netdev - Re: [PATCH v2] net/9p: use a dedicated spinlock for modifying IDR

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]

Message-ID: <YxSDeqn4LrSfSaUs@codewreck.org>
Date:   Sun, 4 Sep 2022 19:52:42 +0900
From:   Dominique Martinet <asmadeus@...ewreck.org>
To:     Tetsuo Handa <penguin-kernel@...ove.SAKURA.ne.jp>
Cc:     Eric Van Hensbergen <ericvh@...il.com>,
        Latchesar Ionkov <lucho@...kov.net>,
        Christian Schoenebeck <linux_oss@...debyte.com>,
        syzbot <syzbot+2f20b523930c32c160cc@...kaller.appspotmail.com>,
        v9fs-developer@...ts.sourceforge.net,
        syzkaller-bugs@...glegroups.com, netdev@...r.kernel.org
Subject: Re: [PATCH v2] net/9p: use a dedicated spinlock for modifying IDR

Tetsuo Handa wrote on Sun, Sep 04, 2022 at 07:02:04PM +0900:
> > In hindsight it's probably faster to just send a diff... Happy to keep
> > you as author if you'd like, or sign off or whatever you prefer -- I
> > wouldn't have guessed what that report meant without you.
> 
> This diff is bigger than I can guess correctness. Maybe v1 patch should be
> applied as a fix for this problem (because including Reported-by: or Fixes:
> likely makes patches be automatically picked up by stable kernel maintainers
> before syzbot tests for a while) and your patch should be applied as an improvement
> (i.e. omit Reported-by: and Fixes: ). You can manually request for inclusion into
> stable kernels after syzbot tested for a while.

Hmm. The diff is bigger but the change really is equivalent: that
client->lock is only ever used in client.c and trans_fd.c, you replaced
all the occurences in client.c (3 locations + init) while I replaced all
the occurences in trans_fd.c (6 locations + init); the end result is
the same of splitting the two locks exactly at the same place; as far as
correctness goes the patches are identical.

The diff is a bit bigger but the result is more maintainable, and both
versions would require trivial context adjustments to backport anyway
because of bd873038aed5 ("net/9p: allocate appropriate reduced message
buffers") which conflict with either version...
I don't think this warrants the overhead of splitting the patch; sorry.

(and anyway Sasha Levin's autopicker seems to pick almost everything 9p,
said bd873038aed5 was backported down to 5.15 so these will have
backport for free on either version)

Back on topic, assuming you don't strongly oppose to keeping my version,
what tags should I add to the patch?
Reported-by: Tetsuo Handa <penguin-kernel@...ove.SAKURA.ne.jp>
undersells your work, but I don't want to add something like
Co-authored-by without your permission.

Thanks,
--
Dominique