From mboxrd@z Thu Jan 1 00:00:00 1970 From: Alex Williamson Subject: Re: [RFC][patch 0/6] pci pass-through support for qemu/KVM on s390 Date: Mon, 08 Sep 2014 08:19:32 -0600 Message-ID: <1410185972.2982.136.camel@ul30vt.home> References: <20140904105223.336503578@de.ibm.com> <54097287.1000006@suse.de> <20140905113954.GB17325@tuxmaker.boeblingen.de.ibm.com> <540A44F2.3040803@suse.de> <540D74F9.50903@redhat.com> Mime-Version: 1.0 Content-Type: text/plain; charset="UTF-8" Content-Transfer-Encoding: 7bit Return-path: In-Reply-To: <540D74F9.50903@redhat.com> Sender: kvm-owner@vger.kernel.org List-Archive: List-Post: To: Paolo Bonzini Cc: Alexander Graf , Frank Blaschka , frank.blaschka@de.ibm.com, qemu-devel@nongnu.org, linux-s390@vger.kernel.org, kvm@vger.kernel.org, aik@ozlabs.ru List-ID: On Mon, 2014-09-08 at 11:20 +0200, Paolo Bonzini wrote: > Il 06/09/2014 01:19, Alexander Graf ha scritto: > >> > 1) interpretive execution of pci load/store instruction. If we use this function > >> > pci access does not get intercepted (no SIE exit) but is handled via microcode. > >> > To enable this we have to disable zpci device and enable it again with information > >> > from the SIE control block. > > Hrm. So how about you create a special vm ioctl for KVM that allows you > > to attach a VFIO device fd into the KVM VM context? Then the default > > would stay "accessible by mmap traps", but we could accelerate it with KVM. > > There is already KVM_DEV_VFIO_GROUP_ADD and KVM_DEV_VFIO_GROUP_DEL. > > Right now, they result in a call to kvm_arch_register_noncoherent_dma or > kvm_arch_unregister_noncoherent_dma, but you can add more hooks. Eric Auger is also working on a patch series to do IRQ forward control on ARM via the kvm-vfio pseudo device, extending the interface to register VFIO device fds. Sounds like that may be a good path to follow here too. Thanks, Alex