From mboxrd@z Thu Jan 1 00:00:00 1970 From: Chuck Lever Subject: Re: RFC: Is SM_SIMU_CRASH (in rpc.statd) needed? Date: Tue, 22 Jun 2010 12:58:24 -0400 Message-ID: <4C20EBB0.9010800@oracle.com> References: <4C20CA12.2070002@oracle.com> <4C20E5D7.80908@RedHat.com> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1; format=flowed Cc: NFSv3 list To: Steve Dickson Return-path: Received: from rcsinet10.oracle.com ([148.87.113.121]:45889 "EHLO rcsinet10.oracle.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751321Ab0FVRA0 (ORCPT ); Tue, 22 Jun 2010 13:00:26 -0400 In-Reply-To: <4C20E5D7.80908-AfCzQyP5zfLQT0dZR+AlfA@public.gmane.org> Sender: linux-nfs-owner@vger.kernel.org List-ID: On 06/22/10 12:33 PM, Steve Dickson wrote: > > > On 06/22/2010 10:34 AM, Chuck Lever wrote: >>> Does anybody know who uses SM_SIMU_CRASH and is it >>> still needed? I'm getting some push back from our >>> security folks to plug this DoS hole up... >> >> How is it a DoS? statd ignores any SM_SIMU_CRASH from a non-loopback >> address that does not use a privileged port. To invoke it, you >> basically have to be a privileged local user. > It happens during an ISS scan.... What were the exact results? >> Right now, no one uses it, but I think we should reserve the right to >> use it in the future. This could be part of a more robust interface >> between lockd and statd. > hmm... how can having a way of bringing down a daemon, make it more robust? SM_SIMU_CRASH is not supposed to bring down statd, it's supposed to trigger sm-notify. If something else is happening, that's a bug.