linux-ext4 - Re: [PATCH 4/9] ext4: Drop pointless IO submission from ext4_bio_write

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: Windows password security audit tool. GUI, reports in PDF.

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

Date:   Thu, 1 Dec 2022 19:10:25 +0530
From:   "Ritesh Harjani (IBM)" <ritesh.list@...il.com>
To:     Jan Kara <jack@...e.cz>
Cc:     Ted Tso <tytso@....edu>, linux-ext4@...r.kernel.org,
        Christoph Hellwig <hch@...radead.org>
Subject: Re: [PATCH 4/9] ext4: Drop pointless IO submission from
 ext4_bio_write_page()

On 22/12/01 11:35AM, Jan Kara wrote:
> On Thu 01-12-22 12:36:55, Ritesh Harjani (IBM) wrote:
> > On 22/11/30 05:35PM, Jan Kara wrote:
> > > We submit outstanding IO in ext4_bio_write_page() if we find a buffer we
> > > are not going to write. This is however pointless because we already
> > > handle submission of previous IO in case we detect newly added buffer
> > > head is discontiguous. So just delete the pointless IO submission call.
> >
> > Agreed. io_submit_add_bh() is anyway called at the end for submitting buffers.
> > And io_submit_add_bh() also has the logic to:
> > 1. submit a discontiguous bio
> > 2. Also submit a bio if the bio gets full (submit_and_retry label).
> >
> > Hence calling ext4_io_submit() early is not required.
> >
> > I guess the same will also hold true for at this place.
> > https://elixir.bootlin.com/linux/v6.1-rc7/source/fs/ext4/page-io.c#L524
>
> So there the submission is needed because we are OOM and are going to wait
> for some memory to free. If we have some bio accumulated, it is pinning
> pages in writeback state and memory reclaim can be waiting on them. So if
> we don't submit, it is a deadlock possibility or at least asking for
> trouble.

Aah! right. I didn't see the ret == -ENOMEM compare there.

Thanks!
-ritesh