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
| ||
|
Message-ID: <20130107195425.GE13659@quack.suse.cz>
Date: Mon, 7 Jan 2013 20:54:25 +0100
From: Jan Kara <jack@...e.cz>
To: Chen Gang <gang.chen@...anux.com>
Cc: Jan Kara <jack@...e.cz>, Theodore Ts'o <tytso@....edu>,
akpm@...ux-foundation.org, linux-ext4@...r.kernel.org
Subject: Re: [Suggestion] fs/ext3: memory leak by calling set_qf_name or
clear_qf_name, many times.
On Fri 04-01-13 13:39:37, Chen Gang wrote:
> 于 2012年12月31日 20:06, Jan Kara 写道:
> > Thanks for report. Yes, memory leak seems to be possible. Attached patch
> > should fix it, I have added it to my tree.
> >
> > Honza
> >
>
> after read your patch, today (although I am not a reviewer).
> is it a possible situation like this (just my imagination):
> 1st let it call clear_qf_name successfully.
> 2nd let it call set_qf_name successfully.
> 3rd call clear_qf_name again ?
Although this is not likely in practice, if you specify mount options
like usrjquota=aquota.user,usrjquota=,usrjquota=aquota.user it can happen.
Thanks for spotting this.
> ...
> if this situation is possible.
> I guess the memory still leak, after apply your patch.
>
> if what I guess is true
> it seems we have to duplicate memory for old_opts in ext3_remount.
> and for clear_qf_name, set_qf_name, ext3_remount:
> use normal memory allocation and free ways instead of current using.
Yes. Updated patch attached.
Honza
--
Jan Kara <jack@...e.cz>
SUSE Labs, CR
View attachment "0001-ext3-Fix-memory-leak-when-quota-options-are-specifie.patch" of type "text/x-patch" (3347 bytes)
Powered by blists - more mailing lists