From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from cn.fujitsu.com ([222.73.24.84]:62110 "EHLO song.cn.fujitsu.com" rhost-flags-OK-FAIL-OK-OK) by vger.kernel.org with ESMTP id S1754748Ab1CCJKw (ORCPT ); Thu, 3 Mar 2011 04:10:52 -0500 Message-ID: <4D6F5B4E.7000903@cn.fujitsu.com> Date: Thu, 03 Mar 2011 17:11:42 +0800 From: Mi Jinlong To: "J. Bruce Fields" CC: NFS Subject: Re: [PATCH 10/17] CLNT: test resues slot ID References: <4D6DE26C.3010103@cn.fujitsu.com> <4D6DE4A5.1050702@cn.fujitsu.com> <20110303061442.GE7724@fieldses.org> In-Reply-To: <20110303061442.GE7724@fieldses.org> Content-Type: text/plain; charset=ISO-8859-1 Sender: linux-nfs-owner@vger.kernel.org List-ID: MIME-Version: 1.0 J. Bruce Fields: > On Wed, Mar 02, 2011 at 02:33:09PM +0800, Mi Jinlong wrote: >> Signed-off-by: Mi Jinlong >> --- >> nfs4.1/server41tests/st_sequence.py | 18 ++++++++++++++++++ >> 1 files changed, 18 insertions(+), 0 deletions(-) >> >> diff --git a/nfs4.1/server41tests/st_sequence.py b/nfs4.1/server41tests/st_sequence.py >> index 4ea5cab..391fc45 100644 >> --- a/nfs4.1/server41tests/st_sequence.py >> +++ b/nfs4.1/server41tests/st_sequence.py >> @@ -363,3 +363,21 @@ def testBadSequenceidAtSlot(t, env): >> >> res = c.c.compound([op.sequence(sid, nfs4lib.dec_u32(seqid), 2, 3, True)]) >> check(res, NFS4ERR_SEQ_MISORDERED) >> + >> +def testReuseSlotID(t, env): >> + """ >> + >> + FLAGS: sequence all >> + CODE: SEQ14 >> + """ > > Missing description? Sorry! I will add it at next version. > >> + c = env.c1.new_client(env.testname(t)) >> + # CREATE_SESSION >> + sess1 = c.create_session() >> + >> + sid = sess1.sessionid >> + res = c.c.compound([op.sequence(sid, 1, 2, 3, True), op.putrootfh()]) >> + check(res) >> + >> + res = c.c.compound([op.sequence(sid, 1, 2, 3, True)]) >> + # should we also check the ERROR of NFS4ERR_SEQ_FALSE_RETRY ?? > > As I understand it, the server isn't required to recognize false > retries, so I don't think we can check for that. > > The only useful thing to check for, I think, is just to send the same > request twice and make sure the requests match. > > Actually, best would be to do this with a non-idempotent operation. For > example, send SEQ+RENAME twice and make sure that both renames succeed. > (Whereas if the server reply cache is broken and the server attempts to > rename the file again, the second rename will fail with NOENT.) Yes, thanks for you advice. I will send a new patch tomorrow. -- ---- thanks Mi Jinlong