Return-Path: Content-Type: text/plain; charset=us-ascii Mime-Version: 1.0 (Mac OS X Mail 9.3 \(3124\)) Subject: Re: GSS sequence number window From: Chuck Lever In-Reply-To: <20170606201504.GH13376@fieldses.org> Date: Tue, 6 Jun 2017 16:16:53 -0400 Cc: Benjamin Coddington , Linux NFS Mailing List Message-Id: References: <63736845-2BD3-4EE1-AC12-0BD21A9ABEF2@oracle.com> <20170530193419.GA9371@fieldses.org> <20170531192231.GA23526@fieldses.org> <28665890-C74A-4319-B42E-475393821EC7@oracle.com> <20170606194158.GG13376@fieldses.org> <4D542D55-DCBA-4838-9DB2-B76B4068783E@oracle.com> <20170606201504.GH13376@fieldses.org> To: "J. Bruce Fields" List-ID: > On Jun 6, 2017, at 4:15 PM, J. Bruce Fields wrote: > > On Tue, Jun 06, 2017 at 03:45:59PM -0400, Chuck Lever wrote: >> >>> On Jun 6, 2017, at 3:41 PM, J. Bruce Fields wrote: >>> >>> On Tue, Jun 06, 2017 at 03:35:23PM -0400, Chuck Lever wrote: >>>> I filed https://bugzilla.linux-nfs.org/show_bug.cgi?id=306 >>>> >>>> To check memory allocation latency, I could always construct >>>> a framework around kmalloc and alloc_page. >>>> >>>> >>>> I've also found some bad behavior around proto=rdma,sec=krb5i. >>>> When I run a heavy I/O workload (fio, for example), every so >>>> often a read operation fails with EIO. I dug into it a little >>>> and MIC verification fails for these replies on the client. >>> >>> Do we still have the problem that the read data can change between the >>> time we calculate the MIC and the time we transmit the data to the >>> client? >> >> I don't see a problem with krb5p, which, if IIUC, would also >> fall victim to this situation, unless there is much stricter >> request serialization going on with krb5p. > > We turn off zero-copy by clearing RQ_SPLICE_OK in the krb5p case. Seems like this is the right answer for krb5i too. Shall I try that? >> Even so, how would I detect if this issue was present? > > Good question. If you knew the data and mic in the bad case, and had > some way to guess what the previous data might have been based on what > you knew about the test, then you could try mic's of likely older > versions of the data and see if you get a match.... That sounds hard. > > --b. -- Chuck Lever