cdrom_id from udev-135 hangs

cdrom_id from udev-135 hangs

[ljb]

I am trying to upgrade an old PC from Slackware 12.1 (udev-118,
Linux-2.6.24.5) to Slackware 12.2 (udev-135, Linux-2.6.27.7). The system
hangs up during boot running cdrom_id. If I disable 60-cdrom_id.rules, or
replace cdrom_id with the one from udev-118, it works.

Running cdrom_id from udev-135 (138 too) logs lots of messages like those
shown below, and never completes.  The drive is an IDE Sony CD-RW CRX175E2,
connected to an Intel SE440BX2 motherboard, and the drive works fine other
than this issue. How do I troubleshoot this?

  kernel: hdc: status error: status=0x59 { DriveReady SeekComplete
    DataRequest Error }
  kernel: hdc: status error: error=0x20 { LastFailedSense=0x02 }
  kernel: ide: failed opcode was: unknown
  kernel: hdc: drive not ready for command
  ...

Re: cdrom_id from udev-135 hangs

[Kay Sievers]

On Tue, Feb 24, 2009 at 03:02, ljb <ljb1813@pobox.com> wrote:
> I am trying to upgrade an old PC from Slackware 12.1 (udev-118,
> Linux-2.6.24.5) to Slackware 12.2 (udev-135, Linux-2.6.27.7). The system
> hangs up during boot running cdrom_id. If I disable 60-cdrom_id.rules, or
> replace cdrom_id with the one from udev-118, it works.
>
> Running cdrom_id from udev-135 (138 too) logs lots of messages like those
> shown below, and never completes.  The drive is an IDE Sony CD-RW CRX175E2,
> connected to an Intel SE440BX2 motherboard, and the drive works fine other
> than this issue. How do I troubleshoot this?
>
>  kernel: hdc: status error: status=0x59 { DriveReady SeekComplete
>    DataRequest Error }
>  kernel: hdc: status error: error=0x20 { LastFailedSense=0x02 }
>  kernel: ide: failed opcode was: unknown
>  kernel: hdc: drive not ready for command

Can you try using libata drivers for it and see if shows the same problem?

Thanks,
Kay

Re: cdrom_id from udev-135 hangs

[ljb]

kay.sievers@vrfy.org wrote:
>...
>
> Can you try using libata drivers for it and see if shows the same problem?

Sorry, I looked and read and I can't find out how to do this. I know my
kernel has libata and ata_piix compiled in, and I see a message from libata
at boot time, but I am pretty sure it isn't being used on this PATA-only PC.
I can't figure out how to get it to use libata for the CD.

Re: cdrom_id from udev-135 hangs

[Bryan Kadzban]

[-- Attachment #1: Type: text/plain, Size: 1186 bytes --]

ljb wrote:
> kay.sievers@vrfy.org wrote:
>> Can you try using libata drivers for it and see if shows the same
>> problem?
> 
> Sorry, I looked and read and I can't find out how to do this. I know
> my kernel has libata and ata_piix compiled in, and I see a message
> from libata at boot time, but I am pretty sure it isn't being used on
> this PATA-only PC. I can't figure out how to get it to use libata for
> the CD.

Disable CONFIG_IDE.  Or disable CONFIG_BLK_DEV_IDE.  Or maybe disable
CONFIG_BLK_DEV_IDECD (but I don't know if the PIIX libata driver will
pick up just the CD drive).  Or disable CONFIG_BLK_DEV_PIIX.  This is in
approximate order of most-general to most-specific options; turning off
the last option will disable the least amount of code.

(I have CONFIG_IDE disabled entirely, but whether you can do that will
depend on your current IDE chipset(s) and their level of libata support...)

There may be ways to do this with kernel options as well, but I don't
know of any offhand.  Check Documentation/kernel-parameters.txt for the
full list of flags; something appropriate may be in there.  (At least
that would let you avoid recompiling.)


[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 260 bytes --]

Re: cdrom_id from udev-135 hangs

[ljb]

bryan@kadzban.is-a-geek.net wrote:
>...
> Disable CONFIG_IDE.  Or disable CONFIG_BLK_DEV_IDE.  Or maybe disable
> CONFIG_BLK_DEV_IDECD (but I don't know if the PIIX libata driver will
> pick up just the CD drive).  Or disable CONFIG_BLK_DEV_PIIX.  This is in
> approximate order of most-general to most-specific options; turning off
> the last option will disable the least amount of code.
>
> (I have CONFIG_IDE disabled entirely, but whether you can do that will
> depend on your current IDE chipset(s) and their level of libata support..)
>
> There may be ways to do this with kernel options as well, but I don't
> know of any offhand.  Check Documentation/kernel-parameters.txt for the
> full list of flags; something appropriate may be in there.  (At least
> that would let you avoid recompiling.)

I was afraid it might take something like that - recompile without IDE or
IDECD. I'll have to think about if I'm brave enough to try that. Also
considering the libata PATA drivers are considered "experimental".

I did look at kernel-parameters.txt before posting to see if there was
something for libata like the old 2.4 option for having ide-scsi handle the 
CD instead of the regular IDE driver.  I didn't find anything there. I did
find this in ide.txt: "ide_core.noprobe=1.0". It causes the IDE driver to
ignore my CD drive (master device, secondary channel). But libata doesn't
pick it up automatically. If there is a way to get it to do so, I haven't
found it.