linux-kernel - Re: [RFC PATCH 2/3] mpage: use bio_for_each_folio_all in mpage_end

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite for Android: free password hash cracker in your pocket

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <7121c937-adea-a495-eb15-8e0d7fe0838c@samsung.com>
Date:   Thu, 16 Mar 2023 11:07:22 +0100
From:   Pankaj Raghav <p.raghav@...sung.com>
To:     Matthew Wilcox <willy@...radead.org>,
        Hannes Reinecke <hare@...e.de>
CC:     <hubcap@...ibond.com>, <senozhatsky@...omium.org>,
        <martin@...ibond.com>, <minchan@...nel.org>,
        <viro@...iv.linux.org.uk>, <brauner@...nel.org>, <axboe@...nel.dk>,
        <akpm@...ux-foundation.org>, <linux-kernel@...r.kernel.org>,
        <linux-block@...r.kernel.org>, <linux-fsdevel@...r.kernel.org>,
        <linux-mm@...ck.org>, <gost.dev@...sung.com>, <mcgrof@...nel.org>,
        <devel@...ts.orangefs.org>
Subject: Re: [RFC PATCH 2/3] mpage: use bio_for_each_folio_all in
 mpage_end_io()

On 2023-03-15 17:08, Matthew Wilcox wrote:
> On Wed, Mar 15, 2023 at 03:52:15PM +0100, Hannes Reinecke wrote:
>> On 3/15/23 13:32, Pankaj Raghav wrote:
>>> Use bio_for_each_folio_all to iterate through folios in a bio so that
>>> the folios can be directly passed to the folio_endio() function.
>>> +	bio_for_each_folio_all(fi, bio)
>>> +		folio_endio(fi.folio, bio_op(bio),
>>> +			    blk_status_to_errno(bio->bi_status));
>>>   	bio_put(bio);
>>>   }
>>
>> Ah. Here it is.
>>
>> I would suggest merge these two patches.
> 
> The right way to have handled this patch series was:
> 
> 1. Introduce a new folio_endio() [but see Christoph's mail on why we
> shouldn't do that]
> 2-n convert callers to use folios directly
> n+1 remove page_endio() entirely.
> 
> Note that patch n+1 might not be part of this patch series; sometimes
> it takes a while to convert all callers to use folios.
> 

This is definitely a much better way to handle these changes. Thanks willy.

> I very much dislike the way this was done by pushing the page_folio()
> call into each of the callers because it makes the entire series hard to
> review.