public inbox for linux-kernel@vger.kernel.org
 help / color / mirror / Atom feed
From: Adrian Bunk <bunk@stusta.de>
To: Mathieu Desnoyers <mathieu.desnoyers@polymtl.ca>
Cc: akpm@linux-foundation.org, linux-kernel@vger.kernel.org
Subject: Re: [patch 1/9] Conditional Calls - Architecture Independent Code
Date: Mon, 4 Jun 2007 21:01:02 +0200	[thread overview]
Message-ID: <20070604190102.GY5500@stusta.de> (raw)
In-Reply-To: <20070530140227.070136408@polymtl.ca>

On Wed, May 30, 2007 at 10:00:26AM -0400, Mathieu Desnoyers wrote:
> Conditional calls are used to compile in code that is meant to be dynamically
> enabled at runtime. When code is disabled, it has a very small footprint.
> 
> It has a generic cond_call version and optimized per architecture cond_calls.
> The optimized cond_call uses a load immediate to remove a data cache hit.
> 
> It adds a new rodata section "__cond_call" to place the pointers to the enable
> value.
>...

I have two questions for getting the bigger picture:

1. How much code will be changed?
Looking at the F00F bug fixup example, it seems we'll have to make
several functions in every single driver conditional in the kernel for 
getting the best performance.
How many functions to you plan to make conditional this way?

2. What is the real-life performance improvement?
That micro benchmarks comparing cache hits with cache misses give great 
looking numbers is obvious.
But what will be the performance improvement in real workloads after the
functions you plan to make conditional according to question 1 have been 
made conditional?

TIA
Adrian

-- 

       "Is there not promise of rain?" Ling Tan asked suddenly out
        of the darkness. There had been need of rain for many days.
       "Only a promise," Lao Er said.
                                       Pearl S. Buck - Dragon Seed


  parent reply	other threads:[~2007-06-04 19:01 UTC|newest]

Thread overview: 57+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2007-05-30 14:00 [patch 0/9] Conditional Calls - for 2.6.22-rc2-mm1 Mathieu Desnoyers
2007-05-30 14:00 ` [patch 1/9] Conditional Calls - Architecture Independent Code Mathieu Desnoyers
2007-05-30 20:32   ` Andrew Morton
2007-05-31 16:34     ` Mathieu Desnoyers
2007-05-31 13:47   ` Andi Kleen
2007-06-05 18:40     ` Mathieu Desnoyers
2007-06-04 19:01   ` Adrian Bunk [this message]
2007-06-13 15:57     ` Mathieu Desnoyers
2007-06-13 21:51       ` Adrian Bunk
2007-06-14 16:02         ` Mathieu Desnoyers
2007-06-14 21:06           ` Adrian Bunk
2007-06-20 21:59             ` Mathieu Desnoyers
2007-06-21 13:00               ` Adrian Bunk
2007-06-21 13:55                 ` Mathieu Desnoyers
2007-05-30 14:00 ` [patch 2/9] Conditional Calls - Hash Table Mathieu Desnoyers
2007-05-30 20:32   ` Andrew Morton
2007-05-31 13:42   ` Andi Kleen
2007-06-01 16:08     ` Matt Mackall
2007-06-01 16:46       ` Mathieu Desnoyers
2007-06-01 17:07         ` Matt Mackall
2007-06-01 17:45           ` Andi Kleen
2007-06-01 18:06             ` Mathieu Desnoyers
2007-06-01 18:49               ` Matt Mackall
2007-06-01 19:35               ` Andi Kleen
2007-06-01 20:33                 ` Mathieu Desnoyers
2007-06-01 20:44                   ` Andi Kleen
2007-06-04 22:26                     ` Mathieu Desnoyers
2007-06-01 18:03           ` Mathieu Desnoyers
2007-05-30 14:00 ` [patch 3/9] Conditional Calls - Non Optimized Architectures Mathieu Desnoyers
2007-05-30 14:00 ` [patch 4/9] Conditional Calls - Add kconfig menus Mathieu Desnoyers
2007-05-30 14:00 ` [patch 5/9] Conditional Calls - i386 Optimization Mathieu Desnoyers
2007-05-30 20:33   ` Andrew Morton
2007-05-31 13:54   ` Andi Kleen
2007-06-05 19:02     ` Mathieu Desnoyers
2007-05-30 14:00 ` [patch 6/9] Conditional Calls - PowerPC Optimization Mathieu Desnoyers
2007-05-30 14:00 ` [patch 7/9] Conditional Calls - Documentation Mathieu Desnoyers
2007-05-30 14:00 ` [patch 8/9] F00F bug fixup for i386 - use conditional calls Mathieu Desnoyers
2007-05-30 20:33   ` Andrew Morton
2007-05-31 21:07     ` Mathieu Desnoyers
2007-05-31 21:21       ` Andrew Morton
2007-05-31 21:38         ` Mathieu Desnoyers
2007-05-30 14:00 ` [patch 9/9] Scheduler profiling - Use " Mathieu Desnoyers
2007-05-30 20:34   ` Andrew Morton
2007-06-01 15:54     ` Mathieu Desnoyers
2007-06-01 16:19       ` Andrew Morton
2007-06-01 16:33         ` Mathieu Desnoyers
2007-05-30 20:59   ` William Lee Irwin III
2007-05-31 21:12     ` Mathieu Desnoyers
2007-05-31 23:41       ` William Lee Irwin III
2007-06-04 22:20         ` Mathieu Desnoyers
2007-05-30 21:44   ` Matt Mackall
2007-05-31 21:36     ` Mathieu Desnoyers
2007-05-31 13:39   ` Andi Kleen
2007-05-31 22:07     ` Mathieu Desnoyers
2007-05-31 22:33       ` Andi Kleen
2007-06-04 22:20         ` Mathieu Desnoyers
  -- strict thread matches above, loose matches on Subject: below --
2007-05-29 18:33 [patch 0/9] Conditional Calls Mathieu Desnoyers
2007-05-29 18:33 ` [patch 1/9] Conditional Calls - Architecture Independent Code Mathieu Desnoyers

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=20070604190102.GY5500@stusta.de \
    --to=bunk@stusta.de \
    --cc=akpm@linux-foundation.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=mathieu.desnoyers@polymtl.ca \
    /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