linux-kernel - Re: [PATCH 1/1]ext4: Fix WARN_ON

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: <20160704142938.GB29557@thunk.org>
Date:	Mon, 4 Jul 2016 10:29:38 -0400
From:	Theodore Ts'o <tytso@....edu>
To:	"Pranay Kr. Srivastava" <pranjas@...il.com>
Cc:	adilger.kernel@...ger.ca, linux-ext4@...r.kernel.org,
	linux-kernel@...r.kernel.org
Subject: Re: [PATCH 1/1]ext4: Fix WARN_ON_ONCE when marking buffer dirty

On Thu, Jun 30, 2016 at 02:12:30PM +0300, Pranay Kr. Srivastava wrote:
> Signed-off-by: Pranay Kr. Srivastava <pranjas@...il.com>

The description for why the change is being made should go in the
commit.  (No need to put the description in a separate cover letter.)
I ended up rewriting the commit description as follows, to make it
much more understandable:

    ext4: Fix WARN_ON_ONCE in ext4_commit_super()

    If there are racing calls to ext4_commit_super() it's possible for
    another writeback of the superblock to result in the buffer being
    marked with an error after we check if the buffer is marked as
    having a write error and the buffer up-to-date flag is set again.
    If that happens mark_buffer_dirty() can end up throwing a
    WARN_ON_ONCE.

    Fix this by moving this check to write before we call
    write_buffer_dirty(), and keeping the buffer locked during this
    whole sequence.

    Signed-off-by: Pranay Kr. Srivastava <pranjas@...il.com>
    Signed-off-by: Theodore Ts'o <tytso@....edu>

Note that the one-line summary needs to carry as much information as
possible so someone who is scanning the commits using git log
--oneline has a chance of understanding it.  This means the high-level
*why* of the commit, not a summary of what the changes in the C code.
Also note the increased context of when the misbehaviour could occur
in the commit description, which was missing in the cover letter.

When I'm processing patches, if I'm in a hurry, patches that require
extra work or which aren't Obviously Right, sometimes get deferred by
a few days.  This patch fell in that category.

Adding to the commit descrtipion additional context and/or
instructions for how to reproduce the problem you are trying to
remediate will often make life much easier for me, and accelerate how
quickly I'll get to your patch.

Cheers,

							- Ted