3D support for Displaylink devices
> The window system needs support for splitting rendering and display. > In X these are currently tied together. The only real obstacle is > fixing this in X. However, this is a lot of work. Dave Airlie has > started working on this, but it's not really usable yet. See: > http://airlied.livejournal.com/71734.html > http://cgit.freedesktop.org/~airlied/xserver/log/?h=drvlayer In the windows world it basically works by 'borrowing' the real graphics card for rendering and then doing the damage list/compression/uplink. So its basically a *very* strange CRTC/scanout. To most intents and purposes plugging one into an i915 say is an extra i915 head, and indeed you can sensibly display the same scanout buffer on both real and link heads. Alan
Questions about libdrm_intel and way to share physical memory between CPU and GPU
On Sat, 28 May 2011 09:54:01 +0100 Chris Wilson wrote: > On Fri, 27 May 2011 14:37:45 -0700, "Segovia, Benjamin" intel.com> wrote: > > Hello gurus, > > > > I have two question mostly regarding libdrm_intel > > > > 1/ What is the difference between drm_intel_bo_map and > > drm_intel_gem_bo_map_gtt ? > bo_map uses the CPU domain, and so is CPU linear (needs sw detiling). > bo_gtt_map uses the uncached [WC] GTT domain, and so is GPU linear > (detiling is performed by the hardware using a fence). > > > 2/ Will it be possible (or is it already possible) to directly share a > > regularly allocated piece of physical memory? Typical use case is the > > following one using OpenCL API: > > Yes. I've proposed a vmap interface to bind user-pages into the GTT, > similar to a completely unused bit of TTM functionality. It seems to me that stolen memory and other things could all be sorted out somewhat if the GEM layer and GEM as shmemfs backing were split apart a bit. A 'privately backed' GEM object wouldn't be able to support flink() but I can't find much else that would break ? Wondering about this for things like the GMA500, and also to get back all that memory the i9xx driver burns on a PC.
Semantics of the 'dumb' interface
I have GEM allocation working on the GMA 500 and I can scribble in a framebuffer (minus an odd 'last page' bug which is an off by one somewhere I imagine) and display it with nice modetest stripes. If I kill the program however it all goes kerblam drm_release calls into fb_release which duely destroys the user frame buffer. This drops the gem object reference and we start to release all the resources but it blows up because the GEM object is still pinned in the GTT as it's still being used as scanout. Where I'm a bit lost right now is understanding what is supposed to happen at this point, and how the scanout is supposed to have ended up cleaned up and reset to the system frame buffer beforehand thus unpinning the scanout from the GTT. Alan
[Bug 31412] radeon memory leak
https://bugzilla.kernel.org/show_bug.cgi?id=31412 --- Comment #15 from Kevin 2011-06-02 17:24:35 --- I'm sorry. I wasn't clear. There is NOT a memory leak. Below is some of my 'free -m' output from above. My question is, why does the 'used -/+ buffers/cache' decrease when I 'echo 3 > /proc/sys/vm/drop_caches'? # free -m total used free sharedbuffers cached Mem: 5920 4318 1602 0113 2465 -/+ buffers/cache: 1740 4180 Swap: 7998 0 7998 # echo 3 > /proc/sys/vm/drop_caches # free -m total used free sharedbuffers cached Mem: 5920225 5695 0 3 12 -/+ buffers/cache:209 5711 Swap: 7998 0 7998 -- Configure bugmail: https://bugzilla.kernel.org/userprefs.cgi?tab=email --- You are receiving this mail because: --- You are watching the assignee of the bug.
Fw: [PATCH] drm: i915: correct return status in intel_hdmi_mode_valid()
On Mon, 30 May 2011 12:48:26 +0200, Nicolas Kaiser wrote: > if (mode->clock < 2) > - return MODE_CLOCK_HIGH; > + return MODE_CLOCK_LOW; Seems obvious to me. Reviewed-by: Keith Packard -- keith.packard at intel.com -- next part -- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 189 bytes Desc: not available URL: <http://lists.freedesktop.org/archives/dri-devel/attachments/20110602/60d5fdbd/attachment.pgp>
3D support for Displaylink devices
Garry, My first name is "PrasannaKumar". I will use my full name to prevent confusion :). I want 3D acceleration for running Compiz or GNOME3 or KWin with composition. Currently windows Displaylink driver compresses and transfers pixel data where there is a change (only damaged area is transferred) to reduce the amount of data transfer. It is able to play HD video without dropping frames. So I think that 3D acceleration and video playback acceleration is possible. High end games cannot be played but the normal 3D and video operations should work without any issues. When displaylink introduces USB 3.0 devices the bandwidth issue will go away (I remember reading in Wikipedia that displaylink is working on a USB 3.0 product). The displaylink framebuffer driver that comes with linux (udlfb) also compresses and transfers only the damaged region to conserve the USB bandwidth. Also CPU usage for doing the compression is very less making it ideal for mobile devices (may be an android mobile). http://www.youtube.com/watch?v=3-bLOc1qnMM=player_embedded shows android mobile with displaylink device. When a mobile phone is able to power a high resolution graphics normal desktops and notebooks can do provide good quality output. PrasannaKumar Muralidharan On 31-05-2011 15:36, Garry Hurley Jr. wrote: > Kumar > > I am going to make the assumption that your culture puts the family > name first, so please excuse me for calling you Kumar if that is not > your given name. > > As to your question, I think I understand what you are asking for and > I was thinking similar things about displaying over ethernet about > five years ago. The problem is complex due to video refresh rates and > the latency of the connection. You would not get the same performance > on a video game, for example, unless you dropped a few dozen frames > per second, since the USB bus is slower than the PCI bus or even the > ISA bus. If you are talking about 3D acceleration, I presume you want > to game with it. The solution may lie in buffering, but again, your > performance would suffer unless you took the quality down a notch. > From the gamers I know, dropping quality for performance is a very > tricky balance. Each one is different about the quality he or she will > allow to be dropped in a game but when that balance is tipped, they > will complain or switch to a different technology. > > I am not saying it is not possible, but I am asking if, knowing this, > you truly feel it is worth the effort to try to implement it. > > Sent from my iPhone > > On May 30, 2011, at 1:30 PM, PRASANNA KUMAR > <mailto:prasanna_tsm_kumar at yahoo.co.in>> wrote: > >> USB graphics devices from displaylink does not have 3D hardware. To >> get 3D effects (compiz, GNOME 3, KWin, OpenGL apps etc) with these >> device in Linux the native (primary) GPU can be used to provide >> hardware acceleration. All the graphics operation is done using the >> native (primary) GPU and the end result is taken and send to the >> displaylink device. Can this be achieved? If so is it possible to >> implement a generic framework so that any device (USB, thunderbolt or >> any new technology) can use this just by implementing device specific >> (compression and) data transport? I am not sure this is the correct >> mailing list. >> >> Thanks, >> Prasanna Kumar >> ___ >> dri-devel mailing list >> dri-devel at lists.freedesktop.org <mailto:dri-devel at >> lists.freedesktop.org> >> http://lists.freedesktop.org/mailman/listinfo/dri-devel -- next part -- An HTML attachment was scrubbed... URL: <http://lists.freedesktop.org/archives/dri-devel/attachments/20110602/5111f69a/attachment.htm>
[Bug 30922] [radeon] Noisy fan with Radeon R600 and Toshiba Satellite L500-164
https://bugzilla.kernel.org/show_bug.cgi?id=30922 --- Comment #6 from Denis Washington 2011-06-02 10:46:39 --- As I have already stated, changing the power management options didn't change anything at all. Even running the "low" profile doesn't stop the fan from cranking up to full speed under normal desktop load and never spinning down again, even if the system is idling. I am also most certain that my laptop has a separate GPU fan; I can, in fact, control the CPU fan per sysfs, but whatever I do, the GPU fan keeps spinning at the same speed, so I guess there must be separate fans for CPU and GPU. -- Configure bugmail: https://bugzilla.kernel.org/userprefs.cgi?tab=email --- You are receiving this mail because: --- You are watching the assignee of the bug.
[Bug 36812] GPU lockup in Team Fortress 2
https://bugs.freedesktop.org/show_bug.cgi?id=36812 --- Comment #8 from Sven Arvidsson 2011-06-02 10:15:55 PDT --- (In reply to comment #7) > With MESA_GLSL=nopt the code that was changed by the bisected commit is not > being executed, so I think the real problem might be somewhere else. I guess > you could try bisecting again with MESA_GLSL=nopt and maybe you'll come up > with > a different bad commit. I might try this. Enrico, do you remember what the good revision you used for the bisect was, 7.10? -- Configure bugmail: https://bugs.freedesktop.org/userprefs.cgi?tab=email --- You are receiving this mail because: --- You are the assignee for the bug.
[Bug 31412] radeon memory leak
https://bugzilla.kernel.org/show_bug.cgi?id=31412 Jesse Zhang changed: What|Removed |Added CC||zh.jesse at gmail.com --- Comment #14 from Jesse Zhang 2011-06-02 05:56:22 --- There is no leaking here. A quick google search turns up http://sourcefrog.net/weblog/software/linux-kernel/free-mem.html. -- Configure bugmail: https://bugzilla.kernel.org/userprefs.cgi?tab=email --- You are receiving this mail because: --- You are watching the assignee of the bug.
[Bug 27517] KMS breaks 3D on R200
https://bugs.freedesktop.org/show_bug.cgi?id=27517 --- Comment #10 from Keith 2011-06-02 03:01:22 PDT --- (In reply to comment #9) > (In reply to comment #8) > > Yes, some time around the introduction of KMS, acceleration stopped > > working. I > > had it fine on Debian before squeeze and Ubuntu Maveric but recently > > upgraded > > to Natty and it stopped working. Starting compiz gives the old > > GLX_EXT_texture_pixmap not available error. It also complains about GL > > version > > 1.4+ whereas glxinfo reports version 1.4 > > Latest updates on Debian squeeze have 3D acceleration bursting into life. > Kernel: 2.6.32-5-686 > X server 1.7.7 That should have been 2D. Also, it was openGL version 1.4 whereas this card only supports 1.3 so everything is as it should be. glxgears still gives low values for FPS. Compiz now works fine. I wish /usr/lib/nux/unity_support_test -p was not such a well kept secret, it would have helped clear up a lot of misconceptions. -- Configure bugmail: https://bugs.freedesktop.org/userprefs.cgi?tab=email --- You are receiving this mail because: --- You are the assignee for the bug.
[Bug 31412] radeon memory leak
https://bugzilla.kernel.org/show_bug.cgi?id=31412 --- Comment #13 from Kevin 2011-06-02 02:57:29 --- So is this something that should be fixed or is everything working as intended? To me, it doesn't seem correct. Unused caches shouldn't be stored in my "used" memory. Is there anything else that I should try to help? Can anyone else reproduce this? I just open large pdfs and scroll through them. -- Configure bugmail: https://bugzilla.kernel.org/userprefs.cgi?tab=email --- You are receiving this mail because: --- You are watching the assignee of the bug.
Re: 3D support for Displaylink devices
Garry, My first name is PrasannaKumar. I will use my full name to prevent confusion :). I want 3D acceleration for running Compiz or GNOME3 or KWin with composition. Currently windows Displaylink driver compresses and transfers pixel data where there is a change (only damaged area is transferred) to reduce the amount of data transfer. It is able to play HD video without dropping frames. So I think that 3D acceleration and video playback acceleration is possible. High end games cannot be played but the normal 3D and video operations should work without any issues. When displaylink introduces USB 3.0 devices the bandwidth issue will go away (I remember reading in Wikipedia that displaylink is working on a USB 3.0 product). The displaylink framebuffer driver that comes with linux (udlfb) also compresses and transfers only the damaged region to conserve the USB bandwidth. Also CPU usage for doing the compression is very less making it ideal for mobile devices (may be an android mobile). http://www.youtube.com/watch?v=3-bLOc1qnMMfeature=player_embedded shows android mobile with displaylink device. When a mobile phone is able to power a high resolution graphics normal desktops and notebooks can do provide good quality output. PrasannaKumar Muralidharan On 31-05-2011 15:36, Garry Hurley Jr. wrote: Kumar I am going to make the assumption that your culture puts the family name first, so please excuse me for calling you Kumar if that is not your given name. As to your question, I think I understand what you are asking for and I was thinking similar things about displaying over ethernet about five years ago. The problem is complex due to video refresh rates and the latency of the connection. You would not get the same performance on a video game, for example, unless you dropped a few dozen frames per second, since the USB bus is slower than the PCI bus or even the ISA bus. If you are talking about 3D acceleration, I presume you want to game with it. The solution may lie in buffering, but again, your performance would suffer unless you took the quality down a notch. From the gamers I know, dropping quality for performance is a very tricky balance. Each one is different about the quality he or she will allow to be dropped in a game but when that balance is tipped, they will complain or switch to a different technology. I am not saying it is not possible, but I am asking if, knowing this, you truly feel it is worth the effort to try to implement it. Sent from my iPhone On May 30, 2011, at 1:30 PM, PRASANNA KUMAR prasanna_tsm_ku...@yahoo.co.in mailto:prasanna_tsm_ku...@yahoo.co.in wrote: USB graphics devices from displaylink does not have 3D hardware. To get 3D effects (compiz, GNOME 3, KWin, OpenGL apps etc) with these device in Linux the native (primary) GPU can be used to provide hardware acceleration. All the graphics operation is done using the native (primary) GPU and the end result is taken and send to the displaylink device. Can this be achieved? If so is it possible to implement a generic framework so that any device (USB, thunderbolt or any new technology) can use this just by implementing device specific (compression and) data transport? I am not sure this is the correct mailing list. Thanks, Prasanna Kumar ___ dri-devel mailing list dri-devel@lists.freedesktop.org mailto:dri-devel@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/dri-devel ___ dri-devel mailing list dri-devel@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/dri-devel
[Bug 27517] KMS breaks 3D on R200
https://bugs.freedesktop.org/show_bug.cgi?id=27517 --- Comment #10 from Keith ke...@backblow-loft.demon.co.uk 2011-06-02 03:01:22 PDT --- (In reply to comment #9) (In reply to comment #8) Yes, some time around the introduction of KMS, acceleration stopped working. I had it fine on Debian before squeeze and Ubuntu Maveric but recently upgraded to Natty and it stopped working. Starting compiz gives the old GLX_EXT_texture_pixmap not available error. It also complains about GL version 1.4+ whereas glxinfo reports version 1.4 Latest updates on Debian squeeze have 3D acceleration bursting into life. Kernel: 2.6.32-5-686 X server 1.7.7 That should have been 2D. Also, it was openGL version 1.4 whereas this card only supports 1.3 so everything is as it should be. glxgears still gives low values for FPS. Compiz now works fine. I wish /usr/lib/nux/unity_support_test -p was not such a well kept secret, it would have helped clear up a lot of misconceptions. -- Configure bugmail: https://bugs.freedesktop.org/userprefs.cgi?tab=email --- You are receiving this mail because: --- You are the assignee for the bug. ___ dri-devel mailing list dri-devel@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/dri-devel
[Bug 30922] [radeon] Noisy fan with Radeon R600 and Toshiba Satellite L500-164
https://bugzilla.kernel.org/show_bug.cgi?id=30922 --- Comment #6 from Denis Washington den...@online.de 2011-06-02 10:46:39 --- As I have already stated, changing the power management options didn't change anything at all. Even running the low profile doesn't stop the fan from cranking up to full speed under normal desktop load and never spinning down again, even if the system is idling. I am also most certain that my laptop has a separate GPU fan; I can, in fact, control the CPU fan per sysfs, but whatever I do, the GPU fan keeps spinning at the same speed, so I guess there must be separate fans for CPU and GPU. -- Configure bugmail: https://bugzilla.kernel.org/userprefs.cgi?tab=email --- You are receiving this mail because: --- You are watching the assignee of the bug. ___ dri-devel mailing list dri-devel@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/dri-devel
[Bug 36812] GPU lockup in Team Fortress 2
https://bugs.freedesktop.org/show_bug.cgi?id=36812 --- Comment #8 from Sven Arvidsson s...@whiz.se 2011-06-02 10:15:55 PDT --- (In reply to comment #7) With MESA_GLSL=nopt the code that was changed by the bisected commit is not being executed, so I think the real problem might be somewhere else. I guess you could try bisecting again with MESA_GLSL=nopt and maybe you'll come up with a different bad commit. I might try this. Enrico, do you remember what the good revision you used for the bisect was, 7.10? -- Configure bugmail: https://bugs.freedesktop.org/userprefs.cgi?tab=email --- You are receiving this mail because: --- You are the assignee for the bug. ___ dri-devel mailing list dri-devel@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/dri-devel
[Bug 31412] radeon memory leak
https://bugzilla.kernel.org/show_bug.cgi?id=31412 --- Comment #15 from Kevin kjs...@gmail.com 2011-06-02 17:24:35 --- I'm sorry. I wasn't clear. There is NOT a memory leak. Below is some of my 'free -m' output from above. My question is, why does the 'used -/+ buffers/cache' decrease when I 'echo 3 /proc/sys/vm/drop_caches'? # free -m total used free sharedbuffers cached Mem: 5920 4318 1602 0113 2465 -/+ buffers/cache: 1740 4180 Swap: 7998 0 7998 # echo 3 /proc/sys/vm/drop_caches # free -m total used free sharedbuffers cached Mem: 5920225 5695 0 3 12 -/+ buffers/cache:209 5711 Swap: 7998 0 7998 -- Configure bugmail: https://bugzilla.kernel.org/userprefs.cgi?tab=email --- You are receiving this mail because: --- You are watching the assignee of the bug. ___ dri-devel mailing list dri-devel@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/dri-devel
Semantics of the 'dumb' interface
I have GEM allocation working on the GMA 500 and I can scribble in a framebuffer (minus an odd 'last page' bug which is an off by one somewhere I imagine) and display it with nice modetest stripes. If I kill the program however it all goes kerblam drm_release calls into fb_release which duely destroys the user frame buffer. This drops the gem object reference and we start to release all the resources but it blows up because the GEM object is still pinned in the GTT as it's still being used as scanout. Where I'm a bit lost right now is understanding what is supposed to happen at this point, and how the scanout is supposed to have ended up cleaned up and reset to the system frame buffer beforehand thus unpinning the scanout from the GTT. Alan ___ dri-devel mailing list dri-devel@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/dri-devel
Re: Questions about libdrm_intel and way to share physical memory between CPU and GPU
On Sat, 28 May 2011 09:54:01 +0100 Chris Wilson ch...@chris-wilson.co.uk wrote: On Fri, 27 May 2011 14:37:45 -0700, Segovia, Benjamin benjamin.sego...@intel.com wrote: Hello gurus, I have two question mostly regarding libdrm_intel 1/ What is the difference between drm_intel_bo_map and drm_intel_gem_bo_map_gtt ? bo_map uses the CPU domain, and so is CPU linear (needs sw detiling). bo_gtt_map uses the uncached [WC] GTT domain, and so is GPU linear (detiling is performed by the hardware using a fence). 2/ Will it be possible (or is it already possible) to directly share a regularly allocated piece of physical memory? Typical use case is the following one using OpenCL API: Yes. I've proposed a vmap interface to bind user-pages into the GTT, similar to a completely unused bit of TTM functionality. It seems to me that stolen memory and other things could all be sorted out somewhat if the GEM layer and GEM as shmemfs backing were split apart a bit. A 'privately backed' GEM object wouldn't be able to support flink() but I can't find much else that would break ? Wondering about this for things like the GMA500, and also to get back all that memory the i9xx driver burns on a PC. ___ dri-devel mailing list dri-devel@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/dri-devel
Re: 3D support for Displaylink devices
The window system needs support for splitting rendering and display. In X these are currently tied together. The only real obstacle is fixing this in X. However, this is a lot of work. Dave Airlie has started working on this, but it's not really usable yet. See: http://airlied.livejournal.com/71734.html http://cgit.freedesktop.org/~airlied/xserver/log/?h=drvlayer In the windows world it basically works by 'borrowing' the real graphics card for rendering and then doing the damage list/compression/uplink. So its basically a *very* strange CRTC/scanout. To most intents and purposes plugging one into an i915 say is an extra i915 head, and indeed you can sensibly display the same scanout buffer on both real and link heads. Alan ___ dri-devel mailing list dri-devel@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/dri-devel
[PATCH] drm/radeon/kms/atom: initialize dig phy a bit later
Commit ac89af1e1010640db072416c786f97391b85790f caused one of the monitors attached to a dual head radeon gpu to have inverted colors (until the first suspend/resume). Initializing dig phy a bit later fixes the problem. --- drivers/gpu/drm/radeon/radeon_display.c |8 1 files changed, 4 insertions(+), 4 deletions(-) diff --git a/drivers/gpu/drm/radeon/radeon_display.c b/drivers/gpu/drm/radeon/radeon_display.c index ae247ee..ddff2cf 100644 --- a/drivers/gpu/drm/radeon/radeon_display.c +++ b/drivers/gpu/drm/radeon/radeon_display.c @@ -1346,10 +1346,6 @@ int radeon_modeset_init(struct radeon_device *rdev) return ret; } - /* init dig PHYs */ - if (rdev-is_atom_bios) - radeon_atom_encoder_init(rdev); - /* initialize hpd */ radeon_hpd_init(rdev); @@ -1359,6 +1355,10 @@ int radeon_modeset_init(struct radeon_device *rdev) radeon_fbdev_init(rdev); drm_kms_helper_poll_init(rdev-ddev); + /* init dig PHYs */ + if (rdev-is_atom_bios) + radeon_atom_encoder_init(rdev); + return 0; } -- 1.7.4.1 ___ dri-devel mailing list dri-devel@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/dri-devel
Re: Fw: [PATCH] drm: i915: correct return status in intel_hdmi_mode_valid()
On Mon, 30 May 2011 12:48:26 +0200, Nicolas Kaiser ni...@nikai.net wrote: if (mode-clock 2) - return MODE_CLOCK_HIGH; + return MODE_CLOCK_LOW; Seems obvious to me. Reviewed-by: Keith Packard kei...@keithp.com -- keith.pack...@intel.com pgplUHlaQ5pZw.pgp Description: PGP signature ___ dri-devel mailing list dri-devel@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/dri-devel
Re: 3D support for Displaylink devices
On Mon, May 30, 2011 at 12:30 PM, PRASANNA KUMAR prasanna_tsm_ku...@yahoo.co.in wrote: USB graphics devices from displaylink does not have 3D hardware. To get 3D effects (compiz, GNOME 3, KWin, OpenGL apps etc) with these device in Linux the native (primary) GPU can be used to provide hardware acceleration. All the graphics operation is done using the native (primary) GPU and the end result is taken and send to the displaylink device. Can this be achieved? If so is it possible to implement a generic framework so that any device (USB, thunderbolt or any new technology) can use this just by implementing device specific (compression and) data transport? I am not sure this is the correct mailing list. fwiw, this situation is not too far different from the SoC world. For example, there are multiple ARM SoC's that share the same IMG/PowerVR core or ARM/mali 3d core, but each have their own unique display controller.. I don't know quite the best way to deal with this (either at the DRM/kernel layer or xorg driver layer), but there would certainly be some benefit to be able to make DRM driver a bit more modular to combine a SoC specific display driver (mostly the KMS part) with a different 2d and/or 3d accelerator IP. Of course the (or some of the) challenge here is that different display controllers might have different memory mgmt requirements (for ex, depending on whether the display controller has an IOMMU or not) and formats, and that the flip command should somehow come via the 2d/3d command stream. I have an (experimental) DRM/KMS driver for OMAP which tries to solve the issue by way of a simple plugin API, ie the idea being to separate the PVR part from the OMAP display controller part more cleanly. I don't think it is perfect, but it is an attempt. (I'll send patches as an RFC, but wanted to do some cleanup first.. just haven't had time yet.) But I'm definitely open to suggestions here. BR, -R Thanks, Prasanna Kumar ___ dri-devel mailing list dri-devel@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/dri-devel ___ dri-devel mailing list dri-devel@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/dri-devel
