From: Oleg Nesterov <oleg@tv-sign.ru>
To: linux-kernel@vger.kernel.org
Cc: Kirill Korotaev <dev@sw.ru>, Andrew Morton <akpm@osdl.org>,
Linus Torvalds <torvalds@osdl.org>
Subject: [PATCH] alternate stack dump fix.
Date: Sun, 03 Oct 2004 20:00:56 +0400 [thread overview]
Message-ID: <41602238.A828A852@tv-sign.ru> (raw)
Hello.
Kirill Korotaev wrote:
>
> This patch fixes incorrect check for stack ptr in
> show_trace()->valid_stack_ptr(). When called from hardirq/softirq
> show_trace() prints "Stack pointer is garbage, not printing trace"
> message instead of call traces.
There is another problem in show_trace(). With CONFIG_FRAME_POINTER
every call to print_context_stack() will now print entire call chain,
switching the stacks transparently, beacause valid_stack_ptr()
now accepts ebp in irq stack.
Then show trace switch the stack, and calls print_context_stack()
again with the same value in ebp, and we have the same dump
after printk(" =======================\n").
What do you think about the following patch?
Against 2.6.9-rc3.
Oleg.
Signed-off-by: Oleg Nesterov <oleg@tv-sign.ru>
--- 2.6.9-rc3/arch/i386/kernel/traps.c~ Sun Oct 3 18:45:52 2004
+++ 2.6.9-rc3/arch/i386/kernel/traps.c Sun Oct 3 19:54:07 2004
@@ -107,36 +107,27 @@ int register_die_notifier(struct notifie
return err;
}
-static int valid_stack_ptr(struct task_struct *task, void *p)
+static int valid_stack_ptr(struct thread_info *tinfo, void *p)
{
- if (p <= (void *)task->thread_info)
- return 0;
- if (kstack_end(p))
- return 0;
- return 1;
+ return p > (void *)tinfo &&
+ p < (void *)tinfo + THREAD_SIZE - 3;
}
-#ifdef CONFIG_FRAME_POINTER
-static void print_context_stack(struct task_struct *task, unsigned long *stack,
+static unsigned long print_context_stack(struct thread_info *tinfo, unsigned long *stack,
unsigned long ebp)
{
unsigned long addr;
- while (valid_stack_ptr(task, (void *)ebp)) {
+#ifdef CONFIG_FRAME_POINTER
+ while (valid_stack_ptr(tinfo, (void *)ebp)) {
addr = *(unsigned long *)(ebp + 4);
printk(" [<%08lx>] ", addr);
print_symbol("%s", addr);
printk("\n");
ebp = *(unsigned long *)ebp;
}
-}
#else
-static void print_context_stack(struct task_struct *task, unsigned long *stack,
- unsigned long ebp)
-{
- unsigned long addr;
-
- while (!kstack_end(stack)) {
+ while (valid_stack_ptr(tinfo, stack)) {
addr = *stack++;
if (__kernel_text_address(addr)) {
printk(" [<%08lx>]", addr);
@@ -144,9 +135,11 @@ static void print_context_stack(struct t
printk("\n");
}
}
-}
#endif
+ return ebp;
+}
+
void show_trace(struct task_struct *task, unsigned long * stack)
{
unsigned long ebp;
@@ -154,11 +147,6 @@ void show_trace(struct task_struct *task
if (!task)
task = current;
- if (!valid_stack_ptr(task, stack)) {
- printk("Stack pointer is garbage, not printing trace\n");
- return;
- }
-
if (task == current) {
/* Grab ebp right from our regs */
asm ("movl %%ebp, %0" : "=r" (ebp) : );
@@ -171,7 +159,7 @@ void show_trace(struct task_struct *task
struct thread_info *context;
context = (struct thread_info *)
((unsigned long)stack & (~(THREAD_SIZE - 1)));
- print_context_stack(task, stack, ebp);
+ ebp = print_context_stack(context, stack, ebp);
stack = (unsigned long*)context->previous_esp;
if (!stack)
break;
next reply other threads:[~2004-10-03 15:54 UTC|newest]
Thread overview: 6+ messages / expand[flat|nested] mbox.gz Atom feed top
2004-10-03 16:00 Oleg Nesterov [this message]
2004-10-03 17:06 ` Andrew Morton
2004-10-04 8:22 ` Oleg Nesterov
2004-10-04 8:27 ` Andrew Morton
2004-10-04 9:28 ` Oleg Nesterov
2004-10-04 8:22 ` Oleg Nesterov
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=41602238.A828A852@tv-sign.ru \
--to=oleg@tv-sign.ru \
--cc=akpm@osdl.org \
--cc=dev@sw.ru \
--cc=linux-kernel@vger.kernel.org \
--cc=torvalds@osdl.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox
all inboxes | Powered by JetHome®