On 2012-08-16 16:34, Avi Kivity wrote: > On 08/16/2012 04:54 PM, Jan Kiszka wrote: >> With this series, we are getting very close to obsoleting qemu-kvm. It >> refactors hw/device-assignment.c and the associated KVM helper functions >> into a form that should allow merging them into QEMU. Once the series is >> acceptable for qemu-kvm, I will break out the necessary uq/master >> patches and push pci-assign to upstream. >> >> The major step of this series is to define a regular set of kvm_device_* >> services that encapsulate classic (i.e. KVM-based, non-VFIO) device >> assignment features and export them to i386 targets only. There will >> never be another arch using them, therefore I pushed them into this >> corner. Moreover, the device assignment device now makes use of the new >> KVM IRQ/MSI routing API and no longer pokes into the internals of that >> layer. Finally, I moved the code into hw/kvm/pci-assign.c, dropped the >> superfluous configure option and did some basic code cleanups (mostly >> coding style) to bring things in shape. >> >> Note that patch 1 is a simple bug fix that should likely be applied for >> qemu-kvm-1.2 independently. >> >> This series depends on [1] and [2] and QEMU upstream (2b97f88c92) being >> merged into qemu-kvm. >> >> Please review. > >>From a quick review it looks ready to merge. Of course I'd appreciate a > review from Alex or Michael as well. Great, thanks. FWIW, as the upstream integration is technically trivial now, I pushed a preview to git://git.kiszka.org/qemu-kvm.git queues/kvm-upstream Will send out once reviews for this series are done. Jan -- Siemens AG, Corporate Technology, CT RTC ITP SDP-DE Corporate Competence Center Embedded Linux -- To unsubscribe from this list: send the line "unsubscribe kvm" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
