linux-ext4 - Re: [PATCH v1 0/4] fs: fix race between llseek SEEK

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] [day] [month] [year] [list]

Date:   Mon, 26 Nov 2018 11:54:06 +0900
From:   Eiichi Tsukata <devel@...ukata.com>
To:     Alexander Viro <viro@...iv.linux.org.uk>
Cc:     andi@...stfloor.org, Chris Mason <clm@...com>,
        Josef Bacik <josef@...icpanda.com>,
        David Sterba <dsterba@...e.com>,
        "Theodore Ts'o" <tytso@....edu>,
        Andreas Dilger <adilger.kernel@...ger.ca>,
        Jaegeuk Kim <jaegeuk@...nel.org>, Chao Yu <yuchao0@...wei.com>,
        Miklos Szeredi <miklos@...redi.hu>,
        Bob Peterson <rpeterso@...hat.com>,
        Andreas Gruenbacher <agruenba@...hat.com>,
        linux-btrfs@...r.kernel.org, linux-ext4@...r.kernel.org,
        linux-f2fs-devel@...ts.sourceforge.net,
        linux-fsdevel@...r.kernel.org, cluster-devel@...hat.com,
        linux-unionfs@...r.kernel.org
Subject: Re: [PATCH v1 0/4] fs: fix race between llseek SEEK_END and write

2018年11月22日(木) 16:06 Al Viro <viro@...iv.linux.org.uk>:
>
> Can you show me where does POSIX/SuS/whatever it's called these days promise
> that kind of atomicity?

No. I couldn't found it.
That's why I previously posted RFC Patch:
https://marc.info/?t=154237277900001&r=1&w=2
I wasn't sure this is a bug in the kernel or not.

> that kind of atomicity?  TBH, I would be very surprised if any Unix promised
> to have file size updated no more than once per write(2).  Any userland code
> that relies on such property is, as minimum, non-portable and I would argue
> that it is outright broken.

Thanks. Now It's clear. It is not a bug in the kernel, but in
userspace if `tail` assumes such
atomicity.

> Note, BTW, that your example depends upon rather non-obvious details of echo
> implementation, starting with the bufferization logics used by particular
> shell.  And AFAICS ash(1) ends up with possibility of more than one write(2)

I've never imagined such a difference in echo implementation, thanks.