/srv/irclogs.ubuntu.com/2010/03/22/#ubuntu-x.txt

Sarvattbryceh: where is the codename attachment to the bug title done in your scripts? it would be a lot more useful to have chipset codenames like RV515 instead of say M64P00:04
Sarvattah fix-titles.py found it00:07
Sarvattbryceh: extending fix-titles to have all ati pci ids and use the chipset codename if you dont mind00:15
Sarvatthelps a lot more tracking down bugs if i know the codename like RV515 and there are a *ton* missing in here00:16
brycehSarvatt, great thanks00:20
brycehyeah the nvidia chipsets are missing there too if you feel ambitious00:20
brycehI run fix-titles.py on -intel manually from time to time, and run it less frequently on -ati00:21
brycehonce you've got your changes in I can re-run it on both (unless you'd rather do it)00:21
brycehI figure eventually we'll want it for nouveau too00:21
RAOFYeah, that'd be useful.00:22
RAOF(Although it's not always obvious which set of codenames to use; many nvidia chips have at least 2 :/)00:23
Sarvattyup will do, 1/4 the way through ati and will do nvidia after00:23
Sarvattnvidia is easy, theres a nice wiki document having them all00:23
Sarvatthttp://nouveau.freedesktop.org/wiki/CodeNames00:24
Sarvatti've got a list of all pci ids somewhere around here to map that to00:24
Sarvattphew, gonna take a break at R500 generation :D00:44
libvSarvatt: check radeonhd code, i spent several days in july 2007 doing the same thing00:54
Sarvattok got everything except cypress hemlock juniper cedar and redwood, not sure on those codenames01:30
Sarvattbryceh: anything special I should do running this? already changed it to dryrun01:32
Sarvatthmm   ---------> Renamed to [RV530] [X1600] jaunty: X session (radeon) slows down, crashes01:35
hyperairSarvatt: i still get my artifacts =\01:48
Sarvattthere we go, all ati bug titles with proper logs tagged with the gpu chipset name, kind of throws off incomplete bugs running this though since it resets expiration time02:37
Sarvattahh darn, things originally filed against xorg dont have PciDisplay: in the description02:39
Sarvattpushed it here but it needs handling for already tagged bugs like cleanup_intel has http://bazaar.launchpad.net/~sarvatt/arsenal/working/revision/49102:56
vishhi.. i'm noticing my screen being badly corrupted with artifacts > and the xorg log is showing :04:08
vish(II) RADEON(0): Printing DDC gathered Modelines:04:08
vish(II) RADEON(0): Modeline "1280x800"x0.0   68.90  1280 1301 1333 1408  800 804 808 816 -hsync -vsync (48.9 kHz)04:08
vish(II) RADEON(0): EDID vendor "QDS", prod id 6504:08
vishbut i'v selected 60kHz in the display settings [which is the default]04:08
vishthis is in Lucid^04:08
vish60*Hz04:10
vishbrb , the screen looks horribly disfigured :(04:10
vishif i take a screenshot , its doesnt show these problems..04:14
RAOFWhich suggests that the problem is in how radeon is driving your screen, as you suspected.04:15
RAOFHas this issue arisen recently?04:17
vishRAOF: hmm , that Xorg log entry might not be the error , but there is no error/message recorded other than those lines..  but yes , this has started recently , i notice the screen corrupted and a restart is the only thing that fixes it and screenshots dont catch the problem :(04:20
RAOFBut a restart fixes it?  It wouldn't seem to be mode-related then.04:21
vishyeah , the restart restores the screen to normal ,  i have never had such bad display , I'v been using Ubuntu for on this system for nearly 2.5yrs04:21
vishlaptop*04:22
RAOFIs there a particular action that you've discovered which triggers the corruption?  Is it only after leaving the computer for a while?  Is it while doing a lot of 3D stuff?  Is it just anytime?04:22
vishi have notice the problems when i return from screensaver[rather after the screen has been switched off during idle].. not while any other usage/3D04:23
RAOFAlways after the return from screensaver?04:24
vishit has been random .. i'm not sure what triggers it.. but i have noticed the problem only after i return from screensaver04:24
vishno errors in the Xorg/sys logs . no messages , nothing is recorded :s04:26
vishi could take foto of the error , but i dont think it would be helpful , not sure where/how i can debug this :(04:28
RAOFA photograph of the problem can be surprisingly useful.04:28
vishok , i'll take one the next time it happens04:29
RAOFThis only happens after a return from screensaver, though?  Is it a particular screensaver?04:29
vishi use the GLXMatrix.. i could try switching to a different screensaver04:30
RAOFMight be worth checking, yeah.04:30
* vish switches screensaver04:32
vishthere has also been another similar error , but it is "fine black noise"[1px black horizontal lines] which also doesnt get caught when using a screenshot as well.. but this started happening several times *while* i was using the system , no fancy 3D stuff , or anything.. this too is solved only by a restart04:35
vishsame no errors/messages :(  04:36
* vish thinks the new drm from .33 kernel might be the cause of these problems 04:36
RAOFIf you think that's the case then you could try booting into the old -15 kernel04:37
vishyeah i should, i didnt notice them earlier , its been happening only for the last couple of weeks.  These random errors are quite bugging :s  if it is something i can reproduce it can be pinned down04:39
vishRAOF: if the bug happens again , anything else i need to log? to narrow down the bug? [other than a foto , which might help the first bug,  but i'm not sure it would help the black noise]04:44
RAOFdmesg & xorg.0.log are the obvious candidates; I'm not aware of anything else for radeon.04:45
vishneat.. thanks [but hope they log errors the next time ;-)]04:47
RAOFIt's quite possible that you can pass some debug parameters to the kernel module, but I'm unfamiliar with them.04:48
RAOFAnd they're likely to be *hugely* verbose.04:49
vishwhom/where can i ask about those debug parameters? [#u-kernel?]04:53
RAOFI'd just wait for the moment.04:57
vishk.. thanks :)04:57
Sarvattvish: in /sys/module/radeon/parameters/ can you tell me what the values for dynclks and new_pll are?04:59
vishSarvatt:  dynclks is -1   and new_pll   is 105:01
Sarvatttry booting with radeon.new_pll=005:01
* Sarvatt thought that was the default...05:02
vishSarvatt: so the  new_pll one i just change to 0   right?05:03
vishor   > radeon.new_pll=005:03
vishah , got it.. you want me to edit the kernel line :)05:03
Sarvattyeah, sorry05:05
Sarvatthttp://bugs.freedesktop.org/show_bug.cgi?id=2635805:05
ubottuFreedesktop bug 26358 in DRM/Radeon "new_pll algo gives flickering LVDS on rv620 (HD 3470 mobility) laptop" [Normal,Resolved: fixed]05:05
Sarvattweird refresh bugs and quirks abound because of new_pll=1..05:06
Sarvattyou're like the third person in here today to say they have a weird undescribable corruption they cant screenshot on radeon, sheesh05:07
vish\o/ i'm not alone ;p  05:08
vishbtw , this is the Xorg.0.log > http://paste.ubuntu.com/399115/05:09
vishXorg.1.log > http://paste.ubuntu.com/399114/05:09
Sarvattyou're the second with an RV515 today too05:12
Sarvattother guy couldn't file a bug because of some KDE problem :(05:12
Sarvatt<Consul_Falx> [18:03:35] +Sarvatt: flickery, snowy or like that ...05:14
Sarvattwhatever that meant :)05:14
vishSarvatt: you want a bug logged for this? 05:16
vish[to switch the default new_pll=0]05:17
Sarvattdoubt thats a good idea globally, probably need to make a quirk for your machine and the hundreds of others that need them too05:18
libvheh, i cannot believe deucher is still struggling with that, i had that fixed before he even started writing his driver.05:18
Sarvattyuck, all the odd flickering bugs the past few days in -ati are all rv515 and rv530 :D05:42
libvm52 or rv515/530?05:43
Sarvatt2xM56P and a x1300 RV515 on this page05:46
Sarvattat least yours doesnt look like this vish - http://launchpadlibrarian.net/41493534/543265_bug_P04.jpg 05:46
libvthat's not a pll issue05:47
libvor...05:47
vish;p 05:47
libvhrm...05:48
RAOFWhee!  I like my framebuffer well sliced!05:48
SarvattEverytime I start Ubuntu 10.04 from Live CD, after few minutes screen start to shaking and i cant do anything (Ctrl+Fn is working, but screen is shaking too).05:48
Sarvatt(description for that last one)05:48
Sarvattwith the screenshot05:49
libvctrl-fn on kms no doubt05:49
superm1Sarvatt, i've seen something similar to that on my netbook too05:49
superm1but i can't come up with a reproducible scenario for it, it's quite random05:50
libvi fear that this photo is that messed up because of the camera05:52
RAOFAnd that the true state of affairs would not include duplicate pieces of screen?05:53
libvnot for a modesetting issues indeed.05:54
Nginterestingly enough, I think I've found a couple of other people seeing similar weird random flickery events on G45s in Lucid09:14
Sarvatttell them to boot with i915.powersave=0 and mention in the bug if that fixes it :)09:18
Sarvattthat was default in -15 but not anymore in -1609:18
tseliotSarvatt: unfortunately I experienced some weird corruptions with nouveau (and I had to update libdrm too)09:20
Sarvatttseliot: i told ya in the email you have to update *everything* in the ppa, not just xserver-xorg-video-nouveau too :D09:20
* tseliot nods09:20
Sarvattgot it running though? didn't work? :(09:21
tseliotwell, it depends on what you mean by "work" ;)09:22
Sarvattwhat kinds of corruptions?09:23
tseliotfirst of all plymouth didn't work (and it used to work) but I got mode setting. Then I could enter the desktop but the menus and the indicator were blank09:25
NgSarvatt: interesting. the problem is that it's *really* intermittent, so all I'd be able to say is that it's not happened for a few days ;)09:26
freeflyinghi all, does xserver-xorg-video-ati works well with xrandr in karmic09:27
tseliotSarvatt: http://albertomilone.com/2010-03-21 19.08.04.jpg09:28
tseliot:-/09:28
tselioterr... http://albertomilone.com/2010-03-21%2019.08.04.jpg09:28
freeflyinga 17" crt monitor has been recoganized as 16" 09:28
Sarvatti haven't seen any flickering bugs on 965+ recently, looked like they ironed them all out but powersave=0 for sure should get rid of them Ng09:28
freeflyingusing xserver-xorg-video-ati in karmic09:28
NgSarvatt: ok, I've passed it on to the other folks and we'll try to start keeping track. Would anything else have changed between -15 and -16 wrt intel? I probably only rebooted into -16 fairly recently09:29
Sarvattyeah *tons*, the entire 2.6.32-2.6.33 drm backport09:29
Ngoh hrm, I rebooted longer ago than I thought09:30
Ng17th I think09:30
Ngok well we'll keep an eye on it and see how it goes :)09:30
=== radoe_ is now known as radoe
tseliotSarvatt: can you add this, please? http://0x04.net/~mwk/0001-drm-nv50-Add-NVA3-support-in-ctxprog-ctxvals-generat.patch10:17
Ngok, I was just using another intel laptop in the office that just had a fresh install of beta1 and it was doing the screen flickering thing just like mind, so I'll reboot wit powersave=0 and see if it goes away10:50
bjsnidertseliot, i noticed last night that nvidia-settings control file still says build on lpia in lucid12:31
tseliotbjsnider: as the rest of the nvidia packages12:32
bjsniderwhy doesn't the build system complain about it?12:32
tseliotI don't know12:34
bjsnideri find it easier and safer to just put "any" int he architecture12:34
tseliotno, why should we?12:38
tjaaltonjust drop lpia12:39
tjaalton"any" doesn't work12:39
tjaaltonno reason to build it on ia64 or sparc12:39
bjsniderwell, when you use the ppa build system, and tell it to build on lpia/lucid, it errors out. the ppa system doesn't build on ia64 or sparc12:40
tjaaltonbut the main buildd's do12:41
bjsnideria64 is the intel itanium?12:41
tjaaltonyep12:41
bjsnideri thoughtt hat was an experiment?12:41
bjsniderare there ia64 production systems out there?12:42
tjaalton"amd64 armel i386 ia64 powerpc sparc", those are the archs that "any" would make it build on12:42
tjaaltonask HP12:42
tjaaltoni guess there are, don't know about ubuntu on them12:43
tjaaltonbut that's besides the point12:43
tjaaltoneven though _your_ ppa packages build fine with "any", there's no need to change the main package to do the same12:43
=== Sinnerman is now known as Cobalt
marcus___hello14:21
marcus___i was curious if it is possible to have the gallium3d nouveau drivers in lucid via the xorg swat/edgers ppa14:21
marcus___can someone offer some insight into this? :)14:22
arabryceh, ping15:46
superm1tseliot, re https://edge.launchpad.net/ubuntu/+source/fglrx-installer/2:8.721-0ubuntu5 , that shouldn't have been necessary once the binaries were deleted from the archive should it?15:53
Sarvatttseliot: uploaded now, sorry had to sleep a bit there :)15:55
tseliotsuperm1: you might want to discuss this with pitti (who requested transitional packages)15:55
Sarvattit's linux-backports-modules-2.6.32_2.6.32-16.999~git20100321.4d950853~xorgedgers15:55
superm1tseliot, ah i'll take a look at bt 15:55
tseliotSarvatt: thanks a lot. Getting some sleep can help from time to time ;)15:55
superm1tseliot, but then those changes just need to be local, and not in the amd phorogit stuff, right?15:55
superm1i'm guessing they help dist-upgrader's is the particular reason15:56
tseliotsuperm1: right15:56
superm1ara, with your proprietary driver testing, is one of your test cases hardy fglrx->lucid fglrx now too?  hardy still had LRM if i'm remembering correctly, i'm wondering how well that would go over15:57
tseliotoh, and BTW suspend/resume works with nvidia now (even here!)15:59
aratseliot, \o/16:00
tseliot:-)16:01
=== yofel_ is now known as yofel
ricotzSarvatt, hello could you please upload a lbm package for 2.6.32-17 as well?16:27
Sarvattis it the default already?16:28
ricotznot yet16:28
Sarvatti dont see a -17 linux-meta16:29
Sarvattyeah not yet sorry, can only have one in there at a time16:29
ricotzSarvatt, ok16:29
Sarvattmight need to update 16 again for tseliot here16:29
ricotzSarvatt, perhaps you can look at this patch http://repo.or.cz/w/mesa/mesa-lb.git/commit/432b8cc477344fc7fcbb24b747d56ab087403d0916:30
ricotzwhich might solves the crash on my nv4916:30
tjaaltonlbm doesn't have nouveau anymore aiui16:31
Sarvatti'm added it back to lbm on edgers since its alot easier to update that16:51
tjaaltonah, k16:53
=== fedoralogger is now known as apachelogger
Sarvattone day i'll work out how to properly create a PPA kernel :D17:00
technovikingMy nvidia driver goes into safe mode when I boot into Lucid. restarting gdm will make the nvidia and X load properly. Anyideas what is causing the problem. https://bugs.edge.launchpad.net/ubuntu/+source/nvidia-graphics-drivers/+bug/53243617:27
ubottuUbuntu bug 532436 in nvidia-graphics-drivers "nvidia driver sometime does not load at boot" [Undecided,New]17:27
tseliottechnoviking: can you reproduce the problem if you uninstall (with --purge) plymouth?17:38
Sarvattlooks like a regression with ati 6.12.192 https://bugs.freedesktop.org/show_bug.cgi?id=2718619:20
ubottuFreedesktop bug 27186 in Driver/Radeon "Visual corruption with new r6xx/r7xx accel code" [Normal,New]19:20
Consul_Falxhello folks19:22
Sarvattradeon.new_pll=0 quirk needed here https://bugs.edge.launchpad.net/ubuntu/+source/xserver-xorg-video-ati/+bug/54304519:45
ubottuUbuntu bug 543045 in xserver-xorg-video-ati "[RV530] screen flickers in lucid lynx" [High,Incomplete]19:45
Sarvatti'm thinking it should be the default for all RV515 and RV530 mobiles19:46
=== radoe_ is now known as radoe
Sarvatti'm just going to tag them needs-pll-quirk to start making a list to give to the kernel guys to selectively use that option. having the generation in the title sure does help spot these19:59
Consul_Falxey folks20:26
Consul_Falxany progress in ati radeon open driver support in lucid?20:27
brycehSarvatt, sounds good20:29
=== KB1JWQ is now known as kb1
=== kb1 is now known as KB1JWQ
tormodanyone who can sponsor bug #287475 please?20:58
ubottuLaunchpad bug 287475 in xserver-xorg-video-sis "ASUS A6000 / SIS M760GX grainy graphic display" [Undecided,Confirmed] https://launchpad.net/bugs/28747520:58
jcristaui can sponsor it to sid :)21:00
tormodjcristau, very well, would you like a debdiff or would you just copy in that patch yourself?21:10
tormodor let me commit it to git :)21:11
vishi'm having a weird problem , the system slows down and everything is slow to react[random]. the mouse pointer has no lag though... earlier i used to SAK and return to session and everything would be fine.. ..  now , i seem to have a workaround of sorts.. when i switch to a guest session and return , everything is normal and works fine... nothing in the logs or .xsession-errors .. where do i need to look to file a bug?21:19
jcristautormod: i'd say just push it to git.d.o.  you have commit access, right?21:21
tormodjcristau, no not to the x packages21:21
vishnot sure if this means anything > [ 5291.324934] __ratelimit: 3 callbacks suppressed  kern.log > http://paste.ubuntu.com/399544/21:22
jcristautormod: should we fix that? :)21:23
tormodjcristau, thanks could be handy :)21:23
jcristautormod: should be done now.  or whenever that thing gets updated.21:29
tormodjcristau, should I keep it "unreleased" or "unstable" ?21:31
jcristaui don't think there'll be any more changes so unstable should be fine.  iirc there's a debian bug that this could fix, too.21:33
tormodthanks, I'll look for it21:33
tjaaltonjust got seven packages synced from unstable21:34
tormodjcristau, actually do you usually add a debian/patches/foo or use git cherrypick there?21:41
jcristautormod: usually cherry-pick -x.  and keep debian/patches/ for not-yet-upstream stuff.  but i don't really mind, debian/patches/ works for this too if you prefer.21:42
brycehhi tormod, do you still plan to update debian's intel-gpu-tools package to latest?21:42
tormodhi bryce, I was thinking to wait for a release.21:44
RAOFMan.  You watch the nouveau bug lists, and note that people hardly ever ask for mmiotraces.  Then we turn nouveau on by default, and all the bugs we send up are “oh, yeah, could you get an mmiotrace of that?  We've never had one for that card before” :X21:52
brycehhehe21:53
alkisgIs Intel 82852/855GM ok with xorg in Lucid? Or I'd better keep running Karmic?21:53
brycehSarvatt, yeah with that pll quirk, as you verify bugs are resolved with that, repoint those bugs to linux for the kernel team.22:16
brycehSarvatt, might be good to also set a milestone and target them to lucid22:17
alkisgAh ok, I was lacking xv and I was afraid that 855GM wasn't working at all in Lucid, but it's just a kms combination problem: https://wiki.edubuntu.org/LucidLynx/ReleaseNotes#No%20Xv%20support%20for%20Intel%2082852/855GM%20video%20chips%20with%20KMS22:19
tormodjcristau, I pushed the cherry-pick. I hope I got that right.22:21
tormodhmm maybe I got the version wrong :(22:23
jcristauah yeah the changelog entry should be merged with the previous one22:23
jcristauDEBCHANGE_RELEASE_HEURISTIC=changelog in ~/.devscripts helps with that22:23
jcristau(if you're using dch)22:23
jcristautormod: didn't you want to close the lp bug in the changelog too?22:25
tormodI thought at the time that 1:0.10.2-2 was already released...22:25
jcristauit says UNRELEASED ;)22:25
tormodjcristau, oh yes, I will add it22:25
tormodjcristau, I was not sure if it was so consistent :) but of course it was22:26
tormodun-goofing it now22:27
tormodjcristau, getting there :) pushed22:32
jcristaulooks good22:35
tormodjcristau, not sure how I update upstream-unstable The Right Way22:36
tormodgit push origin upstream-unstable tells me Everything up-to-date22:36
jcristauyeah upstream-unstable stays on the 0.10.2 tag22:36
tormodyes, it was upstream-experimental I had updated ! time for sleep maybe22:37
jcristauthanks, and good night!22:38
tormodgood night! btw, will you release soon so we can sync in ubuntu instead of patching?22:39
jcristauyes22:39
jcristaueither now, or tomorrow morning when i get to the office.22:40
Sarvatthttp://lists.x.org/archives/xorg-devel/2010-March/006436.html \o/ we have so many bugs about that22:56
jcristauok, tormod's sis fix is in incoming now.23:04
stgraberhey there !23:11
stgraberI have a few thin clients (Netbook like hardware) with Intel 945GM (8086:27a6) showing a LVDS1 output since a few days on Lucid. It used to be a problem back in Jaunty if I remember well where I had to make a few scripts to automatically blacklist that output on non-netbook but it was fixed in karmic/early-lucid.23:12
stgraberI quickly went through a few bugs on LP and it's not very clear if it's a known issue or not and if it's something that we plan on fixing for lucid or not (if we don't, I'll restore the old hack in the ltsp package and upload it to Lucid to fix the issue immediately)23:12
jcristauso the bios pretends there's an lvds?23:15
stgrabermight be, yep. A few thin clients seem to have BIOS reporting LVDS output23:15
stgraberthough usually they are blacklist (in the driver IIRC) because the Chassis type and some other info report that it's desktop hardware and not laptop23:16
RAOFDo they have ACPI lid switches?  i915.ko shouldn't attempt to bring up LVDS screens without a lid-switch, I think.23:17
stgraberRAOF: no /proc/acpi/button/lid (only button/power is there)23:18
RAOFHm.  Maybe that commit isn't in our kernel?23:19
stgraberI'd certainly love to see that check in as it'll solve a lot of issues I have with Atom-based thin clients (exact same hardware as netbooks but used as desktop with a single DVI-I output usually)23:20
jcristauRAOF: 11ba159288f1bfc1a475c994e598f5fe423fde9d reverted that23:23
RAOFjcristau: Ah, ok.23:24
jcristauapparently there's stuff with a lid switch but no lvds23:24
jcristauso the driver trusts the bios.23:24
stgraberhmm, if the issue is stuff with a lid switch and no lvds, how can it hurt to disable LVDS if it's reported by the BIOS and no lid switch is found ?23:25
stgraberI'd guess the issue is for stuff with LVDS but no lid switch23:25
RAOFRight.23:25
jcristaustgraber: there's a quirk list in drivers/gpu/drm/i915/intel_lvds.c where your stuff could be added23:26
jcristaustatic const struct dmi_system_id intel_no_lvds[] = {23:26
jcristau"These systems claim to have LVDS, but really don't"23:26
stgrabergood, only issue is to make a rule that matches my hardware but not a generic netbook then23:26
jcristauwith mac minis, aopen something, and dell studio hybrid in there already.23:27
* stgraber updates his copy of ubuntu-lucid.git23:30
stgraberjcristau: http://paste.ubuntu.com/399595/23:36
stgraberIs that something you can commit yourself or should I open a bug on LP + poke the kernel team about it ?23:37
jcristaushould probably poke your kernel team, or send it upstream to eric@anholt.net, dri-devel@lists.sf.net23:38
stgraberjcristau: ok, I'll open a bug on LP and e-mail the patch to these two e-mails as well23:40

Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!