On Sat, Jun 09, 2007 at 09:07:10PM -0400, Zach wrote: > I just upgraded my kernel in Debian testing release from 2.4.27 to > 2.6.18 and ran into some problems I would appreciate help with. > > I built the kernel from the Debian kernel source package which has > some bug fixes and patches not found in the upstream (official) kernel > sources but I asked around and these don't seem to be related to my > problems. I built the .config from scratch (took 3.5 hours). > Total time for the kernel project was 9 hours heh! > Why build from source? What do you need that the binary kernel images don't provide?
> Problem 1:
> Since devfs has been replaced by udev (dynamic device naming) there is
> no longer a /dev/ttyS0 for my serial modem. Also this was symlinked to
> /dev/modem and that device is also no longer present in 2.6.18. Do I
> just need to change /var/lib/setserial/autoserial.conf so it uses
> /dev/tty0? How can I create a /dev/modem block device?
>
Strange. My udev-based Etch system has a /dev/ttyS0 present. Have you
modified anything in your udev configuration?
> Problem 2:
> Something weird is happening with my CDROM.
>
> netrek:~# mount /dev/hdc /cdrom
> mount: block device /dev/hdc is write-protected, mounting read-only
>
That is normal.
> netrek:~# cd /cdrom
>
> netrek:/cdrom# df
> Filesystem 1K-blocks Used Available Use% Mounted on
> /dev/hda2 9906644 9240328 163080 99% /
> tmpfs 128628 0 128628 0% /lib/init/rw
> udev 10240 44 10196 1% /dev
> tmpfs 128628 0 128628 0% /dev/shm
> /dev/hdc 693330 693330 0 100% /cdrom
>
That is also normal.
> netrek:/cdrom# ls
> 00000001.tmp dxsetup movies01.mix scores.mix sizzle1.vqa
> autoplay.wav install multi.mix setup ts1.dsk
> autorun.exe license.doc readme.doc setup.exe westwood.hlp
> autorun.inf maps01.mix readme.txt sidecd01.mix wolapi
>
It is a directory listing. Nothing looks out of sorts there.
> netrek:/cdrom# ls -l /tmp/license.doc
> -r-xr-xr-x 1 root root 26112 Jun 8 16:02 /tmp/license.doc
>
Everything is OK here as well.
> It took a very long time for such a small file (nearly 40 seconds) and
> /var/log/messages showed:
>
> Jun 8 16:02:38 netrek kernel: hdc: irq timeout: status=0xd0 { Busy }
> Jun 8 16:02:38 netrek kernel: ide: failed opcode was: unknown
> Jun 8 16:02:38 netrek kernel: hdc: ATAPI reset complete
>
That could indicate a damaged CD or failing drive. What is the output
of 'hdparm /dev/hdc' and also 'hdparm -i /dev/hdc'?
> However when I removed the file in /tmp and copied it over again this
> time it worked instantaneously as it should. So i think my ide/ATAPI
> is setup wrong but the .config has so many options i'm not sure which
> ones i may have set wrong, I can send you my kernel .config if that would
> help.
>
It happens instantaneously the second tume because the file is small
enough to fit in the drive's cache. No need to read it back from the
CD all over again.
> Further I noticed some other messages, possibly related, from boot up
> in my /var/log/syslog:
>
> Jun 8 15:47:57 netrek kernel: ide-cd: cmd 0x28 timed out
> Jun 8 15:47:57 netrek kernel: hdc: DMA interrupt recovery
> Jun 8 15:47:57 netrek kernel: hdc: lost interrupt
> Jun 8 15:47:57 netrek kernel: hdc: status error: status=0x58 {
> DriveReady SeekComplete DataRequest }
> Jun 8 15:47:57 netrek kernel: ide: failed opcode was: unknown
> Jun 8 15:47:57 netrek kernel: hdc: drive not ready for command
> Jun 8 15:47:57 netrek kernel: hdc: status timeout: status=0xd0 { Busy }
> Jun 8 15:47:57 netrek kernel: ide: failed opcode was: unknown
> Jun 8 15:47:57 netrek kernel: hdc: DMA disabled
> Jun 8 15:47:57 netrek kernel: hdc: drive not ready for command
> Jun 8 15:47:57 netrek kernel: hdc: ATAPI reset complete
> Jun 8 15:47:59 netrek kernel: Warning: /proc/ide/hd?/settings
> interface is obsolete, and will be removed soon!
>
If this is happening at boot, you have a failing drive or motherboard.
> And what should I do about /proc/ide/hd?/settings?
>
Depends on what is trying to access it.
> Problem 3:
> In my /var/log/syslog I see tons of messages on boot dealing with
> "class_uevent" and "class_device_create_uevent" is this a problem or
> not? Below I have included the logfile excerpt.
>
Not sure.
Regards,
-Roberto
--
Roberto C. Sánchez
http://people.connexer.com/~roberto
http://www.connexer.com
signature.asc
Description: Digital signature
