* Gladimir (listmgr@san.rr.com) [010717 21:31]:
>
> I have tracked this down to the point that I know it is an alias for the
> real time clock, and that it only occurs soon after switching to run-level 6
> and run-level 0. I also know that the alias exists in two files, that I
> know of; /etc/modules.conf and /etc/modutils/arch/i386. Removing the alias
> from i386 has no effect.
>
> I have seen a few people suggesting that I should edit modules.conf, but the
> modules.conf file has comments stating that it should never be edited
> manually. Is it either safe or effective to manually edit this file? If
> not the modules.conf file, then which file do I edit to affect this
> workaround until I can compile rtc in the kernel?
>
> Thanks in advance for any help.
>
> - gladimir
This is most likely being done by /etc/init.d/hwclock.sh. That it
happens with runlevels 0 and 6 would make sense when looking at:
vineet@gobo /etc/init.d % ls /etc/rc?.d/*hwclock.sh
/etc/rc0.d/K25hwclock.sh@ /etc/rc6.d/K25hwclock.sh@
/etc/rcS.d/S50hwclock.sh@
The actual command that tries to load the rtc driver is hwclock
from hwclock(8):
hwclock tries to use /dev/rtc. If it is compiled for a
kernel that doesn't have that function or it is unable to
open /dev/rtc, hwclock will fall back to another method,
if available. On an ISA or Alpha machine, you can force
hwclock to use the direct manipulation of the CMOS regis
ters without even trying /dev/rtc by specifying the
--directisa option.
Looks to me like it tries to access /dev/rtc, which in turn tries to
load the kernel's rtc driver (char-major-10-135).
I saw the same error message with every boot/reboot. My solution was
just to include the rtc driver next time I built my kernel, but you
probably don't need that, and it's not worth recompiling just for
that; try adding the --directisa option in /etc/init.d/hwclock.sh
HTH,
Vineet
Attachment:
pgpwk8SyxN8mC.pgp
Description: PGP signature