From: <bug...@bu...> - 2003-05-21 11:38:34
|
[This e-mail has been automatically generated.] Please do not reply to this email. if you want to comment on the bug, go to the URL shown below and enter your comments there. http://bugs.xfree86.org//cgi-bin/bugzilla/show_bug.cgi?id=271 mh...@ww... changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |mh...@ww... AssignedTo|dev...@bu... |dri- | |de...@li... Status|REOPENED |NEW ------- Additional Comments From mh...@ww... 2003-05-21 07:37 ------- This is likely a DRI related problem, reassigning to DRI project for comment. I agree there's no reason to believe this is distro specific. I will try to reproduce it sometime down the road as well. ------- You are receiving this mail because: ------- You are the assignee for the bug, or are watching the assignee. |
From: <bug...@bu...> - 2003-05-22 23:30:26
|
[This e-mail has been automatically generated.] Please do not reply to this email. if you want to comment on the bug, go to the URL shown below and enter your comments there. http://bugs.xfree86.org//cgi-bin/bugzilla/show_bug.cgi?id=271 ------- Additional Comments From xf...@ri... 2003-05-22 19:29 ------- In response to Mike's comment, I tried it but with the load dri option commented out in XF86Config-4. It still crashes. ------- You are receiving this mail because: ------- You are the assignee for the bug, or are watching the assignee. |
From: <bug...@bu...> - 2003-05-28 15:23:56
|
[This e-mail has been automatically generated.] Please do not reply to this email. if you want to comment on the bug, go to the URL shown below and enter your comments there. http://bugs.xfree86.org//cgi-bin/bugzilla/show_bug.cgi?id=271 ------- Additional Comments From jh...@pl... 2003-05-28 11:23 ------- I also get lockups with euphoria and other GL sreensavers. I have an ATI Rage 128 pro ultra TF (AGP) card. Using redhat linux 9 and the 'r128' driver. I logged into the system, from another one, the last time this occurred and could see the X server almost 100% CPU bound. Running strace seemed to show it looping around signal/sigreturn calls. Sorry, didn't capture the output - I'll do that. ------- You are receiving this mail because: ------- You are the assignee for the bug, or are watching the assignee. |
From: <bug...@bu...> - 2003-05-28 16:09:06
|
[This e-mail has been automatically generated.] Please do not reply to this email. if you want to comment on the bug, go to the URL shown below and enter your comments there. http://bugs.xfree86.org//cgi-bin/bugzilla/show_bug.cgi?id=271 ------- Additional Comments From jh...@pl... 2003-05-28 12:01 ------- Created an attachment (id=222) strace output of X server ------- Additional Comments From jh...@pl... 2003-05-28 12:09 ------- Interestingly the XFree86.0.log file contains: (EE) R128(0): Idle timed out, resetting engine... (EE) R128(0): R128WaitForIdle: CCE reset -1007 (EE) R128(0): R128WaitForIdle: CCE start -1007 (EE) R128(0): Idle timed out, resetting engine... (EE) R128(0): R128WaitForIdle: CCE reset -1007 (EE) R128(0): R128WaitForIdle: CCE start -1007 (EE) R128(0): Idle timed out, resetting engine... (EE) R128(0): R128WaitForIdle: CCE reset -1007 (EE) R128(0): R128WaitForIdle: CCE start -1007 (EE) R128(0): Idle timed out, resetting engine... (EE) R128(0): R128WaitForIdle: CCE reset -1007 (EE) R128(0): R128WaitForIdle: CCE start -1007 This occurred many times whilst the X server was at almost 100% cpu. ------- You are receiving this mail because: ------- You are the assignee for the bug, or are watching the assignee. |
From: <bug...@bu...> - 2003-07-18 14:17:59
|
http://bugs.xfree86.org/show_bug.cgi?id=271 ------- Additional Comments From xf...@ri... 2003-18-07 10:16 ------- The "colorfire" screensaver also does the same thing - it runs for about 2 seconds, and then locks the system hard. According to the documentation, "This didn't take me long to do, and I didn't write more than approximately 100 lines of code" So, it might be a promising test case. Richard -- Configure bugmail: http://bugs.xfree86.org/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are the assignee for the bug, or are watching the assignee. |
From: Dieter <Die...@ha...> - 2003-07-18 18:37:07
|
Am Freitag, 18. Juli 2003 16:16 schrieb bug...@bu...: > http://bugs.xfree86.org/show_bug.cgi?id=271 > > > > > > ------- Additional Comments From xf...@ri... 2003-18-07 10:16 > ------- The "colorfire" screensaver also does the same thing - it runs for > about 2 seconds, and then locks the system hard. According to the > documentation, > > "This didn't take me long to do, and I didn't write more > than approximately 100 lines of code" > > So, it might be a promising test case. Which driver (gfk)? XFree version? No problems, here. SuSE 7.3, KDE 3.1.2, kernel 2.4.22-pre5, DRI CVS r200. -Dieter |
From: Dieter <Die...@ha...> - 2003-07-21 13:59:52
|
Am Samstag, 19. Juli 2003 17:37 schrieb Richard Neill: > Dieter N=FCtzel wrote: > > Am Freitag, 18. Juli 2003 16:16 schrieb bugzilla-daemon@bugs.xfree86.or= g: > >>http://bugs.xfree86.org/show_bug.cgi?id=3D271 > >> > >> > >> > >> > >> > >>------- Additional Comments From xf...@ri... 2003-18-07 > >> 10:16 ------- The "colorfire" screensaver also does the same thing - it > >> runs for about 2 seconds, and then locks the system hard. According to > >> the documentation, > >> > >> "This didn't take me long to do, and I didn't write more > >> than approximately 100 lines of code" > >> > >>So, it might be a promising test case. > > > > Which driver (gfk)? > > XFree version? > > > > No problems, here. > > > > SuSE 7.3, KDE 3.1.2, kernel 2.4.22-pre5, DRI CVS r200. > > > > -Dieter > > r128 driver for an ATI Rage 128 Mobile chip. XF86 4.3-5mdk > I was the original bug reporter. > Shall I add these comments to the XF86 bugzilla? Yes, please. Regards, Dieter |
From: <bug...@bu...> - 2003-07-29 10:39:40
|
http://bugs.xfree86.org/show_bug.cgi?id=271 ------- Additional Comments From jh...@pl... 2003-29-07 06:38 ------- Progress perhaps?! I modified my XF86Config file to include the CCEusecTimeout option. By increasing this I have been able to run both colorfire and euphoria, both at work and at home, with no lockups. Euphoria is a bit 'jerky' though. The relevant section of my work XF86Config contains: Section "Device" Identifier "Videocard0" Driver "r128" VendorName "Videocard vendor" BoardName "ATI Rage 128 Pro Ultra" Option "AGPMode" "4" Option "CCEusecTimeout" "20000" EndSection The default value for CCEusecTimeout is 10000 (10 milliseconds). *HOWEVER* a) I'm not sure what effect this will have on anything else; (b) I use a locked screensaver and when I am prompted for the password it is slow to accept the characters - I assume this is because the screensaver is taking its time (but not timing out!) and hence slowing all the screen operations down. It does work though. I traced the 'Idle timed out, resetting engine' message to the XFree86 source file 'r128_accel.c'. Unfortunately it occurs in two places - I wish programmers would use unique error messages - so to trace this further will require modifying the file and rebuilding XFree86. -- Configure bugmail: http://bugs.xfree86.org/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are the assignee for the bug, or are watching the assignee. |
From: <bug...@bu...> - 2003-07-29 11:23:48
|
http://bugs.xfree86.org/show_bug.cgi?id=271 dw...@mr... changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |dw...@mr... ------- Additional Comments From dw...@mr... 2003-29-07 07:01 ------- *** Bug 344 has been marked as a duplicate of this bug. *** -- Configure bugmail: http://bugs.xfree86.org/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are the assignee for the bug, or are watching the assignee. |
From: <bug...@bu...> - 2003-07-29 15:48:17
|
http://bugs.xfree86.org/show_bug.cgi?id=271 ------- Additional Comments From mi...@da... 2003-29-07 11:47 ------- This makes sense; the chip can apparently be busy for more than 10 milliseconds, and a reset may actually make things worse. I don't think raising the default timeout should have any negative effects except possibly increased latency, and I guess that's preferable to a lockup. :) PS: We don't need to make the error messages unique because we know this must be in R128CCEWaitForIdle(). ;P -- Configure bugmail: http://bugs.xfree86.org/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are the assignee for the bug, or are watching the assignee. |
From: <bug...@bu...> - 2003-07-29 17:22:10
|
http://bugs.xfree86.org/show_bug.cgi?id=271 ------- Additional Comments From mi...@da... 2003-29-07 12:05 ------- Actually, it's 32 (R128_IDLE_RETRY) times 10 or 20 milliseconds... Also, I'm curious: does R128CCEWaitForIdle() deal more gracefully with the timeout if you add a R128CCE_STOP(pScrn, info); before the R128EngineReset(pScrn); ? -- Configure bugmail: http://bugs.xfree86.org/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are the assignee for the bug, or are watching the assignee. |
From: <bug...@bu...> - 2003-07-30 05:59:51
|
http://bugs.xfree86.org/show_bug.cgi?id=271 ------- Additional Comments From kt...@uk... 2003-30-07 01:58 ------- Note that the "radeon" driver has all the problems listed here in this bug too. It is NOT just the "r128" driver. Although it appears "radeon" is a sub- module, needing "r128". This latest "fix" is rejected by the "radeon" driver, to wit: (WW) RADEON(0): Option "CCEusecTimeout" is not used For the record, I have a: (--) Chipset ATI Radeon VE/7000 QY (AGP) found GL-based screensavers will hard lock my system, although the "really slick" ones listed here are pure death, instead of intermittent death. -- Configure bugmail: http://bugs.xfree86.org/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are the assignee for the bug, or are watching the assignee. |
From: <bug...@bu...> - 2003-07-30 11:39:17
|
http://bugs.xfree86.org/show_bug.cgi?id=271 ------- Additional Comments From mi...@da... 2003-30-07 07:38 ------- radeon and r128 are basically independent drivers. The corresponding radeon driver option is called "CPusecTimeout". -- Configure bugmail: http://bugs.xfree86.org/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are the assignee for the bug, or are watching the assignee. |
From: <bug...@bu...> - 2003-08-01 09:00:44
|
http://bugs.xfree86.org/show_bug.cgi?id=271 ------- Additional Comments From jh...@pl... 2003-01-08 04:56 ------- > PS: We don't need to make the error messages unique because we know this must > be in R128CCEWaitForIdle(). ;P > You are truely wise...I am not worthy :-) > Also, I'm curious: does R128CCEWaitForIdle() deal more gracefully with the > timeout if you add a R128CCE_STOP(pScrn, info); before the > R128EngineReset(pScrn); ? > Yes! I modified and rebuilt xfree86 last night at home, adding the line above. I commented out the CCEusecTimeout from my XF86Config file and restarted X. I then ran Euphoria and Colorfire for most of the evening without any lockups. I kept stopping and starting them, and changing their options. No problems at all. The idle/resetting messages still appeared in the log files (although not so many it seemed), but that was all. I had a very quick look at the radeon code, and I assume that a similar thing will be applied to that driver. I have no radeon card so can't test that at all. -- Configure bugmail: http://bugs.xfree86.org/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are the assignee for the bug, or are watching the assignee. |
From: <bug...@bu...> - 2003-08-01 17:08:01
|
http://bugs.xfree86.org/show_bug.cgi?id=271 ------- Additional Comments From xf...@ri... 2003-01-08 12:57 ------- Re #9, I've now tried it. Here's what I found: * With various values of the CCEusecTimeout (5000 -> 1000,000), I get a reduction in frames-per sec of about 3 (Measuring with GL-Snake (which is stable) goes from 30 down to 10 at full-screen). Colorfire really crawls. Euphoria --cubism will usually work OK (and I left it for 10 mins), but it is really jerky (about 1 fps) CPU load remains at full. * This value seems to require a reboot to properly change it. Adding the CCEusecTimeout line to XF86Config-4 and restarting X made the change "take", but to revert to the original settings required a reboot - restarting X didn't suffice (GL-snake was still ~10 fps). This seems Odd. * I did manage to lock Euphoria again. Now with access to another machine to ssh in from, I observe: * SSH from a remote machine works * The local keyboard does not work apart from some Sysrq keys (S,U,B work; R,K do not) * top via ssh reports the euphoria process as defunct, and the X process using 99 % CPU. * killing the x-server (via ssh) will recover the machine. Re #15, that sounds promising. If you'd like more testing, I'll willingly help, although I'd need a few pointers on what to do. On another tack, does anyone know if/how to define extra sysrq combinations and bind them to scripts of my own choice? (Eg, I'd like to have, say Alt-Sysrq-X mapped to "kill -9 `/sbin/pidof X`") Richard -- Configure bugmail: http://bugs.xfree86.org/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are the assignee for the bug, or are watching the assignee. |
From: <bug...@bu...> - 2003-08-01 23:16:20
|
http://bugs.xfree86.org/show_bug.cgi?id=271 ------- Additional Comments From mi...@da... 2003-01-08 19:15 ------- So you're saying it still locks up with a higher timeout for you? Re #15, I hope John can give you instructions. I'm afraid you can't do anything like this with Magic Sysrq - it's a lowlevel kernel debugging feature. However, IIRC Alt-Sysrq-K kills all processes on the current VT, which shouldn't be too different to what you want. :) -- Configure bugmail: http://bugs.xfree86.org/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are the assignee for the bug, or are watching the assignee. |
From: <bug...@bu...> - 2003-08-02 08:31:18
|
http://bugs.xfree86.org/show_bug.cgi?id=271 ------- Additional Comments From kt...@uk... 2003-02-08 04:30 ------- Note that the CPusecTimeout value does not have any significant effect on the radeon driver. Well, it does make things "jerky". But it does not effect or reduce the frequency or severity of hard lockups. I finally had to remove these screensaver programs from my system. Though I love them, they ARE "really slick", but the 20th hard lock is enough. It is clear the radeon driver in 4.3.0 needs serious work. More so than the r128 driver. -- Configure bugmail: http://bugs.xfree86.org/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are the assignee for the bug, or are watching the assignee. |
From: <bug...@bu...> - 2003-08-04 09:27:49
|
http://bugs.xfree86.org/show_bug.cgi?id=271 ------- Additional Comments From jh...@pl... 2003-04-08 05:26 ------- Re: #16 > * With various values of the CCEusecTimeout (5000 -> 1000,000), > But 5000 is less than the default of 10000 so I would expect lockups in this case. > I get a reduction in frames-per sec of about 3 (Measuring with GL-Snake (which > is stable) goes from 30 down to 10 at full-screen). Colorfire really crawls. > But isn't this expected? If you increase the timeout then the driver waits longer for the timeout so the screen 'hangs' until the timeout expires (or the relevant reset code happens) - thus frames per second will drop. Increasing the timeout avoids system lockups but does make the screensavers jerky. > CPU load remains at full. > Yes, these screensavers do seem cpu intensive when they kick in. Some have a 'nice' option to help. > * I did manage to lock Euphoria again. > What value of CCEusecTimeout? What Euphoria options? I'll email you directly what changes I made to XFree86. -- Configure bugmail: http://bugs.xfree86.org/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are the assignee for the bug, or are watching the assignee. |
From: <bug...@bu...> - 2003-08-08 16:48:58
|
http://bugs.xfree86.org/show_bug.cgi?id=271 ------- Additional Comments From id...@us... 2003-08-08 12:47 ------- To make it easier for others, the screensavers in question are available at http://rss-glx.sourceforge.net/. It also doesn't seem to want to build w/gcc 2.96, but gcc 3.3 seems to be okay. FWIW, it doesn't seem to crash on an r200. -- Configure bugmail: http://bugs.xfree86.org/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are the assignee for the bug, or are watching the assignee. |
From: <bug...@bu...> - 2003-08-08 16:55:13
|
http://bugs.xfree86.org/show_bug.cgi?id=271 ------- Additional Comments From ag...@ya... 2003-08-08 12:54 ------- I tried them a while back on my m6 and they seemed to work ok, although I didn't really do extensive testing. -- Configure bugmail: http://bugs.xfree86.org/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are the assignee for the bug, or are watching the assignee. |
From: <bug...@bu...> - 2003-08-14 14:20:36
|
------- Additional Comments From mi...@da... 2003-14-08 10:14 ------- The problem this bug is about is unlikely to be specific to these screensavers. -- Configure bugmail: http://bugs.xfree86.org/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are the assignee for the bug, or are watching the assignee. |
From: <bug...@bu...> - 2003-11-29 02:13:38
|
Please do not reply to this email: if you want to comment on the bug, go to the URL shown below and enter your comments there. http://bugs.xfree86.org/show_bug.cgi?id=271 rw...@gm... changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |rw...@gm... ------- Additional Comments From rw...@gm... 2003-11-28 21:13 ------- Hi, I dunno if this is related but I get these (EE) R128(0): Idle timed out, resetting engine... (EE) R128(0): R128WaitForIdle: CCE reset -1007 (EE) R128(0): R128WaitForIdle: CCE start -1007 Log entries when I enable DRI in my dual screen configuration with a r128 fury pro (AGP) and a Matrox Millenium G450DH (PCI). This occurs during X startup (which hangs then and heavily writes these Log entries). I didn't find much on this issue yet. Any help would be nice! thanks, ron -- Configure bugmail: http://bugs.xfree86.org/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are the assignee for the bug, or are watching the assignee. |
From: <bug...@bu...> - 2003-11-29 02:17:06
|
Please do not reply to this email: if you want to comment on the bug, go to the URL shown below and enter your comments there. http://bugs.xfree86.org/show_bug.cgi?id=271 ------- Additional Comments From rw...@gm... 2003-11-28 21:16 ------- Sorry, I forgot to say that I tried this timeout stuff already - even 80000 didn't help. - ron -- Configure bugmail: http://bugs.xfree86.org/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are the assignee for the bug, or are watching the assignee. |
From: <bug...@bu...> - 2004-03-29 20:31:26
|
Please do not reply to this email: if you want to comment on the bug, go to the URL shown below and enter your comments there. http://bugs.xfree86.org/show_bug.cgi?id=271 ------- Additional Comments From ne...@jm... 2004-03-29 15:03 ------- I have similar problems with these screensavers and my NVIDIA supplied driver - the processes get stuck at 100% and must be "kill -9"ed -- Configure bugmail: http://bugs.xfree86.org/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are the assignee for the bug, or are watching the assignee. |
From: <bug...@bu...> - 2004-03-30 08:19:30
|
Please do not reply to this email: if you want to comment on the bug, go to the URL shown below and enter your comments there. http://bugs.xfree86.org/show_bug.cgi?id=271 ------- Additional Comments From mi...@da... 2004-03-30 03:19 ------- (In reply to comment #25) > I have similar problems with these screensavers and my NVIDIA supplied driver - > the processes get stuck at 100% and must be "kill -9"ed That's a completely different problem. Only nVidia can support their proprietary drivers. -- Configure bugmail: http://bugs.xfree86.org/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are the assignee for the bug, or are watching the assignee. |