linux-kernel - Re: Solid freezes with 2.6.25

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite for Android: free password hash cracker in your pocket

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <20080430165131.7426d5d3@strauss.suse.de>
Date:	Wed, 30 Apr 2008 16:51:31 +0200
From:	Bernhard Walle <bwalle@...e.de>
To:	Gabor Gombas <gombasg@...aki.hu>,
	Andrew Morton <akpm@...ux-foundation.org>,
	linux-kernel@...r.kernel.org, Ingo Molnar <mingo@...e.hu>,
	Thomas Gleixner <tglx@...utronix.de>
Subject: Re: Solid freezes with 2.6.25

Hi,

* Gabor Gombas [2008-04-30 15:38]:
>
> On Tue, Apr 29, 2008 at 05:54:19PM +0200, Bernhard Walle wrote:
> 
> > Hm ..., can you mail /proc/interrupts of the affected system? According
> > to the HPET specification 
> 
> The Intel box:
> 
>            CPU0       CPU1       
>   0:      27346      29145   IO-APIC-edge      timer
>   1:        842        819   IO-APIC-edge      i8042
>   4:          2          2   IO-APIC-edge    
>   6:          0          3   IO-APIC-edge      floppy
>   8:          2          3   IO-APIC-edge      rtc0

Well, looks fine. It's edge triggered. My assumption that the
T1_INT_STS must be cleared is wrong since the complete HPET code
assumes edge-triggered interrupts. To use level-triggered interrupts,
the HPET must be programmed in level triggered mode, i.e. the
Tn_INT_TYPE_CNF bit must be set. Which is never done.

Back to the original problem. Can you please watch the interrupts from
the RTC until the lockup happens with something like

------------------------- 8< -----------------------------------
function get_total_irqs()
{
	sum=0
	for n in $(grep rtc /proc/interrupts | \
				 sed -e 's/.*://g' | \
				 sed -e 's/[a-zA-Z].*//g' ) ; 
		do sum=$[$sum+$n]
	done

	echo $sum
}

last=$(get_total_irqs)
while true ; do
	sum=$(get_total_irqs)
	echo $[$sum-$last]
	last=$sum

	sleep 1
done
------------------------- >8 -----------------------------------

Also, please send the full .config (maybe via private mail if that's
too large for LKML, or upload it).

> > Can you try that patch and see if the IRQ handler is really called
> > repeatedly in that case?
> 
> Unfortunately I won't have time today and will be traveling till the end
> of the week. I'll give it a shot next week.

Ok. Maybe don't use the patch I sent, instead use:



	Bernhard


--- a/arch/x86/kernel/hpet.c
+++ b/arch/x86/kernel/hpet.c
@@ -680,6 +680,12 @@ irqreturn_t hpet_rtc_interrupt(int irq,
        struct rtc_time curr_time;
        unsigned long rtc_int_flag = 0;
 
+       if (printk_ratelimit())
+               printk(KERN_DEBUG "hpet_rtc_interrupt %s%s%s\n",
+                               (hpet_rtc_flags & RTC_UIE) ? "RTC_UIE " : "",
+                               (hpet_rtc_flags & RTC_AIE) ? "RTC_AIE " : "",
+                               (hpet_rtc_flags & RTC_PIE) ? "RTC_PIE " : "");
+
        hpet_rtc_timer_reinit();
        memset(&curr_time, 0, sizeof(struct rtc_time));


--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/