On Sun, 2007-11-11 at 11:02 +0100, Alain Guibert wrote: > Hello Ondrej, Martin, David, and Paul. > > On Thursday, September 20, 2007 at 11:20:15 +0200, Ondrej Certik wrote: > > [bugs #435102, #439593, and #450686] > [hwclock timeouts waiting for RTC interrupt] > > Where is the real problem, that it doesn't work for me without the > > --directisa? Is it a kernel problem, or a hardware problem? > > In bug #450686, Paul pointed the finger at a possible Linux kernel bug > in the /dev/rtc device driver on some hardwares, but this has to be > confirmed. Please, could all of you try the workaround proposed at > <URL:http://bugzilla.kernel.org/show_bug.cgi?id=7014>? In short, you'd > have to recompile the kernel with a different config, completely > removing the old /dev/rtc device driver, and adding the new rtc-cmos > driver instead. Then try the later via hwclock --rtc=/dev/rtc0
Since the Debian kernel images seem to come with rtc-cmos, I removed the
old rtc driver, modprobed the rtc-cmos one, and then hwclock worked
fine, with or without the --rtc option.
So, there are two problems:
* kernel doesn't load the right driver on some systems
* hwclock doesn't ignore buggy drivers and try --directisa
It would be great if both could be fixed.
--
bye,
pabs
http://wiki.debian.org/PaulWise
signature.asc
Description: This is a digitally signed message part
