Skip to contentSkip to conversation

PATCH - ksymoops on Alpha - 2.4.5-ac3

PATCH George France 2 participants 4 messages May 28, 2001 – May 29, 2001 Alpha port

lore
Share & tools
lore
#0 lore raw

Here is a trivial patch that will make ksymoops work again on Alpha.

--George

diff -urN linux-2.4.5-ac3-orig/arch/alpha/kernel/traps.c linux/arch/alpha/kernel/traps.c

arch/alpha/kernel/traps.c +1 -5
@@ -286,11 +286,7 @@
 			continue;
 		if (tmp >= (unsigned long) &_etext)
 			continue;
-		/*
-		 * Assume that only the low 24-bits of a kernel text address
-		 * is interesting.
-		 */
-		printk("%6x%c", (int)tmp & 0xffffff, (++i % 11) ? ' ' : '\n');
+		printk("%16lx%c", tmp);
 #if 0
 		if (i > 40) {
 			printk(" ...");
arch/alpha/kernel/traps.c +1 -5
diff -urN linux-2.4.5-ac3-orig/arch/alpha/kernel/traps.c linux/arch/alpha/kernel/traps.c
@@ -286,11 +286,7 @@
 			continue;
 		if (tmp >= (unsigned long) &_etext)
 			continue;
-		/*
-		 * Assume that only the low 24-bits of a kernel text address
-		 * is interesting.
-		 */
-		printk("%6x%c", (int)tmp & 0xffffff, (++i % 11) ? ' ' : '\n');
+		printk("%16lx%c", tmp);
 #if 0
 		if (i > 40) {
 			printk(" ...");
#1 lore raw
Quoted 1 line · George France

Here is a trivial patch that will make ksymoops work again on Alpha.

Thanks for that. Now if you can just persuade the Alpha people to print the 'Code:' line in the same format as other architectures then ksymoops can decode the instructions as well. If Alpha wants to include its own instruction decoder as well then that is up to them but I would appreciate a standard 'Code:' line being printed first.

#2 lore raw
Quoted 2 lines
Quoted 1 line · George France

Here is a trivial patch that will make ksymoops work again on Alpha.

Cleaner patch.

diff -urN linux-2.4.5-ac3-orig/arch/alpha/kernel/traps.c linux-2.4.5/arch/alpha/kernel/traps.c

arch/alpha/kernel/traps.c +1 -7
@@ -286,17 +286,11 @@
 			continue;
 		if (tmp >= (unsigned long) &_etext)
 			continue;
-		/*
-		 * Assume that only the low 24-bits of a kernel text address
-		 * is interesting.
-		 */
-		printk("%6x%c", (int)tmp & 0xffffff, (++i % 11) ? ' ' : '\n');
-#if 0
+		printk("%lx%c", tmp, ' ');
 		if (i > 40) {
 			printk(" ...");
 			break;
 		}
-#endif
 	}
 	printk("\n");
 }
Quoted 6 lines

Thanks for that. Now if you can just persuade the Alpha people to print the 'Code:' line in the same format as other architectures then ksymoops can decode the instructions as well. If Alpha wants to include its own instruction decoder as well then that is up to them but I would appreciate a standard 'Code:' line being printed first.

Could you send me an oops with the standard 'Code:' line?

Best Regards,

--George

arch/alpha/kernel/traps.c +1 -7
diff -urN linux-2.4.5-ac3-orig/arch/alpha/kernel/traps.c linux-2.4.5/arch/alpha/kernel/traps.c
@@ -286,17 +286,11 @@
 			continue;
 		if (tmp >= (unsigned long) &_etext)
 			continue;
-		/*
-		 * Assume that only the low 24-bits of a kernel text address
-		 * is interesting.
-		 */
-		printk("%6x%c", (int)tmp & 0xffffff, (++i % 11) ? ' ' : '\n');
-#if 0
+		printk("%lx%c", tmp, ' ');
 		if (i > 40) {
 			printk(" ...");
 			break;
 		}
-#endif
 	}
 	printk("\n");
 }
#3 lore raw
Quoted 1 line · George France

Could you send me an oops with the standard 'Code:' line?

arch/mips/kernel/traps.c show_code() is a good example. It prints

Code: xxxxxxxx xxxxxxxx xxxxxxxx <xxxxxxxx> xxxxxxxx xxxxxxxx

So it prints a few words either side of the failing instruction and marks the failure by '<>'.