I've got a similar problem with my mum's dapper installation.
I installed the current 2.6.20-generic kernel from the feisty respository on a dapper machine,
to fix a problem with not detecting a USB scanner at power-up (needed to unplug/replug to see it)
However I found in about 1 in 5 boots the 2.6.20 kernel failed to boot,
with it waiting for ages with:
Begin: Waiting for root file system... ...
whenever this occurred it was preceeded by a:
hda: ERROR PORTS ALREADY IN USE
Its seems it might be a race condition, where the via v82cxxx driver
doesn't always claim /dev/hda before ide-generic does. Which causes the boot to fail.
After a while in the fault contidition it drops into a busybox shell and
I found only/dev/hdc the cdrom drive was listed in /dev
Not sure If I found a fix, but its booted 10 times without problem since I commented the ide-generic and recreated the
initrd file for the 2.6.20 kernel as sort of explained in bug report 76872
I've got a similar problem with my mum's dapper installation.
I installed the current 2.6.20-generic kernel from the feisty respository on a dapper machine,
to fix a problem with not detecting a USB scanner at power-up (needed to unplug/replug to see it)
However I found in about 1 in 5 boots the 2.6.20 kernel failed to boot,
with it waiting for ages with:
Begin: Waiting for root file system... ...
whenever this occurred it was preceeded by a:
hda: ERROR PORTS ALREADY IN USE
Its seems it might be a race condition, where the via v82cxxx driver
doesn't always claim /dev/hda before ide-generic does. Which causes the boot to fail.
After a while in the fault contidition it drops into a busybox shell and
I found only/dev/hdc the cdrom drive was listed in /dev
Not sure If I found a fix, but its booted 10 times without problem since I commented the ide-generic and recreated the
initrd file for the 2.6.20 kernel as sort of explained in bug report 76872