Re: 2.4.0-test11pre2-ac1 and previous problem

From: Andrew Morton (andrewm@uow.edu.au)
Date: Fri Nov 10 2000 - 19:25:00 EST


"Paweł Kot" wrote:
>
> Hello,
>
> I've following error with 2.4.0-test{9|10|pre11pre1-ac1|pre11pre2-ac1}:
>
> NMI Watchdog detected LOCKUP on CPU3, registers:
>
> And then the machine hangs. No response at all.
> Always CPU3 is mentioned.
> The machine is:
> The latest Intel motherboard for 4xCPU (ISP4040)
> 4xPentium III 700 (Xeon)
> 4GB RAM
> mylex raid array (the newest controller)
> eepro100 ethernet card
>
> This machine is running only MySQL database.
>
> What can be wrong?

Oh no. Another one. Could you please try this attached
patch (against test11-pre2) and see if the diagnostics
come out?

--- linux-2.4.0-test11-pre2/arch/i386/kernel/traps.c Fri Nov 10 20:24:02 2000
+++ linux-akpm/arch/i386/kernel/traps.c Sat Nov 11 02:35:25 2000
@@ -382,20 +382,10 @@
         printk("Do you have a strange power saving mode enabled?\n");
 }
 
-#if CONFIG_X86_IO_APIC
-
-int nmi_watchdog = 1;
-
-static int __init setup_nmi_watchdog(char *str)
-{
- get_option(&str, &nmi_watchdog);
- return 1;
-}
-
-__setup("nmi_watchdog=", setup_nmi_watchdog);
-
-extern spinlock_t console_lock, timerlist_lock;
+extern spinlock_t console_lock, timerlist_lock, runqueue_lock;
 static spinlock_t nmi_print_lock = SPIN_LOCK_UNLOCKED;
+extern wait_queue_head_t log_wait;
+static int ignore_spinlocks = -1;
 
 /*
  * Unlock any spinlocks which will prevent us from getting the
@@ -404,9 +394,30 @@
  */
 void bust_spinlocks(void)
 {
+ ignore_spinlocks = smp_processor_id();
+ global_irq_lock = 0;
         spin_lock_init(&console_lock);
         spin_lock_init(&timerlist_lock);
+ spin_lock_init(&runqueue_lock);
+ log_wait.lock = WAITQUEUE_RW_LOCK_UNLOCKED;
 }
+
+int no_spinlocks()
+{
+ return smp_processor_id() == ignore_spinlocks;
+}
+
+#if CONFIG_X86_IO_APIC
+
+int nmi_watchdog = 1;
+
+static int __init setup_nmi_watchdog(char *str)
+{
+ get_option(&str, &nmi_watchdog);
+ return 1;
+}
+
+__setup("nmi_watchdog=", setup_nmi_watchdog);
 
 inline void nmi_watchdog_tick(struct pt_regs * regs)
 {
--- linux-2.4.0-test11-pre2/include/asm-i386/spinlock.h Sun Oct 15 01:27:46 2000
+++ linux-akpm/include/asm-i386/spinlock.h Sat Nov 11 02:17:46 2000
@@ -8,6 +8,8 @@
 extern int printk(const char * fmt, ...)
         __attribute__ ((format (printf, 1, 2)));
 
+extern int no_spinlocks(void);
+
 /* It seems that people are forgetting to
  * initialize their spinlocks properly, tsk tsk.
  * Remember to turn this off in 2.4. -ben
@@ -68,6 +70,8 @@
 static inline int spin_trylock(spinlock_t *lock)
 {
         char oldval;
+ if (no_spinlocks())
+ return 0;
         __asm__ __volatile__(
                 "xchgb %b0,%1"
                 :"=q" (oldval), "=m" (lock->lock)
@@ -85,6 +89,8 @@
                 BUG();
         }
 #endif
+ if (no_spinlocks())
+ return;
         __asm__ __volatile__(
                 spin_lock_string
                 :"=m" (lock->lock) : : "memory");
@@ -149,6 +155,9 @@
         if (rw->magic != RWLOCK_MAGIC)
                 BUG();
 #endif
+ if (no_spinlocks())
+ return;
+
         __build_read_lock(rw, "__read_lock_failed");
 }
 
@@ -158,6 +167,8 @@
         if (rw->magic != RWLOCK_MAGIC)
                 BUG();
 #endif
+ if (no_spinlocks())
+ return;
         __build_write_lock(rw, "__write_lock_failed");
 }
 
--- linux-2.4.0-test11-pre2/arch/i386/kernel/i386_ksyms.c Fri Aug 11 19:06:11 2000
+++ linux-akpm/arch/i386/kernel/i386_ksyms.c Sat Nov 11 02:24:32 2000
@@ -155,3 +155,6 @@
 #ifdef CONFIG_X86_PAE
 EXPORT_SYMBOL(empty_zero_page);
 #endif
+
+EXPORT_SYMBOL(no_spinlocks);
+

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



This archive was generated by hypermail 2b29 : Wed Nov 15 2000 - 21:00:18 EST