/srv/irclogs.ubuntu.com/2013/12/09/#ubuntu-kernel.txt

=== shuduo_afk is now known as shuduo
=== shuduo is now known as shuduo_afk
=== shuduo_afk is now known as shuduo
=== gfrog_afk is now known as gfrog
=== gfrog is now known as gfrog_meeting
=== gfrog_meeting is now known as gfrog
* apw yawns08:47
=== shuduo is now known as shuduo_afk
* cking reboots10:29
caribouapw: morning, remember the small patch you sent my way a while ago (bug 1233175)10:44
ubot2`Launchpad bug 1233175 in linux (Ubuntu Precise) "Kernel panic : mempolicy potential use-after-free on server running mongodb" [High,In progress] https://launchpad.net/bugs/123317510:44
apwcaribou, some yep10:56
apwcaribou, not that i can find it right now, you got a copy?10:59
caribouapw: sure, let me forward your email11:00
apwit must be here somewhere, but if you have that is easier :)11:00
=== smoser` is now known as smoser
apwcking, THP did you notice any increase in resident set on your test rig when running?  just noted the warning in the config option15:20
ckingapw, good question, I need to check for that, I completely overlooked that fact15:21
jdstrandogasawara: hey-- check out this paste: http://paste.ubuntu.com/6546723/. The machine is using wlan0, not eth0 but after a few minutes eth0 shows this weird output even though eth0 isn't being used. This is happening on a Lenovo IdeaPad U460 on precise17:18
jdstrandogasawara: I think this may have started with 56 (I"ve seen it on -56 and -57-- cursory testing of -55 didn't show it)17:19
jdstrandogasawara: have you seen any bug reports with weird networking on recent precise SRU kernels?17:19
ogasawarajdstrand: hrm, I haven't heard of anything.  jsalisbury, bjf: ^^ sounds familiar at alll?17:20
ogasawarajdstrand: sounds fairly reproducible though from your description.  would you have time to run through a few test kernels to bisect it?17:21
jdstrandat some point-- it isn't my system (it is my wife's and she is using it)17:22
bjfogasawara, jdstrand news to me17:22
ogasawarajdstrand: can you do an 'ubuntu-bug linux' from it for me17:23
ogasawarajdstrand: we can check if we have hw in our possession to reproduce17:24
jdstrandogasawara: otoh, is there a trick with doing that from cli only?17:24
ogasawarajdstrand: hrm, not sure there17:24
jdstrandok, let me see what I can do17:24
* jdstrand doesn't have lp credentials on that machine17:24
* ogasawara wonders if we have the same kit in cert to test with17:25
jdstrandok, I just ssh in and do it and it walks me through17:26
rtgit might just be a stable update on that driver in -56.17:27
jdstrandthat is what I was wondering17:28
jdstranderf, lp timeout error17:30
jdstrandogasawara: bug #125925517:31
ubot2`Launchpad bug 1259255 in linux (Ubuntu) "incorrect ifconfig output for unused wired ethernet" [Undecided,New] https://launchpad.net/bugs/125925517:31
ogasawarajdstrand: thanks, we'll see if we can chase this down17:32
jdstrandogasawara: so, I've rebooted her into -55. I used the same networking tests that I did on -56 and -57 and ifconfig still looks ok. I'll keep an eye on it17:37
rtgjdstrand, no changes to atl1c ethernet driver since Ubuntu-3.2.0-55.85. Does it happen right away ?17:38
jdstrandno-- it seems to happen after I connect to a jabber server. I did that on -55 and it is still ok17:38
jdstrandrtg: is atl1c the wired driver?17:39
jdstrandrtg: cause note-- the machine doesn't have a cable connected to it on eth0. it is only using wireless on wlan017:39
rtgjdstrand, but you're connecting via wifi, right ? According to your dmesg it is an atl1c ethernet device. 17:39
rtg[    2.140905] atl1c 0000:04:00.0: PCI INT A -> GSI 18 (level, low) -> IRQ 1817:39
rtg[    2.140918] atl1c 0000:04:00.0: setting latency timer to 6417:39
jdstrandafter using the network via wlan0, after a while eth0's ifconfig is weird17:39
rtgok17:40
jdstrandyes, connecting via wifi17:40
rtgso, its likely something external that changed interrupt routing, or stats updates in the net sub-system, etc. I imagine jsalisbury can help you with a bisect.17:41
jdstrandshe was complaing about networking issues in general on -57-- but she reboots so infrequently it is possible whe was running -55 until today/this weekend17:41
* jdstrand nods17:41
jdstrandI did see a lot of 'net' changes17:41
jdstrandin the changelog17:42
rtgyep, I saw those as well17:42
henrixjdstrand: its weird that you have "[ 5322.316796] device eth0 entered promiscuous mode"17:42
henrixjdstrand: was this some test you exec'ed?17:42
jdstrandno. I used tcpdump on wlan017:42
* jdstrand just checked history17:42
henrixah, ok. that could explain it17:43
* henrix notes that all the ifconfig values are roughly 2^32 (a hint for overflows?)17:43
rtglikely an underflow17:44
jdstrandthis was interesting in -proposed's kernel: net: heap overflow in __audit_sockaddr()17:44
jdstrandbut there were several things in -proposed kernel that looked interesting. I didn't try -proposed yet17:45
jdstrandhrmm, -55 has it too17:46
jdstrandit seemed to take a little longer, but I see it17:47
rtgjdstrand, is it related to promiscuous mode ?17:47
jdstrandI haven't used tcpdump since this reboot17:47
rtgjdstrand, knowing the atl1c driver, this issue might have existed since the first release. there have only been 2 stable updates since then.17:50
jdstrandI wonder if it is in linux-firmware17:50
jdstrandI recorded that I upgraded linux-firmware on 12-0417:50
rtgjdstrand, checking... but I don't recall that it loads firmware.17:51
jdstrand(the date that is)17:51
apwatl1c, is that an ethernet device17:51
apwhaving to rmmod that rings a bell, so i might have one17:51
rtgjdstrand, only firmware load is from your wifi gizmo17:52
rtgjdstrand, just for kicks try installing the Saucy LTS kernel to see if you get the same result17:54
jdstrandI wonder if this is a red herring for her network issues. I guess I should try -23 to see if it was always there and -58 to see if it fixes it17:55
* rtg -> lunch18:29
jdstrandok, -58 exhibits the problem. -23 didn't in the amount of time I tested it. the saucy-lts kernel did not either, but I couldn't test it for very long cause the nvidia driver wasn't behaving18:47
jdstrandI updated the bug18:47
jdstrandright now I put her on -58 and trying to see if her network issues go away18:48
xclaesseAnyone knows what's causing https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1258838 ?18:51
ubot2`Launchpad bug 1258838 in linux (Ubuntu) "[LENOVO 23252EG] suspend/resume failure" [Undecided,Confirmed]18:51
xclaesseit's a really annoying regression in 13.1018:52
xclaesseif I can provide any extra info, please ask18:52
ckingapw, rss updated hugepage info in your inbox18:53
jsalisburyjdstrand, I posed a comment in the bug.  We can either perform a regular bisect to find the commit that introduced the bug, or we can perform a "Reverse" bisect to find what commit fixes this in Saucy.19:35
jdstrandjsalisbury: well, I'm not 100% sure that it is not broken in -23 or fixed in saucy.19:50
jdstrandI'll need to get back to you and let her run -23 for a while (since getting saucy to work for her would take me a bit of fiddling with nvidia)19:50
jsalisburyjdstrand, ok.  We can bisect if there is a way to reliably reproduce the bug.19:51
jdstrandthat's the trick. it seems to show up in -55 - -58 regularly, but I don't have a reduced test case yet19:51
* rtg -> EOD21:03

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