| Sarvatt | bdmurray: http://bugs.freedesktop.org/show_bug.cgi?id=26887 | 00:00 |
|---|---|---|
| ubottu | Freedesktop bug 26887 in DRM/Radeon "fence errors with rs785 and kernel 2.6.33" [Critical,New] | 00:00 |
| bdmurray | will do shortly thanks | 00:02 |
| Sarvatt | if that works thats something to add to the notes for sure | 00:14 |
| Sarvatt | someone else had that problem a month or two ago (vish I think?) | 00:14 |
| bryceh | bdmurray, have you noticed these freezes prior to today? | 00:37 |
| bdmurray | bryceh: no not at all | 00:38 |
| bryceh | bdmurray, do you update daily? | 00:38 |
| bdmurray | bryceh: pretty close | 00:39 |
| bryceh | bdmurray, there was one update to -ati today, although I think it wouldn't cause any user-noticeable effect, it just drops an unnecessary config file | 00:40 |
| bryceh | bdmurray, but it would be helpful if you could downgrade to 1:6.13.0-1ubuntu1 which would rule out that and another change | 00:40 |
| bryceh | my guess is you'll still be able to repro it | 00:41 |
| Sarvatt | yeah theres absolutely no way that has an affect on what he's getting, the 6.12.192-6.13 uprade on the other hand.. | 00:47 |
| bdmurray | we need some what was updated on my system in the last update tool | 00:49 |
| bryceh | yeah | 00:51 |
| Sarvatt | yeah a graphical apt log viewer that shows the history and condenses the info sure would be nice | 00:51 |
| bdmurray | it'd need to show apt and update-manager stuff | 00:51 |
| bryceh | /var/log/dpkg.log shows this info, so as a first order you can look in it | 00:53 |
| bdmurray | synaptic has some strangely incomplete history | 00:53 |
| Sarvatt | someone poke someone interested in making stuff with quickly or something :) | 00:53 |
| bryceh | bdmurray, I've listed out in the bug several X packages which have changed recently that could be worth testing downgrades | 00:53 |
| bryceh | I reviewed the changes for each and nothing lept out at me as likely to cause this problem | 00:54 |
| bryceh | but then, I wouldn't have included the patches if I could spot an obvious problem with htem ;-) | 00:54 |
| Sarvatt | bdmurray: did you try the dynclks change? | 00:54 |
| bdmurray | Sarvatt: no, doing that now | 00:54 |
| bdmurray | Sarvatt: it hasn't crashed yet | 01:04 |
| Sarvatt | good idea, saying that always makes it happen :) | 01:07 |
| bryceh | heh | 01:09 |
| bdmurray | yeah it finally barfed again | 01:19 |
| Sarvatt | wow, you've gotta be kidding me.. looking through server motherboards to see what the common GPU's are there these days and.. matrox g200?! | 04:32 |
| Sarvatt | i had one of those before linux kernel 1.0! :) | 04:33 |
| Sarvatt | the large number of boards with XGI gpu's confuses me, guess thats where they disappeared to | 04:39 |
| vish | Sarvatt: freezes in compiz? i read the backlog , but i dont understand what bdmurray is mentioning , does the system become slow to respond? that i what I'v been experiencing, for a month | 05:01 |
| vish | what i do is restart compiz and everything gets back to normal.. IMO , doesnt seem like a compiz bug , but i dont know where to report it as it is random and i didnt have anything getting logged :( | 05:03 |
| bdmurray | vish: what I was mentioning is a total system crash | 05:24 |
| bdmurray | bryceh: I've narrowed it down | 05:24 |
| Sarvatt | vish: lots of fence timeouts and soft lockups | 05:37 |
| Sarvatt | vish: http://www.google.com/search?sourceid=chrome&client=ubuntu&channel=cs&ie=UTF-8&q=vish+going+to+reset+gpu :D | 05:38 |
| Sarvatt | (first result) | 05:38 |
| Sarvatt | yours looks really different looking at that though | 05:39 |
| Sarvatt | bdmurray: so xserver-xorg-video-ati 1:6.12.192-2ubuntu2 is working ok but 6.13 isn't? | 05:40 |
| vish | hehe , pastebin and google bad ;p | 05:56 |
| vish | Sarvatt: btw , how do i workaround the memory bug in ATI ? Bug 563400 , i dont seem to have the option for BO reuse in driconf | 06:01 |
| ubottu | Launchpad bug 563400 in xserver-xorg-video-intel "Playing flash video causes memory hogging" [Undecided,Confirmed] https://launchpad.net/bugs/563400 | 06:01 |
| Sarvatt | yeah thats an intel thing | 06:02 |
| Sarvatt | i have no idea | 06:02 |
| Sarvatt | sudo apt-get purge adobe-flash would work :D | 06:02 |
| vish | ;p | 06:02 |
| Sarvatt | will keep an eye out about it though | 06:02 |
| Sarvatt | it's only in chromium here | 06:03 |
| Sarvatt | well chrome too | 06:03 |
| vish | hmm ,happens here in firefox and chrome[ium] | 06:03 |
| Sarvatt | haven't worked out running chromium through valgrind yet, darn wrappers | 06:03 |
| vish | Sarvatt: also having this bug > http://www.youtube.com/watch?v=C_ozjS55mW8 , i should file this in xserver-xorg-video-ati or mesa? | 06:06 |
| vish | if i log into a second account its like a disco ;) | 06:07 |
| Sarvatt | hmm what GPU again? | 06:08 |
| Sarvatt | rv515 maybe? | 06:08 |
| Sarvatt | or rv530 | 06:08 |
| Sarvatt | vish: if its either of those boot radeon.newpll=0 | 06:09 |
| vish | i think it is rv530 .. | 06:10 |
| Sarvatt | looks like the pll rework stuff we pulled in isn't exactly fixing things, some people on rv515 and rv530 were getting what you see there on first boot for awhile but now its just guest sessions | 06:10 |
| vish | or rv515 | 06:12 |
| vish | Sarvatt: how do i check the GPU? | 06:12 |
| Sarvatt | yeah all of the bugs closed about flickering on both of those chipsets because of newpll=1 have people saying thats happening with guest sessions now | 06:13 |
| Sarvatt | vish: grep PCI: /var/log/Xorg.0.log | 06:14 |
| vish | weird : $ grep PCI: /var/log/Xorg.0.log | 06:15 |
| vish | (--) PCI:*(0:1:0:0) 1002:7145:1025:0094 ATI Technologies Inc Radeon Mobility X1400 rev 0, Mem @ 0xd0000000/134217728, 0xc8100000/65536, I/O @ 0x00002000/256, BIOS @ 0x????????/131072 | 06:15 |
| Sarvatt | yep thats a RV515 | 06:15 |
| * vish scratches head | 06:16 | |
| vish | Sarvatt: how did you find that? | 06:16 |
| Sarvatt | there were tons of reports with that same pci id :) there's a list here though - http://cgit.freedesktop.org/xorg/driver/xf86-video-ati/tree/src/ati_pciids_gen.h | 06:18 |
| vish | ah :D | 06:18 |
| Sarvatt | 1002 is ati, 7145 is the product id on that list | 06:18 |
| vish | neat , thanks.. | 06:19 |
| === raof is now known as RAOF | ||
| Sarvatt | interested in hearing if radeon.newpll=0 fixes it for you too though | 06:25 |
| Sarvatt | bryceh: has anyone contacted you about these synaptics ClickPad patches? | 06:27 |
| Sarvatt | oh tseliot closed the -synaptics task... | 06:28 |
| Sarvatt | ah found the bug, i thought it was odd the kernel support was brought in this late but not the userspace side but it makes sense now, at least they get 1 button with just the kernel side :) | 06:34 |
| RAOF | What signal does X die with when it dies with an EQ overflow? | 06:43 |
| RAOF | Or do I just get the reporter to set a breakpoint in xorg_backtrace? I guess that'd work. | 06:44 |
| Sarvatt | RAOF: same bug on a NVA5 - http://bugs.freedesktop.org/show_bug.cgi?id=26980#attach_33897 | 06:56 |
| ubottu | Freedesktop bug 26980 in Driver/nouveau "GT230M/nouveau: X server hangs spontaneously" [Normal,New] | 06:56 |
| RAOF | Sarvatt: Right. I was looking for a backtrace, because there appear to be at least three separate bugs there. | 06:57 |
| RAOF | (I mean, there are three separate fedora bugs that it could be) | 06:57 |
| RAOF | Every time I interact with a bugzilla instance I appreciate launchpad just a little bit more. | 07:00 |
| Sarvatt | well its not https://bugzilla.redhat.com/show_bug.cgi?id=566987 since we dont even have pcie aspm enabled | 07:01 |
| ubottu | bugzilla.redhat.com bug 566987 in xorg-x11-drv-nouveau "X freezes the clock stopped, keyboard seemed not to be responding, every so often the mouse would track but no response to clicks." [High,Assigned] | 07:01 |
| RAOF | It's probably not https://bugzilla.redhat.com/show_bug.cgi?id=532579 either. | 07:03 |
| RAOF | :) | 07:03 |
| ubottu | bugzilla.redhat.com bug 532579 in xorg-x11-drv-nouveau "[nouveau] - X crashes before login: [Mat Booth] [9500 GT]" [High,Assigned] | 07:03 |
| Sarvatt | argh, forgot to bump nouveau abi's for .21 | 07:04 |
| RAOF | You know what would make this easier? Having a nouveau_gpu_dump tool like the intel ones. This particular symptom apparrently means “the GPU wedged somehow” | 07:04 |
| Sarvatt | copied the maverick 2.6.34-rc4 kernel into edgers, i'm about to pass out though so not going to get around to updating nouveau and figuring out how to require it | 07:14 |
| Sarvatt | it's going to take 20+ hours to build on amd64 anyway | 07:14 |
| RAOF | :) | 07:14 |
| Sarvatt | just got some chromium codec updates so I think I can guess why :) | 07:15 |
| RAOF | Chromium doesn't take *that* long to build, does it? | 07:24 |
| RAOF | ♪ La la la la / la la la lai / all god's little children, the gotta die ♫. Nick Cave. Such a happy dude. | 07:36 |
| vish | Sarvatt: if i use radeon.newpll=0 i get error [ 10.019888] radeon: Unknown parameter `newpll' | 07:44 |
| vish | and i cant start compiz :( | 07:44 |
| Sarvatt | new_pll=0 | 07:44 |
| vish | oh , k.. | 07:45 |
| * vish tries | 07:45 | |
| Sarvatt | newpll=0 made it not load at all :) | 07:45 |
| Sarvatt | sorry about that if i said that | 07:45 |
| vish | hehe , yeah .. rebooting | 07:45 |
| RAOF | It's a nice trick to *really* prevent the driver being loaded ;) | 07:45 |
| vish | np.. :) | 07:45 |
| vish | Sarvatt: yay , that does it , setting radeon.new_pll=0 solves the guest session problem :) | 07:52 |
| vish | thanks.. | 07:52 |
| === Amaranth_ is now known as Amaranth | ||
| vish | Sarvatt: i think i got hit by bdmurray's bug , http://paste.ubuntu.com/415461/ | 11:22 |
| vish | everything just froze and i couldnt do anything :( | 11:22 |
| vish | that happened with the new_pll=0 , or is this a different bug? | 11:23 |
| vish | the error log just went on repeating and there is a huge log file with the repeat error | 11:23 |
| === |thade| is now known as Alexia_Death_ | ||
| bjsnider | what's the jockey-text command to select nvidia-current? | 14:47 |
| bjsnider | tseliot, what happens if a guy is running the sudo jockey-text -e xorg:nvidia-current command and it is returning unkown driver? | 15:19 |
| Sarvatt | its xorg:nvidia_current | 15:20 |
| * tseliot nods | 15:23 | |
| bjsnider | he says it still returns unknown driver | 15:23 |
| bjsnider | i noticed the description for nvidia-173 and 96 mentions vdpau | 15:26 |
| bjsnider | tseliot, instead of using transitional packages, couldn't you use provides: conflicts: replaces: ? | 15:27 |
| bjsnider | i mean for nvidia-current | 15:27 |
| tseliot | bjsnider: those descriptions are wrong then | 15:28 |
| tseliot | and no, conflicts/replaces won't work in this case | 15:28 |
| bjsnider | it won't upgrade people? | 15:29 |
| tseliot | in short, yes | 15:29 |
| tseliot | make sure that nvidia-common is installed and then type "jockey-text -l" to see what's available | 15:29 |
| bjsnider | he didn't have nvidia-common but he just installed it. | 15:30 |
| bjsnider | does jockey run as a service? would he have to restart? | 15:31 |
| tseliot | just close jockey and make sure that jockey-backend is not running any more, then launch jockey again | 15:32 |
| Sarvatt | vish: thats not good, sounds like we have a serious regression recently then since you're on completely different hardware | 15:45 |
| Sarvatt | bdmurray: did you say you dont have the problem using 6.12.192 but do with 6.13? | 15:45 |
| vish | Sarvatt: yeah , everything was nice initially , only recently as we near release all new bugs seem to occur :( | 15:47 |
| vish | even if GLmartix screensaver is running CPU usage just shoots up | 15:47 |
| vish | rather, with any GL* screensaver | 15:48 |
| vish | even if i preview from screensaver settings , the CPU spikes | 15:50 |
| bdmurray | Sarvatt: yes that is correct | 15:53 |
| Sarvatt | vish: do you have time to mess with it? can you try downgrading -ati to 6.12.192-2ubuntu2 to see if its any different? | 15:53 |
| vish | Sarvatt: downgrade ,by just downloading the 6.12.192-2ubuntu2 and installing it right? i havent tried downgrading earlier | 15:54 |
| Sarvatt | yeah | 15:55 |
| Sarvatt | just that one package | 15:55 |
| bdmurray | it might be in your /var/cache/apt/archives folder too | 15:55 |
| vish | Sarvatt: ok , will do that in a bit , the freeze happened today once at random , but the CPU spike is constantly reproducible , will check it out | 15:56 |
| Sarvatt | i dont see anything between 6.12.192 and 6.13 that should affect rv515 the same way :( | 16:19 |
| Sarvatt | bryceh: 117_fix_crash_with_createglyphset.patch in xserver is the same as 110_findglyphbyhash-fix.patch | 16:28 |
| Sarvatt | how the heck did they both apply | 16:29 |
| Sarvatt | easily I guess - http://paste.ubuntu.com/415626/ | 16:50 |
| jcristau | yay for fuzzy patch applying :) | 16:52 |
| Sarvatt | heads up that lbm-nouveau is now deprecated in xorg-edgers if anyone in here was using it, I'm uploading a xserver-xorg-video-nouveau package now that requires a 2.6.34 based kernel and there is one in the PPA that'll have to be manually installed (2.6.34-1-generic|preempt|generic-pae|whatever) and mainline kernels based on .34+ with the lucid config work also | 17:26 |
| Duke` | wow I see updates on karmic too o/ | 17:30 |
| Sarvatt | i'm leaving all the old lbm-nouveau packages in the PPA so they are removed by ppa-purge properly and put a note on the main page about it | 17:32 |
| Sarvatt | man, rs600 is in horrible shape - http://lists.freedesktop.org/archives/dri-devel/2010-April/000094.html | 18:24 |
| Sarvatt | i think dell latitude XT is the only laptop with it, haven't been able to find another.. lucid isn't bootable without disabling KMS currently on those | 18:29 |
| === KiBi_ is now known as KiBi | ||
| Sarvatt | wiki says abit is the only one that released a motherboard with RS600 since AMD/ATI tried to clear all RS600 parts out of the market after the merger | 18:31 |
| Sarvatt | bryceh: patch series that reverts all of the -ati accel stuff back to 6.12.192-2 status - http://sarvatt.com/downloads/radeon/ | 18:48 |
| vish | Sarvatt: it doesnt let me to downgrade , gdebi keeps saying "Error: A later version is already installed" | 18:51 |
| bryceh | Sarvatt, thanks; do you think we need to go that far? | 18:52 |
| bryceh | Sarvatt, and btw you're a mind-reader, I was *just* contemplating about doing this ;-) | 18:53 |
| bryceh | sweeeet - http://cgit.freedesktop.org/xorg/driver/xf86-video-ati/commit/?id=a69e749d0562887af6bd236c38802472e54640c4 | 19:11 |
| vish | hmm , is it safe to uninstall xserver-xorg-video-ati first and then install the lower version, to downgrade? | 20:01 |
| bryceh | vish, I've posted a ppa with the patches reverted - https://edge.launchpad.net/~bryceharrington/+archive/silver | 20:11 |
| bryceh | vish, please test that | 20:11 |
| bryceh | also note you can downgrade usually by just doing apt-get install foobar=1.2.3 | 20:11 |
| vish | bryceh: ah neat , thanks | 20:12 |
| vish | hmm , the ppa has yet built :s | 20:36 |
| vish | and when i try ~$ sudo apt-get install xserver-xorg-video-ati=1:6.12.192-2ubuntu2 | 20:36 |
| vish | i get E: Version '1:6.12.192-2ubuntu2' for 'xserver-xorg-video-ati' was not found | 20:36 |
| bryceh | vish, ah yeah I think the version has to be present in your apt cache in order to downgrade like that | 20:40 |
| bryceh | anyway, the ppa should be built some time today, I'm sure the buildd's are bogged down with release stuff right now | 20:41 |
| vish | i downloaded and added the deb to the cache as xserver-xorg-video-ati_1%3a6.12.192-2ubuntu2_i386.deb but still it doesnt detect | 20:42 |
| vish | well , i'll just wait | 20:42 |
| Sarvatt | vish just sudo dpkg -i xserver*.deb where you downloaded it | 21:03 |
| * vish tries | 21:05 | |
| vish | yay, downgrading... | 21:05 |
| Sarvatt | woohoo flight booked for UDS :) | 21:07 |
| bryceh | Sarvatt, :-) | 21:09 |
| vish | Sarvatt: i think the memory leak for my ATI happens for any video or any increased graphics usage.. i hadnt played any flash video files but still memory /swap 70% was used :( | 21:11 |
| vish | is used* | 21:11 |
| * vish reboots | 21:15 | |
| vish | bryceh: Sarvatt: downgrading, for sure , prevents high CPU usage by the GL* screensavers | 21:22 |
| * vish now waits and watches an memory problems or fence timeouts | 21:26 | |
| Sarvatt | google earth has been good at breaking radeon wonderfully for a few years now :D | 21:27 |
| Sarvatt | oh sweet, thanks for pushing that ati with the reverts bryce, i was on the road and unable to so i just sent along the patches | 21:29 |
| bryceh | Sarvatt, that was half th ework ;-) | 21:29 |
| bryceh | actually I just pushed it to a ppa; I'd sort of like to get some indication that this is the right way to go before pushing it to lucid | 21:29 |
| Sarvatt | yeah thats what I meant, i wanted to put it in a PPA but couldn't | 21:30 |
| bryceh | go teamwork ;-) | 21:30 |
| Sarvatt | I didn't create the patch series in the wrong order did I? :D | 21:31 |
| bryceh | nope, they all applied just fine | 21:31 |
| Sarvatt | oh nope it built, phew | 21:31 |
| bryceh | yeah in reviewing that, it occurs to me that it was probably unwise for upstream to shove in so many performance reworking patches so close to their release | 21:31 |
| Sarvatt | they pulled an intel :) | 21:32 |
| bryceh | if it were me I would have held those back for putting in after the .0 release | 21:32 |
| bryceh | these silly driver developers :-) | 21:32 |
| bryceh | I don't even want to look at -intel right now | 21:32 |
| Sarvatt | dont think i'd merge that ati uevent patch, the intel one causes lots of problems in fedora.. | 21:35 |
| bryceh | yeah no plans to | 21:35 |
| Sarvatt | at least its upstream now though and likely to get fixed | 21:35 |
| vish | bryceh: your ppa is the same as xserver-xorg-video-ati 1:6.12.192-2ubuntu2 ? or a 6.13 + patches ? | 21:35 |
| bryceh | looks like good xorg-edgers fodder though | 21:35 |
| bryceh | vish, 6.13 - patches actually ;-) | 21:35 |
| Sarvatt | vish: its 6.13+ extra fixes with just the sketcky stuff removed that was added between 6.12.192 and 6.13 | 21:35 |
| vish | hehe ;) | 21:35 |
| bryceh | vish, it is close to 6.12.192 but has some additional upstream changes beyond that (which I think should be quite safe) | 21:36 |
| vish | ah ok. , will test that too | 21:36 |
| Sarvatt | i dont understand why 6.12.192-2 is fine for you too though vish, looked like those commits only affected r600+ | 21:37 |
| bryceh | Sarvatt, what's vish's card? | 21:38 |
| vish | Sarvatt: not sure , but only since the .13 update this week , i have been noticing the screensaver and the memory problems | 21:38 |
| Sarvatt | rv515 (the flicker monster that hes having to use new_pll=0 on still) | 21:39 |
| Sarvatt | only thing i can think of is those extra commits work right with exa from xserver 1.8 and not 1.7.x (fedora is backporting 1.8's exa even in F-12) | 21:39 |
| Sarvatt | http://cgit.freedesktop.org/xorg/xserver/log/exa (all of the stuff from december on isn't in 1.7 branch just about) | 21:42 |
| Sarvatt | vish: so it was hanging with the relocation errors for you too like bdmurray or was it just spewing that and not actually locking up on your rv515? | 21:51 |
| Sarvatt | vish: are you on x64? have you tried bryce's PPA version if not? can you build the package yourself? it's pretty important that we get testing ASAP and theres about a 15 hour queue for x64 | 21:53 |
| Sarvatt | add the ppa and update, sudo apt-get build-dep xserver-xorg-video-ati, apt-get source xserver-xorg-video-ati, then cd to the directory and debuild -uc -us -b to build it | 21:54 |
| Sarvatt | just dug out a machine to throw a HD2400 pro in but i wont be able to try it out until later over the weekend | 21:57 |
| vish | i'm on 32bit , the freeze just locked everything , i couldnt do anything .. | 21:57 |
| vish | i'v added the bryceh's ppa , but since it was still building , i was updating other stuff.. will do the above once the update is done | 21:58 |
| Sarvatt | vish: 32 bit has been built for 2 hours, should be there whenever you are able to try it (and thanks for trying out all of these things) :) | 22:51 |
| vish | Sarvatt: hehe! i was just about to ask you regarding a problem.. when i doing apt-get it pulled source from the main repos ;) | 22:59 |
| vish | nvm , got the ppa now though :) | 22:59 |
| vish | weird , didnt notice the ping so long :s | 23:00 |
| Sarvatt | just pinged you a few minutes ago :) | 23:01 |
| vish | yeah , i didnt notice it 7 mins earlier, probably was busy trying to figure out how to get the source from the ppa ;) | 23:02 |
| Sarvatt | no need to get the source now at least :) | 23:03 |
| vish | yup , just installed , will reboot and keep you informed... | 23:03 |
| vish | neat, with bryceh 's ppa , no problem of screensaver using high CPU.. | 23:14 |
| vish | will watch out for other memory or fence timeouts problems | 23:15 |
| bryceh | vish, had you filed a bug report about your issue? What's the lp#? | 23:15 |
| vish | bryceh: i had filed a bug for the memory problem , but it might be unrelated to Sarvatt's BO reuse > Bug 563400 , i notice the memory leak even without watching flash , havent filed a bug for the screensaver though | 23:17 |
| ubottu | Launchpad bug 563400 in xserver-xorg-video-intel "Playing flash video causes memory hogging" [Undecided,Confirmed] https://launchpad.net/bugs/563400 | 23:17 |
| bryceh | vish, thanks | 23:21 |
| vish | np.. | 23:21 |
| bryceh | uploaded to lucid | 23:23 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!