From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from eggs.gnu.org ([2001:4830:134:3::10]:41910) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1gYos2-000119-IG for qemu-devel@nongnu.org; Mon, 17 Dec 2018 04:07:27 -0500 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1gYory-0001ZH-Kg for qemu-devel@nongnu.org; Mon, 17 Dec 2018 04:07:26 -0500 Date: Mon, 17 Dec 2018 10:07:10 +0100 From: Cornelia Huck Message-ID: <20181217100710.6cb82fd9.cohuck@redhat.com> In-Reply-To: <1544806422-21418-1-git-send-email-pmorel@linux.ibm.com> References: <1544806422-21418-1-git-send-email-pmorel@linux.ibm.com> MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Subject: Re: [Qemu-devel] [PATCH v3] s390x/pci: add common fmb List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: Pierre Morel Cc: walling@linux.ibm.com, borntraeger@de.ibm.com, rth@twiddle.net, david@redhat.com, qemu-s390x@nongnu.org, qemu-devel@nongnu.org, pasic@linux.ibm.com, thuth@redhat.com On Fri, 14 Dec 2018 17:53:41 +0100 Pierre Morel wrote: > After the last review round I worked on endianness. > Doing this I found some errors in the code and in the interpretation > of the documentation. > > The new patch changed the following from previous version: > > In s390-pci-bus: > - Initialize the FMB Format. > > In s390-pci-bus.h > - re-organization of the internal counters, having a table for the > internal counters. > > In s390-pci-inst.c > - Internal counters update (LD/ST/STB/RPCIT) is done always. > even if the FMB if fmb_addr is NULL. > AFAIU this respect the documentation which only states that FMB > update is stopped. > - in mpcifc_service_call(), moved the setting of fmb_addr after > the timer has been stopped. > - fmb_update((), use address_space_stq_be() to handle endianness > when storing the FMB. > - define the format with 32 bits instead of one char and reserved > chars, this is easier to handle the FMB copy. > - No update of the DMA fields inside the FMB, as stipulated by the > documentation when format32 is 0. > > Patch is tested (for the good case) on Z(KVM) and X(TCG). Yes, the values look sane for virtio-pci devices under tcg as well. Will review and hopefully apply later. > > Yi Min Zhao (1): > s390x/pci: add common function measurement block > > hw/s390x/s390-pci-bus.c | 4 +- > hw/s390x/s390-pci-bus.h | 29 ++++++++++ > hw/s390x/s390-pci-inst.c | 141 ++++++++++++++++++++++++++++++++++++++++++++++- > hw/s390x/s390-pci-inst.h | 1 + > 4 files changed, 171 insertions(+), 4 deletions(-) >