[PATCH] x86_64: Rewrite exception stack backtracing

Exceptions and hardware interrupts can, to a certain degree, nest, so when
attempting to follow the sequence of stacks used in order to dump their
contents this has to be accounted for.  Also, IST stacks have their tops
stored in the TSS, so there's no need to add the stack size to get to their
ends.

Minor changes from AK.

Signed-off-by: Jan Beulich <jbeulich@novell.com>
Signed-off-by: Andi Kleen <ak@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

arch/x86_64/kernel/traps.c

@@ -120,95 +120,106 @@ int printk_address(unsigned long address)
 } 
 #endif
 
-unsigned long *in_exception_stack(int cpu, unsigned long stack) 
+static unsigned long *in_exception_stack(unsigned cpu, unsigned long stack,
-{ 
+					unsigned *usedp, const char **idp)
-	int k;
+{
-	for (k = 0; k < N_EXCEPTION_STACKS; k++) {
+	static const char ids[N_EXCEPTION_STACKS][8] = {
-		struct tss_struct *tss = &per_cpu(init_tss, cpu);
+		[DEBUG_STACK - 1] = "#DB",
-		unsigned long start = tss->ist[k] - EXCEPTION_STKSZ;
+		[NMI_STACK - 1] = "NMI",
+		[DOUBLEFAULT_STACK - 1] = "#DF",
+		[STACKFAULT_STACK - 1] = "#SS",
+		[MCE_STACK - 1] = "#MC",
+	};
+	unsigned k;
 
-		if (stack >= start && stack < tss->ist[k])
+	for (k = 0; k < N_EXCEPTION_STACKS; k++) {
-			return (unsigned long *)tss->ist[k];
+		unsigned long end;
+
+		end = per_cpu(init_tss, cpu).ist[k];
+		if (stack >= end)
+			continue;
+		if (stack >= end - EXCEPTION_STKSZ) {
+			if (*usedp & (1U << k))
+				break;
+			*usedp |= 1U << k;
+			*idp = ids[k];
+			return (unsigned long *)end;
+		}
 	}
 	return NULL;
-} 
+}
 
 /*
  * x86-64 can have upto three kernel stacks: 
  * process stack
  * interrupt stack
- * severe exception (double fault, nmi, stack fault) hardware stack
+ * severe exception (double fault, nmi, stack fault, debug, mce) hardware stack
- * Check and process them in order.
  */
 
 void show_trace(unsigned long *stack)
 {
 	unsigned long addr;
-	unsigned long *irqstack, *irqstack_end, *estack_end;
+	const unsigned cpu = safe_smp_processor_id();
-	const int cpu = safe_smp_processor_id();
+	unsigned long *irqstack_end = (unsigned long *)cpu_pda[cpu].irqstackptr;
 	int i;
+	unsigned used = 0;
 
 	printk("\nCall Trace:");
-	i = 0; 
+
-	
+#define HANDLE_STACK(cond) \
-	estack_end = in_exception_stack(cpu, (unsigned long)stack); 
+	do while (cond) { \
-	if (estack_end) { 
+		addr = *stack++; \
-		while (stack < estack_end) { 
+		if (kernel_text_address(addr)) { \
-			addr = *stack++; 
+			/* \
-			if (__kernel_text_address(addr)) {
+			 * If the address is either in the text segment of the \
-				i += printk_address(addr);
+			 * kernel, or in the region which contains vmalloc'ed \
-				i += printk(" "); 
+			 * memory, it *may* be the address of a calling \
-				if (i > 50) {
+			 * routine; if so, print it so that someone tracing \
-					printk("\n"); 
+			 * down the cause of the crash will be able to figure \
-					i = 0;
+			 * out the call path that was taken. \
-				}
+			 */ \
+			i += printk_address(addr); \
+			if (i > 50) { \
+				printk("\n       "); \
+				i = 0; \
+			} \
+			else \
+				i += printk(" "); \
+		} \
+	} while (0)
+
+	for(i = 0; ; ) {
+		const char *id;
+		unsigned long *estack_end;
+		estack_end = in_exception_stack(cpu, (unsigned long)stack,
+						&used, &id);
+
+		if (estack_end) {
+			i += printk(" <%s> ", id);
+			HANDLE_STACK (stack < estack_end);
+			i += printk(" <EOE> ");
+			stack = (unsigned long *) estack_end[-2];
+			continue;
+		}
+		if (irqstack_end) {
+			unsigned long *irqstack;
+			irqstack = irqstack_end -
+				(IRQSTACKSIZE - 64) / sizeof(*irqstack);
+
+			if (stack >= irqstack && stack < irqstack_end) {
+				i += printk(" <IRQ> ");
+				HANDLE_STACK (stack < irqstack_end);
+				stack = (unsigned long *) (irqstack_end[-1]);
+				irqstack_end = NULL;
+				i += printk(" <EOI> ");
+				continue;
 			}
 		}
-		i += printk(" <EOE> "); 
+		break;
-		i += 7;
-		stack = (unsigned long *) estack_end[-2]; 
-	}  
-
-	irqstack_end = (unsigned long *) (cpu_pda[cpu].irqstackptr);
-	irqstack = (unsigned long *) (cpu_pda[cpu].irqstackptr - IRQSTACKSIZE + 64);
-
-	if (stack >= irqstack && stack < irqstack_end) {
-		printk("<IRQ> ");  
-		while (stack < irqstack_end) {
-			addr = *stack++;
-			/*
-			 * If the address is either in the text segment of the
-			 * kernel, or in the region which contains vmalloc'ed
-			 * memory, it *may* be the address of a calling
-			 * routine; if so, print it so that someone tracing
-			 * down the cause of the crash will be able to figure
-			 * out the call path that was taken.
-			 */
-			 if (__kernel_text_address(addr)) {
-				 i += printk_address(addr);
-				 i += printk(" "); 
-				 if (i > 50) { 
-					printk("\n       ");
-					 i = 0;
-				 } 
-			}
-		} 
-		stack = (unsigned long *) (irqstack_end[-1]);
-		printk(" <EOI> ");
-		i += 7;
-	} 
-
-	while (((long) stack & (THREAD_SIZE-1)) != 0) {
-		addr = *stack++;
-		if (__kernel_text_address(addr)) {
-			i += printk_address(addr);
-			i += printk(" "); 
-			if (i > 50) { 
-				printk("\n       ");
-					 i = 0;
-			} 
-		}
 	}
+
+	HANDLE_STACK (((long) stack & (THREAD_SIZE-1)) != 0);
+#undef HANDLE_STACK
 	printk("\n");
 }