Return-Path: Received: from mail-yk0-f178.google.com ([209.85.160.178]:36360 "EHLO mail-yk0-f178.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752308AbbFPMcG (ORCPT ); Tue, 16 Jun 2015 08:32:06 -0400 Received: by ykdr198 with SMTP id r198so11820162ykd.3 for ; Tue, 16 Jun 2015 05:32:05 -0700 (PDT) MIME-Version: 1.0 In-Reply-To: <1857695394.3971135.1434449335470.JavaMail.zimbra@desy.de> References: <1857695394.3971135.1434449335470.JavaMail.zimbra@desy.de> Date: Tue, 16 Jun 2015 08:32:04 -0400 Message-ID: Subject: Re: crash on read overflow From: Trond Myklebust To: "Mkrtchyan, Tigran" Cc: linux-nfs Content-Type: text/plain; charset=UTF-8 Sender: linux-nfs-owner@vger.kernel.org List-ID: Hi Tigran, On Tue, Jun 16, 2015 at 6:08 AM, Mkrtchyan, Tigran wrote: > > > Hi Trond, > > with RHEL6 kernel we observe kernel panic if server reply > more data than client requested. This is obviously fixed in > RHEL7 and upstream. Unfortunately, kernel crashes so badly > that I can't get any stacktrace. My guess is that > commit 64bd577ea0021f5903505de061b3b7d8a785ee94 is the one > which fixes the issue. Can you confirm? > That should help, but you also need to ensure that they have applied commit bfeea1dc1c92 ("SUNRPC: Don't decode beyond the end of the RPC reply message"). You might possibly also need commit 1537693ceaa8 ("SUNRPC: Clean up xdr_set_iov()") and commit 76cacaabf15a ("SUNRPC: xdr_read_pages needs to clear xdr->page_ptr.") since those are all dependencies for bfeea1dc1c92. Cheers Trond