IRC Logs for #crux-devel Wednesday, 2009-09-09

*** Rotwang has joined #crux-devel00:02
*** sepen has joined #crux-devel01:07
sepentilman, pdbcacher worked fine01:07
*** tilman has quit IRC01:08
*** Rotwang has quit IRC01:11
*** tilman has joined #crux-devel01:11
*** Rotwang has joined #crux-devel01:12
*** mike_k has joined #crux-devel01:26
*** mike_k has quit IRC01:33
*** Rotwang has quit IRC02:00
*** rantolo1 has joined #crux-devel02:34
*** Rotwang has joined #crux-devel02:45
*** Rotwang has quit IRC02:58
*** sepen has quit IRC05:39
*** jaeger has quit IRC05:39
*** jaeger has joined #crux-devel05:40
*** sepen has joined #crux-devel05:50
*** mike_k has joined #crux-devel07:40
sepenjaeger, did you have a log from the iso's building process? or a partial log of the make iso's stage?07:59
sepenI can't find lib/modules/2.6.30.5/kernel/fs on the initramfs, is this right?08:00
sepenalso I found that using the unsquashfs in the iso:08:22
sepen$ tools/unsquashfs08:22
sepenFATAL: kernel too old08:22
sepenSegmentation fault08:22
sepenI think this could break the make_usbdisk on my host08:22
sepen*err is not used by the script08:23
jaegersepen: the fs drivers are in the squashfs08:42
jaegeralso unsquashfs executes fine for me in the iso tree, not sure about the installed version though I expect it would be the same08:43
sepenhmm08:48
*** jue has joined #crux-devel08:49
sepenjaeger, I was wondering if I need an extra fs driver (from the crux.squashfs image) to mount the usbdisk after boot08:50
* jue reads the log08:50
juesepen: all of that works for me08:53
sepenjust trying to find for what I can't mount the pendrive after booting from it (required for mounting the squashfs image too for pivoting to /newroot, and etc.)08:54
sepenthanks jue, I'm still working on the usbdisk stuff08:54
jueyeah, thought so08:55
juebut I can't reproduce the problems/errors you described08:56
sepennah' the problems where only tests08:57
sepenthe question is why I can't mount the usbdisk when booting from it? I can mount it when booting from the cdrom08:58
juehmm, you said that tools/unsquashfs segfaults08:58
jueand that you are missing a kernel-module dir ;)08:58
sepenno no, its not related to unsquashfs08:59
jaegerI'm not familiar with your usbdisk stuff but if you can list all the steps that happen we could figure out what's happening, probably09:00
sepenwhen syslinux boots the usbdisk, it puts a kernel image and initramfs on your ram, so from it the init script tries to mount the boot-media device09:00
sepenis that right?09:00
jaegerI can't speak for the usbdisk but that's basically what happens with the CD boot09:01
sepenyeah, its the same09:01
sepenjust change isolinux by syslinux09:01
jaegerthe initramfs consists of the modules needed to find the cdrom and script to get started... there can be others but that's where it starts09:01
*** pitillo_ has joined #crux-devel09:02
sepenwell, in this step the init (from your ramdisk) tries to mount the cdrom on the iso, the usbdisk's init tries to mount the usbdisk media09:02
sepenjaeger, its similar09:02
sepenbut at this point, I can't mount the usbdisk (required to get crux.squashfs and to continue to boot process)09:03
jaegerIs your init pasted somewhere I can view it?09:04
sepenyeah09:07
sepen1sec09:07
sepenhttp://dpaste.com/91423/09:08
sepenand this is the 'diff' from the original http://dpaste.com/91424/09:09
sepenfrom the diff, line 25 I think is the more important (USB_DEVICES)09:10
jaegerAnd where does it fail? What's the error?09:15
sepencan't 'find_and_mount_usbdisk'09:16
jaegerthat's literally what the error says?09:16
sepen"Spawning a shell for you to attempt to fix this problem...."09:17
jaegerok, and when you try to manually mount the usbdisk, what error do you get?09:18
sepenfirst, dmesg show 2 different devices, sda and sdb09:18
sepenbut seems that the kernel is mixing both of them09:18
sepenthen when trying to mount manually I can't remember the error exactly but I can't mount any of them09:19
sepenthats for what I think I need more fs or something09:20
jaegerIf the kernel is truly mixing devices I'd say you have bad hardware, I can't think of a situation where it would do that with block devices intentionally09:20
sepencat /proc/filesystem has only a couple of them09:20
jaegerWhich filesystem is used on the usbdisk?09:20
sepenjaeger, I though so but I tested it on 2 boxes with 3 different usbdisk09:20
*** pitillo_ has quit IRC09:21
jaegerWhat does dmesg actually say about both devices?09:21
sepenthe usbdisk is in superfloppy format, dosfs for the whole sda, or sdb09:21
*** pitillo_ has joined #crux-devel09:21
jaegerooh, I just thought of something, is this a U3-formatted usbdisk?09:21
sepenjaeger, I can't obtain a capture of dmesg, just waiting to arrive to home to take a photo or something09:21
sepenhmm09:22
jaeger"A U3 flash drive presents itself to the host system as a USB hub with a CD drive and standard USB mass storage device attached."09:22
sepenI formated with mkdosfs09:22
jaegerthe fake cd drive would still show up09:22
jaegerbut theoretically it should work anyway as long as the script does the right thing09:23
sepenjaeger, but I didn't have errors in the past09:23
sepenthe usbdisk worked fine with 2.509:23
jaegerdid you compare the kernel configs between 2.5 and 2.6?09:23
jaegerat least the filesystems part09:23
jaegerMy guess would be that you used to have CONFIG_MSDOS_FS=y in the config09:24
jaegernot familiar with superfloppy09:26
jaegera quick google search leads me to believe it would only create sda or sdb, not both09:27
sepenhmmm09:35
sepenjaeger, I'll look in the kernel config changes09:36
sepenthanks for the support, I apreciated it so much09:36
jaegernp, hope it helps09:37
*** pitiIIo has joined #crux-devel09:37
*** pitillo_ has left #crux-devel09:38
*** tilman has quit IRC09:57
*** tilman has joined #crux-devel09:58
*** pitiIIo_ has joined #crux-devel10:07
*** pitiIIo has quit IRC10:08
sepenuff, lot of changes between kernel series http://dpaste.com/91454/10:21
sepenI'll try to investigate a bit more, now time to go home10:22
sepenbbl10:22
*** sepen has quit IRC10:23
*** pitiIIo_ has quit IRC10:25
*** pitiIIo has joined #crux-devel10:29
*** treach has joined #crux-devel10:45
*** Rotwang has joined #crux-devel10:48
*** pitiIIo_ has joined #crux-devel10:49
*** pitiIIo has quit IRC10:56
*** rantolo1 has quit IRC11:24
*** sepen has joined #crux-devel11:42
*** sepen has joined #crux-devel11:47
*** pitiIIo has joined #crux-devel11:49
*** pitiIIo_ has quit IRC12:03
*** sepen has joined #crux-devel12:16
*** pitiIIo has quit IRC13:28
*** pitiIIo has joined #crux-devel13:32
*** pitiIIo has quit IRC14:23
*** pitiIIo has joined #crux-devel14:29
*** tilman has quit IRC14:38
*** tilman has joined #crux-devel14:45
*** mike_k has quit IRC14:47
*** pitiIIo_ has joined #crux-devel15:01
*** pitiIIo has quit IRC15:10
*** pitiIIo_ has quit IRC15:10
*** pitiIIo has joined #crux-devel15:14
*** tilman has quit IRC15:26
*** tilman has joined #crux-devel15:31
*** Rotwang has quit IRC16:24
*** pitiIIo has quit IRC17:19
*** jue has quit IRC17:51
*** j^2 has quit IRC18:00
rehabdollisnt giflib a replacement for libungif?18:05
treach"It is API and ABI compatible with libungif which was in wide use while the LZW compression algorithm was patented." I'd say libungif was the replacement. :>18:13
treachfwiw, at least suse uses giflib rather than libungif these days.18:14
*** sepen has quit IRC18:17
*** treach has quit IRC18:21
rehabdollyeah the patents expired18:22
rehabdollso seems giflib is back! :)18:22
rehabdollbtw, new firefox...18:23

Generated by irclog2html.py 2.11.0 by Marius Gedminas - find it at mg.pov.lt!