/srv/irclogs.ubuntu.com/2010/03/26/#ubuntu-kernel.txt

Sarvattis it a known issue that theres no linux-backports-modules for 2.6.32-17 in the archives yet?02:41
=== kamalm is now known as kamalm-away
MTecknologySarvatt: it's in the pocket02:59
MTecknologyso.. it shoulda been there already02:59
JohnFluxwoohoo!  Suspend works now on my machine :-)04:25
JohnFlux It didn't work in 9.10  but does in 10.0404:25
JohnFluxI love you guys! :)04:25
JohnFluxWhy do drivers released by companies have such poor quality?04:43
johanbrif you're talking about closed-source drivers, it seems like the authors often have a less than perfect understanding of how to interface with the kernel properly04:44
JohnFluxeven GPL'ed drivers04:44
JohnFluxThe released ralink driver doesn't compile because they removed MODULE_LICENSE("GPL") at the last moment04:44
JohnFluxit contains a serious flaw that lets root the system remotely04:45
johanbroh that... yes, ralink is a mess04:45
JohnFluxIt spams the syslog with the message "#"  once for every packet!!04:45
JohnFluxThe email addy bounces04:45
JohnFluxbut I've seen this mess elsewhere04:46
JohnFluxI worked for a company writing video drivers for SGX04:46
JohnFluxpeople keep pushing us to release the code for the driver..  but there's a good reason we don't show anyone it.. :-D04:47
johanbrright :)04:47
johanbrI think opensource encourages good programming practices04:47
JohnFluxwe used RCS, and actually sent customer patches as a Word document!04:48
JohnFluxwith instructions to copy and paste the text, then manually fix the smart quotes..04:48
johanbryikes :)04:49
JohnFluxwhen I left, last year, they were planning to upgrade to CVS04:49
JohnFluxthey still haven't though :-D04:49
johanbrmaybe they got rid of the Word patches at least :)04:50
JohnFluxno, I tried to at least automate the process..  but no luck04:50
syn-ackSeriously, *word*?05:09
syn-ackman, I could understand something like notepad or wordpad but *word*?!!? /me shudders05:09
crimsunbjf: patch for #303789 sent to stable and upstream05:46
=== lexical is now known as kengyu
amitksmb: the d-i patch, shouldn't it say modularise instead of de-modularise?08:24
smbamitk, Yes Andy should have known better. Thats how those patches are now in the repo. I won't rewrite history08:25
* amitk doubts andy's english skills, he's english after all 08:26
smb(and actually we all failed to spot it in looking at the patches. /me included)08:26
smbamitk, MY language skill get bad in any language given the time of day is late enough. ;-P08:27
amitkthere has to be a better way to create our d-i modules files automagically - comparing config file changes to d-i changes is painful08:31
amitksmb: do you know if PATA_SIS == module(pata_sl82c105.ko)?08:36
amitkI guess I can look at the Kconfig help text08:37
smbamitk, I created the list from the modules I found on my build system.08:37
smbThere are a few that are different between amd64 and i386 but I hope this catches all08:37
amitksmb: some discrepancies08:40
amitkPATA_SIS is enabled in the config, but not in d-i08:40
smbOk, I should add that then08:41
amitkIn Sata:08:42
smbamitk, Hm, pata_sis? I only see sata_sis.ko08:42
amitksmb: I see a PATA_SIS enabled in debian.master/config/config.common.ubuntu08:43
amitksata looks good08:45
smbamitk, For some reason the PATA_SIS ends up being built in...08:46
smbamitk, Need to check the lower config files08:46
amitksmb: you're right, I missed the =y, so yes it is built in08:47
amitkso according to config, the d-i is correct.08:47
smbamitk, Though I wonder whether this should also becom =m08:47
amitkwhether or not PATA_SIS should be built-in or not is a separate story08:47
smbyeah08:47
smbI make a sticky-note for apw08:47
smbamitk, I just want that d-i change out and in today as this breaks a good deal of things08:48
amitksmb: ack away08:51
smbamitk, Many thanks08:51
JohnFlux_smb: Hey man09:08
smbJohnFlux_, mornin09:08
JohnFlux_smb: I'm going to ask a very stupid question..  what does "Fix Committed" mean ? :-)09:08
JohnFlux_smb: This means it's now a patch in the ubuntu kernel, right?  But not yet accepted by the staging drivers guys upstream?09:09
smbJohnFlux_, :) That I put the patch into our git tree. When it gets uploaded and released it will become fix released09:09
smbThe status has nothing to do with upstream. Just internal workflow09:09
JohnFlux_smb: got ya.  We added 0x0411, 0x015D     but I'm seeing the occasional post about the same thing also being   0x0411, 014F09:12
JohnFlux_smb: e.g.  http://forums.ncix.com/forums/index.php?mode=showthread&msg_id=2026623&threadid=2026623&forum=105&product_id=37959&msgcount=0&overclockid=009:12
smbJohnFlux_, I plan to do an upload late today. So it should become available latest next week. For some things I need other people to do it, so it can get delayed.09:12
JohnFlux_smb: Should I update  https://help.ubuntu.com/community/HardwareSupportComponentsWirelessNetworkCardsBuffalo   to change the status to working?09:13
smbJohnFlux_, If there are other IDs, those should get filed as separate bugs. Nothing worse that one bug with plenty of different "oh and this ID too"09:13
smbJohnFlux_, Maybe wait for the official kernel to be available for that09:14
smbJust being paranoid. ;-)09:14
JohnFlux_smb: hum, it seems easier to just add them now and see if we get bug reports about them09:15
smbJohnFlux_, And again, the official kernel will not contain the ID for that ralink 2070L09:15
JohnFlux_official ubuntu kernel, or official linus's kernel?09:15
JohnFlux_oh right, the 207009:16
JohnFlux_Hopefully I can get someone else to confirm it works in time09:16
JohnFlux_but..   what's the harm?09:16
JohnFlux_we know that this is the driver that should be used09:16
JohnFlux_we know that it _should_ work09:16
JohnFlux_and we know that without the change, it certaintly won't work09:16
JohnFlux_so what's the harm in just adding it?09:16
smbJohnFlux_, Exactly _should_. I like to be _certain_09:17
JohnFlux_smb: I can be _certain_ that it won't work without it09:17
JohnFlux_smb: how's that for certain? :P09:17
smbJohnFlux_, Not enough to convince me. :-P09:17
JohnFlux_smb: I don't really get why not :P  Seems a very obvious choice between "certaintly won't work" and "probably should work"09:18
smbJohnFlux_, But I won't carry a patch that _might_ work. If it _does_ work it should go upstream, but I won't bother Greg with something that maybe works09:18
JohnFlux_okay09:19
amitkJohnFlux_: we don't carry "probably works" patches. They has to be verification that it works.09:22
JohnFlux_is there a way to force a driver to work anyway for a given usb id?09:25
smbJohnFlux_, I think there is a bind file in sysfs for each driver, which you can echo IDs into...09:29
amitkJohnFlux_: http://lwn.net/Articles/143397/09:31
amitkJohnFlux_: apologies, that was for driver binding, this one is for adding a new ID: http://www.ha19.no/usb/09:32
smbamitk, Good find. I just was about to mention new_id09:33
JohnFlux_very interesting09:33
JohnFlux_smb: http://ubuntuforums.org/showthread.php?p=9030092#post9030092   someone said that your kernel doesn't boot for them?13:16
JohnFlux_smb: they mentioned that they have an eepc though13:16
JohnFlux_'mounted root fs can not mount /dev/pts and something else cant mount rootfs'13:16
JohnFlux_Does this error mean anything to anyone ?13:17
smbIt seems to be unable to find or mount the root file system and another virtual fs. Why is hard to say. Have you tried the kernel I made?13:19
smbProbably not as you had a working kernel before13:19
JohnFlux_smb: sorry, this is what someone else said13:21
JohnFlux_smb: I haven't actually tried your kernel :-)13:21
JohnFlux_smb: I pasted that error from that link I just gave13:22
smbBut the information is not very detailed. Could be something gone wrong installing. Could be something in the build. But its just the usual build with two more patches13:22
smbI have seen that. But its hard to tell more13:22
JohnFlux_smb: might be something related to the EEPC ?13:23
smbI would doubt it. If they had been running Ubuntu before...13:24
cndhow do you delete a chroot after you don't need it anymore?13:27
* cnd doesn't need lucid chroot now that I'm running lucid13:27
smbcnd, You could rm -rf and remove the entry from the config. But I would still keep it as its a better controlled environment13:29
cndsmb, good point, but not good enough when I need to conserve space :)13:30
cndsince I do most of my kernel builds on emerald.mills anyways13:31
cndwell, all my builds really13:31
=== sconklin-gone is now known as sconklin
smbcnd, How can one be so tight on space. I hardly was able to by a laptop hardrive below 250G nowadays. :-P13:32
smbcnd, But as said. You can edit /etc/schroot/schroot.conf and remove the lucid entry13:32
smband then just remove the directory where you installed the chroot13:33
pmatulisare there recent kernels (PPA) i can try on 8.04? there is "Using newer kernels on LTS releases" in LP but it says the project is "closed"13:33
smbcnd, Just make double sure, that nothing has bind mounted /home somewhere in it13:33
cndsmb: macbook has 120GB, ~80 is used by os x (which I don't really use anymore, but need to keep around), and I have the rest split between main partition and test partition13:33
cndsmb: good call13:33
smbpmatulis, I don't think so.13:34
ogasawaramanjo: https://bugs.edge.launchpad.net/ubuntu/+source/linux/+bug/507148 - will you be able to test the upstream patches?14:00
ubot3Malone bug 507148 in linux "[lucid] desktop runs out of video memory on ATI Radeon Mobility 7500" [High,Confirmed] 14:00
smbmanjo, Have you tested with the -17 kernel on that?14:03
Keybukso tempted to do a companion blog to cking's blog of awesome14:09
Keybuk"system calls and C library functions you didn't know about"14:10
Keybuktoday I found statfs()14:10
ckingKeybuk, I only blog about obscure stuff so I can google search it later :-)14:10
Keybukcking: and as a result, your blog is one of the most useful ones in the known universe ;-)14:11
ckingyou're too kind14:11
Keybukmore particularly, I know this trick14:13
Keybukstat("/") => info14:13
Keybukstat("/dev") => dev_info14:13
Keybukif info.st_dev != dev_info.st_dev:14:13
Keybuk  then dev is mounted14:13
cndsmb, so about that acpi patch that upstream isn't biting on, I still think it should go into lucid14:13
Keybukbetter trick, just statfs("/dev") :-)14:13
cndsmb: http://patchwork.ozlabs.org/patch/47164/14:13
smbcnd, I think I remember the one about the warning level for those resource conflicts14:14
cndsmb: yep14:14
smbcnd, Hasn't mjg59 acked that sort of? No I guess just said it would be ok14:14
cndsmb: right, he just seemed to agree in principle at the time14:15
cndbut I'd be happy to have mjg59 ack it (wink wink)14:15
cnd:)14:16
smbcnd, If nothing happens, we might go and take it as a sauce patch for now. But I guess I will leave that to the time apw gets back14:16
cndsmb: ok, just want to make sure it's not left until it becomes too late for inclusion14:17
smbcnd, No its still clearly visible on our list 14:17
cndI'm just new to the process, so I figure it's better to be safe than sorry14:18
ckingKeybuk, that's a trick worth documenting14:26
=== zul is now known as sorensbot
=== sorensbot is now known as zul
cndamitk: what was your methodology for determining which power saving features to put into powersave-policy?14:42
cndI'm curious if there's anything else that didn't do much for you, but maybe works well for others?14:42
cndoverall though, I'm not seeing a huge difference in power consumption with the extra policy scripts14:43
amitkcnd: the first step was just to make laptop-model-tools redundant. There are a few more script in laptop mode tools that were very hacking and/or too intrusive14:46
amitks/hacking/hackish14:46
cndamitk: so we've basically taken everything from laptop-mode-tools that seems reasonable?14:46
amitkcnd, yes IMO. I'll send you an email listing why I didn't pick the others.14:48
cndamitk: thanks14:48
=== kamalm-away is now known as kamalm
=== sconklin is now known as sconklin-afk
ogasawaraJFo: can you add bug 532374 and bug 548513 to our list for monday15:08
ubot3Malone bug 532374 in linux "Lenovo Thinkpads with Core i5 and i7 suspend/resume (with kernel oops) once then fail horribly on next suspend" [Critical,Confirmed] https://launchpad.net/bugs/53237415:08
ubot3Malone bug 548513 in linux "Firewire disks not working under 10.04" [Critical,Confirmed] https://launchpad.net/bugs/54851315:08
JFoogasawara, will do15:08
smbcking, Isn't that your pet area nowadays? ^15:09
JFoogasawara, done15:09
ogasawaraJFo: thanks. robbiew asked that we get them on our radar.15:09
JFono problem15:09
ckingsmb, yes jerone asked me to eyeball that when I was on vacation. It's on my ever growing list15:10
JFoand I just made the list huge15:11
JFoKernel Regression Bug Day schedule announced in e-mail15:11
smbcking, I wonder whether we might be allowed to volunteer you to have a look after that15:11
ckingplease do15:11
smbJFo, ^15:12
JFook, for the bug call on Monday?15:12
pgranercnd: you want to save power? Turn of the bling, doing that on my netbook saved almost 2watts15:12
cndpgraner: bling?15:12
JFosmb, that for the firewire one?15:12
pgranercnd: bling == compiz15:12
smbJFo, If its on that list cking would have the best insight I think (no the i7)15:12
pgranercnd: putting the gpu into 2d mode makes a big difference15:13
cndpgraner: I have a feeling people would get upset if we started flipping compiz on and off every time you plug in and out of ac15:13
* cking notes that disabling video and ssh'ing in helps save power too :-)15:13
pgranercnd: yea since now it tends to gather you desktops onto one15:13
pgranercking: you're so damn old school15:14
pgranercking: X for me is nothing more than a VT arranging utility15:14
cndpgraner: do you know how to disable compiz and 3d on demand?15:14
ckingpgraner, I'm happy with a serial console ;-)15:14
cndI'd like to try it personally just to see how effective it is15:14
pgranercnd: nope, talk to bryceh 15:14
pgranercnd: or one of the desktop guys15:14
JFosmb, ok15:15
cndcompiz for netbook edition seems overkill if I can get 2W back15:15
cndsince it only uses 10W right now anyways15:15
ckinghttp://smackerelofopinion.blogspot.com/2009/11/improving-battery-life-on-hp-mini.html is my powersaving hack15:15
cndand it should help with flash15:15
ckingand also: http://smackerelofopinion.blogspot.com/2009/06/reducing-wifi-beacon-interval-to-save.html15:16
pgranercnd: if your running compiz you can use: metacity --replace  & compiz --replace to switch between them15:20
pgranercnd: I'm supposing you want to test it in the scripts?15:21
cndpgraner: nope, just running metacity15:21
cndpgraner: no, just for personal usage15:21
cndand documenation15:21
pgranercnd: well thats how you switch inbetween them15:21
cndok15:22
crimsunamitk: / cnd: hmm, I see (using cnd's pm-utils-powersave-policy 0.4~powersave2) that powerdown is now enabled for all hda when on battery -- that annoying pop will return for non-IDT/Sigmatel codecs15:27
crimsunthat was the reason for the earlier check for the idt module15:27
cndcrimsun: yeah, Keybuk mentioned it, and I forgot to remove the script in powersave215:27
crimsunah ,ok15:27
cndcrimsun: unless there's some easy way to check for good codecs in the script?15:28
cndthough in that case, it should just be on all the time15:28
crimsuncnd: yes, in the original, I check for the existence of the module dir in /sys15:28
cndand besides, this is targetted at M, at which point the popping should be sorted out for all the codecs right?15:28
crimsunsec, will pull the diff15:28
crimsuncnd: ah, so it's not for Lucid?15:29
cndcrimsun: no, it's too late for Lucid15:29
crimsunah, then that's no big deal :-)15:29
crimsuncnd: yeah, just a simple [ -e $CODEC -a -w $PD ] currently (where $CODEC is /sys/module/snd_hda_codec_idt)15:31
cndcrimsun: so in M, the popping will be fixed for all codecs?15:32
crimsuncnd: kernel version-dependent, but yes15:33
cndcrimsun: what do you mean?15:33
crimsuncnd: well, there's the assumption that 2.6.32-2.6.34 won't be chosen15:34
cndcrimsun: ok?15:35
crimsuncnd: so with a sufficiently new kernel, then yes, those symptoms are fixed15:35
cndhow new are we talking? .34 isn't out yet15:35
cndis there some patch waiting in the wings for the .35 merge window?15:36
crimsunyes, I have updates for the conexant and realtek ones15:36
crimsunthere were a number of realtek ids added since 2.6.3215:36
cndok15:37
cndeven if M is based on .34, we'd likely take a look at the patches anyways15:37
cndespecially if it's just a matter of ids at that point15:37
crimsunok15:38
cndamitk: what about laptop_mode itself?15:40
cnddoesn't laptop-mode-tools enable that? pm-powersave doesn't yet15:40
margaHi!  I'm a Debian person (user, developer, etc), but I'm trying to help an Ubuntu user, that has a 32bit user space Ubuntu installation, with a 64bit processor.  In Debian we have an -amd64 flavor, for 64 bits processors with the i386 userspace.  By looking at the Ubuntu flavors for i386, the one that is 64bits -apparently- is -server... What is special about -server? Can it be used by a 64bit desktop?15:46
amitkcnd: yeah, that should be enabled too15:49
cndamitk: I reread the spec, and laptop_mode shouldn't be enabled by default per Ted T'so due to interactions with ext416:07
cndso we look good there16:07
cndamitk: thanks for the writeup!16:18
cndhelps explain many questions I had16:18
nosse1Hi. I'm trying to build a kernel for an ARM target. Where can I find a "standard" set of kernel config which represents what ubuntu requires from the kernel?16:31
manjoogasawara, on https://bugs.edge.launchpad.net/ubuntu/+source/linux/+bug/507148 building and testing the kernel on jdstrand's laptop 16:40
ubot3Malone bug 507148 in linux "[lucid] desktop runs out of video memory on ATI Radeon Mobility 7500" [High,Confirmed] 16:40
ogasawaramanjo: cool, thanks16:40
manjoogasawara, was busy with several audio bugs ... so results will be after lunch ... 16:40
ogasawaramanjo: k thanks, just mainly wanted to know the status for the release meeting16:41
manjoogasawara, ah ok... 16:41
crimsunmanjo: for #548371, your Mic Boosts are still set at zero16:53
crimsun(...unless you're just pasting outdated alsa-info output?)16:53
manjocrimsun, no the last one I posted was with  ubuntu audio dev ppa 16:54
crimsunmanjo: I mean that you're running the script when the capture is not occurring16:55
crimsunsame for #528719, BTW16:55
manjohmmm so what is that amixer option to boost it ? 16:57
manjocrimsun, ^  ?16:57
crimsunmanjo: (sorry, I'm at a conference currently) for which bug?17:11
manjoshould I do amixer set 'Capture',1 cap17:12
crimsunmanjo: for #528719, it's amixer set 'Mic Boost' 100%,100%17:12
manjook17:13
crimsunsimilar for #548371 but you also need 'Front Mic Boost'17:13
manjocrimsun, after the boost I still get white noise 17:17
manjocrimsun, I hear my voice very very faintly.17:17
crimsunmanjo: for which hw?17:19
manjoacer17:19
manjo#52871917:19
manjocrimsun, on the aspire1 17:22
manjoSimple mixer control 'Mic Boost',017:22
manjo  Capabilities: volume penum17:22
manjo  Playback channels: Front Left - Front Right17:22
manjo  Capture channels: Front Left - Front Right17:22
manjo  Limits: 0 - 317:22
manjo  Front Left: 3 [100%]17:22
manjo  Front Right: 0 [0%]17:22
manjohow do I boost  Front Right: 0 [0%] to 100% ? 17:22
crimsunmanjo: err, did you use the syntax above?17:22
manjoyes17:23
crimsunhuh. Well, just 100%17:23
manjook that made both 100%17:24
crimsunbah, bug in amixer :-(17:24
manjook tried after setting that.... and again .. white noise with faint recording of the sound 17:25
manjocrimsun, btw the bug is only when using external mic, and does not occur with built in mic 17:26
manjocrimsun, I have used 2 diff mics to make sure its not the mic hw17:26
crimsunmanjo: ok, please update the bug reports, respectively. I'm having a difficult time tracking irc and conference simultaneously, sorry17:28
manjocrimsun, no problem! will update17:28
crimsunI should have a free stretch in two hours17:28
=== Hedge|Hog is now known as Hedgehog
=== Hedgehog is now known as Guest23197
=== sconklin-afk is now known as sconklin
cndbryceh: I'm looking at bug 54474121:19
ubot3Malone bug 544741 in linux "[X700] KMS, amd64: Kernel panic while trying to launch system > preferences > appearance" [High,Triaged] https://launchpad.net/bugs/54474121:19
cndI saw your email about edid quirks21:19
cndactually, let me read the email once again, and I'll get back to you21:20
mozmckHow do I create a new flavour of the lucid kernel?22:55

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