[R100 7200] lockups during/after dri

[Dave Gilbert]

Hi,
  (Hoary, Device database id: 20bff8ff4c11d05b1b87f7dc6f022115)
  I have a Radeon 7200 and find while 3d/dri works fine at startup after some
point (that seems random)
it breaks. If any 3D app starts after that point there is no response to
clicking on anything (of course
the wonders of it first happening in my screen saver - euphoria - were that I
could do nothing); remotely
logging in and killing the offending application frees the machine up.

A similar bug occured to me on debian/sid with xfree in which after the failure
point no 3d would get drawn - but at least
it wouldn't stop anything else.

(I'll add an lspci, conf file and log straight after I post this).
[lspci]
0000:00:00.0 Host bridge: Advanced Micro Devices [AMD] AMD-760 MP [IGD4-2P] System Controller (rev 11)
     Subsystem: Ensoniq Creative Sound Blaster AudioPCI128
0000:01:05.0 VGA compatible controller: ATI Technologies Inc Radeon R100 QD [Radeon 7200] (prog-if 00 [VGA])
     Subsystem: ATI Technologies Inc Radeon 7000/Radeon
0000:00:00.0 0600: 1022:700c (rev 11)
     Subsystem: 1274:2000

[Dave Gilbert (ubuntu-treblig)]

Created an attachment (id=2069)
Xorg.conf

[Dave Gilbert (ubuntu-treblig)]

Created an attachment (id=2070)
lspci -v and -vn

[Dave Gilbert (ubuntu-treblig)]

Created an attachment (id=2071)
Xorg.0.log (failure hasn't yet occurred during this boot)

[Daniel Stone (daniels)]

Upstream know about this and are working on a fix.

[Joe-ankhcraft (joe-ankhcraft)]

This desktop is an Athlon XP 1800+, w/ a VIA KT400 chipset on the motherboard.
The card I've got in there is a Radeon RV100 QY [Radeon 7000/VE] with 64 megs of
DDR sdram.

When I run glxgears w/ DRI enabled, the entire system locks up when the mouse is
moved, unless I set SWcursor on. By the entire system, I mean everything, not
just X. The only recourse is to press the Reset switch.

Although I'm glad I thought to try setting SWcursor on, this still is just a
hack to workaround a bug somewhere. Plus, the cursor disappears when I move it
over the glxgears window.

[Dmitry-butskoy (dmitry-butskoy)]

  The same for me (with the same card).
  Similar lock up is when I try to watch TV using "xawtv". Disabling dri helps
in this case (as for glxgears too), but using software cursor does not...

[Thomas Hood (jdthood)]

*** This bug has been marked as a duplicate of 2362 ***

[Eric Anholt (eric-anholt)]

Reopening bug because it is definitely not clear that this is the same thing
being referred to by the submitter of #2362.

[Thomas Hood (jdthood)]

See also #2581.

[Dave Gilbert (ubuntu-treblig)]

(In reply to comment #4)
> Upstream know about this and are working on a fix.

Is there a refernce in the upstream bugtracker for this that I can follow?

[Mikeklem (mikeklem)]

I am willing to donate a Radeon 7000/VE card to get this problem resolved,
please contact mikeklem [at] csh.rit.edu

[Michel-tungstengraphics (michel-tungstengraphics)]

Does this still happen with current versions of xf86-video-ati and Mesa?

[Erik Andrén (erik-andren)]

ping!

[Dave Gilbert (ubuntu-treblig)]

This still happens on Dapper.

[Dave Gilbert (ubuntu-treblig)]

I originally reported this way back in Hoary when it was in xfree86 I think.

[Paul Dufresne (paulduf)]

Marking as confirmed, because I have the feeling it is a duplicate of bug #16873. I did asked there someone to confirm.

[Timo Jyrinki (timo-jyrinki)]

In bug #16873 it's said that this is finally fixed in Ubuntu 6.10. Could someone confirm and mark both as fixed?

[Dave Gilbert (ubuntu-treblig)]

I'd love to say it was fixed - unfortunately it still happens on edgy for me.
I've just checked it and it still happens as per my original report on hoary.

Dave

[Timo Aaltonen (tjaalton)]

could this be marked as duplicate of 63503, since I already managed to file seven dupes for it :)

the issue is far from fixed. Maybe works for some cards but certainly not my 8500.

[Timo Jyrinki (timo-jyrinki-hut)]

This is interesting in that the bug is "null" status/resolution state currently. Trying to change something, sorry for the noise. Meanwhile, you (Joe, Dmitry, Mike) could tell if you are still seeing this with X.org 7.2.

[Dmitry-butskoy (dmitry-butskoy)]

Unfrotunately, the solution pointed out in bug #5986 (comes from Debian) does not help in my case:

ATI Radeon RV100 QY [Radeon 7000/VE]
xawtv-3.95, xorg-x11-drv-ati-6.5.8.0

Surely I've checked out not the patch itself, but its idea -- to comment out the fragment in src/radeon_accelfuncs.c :
> a->SubsequentSolidTwoPointLine
> = FUNC_NAME(RADEONSubsequentSolidTwoPointLine);
It seems it does NOT help...

[Dmitry-butskoy (dmitry-butskoy)]

BTW,
I'm free enough now to check some further ideas for xorg-x11-drv-ati
(Thanks x.org source is now splitted! ;) )

[Daniel Stone (daniels)]

Sorry about the phenomenal bug spam, guys. Adding xorg-team@ to the QA contact so bugs don't get lost in future.

[Timo Jyrinki (timo-jyrinki)]

I separated ths bug now again, as it seems the fix (patch for line acceleration) discussed in bug 63503 only fixes the problems with 9000/9100/9200-series, which means this bug is a different one.

It would be nice if someone with a r100 card (Radeon 7000 series or a Radeon Mobility M6) could try these:

1. Some known problems are fixed in Ubuntu 7.04 development version Herd-4, so the first thing one should do is to make sure you still can hang the machine with Herd-4. Success reports plus card model are welcome. http://cdimage.ubuntu.com/releases/feisty/herd-4/. Herd 5 will be available at http://cdimage.ubuntu.com/releases/feisty/herd-5 on Thursday.

2. If you manage to see the hangs with the Ubuntu development version, you could try the latest GIT version of xserver-xorg-video-ati next. It has even more fixes to hanging problems. I put compilation instructions (typo in the apt-get command though) near the end of bug 28925.

[Michel-tungstengraphics (michel-tungstengraphics)]

(In reply to comment #9)
> Unfrotunately, the solution pointed out in bug #5986 (comes from Debian) does
> not help in my case:

Yes, that's just a workaround for a very specific problem (lockups with CP line acceleration on RV280s).

> ATI Radeon RV100 QY [Radeon 7000/VE]
> xawtv-3.95, xorg-x11-drv-ati-6.5.8.0

If your problem only occurs with xawtv, it's probably not the same issue as reported here originally either.

Still waiting for updates on whether the original problem reported here persists with current releases.

[Dmitry-butskoy (dmitry-butskoy)]

Well,
I've upgraded my xorg-x11-drv-ati (xf86-video-ati) from 6.5.8.0 to 6.6.3 (have compiled it under the old Fedora Core 5 environment, i.e. xorg-x11-server-sdk-1.0.1 , with some little adaptation).

It seems that the issue is gone. Already 15 minutes with hard tests I cannot cause my compurter to freeze ;)

Note, that the issue still was present in version 6.6.1 .

Also note: to compile the 6.6.3 version under old plus "exa < 2.0" servers's environment, I have to apply the patch (attached) and cause 'USE_EXA' macro to be set anyway (because of some inaccuracy in ifdef code usage).

[Dmitry-butskoy (dmitry-butskoy)]

Created an attachment (id=8904)
patch to compile it under Fedora Core 5

Besides this patch, you have to use "make CC='gcc -DUSE_EXA'" to cause USE_EXA macro to be set, else it is failed to compile.

[Michel-tungstengraphics (michel-tungstengraphics)]

(In reply to comment #13)
> It seems that the issue is gone. Already 15 minutes with hard tests I cannot
> cause my compurter to freeze ;)

You're saying even xawtv survives now? :)

> Note, that the issue still was present in version 6.6.1 .

Yes, the fix for this only went into 6.6.2.

Closing, feel free to reopen if you still get lockups with a current driver on a Radeon 7000.

[Dmitry-butskoy (dmitry-butskoy)]

> You're saying even xawtv survives now? :)
Yep!

xawtv still works... :)

[Dave Gilbert (ubuntu-treblig)]

Hi,
  I upgraded to Feisty yesterday (dist-upgrade from edgy) and I've got to say this does look promising; I've run google earth a few times, and blender and a few other things and it's survived. However I did have days where it has worked previously; so it's a matter of giving it a little time to see how it plays out.

  Lets hope this one is finally squashed!

Dave

[Dave Gilbert (ubuntu-treblig)]

I'm afraid I spoke too soon.
In general use on the desktop it has been fine (I even had compiz running).
However, I had left euphoria screensaver running overnight and I've come back to it twice now humg; this time I ssh'd in and found it was a hung X server; killing the X server got me control back.
I've included two Xorg.0.log files; the 'hung' one is the one where it was hung - last thing is a power management command. (To make it clear, it has managed power saving and running this screensaver - so it's not an everytime). In this hung state the X server wouldn't let me attach with a gdb or allow me to strace it to see what it was up to.

I've also included the log from me trying to restart it - it complains of EngineRestore 'Idle timed out'.

Dave

[Dave Gilbert (ubuntu-treblig)]

and here is the restart one

[Timo Jyrinki (timo-jyrinki-hut)]

Reopening because a mobility m6 (ie. 7000-series) crash is still being reported with the 6.6.3 driver at https://bugs.launchpad.net/xserver-xorg-driver-ati/+bug/15219. X.org 7.2, Mesa 6.5.2, xserver-xorg-video-ati 6.6.3 (with a few additional patches from git).

Dave Gilbert says there:
---
"In general use on the desktop it has been fine (I even had compiz running).
However, I had left euphoria screensaver running overnight and I've come back to it twice now humg; this time I ssh'd in and found it was a hung X server; killing the X server got me control back.
I've included two Xorg.0.log files; the 'hung' one is the one where it was hung - last thing is a power management command. (To make it clear, it has managed power saving and running this screensaver - so it's not an everytime). In this hung state the X server wouldn't let me attach with a gdb or allow me to strace it to see what it was up to.

I've also included the log from me trying to restart it - it complains of EngineRestore 'Idle timed out'."

[Timo Jyrinki (timo-jyrinki-hut)]

Created an attachment (id=9280)
Log from when X was hung

Copied from ubuntu's bug reporting system

[Timo Jyrinki (timo-jyrinki-hut)]

Created an attachment (id=9281)
X failing to restart

Also copied.

[Freedesktop-treblig (freedesktop-treblig)]

(In reply to comment #17)
> Reopening because a mobility m6 (ie. 7000-series) crash is still being reported

Just to be clear my bug reported there is a Radeon 7200 (R100 QD).

Dave

[Timo Jyrinki (timo-jyrinki)]

I put your comment and logs to the upstream bug concerning radeon 7000 and mobility m6 series, reopening the upstream bug at the same time. If you can, please follow if there are any further questions there in the future...

[Dmitry-butskoy (dmitry-butskoy)]

> killing the X server got me control back.
It seems to be some another issue, since the entire system is not freezed (it's possible to login remote usung ssh) ...

[Timo Jyrinki (timo-jyrinki)]

Could someone test if these problems still occur with Ubuntu 7.10? I have one laptop with Mobility M6 that runs Ubuntu 7.10 fluently with 3D enabled.

[Dave Gilbert (ubuntu-treblig)]

Still seems to happen for me on Gutsy; (and I've updated my machine to latest BIOS as well to fix another issue).

For this test I ran running google-earth for a while seemed OK,
Then I ran planet penguin racer, got a lock in the racer, couldn't
gdb or strace the process; -9 killed it, X left in odd resoluiton

Restarted X, looks OK - but all GL apps
(glxgears, gltron etc) all just hang on startup.

I did a physical reset of the machine and it worked again - glxgears worked,
managed a few seconds in 'torcs' (racing game).

Dave

[Bugs-freedesktop-org (bugs-freedesktop-org)]

(In reply to comment #21)
> > killing the X server got me control back.
> It seems to be some another issue, since the entire system is not freezed (it's
> possible to login remote usung ssh) ...

dupe of bug 2581?

[Bryce Harrington (bryce)]

I don't know if this will fix this specific issue, but I've backported a bunch of high importance patches that upstream recommended, that fix problems sort of like this one, so I think it would be worth the time to test. Please try this .deb and report what you find:

http://people.ubuntu.com/~bryce/Testing/ati/

If we can determine that the patches in this deb fix this bug, they may be candidates for backporting to Hardy.

[Bugs-freedesktop-org (bugs-freedesktop-org)]

*** Bug 2581 has been marked as a duplicate of this bug. ***

[Brian Murray (brian-murray)]

Thanks for taking the time to report this bug and helping to make Ubuntu better. In the development cycle for Intrepid there have been some vast improvements in the open source ati video driver and we could use your help testing them. Could you please download the latest Alpha CD image of Intrepid and test this particular bug just using the Live CD? You can find the latest image at http://www.ubuntu.com/testing . Your testing can help make Ubuntu and the open source ati driver even better! Thanks in advance.

[Dave Gilbert (ubuntu-treblig)]

I've just tried the Alpha 4 live CD; and it's having video problems that maybe related to this or maybe something different.

I get the gdm login, but if I go into GNOME I get a Hardy wallpaper and then the screen goes all-white - with the mouse cursor on; I can move the cursor but I don't see anything else.

Failsafe-gnome exhibits the same behaviour (but failsafe seems to include compiz which seems IMHO a very bad idea)

Failsafe-xterm works. A glxgears from there seems to be OK.

metacity --replace from a vt fixes this.

I'm not sure if this should be filed as a separate report or if it's all part of the same thing.

Dave

[Timo Jyrinki (timo-jyrinki-hut)]

Could you by any chance try for example latest Ubuntu 8.10 alpha from http://cdimage.ubuntu.com/releases/8.10/ ? It includes all the latest fixes done with help of AMD releasing documentation, and is known to fix many problems. Dunno if Radeon 7000 series are affected or not.

[Freedesktop-treblig (freedesktop-treblig)]

I've tried on Ubuntu Intrepid 8.10 Alpha 4 and haven't had much look;
it's failing to open AGP so disabling DRI.

(It's also giving me a completely white screen if Compiz is on - but I guess that might be separate).

I've added a comment to the Ubuntu tracker of my matching Ubuntu bug; this is on a Tyan S2460/Dual athlon 760MP with Radeon R100QD/7200 - others may have more look.

[agd5f (agd5f)]

(In reply to comment #25)
> I've tried on Ubuntu Intrepid 8.10 Alpha 4 and haven't had much look;
> it's failing to open AGP so disabling DRI.
>

Can you attach your xorg log?

> (It's also giving me a completely white screen if Compiz is on - but I guess
> that might be separate).
>

Compiz requires the DRI. It won't work without it.

[Freedesktop-treblig (freedesktop-treblig)]

I've got the X.org logs and a boot messages in the matching ubuntu bug:

https://bugs.launchpad.net/xserver-xorg-driver-ati/+bug/15219

X.org log:

http://launchpadlibrarian.net/17039039/Xorg.0.log.intrepid

messages:
http://launchpadlibrarian.net/17039052/messages-gallifrey

I realise compiz needs DRI/3D - but leaving a fully white screen without anyway to get out of it was not what I was expecting.
(the mouse cursor still moved - so the X server was still alive).

(I got KDE4.1 to come up - it offered the choice to switch to Xrender rather than OpenGL - in OpenGL mode it also got white screen).

[Dave Gilbert (ubuntu-treblig)]

OK, just a follow up to this - I see the same behaviour on a full installer ather than running from the CD; however looking at the log files it looks like Intrepid isn't enabling AGP for some reason - I'm not sure whether to kick this off as a separate Intrepid bug or just keep it with this one.

I've included the Xorg.0.log from the full installation version - note the 'AGP not available' and associated messages.

[Dave Gilbert (ubuntu-treblig)]

and the /var/log/messages associated with it

[Dave Gilbert (ubuntu-treblig)]

I was about to add a comment saying that I had DRI back in the latest Intrepid update (as of Sat 30th about 11pm - kernel 2.6.27-2) - but then I noticed it had switched me to AIGLX and even though glxinfo showed direct rendering it was in fact using the 'DRISWRAST GL' module in the X server and still had the same AGP disabled message.

[Bryce Harrington (bryce)]

Based on the description of the symptoms and especially that it seems related to DRI and AGP, it sounds like this might be caused by an incorrect AGPMode setting, which is relevant when DRI is enabled. You should test if changing this makes the issue go away. In your /etc/X11/xorg.conf set AGPMode to various values like this:

Section "Device"
   ...
   Option "AGPMode" "2"
EndSection

Possible values include 1, 2, 4, 8. You can see what it's currently set to by looking in /var/log/Xorg.0.log. It's worthwhile to test all four values even once you find one that works. If this takes care of it for you, we may be able to establish a quirk to set it for your hardware combo, if you can provide the following data:

  * AGPMode value(s) that work
  * Make/Model of laptop or motherboard
  * Output of lspci -vvnn
  * Is the system all factory hardware, or have any parts been replaced?
  * Is there an AGP Mode in the system BIOS?
    - If so, is it set to the factory default?

For more details on this, please see the "ATI AGP Mode Quirk" section at https://wiki.ubuntu.com/X/Quirks

[Dennis Schridde (devurandom)]

Thanks a lot!

I had this issue on Debian Sid/Unstable with a Radeon M7 and AGPMode=2 fixed it.

IBM ThinkPad R40

01:00.0 VGA compatible controller [0300]: ATI Technologies Inc Radeon Mobility M7 LW [Radeon Mobility 7500] [1002:4c57] (prog-if 00 [VGA controller])

OpenGL vendor string: Tungsten Graphics, Inc.
OpenGL renderer string: Mesa DRI R100 (RV200 4C57) 20090101 AGP 2x x86/MMX/SSE2 TCL
OpenGL version string: 1.3 Mesa 7.6

[Dennis Schridde (devurandom)]

Hm, sorry, no, I take that back.
While initialy neverball worked, warzone2100 crashed, being unable to allocate a texture, and now I am at a kernel panic again (in init: general protection, attempted to kill init). (Before that happens, I get weird ELF "inconsistency" errors when I try to start programs, like assertions failing, or sections being missing.)

[Corbin Simpson (mostawesomedude)]

Closing; not sure if FIXED or WONTFIX is more accurate, since the code has changed so much since the last substantial report. If there's still problems with a current stack, please open a new bug.

[Bryce Harrington (bryce)]

[This is an automatic notification.]

Hi Dave,

This bug was reported against an earlier version of Ubuntu, can you
test if it still occurs on Lucid?

Please note we also provide technical support for older versions of
Ubuntu, but not in the bug tracker. Instead, to raise the issue through
normal support channels, please see:

    http://www.ubuntu.com/support

If you are the original reporter and can still reproduce the issue on
Lucid, please run the following command to refresh the report:

  apport-collect 15219

If you are not the original reporter, please file a new bug report, so
we can work with you as the original reporter instead (you can reference
bug 15219 in your report if you think it may be related):

  ubuntu-bug xorg

If by chance you can no longer reproduce the issue on Lucid or if you
feel it is no longer relevant, please mark the bug report 'Fix Released'
or 'Invalid' as appropriate, at the following URL:

  https://bugs.launchpad.net/ubuntu/+bug/15219

[Vish (vish)]

Thank you for taking the time to report this bug and helping to make Ubuntu better. We are closing this bug report because it lacks the information we need to investigate the problem, as described in the previous comments. Please reopen it if you can give us the missing information, and don't hesitate to submit bug reports in the future.
To reopen the bug report you can click on the current status, under the Status column, and change the Status back to "New".