Re: [Qemu-devel] [PATCH v2 0/3] Add support for for GICv2m and MSIs to arm-virt

[Christoffer Dall]

On Wed, May 06, 2015 at 05:39:28PM +0100, Peter Maydell wrote:
> On 6 May 2015 at 17:33, Peter Maydell <address@hidden> wrote:
> > On 27 April 2015 at 18:31, Christoffer Dall <address@hidden> wrote:
> >> Now when we have a host generic PCIe controller in the virt board, it
> >> would be nice to be able to use MSIs so that we can eventually enable
> >> VHOST with KVM.
> >>
> >> With these patches you can use MSIs with TCG and with KVM, but you still
> >> need some fixes for the mapping of the IRQ index to the GSI number for
> >> IRQFD to work.  A separate series that enables IRQFD and vhost
> >> is available: "ARM adaptations for vhost irqfd setup"
> >> https://lists.gnu.org/archive/html/qemu-devel/2015-04/msg01054.html)
> >>
> >> Tested with KVM on XGene and with TCG by configuring a virtio-pci
> >> network adapter for the guest and verifying MSIs going through as
> >> expected.
> >
> > You forgot to change the QOM device name to 'arm-gicv2m', but
> > I'll fix that up as I apply this to target-arm.next.
> 
> ...except this series breaks booting of a linux guest using PCI
> on the virt board with aarch32:
> 
> PCI host bridge /address@hidden ranges:
>    IO 0x3eff0000..0x3effffff -> 0x00000000
>   MEM 0x10000000..0x3efeffff -> 0x10000000
> pci-host-generic 3f000000.pcie: PCI host bridge to bus 0000:00
> pci_bus 0000:00: root bus resource [bus 00-0f]
> pci_bus 0000:00: root bus resource [io  0x0000-0xffff]
> pci_bus 0000:00: root bus resource [mem 0x10000000-0x3efeffff]
> PCI: bus0: Fast back to back transfers disabled
> pci 0000:00:01.0: of_irq_parse_pci() failed with rc=-22
> pci 0000:00:02.0: of_irq_parse_pci() failed with rc=-22
> pci 0000:00:02.0: BAR 6: assigned [mem 0x10000000-0x1003ffff pref]
> pci 0000:00:01.0: BAR 1: assigned [mem 0x10040000-0x10040fff]
> pci 0000:00:02.0: BAR 1: assigned [mem 0x10041000-0x10041fff]
> pci 0000:00:01.0: BAR 0: assigned [io  0x1000-0x103f]
> pci 0000:00:02.0: BAR 0: assigned [io  0x1040-0x105f]
> virtio-pci 0000:00:01.0: enabling device (0100 -> 0103)
> virtio-pci 0000:00:02.0: enabling device (0100 -> 0103)
> virtio_blk: probe of virtio32 failed with error -22
> virtio_net: probe of virtio33 failed with error -22
> 
> (and without virtio-blk we don't mount our rootfs).
> 
Shanker figured out that this was due to me changing address-cells and
size-cells in the gic node and breaking the irq-map in the DT and
provided a fix.

I will send a new series.

Thanks,
-Christoffer