dandel_ | hmm... i found a bug with the ext3 subsystem used for the virtual partition install of ubuntu 9.04 :/ (when using kernel 2.6.32-rc7) | 00:34 |
---|---|---|
=== dandel_ is now known as dandel | ||
dtchen_ | reported upstream? | 00:35 |
dandel | somewhat. | 00:35 |
dandel | i'm trying to reopen a bug that got closed ><; | 00:35 |
dandel | i haft to figure out how to over ethernet send the debug ><; (the crash happens upon resume) | 00:35 |
dandel | 2.6.28 (which is used in 9.04 by default does not display this bug, but the system still crashes shortly afterwards) | 00:36 |
* dandel checks 2.6.32-rc6 for bug. | 00:37 | |
dandel | that also displays the bug (go figure) | 00:39 |
dtchen_ | and rc8? | 00:42 |
dtchen_ | rc7* | 00:42 |
dandel | i'm testing 2.6.31 ><; | 00:42 |
dandel | 2.6.31 also does the crash on resume 0o | 00:43 |
dandel | i know it's some sort of module change... the crash does not occur when i suspend with testing via s2ram on init=/bin/bash | 00:45 |
dandel | it'd be excellent if i could get the kernel dmesg stuff to go to my tower (the laptop lacks serial ports tho) | 00:51 |
maco | is one of your USB ports a debug port? | 00:55 |
dandel | no clue. | 00:55 |
dandel | however, the kernel did not panic, so the log is scrollable (although it is a little short) | 00:56 |
dandel | how can i make the usb port a debug port? | 00:57 |
maco | http://www.coreboot.org/EHCI_Debug_Port | 00:58 |
dandel | Yes, the usb does support debug 0o' | 01:00 |
maco | this time, im bookmarking that link. i had to search through last december's chatlogs to find it :P | 01:01 |
dandel | so how do i run the device as a debug port then? | 01:02 |
dandel | (Keep in mind that i have 2 EHCI controllers, and both of them support debug) | 01:02 |
* dandel makes note that both the laptop and tower have debug port capability on the usb. | 01:08 | |
maco | ive never done it myself. the $83 pricetag on that hardware was a bit steep for me | 01:09 |
dandel | ><; | 01:09 |
dandel | any idea on how to get the debug to go over ethernet? | 01:09 |
maco | spose you could ssh in before it breaks... | 01:11 |
dandel | can't | 01:11 |
maco | right....suspend/resume | 01:11 |
maco | i wasnt thinking | 01:11 |
dandel | hmm... my only option now is to start gutting module trees with the modprobe blacklisting. | 01:33 |
dandel | ok... got the crash down to about 15 modules which i can't unload. | 02:39 |
hagabaka | why doesn't the kernel ppa work as a repository now? | 03:41 |
asabil | hi all | 12:08 |
teabag | hello all, I am getting this error during the makefile process | 14:28 |
teabag | http://paste.ubuntu.com/319311/ | 14:28 |
teabag | can anyone plz tell me what is wrong and how to fix it ? | 14:28 |
teabag | ah nevermind | 14:39 |
teabag | go it | 14:39 |
=== teabag_ is now known as teabag | ||
=== teabag is now known as pronoy | ||
=== eakron__ is now known as eakron | ||
_Narc_ | Hello, Kernel Wizards. I got a little something to ask you. I had/I'm having/ a hard time figuring out if a bug is coming form the kernel or from something else. It's network related, especially to the r8169 module... Anyone I could explain to ? Thanks | 20:36 |
dtchen_ | _Narc_: please don't ask to ask. | 21:41 |
dtchen_ | _Narc_: and, it's better if you have a bug report reference. | 21:41 |
_Narc_ | dtchen_: Sorry. The problem is I don't have a bug report yet because I waited to have an advise here to submit it under the right package. | 22:27 |
BenC | _Narc_: you can file a bug report against just ubuntu and specify the package later | 22:28 |
BenC | _Narc_: since you know it is in ubuntu, you know you will have to file a bug report anyway, so best to do it from the start and narrow down the exact package after it's been triaged a bit | 22:29 |
_Narc_ | BenC: Good idea, I didn't think about that. I came here to ask because someone on ubuntu-bugs told me it was weird and to ask to someone who knows the kernel better. But I'll do that, thanks. | 22:30 |
BenC | _Narc_: you can still explain here if you can provide enough detail | 22:30 |
_Narc_ | BenC: Yes, I even have too much of them. Too make a long story short, it's been months that I experience network hangs/ deconnexions when the network was under heavy load, using bittorrent (Miro) for example. I tested with a different router, changed nothing. Then I looked in syslog and found this : http://paste.ubuntu.com/318620/ . | 22:35 |
BenC | _Narc_: not so sure that's a bug at all... | 22:37 |
BenC | _Narc_: if anything, it is kernel, but just saying that the driver may be acting as expected | 22:38 |
_Narc_ | With different people on ubuntu-bugs, we thought it was a r8169 issue. But now I discovered that changing the default port in Miro stops it. So, I need your help determining if it's Miro who's talking nonsense to the kernel or the kernel who can't handle it. | 22:38 |
_Narc_ | Oh really | 22:38 |
_Narc_ | Two people on ubuntu-bugs told me it was an Oops | 22:38 |
BenC | it's not an oops at all | 22:38 |
_Narc_ | Oh | 22:38 |
BenC | it's a WARN_ON() in the driver | 22:38 |
BenC | oops == crash, WARN_ON() == driver got something unexpected from the hardware and tried to handle it | 22:39 |
BenC | likely in this case, that means resetting the device | 22:39 |
_Narc_ | Ok, it means it's just telling that something is wrong, but acting as expected ? | 22:39 |
_Narc_ | Ok | 22:39 |
BenC | that's an initial guess | 22:40 |
_Narc_ | It's really something already :) | 22:40 |
_Narc_ | Do you know if the SYN flood messages just at the beginning could trigger this ? | 22:40 |
_Narc_ | Or is it unrelated | 22:40 |
_Narc_ | Because I had a lot of them. | 22:41 |
BenC | I believe it is directly related | 22:41 |
BenC | need to find out why the kernel thinks it is getting syn-flooded | 22:42 |
_Narc_ | Oh. That's interesting, because that's what I thought, and I unabled the syn flood cookies, and it still hanged. But changing Miro's default port from 8500 to 51413 solves it. So, I wondered why one and not the other... | 22:43 |
_Narc_ | I say Miro specifically because it's the only app who's doing this. On this port. And the syslog message I showed you does not appear at each network hang. | 22:46 |
_Narc_ | So, hardware, kernel, miro... I'm not good enough to figure out. Ask me if you need any other info. | 22:47 |
BenC | _Narc_: do not disable syn cookies, that causes more problems | 22:48 |
BenC | _Narc_: quite possible you need to lower some limits in Miro so that it doesn't cause flooding on your connecting | 22:48 |
_Narc_ | Yes, that's what I did. But the strange thing is that I only get the syn warnings when miro's port range are 8500-8600. So, it's not a kernel issue you say, because it's acting brave and try to handle miro's/my hardware crappy packets, right ? | 22:52 |
BenC | _Narc_: I believe so | 23:00 |
_Narc_ | Ok, thank you very much then. I'm starting to understand this more. It's probably a Miro bug then. Because my hardware is not exotic. | 23:02 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!