[v4,13/16] x86/dumpstack: Try harder to get a call trace on stack overflow
diff mbox

Message ID f76b49739625fce1264428155540ca7161925511.1466741835.git.luto@kernel.org
State New
Headers show

Commit Message

Andy Lutomirski June 24, 2016, 4:23 a.m. UTC
If we overflow the stack, print_context_stack will abort.  Detect
this case and rewind back into the valid part of the stack so that
we can trace it.

Signed-off-by: Andy Lutomirski <luto@kernel.org>
---
 arch/x86/kernel/dumpstack.c | 9 ++++++++-
 1 file changed, 8 insertions(+), 1 deletion(-)

Comments

Josh Poimboeuf June 24, 2016, 3:35 p.m. UTC | #1
On Thu, Jun 23, 2016 at 09:23:08PM -0700, Andy Lutomirski wrote:
> If we overflow the stack, print_context_stack will abort.  Detect
> this case and rewind back into the valid part of the stack so that
> we can trace it.
> 
> Signed-off-by: Andy Lutomirski <luto@kernel.org>

Reviewed-by: Josh Poimboeuf <jpoimboe@redhat.com>
Andy Lutomirski June 26, 2016, 4:59 p.m. UTC | #2
On Fri, Jun 24, 2016 at 8:35 AM, Josh Poimboeuf <jpoimboe@redhat.com> wrote:
> On Thu, Jun 23, 2016 at 09:23:08PM -0700, Andy Lutomirski wrote:
>> If we overflow the stack, print_context_stack will abort.  Detect
>> this case and rewind back into the valid part of the stack so that
>> we can trace it.
>>
>> Signed-off-by: Andy Lutomirski <luto@kernel.org>
>
> Reviewed-by: Josh Poimboeuf <jpoimboe@redhat.com>
>

FWIW, I'm making a trivial change here for v4: task->stack ->
task_stack_page(task).  Since it seems inconsequential, I'm keeping
your reviewed-by.

> --
> Josh

Patch
diff mbox

diff --git a/arch/x86/kernel/dumpstack.c b/arch/x86/kernel/dumpstack.c
index 4592bc4ed3e1..4538f7ca9072 100644
--- a/arch/x86/kernel/dumpstack.c
+++ b/arch/x86/kernel/dumpstack.c
@@ -87,7 +87,7 @@  static inline int valid_stack_ptr(struct task_struct *task,
 		else
 			return 0;
 	}
-	return p > t && p < t + THREAD_SIZE - size;
+	return p >= t && p < t + THREAD_SIZE - size;
 }
 
 unsigned long
@@ -98,6 +98,13 @@  print_context_stack(struct task_struct *task,
 {
 	struct stack_frame *frame = (struct stack_frame *)bp;
 
+	/*
+	 * If we overflowed the stack into a guard page, jump back to the
+	 * bottom of the usable stack.
+	 */
+	if ((unsigned long)task->stack - (unsigned long)stack < PAGE_SIZE)
+		stack = (unsigned long *)task->stack;
+
 	while (valid_stack_ptr(task, stack, sizeof(*stack), end)) {
 		unsigned long addr;