Re: Realtime Preemption, 2.6.12, Beginners Guide?

[Ingo Molnar]

* Ingo Molnar <mingo@xxxxxxx> wrote:

> does the patch below help? We initialized the timestamps to 0, but 
> with jiffies starting out negative, that means a ~5 minutes gap until 
> we first reach a value of 0. That would explain the messages. The only 
> thing it doesnt explain, why did this only trigger on your box?

here's an updated patch - it will print out all timestamps too. (you'll 
have to revert all previous softlockup patches first, via patch -R.)

	Ingo

Index: linux/kernel/softlockup.c
===================================================================
--- linux.orig/kernel/softlockup.c
+++ linux/kernel/softlockup.c
@@ -16,9 +16,9 @@
 
 static DEFINE_RAW_SPINLOCK(print_lock);
 
-static DEFINE_PER_CPU(unsigned long, timeout) = 0;
-static DEFINE_PER_CPU(unsigned long, timestamp) = 0;
-static DEFINE_PER_CPU(unsigned long, print_timestamp) = 0;
+static DEFINE_PER_CPU(unsigned long, timeout) = INITIAL_JIFFIES;
+static DEFINE_PER_CPU(unsigned long, timestamp) = INITIAL_JIFFIES;
+static DEFINE_PER_CPU(unsigned long, print_timestamp) = INITIAL_JIFFIES;
 static DEFINE_PER_CPU(struct task_struct *, watchdog_task);
 
 static int did_panic = 0;
@@ -65,8 +65,8 @@ void softlockup_tick(void)
 		per_cpu(print_timestamp, this_cpu) = timestamp;
 
 		spin_lock(&print_lock);
-		printk(KERN_ERR "BUG: soft lockup detected on CPU#%d!\n",
-			this_cpu);
+		printk(KERN_ERR "BUG: soft lockup detected on CPU#%d! %ld-%ld(%ld)\n",
+			this_cpu, jiffies, timestamp, timeout);
 		dump_stack();
 #if defined(__i386__) && defined(CONFIG_SMP)
 		nmi_show_all_regs();
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/