Hi Jean-Philippe,
One comment below.
Jean-Philippe Brucker <jean-philippe.brucker@xxxxxxx> writes:
> In some cases device regions don't support mmap. They can still be made
> available to the guest by trapping all accesses and forwarding reads or
> writes to VFIO. Such regions may be:
>
> * PCI I/O port BARs.
> * Sub-page regions, for example a 4kB region on a host with 64k pages.
> * Similarly, sparse mmap regions. For example when VFIO allows to mmap
> fragments of a PCI BAR and forbids accessing things like MSI-X tables.
> We don't support the sparse capability at the moment, so trap these
> regions instead (if VFIO rejects the mmap).
>
> Signed-off-by: Jean-Philippe Brucker <jean-philippe.brucker@xxxxxxx>
> ---
> v3->v4: new.
> ---
> include/kvm/vfio.h | 3 +
> vfio/core.c | 165 +++++++++++++++++++++++++++++++++++++++++++++++------
> vfio/pci.c | 41 +++++++------
> 3 files changed, 176 insertions(+), 33 deletions(-)
>
[...]
> diff --git a/vfio/pci.c b/vfio/pci.c
> index 698b19582143..9c3ef2f1a373 100644
> --- a/vfio/pci.c
> +++ b/vfio/pci.c
[...]
> @@ -785,6 +788,7 @@ static int vfio_pci_configure_bar(struct kvm *kvm, struct vfio_device *vdev,
> size_t nr)
> {
> int ret;
> + u32 bar;
> size_t map_size;
> struct vfio_pci_device *pdev = &vdev->pci;
> struct vfio_region *region = &vdev->regions[nr];
> @@ -792,6 +796,10 @@ static int vfio_pci_configure_bar(struct kvm *kvm, struct vfio_device *vdev,
> if (nr >= vdev->info.num_regions)
> return 0;
>
> + bar = pdev->hdr.bar[nr];
> +
> + region->vdev = vdev;
> + region->is_ioport = !!(bar & PCI_BASE_ADDRESS_SPACE_IO);
> region->info = (struct vfio_region_info) {
> .argsz = sizeof(*region),
sizeof(region->info)?
Otherwise the patch looks good.
Thanks,
Punit
> .index = nr,
> @@ -819,14 +827,13 @@ static int vfio_pci_configure_bar(struct kvm *kvm, struct vfio_device *vdev,
> }
> }
>
> - /* Grab some MMIO space in the guest */
> - map_size = ALIGN(region->info.size, PAGE_SIZE);
> - region->guest_phys_addr = pci_get_io_space_block(map_size);
> + if (!region->is_ioport) {
> + /* Grab some MMIO space in the guest */
> + map_size = ALIGN(region->info.size, PAGE_SIZE);
> + region->guest_phys_addr = pci_get_io_space_block(map_size);
> + }
>
> - /*
> - * Map the BARs into the guest. We'll later need to update
> - * configuration space to reflect our allocation.
> - */
> + /* Map the BARs into the guest or setup a trap region. */
> ret = vfio_map_region(kvm, vdev, region);
> if (ret)
> return ret;
