| Message ID | 156418830210.10856.17740359763468342629.stgit@gimli.home |
|---|---|
| Headers | show |
| Series | PCI DMA alias support | expand |
On Fri, Jul 26, 2019 at 06:55:27PM -0600, Alex Williamson wrote: > Please see patch 1/ for the motivation and utility of this series. > This v1 submission improves on the previous RFC with revised commit > logs, comments, and more testing, and the missing IVRS support for DMA > alias ranges is now included. Testing has been done with Linux guests > with both SeaBIOS and OVMF with configurations of intel-iommu and > amd-iommu. Intel-iommu testing includes device assignment, amd-iommu > is necessarily limited to emulated devices with interrupt remapping > disabled and iommu=pt in the guest (enabling interrupt remapping or > disabling guest passthrough mode fails to work regardless of this > series). This series is NOT intended for QEMU v4.1. Thanks, > > Alex series looks good to me. pls ping when 4.1 is out and I'll queue it. > --- > > Alex Williamson (2): > pci: Use PCI aliases when determining device IOMMU address space > hw/i386: AMD-Vi IVRS DMA alias support > > > hw/i386/acpi-build.c | 127 +++++++++++++++++++++++++++++++++++++++++++++++--- > hw/pci/pci.c | 43 ++++++++++++++++- > 2 files changed, 160 insertions(+), 10 deletions(-)
On Mon, 29 Jul 2019 15:15:29 -0400 "Michael S. Tsirkin" <mst@redhat.com> wrote: > On Fri, Jul 26, 2019 at 06:55:27PM -0600, Alex Williamson wrote: > > Please see patch 1/ for the motivation and utility of this series. > > This v1 submission improves on the previous RFC with revised commit > > logs, comments, and more testing, and the missing IVRS support for DMA > > alias ranges is now included. Testing has been done with Linux guests > > with both SeaBIOS and OVMF with configurations of intel-iommu and > > amd-iommu. Intel-iommu testing includes device assignment, amd-iommu > > is necessarily limited to emulated devices with interrupt remapping > > disabled and iommu=pt in the guest (enabling interrupt remapping or > > disabling guest passthrough mode fails to work regardless of this > > series). This series is NOT intended for QEMU v4.1. Thanks, > > > > Alex > > > series looks good to me. > pls ping when 4.1 is out and I'll queue it. Here's the requested ping :) If you'd like a re-posting or comment update, just say so. I think Peter was ultimately satisfied enough to not request a re-spin for comments alone. Thanks, Alex > > --- > > > > Alex Williamson (2): > > pci: Use PCI aliases when determining device IOMMU address space > > hw/i386: AMD-Vi IVRS DMA alias support > > > > > > hw/i386/acpi-build.c | 127 +++++++++++++++++++++++++++++++++++++++++++++++--- > > hw/pci/pci.c | 43 ++++++++++++++++- > > 2 files changed, 160 insertions(+), 10 deletions(-)
On Mon, Aug 19, 2019 at 03:23:16PM -0600, Alex Williamson wrote: > On Mon, 29 Jul 2019 15:15:29 -0400 > "Michael S. Tsirkin" <mst@redhat.com> wrote: > > > On Fri, Jul 26, 2019 at 06:55:27PM -0600, Alex Williamson wrote: > > > Please see patch 1/ for the motivation and utility of this series. > > > This v1 submission improves on the previous RFC with revised commit > > > logs, comments, and more testing, and the missing IVRS support for DMA > > > alias ranges is now included. Testing has been done with Linux guests > > > with both SeaBIOS and OVMF with configurations of intel-iommu and > > > amd-iommu. Intel-iommu testing includes device assignment, amd-iommu > > > is necessarily limited to emulated devices with interrupt remapping > > > disabled and iommu=pt in the guest (enabling interrupt remapping or > > > disabling guest passthrough mode fails to work regardless of this > > > series). This series is NOT intended for QEMU v4.1. Thanks, > > > > > > Alex > > > > > > series looks good to me. > > pls ping when 4.1 is out and I'll queue it. > > Here's the requested ping :) If you'd like a re-posting or comment > update, just say so. I think Peter was ultimately satisfied enough to > not request a re-spin for comments alone. Thanks, Hi, Michael, Do you plan to merge this series? Or do you still wait for an ack? In all cases: Reviewed-by: Peter Xu <peterx@redhat.com> Thanks,
On Mon, 14 Oct 2019 10:59:57 +0800 Peter Xu <peterx@redhat.com> wrote: > On Mon, Aug 19, 2019 at 03:23:16PM -0600, Alex Williamson wrote: > > On Mon, 29 Jul 2019 15:15:29 -0400 > > "Michael S. Tsirkin" <mst@redhat.com> wrote: > > > > > On Fri, Jul 26, 2019 at 06:55:27PM -0600, Alex Williamson wrote: > > > > Please see patch 1/ for the motivation and utility of this series. > > > > This v1 submission improves on the previous RFC with revised commit > > > > logs, comments, and more testing, and the missing IVRS support for DMA > > > > alias ranges is now included. Testing has been done with Linux guests > > > > with both SeaBIOS and OVMF with configurations of intel-iommu and > > > > amd-iommu. Intel-iommu testing includes device assignment, amd-iommu > > > > is necessarily limited to emulated devices with interrupt remapping > > > > disabled and iommu=pt in the guest (enabling interrupt remapping or > > > > disabling guest passthrough mode fails to work regardless of this > > > > series). This series is NOT intended for QEMU v4.1. Thanks, > > > > > > > > Alex > > > > > > > > > series looks good to me. > > > pls ping when 4.1 is out and I'll queue it. > > > > Here's the requested ping :) If you'd like a re-posting or comment > > update, just say so. I think Peter was ultimately satisfied enough to > > not request a re-spin for comments alone. Thanks, > > Hi, Michael, > > Do you plan to merge this series? Or do you still wait for an ack? > > In all cases: > > Reviewed-by: Peter Xu <peterx@redhat.com> Thanks Peter, somehow I had it in my head that MST has already queued these, but it seems not. There's been hardly any churn in these files upstream, so please consider these patches in play. Thanks, Alex
Please see patch 1/ for the motivation and utility of this series. This v1 submission improves on the previous RFC with revised commit logs, comments, and more testing, and the missing IVRS support for DMA alias ranges is now included. Testing has been done with Linux guests with both SeaBIOS and OVMF with configurations of intel-iommu and amd-iommu. Intel-iommu testing includes device assignment, amd-iommu is necessarily limited to emulated devices with interrupt remapping disabled and iommu=pt in the guest (enabling interrupt remapping or disabling guest passthrough mode fails to work regardless of this series). This series is NOT intended for QEMU v4.1. Thanks, Alex --- Alex Williamson (2): pci: Use PCI aliases when determining device IOMMU address space hw/i386: AMD-Vi IVRS DMA alias support hw/i386/acpi-build.c | 127 +++++++++++++++++++++++++++++++++++++++++++++++--- hw/pci/pci.c | 43 ++++++++++++++++- 2 files changed, 160 insertions(+), 10 deletions(-)