linux-mips
[Top] [All Lists]

Re: [PATCH -v4 9/9] tracing: add function graph tracer support for MIPS

To: rostedt@goodmis.org
Subject: Re: [PATCH -v4 9/9] tracing: add function graph tracer support for MIPS
From: David Daney <ddaney@caviumnetworks.com>
Date: Wed, 21 Oct 2009 10:48:50 -0700
Cc: Wu Zhangjin <wuzhangjin@gmail.com>, linux-kernel@vger.kernel.org, linux-mips@linux-mips.org, Thomas Gleixner <tglx@linutronix.de>, Ralf Baechle <ralf@linux-mips.org>, Nicholas Mc Guire <der.herr@hofr.at>
In-reply-to: <1256145813.18347.3210.camel@gandalf.stny.rr.com>
Original-recipient: rfc822;linux-mips@linux-mips.org
References: <028867b99ec532b84963a35e7d552becc783cafc.1256135456.git.wuzhangjin@gmail.com> <2f73eae542c47ac5bbb9f7280e6c0271d193e90d.1256135456.git.wuzhangjin@gmail.com> <3f0d3515f74a58f4cfd11e61b62a129fdc21e3a7.1256135456.git.wuzhangjin@gmail.com> <ea8aa927fbd184b54941e4c2ae0be8ea0b4f6b8a.1256135456.git.wuzhangjin@gmail.com> <96110ea5dd4d3d54eb97d0bb708a5bd81c7a50b5.1256135456.git.wuzhangjin@gmail.com> <5dda13e8e3a9c9dba4bb7179183941bda502604f.1256135456.git.wuzhangjin@gmail.com> <af3ec1b5cd06b6f6a461c9fa7d09a51fabccb08d.1256135456.git.wuzhangjin@gmail.com> <a6f2959a69b6a77dd32cc36a5c8202f97d524f1e.1256135456.git.wuzhangjin@gmail.com> <53bdfdd95ec4fa00d4cc505bb5972cf21243a14d.1256135456.git.wuzhangjin@gmail.com> <1256141540.18347.3118.camel@gandalf.stny.rr.com> <4ADF38D5.9060100@caviumnetworks.com> <1256143568.18347.3169.camel@gandalf.stny.rr.com> <4ADF3FE0.5090104@caviumnetworks.com> <1256145813.18347.3210.camel@gandalf.stny.rr.com>
Sender: linux-mips-bounce@linux-mips.org
User-agent: Thunderbird 2.0.0.21 (X11/20090320)
Steven Rostedt wrote:
On Wed, 2009-10-21 at 10:07 -0700, David Daney wrote:

I have not used -pg, so I don't know for sure, I think all it does is add the calls to _mcount. Someone could investigate -fno-omit-frame-pointer, with that you may be able to use:

Note, -pg assumes -fno-omit-frame-pointer, since -fomit-frame-pointer
and -pg are incompatible.

    move    s8,sp

To identify function prologs, but it would still be ad hoc, as modern versions of GCC will reorder instructions in the prolog for better scheduling.

I'll have to search the ABI documentation about calling _mcount in MIPS.
There are assumptions with _mcount that are made. It may very well be
safe to assume that the move s8,sp will always be there before an mcount
call.


Although I am quite interested in ftrace for MIPS, I a haven't spent the effort to try to figure out how many levels of backtrace we need to be generating here.

If you only need the caller's address, that is passed in register 'at'. If you need more than that, then things get tricky. I would do one of two things:

1) Use the mechanism used by the OOPS dump.

2) Implement back traces based on DWARF2 unwinding data that is generated by the compiler.

David Daney

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