This series addresses a peculiarity of the current VGIC
implementation, namely that we don't support interrupt grouping.
KVM either implements a GICv2 without support for the security
extensions, or a GICv3 with DS=1. For GICv2, on systems without the
security extensions, group 0 interrupts can be configured to be either
signalled as FIQs or as IRQs by the VM, whereas group 1 interrupts are
always IRQs. For GICv3, with DS=1, group 1 interrupts are always IRQs
and group 0 interrupts are always FIQs, and there is no concept of
secure vs. non-secure group 1 interrupts when DS=1.
We were treating all interrupts on GICv2 as group 0, but yet telling the
guest that they were group 1. The first patch changes this behavior,
which seems to have no effect on no known guests, but still.
The remaining patches introduce proper interrupt grouping support, along
with MMIO accessors for the VM and userspace to retrieve and set the
which group SGIs, PPIs, and SPIs belong to. LPIs are always group 1
interrupts as per the architecture, and there is no way to modify this
configuration (no IGROUPR registers for LPIs or equivalent ITS
commands).
Tested on Seattle, TX1, and the foundation model. I've run a
GICv2 guest on a GICv3 host on the foundation model. I've tested
migration on Seattle, and migration from an old kernel to a newer kernel
with an unmodified QEMU binary continues to work. I also tested a
modified userspace that sets some interrupts to group 1, and that is
reflected in the vgic debug output.
Applies to v4.18-rc5.
Changes since v3:
- Rebased to v4.18-rc5
- Reworked documentation to clarify order of register restore and
include GICv3
- Re-ordered patches to add infrastructure for returning uaccess write
errors first, and factored out GIC_IIDR write checking to a separate
patch.
- Added a return 0; in the GICv2 GICD_IIDR uaccess write handler and
fixed a typo in the comment.
Changes since v2:
- Patch 6 doesn't allow configuraiton of groups from userspace for
GICv2.
- Patches 7 through 9 are new in v3 of this series, and prevents
userspace from restoring an incompatible IIDR without seeing an error
in the future, and supports legacy userspace migration from old to
new kernels.
Changes since v1:
- Bumped implementation revision field in GICD_IIDR along with changes
- Changed logic in init code to correctly detect vgic emulation type
- Rebased to v4.18-rc3
Christoffer Dall (10):
KVM: arm/arm64: vgic: Define GICD_IIDR fields for GICv2 and GIv3
KVM: arm/arm64: vgic: Keep track of implementation revision
KVM: arm/arm64: vgic: GICv2 IGROUPR should read as zero
KVM: arm/arm64: vgic: Add group field to struct irq
KVM: arm/arm64: vgic: Signal IRQs using their configured group
KVM: arm/arm64: vgic: Permit uaccess writes to return errors
KVM: arm/arm64: vgic: Return error on incompatible uaccess GICD_IIDR
writes
KVM: arm/arm64: vgic: Allow configuration of interrupt groups
KVM: arm/arm64: vgic: Let userspace opt-in to writable v2 IGROUPR
KVM: arm/arm64: vgic: Update documentation of the GIC devices wrt IIDR
Documentation/virtual/kvm/devices/arm-vgic-v3.txt | 8 +++
Documentation/virtual/kvm/devices/arm-vgic.txt | 15 +++---
include/kvm/arm_vgic.h | 7 +++
include/linux/irqchip/arm-gic-v3.h | 10 ++++
include/linux/irqchip/arm-gic.h | 11 ++++
virt/kvm/arm/vgic/vgic-debug.c | 8 +--
virt/kvm/arm/vgic/vgic-init.c | 20 ++++++-
virt/kvm/arm/vgic/vgic-its.c | 1 +
virt/kvm/arm/vgic/vgic-mmio-v2.c | 63 ++++++++++++++++++++---
virt/kvm/arm/vgic/vgic-mmio-v3.c | 53 ++++++++++++++-----
virt/kvm/arm/vgic/vgic-mmio.c | 56 ++++++++++++++++++--
virt/kvm/arm/vgic/vgic-mmio.h | 25 ++++++---
virt/kvm/arm/vgic/vgic-v2.c | 3 ++
virt/kvm/arm/vgic/vgic-v3.c | 6 +--
14 files changed, 237 insertions(+), 49 deletions(-)
--
2.7.4
