This is the mail archive of the
systemtap@sourceware.org
mailing list for the systemtap project.
Re: [PATCH] systemtap: fix up on_each_cpu() for kernels 2.6.26+
- From: James Bottomley <James dot Bottomley at HansenPartnership dot com>
- To: systemtap at sourceware dot org
- Cc: linux-kernel <linux-kernel at vger dot kernel dot org>
- Date: Thu, 17 Jul 2008 11:53:11 -0500
- Subject: Re: [PATCH] systemtap: fix up on_each_cpu() for kernels 2.6.26+
- Dkim-signature: v=1; a=rsa-sha256; c=simple/simple; d=hansenpartnership.com; s=2007; t=1216313593; bh=VI34UfCiXlX7QoGg4ZauZ71DMJzZysY7W4xUj2eJHw o=; l=823; h=Subject:From:To:Cc:In-Reply-To:References:Content-Type: Date:Message-Id:Mime-Version:Content-Transfer-Encoding; b=eQG91Vzt q1bbVKCfk0f7vMqPhYeQm5DyLoqmbJ69V22yaEIIRauZHm/gCgn9rmwh4TW3+adA3xi K70Bwz3zKBz5UVjNz7E6STgq4aUZd5lei7r4WiES2zdrJ6IhmYuyOtcOwVzdXDRAKHb T3XbIFnGZ9W+TJ1CfNEVVyjH+jtcs=
- Dkim-signature: v=1; a=rsa-sha256; c=simple/simple; d=hansenpartnership.com; s=2007; t=1216313592; bh=VI34UfCiXlX7QoGg4ZauZ71DMJzZysY7W4xUj2eJHw o=; l=823; h=Subject:From:To:Cc:In-Reply-To:References:Content-Type: Date:Message-Id:Mime-Version:Content-Transfer-Encoding; b=PAIi62Zk d6bolpRXquoOQJUahGSd2j3MfhTjtXIvel1cQMhH4GVSZwdymDGLrH9NJyxqrjsNX/X 9LJq2Y3VzJj0dXy8j9cvue0BLZzLuvk9cMROyndbvAddMJSlxbMbv1qt1OuP07U+JG0 44+jvlpyj/pa97/AtrTfgh2fStIEg=
- References: <1216313476.5515.16.camel@localhost.localdomain>
On Thu, 2008-07-17 at 11:51 -0500, James Bottomley wrote:
> In kernel 2.6.26, this patch
>
> commit 15c8b6c1aaaf1c4edd67e2f02e4d8e1bd1a51c0d
> Author: Jens Axboe <jens.axboe@oracle.com>
> Date: Fri May 9 09:39:44 2008 +0200
>
> on_each_cpu(): kill unused 'retry' parameter
>
> means that runtime/time.c is now using the wrong calling conventions.
> Fix this up and surround it by kernel versioning #ifdefs.
By the way, this is a classic illustration of the fragility problem in
holding the systemtap runtime outside of the kernel. If it had been
in-kernel, all this would be fixed up and running and no-one would even
have noticed.
At least with changes in argument numbers, the compile breaks ... it
would have been a lot nastier if one of the variables simply changed
meaning ...
James