[02:02] <fabbione> BenC: ping?
[02:02] <BenC> fabbione: yo
[02:03] <fabbione> BenC: hey man.. i think something on i386 kernels is broken badly
[02:03] <BenC> fabbione: Can you describe?
[02:03] <fabbione> trace -o /home/fabbione/foo.strace /bin/ls
[02:03] <fabbione> umovestr: Input/output error
[02:03] <fabbione> bin    dev   initrd      lib64       mirrors  proc  sbin  tmp  vmlinuz
[02:03] <fabbione> boot   etc   initrd.img  lost+found  mnt      raid  srv   usr
[02:03] <fabbione> cdrom  home  lib         media       opt      root  sys   var
[02:04] <fabbione> make that a strace
[02:04] <fabbione> i get a lot of errors (also confirmed by different people on #u-devel) like that umovestr
[02:04] <fabbione> it seems to happen only i386
[02:04] <BenC> is this the -12 kernel?
[02:05] <fabbione> if you strace something bigger you might see also other errors like ptrace: Input/Output error
[02:05] <fabbione> confirmed to be there from at least .20-9
[02:05] <fabbione> but yes. i am specifically running .20-12
[02:05] <BenC> hmm
[02:11] <BenC> cool, I have -6 and -7 on this box, let me check those
[02:12] <fabbione> ok
[02:19] <BenC> fabbione: Same on -7
[02:19] <fabbione> ok
[02:29] <Keybuk> BenC: random question about the kernel, if you have time to answer it
[02:30] <BenC> Keybuk: Sure
[02:30] <Keybuk> disk write cache; does that get flushed on shutdown?
[02:30] <Keybuk> specifically for IDE
[02:30] <mjg59> Yes
[02:31] <Keybuk> can you cite a reference in the source; I couldn't find out for definite where it happened
[02:31] <mjg59> Unless you're seeing massive filesystem corruption, in which case the answer is probably no (but shouldb be yes)
[02:31] <Keybuk> I can see it happens when you reboot, but not halt or poweroff
[02:31] <Mithrandir> hm, maybe that's the reason why I've seen some ext3 corruption in the past.
[02:32] <mjg59> Keybuk: Hm. In the poweroff case, I suspect that it's supposed to be done when the drive's powered down
[02:32] <Keybuk> but I can't see anything that even powers down the drive :p
[02:33] <mjg59> Keybuk: Just to clarify - do you mean drivers/ide or libata?
[02:33] <BenC> fabbione: Same as far back as -6
[02:33] <Keybuk> drivers/ide
[02:34] <Keybuk> reason for asking:
[02:34] <Keybuk> the sysvinit halt binary deliberately standbys IDE disks by iterating /proc/ide
[02:34] <Keybuk> and the manpage claims this is because the Linux kernel doesn't
[02:34] <mjg59> Hm. It's conceivable.
[02:34] <Keybuk> this seems somewhat silly to me; and causes problems when you try and run halt on things like <hardware from Sony>s which don't have an IDE controller
[02:34] <mjg59> (Go go stupid workarounds)
[02:35] <Keybuk> I could believe that the manpage is way out of date
[02:35] <Keybuk> or if not, and it's a simple patch, it'd seem sweeter to fix the kernel?
[02:35] <mjg59> Yeah
[02:50] <ivoks> mjg59: do we have any policy for 
[02:50] <ivoks> mjg59: doh.. STOP_SERVICES in /etc/default/acpi-support
[02:50] <mjg59> Yes. Don't add things to it.
[02:51] <ivoks> ok :)
[02:51] <kylem> Keybuk, are you complaining about what? sata/scsi?
[02:51] <mjg59> kylem: drivers/ide
[02:52] <kylem> ah.
[02:52] <kylem> well. turn off your write cache. it's a bad idea anyway.
[02:52] <Keybuk> kylem: is it off by default? :p
[02:52] <kylem> it should be on most disks.
[02:53] <kylem> unless hdd folks have gotten really ballsy in the last couple years.
[02:54] <kylem> SCSI device sda: write cache: enabled, read cache: enabled, doesn't support DPO or FUA
[02:54] <kylem> huh. poop.
[02:54] <Keybuk>        The  -h  flag  puts  all  harddisks in standby mode just before halt or
[02:54] <Keybuk>        poweroff. Right now this is only implemented for  IDE  drives.  A  side
[02:54] <Keybuk>        effect  of putting the drive in standby mode is that the write cache on
[02:54] <Keybuk>        the disk is flushed. This is important for IDE drives, since the kernel
[02:54] <Keybuk>        doesnt flush the write-cache itself before poweroff.
[02:54] <Keybuk> (is the sysvinit halt manpage reference)
[02:54] <kylem> anyway, if you're using libata-pata, the scsi layer will send a SYNC CACHES command which should end up in libata and be issues properly to the drive.
[02:55] <kylem> Keybuk, interesting.
[04:22] <seb128> hi
[04:23] <seb128> I'm trying to rebuild a feisty kernel with the 8390.c from edgy following the instructions on https://wiki.ubuntu.com/KernelCustomBuild
[04:23] <seb128> I've apt-get install linux-source, copied the tar.bz2 somewhere, untar-ed it, copied the .c and now I'm running "AUTOBUILD=1 fakeroot debian/rules binary-debs" which doesn't work
[04:24] <seb128> "make: *** No rule to make target `binary-debs'.  Stop."
[04:24] <seb128> any idea of what I'm doing wrong? ;)
[04:31] <abogani> seb128: That procedure works only with original git sources....
[04:32] <seb128> abogani: the wiki page is not clear then ;)
[04:33] <abogani> seb128: exactly :-)
[04:33] <seb128> ok, using make-kpkg then ;)
[04:33] <seb128> thank you
[04:33] <abogani> seb128: Nothing
[04:35] <Simira> does anyone know the bug number on "fan doesn't work after resume"? It's a kernel bug with hp laptops, that's reappeared. 
[04:38] <BenC> Simira: It should be fixed in 2.6.20-12 kernels
[04:43] <Simira> BenC: I just tested on daily
[04:43] <Simira> BenC: and it doesn't work for me
[04:43] <BenC> Does it have -12 kernel?
[04:43] <Simira> yup
[04:44] <BenC> I included all the mentioned patch sets to fix that, so if something else is needed, someone with that hw will have to work out what it is
[04:44] <BenC> You've got about 3 weeks to do it :)
[04:45] <Simira> I'll test again. We're two testing this spesific model, so we should manage to work it out
[04:46] <BenC> Simira: If you need to find the bug, do an advanced search on linux-source-2.6.20 and include "Fix Released" bugs
[04:46] <Simira> BenC: I'll do that. Thanks
[04:49] <seb128> anything else I can try about bug #87078? 2.6.20 doesn't build with the 8390.c from 2.6.17
[04:49] <seb128> and the "irqpoll" option makes the box freeze during desktop login
[05:06] <abogani> seb128: If you use apt-get source linux-source-2.6.20 you can use "AUTOBUILD.... etc etc" as reported in the wiki page
[05:07] <seb128> abogani: ok, thank you
[07:05] <BenC> fabbione: FYI, I get the strace problem all the way back to -6
[08:01] <BenC> mjg59: ping
[08:03] <mjg59> BenC: Hi
[08:04] <BenC> mjg59: Hey, I want to enable the tifm_ms+memstick driver you sent me, but it's missing linux/memstick.h
[08:04] <mjg59> BenC: Ok. I'll try to sort that.
[08:04] <BenC> mjg59: Thanks
[08:05] <fabbione> BenC: ok
[08:20] <Lure> Simira: I am currently building kernel with additional patches that may fix fan & resume issues on hp laptops
[08:20] <Lure> Simira: will document in bug what I find
[08:21] <Lure> cd
[08:21] <orangey> hey all!
[08:21] <Lure> Simira: related bugs are 63123, 75398 & 74877
[08:21] <orangey> Ben Collins has accepted a patch into the kernel that makes ide-acpi.c, however, it is not enabled in the kernel binaries
[08:22] <orangey> Lure: I'm the dude who submitted the unified patch..
[08:22] <orangey> Lure: I saw your comments.
[08:22] <Lure> orangey: yep, one patch from your diff is missing, I have traced most fixes back in upstream git
[08:22] <orangey> re config for kernel: what config do I play with to patch? I want to enable IDEACPI
[08:22] <orangey> Lure: awesome.
[08:23] <orangey> Lure: you're talking about 37?
[08:23] <BenC> orangey: In feisty, it's enabled
[08:23] <BenC> at least it should be
[08:24] <Lure> BenC: is it possible to build generic without lowlatency (building with flavour=generic)?
[08:24] <BenC> $ rgrep BLK_DEV_IDEACPI debian/config/debian/config/ia64/config:CONFIG_BLK_DEV_IDEACPI=y
[08:24] <BenC> debian/config/i386/config:CONFIG_BLK_DEV_IDEACPI=y
[08:24] <BenC> debian/config/amd64/config:CONFIG_BLK_DEV_IDEACPI=y
[08:24] <Lure> BenC: it is enabled
[08:24] <BenC> in dact, it is enabled
[08:24] <BenC> fact
[08:24] <BenC> Lure: rm -f debian/config/*/lowlatency
[08:24] <orangey> BenC: ah. ok.
[08:24] <Lure> BenC: thanks - will shorten my build by half ;-)
[08:25] <orangey> BenC: it seems I don't have the most recent source, then. one moment.
[08:25] <BenC> orangey: grep IDEACPI /boot/config-`uname -r`
[08:26] <orangey> BenC: yep. there!
[08:26] <BenC> orangey: Then you have nothing to worry about :)
[08:26] <orangey> BenC: my bad. In that case, there is some other reason I can't figure out for why my system will no longer resume with this kernel.
[08:27] <BenC> orangey: Please file a bug report on linux-source-2.6.20 with as much information as possible
[08:27] <orangey> BenC: will do as soon as I have info : )
[08:27] <BenC> like what kernel version suspend/resume worked with, what type of resume (mem/disk), etc.
[08:35] <Lure> BenC: is there any way to override ABI check or do I need to bump ABI number on my local git clone?
[08:35] <orangey> hmm..
[08:36] <orangey> It seems my conflicting problems have hit..
[08:36] <BenC> Lure: debian/rules .... skipabi=true
[08:36] <BenC> Lure: See topic, all this info is under our wiki
[08:37] <Lure> BenC: thanks for help and sorry for bothering you
[08:37] <BenC> Lure: No problem at all, just figured there may be other things in the wiki that could help you
[08:38] <Lure> BenC: it helped a lot (actually I was lurking when you had online session), it is just a lot to digest ;-)
[08:39] <BenC> Lure: The stuff in KernelTeam/KnowledgeBase is probably most interesting for what you're doing
[08:39] <BenC> custom kernel build bits
[08:39] <orangey> alright.. time to reboot and retest. I think I have the issue figured out.
[08:39] <orangey> Thank you Ben.
[08:39] <orangey> Amazing work, btw.
[08:40] <orangey> It's my first interaction at all with kernel stuff or patching, and you have made it pleasant via your patience.
[08:40] <orangey> thank you.
[08:40] <orangey> see y'all : )
[08:40] <BenC> orangey: Np
[08:43] <BenC> crimsun: ping
[09:05] <Nafallo> nice
[09:16] <orangey> hey all.
[09:16] <orangey> back.
[09:18] <orangey> Alright. I think I have a good idea of the problem. For whatever reason, it is ata_piix that stops my NC4200 from booting. It was noted that it had to be disabled for suspend / resume to work, however that leaves me victim of the ALERT! no /dev/by-uuid-whatever problem, where I can't boot..
[09:18] <orangey> now..
[09:18] <orangey> any ideas how I could try to track down the problem more than that?
[09:19] <orangey> or must this work be done via bug tracking?
[09:23] <orangey_> bah.. Just got disconnected, so repasting the message:
[09:23] <orangey_> Alright. I think I have a good idea of the problem. For whatever reason, it is ata_piix that stops my NC4200 from booting. It was noted that it had to be disabled for suspend / resume to work, however that leaves me victim of the ALERT! no /dev/by-uuid-whatever problem, where I can't boot.. now.. any ideas how I could try to track down the problem more than that? or must this work be done via bug tracking?
[09:23] <orangey_> rather, ata_piix stops the NC4200 from Resuming after suspend.
[09:46] <crimsun> BenC: pong
[09:47] <BenC> crimsun: Hey, just FYI, I'm assigning all sound/audio related bug to the ubuntu-audio group
[09:47] <BenC> crimsun: In case you go wondering around for audio bugs filed on linux-source-2.6.20 :)
[09:47] <crimsun> BenC: yep, according to the triaging practices
[09:48] <BenC> crimsun: Ah, I forgot that I told you this already
[09:48] <BenC> I'm burned out from bugs today, brain is failing me right now
[09:49] <crimsun> heh, I hear ya
[10:11] <Simira> Lure, BenC: we are two people who have tested the daily(-12) kernel twice, and the fans don't work. :(
[10:11] <Lure> Simira: right, I am testing some more patches from upstream - first try did not help :-(
[10:11] <BenC> Simira: If you can build kernels, then perhaps start looking into the patches geared that the type of laptop
[10:12] <Simira> BenC: not really, but I might be able to get some competent guidance 
[10:12] <Mithrandir> BenC: I could try git-bisecting on it.
[10:13] <Mithrandir> (Simira is my wife)
[10:28] <BenC> Mithrandir: I'm not sure there's a case of it ever working to bisect against, other than edgy...or do you have a known working feisty kernel?
[10:29] <Mithrandir> BenC: -9 apparently worked.
[10:29] <BenC> Ok, then a bisect would be greatly appreciated
[11:11] <Whoopie> I have problems with running latest kernel in a vmware with the virtual cd-rom. As I can't see any obvious messages in dmesg, I disabled the cd-rom and now it works. Seems to be related to ata driver. Did anybody see the same behaviour?
[11:12] <crimsun> there are certain known regressions from -10.17 and -11.18
[11:12] <crimsun> without additional information, it's all hand-waving
[11:14] <Whoopie> crimsun: but it worked with -11
[11:15] <Whoopie> crimsun: I know that one need more information. i just thought I give it a try. ;)
[11:22] <kylem> Benbot4000
[11:23] <ivoks> crimsun: if you need any help with muting not working, i'm here :)
[11:24] <crimsun> ivoks: I'm working through an alsa issue atm
[11:25] <ivoks> that's why i say; i'm here to help as much as i can
[11:25] <crimsun> I'm a poor multitasker; patience, please.
[11:26] <ivoks> :)