[12:16] kylem: works on my pata_amd [12:16] lazka, er, maybe? [12:16] kylem, what should be the dev name? [12:16] mount -t iso9660 /dev/scd0 /mount [12:16] FYI my problem persists with a new install from today's daily build. [12:17] /dev/cdrom [12:17] trejack: Try: http://people.ubuntu.com/~kyle/kernels/feisty/linux-image-2.6.20-15-generic_2.6.20-15.27_i386.deb [12:17] trejack, what problem? [12:17] kylem: trejack was also having the pata_amd issue [12:17] ok [12:17] no medium found.. [12:17] lazka: do you have a disk in there? [12:17] I can't. Wiped my working 2.6.20-17 install [12:18] yeah, i will look for a cleaner one.. [12:18] Wait a sec - I can chroot in from my edgy [12:18] is it the correct device, laska? ... could be scd1, too, if you got more [12:18] I'll be back [12:19] ok, /dev/cdrom is my dvd-drive [12:19] and its working [12:20] but i cant find the lite-on === heno goes to find more testers === trejack [n=Tre@24-119-246-56.cpe.cableone.net] has left #ubuntu-kernel [] [12:21] kylem, it makes some noise but doesnt spin up [12:21] [ 30.987575] scsi 1:0:1:0: CD-ROM PLEXTOR DVD-ROM PX-116A2 1.00 PQ: 0 ANSI: 5 [12:22] that's the only thing I see connected [12:22] it only found one ATAPI device [12:22] There is only /dev/sr0, no /dev/sr1 [12:22] and didn't even whine about the rest [12:22] are you sure it's cabled properly? === defendguin [n=supertux@ip70-180-86-93.no.no.cox.net] has joined #ubuntu-kernel [12:23] kylem: are you building an amd64 version for testing as well? [12:24] kylem, should be, i'll check with 20.13 again if it still works there, k? [12:24] heno: I am, I'll upload it when it's done [12:24] ok [12:24] ok, cool === ..[topic/#ubuntu-kernel:BenC] : Ubuntu kernel development discussion ONLY | Kernel Wiki: https://wiki.ubuntu.com/KernelTeam | 2.6.20-15.25 Uploaded | URGENT: Currently debugging pata_amd problems in 2.6.20-15.25 (do not report problems unless you have -15.25). Kernel that should fix it: http://people.ubuntu.com/~kyle/kernels/feisty/ | Kernel that is known to fix sata_nv and HPA problems: http://people.ubuntu.com/~bcollins/kernels/feisty-release/ | Please report success/fai === ..[topic/#ubuntu-kernel:BenC] : Ubuntu kernel development discussion ONLY | Kernel Wiki: https://wiki.ubuntu.com/KernelTeam | 2.6.20-15.25 Uploaded | URGENT: Currently debugging pata_amd problems in 2.6.20-15.25 (do not report problems unless you have -15.25). Kernel that should fix it: http://people.ubuntu.com/~kyle/kernels/feisty/ | Kernel that is known to fix sata_nv and HPA problems: http://people.ubuntu.com/~bcollins/kernels/feisty-release/ === lazka [n=lazka@85-125-223-239.dynamic.xdsl-line.inode.at] has joined #ubuntu-kernel === infinity2 [n=adconrad@cerberus.0c3.net] has joined #ubuntu-kernel [12:27] kylem, bad news, works ok with 20.13 [12:28] lazka: can you pastebin dmesg from .13? [12:29] http://pastebin.ubuntu-nl.org/15681/ [12:29] [ 32.828189] hdc: LITE-ON LTR-52246S, ATAPI CD/DVD-ROM drive [12:29] kylem: amd64 is uploading to my ~/, I'll let you put it next to yours [12:29] hdc?! [12:30] you must have shuffled pci ids? [12:30] umh what? [12:30] brb. [12:30] -13 -> -14 == ata update [12:30] probably moved from amd ide to pata_amd [12:30] pata_amd still used hd? in -13 for me aswell [12:31] or that :-) [12:31] brb === mxpxpod [n=bryan@unaffiliated/mxpxpod] has joined #ubuntu-kernel [12:34] it did say that in the bugreport... === trejack [n=Tre@24-119-246-56.cpe.cableone.net] has joined #ubuntu-kernel === infinity3 [n=adconrad@cerberus.0c3.net] has joined #ubuntu-kernel [12:45] kylem: if the problem we're having is just in pata_amd, maybe switching back to the amd IDE driver is best [12:45] looks like the quick fix didn't take care of all the problems [12:49] No good for me, either. (Iused the -15.27 kernel). The dmesg is at http://librarian.launchpad.net/7331329/dmesg2.txt [12:50] still showing exceptions for trejack too [12:50] Don't know if this will help or not, but when I booted into the live cd, it at first couldn't use any of my existing partitions. [12:51] I reformated the sda3 one using gparted, and it completed in seconds, and my sda1 partition was suddenly available, as well [12:52] trejack: doing a new build to revert back to amd74xx driver [12:52] lazka, moodydeath: you guys willing to test another kernel in about 30 minutes? [12:53] BenC: yep [12:53] I'll check back in 30. Thanks [12:53] yep [12:53] thanks, build started, will post here when it's done === trejack [n=Tre@24-119-246-56.cpe.cableone.net] has left #ubuntu-kernel [] === MrNOKIA [n=user@86.121.181.123] has joined #ubuntu-kernel === ..[topic/#ubuntu-kernel:BenC] : Ubuntu kernel development discussion ONLY | Kernel Wiki: https://wiki.ubuntu.com/KernelTeam | 2.6.20-15.25 Uploaded | URGENT: Currently debugging pata_amd problems in 2.6.20-15.25 (do not report problems unless you have -15.25). Test kernel is building, will be done at around 23:20 UTC | Kernel that is known to fix sata_nv and HPA problems: http://people.ubuntu.com/~bcollins/kernels/feisty-release/ [01:09] is this the new one ? [01:09] moodydeath: no [01:09] kk :) [01:09] ah ... should read the whole topic ... -_- [01:10] there is so much crap on TV at night :P omg === trex37 [n=tre@24-119-246-56.cpe.cableone.net] has joined #ubuntu-kernel [01:21] builds are done, getting ready to upload [01:22] hey BenC === trex37 [n=tre@24-119-246-56.cpe.cableone.net] has left #ubuntu-kernel [] === trex37 [n=tre@24-119-246-56.cpe.cableone.net] has joined #ubuntu-kernel [01:22] apparently the mmc card reader fix isn't working [01:23] it was reported to work for many people [01:23] hmmm [01:23] it doesn't resolve all mmc issues [01:23] i wish i had a card on me [01:23] it was just one major regression [01:24] 53b205c5fbe264a700454a4f5416d0a8 linux-image-2.6.20-15-generic_2.6.20-15.26_amd64.deb [01:24] 7a5f4a0801be29c9c614b3d1a673edf2 linux-image-2.6.20-15-generic_2.6.20-15.26_i386.deb === john_brown_jr [n=andris@balticom-130-178.balticom.lv] has joined #ubuntu-kernel === ..[topic/#ubuntu-kernel:BenC] : Ubuntu kernel development discussion ONLY | Kernel Wiki: https://wiki.ubuntu.com/KernelTeam | 2.6.20-15.25 Uploaded | URGENT: Currently debugging pata_amd problems in 2.6.20-15.25 (do not report problems unless you have -15.25). Test kernel: http://people.ubuntu.com/~bcollins/kernels/feisty-release/ | Kernel that is known to fix sata_nv and HPA problems: http://people.ubuntu.com/~bcollins/kernels/feisty-release/ [01:24] http://people.ubuntu.com/~bcollins/kernels/feisty-release/ [01:24] dl running :D [01:25] note that kernel changes pata_amd back to amd74xx, so your devices will go to hdX [01:25] moodydeath: it's still being uploaded [01:25] hmm [01:25] probably be completed in about 5 minutes [01:25] probably less [01:25] is it a 32 or a 64 bit version ? [01:25] i'll f5 ;) === john_brown_jr [n=andris@balticom-130-178.balticom.lv] has left #ubuntu-kernel [] [01:27] i'm sorry, i must be missing something [01:27] i'm using linux-image-2.6.20-15-generic_2.6.20-15.27_i386.deb [01:27] what's the file size? [01:28] the new file is .26 ? [01:28] 21 Mb [01:29] it's ready [01:29] yes,but... [01:29] why is it .26 now ? [01:30] i.m on .27 for about an hour [01:31] .27 was a version from kyle [01:31] kile ? [01:31] kyle ? === trex37 [n=tre@24-119-246-56.cpe.cableone.net] has left #ubuntu-kernel [] [01:33] BenC: kernel boots ... as fast as -12 did [01:33] testing drives [01:34] moodydeath, what is kyle ? should i try the new kernel now ? [01:34] scd0 & scd1 working :D === lazka [n=lazka@85-124-41-166.dynamic.xdsl-line.inode.at] has joined #ubuntu-kernel [01:35] http://people.ubuntu.com/~kyle/kernels/feisty/ <-- here is the .27 ... but i think the .26 which was released some minutes ago is the new one [01:36] Ubuntu 2.6.20-15.26-generic [01:36] http://pastebin.ubuntu-nl.org/15700/ [01:36] everything back to normal [01:37] http://pastebin.ubuntu-nl.org/15701 [01:37] me, too :D [01:38] BenC: is it intentional that ivtv isn't working with your builds? [01:38] it worked with kylem's build === trejack [n=tre@24-119-246-56.cpe.cableone.net] has joined #ubuntu-kernel [01:39] Yay! That one did it [01:39] what do you need from me? [01:44] lazka: no, I didn't change that [01:44] trejack: dmesg please [01:44] BenC: what do you mean? [01:44] lazka: ivtv [01:45] lazka: from your log earlier it looked like you were just missing the firmware [01:45] iirc [01:45] BenC: your build is 2 MB smaller.. [01:45] lazka: could be a left out the firmware [01:45] quirk in the build [01:46] thats what i mean.. [01:46] lazka: try copying it from /lib/firmware/linux-image-2.6.20-13-generic/ [01:46] it wont happen in the final build [01:46] ok, thanks [01:46] I trimmed some things out of the build to speed things up [01:46] mainly linux-headers, linux-debug, and apparently firmware [01:47] ah, i see, however.. i'm going to bed now, already 2 am here. [01:47] lazka: thanks [01:47] gn8 all [01:47] Mithrandir, cjwatson: The three testers we had that reported the problem have reported success [01:47] the fix was reverting back to amd74xx IDE driver instead of pata_amd [01:47] right, do you have the diff or is it in git? [01:48] so it's a relatively safe and isolated fix [01:48] amd74xx was actually being used up until about 5 weeks ago [01:48] I'm actually just going to bed, but I am happy to be woken up once it's uploaded [01:48] will be done in about 10 minutes [01:48] ok [01:48] send me an SMS on my mobile number? [01:49] cjwatson: Ok [01:49] cjwatson: wait, I can't send Intl SMS [01:50] unless you have an email that redirects [01:51] trejack, lazka, moodydeath: thanks for the testing! [01:51] np :) [01:51] np ? [01:51] the past few days, we've been very fortunate to have folks coming around to bear the burden of our test builds [01:51] MrNOKIA: np == no problem [01:52] MrNOKIA: were you having the pata_amd problems? [01:52] will there be something like a final version, that will need a test ? [01:52] sorry [01:52] i knew that [01:52] moodydeath: by tomorrow there should be something in your update manager :) [01:52] I guess [01:52] MrNOKIA: could you test the kernel in the topic URL? [01:53] I'm sure it will work for you, if that's in fact the issue you were having [01:53] I could'nt boot the new kernel untill the .25 version came up [01:53] http://librarian.launchpad.net/7331974/dmesg3.txt [01:53] i posted info on the dedicated thread [01:53] hope it helped [01:54] my nick is pretty much the same as on ubuntuforums :) [01:54] MrNOKIA: ok, thanks [01:54] trejack: excellent, thanks [01:54] cjwatson, Mithrandir: kernel away [01:54] Thank you! [01:55] I must say I was intimidated by all of this at first, but it's been a pleasure... [01:55] BenC: a new kernel ? [01:56] i see only the 15 april .26 version === trejack [n=tre@24-119-246-56.cpe.cableone.net] has left #ubuntu-kernel [] [02:08] got it all fixed? === gnomefreak [n=gnomefre@ubuntu/member/gnomefreak] has joined #ubuntu-kernel [02:23] defendguin: looks that way [02:23] MrNOKIA: the one in my dir in the URL is the same as the one I uploaded (sans some firmware) === reitblatt [n=mark@resnet-50-180.dorm.utexas.edu] has joined #ubuntu-kernel === philwyett [n=philip@bb-87-81-146-45.ukonline.co.uk] has joined #ubuntu-kernel === _4strO [n=cedric@AStrasbourg-253-1-70-31.w90-26.abo.wanadoo.fr] has joined #ubuntu-kernel === johanbr [n=j@blk-224-156-151.eastlink.ca] has joined #ubuntu-kernel === Eruantalon [n=hans@5634185c.rev.stofanet.dk] has joined #ubuntu-kernel === gnufied [n=gnufied@59.144.3.126] has joined #ubuntu-kernel [04:17] extended attributes are enabled on stock ubuntu dapper kernels? [04:17] if yes, then how do i find out? [04:19] grep -i xattr /boot/config-$(uname -r) [04:19] thanks, so its enabled. === gnufied [n=gnufied@59.144.3.126] has left #ubuntu-kernel ["Leaving"] === yoasif [n=yoasif@pool-71-167-74-163.nycmny.east.verizon.net] has joined #ubuntu-kernel [04:54] i just saw the post on ubuntuforums about the sata_nv issue... how can i help? [04:54] Should be fixed now [04:55] See the topic [04:55] ty [04:56] hmm, 15-27 is listed on ubuntuforums, should i get that, or the one in the topic? [04:56] (i am experiencing the PATA issues) [04:57] 15.27 is the latest, I believe [04:57] It's worth a go [04:57] ok [05:00] mjg59: do you have any ideas what the problem might be if every once in a while, the screen goes all wonky (the screen gets compressed down, and a load of copies of the screen get patterned, and skewed) [05:00] ssh doesn't work, and i've tried just using vesa [05:00] No [05:01] heh... i'm really wondering how i can describe the issue properly. [05:18] sounds a little like your graphics card aesploded [05:18] it's onboard [05:18] and a pretty new mobo === defendguin [n=supertux@ip70-180-86-93.no.no.cox.net] has joined #ubuntu-kernel === tuxmaniac [n=aanjhan@unaffiliated/tuxmaniac] has joined #ubuntu-kernel === anibal [n=anibal@debian/developer/anibal] has joined #ubuntu-kernel === keturn [n=kevint@pdpc/supporter/sustaining/keturn] has left #ubuntu-kernel ["Client] === tuxmaniac [n=aanjhan@unaffiliated/tuxmaniac] has joined #ubuntu-kernel === tuxmaniac [n=aanjhan@unaffiliated/tuxmaniac] has joined #ubuntu-kernel [06:29] -15.27 kernel belatedly accepted - builds will start running as soon as I can manage it [06:30] thanks colin [06:32] http://librarian.launchpad.net/7331974/dmesg3.txt (from trejack earlier) has some DriveReady SeekComplete Error stuff before the kernel finally gets hda going - is that unrelated? [06:32] unlikely [06:33] oh, yeah [06:33] Likely that it's unrealted, by the looks of it [06:33] i thought hda was the missing cdrom [06:33] BadCRC is either complete timing fuckup, or bad cable [06:33] likely a 40pin cable [06:34] But, hell, drivers/ide has never really worked [06:34] hda has a partition table in that dmesg so it's not a cdrom [06:34] right [06:34] It ends up coming up in pio mode [06:34] works by magic [06:34] sheer force of will [06:34] Not ideal, but it'll do [06:34] ha [06:34] hmm, mark lord lives in the same city as me [06:35] i can walk over and beat him up if you want [06:35] :P [06:35] ok, publisher's running, I'm off back to sleep [06:35] Oh, that system looks shitted anyway [06:35] cjwatson, night [06:35] [ 20.079279] ..MP-BIOS bug: 8254 timer not connected to IO-APIC [06:35] [ 20.079317] ...trying to set up timer (IRQ0) through the 8259A ... failed. [06:35] [ 20.079320] ...trying to set up timer as Virtual Wire IRQ... failed. [06:35] [ 20.119013] ...trying to set up timer as ExtINT IRQ... works. [06:35] Never seen that on an nvidia board before [06:36] phone me if it all blows up, and I'll try not to fall straight back to sleep again :-/ [06:36] Mithrandir has it set up so that the kernel builds on palmer this time [06:37] cjwatson: The errors there don't look related to any of the HPA/libata stuff [06:38] ok, thanks === bdgraue [n=bdgraue@dynadsl-080-228-82-061.ewetel.net] has joined #ubuntu-kernel === bdgraue [n=bdgraue@dynadsl-080-228-82-061.ewetel.net] has joined #ubuntu-kernel [07:24] cjwatson: linux-image-2.6.20-15-generic_2.6.20-15.26_i386.deb works for me === ^^CatTuX^^ [n=^^CatTuX@137.101.129.63] has joined #UBUNTU-KERNEL === tritium_ [n=tritium@ubuntu/member/tritium] has joined #ubuntu-kernel === Lure [n=lure@clj46-234.dial-up.arnes.si] has joined #ubuntu-kernel [08:40] kylem: my server seems to have kept itself from hanging after I removed irqpoll :-). === \sh [n=shermann@server3.servereyes.de] has joined #ubuntu-kernel [09:17] BenC: I've emailed you a tested and verified patch for #105582. Please consider applying it to avoid a regression from dapper. [09:19] crimsun: too late, but it should go into an SRU [09:20] Mithrandir: whatever you guys [have] decide[d] [09:21] crimsun: sorry. :-/ === tuxmaniac [n=aanjhan@unaffiliated/tuxmaniac] has joined #ubuntu-kernel === benh [n=benh@ozlabs.org] has joined #ubuntu-kernel [09:39] crimsun: here ? :-) [09:39] yep. [09:39] ok, so it's indeed Version: 2.6.20-15.25 [09:40] and it blows up when moving the mouse [09:40] on a Quad G5 [09:40] linux-image-2.6.20-15-powerpc64-smp, in console mode, minimum install (no X, no gpm) [09:40] I don't have a hardcopy of the crash yet, it crashes at irq time, so no log [09:40] backtrace shows that it comes from usb -> input -> evdev [09:41] and it might be blowing up in do_gettimeofday() which would be VERY strange [09:41] are there some known ubuntu patches in that area ? [09:43] crimsun: yup, I think BenC is asleep :-( pung him already on some other channel, didn't help === `sam` [n=sam@CPE-75-81-234-51.we.res.rr.com] has joined #ubuntu-kernel [09:44] I'm a bit lost with launchpad, how do I look at the bugs reported for a given package ? [09:44] https://launchpad.net/ubuntu/+source/linux-source-2.6.20 seems to indicate that a .27 is out [09:44] though it hasn't hit au.archive.ubuntu.com [09:45] might be worth for me trying to dig that and check before filing a bug report [09:45] https://launchpad.net/ubuntu/+source/linux-source-2.6.20/+bugs [09:46] hrm.. I don't see it [09:46] me gets the source and have a look how on earch do_gettimeofday() could crash ... [09:48] -15.27 probably won't hit au.archive for another couple of hours, but you can get it off archive.ubuntu.com or launchpad. [09:49] yup, I'll try that too [09:50] https://launchpad.net/+builds/+build/318987/linux-image-2.6.20-15-powerpc64-smp [09:51] not on archive's Package file yet, I'll wait and look at hte source in the meantime [09:51] ok [09:51] that would do too I suppose :-0 [09:51] I really can't believe we wouldn't have massive amounts of bug reports about that bug, unless there's something special with your hardware. [09:52] nothing special that I can see... just a fairly ordinary quad G5 [09:52] <`sam`> does -25.27 need testing for sata_nv or is it known to work? because i'm installing now... [09:52] anyway, have to run, will be back later & investigate [09:53] have 2 disks though and using an lvm but heh... that shouldn't matter :-) [09:53] 15-26 is running fine with sata_nv [09:53] -15.26 i mean [09:54] `sam`: we want as much testing as we can get on the latest kernel (-15.27). Even if previous ones worked fine, we want to make sure there are no regressions. [09:54] `sam`: so if you have a bit of time to spare, it's much appreciated. [09:54] Mithrandir: i'll try the new -15.27 too [09:55] <`sam`> ok, i've had -15.25 installed but have still been running -14.23... going to reboot now and see === mcarroll [n=mc@S010600134642ba3b.vc.shawcable.net] has joined #ubuntu-kernel === `sam` [n=sam@CPE-75-81-234-51.we.res.rr.com] has joined #ubuntu-kernel [10:03] <`sam`> well -15.27 boots for me [10:03] <`sam`> took me a few minutes because i had changed my xorg.conf about 30 minutes ago and had messed it up [10:04] `sam`: good to hear, CD-ROM/DVD-ROM and all works too? [10:06] <`sam`> yeah i just tested a dvd and it started playing in totem [10:07] great [10:07] please tell us if you have other (kernel-related) regressions? [10:10] Mithrandir: where can i get the 15-27? [10:10] <`sam`> like what? anything specific you want me to check out? === mcarroll [n=mc@S010600134642ba3b.vc.shawcable.net] has left #ubuntu-kernel [] [10:12] are the 2.6.20-15.27 not available for i386 yet? [10:13] <`sam`> bdgraue, have you checked in update-manager? i just found it there about 25 minutes ago [10:14] `sam`: not here i have archive.ubuntu.com [10:14] bdgraue: are you on i386? If so, the kernel is not done yet. [10:15] ok then [10:15] i'll wait :-) [10:16] palmer is cooking hot ;) [10:17] Status: Currently building [10:17] :-) === heno [n=henrik@ubuntu/member/heno] has joined #ubuntu-kernel [10:17] <`sam`> Mithrandir, was the drive detection the main problem? if there's anything else i could test i've got time... even though i don't know what all i'm looking for :) [10:21] `sam`: yes, the problems were all related to drive detection and also some HPA. [10:22] can you run dmesg | grep hpa in a shell and paste the contents somewhere? [10:23] <`sam`> it's only 2 lines, can i paste it here? [10:24] sure [10:24] <`sam`> [ 24.478810] ata1.00: ata_hpa_resize 1: sectors = 312500000, hpa_sectors = 312500000 [10:24] <`sam`> [ 24.490762] ata1.00: ata_hpa_resize 1: sectors = 312500000, hpa_sectors = 312500000 [10:25] ok, and no scary errors of any kind in dmesg? [10:27] <`sam`> i don't think so, but i don't know what this is: [10:27] <`sam`> [ 25.958194] ATA: abnormal status 0x7F on port 0x0000000000010967 [10:29] hm [10:30] not great, but if it all works, I wouldn't worry too much [10:37] <`sam`> what is the hpa exactly? i'm just finding stuff about thinkpads... [10:37] Mithrandir: crash still happens with .27 [10:37] Mithrandir: will investigate [10:39] Mithrandir: do you guys have extra patches on top of mainline like some of the -rt or clock sources stuff ? some of that is known to introduce fancy breakage on ppc [10:40] (still d/l the source on my crappy link here) [10:40] `sam`: http://en.wikipedia.org/wiki/Host_Protected_Area [10:40] benh: we have lowlatency, but it's not enabled in the default kernels. [10:40] k [10:41] our kernel's in git if that helps [10:41] benh: I'm not a kernel hacker, so I'm not sure what other patches have been brought in. [10:41] ok, because the scary thing is ... there is -no way- such an error could happen in gettimeofday on normal mainline [10:42] unless there have been kernel memory corruption :-( [10:42] http://git.kernel.org/?p=linux/kernel/git/bcollins/ubuntu-feisty.git;a=summary [10:43] ... or it's passed a crap argument [10:43] which is probably the likely cause... need to disasm to see where exactly the crash occurs === john_brown_jr [n=andris@balticom-130-178.balticom.lv] has joined #ubuntu-kernel [10:43] Mithrandir: -15.27/i386 is in accepted, if you hadn't already noticed; feel like accepting d-i? [10:43] I wounder if stupid mousemu could be causing it [10:48] cjwatson: ah, just got there then, I'll accept it [10:49] Mithrandir: ok, if you kill mouseemu, it doesn't crash [10:49] is mouseemu kernel mode? [10:49] Mithrandir: at this point, I'm about 99% convinced it's some crap in uinput vs. 32 bits apps on 64 bits kernels [10:50] Mithrandir: no, but it uses some whacky kernel interfaces (uinput) which I wouldn't be surprised at all is broken === ivoks [n=ivoks@5-51.dsl.iskon.hr] has joined #ubuntu-kernel [10:50] Mithrandir: there is a long tradition of brokenness in the input layer userland interfaces, especially vs 32/64 bits [10:50] hmkay. [10:50] Mithrandir: it's likely that running a 32 bits mouseemu on a 64 bits x86 will crash too [10:51] Mithrandir: I'll have to investigate in more details but can't do that tonight, in the meantime, it might be worth (if possible) to disable mouseemu by default on powerpc64 [10:51] we don't support 64 bit kernels and 32 bit userland on x86_64, so that's not really that important. [10:51] Mithrandir: hehe, fair enough :-) [10:51] "you run this and it breaks? You get to keep both pieces." [10:52] Mithrandir: well, I know powerpc isn't officially supported anymore.. but if you guys want to avoid that problem, the short term band-aid is to disable mouseemu on 64 bits powerpc's [10:52] cjwatson: ^^ ; your code, your call, I'm not sure what's a reasonable direction to go in from here. [10:52] Mithrandir: want me to file a launchpad entry ? [10:52] benh: yes, a bug would be good. [10:52] Mithrandir: I'll try to debug that properly later, I'll be away for most of next week though [10:52] benh: yes, I'm trying to think of a way to not end up saying "sorry, ppc is SOL". [10:52] Mithrandir: :-) [10:53] Mithrandir: well... it would hit ps3 [10:53] it would, which is bad. [10:53] I'm wondering if we can do it in an SRU or not.. [10:53] I'd say just don't install mousemu by default on 64 bits ppc's ... I suppose you install it by default bcs of 1 buttons apple mice... [10:54] anyway, I'll file a launchpad entry tonight so it's at least tracked before I go [10:54] I'll try to find out what's wrong in the kernel later === MrNOKIA Offtopic: update available for Ubuntu Forums Firefox extension /Offtopic === `sam` [n=sam@CPE-75-81-234-51.we.res.rr.com] has left #ubuntu-kernel ["Leaving"] [10:56] Mithrandir: wah, need to do that before d-i goes in [10:56] Mithrandir: can you reject it temporarily? [10:58] benh: one-button mice> right, the old sysctl approach isn't available on x86 for Intel Macs and I wanted to have the same scheme across architectures [10:58] it was working fine for me on 32-bit powerpc :-/ [10:58] cjwatson: faiir enough [10:58] cjwatson: yeah, well, uinput is a pile of poo unfortunately [10:58] cjwatson: I'll check if the problem is still in upstream 2.6.21-whatever-git-of-the-day first [10:58] Mithrandir: I've rejected debian-installer so that I can squeeze hw-detect in front [10:59] benh: the powerpc64 option sounds reasonable [10:59] another option might be to build it as a 64 bits binary :-) if the bug is what I think, that might work [10:59] can't do that at this point [10:59] it's the same binary package on powerpc32 and powerpc64, would involve massive fiddling [10:59] also we're right at the wire [11:00] literally if this had been half an hour later it would have been too late ;) [11:00] heh [11:00] I knew it was a good idea to try out fesity on this g5 :-) [11:01] meh, this means I have to figure out whether the running system is ppc64 [11:01] I've been running it on the laptop for some time, had minor issues, but overall good [11:01] arch_get_kernel_flavour () { [11:01] CPU=`grep '^cpu[[:space:] ] *:' "$CPUINFO" | head -n1 | cut -d: -f2 | sed 's/^ *//; s/[, ] .*//' | tr A-Z a-z` [11:01] case "$CPU" in [11:01] cjwatson: ugly way: check /proc/ppc64 :-) [11:01] power3|i-star|s-star|power4|power4+|ppc970*|power5|power5+|cell) [11:01] family=powerpc64 [11:01] nah [11:01] will have to duplicate crap like that I guess [11:01] don't test the cpu model [11:01] benh: huh, that actually exists? [11:01] that's too ugly for words [11:01] I know, but it's all we had [11:01] cjwatson: it does in 2.6.20 :-0 [11:01] neat [11:02] let me dbl check it does on your kernel... [11:02] the code's there, but I'm no expert [11:03] it does [11:03] either that or you check for uname -a containing powerpc64 [11:03] I'm sure there are better ways :-) [11:03] but you should never have to check for individual cpu models [11:03] anyway, anything that works for you... [11:04] might change that at some point, cpuinfo testing is useful though because of the automatic test framework there [11:04] could probably teach it about uname output if it doesn't know already [11:04] anyway, I'll [ ! -d /proc/ppc64 ] in hw-detect, thanks [11:06] yah, it's dir with various crap in it [11:06] we can't really remove it because of some stupid stuff relying on it [11:07] so it will stay a bit longer... [11:07] anyway, use whatever is best for you [11:07] btw, I was impressed how good the installer was vs. setting up yaboot [11:07] I was installing on sdb, with lvm & all, sda had my debian stuff etc.. [11:07] and the installer figured it all out, and setup yaboot.conf entries for my sda stuff too, neat ! [11:08] Mithrandir: could you review hw-detect 1.45ubuntu5 to work around benh's bug? [11:08] then binaries of that need to be accepted before re-accepting debian-installer [11:08] not in the queue yet, but yes, will do in two minutes [11:08] benh: yeah, the os-prober stuff for detecting other systems is nice [11:09] still keep getting the odd report of ofpath glitches [11:09] yeah, for some reason, I didn't hit those tho [11:09] (I was expecting to :-) [11:09] I think it re-used my boot partition on sda which is probably why it worked [11:10] we do have a couple of ofpath patches, think I sent them to Debian a while back [11:10] not entirely sure they're right on all systems though [11:11] I'm filing a launchpad entry for the kernel bug so we can track it [11:11] had a report the other day which suggests not, though I haven't seen the /proc/device-tree/ tarball for it yet [11:12] thanks. so you reckon it's lack of 64/32-bit thunking in the uinput ioctls? [11:13] it's either that or a bug in the thunking... either way, it shouldn't crash the kernel :-) [11:13] are those ioctls available to the user? If so, we should get this fixed in a security upload. [11:13] Mithrandir: yes, mouseemu is userspace [11:13] oh to non-root you mean? [11:13] well, non-root user. "root can crash the kernel" isn't very interesting. [11:14] I suspect it might use read/write instead of ioctl's ... but I'll confirm all that when I have time to dig, after I'm back from Perth next week [11:14] probably to anyone who can read/write /dev/input/* [11:14] depends on the permission on uinput [11:14] benh: ioctl to set up the virtual device, read/write to interact with it [11:14] uinput is a hack that allows to filter input events from userspace [11:14] crw-rw---- 1 root root 10, 223 2007-04-12 10:28 uinput [11:14] yeah [11:14] at the moment [11:14] ok, so it's not -too- bad... still worth fixing at one point [11:15] sure, it's worth fixing, but not getting into (another) panic about. [11:15] either a small fix if easy, or just forbid read/write on 32 bits if not a simple fix [11:15] sure [11:15] https://bugs.launchpad.net/ubuntu/+source/linux-source-2.6.20/+bug/106723 [11:16] thanks [11:17] the input layer 32/64 its thunking is known to be a total PITA [11:17] anyway, I'll dig when I can, not enabling mouseemu by default on powerpc64 is a good enough workaround for now as far as I'm concerned === john_brown_jr [n=andris@balticom-130-178.balticom.lv] has left #ubuntu-kernel [] [11:17]