I can't really comment on patch 1, but 2 and 3 are Reviewed-by: Jason Ekstrand On Fri, Oct 31, 2014 at 12:05 PM, Matt Turner wrote: > Last use was in shader_time. > --- > src/mesa/drivers/dri/i965/brw_fs.cpp | 17 - > src/mesa/drivers/dri/i965/brw_fs.h

Re: [Mesa-dev] [PATCH] gallium/docs: fix NRM, NRM4 docs

2014-10-31 Thread Jose Fonseca
Looks good. AFAICT these opcodes are never generated, and no driver except llvmpipe/softpipe/svga supports them. The performance is dominated by the square root reciprocate which has its own opcode, RSQ, so personally I'm for removing these opcodes. Jose

Re: [Mesa-dev] [PATCH] gallium/docs: fix NRM, NRM4 docs

2014-10-31 Thread Brian Paul
DX9 ps3/vs3 has a NRM (3-component) instruction so there's that potential use, I guess. -Brian On 10/31/2014 04:52 PM, Jose Fonseca wrote: Looks good. AFAICT these opcodes are never generated, and no driver except llvmpipe/softpipe/svga supports them. The performance is dominated by the sq

[Mesa-dev] [PATCH 2/2] i965: Remove non-existent vertical strides from array.

2014-10-31 Thread Matt Turner
These never existed, as far as I can tell. --- src/mesa/drivers/dri/i965/brw_eu_emit.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/src/mesa/drivers/dri/i965/brw_eu_emit.c b/src/mesa/drivers/dri/i965/brw_eu_emit.c index bb6334b..b611e77 100644 --- a/src/mesa/drivers/dri/i96

[Mesa-dev] [PATCH 1/2] i965: Convert stride/width/execution size macros into enums.

2014-10-31 Thread Matt Turner
--- src/mesa/drivers/dri/i965/brw_defines.h | 61 ++--- 1 file changed, 33 insertions(+), 28 deletions(-) diff --git a/src/mesa/drivers/dri/i965/brw_defines.h b/src/mesa/drivers/dri/i965/brw_defines.h index a6807ae..0116347 100644 --- a/src/mesa/drivers/dri/i965/brw_d

[Mesa-dev] [RFC 5/9] i965/cfg: Add functions to get first and last non-CF instructions.

2014-10-31 Thread Matt Turner
--- src/mesa/drivers/dri/i965/brw_cfg.h | 74 + 1 file changed, 74 insertions(+) diff --git a/src/mesa/drivers/dri/i965/brw_cfg.h b/src/mesa/drivers/dri/i965/brw_cfg.h index e6403bd..c2029cc 100644 --- a/src/mesa/drivers/dri/i965/brw_cfg.h +++ b/src/mesa/drive

[Mesa-dev] [RFC 7/9] i965/fs: Add pass to combine immediates.

2014-10-31 Thread Matt Turner
--- src/mesa/drivers/dri/i965/Makefile.sources | 1 + src/mesa/drivers/dri/i965/brw_fs.cpp | 2 + src/mesa/drivers/dri/i965/brw_fs.h | 1 + .../drivers/dri/i965/brw_fs_combine_constants.cpp | 222 + 4 files changed, 226 insertions(+)

[Mesa-dev] [RFC 2/9] i965/cfg: Calculate the immediate dominators.

2014-10-31 Thread Matt Turner
--- src/mesa/drivers/dri/i965/brw_cfg.cpp | 68 +- src/mesa/drivers/dri/i965/brw_cfg.h| 7 ++- .../drivers/dri/i965/brw_dead_control_flow.cpp | 5 +- 3 files changed, 75 insertions(+), 5 deletions(-) diff --git a/src/mesa/drivers/dri/i965/brw

[Mesa-dev] [RFC 8/9] i965/fs: Allow immediates in MAD and LRP instructions.

2014-10-31 Thread Matt Turner
And then the opt_combine_constants() pass will pull them out into registers. --- .../drivers/dri/i965/brw_fs_combine_constants.cpp | 22 +++--- .../drivers/dri/i965/brw_fs_copy_propagation.cpp | 6 ++ 2 files changed, 25 insertions(+), 3 deletions(-) diff --git a/src/mesa/

[Mesa-dev] [RFC 4/9] i965/cfg: Add function to generate a dot file of the dominator tree.

2014-10-31 Thread Matt Turner
--- src/mesa/drivers/dri/i965/brw_cfg.cpp | 10 ++ src/mesa/drivers/dri/i965/brw_cfg.h | 1 + 2 files changed, 11 insertions(+) diff --git a/src/mesa/drivers/dri/i965/brw_cfg.cpp b/src/mesa/drivers/dri/i965/brw_cfg.cpp index cc0317d..65e2d37 100644 --- a/src/mesa/drivers/dri/i965/brw_

[Mesa-dev] [RFC 3/9] i965/cfg: Add function to generate a dot file of the CFG.

2014-10-31 Thread Matt Turner
--- src/mesa/drivers/dri/i965/brw_cfg.cpp | 14 ++ src/mesa/drivers/dri/i965/brw_cfg.h | 1 + 2 files changed, 15 insertions(+) diff --git a/src/mesa/drivers/dri/i965/brw_cfg.cpp b/src/mesa/drivers/dri/i965/brw_cfg.cpp index 622b7de..cc0317d 100644 --- a/src/mesa/drivers/dri/i965/

[Mesa-dev] [RFC 6/9] i965/fs: Remove force_writemask_all assertion for execsize < 8.

2014-10-31 Thread Matt Turner
This doesn't seem to be necessary. --- src/mesa/drivers/dri/i965/brw_fs_generator.cpp | 1 - 1 file changed, 1 deletion(-) diff --git a/src/mesa/drivers/dri/i965/brw_fs_generator.cpp b/src/mesa/drivers/dri/i965/brw_fs_generator.cpp index 401a5ed..36c11c2 100644 --- a/src/mesa/drivers/dri/i965/br

[Mesa-dev] [RFC 1/9] i965/cfg: Allow cfg::dump to be called without a visitor.

2014-10-31 Thread Matt Turner
The fs_visitor's dump_instruction() implementation calls cfg_t() indirectly through calculate_live_intervals, so if you have an infinite loop in the CFG code, you can't call cfg::dump(fs_visitor *) to debug it. --- src/mesa/drivers/dri/i965/brw_cfg.cpp | 3 ++- 1 file changed, 2 insertions(+), 1 d

[Mesa-dev] [RFC 0/9] i965/fs: Combine constants and unconditionally emit MADs

2014-10-31 Thread Matt Turner
Three-source instructions on i965 have an annoying property that they cannot use immediate operands. They've do have the alluring property that they perform multiple operations in basically the same number of cycles as any other instruction. But when your arguments are immediates we decided that a

[Mesa-dev] [RFC 9/9] i965/fs: Emit MAD instructions when possible.

2014-10-31 Thread Matt Turner
Previously we didn't emit MAD instructions since they cannot take immediate arguments, but with the opt_combine_constants() pass we can handle this properly. --- src/mesa/drivers/dri/i965/brw_fs_fp.cpp | 11 --- src/mesa/drivers/dri/i965/brw_fs_visitor.cpp | 5 - 2 files changed,

Re: [Mesa-dev] [RFC 0/9] i965/fs: Combine constants and unconditionally emit MADs

2014-10-31 Thread Connor Abbott
On Fri, Oct 31, 2014 at 9:27 PM, Matt Turner wrote: > Three-source instructions on i965 have an annoying property that they > cannot use immediate operands. They've do have the alluring property > that they perform multiple operations in basically the same number of > cycles as any other instructi

Re: [Mesa-dev] [RFC 7/9] i965/fs: Add pass to combine immediates.

2014-10-31 Thread Connor Abbott
On Fri, Oct 31, 2014 at 9:27 PM, Matt Turner wrote: > --- > src/mesa/drivers/dri/i965/Makefile.sources | 1 + > src/mesa/drivers/dri/i965/brw_fs.cpp | 2 + > src/mesa/drivers/dri/i965/brw_fs.h | 1 + > .../drivers/dri/i965/brw_fs_combine_constants.cpp

Re: [Mesa-dev] Mesa 10.4 release plan strawman

2014-10-31 Thread Emil Velikov
On 22/10/14 22:14, Emil Velikov wrote: > Hi all, > > I was wondering earlier "how far are we until the 10.4 release" and it > hit me... there isn't much left. So in order to stick with the original > three month release schedule here is my proposal. > > November 14th 2014 - Feature freeze/Release

Re: [Mesa-dev] [PATCH 1/2] glsl: Emit mul instead of dot if only one component left.

2014-10-31 Thread Ian Romanick
Series is Reviewed-by: Ian Romanick On 10/31/2014 10:35 AM, Matt Turner wrote: > Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=85683 > Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=85691 > --- > src/glsl/opt_algebraic.cpp | 5 - > 1 file changed, 4 insertions(+), 1 deletio

Re: [Mesa-dev] [demos PATCH] glxinfo/wglinfo: reverse order of the gl_versions[] array

2014-10-31 Thread Ian Romanick
Reviewed-by: Ian Romanick On 10/31/2014 08:54 AM, Brian Paul wrote: > By going from newest to oldest we can simplify the glxinfo/wglinfo > loops over the possible GL versions. > --- > src/wgl/wglinfo.c |2 +- > src/xdemos/glinfo_common.h | 40 +-