Re: [PATCH] mips: irq: add stackoverflow detection

Previous thread: [PATCH] perf: fix possible divide-by-zero in perf_swevent_overflow() by Dongdong Deng on Thursday, August 26, 2010 - 5:07 am. (9 messages)

Next thread: [Patch 00/25] GRU - GRU Updates - Production Driver by steiner on Thursday, August 26, 2010 - 6:19 am. (26 messages)
From: Adam Jiang
Date: Thursday, August 26, 2010 - 6:19 am

Add stackoverflow detection to mips arch
---
 arch/mips/Kconfig.debug |    7 +++++++
 arch/mips/kernel/irq.c  |   19 +++++++++++++++++++
 2 files changed, 26 insertions(+), 0 deletions(-)

diff --git a/arch/mips/Kconfig.debug b/arch/mips/Kconfig.debug
index 43dc279..f1a00a2 100644
--- a/arch/mips/Kconfig.debug
+++ b/arch/mips/Kconfig.debug
@@ -67,6 +67,13 @@ config CMDLINE_OVERRIDE
 
 	  Normally, you will choose 'N' here.
 
+config DEBUG_STACKOVERFLOW
+	bool "Check for stack overflows"
+	depends on DEBUG_KERNEL
+	help
+	  This option will cause messages to be printed if free stack space
+	  drops below a certain limit.
+
 config DEBUG_STACK_USAGE
 	bool "Enable stack utilization instrumentation"
 	depends on DEBUG_KERNEL
diff --git a/arch/mips/kernel/irq.c b/arch/mips/kernel/irq.c
index c6345f5..8fdf79e 100644
--- a/arch/mips/kernel/irq.c
+++ b/arch/mips/kernel/irq.c
@@ -151,6 +151,25 @@ void __init init_IRQ(void)
 #endif
 }
 
+#ifdef DEBUG_STACKOVERFLOW
+static inline void check_stack_overflow(void)
+{
+	long sp;
+
+	asm volatile("move %0, $sp" : "=r" (sp));
+	sp = sp & (THREAD_SIZE-1);
+
+	/* check for stack overflow: is there less than 2KB free? */
+	if (unlikely(sp < (sizeof(struct thread_info) + 2048))) {
+		printk("do_IRQ: stack overflow: %ld\n",
+		       sp - sizeof(struct thread_info));
+		dump_stack();
+	}
+}
+#else
+static inline void check_stack_overflow(void) {}
+#endif
+
 /*
  * do_IRQ handles all normal device IRQ's (the special
  * SMP cross-CPU interrupts have their own specific
-- 
1.7.1

--

From: Sergei Shtylyov
Date: Thursday, August 26, 2010 - 12:00 pm

Hello.



    You've dropped call check_stack_overflow() call in do_IRQ() -- was that 
intentional?

WBR, Sergei
--

From: Adam Jiang
Date: Thursday, August 26, 2010 - 4:30 pm

Of course NO. It is awful, I have say.

I spent a lot of time to struggle with git yesterday. When I revised the
coding style problem and tried to format-patch to master branch, then I
got 2 patches but not all in one. How could I rebase my working copy and
get an all-in-one patch of this?

I tried to revert my commit and redo this patch, then I lost to calling
it in do_IRQ();

Best regards,
--

Previous thread: [PATCH] perf: fix possible divide-by-zero in perf_swevent_overflow() by Dongdong Deng on Thursday, August 26, 2010 - 5:07 am. (9 messages)

Next thread: [Patch 00/25] GRU - GRU Updates - Production Driver by steiner on Thursday, August 26, 2010 - 6:19 am. (26 messages)