[01:05] Hi [01:07] heya [01:07] if realloc crashes, couldn't you try valgrind? [01:09] FWIW, I'm continuing with the tail-end of the respins; it doesn't lose us that much even if we end up respinning for this [01:09] And it will let Kubuntu get going with validation [01:10] but what exactly is the issue now? I haven't really followed it as I was trying to fall asleep [01:11] bug 1124660 [01:11] bug 1124660 in xorg (Ubuntu) "Precise 20120213 i386 live session fails in virtualbox" [Undecided,New] https://launchpad.net/bugs/1124660 [01:12] As I think Sarvatt pasted, my take is that I don't currently see how the /dev/port patch could have changed memory allocation behaviour, and thus there is a tempting working hypothesis that says that this was lurking all along and that it's possible that we've just gone from fatal server error to a realloc crash, in which case who really cares [01:12] But Sarvatt is right to flag a possible regression and I think we should make sure we understand it [01:12] backport X stack + vesa is working fine with the newer libpciaccess on intel, http://paste.ubuntu.com/1647335/ [01:13] cjwatson, yes seems plausible [01:13] sleep> tell me about it, I got three hours last night and I've been up for 17 hours now [01:13] i've only been able to reproduce it in virtualbox so far [01:13] Sarvatt: what makes you think it's an abi break though? [01:13] I don't see anything in the patch that owuld indicate that [01:14] mlankhorst, the stacktrace looked like what you might get from an abi break. just a guess though. [01:15] meh valgrind it [01:15] * mlankhorst checks if virtualbox works [01:16] nope, guess ill netboot and do it over nfs on a crappy 100 mbit :/ [01:18] Isn't the failure specific to vbox hardware? [01:18] probably [01:18] plars might still be around and walkable through it [01:18] 23:58 * plars -> supper, then back for more iso testing [01:18] but i can run it in a virtualbox over netboot [01:18] that was 80 minutes ago [01:18] ah [01:19] it's also very slow but I'm not really awake anyway [01:20] hmm, but vesa works when booting the livecd, its just when you click try ubuntu that there is a crash [01:22] The effect of "Try Ubuntu" is to fall out of ubiquity-dm and let lightdm start [01:22] So it'll be tearing down the server and starting a new one after whatever random plumbing wibble happens around there [01:22] ^- technical term [01:23] Heh [01:26] * mlankhorst tries hard not to fall asleep [01:26] cd is http://cdimage.ubuntu.com/precise/daily-live/current/precise-desktop-i386.iso ? [01:26] Yes [01:27] * RAOF wonders what will happen first: the fibre-to-the-home rollout hitting his house, or that iso downloading. [01:29] RAOF: axel? [01:29] great, virtualbox doesn't work on lts kernel [01:30] Sarvatt: ??? [01:31] axel http://cdimage.ubuntu.com/precise/daily-live/current/precise-desktop-i386.iso [01:31] ugh [01:31] axel -n 32 if its slow? :P [01:31] fine try 2 from quantal itself [01:32] looks like virtualbox hates me today [01:32] Sarvatt: Ah. The "make the server admin hate you" application :) [01:43] vbox is only mildly slow [01:46] blah [01:46] cd panics on me [01:48] virtualbox from quantal [01:51] Hm. Seems to work fine in kvm, too. [01:53] second try starts normally [01:53] I'm just going to pretend I didn't see a kernel panic [01:57] :) [01:57] same thing happening here though [02:00] The realloc crash? [02:01] well hitting failsafe-mode but really too tired to look atm, i dont think its a regression though, just something that wasn't hit before [02:02] OK; thanks for the investigation, at present I'm minded to not let this block [02:02] downgrading libpciaccess just gives me the black screen and [ 206.226] (II) VESA(0): VESA BIOS not detected this "fixed", not even failsafe [02:03] cjwatson, thanks. should the issue get mentioned in some release notes maybe? [02:04] I would welcome an X developer writing an appropriate note in https://wiki.ubuntu.com/PrecisePangolin/ReleaseNotes/CommonInfrastructure [02:04] << dlrrp [02:05] on it [02:07] This sounds like a problem I used to run into sometimes on previous distros, not in a VM but on a couple of different laptops. I found that if I clicked on "Try Ubuntu" it would sometimes hang, but if I closed the window by clicking on (X), I would get back to an X session every time. [02:09] Sometimes enters failsafe-x mode when booting into Live Session in virtualbox. (LP: #1124660) [02:09] Launchpad bug 1124660 in xorg (Ubuntu) "Precise 20120213 i386 live session fails in virtualbox" [Undecided,New] https://launchpad.net/bugs/1124660 [02:09] more to say than that? [02:10] what if you try libpciaccess from quantal? [02:11] bryce: on i386 only, can be worked around by using amd64? [02:11] Sarvatt, ok [02:12] mlankhorst: trying that now [02:12] alrighty, done. [02:13] mlankhorst: that works! [02:20] http://cgit.freedesktop.org/xorg/lib/libpciaccess/patch/?id=a798395a1bfd9d06d40e2d8d14377a156c94429a ...? [02:21] bryce: Well, the problem is also that failsafe-X mode doesn't work, AIUI from the bug [02:21] Sarvatt: sounds likely [02:21] thats it [02:22] with that commit added it works again [02:22] Sounds possibly lucky that amd64 works [02:22] cjwatson, ah right [02:22] OK, let's have that patch for 12.04.3 then, thanks :) [02:23] SRU after, if someone installs with updates checked it'll pull it in and work after boot? [02:23] or maybe its reallocating SOMEONE ELSE's allocation [02:23] alrighty! [02:23] ;P [02:23] Sarvatt: I'm not in much of a panic about virtualbox not working - dailies aren't a terrible option there [02:23] * Sarvatt nods [02:23] The main value is helping our own testers [02:24] really sorry for all the trouble here [02:24] But I think they might actually crucify me if I respun at this point for only that :) [02:24] Don't worry, it doesn't look like you broke anything === tomreyn_ is now known as tomreyn === Quintasan_ is now known as Quintasan [12:18] morning [16:43] I have read that ubuntu 12.04.2 comes with kernel 3.5. Will it work with proprietary ati drivers for radeon 4000 series as they are not compatible with quantal? [16:44] if you use the old xserver/kernel sure [16:45] which is what you keep if you upgrade from 12.04.1 to 12.04.2 [16:45] mlankhorst: just old xorg or kernel too? [16:46] the combination old xserver + new kernel is unsupported [16:46] it will probably work, but it may set your cat on fire [16:48] mlankhorst: so that means I just have to work with kernel 3.2. Probably but the kittens will extinguish it [16:49] lets hope so :) [16:49] mlankhorst: yeah :) [16:50] mlankhorst: amd did a update to its legacy drivers to version 13.1. Do they work with 3.5 and new xorg? [17:07] mlankhorst, prime helpers have landed in drm-next. [17:07] starks: ah, still not that useful [17:07] needs fencing? [17:08] doesn't actually do anything [17:08] at least not what I care about [17:08] can nvidia release a driver against 3.9? [17:08] makes me wonder why nvidia cares though [17:08] I'm not nvidia === yofel_ is now known as yofel [20:23] what was the metapackage that reverts to the original precise stack? [20:24] xserver-xorg-lts-precise? [20:24] that was it [20:25] step 4 is unneeded [20:26] we use autobind patch for now