All of lore.kernel.org
 help / color / mirror / Atom feed
From: Robert Richter <robert.richter@amd.com>
To: Ingo Molnar <mingo@elte.hu>
Cc: LKML <linux-kernel@vger.kernel.org>,
	oprofile-list <oprofile-list@lists.sourceforge.net>
Subject: Re: [PATCH 4/5] oprofile: Remove exit function for timer mode
Date: Wed, 19 Oct 2011 18:33:28 +0200	[thread overview]
Message-ID: <20111019163328.GG24346@erda.amd.com> (raw)
In-Reply-To: <20111018061346.GD16304@elte.hu>

On 18.10.11 02:13:47, Ingo Molnar wrote:
> 
> * Robert Richter <robert.richter@amd.com> wrote:
> 
> >  	/* failed */
> > -	if (timer_mode)
> > -		oprofile_timer_exit();
> > -	else
> > +	if (!timer_mode)
> >  		oprofile_arch_exit();
> 
> > -	if (timer_mode)
> > -		oprofile_timer_exit();
> > -	else
> > +	if (!timer_mode)
> >  		oprofile_arch_exit();
> 
> ... and the ugliness didnt improve much here either.
> 
> Any reason why oprofile_arch_exit() couldnt check timer_mode?

The two fixes are the smallest possible solution for stable kernels.
... and they don't look nice.

Moving the check to oprofile_arch_exit() would mean to make timer_mode
global and to touch all oprofile_arch_exit() functions for all archs.
But we could get rid of timer_mode if the exit function would be part
of struct oprofile_operations. For this we would have to touch the
code for all architectures too, and I actually want to avoid this. But
it would be a nice solution.

Anyway, I have some improvements in patch #5 that make the code looks a
bit better. See below for a preview, will post a v2 patch set after
testing. As said, I could remove variable timer_mode completely by
moving exit functions to oprofile_operations. This could be on top of
the changes below.

-Robert



diff --git a/drivers/oprofile/oprof.c b/drivers/oprofile/oprof.c
index f7cd069..d2994e7 100644
--- a/drivers/oprofile/oprof.c
+++ b/drivers/oprofile/oprof.c
@@ -246,26 +246,23 @@ static int __init oprofile_init(void)
        int err;
 
        /* always init architecture to setup backtrace support */
+       timer_mode = 0;
        err = oprofile_arch_init(&oprofile_ops);
-
-       timer_mode = err || timer;      /* fall back to timer mode on errors */
-       if (timer_mode) {
-               if (!err)
-                       oprofile_arch_exit();
-               err = oprofile_timer_init(&oprofile_ops);
-               if (err)
-                       return err;
+       if (!err) {
+               if (!timer && oprofilefs_register())
+                       return 0;
+               oprofile_arch_exit();
        }
 
-       err = oprofilefs_register();
-       if (!err)
-               return 0;
-
-       /* failed */
-       if (!timer_mode)
-               oprofile_arch_exit();
+       /* setup timer mode */
+       timer_mode = 1;
+       err = op_nmi_timer_init(&oprofile_ops);
+       if (err)
+               err = oprofile_timer_init(&oprofile_ops);
+       if (err)
+               return err;
 
-       return err;
+       return oprofilefs_register();
 }
 
 




-- 
Advanced Micro Devices, Inc.
Operating System Research Center


  reply	other threads:[~2011-10-19 16:35 UTC|newest]

Thread overview: 13+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2011-10-17 14:26 [PATCH 0/5] oprofile: Updates and fixes for timer modes Robert Richter
2011-10-17 14:27 ` [PATCH 1/5] oprofile, x86: Add kernel parameter Robert Richter
2011-10-17 14:27 ` [PATCH 2/5] oprofile: Fix crash when unloading module (hr timer Robert Richter
2011-10-18  6:12   ` Ingo Molnar
2011-10-17 14:27 ` [PATCH 3/5] oprofile, x86: Fix crash when unloading module (nmi Robert Richter
2011-10-17 14:27 ` [PATCH 4/5] oprofile: Remove exit function for timer mode Robert Richter
2011-10-18  6:13   ` Ingo Molnar
2011-10-19 16:33     ` Robert Richter [this message]
2011-10-17 14:27 ` [PATCH 5/5] oprofile, x86: Reimplement nmi timer mode using perf Robert Richter
2011-10-18 10:59   ` Peter Zijlstra
2011-10-19 16:52     ` Robert Richter
2011-10-19 18:07   ` [PATCH 4/5] oprofile: Remove exit function for timer mode Robert Richter
2011-10-17 15:22 ` [PATCH 0/5] oprofile: Updates and fixes for timer modes Robert Richter

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=20111019163328.GG24346@erda.amd.com \
    --to=robert.richter@amd.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=mingo@elte.hu \
    --cc=oprofile-list@lists.sourceforge.net \
    /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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.