Message-ID: <4D15545A.3020800@suse.com>
Date: Fri, 24 Dec 2010 21:18:02 -0500
From: Jeff Mahoney <jeffm@suse.com>
User-Agent: Mozilla/5.0 (X11; U; Linux x86_64; en-US; rv:1.9.2.12) Gecko/20101026 SUSE/3.1.6 Thunderbird/3.1.6
MIME-Version: 1.0
To: Dan Carpenter <error27@gmail.com>, David Miller <davem@davemloft.net>,
        linux-kernel@vger.kernel.org, balbir@linux.vnet.ibm.com,
        Andrew Morton <akpm@linux-foundation.org>
Subject: Re: taskstats alignment...
References: <20101223.093022.193727098.davem@davemloft.net> <4D14EA49.3070005@suse.com> <20101224211529.GA19019@bicker>
In-Reply-To: <20101224211529.GA19019@bicker>
Content-Type: text/plain; charset=ISO-8859-1
Content-Transfer-Encoding: 7bit
Sender: linux-kernel-owner@vger.kernel.org
Content-Length: 2437
Lines: 70

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On 12/24/2010 04:16 PM, Dan Carpenter wrote:
> On Fri, Dec 24, 2010 at 01:45:29PM -0500, Jeff Mahoney wrote:
> v> -----BEGIN PGP SIGNED MESSAGE-----
>> Hash: SHA1
>>
>> On 12/23/2010 12:30 PM, David Miller wrote:
>>>
>>> Re: commit 4be2c95d1f7706ca0e74499f2bd118e1cee19669
>>>
>>> Pretty much every 64-bit architecture other than
>>> powerpc64 and x86-64 needs that code, not just
>>> IA64.
>>>
>>> Better check would be:
>>>
>>> CONFIG_64BIT && !CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS
>>>
>>> Otherwise we'll be twiddling that ifdef endlessly as each
>>> and every other 64-bit platform bumps into this issue.
>>>
>>> So please could you change this to use a more sane check?
>>
>> I don't have an objection to it, but I've been pushing that we make the
>> change universal from the beginning of the discussion.
>>
>> The issue is that it causes breakage on apps that aren't following the
>> interface properly. iotop, in particular, has hard-coded offsets into
>> the packet to fish out the taskstats structure.
>>
>> So, if the goal of not breaking x86_64 is good enough, I'm fine with
>> this change.
> 
> Didn't you say something along the lines that if they don't have
> CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS then there is a warning message
> printed in dmesg?  I thought that was what prompted you to change the
> alignment in the first place.  It sound like those arches are already
> broken so David's suggestion would be a clear improvement over the 
> current code.

Yes, that is the original reason for the patch. The discussion was
surrounding which arches to break, since I made it universal. Changing
it to CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS (didn't know that existed
before) makes sense.

> BTW, since you're redoing the patch, it would be good if you pasted the
> warning message into the changelog.

Ok, that's easy enough.

- -Jeff

- -- 
Jeff Mahoney
SUSE Labs
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2.0.16 (GNU/Linux)
Comment: Using GnuPG with SUSE - http://enigmail.mozdev.org/

iEYEARECAAYFAk0VVFgACgkQLPWxlyuTD7JD2gCgi99mSqA29g5k9b9yj7od67mk
rkAAnR5Stz8fY4eYZbRJetz+fyH8D3iz
=zgux
-----END PGP SIGNATURE-----
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/