Re: no cd/dvd devices detected after kernel change
On Tue, Jul 29, 2008 at 07:51:04PM +0100, Joe Hickey wrote:
> > > Using 2.6.24-1-686, my two IDE-connected cd/dvd drives are detected as
> > > /dev/sr0 and /dev/sr1, symlinked to cdrom0 and cdrom1, and work perfectly.
> > > When I boot 2.6.25-2-686-bigmem, no entries for these devices appear in
> > > /dev.
>
> <snip dmesg and lsmod output>
>
> > I'm not sure this is relevant but when I modinfo on ide_cd_mod it shows :
> >
> > filename: /lib/modules/2.6.25-2-686/kernel/drivers/ide/ide-cd_mod.ko
> > license: GPL alias: ide-cd alias: ide:*m-cdrom*
> > description: ATAPI CD-ROM Driver depends: ide-core,cdrom vermagic:
> > 2.6.25-2-686 SMP mod_unload 686 parm: ignore:charp
> >
> > Note the description : *ATAPI* CD-ROM Driver. In the dmesg of your bigmem
> > kernel your drivers seems to be recognize as SATA, not ATAPI.
> >
> > There's nothing about GSA-H55L in your dmesg ? Because on your normal kernel
> > there is the name of the drives, not in the bigmem kernel. Is the lspci
> > output the same for both kernels ?
> >
> > Thomas Preud'homme
>
> Thanks for your reply. GSA-H55L doesn't appear in dmesg for the bigmem kernel,
> nor do any references to DVD drives. Both appear when booting 2.6.4-1-686.
>
> Comparing outputs of lspci for the two kernels, there is a difference for the
> following device: 03:00.0 IDE interface [0101]: Marvell Technology Group Ltd.
> 8 Subsystem: ASUSTeK Computer Inc. Device [1043:82a2]
>
> In the standard issue kernel, lspci reports: Kernel driver in use:
> pata_marvell Kernel modules: pata_marvell
>
> But the bigmem kernel: Kernel driver in use: ahci Kernel modules: ahci,
> pata_marvell
>
> Might this be the issue? I tried adding ahci to /etc/modprobe.d/blacklist and
> rebooting the bigmem kernel, but the dvd devices are still not detected.
>
I solved this problem by adding 'pata_marvell' to /etc/initramfs-tools/modules
and then rebuilding the initial boot image by typing 'sudo update-initramfs -u'.
My guess is that pata_marvell wasn't being loaded early enough. Is
this behaviour a bug/regression in the kernel, or just a quirk of my setup?
--
Joe Hickey
Reply to: