From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from eggs.gnu.org ([2001:4830:134:3::10]:41549) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1VnnI5-0005Sz-Uj for qemu-devel@nongnu.org; Tue, 03 Dec 2013 05:33:23 -0500 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1VnnHz-0007Y7-VJ for qemu-devel@nongnu.org; Tue, 03 Dec 2013 05:33:17 -0500 Received: from mx1.redhat.com ([209.132.183.28]:40929) by eggs.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1VnnHz-0007Xu-Nv for qemu-devel@nongnu.org; Tue, 03 Dec 2013 05:33:11 -0500 Message-ID: <529DB35D.5070408@redhat.com> Date: Tue, 03 Dec 2013 18:33:01 +0800 From: Fam Zheng MIME-Version: 1.0 References: <1385430206-26900-1-git-send-email-famz@redhat.com> <1385430206-26900-4-git-send-email-famz@redhat.com> <529D74AC.5060008@linux.vnet.ibm.com> <529D948B.3030605@redhat.com> <20131203092801.GC9956@dhcp-200-207.str.redhat.com> In-Reply-To: <20131203092801.GC9956@dhcp-200-207.str.redhat.com> Content-Type: text/plain; charset=GB2312 Content-Transfer-Encoding: quoted-printable Subject: Re: [Qemu-devel] [PATCH v4 3/6] qemu-iotests: Add _default_cache_mode and _supported_cache_modes List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: Kevin Wolf Cc: Wenchao Xia , stefanha@redhat.com, qemu-devel@nongnu.org On 2013=C4=EA12=D4=C203=C8=D5 17:28, Kevin Wolf wrote: > Am 03.12.2013 um 09:21 hat Fam Zheng geschrieben: >> On 2013=C4=EA12=D4=C203=C8=D5 14:05, Wenchao Xia wrote: >> >>> >>> Why forbid mode =3D writeback? >>> >> >> These test cases used to run with only possibly "writethrough", or "no= ne". >> And they don't work with writeback, at least in my case. So I didn't a= dd >> other modes here. >=20 > I suspect that the tests allowing writethrough might also work with > directsync. Did you give that one a try? >=20 I guess so. Except for my /tmp is tmpfs and complains about O_DIRECT. It works for me with "TMPDIR=3D/var/tmp". > 026 has different output for wt and wb modes, this is why writeback > fails. It should probably check against 026.out.nocache (which would > have to be renamed as 026.out.wb) instead of 026.out. That's something > for a separate series, though. >=20 Yes, good idea. I'll do this and the directsync including in the next series. Fam