[PATCH] ppc64: Oops cleanup

From: Anton Blanchard <anton@samba.org>

Oops cleanup:

- Move prototypes into system.h
- Move the debugger hooks into die, all the calls sites were calling them.
- Handle bad values passed to prregs

4 files changed, 50 insertions, 47 deletions

diff --git a/arch/ppc64/kernel/traps.c b/arch/ppc64/kernel/traps.c
index 6074cae369e9..82d9b8b24e61 100644
--- a/arch/ppc64/kernel/traps.c
+++ b/arch/ppc64/kernel/traps.c
@@ -37,9 +37,6 @@
 #include <asm/processor.h>
 #include <asm/ppcdebug.h>
 
-extern int fix_alignment(struct pt_regs *);
-extern void bad_page_fault(struct pt_regs *, unsigned long, int);
-
 #ifdef CONFIG_PPC_PSERIES
 /* This is true if we are using the firmware NMI handler (typically LPAR) */
 extern int fwnmi_active;
@@ -67,11 +64,17 @@ EXPORT_SYMBOL(__debugger_fault_handler);
 
 static spinlock_t die_lock = SPIN_LOCK_UNLOCKED;
 
-void die(const char *str, struct pt_regs *regs, long err)
+int die(const char *str, struct pt_regs *regs, long err)
 {
 	static int die_counter;
 	int nl = 0;
 
+	if (debugger_fault_handler(regs))
+		return 1;
+
+	if (debugger(regs))
+		return 1;
+
 	console_verbose();
 	spin_lock_irq(&die_lock);
 	bust_spinlocks(1);
@@ -126,15 +129,16 @@ void die(const char *str, struct pt_regs *regs, long err)
 		panic("Fatal exception");
 	}
 	do_exit(SIGSEGV);
+
+	return 0;
 }
 
 static void
 _exception(int signr, siginfo_t *info, struct pt_regs *regs)
 {
 	if (!user_mode(regs)) {
-		if (debugger(regs))
+		if (die("Exception in kernel mode", regs, signr))
 			return;
-		die("Exception in kernel mode", regs, signr);
 	}
 
 	force_sig_info(signr, info, current);
@@ -188,8 +192,7 @@ SystemResetException(struct pt_regs *regs)
 	}
 #endif
 
-	if (!debugger(regs))
-		die("System Reset", regs, 0);
+	die("System Reset", regs, 0);
 
 	/* Must die if the interrupt is not recoverable */
 	if (!(regs->msr & MSR_RI))
@@ -246,9 +249,6 @@ static int recover_mce(struct pt_regs *regs, struct rtas_error_log err)
  *
  * On hardware prior to Power 4 these exceptions were asynchronous which
  * means we can't tell exactly where it occurred and so we can't recover.
- *
- * Note that the debugger should test RI=0 and warn the user that system
- * state has been corrupted.
  */
 void
 MachineCheckException(struct pt_regs *regs)
@@ -266,12 +266,11 @@ MachineCheckException(struct pt_regs *regs)
 	}
 #endif
 
-	if (debugger_fault_handler(regs))
-		return;
-	if (debugger(regs))
-		return;
+	die("Machine check", regs, 0);
 
-	die("Machine check in kernel mode", regs, 0);
+	/* Must die if the interrupt is not recoverable */
+	if (!(regs->msr & MSR_RI))
+		panic("Unrecoverable Machine check");
 }
 
 void
@@ -397,9 +396,6 @@ ProgramCheckException(struct pt_regs *regs)
 {
 	siginfo_t info;
 
-	if (debugger_fault_handler(regs))
-		return;
-
 	if (regs->msr & 0x100000) {
 		/* IEEE FP exception */
 
@@ -438,16 +434,18 @@ ProgramCheckException(struct pt_regs *regs)
 	}
 }
 
-void
-KernelFPUnavailableException(struct pt_regs *regs)
+void KernelFPUnavailableException(struct pt_regs *regs)
 {
-	die("Unrecoverable FP Unavailable Exception in Kernel", regs, 0);
+	printk(KERN_EMERG "Unrecoverable FP Unavailable Exception "
+			  "%lx at %lx\n", regs->trap, regs->nip);
+	die("Unrecoverable FP Unavailable Exception", regs, SIGABRT);
 }
 
-void
-KernelAltivecUnavailableException(struct pt_regs *regs)
+void KernelAltivecUnavailableException(struct pt_regs *regs)
 {
-	die("Unrecoverable VMX/Altivec Unavailable Exception in Kernel", regs, 0);
+	printk(KERN_EMERG "Unrecoverable VMX/Altivec Unavailable Exception "
+			  "%lx at %lx\n", regs->trap, regs->nip);
+	die("Unrecoverable VMX/Altivec Unavailable Exception", regs, SIGABRT);
 }
 
 void
@@ -539,7 +537,6 @@ void unrecoverable_exception(struct pt_regs *regs)
 {
 	printk(KERN_EMERG "Unrecoverable exception %lx at %lx\n",
 	       regs->trap, regs->nip);
-	debugger(regs);
 	die("Unrecoverable exception", regs, SIGABRT);
 }
 
@@ -551,7 +548,6 @@ void kernel_bad_stack(struct pt_regs *regs)
 {
 	printk(KERN_EMERG "Bad kernel stack pointer %lx at %lx\n",
 	       regs->gpr[1], regs->nip);
-	debugger(regs);
 	die("Bad kernel stack pointer", regs, SIGABRT);
 }
 
diff --git a/arch/ppc64/mm/fault.c b/arch/ppc64/mm/fault.c
index 4fd5100acff1..73712c143825 100644
--- a/arch/ppc64/mm/fault.c
+++ b/arch/ppc64/mm/fault.c
@@ -37,8 +37,6 @@
 #include <asm/system.h>
 #include <asm/uaccess.h>
 
-void bad_page_fault(struct pt_regs *, unsigned long, int);
-
 /*
  * The error_code parameter is
  *  - DSISR for a non-SLB data access fault,
@@ -177,10 +175,8 @@ do_sigbus:
  * It is called from do_page_fault above and from some of the procedures
  * in traps.c.
  */
-void
-bad_page_fault(struct pt_regs *regs, unsigned long address, int sig)
+void bad_page_fault(struct pt_regs *regs, unsigned long address, int sig)
 {
-	extern void die(const char *, struct pt_regs *, long);
 	const struct exception_table_entry *entry;
 
 	/* Are we prepared to handle this fault?  */
@@ -190,7 +186,5 @@ bad_page_fault(struct pt_regs *regs, unsigned long address, int sig)
 	}
 
 	/* kernel has accessed a bad area */
-	if (debugger(regs))
-		return;
 	die("Kernel access of bad area", regs, sig);
 }
diff --git a/arch/ppc64/xmon/xmon.c b/arch/ppc64/xmon/xmon.c
index 8bf490b348f1..90fe14bacc59 100644
--- a/arch/ppc64/xmon/xmon.c
+++ b/arch/ppc64/xmon/xmon.c
@@ -542,8 +542,7 @@ cmds(struct pt_regs *excp)
 			symbol_lookup();
 			break;
 		case 'r':
-			if (excp != NULL)
-				prregs(excp);	/* print regs */
+			prregs(excp);	/* print regs */
 			break;
 		case 'e':
 			if (excp == NULL)
@@ -966,8 +965,7 @@ static void backtrace(struct pt_regs *excp)
 
 spinlock_t exception_print_lock = SPIN_LOCK_UNLOCKED;
 
-void
-excprint(struct pt_regs *fp)
+void excprint(struct pt_regs *fp)
 {
 	unsigned long flags;
 
@@ -1002,21 +1000,31 @@ excprint(struct pt_regs *fp)
 	spin_unlock_irqrestore(&exception_print_lock, flags);
 }
 
-void
-prregs(struct pt_regs *fp)
+void prregs(struct pt_regs *fp)
 {
 	int n;
 	unsigned long base;
 
 	if (scanhex((void *)&base))
 		fp = (struct pt_regs *) base;
-	for (n = 0; n < 16; ++n)
-		printf("R%.2ld = %.16lx   R%.2ld = %.16lx\n", n, fp->gpr[n],
-		       n+16, fp->gpr[n+16]);
-	printf("pc  = %.16lx   msr = %.16lx\nlr  = %.16lx   cr  = %.16lx\n",
-	       fp->nip, fp->msr, fp->link, fp->ccr);
-	printf("ctr = %.16lx   xer = %.16lx   trap = %8lx\n",
-	       fp->ctr, fp->xer, fp->trap);
+
+	if (setjmp(bus_error_jmp) == 0) {
+		__debugger_fault_handler = handle_fault;
+		sync();
+		for (n = 0; n < 16; ++n)
+			printf("R%.2ld = %.16lx   R%.2ld = %.16lx\n", n,
+			       fp->gpr[n], n+16, fp->gpr[n+16]);
+		printf("pc  = %.16lx   msr = %.16lx\nlr  = %.16lx   "
+		       "cr  = %.16lx\n", fp->nip, fp->msr, fp->link, fp->ccr);
+		printf("ctr = %.16lx   xer = %.16lx   trap = %8lx\n",
+		       fp->ctr, fp->xer, fp->trap);
+
+		sync();
+		/* wait a little while to see if we get a machine check */
+		__delay(200);
+	} else {
+		printf("*** Error reading regs\n");
+	}
 }
 
 void
diff --git a/include/asm-ppc64/system.h b/include/asm-ppc64/system.h
index 9d732aa1256e..152a59acbc35 100644
--- a/include/asm-ppc64/system.h
+++ b/include/asm-ppc64/system.h
@@ -94,7 +94,12 @@ static inline int debugger_dabr_match(struct pt_regs *regs) { return 0; }
 static inline int debugger_fault_handler(struct pt_regs *regs) { return 0; }
 #endif
 
+extern int fix_alignment(struct pt_regs *regs);
+extern void bad_page_fault(struct pt_regs *regs, unsigned long address,
+			   int sig);
 extern void show_regs(struct pt_regs * regs);
+extern int die(const char *str, struct pt_regs *regs, long err);
+
 extern void flush_instruction_cache(void);
 extern int _get_PVR(void);
 extern void giveup_fpu(struct task_struct *);