Re: [Dri-devel] OS Independent Templates
I'm pretty sure the problems with the M6 chip aren't the fault of the bsd-3-0-0-branch changes, but rather with the 2d 3d drivers themselves. Are there any other reasons why the bsd-3-0-0 branch shouldn't merge to the trunk? Keith --- This sf.net email is sponsored by:ThinkGeek No, I will not fix your computer. http://thinkgeek.com/sf ___ Dri-devel mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/dri-devel
Re: [Dri-devel] OS Independent Templates
I did some more testing tonight, and I can't find anything that concerns me. I say go for it :-) Adam On Wed, 3 Jul 2002, Keith Whitwell wrote: I'm pretty sure the problems with the M6 chip aren't the fault of the bsd-3-0-0-branch changes, but rather with the 2d 3d drivers themselves. Are there any other reasons why the bsd-3-0-0 branch shouldn't merge to the trunk? Keith --- This sf.net email is sponsored by:ThinkGeek No, I will not fix your computer. http://thinkgeek.com/sf ___ Dri-devel mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/dri-devel --- This sf.net email is sponsored by:ThinkGeek No, I will not fix your computer. http://thinkgeek.com/sf ___ Dri-devel mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/dri-devel
Re: [Dri-devel] OS Independent Templates
On Tue, 2002-07-02 at 00:51, Keith Whitwell wrote: - one user I was talking with said he had glxgears above 2000fps (was it 2600?) on a 1700xp and radeon 7500 with 4.2.0/drm-kmod and it dropped to 1900 with bsd-3-0-0-branch. I wouldn't necessarily blame bsd-3-0-0 for that -- there have been major changes to the radeon driver since 4.2 -- if he could retest against the trunk that would be more interesting. In any case I'm not too concerned about gears performance, it rarely reflects anything interesting. It has for us, at least. It pointed out the deficiency that led to the addition of mtrr in bsd, for example. If there was any other tool that all the users of XFree86 had, we would gladly use that instead, but as it is it's a quick measure that can show differences between people's dri status. I understand worrying abotu glxgears can be bad, but if other users are seeing drops in rate and I (with almost exactly the same hardware) am seeing increases, that's significant to me. I'm mostly concerned with that last point, but I still don't think it should stop a merge. As a side note, I would prefer to see someone else commit the actual merge if possible, because I don't feel proficient enough with CVS yet. It isn't too hard -- the CVS policies document on dri.sf.net gives a very good step-by-step guide to this. Just test where it says to test... Read it see if you're still unsure. Okay, I started trying that even before your mail came. Basically I have a clean copy of bsd-3-0-0-branch, and I did a cvs update -d -j HEAD in it (I wasn't planning on committing, just finding what crashes had been fixed). It merged a bunch of stuff, but also created conflicts in many files that had no bsd-3-0-0-branch specific changes to them. Is there a better way to go about this? Is there any way to diff to HEAD easily? When I do, for example, cvs diff -u -rHEAD lib/GL/mesa/src/drv/radeon/radeon_maos_verts.c in bsd-3-0-0-branch it returns nothing, though there is a diff to HEAD there (when looking at cvs log and then supplying -r1.4). Same thing happens if I add a -rbsd-3-0-0branch before the other -r arg. I would like to be able to easily cvs diff directories between branches. I thought from reading the manpage I could do this, but it's not working. -- Eric Anholt [EMAIL PROTECTED] http://lclark.edu/~eta/dri/ --- This sf.net email is sponsored by:ThinkGeek Welcome to geek heaven. http://thinkgeek.com/sf ___ Dri-devel mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/dri-devel
Re: [Dri-devel] OS Independent Templates
I'm running Debian(Linux) Woody and I just built the bsd-3-0-0 branch and it workes on my Radeon 7500 QW. --- Jens Owen [EMAIL PROTECTED] wrote: Eric Anholt wrote: Concerns at this point: - I'm still noticing some problems with my Radeon 7500 with DRI (the last lockup was dragging a glxgears around on top of another one). There has been talk of 7500 issues, though. - clipping issues on r128. r128 I know has problems on linux anyway. - one user I was talking with said he had glxgears above 2000fps (was it 2600?) on a 1700xp and radeon 7500 with 4.2.0/drm-kmod and it dropped to 1900 with bsd-3-0-0-branch. Consider this a call for Testers! I think our community testers would rather try this out on the bsd-3-0-0 branch than after it hits the stable trunk. -- /\ Jens Owen/ \/\ _ [EMAIL PROTECTED] /\ \ \ Steamboat Springs, Colorado --- This sf.net email is sponsored by:ThinkGeek Welcome to geek heaven. http://thinkgeek.com/sf ___ Dri-devel mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/dri-devel __ Do You Yahoo!? Sign up for SBC Yahoo! Dial - First Month Free http://sbc.yahoo.com --- This sf.net email is sponsored by:ThinkGeek Welcome to geek heaven. http://thinkgeek.com/sf ___ Dri-devel mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/dri-devel
Re: [Dri-devel] OS Independent Templates
On Mon, 1 Jul 2002, Jens Owen wrote: Eric Anholt wrote: Concerns at this point: - I'm still noticing some problems with my Radeon 7500 with DRI (the last lockup was dragging a glxgears around on top of another one). There has been talk of 7500 issues, though. - clipping issues on r128. r128 I know has problems on linux anyway. - one user I was talking with said he had glxgears above 2000fps (was it 2600?) on a 1700xp and radeon 7500 with 4.2.0/drm-kmod and it dropped to 1900 with bsd-3-0-0-branch. Consider this a call for Testers! I think our community testers would rather try this out on the bsd-3-0-0 branch than after it hits the stable trunk. I have three machines that I can test this on, one with a 7500, one with a 7200, and one with a Radeon Mobility M6. It'll take me a day or two to effectively test on all of these machines. Eric, can you give me a quick status of the DRI on NetBSD and OpenBSD? Adam --- This sf.net email is sponsored by:ThinkGeek Welcome to geek heaven. http://thinkgeek.com/sf ___ Dri-devel mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/dri-devel
Re: [Dri-devel] OS Independent Templates
Alright... Here's the first results :-) With a Radeon Mobility M6 and the bsd-3-0-0 branch, I get X locking up when I run any GL application. If I launch glxgears (or any other gl app that I've tried so far) from an xterm, I get: disabling TCL support radeonUpdatePageFlipping allow 0 current 0 I get the outline of the window for the application, but that's it. Everything window below the application is still visible. I can sitll move my mouse around, but the keyboard won't respond, and I can't kill the GL app or the X server. I have to ssh in to the machine and reboot. If I set LIBGL_DEBUB to verbose, I get some extra info, but nothing unusual or particularly telling. Unfortunately, I haven't actually updated the main branch of the DRI on this machine in at least a couple of weeks, so I'm not positive if this problem is specific to the bsd branch. I'll pull the main branch in a little bit and give it a whirl. On to the 7200... Everything seemed to work fine when I first started... glxgears gave comparable speeds to the main branch. A variety of the GL xscreensaver apps ran smoothly. gltron ran fine, as did tuxkart. I don't have any really big GL games (ie. games from Loki) on this machine to test with. I did manage to get a lockup (similar to the one I mentioned above) when I started engine (xscreensaver app) and then started circuit (xscreensaver app, too) at the same time. However, I didn't have that same problem when I started up glxgears twice. Anyway, this particular machine is my machine at work, so I'll have to stop testing for the time being. I'll try to get to the 7500 tonight. Adam On Tue, 2 Jul 2002, Adam K Kirchhoff wrote: On Mon, 1 Jul 2002, Jens Owen wrote: Eric Anholt wrote: Concerns at this point: - I'm still noticing some problems with my Radeon 7500 with DRI (the last lockup was dragging a glxgears around on top of another one). There has been talk of 7500 issues, though. - clipping issues on r128. r128 I know has problems on linux anyway. - one user I was talking with said he had glxgears above 2000fps (was it 2600?) on a 1700xp and radeon 7500 with 4.2.0/drm-kmod and it dropped to 1900 with bsd-3-0-0-branch. Consider this a call for Testers! I think our community testers would rather try this out on the bsd-3-0-0 branch than after it hits the stable trunk. I have three machines that I can test this on, one with a 7500, one with a 7200, and one with a Radeon Mobility M6. It'll take me a day or two to effectively test on all of these machines. Eric, can you give me a quick status of the DRI on NetBSD and OpenBSD? Adam --- This sf.net email is sponsored by:ThinkGeek Welcome to geek heaven. http://thinkgeek.com/sf ___ Dri-devel mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/dri-devel
Re: [Dri-devel] OS Independent Templates
On Tue, 2 Jul 2002, Adam K Kirchhoff wrote: Alright... Here's the first results :-) With a Radeon Mobility M6 and the bsd-3-0-0 branch, I get X locking up when I run any GL application. If I launch glxgears (or any other gl app that I've tried so far) from an xterm, I get: disabling TCL support radeonUpdatePageFlipping allow 0 current 0 I get the outline of the window for the application, but that's it. Everything window below the application is still visible. I can sitll move my mouse around, but the keyboard won't respond, and I can't kill the GL app or the X server. I have to ssh in to the machine and reboot. If I set LIBGL_DEBUB to verbose, I get some extra info, but nothing unusual or particularly telling. Unfortunately, I haven't actually updated the main branch of the DRI on this machine in at least a couple of weeks, so I'm not positive if this problem is specific to the bsd branch. I'll pull the main branch in a little bit and give it a whirl. Well, I've got good news and bad news. The good news is that the bsd branch performs the same on my Mobility M6 as the DRI trunk does. The bad news is that the DRI trunk has the same lock-up problem as the bsd branch. Adam --- This sf.net email is sponsored by:ThinkGeek Welcome to geek heaven. http://thinkgeek.com/sf ___ Dri-devel mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/dri-devel
Re: [Dri-devel] OS Independent Templates
Adam K Kirchhoff wrote: Alright... Here's the first results :-) With a Radeon Mobility M6 and the bsd-3-0-0 branch, I get X locking up when I run any GL application. If I launch glxgears (or any other gl app that I've tried so far) from an xterm, I get: disabling TCL support radeonUpdatePageFlipping allow 0 current 0 I get the outline of the window for the application, but that's it. Everything window below the application is still visible. I can sitll move my mouse around, but the keyboard won't respond, and I can't kill the GL app or the X server. I have to ssh in to the machine and reboot. If I set LIBGL_DEBUB to verbose, I get some extra info, but nothing unusual or particularly telling. Unfortunately, I haven't actually updated the main branch of the DRI on this machine in at least a couple of weeks, so I'm not positive if this problem is specific to the bsd branch. I'll pull the main branch in a little bit and give it a whirl. It's possible that the trunk is doing the same thing now. I didn't know we *ever* ran correctly on mobility chips. On to the 7200... Everything seemed to work fine when I first started... glxgears gave comparable speeds to the main branch. A variety of the GL xscreensaver apps ran smoothly. gltron ran fine, as did tuxkart. I don't have any really big GL games (ie. games from Loki) on this machine to test with. I did manage to get a lockup (similar to the one I mentioned above) when I started engine (xscreensaver app) and then started circuit (xscreensaver app, too) at the same time. However, I didn't have that same problem when I started up glxgears twice. Anyway, this particular machine is my machine at work, so I'll have to stop testing for the time being. Hmm. --- This sf.net email is sponsored by:ThinkGeek Welcome to geek heaven. http://thinkgeek.com/sf ___ Dri-devel mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/dri-devel
Re: [Dri-devel] OS Independent Templates
On Tue, 2002-07-02 at 16:19, Keith Whitwell wrote: Adam K Kirchhoff wrote: Alright... Here's the first results :-) With a Radeon Mobility M6 and the bsd-3-0-0 branch, I get X locking up when I run any GL application. If I launch glxgears (or any other gl app that I've tried so far) from an xterm, I get: disabling TCL support radeonUpdatePageFlipping allow 0 current 0 I get the outline of the window for the application, but that's it. Everything window below the application is still visible. I can sitll move my mouse around, but the keyboard won't respond, and I can't kill the GL app or the X server. I have to ssh in to the machine and reboot. If I set LIBGL_DEBUB to verbose, I get some extra info, but nothing unusual or particularly telling. Unfortunately, I haven't actually updated the main branch of the DRI on this machine in at least a couple of weeks, so I'm not positive if this problem is specific to the bsd branch. I'll pull the main branch in a little bit and give it a whirl. It's possible that the trunk is doing the same thing now. I didn't know we *ever* ran correctly on mobility chips. Could the problem simply be that the 3D driver unconditionally omits RADEON_TCL_BYPASS for SET_SE_CNTL_STATUS in radeon_state_init.c? BTW you keep talking of the 'Mobility chips', but only the M6 is problematic, the M7 has a TCL unit and works great. :) -- Earthling Michel Dänzer (MrCooper)/ Debian GNU/Linux (powerpc) developer XFree86 and DRI project member / CS student, Free Software enthusiast --- This sf.net email is sponsored by:ThinkGeek No, I will not fix your computer. http://thinkgeek.com/sf ___ Dri-devel mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/dri-devel
Re: [Dri-devel] OS Independent Templates
So I did some quick testing on the Radeon 7500... I won't be able to do any thorough testing till tomorrow, probably, but this is what I've found so far... For the most part, everything seems to run smoothly. I ran a number of gl apps without any issues. I was unable to replicate the lockup from dragging one glxgears window around on top of another, though I did this for at least a minute. I was able to replicate the same problem I had with the 7200: starting up circuit and engine caused the X server to lockup. With the 7200 it happened immediately, but with the 7500 it didn't happen till I clicked on the frame of one of the windows, thinking I'd drag it around. I'll try to do some more robust testing with that card tonight or tomorrow. Adam On Tue, 2 Jul 2002, Adam K Kirchhoff wrote: On Tue, 2 Jul 2002, Adam K Kirchhoff wrote: Alright... Here's the first results :-) With a Radeon Mobility M6 and the bsd-3-0-0 branch, I get X locking up when I run any GL application. If I launch glxgears (or any other gl app that I've tried so far) from an xterm, I get: disabling TCL support radeonUpdatePageFlipping allow 0 current 0 I get the outline of the window for the application, but that's it. Everything window below the application is still visible. I can sitll move my mouse around, but the keyboard won't respond, and I can't kill the GL app or the X server. I have to ssh in to the machine and reboot. If I set LIBGL_DEBUB to verbose, I get some extra info, but nothing unusual or particularly telling. Unfortunately, I haven't actually updated the main branch of the DRI on this machine in at least a couple of weeks, so I'm not positive if this problem is specific to the bsd branch. I'll pull the main branch in a little bit and give it a whirl. Well, I've got good news and bad news. The good news is that the bsd branch performs the same on my Mobility M6 as the DRI trunk does. The bad news is that the DRI trunk has the same lock-up problem as the bsd branch. Adam --- This sf.net email is sponsored by:ThinkGeek Welcome to geek heaven. http://thinkgeek.com/sf ___ Dri-devel mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/dri-devel --- This sf.net email is sponsored by:ThinkGeek No, I will not fix your computer. http://thinkgeek.com/sf ___ Dri-devel mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/dri-devel
