/srv/irclogs.ubuntu.com/2009/07/20/#ubuntu-kernel.txt

=== MobileMyles6o7 is now known as TwoToneSpirit
NCommanderapw, do you still need to forward the SPARC patches? (it hit the queue, but I didn't get the bounce until I took off for the weekend)02:32
cooloneyericm, Jaunty 2.6.28 kernel already merged that patch?06:12
cooloneyericm,  do you know the SHA of the commit?06:12
ericmcooloney: please check arch/x86/kernel/cpu/intel.c: line 3906:12
ericmcooloney: seems to be merged via the stable tree06:13
cooloneyericm, that is OK06:13
ericmthe change is brought up by the fix06:14
cooloneyericm, we can change the status to Fixed in Release06:14
cooloneyericm, cool06:14
ericmcooloney: now your turn ;-)06:14
cooloneyericm, ok, no problem. 06:14
cooloneyericm, thanks for taking a look at this06:14
ericmcooloney: you are welcome, bro06:15
kdubhello everyone10:43
=== lamont` is now known as lamont
apwkdub, hi12:31
* apw pokes apw_12:39
smbapw, talking to yourself again. :-P12:43
* apw_ waves to cking 12:43
apwheh yeah ... just got wireless working12:43
* Keybuk tries to figure out TRACE_EVENT13:43
rtg_amitk, are you accumulating all of the Karmic  ARM patches from the list? 14:08
Keybukrtg_: remind me, is there an easy way to build an i386 kernel on an amd64?14:23
rtg_Keybuk, other then using a chroot? not that I know of.14:24
Keybukthat's what I thought14:24
Keybuklet's see if this works14:26
KeybukI've rewritten the old sreadahead ftrace-based open() patch14:26
Keybukit now uses TRACE_EVENT14:26
KeybukI'm hoping it'll be much easier to get upstream like that ;-)14:26
Keybuk(assuming that the syscall tracer stuff doesn't make it)14:28
Keybukooh, it compiled :D14:30
rtg_Keybuk, is anyone championing the syscall tracer stuff ?14:31
Keybukno idea14:32
rtg_Keybuk, couldn't get it in Andrew's tree?14:32
Keybukwho couldn't/14:33
* Keybuk doesn't follow LKML remember14:33
KeybukI only have 24 hours in one day14:33
Keybuklast I heard about the syscall tracer stuff, Ingo had said he though it was the right thing to do14:34
rtg_Keybuk, he's also another avenue for getting a patch like that accepted.14:35
Keybuk(the syscall tracer stuff not being my work :p)14:35
rtg_oh, I thought it was. I get so confused.14:35
Keybuk:)14:36
Keybukin jaunty we have a patch in our kernel tree which adds a trace for the open() syscall using ftrace14:36
Keybukthat patch doesn't apply in karmic, so I've rewritten it to use TRACE_EVENT14:36
Keybukwe'll obviously want to include that in our tree if it works14:36
Keybukand I'll forward it upstream as well14:36
rtg_ok14:37
Keybukbut I'm mindful that there's separate work upstream to provide a generic all-syscall tracer using tracepoints, which would supercede it anyway14:37
rtg_its probably time to get it into Karmic 'cause whatever happens upstream won't appear (for us) until LL14:37
Keybukbut at least, by implementing it on top of trace_event, it'll be compatible :P14:37
Keybukright, I only wrote it today :)14:38
Keybukam doing the test compile now <g>14:38
rtg_slacker14:38
Keybukhah14:38
Keybukhmm, didn't quite work14:41
Keybukoh, no, I just didn't copy a file over14:44
Keybukhmm, no I didn't14:49
=== pace_t_zulu_ is now known as pace_t_zulu
* Keybuk is clearly missing some glue14:55
Keybukhow do I ignore ABI changes again?16:42
Keybukoh, I just fix the changelog entry, that's right16:43
rtg_Keybuk, you can also 'echo 1 > debian/abi/2.6.31*/ignore'16:44
rtg_rather, 'echo 1 > debian/abi/2.6.31*/ARCH/ignore'16:45
Keybukbtw, is it me or is the kernel package missing build-deps these days?16:47
Keybukneither kernel-wedge nor makedumpfile are in its deps16:47
rtg_Keybuk, hmm16:48
rtg_Keybuk, Build-Depends: debhelper (>= 3), module-init-tools, kernel-wedge (>= 2.24ubuntu1), makedumpfile [amd64 i386 lpia], device-tree-compiler [powerpc]16:48
Keybukrtg_: ah, but debian/control is an empty file16:50
Keybukthat's what breaks it16:51
rtg_debian/rules clean16:51
Keybukso deps work from the source but not from git?16:51
Keybukrtg_: clean depends on kernel-wedge <g>16:51
rtg_Keybuk, debian/control is generated16:51
rtg_Keybuk, the source package has a correct debian/control, but not the bits that are checked out from git16:52
Keybukyeah16:52
KeybukCould not open /home/scott/co/linux-karmic/debian/linux-image-2.6.31-4-generic/lib/modules/2.6.31-4-generic/modules.alias at debian/tests/check-aliases line 10.16:52
Keybukweird16:52
Keybukprobably an issue with changing the version part way though the build ;)16:53
rtg_Keybuk, likely. you're better off using an ignore file.16:54
Keybukwildo16:56
amitkrtg_: I am (collecting all the ARM-related patches, that is). They are at git://kernel.ubuntu.com/amitk/ubuntu-fsl-2.6.31.git16:57
rtg_amitk, shouldn't we have an official topic branch in the Karmic repo?17:01
=== bjf-afk is now known as bjf
bjfrtg_, we are using a topic branch named "arm" in the Jaunty tree17:02
Keybuk# grep open /sys/kernel/debug/tracing/available_events17:10
KeybukADMIRAL!  THERE BE TRACERS HERE!17:10
Keybuk\o/17:10
Keybuk# tail trace17:13
Keybuk  tail-2176  [001]    58.986011: do_sys_open: "/etc/ld.so.cache" 0 017:13
Keybuk  tail-2176  [001]    58.986054: do_sys_open: "/lib/tls/i686/cmov/libc.so.6" 0 26004517:13
Keybuk  tail-2176  [001]    58.986149: do_sys_open: "/dev/urandom" 0 2677660432017:13
Keybukyay17:13
Keybuknot sure why the junk at the end, admittedly17:13
Keybukah, random mode when opening without O_CREAT of course17:17
Keybukrtg_: incoming patch for you17:20
rtg_Keybuk, cool.17:20
apwjjohansen, did we decide that we were happy to enable apparmour now, as the ecryptfs issue was already in jaunty and the static IP addy thing there was going to be a work around in the definition17:55
jjohansenapw: nope, but I think it is worth doing17:56
apwi think it was a patch for that functonality wasn't it.  we should get that up onto the mailing list for inclusion17:57
jjohansenapw: yeah, it adds a config option to set a default security module17:57
apwcool.  so lets get that up soon and i can get it tested in my ppa17:58
jjohansenapw: okay, I will commit and push17:59
apwsounds good18:00
=== manjo is now known as manjo`away
=== manjo`away is now known as manjo
=== manjo is now known as manjo`away
jjohansenapw: request sent18:37
apwjjohansen, thanks18:37
rtg_apw, any erason why CONFIG_SSB and CONFIG_B44 shouldn't be common across all arches/flavours?18:43
apwprobabally not ... shall i sort them out as i have been doing that merging?18:44
apwand i am about to do an AA config option too :)18:44
rtg_apw, I've already got that committed locally and am build testing.18:44
apwahh ok ...18:44
rtg_the CONFIG_SSB/CONFIG_B44 patch (I mean)18:44
apwlet me know when you have it pushed so i can work from the top of that on the next config thing18:45
apwthey tend to collide like a pig18:45
rtg_k, about 30 minutes.18:45
rtg_I'm also getting Keybuk strace patch18:45
rtg_er, trace event patch18:45
apwsounds good to me18:46
smbrtg_, apw Probably the only risk is for bcmwl but jockey can sort that18:46
apwsmb, whats that risk?18:47
rtg_smb, its already built for the distro arches anyway. so jockey already has to handle it.18:47
apwoh i see having it turned on you mean18:47
apwyeah its arm which is mostly the one which isn't common18:47
smbapw, ssb has the tendency to grab all ssb capable devices (not necessarily only those it has backends for)18:47
apwahh i see18:48
smbapw, So the wl driver wont work until you remove ssb (plus those using it) then modprobe wl and the again all the others18:48
smbwhich is the modprobe rule jockey instals18:49
apwahh hence when i installel the wl thing i had to reboot to get it to work18:49
smbor remove b44+ssb then modprobe wl and probe the other like I did 18:49
smbexactly18:49
rtg_apw, karmic pushed19:21
rtg_Keybuk, what would be a good package for an rfkill application? its needed for acpi-support. perhaps wireless-tools ?19:32
Keybukyeah something like that19:36
rtg_Keybuk, so, just drop it in and send a patch to the debian maintainer?19:36
Keybukyup19:40
NCommanderapw, you around? (I'm guessing my SPARC config patch never materialized on the mailing list since i don't see it in history)19:53
apwnot seen it no19:53
apwif you want to send it again and cc: me then i can follow up on it19:53
NCommanderapw, *sigh*, ok, I have a copy here19:53
apwthese things happen19:54
apwi assume its not huge?19:54
NCommanderapw, not too big19:54
NCommanderapw, forwarded. If it fails to apply, then we've got a problem; my home router has failed, and I can't access my SPARC box remotely (and unless my super can restart the router which is a longshot), I probably won't be able to fix the problem until I return to Rochester in ~3 weeks19:55
apwlets hope it applies19:56
NCommanderapw, yeah .... I can try working on faure if it doesn't, but I obviously won't be able to test the results :-/19:57
NCommanderO_O;;;;  http://news.slashdot.org/story/09/07/20/1643251/Microsoft-Releases-Linux-Device-Drivers-As-GPL?art_pos=419:57
apwwell sparc wasn't building last i looked 19:57
* NCommander feels its a very very cold day in hell ...19:57
apwits an odd move, lets hope its the open move we would want it to be19:58
NCommanderIndeed19:59
NCommanderBut still, Microsoft committing kernel patches ...19:59
apwi like the chosen tags19:59
NCommanderNext thing you'll know, Apple is going to dump Darwin and move to a LInux base19:59
apwheh ... that really would imply flying pigs20:00
apwNCommander, you are in luck it applied20:03
NCommander\ooooooooo/!20:04
=== manjo`away is now known as manjo
NCommanderThat makes me feel better20:04
apwas it only can affect the ports its pretty safe20:04
apw(for me)20:04
NCommanderMy issue w.r.t. to the new layout of debian/config is one config change breaks all other pending config patches20:04
NCommanderBut on the flip side, its nicer to see what is common to all architectures ...20:04
apwyes and no, not as much as you might think20:04
apwyour change is an issue as its a big mother20:04
apwand changes a lot of the file20:05
* NCommander notes the ports configs are very very unhappy in general as it stands20:05
NCommanderWell, until now20:05
apwbut there was an apparmor related changed between your change being diffed and applied by me20:05
apwand they didn't collide20:05
NCommanderThats a nice warm fuzzy feeling20:05
apwthe ones where you do what you did there, take an arch from BUST to WORKING20:06
apwis going to run that risk, but generally we change one option at a time and the collisions are managable20:06
apwi am also sure this isn't the last improvement to config management in my lifetime either20:06
NCommanderWell, it just needed one big change to make everything happy20:07
NCommanderI'm kinda wondering how the configs got this broken though, I thought I fixed all the -ports configs in the jaunty cycle :-/20:07
apwcannot say, perhaps the wrong ones got pulled to karmic when they merged into the kernel package not sure20:09
NCommanderpossibly20:14
NCommanderOh well, fixed now20:14
apwheres hoping ... 20:15
manjodinh, did u get a chance to repond to amits question regarding the USB patch ? 20:45
rtg_apw, I figured out where NCommander's patch went. I'd forgotten to moderate the list this morning.20:59
apwrtg_, ahh did wonder ... it looked benign to us so i have applied it21:03
* NCommander doesn't understand why he keeps hitting the moderation queue21:07
NCommanderI'm subscribed, and @ubuntu.com addresses should be whitelisted21:07
apwmaybe the size limit is very small21:07
rtg_apw, 40k21:10
dinhmanjo: I responded to amitk's email with a condensed patch that only fixes USB...did you not see it?21:43
manjodinh, ah I must have missed it 21:43
manjodinh, thanks for the update sir21:43
dinhsure thang21:44

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