headers
Brian Paul | 1 Sep 2009 01:11
Favicon

Re: New arb_sync branch

Ian Romanick wrote:
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
> 
> I just pushed a branch that implements the new GL_ARB_sync extension.
> It is currently only supported for the software rasterizer, but I
> believe the Eric and I will be able to add support to the Intel drivers
> (that use GEM) tomorrow.  I am hoping to merge this branch to master in
> the next few days.
> 
> Before I merge this branch, I'd like a little review.  In particular,
> I'd like for folks to review the locking around ctx->Shared.  I'm 90%
> confident that what I have implemented gives the correct semantic, but
> I'd really like to have some more eyes on it.

Looks OK.  Though, someday I'd like to review the use of 
ctx->Shared->Mutex of everywhere.  I think we should probably have a 
per-object type lock intead of just one big lock.  But multi-context, 
multi-threaded apps aren't too common so I'm not in a big hurry.

Other than that, I have just a few minor suggestions:

1. Instead of (*ctx->Driver.DeleteSyncObject)(ctx, syncObj);
use ctx->Driver.DeleteSyncObject(ctx, syncObj);  Etc.

2. Replace:  const GLsizei copy_count = (size > bufSize) ? bufSize : size;
with:  const GLsizei copy_count = MIN2(size, bufSize);

3. Replace double blank lines in functions with a single blank line.
(Continue reading)

Permalink | Reply | 0 comments |
headers
Jesse Barnes | 1 Sep 2009 03:21
Favicon

RFC: buffer swaps in a composited environment

In preparation for our Plumbers session on this topic, Kristian, Ian
and myself have had several discussions about how to handle the various
GL swap & sync extensions in a DRI2 composited environment.

Today after some discussion I put together this page:
http://dri.freedesktop.org/wiki/CompositeSwap

I think it captures most of the issues and extensions we're worried
about, but there may be others.  So if you're interested in this topic
please check it out and help us fill in any holes.

I'm especially interested in hearing from app, toolkit and window
manager/compositor developers here.

Thanks,
--

-- 
Jesse Barnes, Intel Open Source Technology Center

------------------------------------------------------------------------------
Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day 
trial. Simplify your report design, integration and deployment - and focus on 
what you do best, core application coding. Discover what's new with 
Crystal Reports now.  http://p.sf.net/sfu/bobj-july
Permalink | Reply | 0 comments |
headers
Keith Whitwell | 1 Sep 2009 09:48
Favicon

Re: Mesa (master): llvmpipe: Texture cache is read-only. No need to flush.

I'm not sure this is valid.  The cache is read-only, but textures can
get modified by other actions, rendering the cached copy invalid -
that's the point of the flush.

Keith

On Sun, 2009-08-30 at 04:37 -0700, Jose Fonseca wrote:
> Module: Mesa
> Branch: master
> Commit: ffe1b7bad82726da44bf91bd61ca62e5cf6645de
> URL:    http://cgit.freedesktop.org/mesa/mesa/commit/?id=ffe1b7bad82726da44bf91bd61ca62e5cf6645de
> 
> Author: José Fonseca <jfonseca <at> vmware.com>
> Date:   Sun Aug 30 12:03:49 2009 +0100
> 
> llvmpipe: Texture cache is read-only. No need to flush.
> 
> ---
> 
>  src/gallium/drivers/llvmpipe/lp_flush.c     |    6 ----
>  src/gallium/drivers/llvmpipe/lp_tex_cache.c |   39 ---------------------------
>  src/gallium/drivers/llvmpipe/lp_tex_cache.h |    3 --
>  3 files changed, 0 insertions(+), 48 deletions(-)
> 
> diff --git a/src/gallium/drivers/llvmpipe/lp_flush.c b/src/gallium/drivers/llvmpipe/lp_flush.c
> index 44b4696..b5c1c95 100644
> --- a/src/gallium/drivers/llvmpipe/lp_flush.c
> +++ b/src/gallium/drivers/llvmpipe/lp_flush.c
>  <at>  <at>  -51,12 +51,6  <at>  <at>  llvmpipe_flush( struct pipe_context *pipe,
>
(Continue reading)

Permalink | Reply | 1 comment |
headers
Chia-I Wu | 1 Sep 2009 12:22
Picon

Re: Dispatch table in OpenGL ES state tracker

On Tue, Sep 01, 2009 at 06:08:06PM +0800, Chia-I Wu wrote:
> Most of the time was spent on writing XMLs for OpenGL ES 1.x and 2.x.  I
> want to talk about how I did it.  Feedbacks would be appreciated.
I have uploaded them to http://0xlab.org/~olv/mesa-es/.

--

-- 
Regards,
olv

------------------------------------------------------------------------------
Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day 
trial. Simplify your report design, integration and deployment - and focus on 
what you do best, core application coding. Discover what's new with 
Crystal Reports now.  http://p.sf.net/sfu/bobj-july
Permalink | Reply | 0 comments |
headers
José Fonseca | 1 Sep 2009 13:27
Favicon

Re: Mesa (master): llvmpipe: Texture cache is read-only. No need to flush.

That's taken care of in

void
lp_tex_tile_cache_validate_texture(struct llvmpipe_tex_tile_cache *tc)
{
   if (tc->texture) {
      struct llvmpipe_texture *lpt = llvmpipe_texture(tc->texture);
      if (lpt->timestamp != tc->timestamp) {
         /* texture was modified, invalidate all cached tiles */
         uint i;
         _debug_printf("INV %d %d\n", tc->timestamp, lpt->timestamp);
         for (i = 0; i < NUM_ENTRIES; i++) {
            tc->entries[i].addr.bits.invalid = 1;
         }

         tc->timestamp = lpt->timestamp;
      }
   }
}

timestamp is incremented whenever the textures is written by CPU or GPU.

I also tested gearbox.

Jose

On Tue, 2009-09-01 at 00:48 -0700, Keith Whitwell wrote:
> I'm not sure this is valid.  The cache is read-only, but textures can
> get modified by other actions, rendering the cached copy invalid -
> that's the point of the flush.
(Continue reading)

Permalink | Reply | 0 comments |
headers
Maxim Levitsky | 1 Sep 2009 03:20
Picon

i965 support broken

Hi,

I have just updated graphics stack on my i965 based system.

Now, I notice that in many 3d applications, (I test few games) screen
flashes and I see for a moment few garbage triangles.

ppracer flashes its main screen in above way, suspetuxkart shows that
issue, after few minutes of playing, etc...

a quick bisect revealed unfortunately that last *good* commit is

20d9204fbd71aebf870834b612579419d2c278b5 radeon: fix max indx/vertex emission due to state checker

After that, 3d output went completely black, and I didn't yet traced it
back when it started working again, the way it does in head.

commit d6b8664e3cd37c081cb1dd3d6cd5ffdac1813dac swrast: minor code consolidation

still shows black output

master commit on my system is c4a3e036ed1c755a291018251c4f55c45ac17079

could you also look at 
https://bugs.freedesktop.org/show_bug.cgi?id=23254

pretty much even now, if I revert commit 

0f328c90dbc893e15005f2ab441d309c1c176245 i965: Fall back or appropriately adjust offsets of drawing
to tiled regions.
(Continue reading)

Permalink | Reply | 0 comments |
headers
bugzilla-daemon | 1 Sep 2009 16:56

[Bug 21267] progs/tests: various issues

http://bugs.freedesktop.org/show_bug.cgi?id=21267

--- Comment #10 from Brian Paul <brian.paul <at> tungstengraphics.com>  2009-09-01 07:56:17 PST ---
I've restored the disabled exit() call in fbotest1.c so that should no longer
segfault.

As for api_speed.c, it's old and doesn't compile any more so just ignore it.

--

-- 
Configure bugmail: http://bugs.freedesktop.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug.

------------------------------------------------------------------------------
Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day 
trial. Simplify your report design, integration and deployment - and focus on 
what you do best, core application coding. Discover what's new with 
Crystal Reports now.  http://p.sf.net/sfu/bobj-july
Permalink | Reply | 0 comments |
headers
bugzilla-daemon | 1 Sep 2009 17:00

[Bug 3050] BadLength error when running glitz apps on top of software Mesa

http://bugs.freedesktop.org/show_bug.cgi?id=3050

Alan Coopersmith <alan.coopersmith <at> sun.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Blocks|5799                        |

--

-- 
Configure bugmail: http://bugs.freedesktop.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug.

------------------------------------------------------------------------------
Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day 
trial. Simplify your report design, integration and deployment - and focus on 
what you do best, core application coding. Discover what's new with 
Crystal Reports now.  http://p.sf.net/sfu/bobj-july
Permalink | Reply | 0 comments |
headers
bugzilla-daemon | 1 Sep 2009 17:08

[Bug 21267] progs/tests: various issues

http://bugs.freedesktop.org/show_bug.cgi?id=21267

Fabio <fabio.ped <at> libero.it> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED

--- Comment #11 from Fabio <fabio.ped <at> libero.it>  2009-09-01 08:08:30 PST ---
(In reply to comment #10)
> I've restored the disabled exit() call in fbotest1.c so that should no longer
> segfault.
> 
> As for api_speed.c, it's old and doesn't compile any more so just ignore it.
> 

OK, closing.

--

-- 
Configure bugmail: http://bugs.freedesktop.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug.

------------------------------------------------------------------------------
Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day 
trial. Simplify your report design, integration and deployment - and focus on 
what you do best, core application coding. Discover what's new with 
Crystal Reports now.  http://p.sf.net/sfu/bobj-july
(Continue reading)

Permalink | Reply | 0 comments |
headers
Maxim Levitsky | 2 Sep 2009 00:13
Picon

Re: i965 support broken

Added CC.
> Hi,
> 
> I have just updated graphics stack on my i965 based system.
> 
> Now, I notice that in many 3d applications, (I test few games) screen
> flashes and I see for a moment few garbage triangles.
> 
> ppracer flashes its main screen in above way, suspetuxkart shows that
> issue, after few minutes of playing, etc...
> 
> 
> a quick bisect revealed unfortunately that last *good* commit is
> 
> 20d9204fbd71aebf870834b612579419d2c278b5 radeon: fix max indx/vertex emission due to state checker
> 
> 
> After that, 3d output went completely black, and I didn't yet traced it
> back when it started working again, the way it does in head.
> 
> commit d6b8664e3cd37c081cb1dd3d6cd5ffdac1813dac swrast: minor code consolidation
> 
> still shows black output
> 
> 
> master commit on my system is c4a3e036ed1c755a291018251c4f55c45ac17079
> 
> 
> could you also look at 
> https://bugs.freedesktop.org/show_bug.cgi?id=23254
(Continue reading)

Permalink | Reply | 0 comments |

Gmane