On Fri, Jul 19, 2013 at 12:53:49AM +0200, Borislav Petkov wrote:
> > I assumed it was some sort of BIOS issue where any modification of the
> > RTC_AIE bit caused the alarm irq line to be left high(or something
> > like that) that triggered the immediate power-on on shutdown. But I've
> > not been able to dig down on this.
>
> Ha, this actually fits like an ass on a bucket (don't ask - German
> proverb :-)).
>
> If what you're saying is actually the case, then this explains why not
> writing to 0xb doesn't cause the alarm irq to fire.
>
> Btw, in the trace above we do the disabling twice. Once from
> rtc_timer_remove() and then again from rtc_timer_do_work().
>
> So, if we disable it once and we touch RTC_AIE again causing the second
> time to rearm the alarm irq, this would explain the issue. Which reminds
> me:
>
> Maybe we should read out the alarm interrupt first and disable it only
> if it is enabled - that would save us the modification of RTC_AIE. Cool,
> I'll try that tomorrow.
Well, the below seems to do the trick. But since I don't trust the BIOS
in any way, I'll run it a couple more days here. Btw, I think we should
commit this regardless, as it saves us unneeded writes:
---
diff --git a/drivers/rtc/rtc-cmos.c b/drivers/rtc/rtc-cmos.c
index be06d7150de5..bb265f1651e7 100644
--- a/drivers/rtc/rtc-cmos.c
+++ b/drivers/rtc/rtc-cmos.c
@@ -304,6 +304,9 @@ static void cmos_irq_enable(struct cmos_rtc *cmos, unsigned
char mask)
rtc_control = CMOS_READ(RTC_CONTROL);
cmos_checkintr(cmos, rtc_control);
+ if (rtc_control == mask)
+ return;
+
rtc_control |= mask;
CMOS_WRITE(rtc_control, RTC_CONTROL);
hpet_set_rtc_irq_bit(mask);
@@ -316,6 +319,10 @@ static void cmos_irq_disable(struct cmos_rtc *cmos,
unsigned char mask)
unsigned char rtc_control;
rtc_control = CMOS_READ(RTC_CONTROL);
+
+ if (!(rtc_control & mask))
+ return;
+
rtc_control &= ~mask;
CMOS_WRITE(rtc_control, RTC_CONTROL);
hpet_mask_rtc_irq_bit(mask);
--
--
Regards/Gruss,
Boris.
Sent from a fat crate under my desk. Formatting is fine.
--
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majord...@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
