linux-mips
[Top] [All Lists]

Re: stack backtrace

To: fxzhang@ict.ac.cn
Subject: Re: stack backtrace
From: Atsushi Nemoto <anemo@mba.ocn.ne.jp>
Date: Mon, 03 Apr 2006 01:17:11 +0900 (JST)
Cc: linux-mips@linux-mips.org
In-reply-to: <442FE1CA.4030905@ict.ac.cn>
Original-recipient: rfc822;linux-mips@linux-mips.org
References: <442FE1CA.4030905@ict.ac.cn>
Sender: linux-mips-bounce@linux-mips.org
>>>>> On Sun, 02 Apr 2006 22:38:02 +0800, Fuxin Zhang <fxzhang@ict.ac.cn> said:

>     Instead for my need I just hack up a simple version of way 1, with
> frame pointer kept on: CONFIG_FRAME_POINTER.

> BTW:It seems nobody use this option for MIPS? Is it dangerous? The size
> and performance overhead should be barable at most time for debugging?

> here is the code patch(just for reference), it depends on
> CONFIG_KALLSYMS too.

The get_frame_info() in process.c in kernel 2.6.16 no longer depends
on a frame pointer.  It would fit your needs better.  I think you can
use it with slight modifications.

BTW, Is there any point using -fno-omit-frame-pointer on MIPS now?
CONFIG_SCHED_NO_NO_OMIT_FRAME_POINTER=y is better for MIPS, isn't it?

---
Atsushi Nemoto

<Prev in Thread] Current Thread [Next in Thread>