[Top] [All Lists]

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

To: Patrik Kluba <>
Subject: Re: [PATCH -v4 9/9] tracing: add function graph tracer support for MIPS
From: Wu Zhangjin <>
Date: Sun, 25 Oct 2009 22:22:25 +0800
Cc:, GCC Patches <>, Adam Nemet <>,,, Thomas Gleixner <>, Ralf Baechle <>, Nicholas Mc Guire <>, Richard Sandiford <>, David Daney <>
Dkim-signature: v=1; a=rsa-sha256; c=relaxed/relaxed;; s=gamma; h=domainkey-signature:received:received:subject:from:reply-to:to:cc :in-reply-to:references:content-type:organization:date:message-id :mime-version:x-mailer:content-transfer-encoding; bh=Sau/RyEn42TsgodGkilMI9ZUiNRa8ZzmJDiHudxPAWc=; b=Dbg9cyK6xieS8w24+P/YHb7CCs5kwyRMrjWGzeuALFiesQfmYjQO0B4qzMvmpQKUVe Ti6wxySAU6jJBGGT9rOELpIpEFOt/Fh8vbQ6P/izzD2DGeEgaypvdWLWX0luEuK3Stez pZYhlDZbNP+jxGXUrw9PWMYjD90Zdwfz9FGnA=
Domainkey-signature: a=rsa-sha1; c=nofws;; s=gamma; h=subject:from:reply-to:to:cc:in-reply-to:references:content-type :organization:date:message-id:mime-version:x-mailer :content-transfer-encoding; b=tNjvgXPO1zu1eQPVEbsJh1xkq6AbdXIFfyhfFI2yhXFu7/qYAQoGeYKXS6G74h4EdN jjgnL8TskI+S1Fx8KsDCjbUtFnre6bCtq0a9fkrETMnUjHTEmzz3DC50ShkuO5ZOqh+U m0QF37sMuTpHeeWcZb4qI0aiI+SLkaaqyJiak=
In-reply-to: <>
Organization: DSLab, Lanzhou University, China
Original-recipient: rfc822;
References: <> <> <> <> <> <> <> <> <1256467717.6143.13.camel@falcon> <>
On Sun, 2009-10-25 at 14:37 +0100, Patrik Kluba wrote:
> On Sun, Oct 25, 2009 at 11:48 AM, Wu Zhangjin <> wrote:
> >
> > do you mean if enabling PROFILE_BEFORE_PROLOGUE, there will be some
> > problems with module support using -mlong-calls?
> >
> No, there are no problems. I've tested it on friday, and function
> graph tracing was working correctly.
> I meant to say that 4.2.1 we use does not generate correct profile
> calls from kernel modules. Maybe this issue was fixed in newer
> releases, I did not check. I've applied a patch (don't remember where
> have I found that, maybe it was created by you) to our toolchain
> several months ago.

I have never sent a patch to gcc before :-) but perhaps somebody have
fixed it for us. so, the left job is hoping somebody enable
PROFILE_BEFORE_PROLOGUE for MIPS in the next version of gcc if there is
no side effect, and then we can hijack the return address of non-leaf &
leaf function directly in the same way in _mcount.

> I was thinking about dynamic tracing, and I think a toolchain patch
> can be avoided completely. We only need to make difference between
> "jal _mcount" and "jalr v1"-style mcount calls when replacing them
> with "nop" instructions in the code-patching function called by
> ftrace_convert_nops(). This can be done in 2 ways:
> 1) keeping old instructions - takes extra memory, not an option
> 2) using 2 separate instructions to replace with. One of them could be
> the normal NOP instruction, which expands to "sll r0, r0, 0". For the
> other we could use "sll r0, r0, 1" but as it has already special
> meaning (SSNOP) a better candidate could be something like "sll r1,
> r1, 0". This way we can decide which instruction to patch in when
> tracing is enabled for a function, eg. when the code patcher
> encounters a "sll r0, r0, 0" it emits a function call using JAL and
> when it encounters "sll r1, r1, 0" it emits a function call using
> "JALR v1".

If only thinking about dynamic tracing, no patch for gcc needed,
-mlong-calls is enough, I have done it via a "stupid" trick, will send
the patchset out asap :-)

        Wu Zhangjin

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