Back to Home / #uml / 2007 / 12 / Prev Day | Next Day
#uml IRC Logs for 2007-12-05

---Logopened Wed Dec 05 00:00:37 2007
---Daychanged Wed Dec 05 2007
00:00|-|dang [~dang@nemesis.fprintf.net] has quit [Ping timeout: 480 seconds]
03:04|-|balbir [~balbir@122.167.177.89] has quit [Ping timeout: 480 seconds]
03:15|-|balbir [~balbir@122.167.180.62] has joined #uml
03:16|-|aroscha [~aroscha@chello213047053193.30.11.tuwien.teleweb.at] has joined #uml
03:49|-|Ancalagon [~PtitKosmi@215.208-78-194.adsl-fix.skynet.be] has joined #uml
04:10|-|aroscha [~aroscha@chello213047053193.30.11.tuwien.teleweb.at] has quit [Quit: aroscha]
04:12|-|ftumch [~James@james.1ec.aaisp.net.uk] has joined #uml
04:45|-|ftumch [~James@james.1ec.aaisp.net.uk] has quit [Quit: Goodbye.]
04:45|-|ftumch [~James@james.1ec.aaisp.net.uk] has joined #uml
04:46|-|Ancalagon [~PtitKosmi@215.208-78-194.adsl-fix.skynet.be] has quit [Quit: ChatZilla 0.9.79 [Firefox 2.0.0.8/2007100400]]
06:56|-|IntuitiveNipple [~TJ@alexandros.tjworld.net] has joined #uml
08:22|-|dgraves_ [~asdf@inet-netcache3-o.oracle.com] has quit [Ping timeout: 480 seconds]
08:59|-|dang [~dang@aa-redwall.nexthop.com] has joined #uml
10:25|-|jdike [~jdike@pool-71-248-190-161.bstnma.fios.verizon.net] has joined #uml
10:25<jdike-#uml->>Hi guys
10:29[~]linbot #uml dispenses breakfast#uml-> dispenses breakfast
10:30[~]jdike #uml runs#uml-> runs
10:37<dgraves-#uml->>morning jdike. :)
10:37<dgraves-#uml->>runnig already? :)
10:37<jdike-#uml->>how's it going?
10:38<jdike-#uml->>yeah, getting jumpy
10:38<dgraves-#uml->>pretty good. i'm starting from the top today again. Gonna be more careful with the quilt patches, went to back out some after i hit an error, and they didn't back out cleanly.
10:38<dgraves-#uml->>i *think* its something in the file system patches. I know that RH changes the file ops in 2.6.18, but they didn't make vanilla till like 2.6.19.
10:39<dgraves-#uml->>i got through... 200 patches before that happened though?
10:40<dgraves-#uml->>how goes it with you?
10:42<jdike-#uml->>OK
10:42<jdike-#uml->>updating to the latest mm atm
10:50<jdike-#uml->>now let's see how well it builds
10:51<jdike-#uml->>wow
10:52<jdike-#uml->>someone changed all the field names in sigcontext
10:56<jdike-#uml->>fix that and things look OK
11:10<dgraves-#uml->>::L::
11:10<dgraves-#uml->>they had nothing better to do? :)
11:11<jdike-#uml->>x86 unification
11:13<dgraves-#uml->>oh, irght.
11:13<dgraves-#uml->>i remember reading about that.
11:13<dgraves-#uml->>will that affect uml much? Make it easier, perhaps?
11:13<jdike-#uml->>right now, it's just causing trouble
11:17<dgraves-#uml->>:(
11:19<jdike-#uml->>let's see what the x86_64 side of things looks like
11:22<jdike-#uml->>same thing
11:31<jdike-#uml->>patch off to akpm
11:57|-|hfb [~hfb@pool-71-106-219-180.lsanca.dsl-w.verizon.net] has joined #uml
11:59[~]jdike #uml bisects -mm again#uml-> bisects -mm again
12:01|-|balbir [~balbir@122.167.180.62] has quit [Read error: Operation timed out]
12:14|-|balbir [~balbir@122.167.196.19] has joined #uml
12:18|-|ram [~ram@pool-96-225-204-220.ptldor.fios.verizon.net] has joined #uml
12:32|-|Marquel [~Marquel@port-2766.pppoe.wtnet.de] has joined #uml
12:32<Marquel-#uml->>good evening...
12:39<jdike-#uml->>didn't have to bisect anything - I broke the filesystem
12:58<Marquel-#uml->>can someone help me building a uml-kernel from source? linux-2.6.23 and -2.6.24 are failing with "asm/user.h: File not found."
13:01<karol-#uml->>Marquel: on the phone. be with you within an hour.
13:07<dgraves-#uml->>Marquel, what was your build command line?
13:08<Marquel-#uml->>make linux ARCH=um
13:12<dgraves-#uml->>Marquel, did you do a 'make ARCH=um mrproper defconfig' first?
13:12<jdike-#uml->>current 2.6.24-rc has that fixed
13:13<Marquel-#uml->>which is current rc?
13:13<Marquel-#uml->>defconfig?
13:15<jdike-#uml->>2.6.24-rc4 right now
13:15<Marquel-#uml->>so 2.6.22/23 has a problem?
13:16<jdike-#uml->>more a misunderstanding :-)
13:16<Marquel-#uml->>huh?
13:17<jdike-#uml->>I used to include libc's asm/page.h and asm/user.h, but they're not really intended for processes to include
13:17<jdike-#uml->>so distros have been getting rid of them
13:17<jdike-#uml->>and I've been making UML not include them
13:17<Marquel-#uml->>i guess there's no patch for 2.6.23 then?
13:18<jdike-#uml->>there are
13:19<Marquel-#uml->>*getting big eyes looking around* wherewherewhere? *sniffing like a dog* ;)
13:19<jdike-#uml->>hehe
13:19[~]dgraves #uml throws marquel a bone. :)#uml-> throws marquel a bone. :)
13:19<jdike-#uml->>dgraves, can you show him the patches?
13:19<Marquel-#uml->>*catch*
13:19<dgraves-#uml->>::ROFL::
13:19<dgraves-#uml->>that would be the git commits, yes, jdike?
13:20<jdike-#uml->>yeah
13:20<jdike-#uml->>you have them in one place
13:20<jdike-#uml->>I'd have to go digging them up again
13:20<dgraves-#uml->><jdike> it's commit 71f926f2ea61994470a53c9e11d3ef993197cada in mainline
13:20<dgraves-#uml->><jdike> and ee1eca5d2493026affbbc91b228dd00879484687
13:20|-|kokoko1 [~Slacker@203.148.65.19] has joined #uml
13:20<jdike-#uml->>that's asm/page.h
13:21|-|krau [~cktakahas@200.184.118.132] has quit [Quit: Varei!!!]
13:21<jdike-#uml->>you should have something for asm/user.h too
13:21|-|krau [~cktakahas@200.184.118.132] has joined #uml
13:21<kokoko1-#uml->>Hello jdike
13:22<jdike-#uml->>hey dude
13:22[~]dgraves #uml doesn't recall user.h.#uml-> doesn't recall user.h.
13:22<dgraves-#uml->>i don't think i do. i think EL5 got rid of that...
13:22<dgraves-#uml->>the only other patches I have from you are the utrace ones.
13:23<jdike-#uml->>hmm, OK
13:23<jdike-#uml->>hold on
13:24<kokoko1-#uml->>dgraves, long time no see :)
13:25<Marquel-#uml->>so to sum it up whether i like it or not i either have to use 2.6.24-rc4 or get git to patch my 2.6.23?
13:27<jdike-#uml->>asm/user.h removal is commit 189872f968def833727b6bfef83ebd7440c538e6
13:27<jdike-#uml->>you can get them from kernel.org, I think
13:28<Marquel-#uml->>i think so too, i just don't like the idea of running my single-point-of-internet-connection-failure on a release candidate ;)
13:28<jdike-#uml->>OK - these are fairly low-risk patches, so applying them to 2.6.23 is probably your best bet
13:29<Marquel-#uml->>there's no way to download them as a unified diff, is there?
13:32<jdike-#uml->>http://rafb.net/p/PawRam38.txt
13:33<dgraves-#uml->>kokoko1, yeah, how's it going?
13:33<Marquel-#uml->>jdike: where do i send the cake?
13:33<dgraves-#uml->>jdike, I'm making a note of that. :)
13:34<jdike-#uml->>http://rafb.net/p/vfWrQI62.txt
13:34<kokoko1-#uml->>dgraves, fine thanks
13:34<dgraves-#uml->>excellent!
13:34<dgraves-#uml->>you're missing all the fun!
13:34<jdike-#uml->>http://rafb.net/p/LUCs1r64.txt
13:34<dgraves-#uml->>i'm trying to make a UML EL5 work. :)
13:34<jdike-#uml->>those should do it
13:35<kokoko1-#uml->>ahaha sure it sound lot of fun. ;)
13:36<kokoko1-#uml->>here i am waiting for daring fsck to finish
13:36<kokoko1-#uml->>var has gone 220 days without being checked, check forced
13:36<jdike-#uml->>it sounds more mechanical right now
13:36<dgraves-#uml->>i see. you only come see us when you're bored, kokoko1. :)
13:38<kokoko1-#uml->>oh not really
13:38[~]dgraves #uml knows all about mechanical parts of uml, eh, jdike? :)#uml-> knows all about mechanical parts of uml, eh, jdike? :)
13:38[~]kokoko1 #uml fsck is finished #uml-> fsck is finished
13:39<jdike-#uml->>you do now
13:40<dgraves-#uml->>::L::
13:46|-|tyler29 [~tyler@ARennes-257-1-71-181.w81-53.abo.wanadoo.fr] has joined #uml
13:48<karol-#uml->>Marquel: I see everyone else beat with helping you by the time I finished my talk. :)
13:49<Marquel-#uml->>karol: which doesn't mean i wouldn't be grateful for your effort ;)
13:49<karol-#uml->>I try to help everyone who comes here because my coding ain't that great yet. So if you have a problem feel free to come here and ask, I will try to help as best as I can.
13:50<dgraves-#uml->>jdike, looks like the old guard of kokoko1, me and the_hydra has changed into karol and caker. :)
13:51<jdike-#uml->>he hypothetically would be grateful for the help you hypothetically would have provided
13:51<dgraves-#uml->>::ROFL::
13:51<dgraves-#uml->>i'm too busy making mad chedda to help. ;)
13:51<dgraves-#uml->>and by making mad chedda, i mean, of course, disembowling uml. :(
13:52<karol-#uml->>jdike: Today during debugging the stack crash thingy, I found something interesting. I am not sure of my findings yet, but... It seems like the thing switch_threads to one thing and then does it backwards. I kinda got the impression like it only puts the printed task into context for a split second.
13:53<jdike-#uml->>you're right
13:53<karol-#uml->>Right. Thanks for confirming.
13:53<jdike-#uml->>it's done completely under the covers, hidden from the main scheduler
13:53<karol-#uml->>Yup, I noticed that...
13:53<karol-#uml->>Ok, so my sanity check is complete.
13:53<jdike-#uml->>I thought it was a cute trick
13:54<karol-#uml->>I guess you put a breakpoint to debug something with stack?
13:54<dgraves-#uml->>:)
13:54<dgraves-#uml->>you're smart, karol.
13:54<jdike-#uml->>right
13:54<karol-#uml->>Yeah, it is a pretty cute trick, it just bothers me how it's real purpose ain't documented.
13:54<dgraves-#uml->>jdike had to explain that to me, and I wrote it. :)
13:55<karol-#uml->>I would separate the printing of the stack and switching to a task into two things.
13:55<karol-#uml->>I guess I should document the stack's other use then, because it was a real surprise to discover that you use it for debugging.
13:56<karol-#uml->>I will leave the splitting for my own entertainment and ease of debugging.
13:57<jdike-#uml->>well, it made sense for it to do something
13:57<jdike-#uml->>and having it dump a stack seemed reasonable
13:59[~]dgraves #uml would be interested to see karol's patch. :)#uml-> would be interested to see karol's patch. :)
13:59<jdike-#uml->>anyone have a box with a non-standard VM split and want to test a little proggie?
13:59<karol-#uml->>Yup.
13:59<karol-#uml->>I'm on a strange memsplit.
14:00<karol-#uml->>(and I hate defconfig sometimes)
14:00<jdike-#uml->>compile and run http://rafb.net/p/J57Beb68.txt
14:00<karol-#uml->>Gotcha.
14:00<jdike-#uml->>and tell me what it says
14:00<jdike-#uml->>and also tell me if it's right :-)
14:01<karol-#uml->>Address space top 0xb0000000
14:01<karol-#uml->>Let me see if it is right now.
14:01<karol-#uml->>Runtime memsplit discovery?
14:01<jdike-#uml->>yup
14:01<karol-#uml->>Wow, what a godsend.
14:02<karol-#uml->>Thank you thank you thank you thank you thank you.
14:02<jdike-#uml->>need to test it on x86_64
14:02[~]dgraves #uml only has 1 box x86_64, and its a normal split.#uml-> only has 1 box x86_64, and its a normal split.
14:03<karol-#uml->>*sigh* I need to find a garbage dump with computers one day... Lack of machines has been biting me recently.
14:03<jdike-#uml->>x86_64 doesn't have odd VMSPLITs
14:03<jdike-#uml->>only i386
14:03<jdike-#uml->>and it doesn't work on x86_64
14:04<jdike-#uml->>I don't think it matters
14:04<jdike-#uml->>I can make it i386-only
14:04<dgraves-#uml->>karol, i should have given you some. my wife made me get rid of some. :)
14:04<karol-#uml->>jdike: Yes, the thing is consistent with UML config.
14:04<karol-#uml->>Want me to try a few more memsplits?
14:04<jdike-#uml->>sure
14:05<karol-#uml->>Ok. This will require reboots to this machine, so it will take a while.
14:05<jdike-#uml->>np
14:05<karol-#uml->>And new kernels. I love ccache.
14:05<jdike-#uml->>play with it when it's convenient and let me know if it messed up
14:06[~]dgraves #uml supposes we'll see karol bouncing for a bit.#uml-> supposes we'll see karol bouncing for a bit.
14:06<dgraves-#uml->>just run vmware, karol! you can test uml inside it. :)
14:06<karol-#uml->>Not open source. :)
14:06[~]dgraves #uml wonders about running uml to test the new kernel split.#uml-> wonders about running uml to test the new kernel split.
14:06<dgraves-#uml->>karol, its free. :)
14:06<dgraves-#uml->>and some of it is.
14:06<karol-#uml->>You have a OSS zealot on your hands here.
14:06<dgraves-#uml->>::ROFL::
14:07<dgraves-#uml->>i could offer you xen, from Oracle? :)
14:07<karol-#uml->>I don't like Xen, I want my host not to be a virtual machine.
14:07<karol-#uml->>Picky, ain't I? :)
14:07<dgraves-#uml->>yeah, that's a disadvantage.
14:08<dgraves-#uml->>oh, jdike, i wrote the EL guys at Oracle the other day, asking for UML support in the EL kernel...
14:08<jdike-#uml->>hehe
14:08<dgraves-#uml->>they didn't sound to open to the idea, suggesting that I use xen instead.
14:08<jdike-#uml->>cool
14:08<dgraves-#uml->>I'm trying. :)
14:09<dgraves-#uml->>i've talked with some of my group, and they're not thrilled with the idea either.
14:09<dgraves-#uml->>same thing that karol said, we don't like our host being a virtual box.
14:09<jdike-#uml->>which idea?
14:09<dgraves-#uml->>moving to xen.
14:09<jdike-#uml->>yeah
14:09<dgraves-#uml->>they like uml.
14:10|-|kos_tom [~thomas@col31-3-82-247-183-72.fbx.proxad.net] has joined #uml
14:10<dgraves-#uml->>so it may be a fight for our group. :)
14:10<dgraves-#uml->>i guess it depends on how much time they want me to take building kernels. :)
14:10[~]dgraves #uml is getting good at this.#uml-> is getting good at this.
14:10<dgraves-#uml->>in 30 min, i've replicated my work from most of yesterday.
14:10<dgraves-#uml->>and its better done with quilt correctly applied. :)
14:12<karol-#uml->>2.6.23-rc7... Wow. Very damn stable. I have been using it all along, never had a problem. I really like the 23 kernel line.
14:12<karol-#uml->>Building a new rc7 kernel now.
14:12|-|tyler29 [~tyler@ARennes-257-1-71-181.w81-53.abo.wanadoo.fr] has quit [Ping timeout: 480 seconds]
14:17<dgraves-#uml->>jdike, , i will say that oracle is cool that way. I can just email VP's out of left field, and they'll answer back.
14:21|-|Magotari [~karol@chello089076073248.chello.pl] has joined #uml
14:21|-|karol [~karol@chello089076073248.chello.pl] has quit [Quit: leaving]
14:21<Magotari-#uml->>Thankfully I do have a second computer.
14:22[~]dgraves #uml nukes Magotari 's second machine with boston creme pies.#uml-> nukes Magotari 's second machine with boston creme pies.
14:22<dgraves-#uml->>not anymore you don't! :)
14:27<Magotari-#uml->>jdike: For 2/2G optimal I got 0x78000000
14:28<Marquel-#uml->>*retrying to build*
14:29<Magotari-#uml->>I don't even think UML supports that memsplit, so can't confirm there.
14:29|-|tyler29 [~tyler@ARennes-257-1-176-171.w86-214.abo.wanadoo.fr] has joined #uml
14:29<Magotari-#uml->>Building a 1/3 now.
14:35<Marquel-#uml->>jdike: erm... it seems i need a patch specifically for arch/um/sys-i386/ptrace_user.c, which still references asm/user.h. is it sufficient to just drop the #include or do i need something more?
14:36<Magotari-#uml->>1/3 gave me: 0x40000000. This is consistent with UML's config.
14:37<Magotari-#uml->>Marquel: Just try dropping it, might work. Worked for me a few times. Not sure if with .24 or .23.
14:37<Magotari-#uml->>Just comment it out for now.
14:37<Marquel-#uml->>.23
14:38<Marquel-#uml->>compiles... *narf*
14:39<jdike-#uml->>Magotari, it does
14:39<jdike-#uml->>didn't, but I fixed that, and I thought you were the one complaining about it
14:41<jdike-#uml->>Marquel, it does?
14:41<jdike-#uml->>it doesn't now and I don't see any patches to it which changed that
14:42<Marquel-#uml->>jdike: after commenting #include <asm/user.h> in arch/um/sys-i386/ptrace_user.c it compiles so far.
14:42|-|ferret_0567 [~travis@72.191.26.86] has joined #uml
14:42<Marquel-#uml->>not yet completed, but still working on the problem.
14:42<Magotari-#uml->>jdike: I still need to test normal memsplit, but phone time is here.
14:43<jdike-#uml->>found it
14:43<jdike-#uml->>removed asm/user.h from there as part of a more general cleanup
14:43<jdike-#uml->>Just deleting the include should be fine
14:44<ferret_0567-#uml->>I can't get UML to load a gzipped cpio archive for a initrd
14:44<ferret_0567-#uml->>CONFIG_BLK_DEV_RAM=y, CONFIG_BLK_DEV_INITRD=y
14:44<jdike-#uml->>what's the command line?
14:45<ferret_0567-#uml->>[42949372.960000] Kernel command line: mem=256M ubda=/EXTRASTORAGE/SMGL-root.ext2-loopmount root=/dev/ubda eth0=tuntap,,,192.168.1.105 initrd=/boot/initrd-smgl-no-udev root=98:0
14:45<Marquel-#uml->>.oO(and all this just to get IPv6 into my home-lan.... i must be ... mad or something...)
14:45<caker-#uml->>ferret_0567: UML wants the initrd to exist on the host filesystem
14:45<jdike-#uml->>no it doesn't
14:45<ferret_0567-#uml->>caker: it does
14:45<jdike-#uml->>whoops
14:46<jdike-#uml->>caker's right
14:46[~]caker #uml bonks jdike #uml-> bonks jdike
14:46<jdike-#uml->>ow
14:46<jdike-#uml->>any boot output relating to the initrd?
14:47<ferret_0567-#uml->>no
14:48<ferret_0567-#uml->>I do get this though: [42949373.510000] RAMDISK driver initialized: 16 RAM disks of 4096K size 1024 blocksize
14:48<jdike-#uml->>can you paste the boot output someplace?
14:48<ferret_0567-#uml->>yes
14:49|-|tyler29 [~tyler@ARennes-257-1-176-171.w86-214.abo.wanadoo.fr] has quit [Ping timeout: 480 seconds]
14:50<dgraves-#uml->>jdike, i didn't get caught on asm/user.h, since i got narfed on some other header issues, so i munged with what was in 2.6.23.9 till it worked in my tree.
14:50<ferret_0567-#uml->>http://rafb.net/p/1LN3Jl33.txt
14:52<jdike-#uml->>OK, how do you feel about gdb-ing this?
14:54<IntuitiveNipple-#uml->>ferret_0567: What does "ls -l /boot/initrd-smgl-no-udev" report?
14:54<ferret_0567-#uml->>-rw-r--r-- 1 root root 2634128 2007-12-05 00:34 /boot/initrd-smgl-no-udev
14:55<IntuitiveNipple-#uml->>size looks okay
14:55<dgraves-#uml->>ferret_0567, are you root?
14:55<ferret_0567-#uml->>dgraves: no
14:56<ferret_0567-#uml->>I started the UML as my user, travis
14:56<jdike-#uml->>dgraves is right on top of permissions now
14:57<ferret_0567-#uml->>let me start the UML as root
14:57<dgraves-#uml->>look, i know its only supposed to be reading from it, but how about changing the perms on it? :)
14:57<ferret_0567-#uml->>ok...
14:58<dgraves-#uml->>or do uml as root. either way. :)
14:59<ferret_0567-#uml->>still no change
14:59<ferret_0567-#uml->>with -rw-rw-rw-
15:00<ferret_0567-#uml->>trying again with 777
15:00<ferret_0567-#uml->>nope...ok, I'll try with root
15:02<ferret_0567-#uml->>even with root priviledges it still doesn't work
15:04<IntuitiveNipple-#uml->>ferret_0567: run a 2nd console, with "inotifywait -m /boot" then start uml, and then check the inotifywait to ensure the initrd file is being accessed
15:05<jdike-#uml->>or just strace UML
15:06<ferret_0567-#uml->>it's not even being accessed
15:06<IntuitiveNipple-#uml->>Yeah... strace will throw a lot out though, inotify will just watch the file he's having issues with
15:06<jdike-#uml->>strace -e open will be pretty specific
15:06|-|tyler29 [~tyler@ARennes-257-1-153-182.w86-214.abo.wanadoo.fr] has joined #uml
15:12[~]dgraves #uml isn't sure to bet on inotifywait or strace in this fight.#uml-> isn't sure to bet on inotifywait or strace in this fight.
15:13<dgraves-#uml->>okay, jdike. 95 patches in again. :)
15:13<dgraves-#uml->>was right around this point where it failed yesterday. :-0
15:16[~]dgraves #uml is about to change the inode size files. or as the patch says, inode_diet.#uml-> is about to change the inode size files. or as the patch says, inode_diet.
15:18<jdike-#uml->>that shouldn't affect UML
15:21<dgraves-#uml->>jdike, yeah, i know. but they moved some of the ops around, and stuff too, i think. again, it shouldn't break uml, but this is around where it broke the other day.
15:25<ferret_0567-#uml->>I serios
15:26<ferret_0567-#uml->>I seriously don't think anything is wrong with my kernel config, but, I can't get the initrd to load
15:27<jdike-#uml->>want to gdb it?
15:27<ferret_0567-#uml->>me?
15:28<jdike-#uml->>yup
15:28<ferret_0567-#uml->>I don't know how to gdb the kernel
15:28<IntuitiveNipple-#uml->>you could put a breakpoint on read_initrd()
15:29<jdike-#uml->>yup
15:29<IntuitiveNipple-#uml->>The code is in arch/um/kernel/initrd.c
15:30<jdike-#uml->>and uml_initrd_setup
15:30<ferret_0567-#uml->>I don't think I have the debugging symbols in my kernel
15:30<jdike-#uml->>that's under Kernel Hacking
15:30<ferret_0567-#uml->>I know
15:38<ferret_0567-#uml->>I selected "Compile the kernel with debug info", and I will not run strip on the resulting linux kernel
15:39<jdike-#uml->>OK
15:42<Marquel-#uml->>jdike: it compiled now. :)
15:42<jdike-#uml->>good
15:42<Marquel-#uml->>now i have to go away, need to reboot the router ;)
15:43<caker-#uml->>UML hangs early in boot if the initrd is zero length
15:43<jdike-#uml->>got a stack?
15:51<ferret_0567-#uml->>the kernel just finished compiling
15:54<jdike-#uml->>Locating the top of the address space ... 0xffc00000
15:54<jdike-#uml->>cool
15:57|-|Marquel [~Marquel@port-2766.pppoe.wtnet.de] has quit [Ping timeout: 480 seconds]
15:59<ferret_0567-#uml->>Function "read_initrd()" not defined.
15:59<ferret_0567-#uml->>Make breakpoint pending on future shared library load? (y or [n]) y
16:00<ferret_0567-#uml->>also made a breakpoint on uml_initrd_setup()
16:01<ferret_0567-#uml->>Error accessing memory address 0x60011000: Input/output error.
16:02<IntuitiveNipple-#uml->>read_intitrd() should be compiled in
16:03<ferret_0567-#uml->>I didn't run the program yet when I made that breakpoint
16:03<jdike-#uml->>did it ask the same about uml_initrd_setup?
16:03<ferret_0567-#uml->>yes
16:03<IntuitiveNipple-#uml->>how did you start gdb?
16:03<IntuitiveNipple-#uml->>gdb --args linux mem= .... ?
16:03<ferret_0567-#uml->>yes
16:04<jdike-#uml->>somehow, the initrd stuff isn't there
16:04<IntuitiveNipple-#uml->>and you're sure the exectuable you are starting, is the same one you just built?
16:04<jdike-#uml->>Locating the top of the address space ... 0xc0000000
16:04<jdike-#uml->>and that works a bit better
16:04<ferret_0567-#uml->>IntuitiveNipple: yes
16:05<jdike-#uml->>look at your build log and see if arch/um/kernel/initrd.o is built
16:05<IntuitiveNipple-#uml->>what does linux.uml-2.6.23.9 --showconfig | grep -i initrd show?
16:05<ferret_0567-#uml->>jdike: sorry, I don't have a build log :(
16:05<jdike-#uml->>why not?
16:05<jdike-#uml->>you shove it all down /dev/null?
16:05<ferret_0567-#uml->>no
16:06<ferret_0567-#uml->>I just didn't put the output to a file
16:06<jdike-#uml->>you have it in an xterm?
16:06<IntuitiveNipple-#uml->>$ ./linux01 --showconfig | grep -i initrd
16:06<IntuitiveNipple-#uml->># CONFIG_BLK_DEV_INITRD is not set
16:06<jdike-#uml->>Ho
16:06<ferret_0567-#uml->>CONFIG_BLK_DEV_INITRD=y
16:07<jdike-#uml->>whoops
16:07<IntuitiveNipple-#uml->>lol
16:07<IntuitiveNipple-#uml->>We don't look that alike do we?
16:07<ferret_0567-#uml->>nope
16:07[~]jdike #uml gets excited easily#uml-> gets excited easily
16:07[~]IntuitiveNipple #uml backs away#uml-> backs away
16:07<jdike-#uml->>HAHA
16:08<IntuitiveNipple-#uml->>ok, so, check for that initrd.o :)
16:08<jdike-#uml->>touch arch/um/kernel/initrd.c and rebuild, and see if it gets compiled
16:08<jdike-#uml->>or just do what IntuitiveNipple says
16:08<ferret_0567-#uml->>/EXTRASTORAGE/linux-2.6.23.9/arch/um/kernel/initrd.o
16:08<jdike-#uml->>and I can't believe I just said that
16:09<ferret_0567-#uml->>and, there is a initrd.c for ARCH um
16:09<IntuitiveNipple-#uml->>before you built this all, did you do a clean?
16:09<jdike-#uml->>that's going to show up one day, completely out of context
16:09<ferret_0567-#uml->>IntuitiveNipple: I did a make mrproper, for sure
16:09<IntuitiveNipple-#uml->>ok
16:09<ferret_0567-#uml->>it wouldn't build without doing that
16:10<IntuitiveNipple-#uml->>just making sure we're not talking at cross purposes
16:10<ferret_0567-#uml->>make mrproper and make mrproper ARCH=um
16:11<IntuitiveNipple-#uml->>what does " grep -i initrd System.map" report?
16:13<ferret_0567-#uml->>http://rafb.net/p/bHNDDY46.html
16:15<jdike-#uml->>everything's there
16:15<jdike-#uml->>why doesn't gdb know about them
16:15|-|Marquel [~Marquel@port-108.pppoe.wtnet.de] has joined #uml
16:15<ferret_0567-#uml->>should I run the program before setting a breakpoint or not?
16:16<jdike-#uml->>no
16:16<ferret_0567-#uml->>simple:
16:16<jdike-#uml->>doesn't matter
16:16<ferret_0567-#uml->>don't use any "()"'s when defining the breakpoint
16:16<jdike-#uml->>wait
16:17<ferret_0567-#uml->>I just found that out
16:17<IntuitiveNipple-#uml->>lol
16:17<jdike-#uml->>I was just looking at that
16:17<IntuitiveNipple-#uml->>ha! I always type ()'s to be clear it is a function name, doesn't mean you need to use them :)
16:17<caker-#uml->>IntuitiveNipple()
16:18<ferret_0567-#uml->>(gdb) br read_initrd
16:18<ferret_0567-#uml->>Breakpoint 2 at 0x6000638d: file arch/um/kernel/initrd.c, line 19.
16:18<ferret_0567-#uml->>I setup another one at uml_initrd_setup
16:19<ferret_0567-#uml->>run:... Cannot insert breakpoint -3.\nError accessing memory address 0x60011000: Input/output error.
16:19<ferret_0567-#uml->>do I need to run gdb as root?
16:19<jdike-#uml->>no
16:20<jdike-#uml->>read_initrd should be OK
16:21<IntuitiveNipple-#uml->>do 'start' then 'continue'
16:22<IntuitiveNipple-#uml->>that should then stop at the entry to read_initrd()
16:23<jdike-#uml->>what does start do?
16:23<jdike-#uml->>that's useful
16:23<IntuitiveNipple-#uml->>takes it to the beginning of main
16:24<jdike-#uml->>much better than my technique of 'b main; run'
16:24<IntuitiveNipple-#uml->>:)
16:24<ferret_0567-#uml->>I did "start" and then "continue" and it still gives the same error
16:24<IntuitiveNipple-#uml->>back out, quit gdb, start again
16:25<IntuitiveNipple-#uml->>that way you know it's a clean slate
16:25<ferret_0567-#uml->>ok
16:25<IntuitiveNipple-#uml->>gdb --args etc.
16:27<ferret_0567-#uml->>ok, I started gdb again, set a breakpoint at read_initrd, and then started the program using run, then I did start; continue
16:28<IntuitiveNipple-#uml->>you shouldn't need 'run' if using 'start'
16:28<IntuitiveNipple-#uml->>'start' will execute the initialisation routines and then break at the entry to main(), from where you can do 'continue' to go until the breakpoint(s) are hit
16:29<IntuitiveNipple-#uml->>'run' will take you to the first breakpoint, or should
16:30<ferret_0567-#uml->>well, I'm ending up having to type continue over and over again, each time getting "Cannot insert breakpoint -3.\nError accessing memory address 0x60011000: Input/output error."
16:30<jdike-#uml->>what breakpoint is that?
16:30<IntuitiveNipple-#uml->>That address doesn't seem 'right' having looked at the System.map
16:31<jdike-#uml->>if it's uml_initrd_setup, then just delete it
16:31<jdike-#uml->>d 3
16:32<ferret_0567-#uml->>http://rafb.net/p/9owSaT80.html
16:32<jdike-#uml->>can you run with those two?
16:35<ferret_0567-#uml->>it doesn't really go too far with those two when I execute the run command
16:35<jdike-#uml->>same errors?
16:36<ferret_0567-#uml->>no
16:36<jdike-#uml->>what happens?
16:36<ferret_0567-#uml->>not with the run command, but with the continue command
16:36[~]dgraves #uml is near (w\in 2 patches of the bad patch that makes it panic!#uml-> is near (w\in 2 patches of the bad patch that makes it panic!
16:36<IntuitiveNipple-#uml->>When you use run, does it stop at read_initrd ?
16:37<ferret_0567-#uml->>nope
16:40<ferret_0567-#uml->>http://rafb.net/p/by8aO082.html
16:40<jdike-#uml->>just say YES
16:41<ferret_0567-#uml->>ok, I did that
16:41<ferret_0567-#uml->>same error :(
16:41<jdike-#uml->>paste it
16:42<ferret_0567-#uml->>http://rafb.net/p/4VsHuN76.html
16:42<jdike-#uml->>so you're not trying to break on uml_initrd_setup at all?
16:43<jdike-#uml->>and it's giving you that?
16:43<ferret_0567-#uml->>yes
16:43<jdike-#uml->>do a 'bt' at this point
16:45<IntuitiveNipple-#uml->>that address, 60011000 looks to be at the border of the t and T sections
16:46<jdike-#uml->>I thought it was suspiciously round
16:46<IntuitiveNipple-#uml->>Can you search for that address with grep in the System.map?
16:47<IntuitiveNipple-#uml->>Or, even better, open System.map with a text editor and find the bit that covers that address and paste the addresses/functions around that spot for us
16:48<IntuitiveNipple-#uml->>What he needs do is install AjaxTerm on there, so we can play via the web :)
16:49<jdike-#uml->>I had a gdbbot at one point which would allow someone to control gdb via IRC
16:49[~]dgraves #uml is intrigued by this. #uml-> is intrigued by this.
16:49<dgraves-#uml->>its like a live debugging tutorial! :)
16:50<IntuitiveNipple-#uml->>lol
16:50<IntuitiveNipple-#uml->>Well, I'd like one where we could watch a session as it happens, even if not control it
16:51<IntuitiveNipple-#uml->>Remote blind debugging is just a hassle, things so easy to misunderstand
16:51<IntuitiveNipple-#uml->>Ha! VNC :p
16:51<IntuitiveNipple-#uml->>messing about with pastebins, etc, grrrr
16:52<dgraves-#uml->>IntuitiveNipple, i've done vnc debugging a bunch, frankly.
16:52<jdike-#uml->>http://rafb.net/p/vESyyB59.txt
16:53<IntuitiveNipple-#uml->>"Vlad the Debugger" lol!
16:55<dgraves-#uml->>::LOL::
16:57<jdike-#uml->> /j #umldebug
16:58|-|dang [~dang@aa-redwall.nexthop.com] has quit [Quit: Leaving.]
16:59<dgraves-#uml->>jdike, you'll be happy to know I caught one of the offending patches that makes uml go splat.
17:00<dgraves-#uml->>the reiserfs superblock patch!
17:04<jdike-#uml->>ferret_0567, want one of us to play with your gdb?
17:05<dgraves-#uml->>and by one of us, jdike is volunteering. :)
17:05<jdike-#uml->>hehe
17:06<Magotari-#uml->>Fuck. Can't work anymore. Bad case of the broken heart. My assumptions were so wrong... I'm outta here for the day.
17:06|-|Magotari [~karol@chello089076073248.chello.pl] has quit [Quit: leaving]
17:06<dgraves-#uml->>:(
17:06<dgraves-#uml->>that didn't sound good.
17:07<IntuitiveNipple-#uml->>Did UML cheat on him? or was that referring to personal life?
17:08<dgraves-#uml->>::L::
17:08<dgraves-#uml->>jdike: http://rafb.net/p/P6qN2S72.html
17:08<dgraves-#uml->>there's the patch. Its the same stack trace that my EL5 UML ran into.
17:08<dgraves-#uml->>but i have no idea why that patch would cause it.
17:09<jdike-#uml->>you're not trying to boot reiserfs?
17:11|-|kos_tom [~thomas@col31-3-82-247-183-72.fbx.proxad.net] has quit [Quit: I like core dumps]
17:13<dgraves-#uml->>ext3
17:13<dgraves-#uml->>Oracle_EL5_10GB_64bit: Linux rev 1.0 ext3 filesystem data (needs journal recovery) (large files)
17:14<dgraves-#uml->>so, no.
17:14[~]dgraves #uml wonders if its asking reiser if the sb is its...#uml-> wonders if its asking reiser if the sb is its...
17:14<jdike-#uml->>can you configure reiserfs out?
17:15<dgraves-#uml->>hmmm... yeah, probably. i've just been using defconfig.
17:16[~]dgraves #uml wonders what happens if he just reverts the patch.#uml-> wonders what happens if he just reverts the patch.
17:16<IntuitiveNipple-#uml->>dgraves: context, if you don't have it already http://lkml.org/lkml/2006/8/9/129
17:17<dgraves-#uml->>configuring Reiser out works, jdike.
17:18<dgraves-#uml->>thanks, IntuitiveNipple
17:19<dgraves-#uml->>looks like i've got the second (supposedly) working patch though.
17:21<jdike-#uml->>it fixed a NULL dereference
17:21<jdike-#uml->>not can't-find-your-filesystem
17:21<dgraves-#uml->>::L::
17:21<dgraves-#uml->>yeah.
17:25<Marquel-#uml->>okay, thx guys, it works now. and except for that single file, the problem is fixed in 2.6.23
17:29<dgraves-#uml->>grr... now its not hitting those breakpoints.
17:29<dgraves-#uml->>gdb claims we never hit mount_block_root or mount_root or reiserfs_put_super
17:29<dgraves-#uml->>put_STUPID!!!
17:43<dgraves-#uml->>jdike: taking out the reiserfs from the original EL5 tree we were working on, and 64bit UML works.
17:52<jdike-#uml->>cool
17:52<Marquel-#uml->>have good night
17:52|-|Marquel [~Marquel@port-108.pppoe.wtnet.de] has left #uml [scrubbing cyberspace clean with tiny broadcast bubbles]
17:53<dgraves-#uml->>jdike, so, i dunno if i'm gonna hunt it more. I've got some qualifying tests to run against it, and we'll see.
17:53<dgraves-#uml->>i don't understand why the patch did that, but hey.
17:53<dgraves-#uml->>i'll assume you don't have any objections, of course. :)
17:53<jdike-#uml->>it works
17:53<jdike-#uml->>don't ask too many questions
17:54<dgraves-#uml->>::L::
17:54<dgraves-#uml->>very true, to a certain extent. :)
17:54<dgraves-#uml->>i've got to be reasonably certain it won't fall over when doing development... :)
17:58<dgraves-#uml->>jdike, thanks again for all your help.
17:58<dgraves-#uml->>i really appreciate it.
17:58<dgraves-#uml->>want some christmas cookies? :)
17:58<jdike-#uml->>sure
17:58<dgraves-#uml->>send me an address, i'll ship em to you. homemade. requests if you want.
17:59[~]jdike #uml noted no "" around cookies#uml-> noted no "" around cookies
17:59<jdike-#uml->>18 Roxbury Dr, Littleton, MA 01460
17:59[~]dgraves #uml isn't quite sure if that's to mean money, or other offers of "cookies".#uml-> isn't quite sure if that's to mean money, or other offers of "cookies".
18:00<dgraves-#uml->>:)
18:00|-|tyler29 [~tyler@ARennes-257-1-153-182.w86-214.abo.wanadoo.fr] has quit [Ping timeout: 480 seconds]
18:13|-|tyler29 [~tyler@ARennes-257-1-95-23.w81-48.abo.wanadoo.fr] has joined #uml
18:37|-|tyler29 [~tyler@ARennes-257-1-95-23.w81-48.abo.wanadoo.fr] has quit [Remote host closed the connection]
18:39|-|hfb [~hfb@pool-71-106-219-180.lsanca.dsl-w.verizon.net] has quit [Quit: Leaving]
18:55|-|IntuitiveNipple [~TJ@alexandros.tjworld.net] has quit [Quit: The only intuitive interface is the nipple; everything else is learned]
18:56|-|aroscha [~aroscha@chello213047053193.30.11.tuwien.teleweb.at] has joined #uml
19:03|-|ferret_0567 [~travis@72.191.26.86] has quit [Quit: leaving]
19:07|-|ferret_0567 [~travis@72.191.26.86] has joined #uml
19:07|-|ferret_0567 [~travis@72.191.26.86] has quit []
19:11|-|jdike [~jdike@pool-71-248-190-161.bstnma.fios.verizon.net] has quit [Quit: Leaving]
19:55|-|aroscha [~aroscha@chello213047053193.30.11.tuwien.teleweb.at] has quit [Quit: aroscha]
20:16|-|ferret_0567 [~travis@cpe-72-191-26-86.satx.res.rr.com] has joined #uml
20:46|-|ferret_0567 [~travis@cpe-72-191-26-86.satx.res.rr.com] has quit [Remote host closed the connection]
22:14|-|ferret_0567 [~travis@72.191.26.86] has joined #uml
23:09|-|kokoko1 [~Slacker@203.148.65.19] 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
---Logclosed Thu Dec 06 00:00:28 2007