So far, the VGIC data structures have been statically sized, meaning that we always have to support more interrupts than we actually want, and more CPU interfaces than we should. This is a waste of resource, and is the kind of things that should be tuneable. This series addresses that issue by changing the data structures to be dynamically allocated, and adds a new configuration attribute to allocate the number of interrupts. When the attribute is not used, we fallback to the old behaviour of allocating a fixed number of interrupts. The last patch of the series is a bit out of context, but tends to fit well here code-wise. It solves an interesting issue having to do with the placement of the GICV interface in Stage-2 when using 64k pages (if the HW is not 64k aligned, we need to tell userspace about the "sub-page offset" so it can correctly place the guest's GICC region). This series is also the base for Andre Przywara's GICv3 distributor emulation code (which can support far more than 8 vcpus and 1020 interrupts). This has been tested on both ARM (TC2) and arm64 (model). Marc Zyngier (9): KVM: ARM: vgic: plug irq injection race arm/arm64: KVM: vgic: switch to dynamic allocation arm/arm64: KVM: vgic: Parametrize VGIC_NR_SHARED_IRQS arm/arm64: KVM: vgic: kill VGIC_MAX_CPUS arm/arm64: KVM: vgic: handle out-of-range MMIO accesses arm/arm64: KVM: vgic: kill VGIC_NR_IRQS arm/arm64: KVM: vgic: delay vgic allocation until init time arm/arm64: KVM: vgic: make number of irqs a configurable attribute arm64: KVM: vgic: deal with GIC sub-page alignment arch/arm/include/uapi/asm/kvm.h | 2 + arch/arm/kvm/arm.c | 10 +- arch/arm64/include/uapi/asm/kvm.h | 2 + include/kvm/arm_vgic.h | 54 +++--- virt/kvm/arm/vgic.c | 357 ++++++++++++++++++++++++++++++++------ 5 files changed, 339 insertions(+), 86 deletions(-) -- 1.8.3.4 -- To unsubscribe from this list: send the line "unsubscribe kvm" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
