On Mon, 5 Sep 2005, Atsushi Nemoto wrote: > 3. We should protect rtc_set_mmss() call during get_rtc_time or > set_rtc_time (this is not your patch's fault). > > How about this patch? The rtc_lock could be manipulated in each > RTC-dependent routines, but I choose a simple way for now. That's how other architectures do this, see e.g. "arch/alpha/kernel/time.c". Why should we be different, even for now? Also the call is named rtc_set_mmss() for an unknown reason while all the others have set_rtc_mmss(). Maciej
