From: Cornelia Huck Date: Tue, 22 Jan 2019 16:46:45 +0000 Subject: Re: [RFC v1 2/2] vfio-ccw: Don't exit early if state of the vfio-ccw subchannel is not idle Message-Id: <20190122174645.141d8c3b.cohuck@redhat.com> In-Reply-To: <20190122174307.0977f475.cohuck@redhat.com> References: <20190122174307.0977f475.cohuck@redhat.com> To: linux-s390@vger.kernel.org, kvm@vger.kernel.org List-ID: On Tue, 22 Jan 2019 10:19:29 -0500 Farhan Ali wrote: > The test setup basically runs fio (sequential, random) read and write > workloads for a long period of time. I run each workload at least for 5 > minutes and keep running them in a continuous loop for few hours. Ok, this is more intensive testing than what I'm doing :) (basically some dd and some tunedasd calls) > > Also I am testing with Hyper PAV, so I run the workloads with different > number of HPAVs for a base device. Ok, I don't have PAV in my setup.