From mboxrd@z Thu Jan 1 00:00:00 1970 From: Richard Weinberger Subject: Re: [PATCH][Resend v2] Fix infinite loop in search_binary_handler() Date: Tue, 5 Jul 2011 14:21:02 +0200 Message-ID: <201107051421.02658.richard@nod.at> References: <1309779003-8668-1-git-send-email-richard@nod.at> <201107051155.18465.richard@nod.at> <201107052102.BHE69234.OLQOVMJFOFFHSt@I-love.SAKURA.ne.jp> Mime-Version: 1.0 Content-Type: Text/Plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit Cc: viro@zeniv.linux.org.uk, akpm@linux-foundation.org, linux-kernel@vger.kernel.org, linux-fsdevel@vger.kernel.org To: Tetsuo Handa Return-path: In-Reply-To: <201107052102.BHE69234.OLQOVMJFOFFHSt@I-love.SAKURA.ne.jp> Sender: linux-kernel-owner@vger.kernel.org List-Id: linux-fsdevel.vger.kernel.org Am Dienstag 05 Juli 2011, 14:02:34 schrieb Tetsuo Handa: > Richard Weinberger wrote: > > But IMHO adding a new attribute to task_struct is a bit overkill. > > Why is your variant better than my strcmp() in fs/exec.c? > > Say, there are multiple /sbin/modprobe > > /sbin/modprobe > /var/chroot/sbin/modprobe > > and only /var/chroot/sbin/modprobe needs request_module() whereas > /sbin/modprobe does not need request_module(). Why do we need to make > execl("/sbin/modprobe", "--help", NULL) from chroot("/var/chroot/") fail > by denying request_module() that does not cause recursion? *headdesk*, bprm->filename will contain "/sbin/modprobe" and the strcmp() will falsely match. Thanks for the pointer! //richard