| --- | Log | opened Thu Jul 24 00:06:22 2008 |
| 01:24 | -!- | ram [~ram@pool-71-245-96-80.nycmny.fios.verizon.net] has quit [Quit: Leaving] |
| 01:52 | -!- | Magotari_ [~karol@abhw207.neoplus.adsl.tpnet.pl] has joined #uml |
| 01:54 | -!- | Magotari [~karol@abil39.neoplus.adsl.tpnet.pl] has quit [Ping timeout: 480 seconds] |
| 02:11 | -!- | silug [~steve@ppp-70-225-32-179.dsl.covlil.ameritech.net] has quit [Ping timeout: 480 seconds] |
| 02:11 | -!- | silug [~steve@ppp-70-225-32-179.dsl.covlil.ameritech.net] has joined #uml |
| 02:33 | -!- | aroscha [~aroscha@chello213047053193.30.11.tuwien.teleweb.at] has joined #uml |
| 02:33 | -!- | aroscha [~aroscha@chello213047053193.30.11.tuwien.teleweb.at] has quit [] |
| 02:35 | -!- | aroscha [~aroscha@chello213047053193.30.11.tuwien.teleweb.at] has joined #uml |
| 03:00 | -!- | low [~low@LSt-Amand-152-31-5-221.w82-127.abo.wanadoo.fr] has joined #uml |
| 03:01 | -!- | Basic [~Basic@c-75-73-131-27.hsd1.mn.comcast.net] has quit [Quit: Basic] |
| 04:10 | -!- | aroscha [~aroscha@chello213047053193.30.11.tuwien.teleweb.at] has quit [Quit: aroscha] |
| 04:35 | -!- | aroscha [~aroscha@nat.labs.nic.at] has joined #uml |
| 05:44 | -!- | Magotari_ [~karol@abhw207.neoplus.adsl.tpnet.pl] has quit [Read error: Connection reset by peer] |
| 09:03 | -!- | dang [~dang@aa-redwall.ghs.com] has joined #uml |
| 09:06 | -!- | jdike [~jdike@pool-72-70-38-200.bstnma.fios.verizon.net] has joined #uml |
| 09:06 | <jdike:#uml> | Hi guys |
| 09:11 | -!- | aroscha [~aroscha@nat.labs.nic.at] has quit [Quit: aroscha] |
| 10:05 | -!- | aroscha [~aroscha@nat.labs.nic.at] has joined #uml |
| 10:56 | -!- | hfb [~hfb@pool-71-118-254-245.lsanca.dsl-w.verizon.net] has joined #uml |
| 11:03 | -!- | aroscha [~aroscha@nat.labs.nic.at] has quit [Ping timeout: 480 seconds] |
| 11:13 | -!- | low [~low@LSt-Amand-152-31-5-221.w82-127.abo.wanadoo.fr] has quit [Quit: bbl] |
| 11:22 | <huslu:#uml> | jdike: i was able to reproduce that crash on halt on a different box (but which had most up todate libraries) |
| 11:22 | <jdike:#uml> | get a stack trace from gdb? |
| 11:22 | <huslu:#uml> | no, i do not know how to use it. |
| 11:23 | <jdike:#uml> | before you halt the UML, in another window, |
| 11:23 | <jdike:#uml> | gdb linux linux-pid |
| 11:24 | <jdike:#uml> | b panic |
| 11:24 | <jdike:#uml> | c |
| 11:24 | <jdike:#uml> | then shut it down and wait for the panic breakpoint to be hit |
| 11:24 | <jdike:#uml> | then bt to gdb |
| 11:28 | <jdike:#uml> | I still suspect the library, BTW |
| 11:30 | -!- | aroscha [~aroscha@chello213047053193.30.11.tuwien.teleweb.at] has joined #uml |
| 11:36 | <huslu:#uml> | "function 'panic' not defined" |
| 11:37 | <jdike:#uml> | bs |
| 11:37 | <jdike:#uml> | you gave the path to the linux binary? |
| 11:37 | <jdike:#uml> | on the gdb command line? |
| 11:37 | <huslu:#uml> | no, it's run as ./linux |
| 11:38 | <jdike:#uml> | so, gdb ./linux linux-pid |
| 11:38 | <jdike:#uml> | is what you did? |
| 11:38 | <jdike:#uml> | do you have debugging symbols enabled? |
| 11:38 | <huslu:#uml> | just gdb linux <pid> as earlier |
| 11:39 | <jdike:#uml> | CONFIG_DEBUG_INFO enabled? |
| 11:39 | <jdike:#uml> | and CONFIG_DEBUG_KERNEL, I guess |
| 11:39 | <huslu:#uml> | not sure, but i was able to do panic breakpoint with full path to my binary |
| 11:40 | <jdike:#uml> | hmm, OK |
| 11:41 | <jdike:#uml> | gdb might search the path, I guess |
| 11:41 | <jdike:#uml> | and might find a /usr/bin/linux before ./linux depending on your path |
| 11:43 | <huslu:#uml> | ok, i gave 'bt' command to gdb |
| 11:44 | <huslu:#uml> | http://pastebin.ca/1081929, but not much to see |
| 11:44 | <huslu:#uml> | it's kinda stopped now, anything else i should do? |
| 11:45 | -!- | jdike [~jdike@pool-72-70-38-200.bstnma.fios.verizon.net] has quit [Quit: Leaving] |
| 11:46 | -!- | jdike [~jdike@pool-72-70-38-200.bstnma.fios.verizon.net] has joined #uml |
| 11:47 | <jdike:#uml> | keyboard keeps dying |
| 11:47 | <jdike:#uml> | huslu, handle SIGSEGV pass nostop noprint |
| 11:47 | <jdike:#uml> | then c |
| 11:47 | <jdike:#uml> | and shut it down |
| 11:47 | <huslu:#uml> | i have it a halt command already |
| 11:50 | <huslu:#uml> | hm, debug seems not to be turned on. |
| 11:50 | <huslu:#uml> | but it didn't shut down, after the normal crash messages it's kinda stopped again, what should i do now? |
| 11:52 | <huslu:#uml> | after continuing it says: |
| 11:52 | <huslu:#uml> | Breakpoint 1, 0x08232509 in panic () |
| 11:52 | <huslu:#uml> | Current language: auto; currently asm |
| 11:52 | <huslu:#uml> | and gives me prompt in gdb |
| 11:53 | <jdike:#uml> | should have given you a line number |
| 11:53 | <jdike:#uml> | what does bt say? |
| 11:54 | <huslu:#uml> | it was in my earlier pastebin link |
| 11:54 | <jdike:#uml> | you hadn't hit panic at that point |
| 11:54 | <jdike:#uml> | just the segfault |
| 12:13 | -!- | Hunger [QuakeNet@Hunger.hu] has joined #uml |
| 12:15 | <huslu:#uml> | http://pastebin.ca/1081962 |
| 12:15 | <jdike:#uml> | OK check the above-mentioned config settings |
| 12:22 | <huslu:#uml> | debug stuff is not there, recompiling static kernel with debug options |
| 12:24 | <huslu:#uml> | does slub need to be debugged? |
| 12:24 | <jdike:#uml> | I suppose it wouldn't hurt |
| 12:25 | <jdike:#uml> | but that hasn't been useful so far |
| 12:27 | -!- | karol [~karol@abhq82.neoplus.adsl.tpnet.pl] has joined #uml |
| 12:27 | -!- | karol is now known as Magotari |
| 12:40 | <huslu:#uml> | jdike: now the kernel is ~30mb and it crashes right in the beginning after vde |
| 12:40 | <jdike:#uml> | OK, that's different |
| 12:41 | <huslu:#uml> | yup, only changed debug stuff |
| 12:41 | <jdike:#uml> | what do you get for a stack trace |
| 12:41 | <jdike:#uml> | not surprising though, memory corruption will move around due to other random things |
| 12:42 | <huslu:#uml> | http://pastebin.ca/1081981 |
| 12:42 | -!- | Basic [~Basic@gatekeeper.real-time.com] has joined #uml |
| 12:45 | <jdike:#uml> | have you brought this to the attention of the VDE people? |
| 12:45 | <huslu:#uml> | no, not yet. |
| 13:00 | -!- | aroscha [~aroscha@chello213047053193.30.11.tuwien.teleweb.at] has quit [Quit: aroscha] |
| 13:06 | <huslu:#uml> | http://pastebin.ca/1081999 uml kernel crash with no slub debug info compiled in. crash happened after swap |
| 13:09 | <jdike:#uml> | these are all memory corruption with the stack not telling us anything |
| 13:11 | -!- | ram [~ram@pool-71-245-96-80.nycmny.fios.verizon.net] has joined #uml |
| 13:12 | <huslu:#uml> | k, i guess i'm back to using what has worked before. |
| 13:44 | -!- | vandebo [~vandebo@abydos.nerdbox.net] has joined #uml |
| 13:52 | <huslu:#uml> | jdike: any info now possible to get out of the kernel binary that has all debug info in? |
| 13:52 | <huslu:#uml> | i mean during the crash that is. |
| 13:52 | <jdike:#uml> | nothing comes to mind |
| 13:53 | <jdike:#uml> | slab debugging was my best hope |
| 13:53 | <huslu:#uml> | gdb won't help? |
| 13:53 | <jdike:#uml> | do you want to step through libvde and see where it's smashing memory? |
| 13:53 | <jdike:#uml> | coz that'll work if you feel like it |
| 13:55 | <jdike:#uml> | the other thing to do is just look at the code, make sure there's a 1-1 correspondence between allocs and frees and that there are no buffer overruns |
| 13:55 | <jdike:#uml> | and no chasing stray pointers |
| 13:55 | <huslu:#uml> | hm, that would be like asking three year old to troubleshoot a car engine. (at least if you ask me) |
| 13:55 | <jdike:#uml> | hehe |
| 13:56 | <jdike:#uml> | give junior the screwdriver and stand back |
| 13:56 | <jdike:#uml> | another thing you might try is build something else against libvde and run it under valgrind |
| 13:57 | <jdike:#uml> | a normal process-type thing |
| 13:59 | <huslu:#uml> | there was some trick you gould ld preload vdetaplib and then if an app wants to use a tap it gets a place in vde switch or something. i wonder if that would crash. |
| 14:01 | <huslu:#uml> | so the gdb thing that i did earlier would not help now that linux kernel has debug info in? (altho it crashes right in the beginning)? |
| 14:02 | <jdike:#uml> | I wanted to see if gdb showed anything on the stack that the in-kernel stack tracer didn't |
| 14:05 | -!- | aroscha [~aroscha@chello213047053193.30.11.tuwien.teleweb.at] has joined #uml |
| 14:07 | <huslu:#uml> | i didn't have debug info compiled in when i was using gdb earlier, so now it has debug info in, but it crashes right in the beginning |
| 14:08 | <huslu:#uml> | i don't know how to get gdb into mix since i can't see the process id so fast as to use it before the crash |
| 14:09 | <huslu:#uml> | either way, if there's no point of trying gdb now, that would make sense too |
| 14:28 | <Magotari:#uml> | Now, I'm not an uml expert, and I have no idea how to gdb it, but when I was debugging a normal program all I had to do is to point gdb at the executable. |
| 14:28 | <Magotari:#uml> | No need for process id then. |
| 14:57 | -!- | ram [~ram@pool-71-245-96-80.nycmny.fios.verizon.net] has quit [Ping timeout: 480 seconds] |
| 15:01 | <jdike:#uml> | exactly |
| 15:37 | <huslu:#uml> | jdike: http://pastebin.ca/1082152 if that helps any. i don't know what i was doing anyways |
| 15:49 | <jdike:#uml> | No, not really |
| 15:50 | <jdike:#uml> | probably the best thing you could do to track this down is to valgrind some other thing which uses libvde |
| 15:51 | <jdike:#uml> | doesn't require any ability to deal with code |
| 15:51 | -!- | low [~low@low.ath.cx] has joined #uml |
| 16:07 | -!- | ram [~ram@bi01p1.co.us.ibm.com] has joined #uml |
| 16:31 | -!- | aroscha [~aroscha@chello213047053193.30.11.tuwien.teleweb.at] has quit [Quit: aroscha] |
| 17:12 | -!- | dang [~dang@aa-redwall.ghs.com] has quit [Quit: Leaving.] |
| 17:31 | -!- | low [~low@low.ath.cx] has quit [Quit: bbl] |
| 19:08 | -!- | Basic [~Basic@gatekeeper.real-time.com] has quit [Quit: Linghu - Level 70 Hunter - Executus US] |
| 19:49 | -!- | hfb [~hfb@pool-71-118-254-245.lsanca.dsl-w.verizon.net] has quit [Quit: Leaving] |
| 21:25 | -!- | jdike [~jdike@pool-72-70-38-200.bstnma.fios.verizon.net] has quit [Quit: Leaving] |
| 21:55 | -!- | ram [~ram@bi01p1.co.us.ibm.com] has quit [Ping timeout: 480 seconds] |
| 23:59 | -!- | VS_ChanLog [~stats@ns.theshore.net] has left #uml [Rotating Logs] |
| 23:59 | -!- | VS_ChanLog [~stats@ns.theshore.net] has joined #uml |
| --- | Log | closed Fri Jul 25 00:00:10 2008 |
![[FSF Associate Member]](/assets/fsf.png){kind=small}
