From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Subject: Re: [Qemu-devel] [PATCH v4 1/6] vfio-ccw: make it safe to access channel programs From: Eric Farman References: <20190301093902.27799-1-cohuck@redhat.com> <20190301093902.27799-2-cohuck@redhat.com> <9e81af36-ebd2-671b-5256-90e8efaad6f2@linux.ibm.com> <20190408190747.12e3618b.cohuck@redhat.com> <20190410013434.7cea1971@oc2783563651> <4a692c11-edb2-93f9-7d8e-96a73a7ecdae@linux.ibm.com> Date: Thu, 11 Apr 2019 17:27:42 -0400 MIME-Version: 1.0 In-Reply-To: <4a692c11-edb2-93f9-7d8e-96a73a7ecdae@linux.ibm.com> Content-Type: text/plain; charset="utf-8"; format="flowed" Content-Language: en-US Content-Transfer-Encoding: 8bit Message-Id: <571b9f8c-cfb0-704e-ded4-8661a8136876@linux.ibm.com> Sender: kvm-owner@vger.kernel.org List-Archive: List-Post: To: Halil Pasic , Cornelia Huck Cc: Farhan Ali , linux-s390@vger.kernel.org, Alex Williamson , Pierre Morel , kvm@vger.kernel.org, qemu-devel@nongnu.org, qemu-s390x@nongnu.org List-ID: On 4/10/19 10:59 PM, Eric Farman wrote: > > > On 4/9/19 7:34 PM, Halil Pasic wrote: >> On Mon, 8 Apr 2019 19:07:47 +0200 ...snip... >> I'm just running fio on a pass-through DASD and on some virto-blk disks >> in parallel. My QEMU is today's vfio-ccw-caps from your repo. >> >> I see stuff like this: >> qemu-git: vfio-ccw: wirte I/O region failed with errno=16[1811/7332/0 >> iops] [eta 26m:34s] > > Without knowing what the I/O was that failed, this is a guessing game. > But I encountered something similar just now running fio. > > qemu: > 2019-04-11T02:06:09.524838Z qemu-system-s390x: vfio-ccw: wirte I/O > region failed with errno=16 ...snip... > From the associated I/O, I think this is fixed by a series I am nearly > ready to send for review.  I'll try again with those fixes on top of the > two series here, and report back. So, I've run enough combinations to feel comfortable saying that the error (EBUSY) I saw last night (and presumably the one Halil saw) exists in today's code and is not introduced by this series. It also appears to be addressed by one of the patches in a series I'm working on, but which that series still has some further problems. Sigh, there are too many branches and too many interrupts. - Eric