Here's a handful of random fixes I've queued locally that didn't have
a chance to make it in 4.11.
The first two patches avoid stack-protector messing with the HYP code,
as this ends up being a complete disaster.
The following two patches fix a bug introduced in the new vgic, where
we may queue HW interrupts with the Pending+Active state, which is
illegal.
The final patch fixes a misinterpretation of the spec, where we
compute the number of APxRn register based on the number of priorities
instead of using the number of preemption levels.
I've tagged the first 4 patches for stable, given that we're doing
something potentially harmful. The last patch is more of a theoretical
issue at this stage, so probably need for a backport.
Marc Zyngier (5):
arm64: KVM: Do not use stack-protector to compile EL2 code
arm: KVM: Do not use stack-protector to compile HYP code
KVM: arm/arm64: vgic-v2: Do not use Active+Pending state for a HW
interrupt
KVM: arm/arm64: vgic-v3: Do not use Active+Pending state for a HW
interrupt
KVM: arm/arm64: vgic-v3: Use PREbits to infer the number of
ICH_APxRn_EL2 registers
arch/arm/kvm/hyp/Makefile | 2 ++
arch/arm64/kvm/hyp/Makefile | 2 ++
virt/kvm/arm/hyp/vgic-v3-sr.c | 18 +++++++++---------
virt/kvm/arm/vgic/vgic-v2.c | 7 +++++++
virt/kvm/arm/vgic/vgic-v3.c | 7 +++++++
5 files changed, 27 insertions(+), 9 deletions(-)
--
2.11.0
