Hello folks,
here's what I've got trying to debug with kgdb on powerpc target (Sandpoint)
with the latest patches:
BUG: soft lockup detected on CPU#0!
Call Trace:
[C028BCA0] [C0008428] show_stack+0x3c/0x194 (unreliable)
[C028BCD0] [C00400E0] softlockup_tick+0x94/0xc4
[C028BCF0] [C002858C] run_local_timers+0x18/0x28
[C028BD00] [C00285CC] update_process_times+0x30/0x7c
[C028BD10] [C000D0F0] timer_interrupt+0xd4/0x508
[C028BD90] [C0010648] ret_from_except+0x0/0x14
--- Exception: 901 at breakpoint+0xb4/0xcc
LR = kgdb8250_interrupt+0x74/0xa8
[C028BE50] [C0022FA4] irq_exit+0x48/0x58 (unreliable)
[C028BE60] [C0131500] kgdb8250_interrupt+0x74/0xa8
[C028BE70] [C004041C] handle_IRQ_event+0x64/0xb8
[C028BE90] [C0041D94] handle_level_irq+0x9c/0x13c
[C028BEB0] [C0018008] sandpoint_8259_cascade+0x7c/0xb0
[C028BED0] [C0005F00] do_IRQ+0x9c/0xc0
[C028BEE0] [C0010648] ret_from_except+0x0/0x14
--- Exception: 501 at cpu_idle+0x48/0xe8
LR = cpu_idle+0xd8/0xe8
[C028BFB0] [C0003F8C] rest_init+0x28/0x38
[C028BFC0] [C02336B0] start_kernel+0x1b8/0x220
[C028BFF0] [00003860] 0x3860
I've been poking around that a bit, and the problem seems to be introduced by
core.patch. The old (working) version of core.patch did the following:
void do_timer(struct pt_regs *regs)
{
+ int this_cpu = smp_processor_id();
jiffies_64++;
/* prevent loading jiffies before storing new jiffies_64 value. */
barrier();
update_times();
- softlockup_tick(regs);
+
+#ifdef CONFIG_KGDB
+ if(!atomic_read(&kgdb_sync_softlockup[this_cpu]))
+#endif
+ softlockup_tick(regs);
+
}
whereas the new one does the following:
void do_timer(struct pt_regs *regs)
{
+ int this_cpu = smp_processor_id();
jiffies_64++;
/* prevent loading jiffies before storing new jiffies_64 value. */
barrier();
update_times();
+
+#ifdef CONFIG_KGDB
+ if(!atomic_read(&kgdb_sync_softlockup[this_cpu]))
+#endif
+ softlockup_tick();
}
which of course isn't equivalent.
AFAICS, softlockup_tick() has moved to run_local_timers() so the current
core.patch introduces second (wrong) softlockup_tick() call and doesn't modify
the first one not to happen when kgdb_sync_softlockup is set. The latter one
was triggering the soft lockup message I've been getting.
So, please find the fix for core.patch as I see it inlined below, any comments
are welcome.
Vitaly
kernel/timer.c | 10 ++++------
1 file changed, 4 insertions(+), 6 deletions(-)
Signed-off-by: Vitaly Wool <[EMAIL PROTECTED]>
diff -u linux-2.6.git/kernel/timer.c powerpc.git/kernel/timer.c
--- linux-2.6.git/kernel/timer.c 2006-08-11 19:55:59.000000000 +0400
+++ powerpc.git/kernel/timer.c
@@ -1258,7 +1258,11 @@
*/
void run_local_timers(void)
{
+ int this_cpu = smp_processor_id();
raise_softirq(TIMER_SOFTIRQ);
+#ifdef CONFIG_KGDB
+ if(!atomic_read(&kgdb_sync_softlockup[this_cpu]))
+#endif
softlockup_tick();
}
@@ -1284,16 +1288,10 @@
void do_timer(struct pt_regs *regs)
{
- int this_cpu = smp_processor_id();
jiffies_64++;
/* prevent loading jiffies before storing new jiffies_64 value. */
barrier();
update_times();
-
-#ifdef CONFIG_KGDB
- if(!atomic_read(&kgdb_sync_softlockup[this_cpu]))
-#endif
- softlockup_tick();
}
#ifdef __ARCH_WANT_SYS_ALARM
-------------------------------------------------------------------------
Using Tomcat but need to do more? Need to support web services, security?
Get stuff done quickly with pre-integrated technology to make your job easier
Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo
http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&dat=121642
_______________________________________________
Kgdb-bugreport mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/kgdb-bugreport
