From mboxrd@z Thu Jan 1 00:00:00 1970 From: Al Viro Subject: Re: [PATCH 26/26] audit: collect path information when possible Date: Thu, 17 Nov 2011 22:47:03 +0000 Message-ID: <20111117224703.GZ2203@ZenIV.linux.org.uk> References: <20111117220244.23481.96785.stgit@paris.rdu.redhat.com> <20111117220510.23481.42750.stgit@paris.rdu.redhat.com> Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit Return-path: Content-Disposition: inline In-Reply-To: <20111117220510.23481.42750.stgit@paris.rdu.redhat.com> List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Sender: linux-audit-bounces@redhat.com Errors-To: linux-audit-bounces@redhat.com To: Eric Paris Cc: linux-audit@redhat.com List-Id: linux-audit@redhat.com On Thu, Nov 17, 2011 at 05:05:11PM -0500, Eric Paris wrote: > For _at type syscalls (like openat) we do not collect any information about > the dfd. This patch grabs a reference to the path of all fd's passed to > the kernel. We free those on syscall exit. We will then output those paths > as inode records and use the path information to generate better pathnames if > possible. I think this is bogus. If nothing else, if you want dfd, then by damn collect that information *when* *you* *are* *starting* *a* *lookup*. Not on every bleeding fget(), no matter why and by whom had it been called. FWIW, hooking into getname() also had always looked wrong, for much the same reasons...