This series allows to optimize the deactivation of virtual interrupts associated to a vfio platform device IRQ. Let's call this optimization: ARM IRQ forwarding. Without that optimization the deactivation of the physical IRQ is done by the host and the deactivation of the virtual IRQ, by the guest, is trapped by KVM with a maintenance IRQ. With ARM IRQ forwarding, the physical IRQ is automatically deactivated by the GIC interrupt controller when the guest deactivates the virtual IRQ. The series enables the auto-negotiation of this optimization, on ARM. The auto-negotiation is coordinated by the IRQ bypass manager. The VFIO platform driver and KVM/irqfd cooperate to turn that feature on. The series implements the setup of IRQ forwarding and adapt the existing VGIC state machine so that it works. When the IRQ is forwarded, the VFIO platform driver does not need to mask the physical IRQ anymore before signaling the eventfd. Indeed genirq lowers the running priority, enabling other physical IRQ to hit except that one. The injection still is based on irqfd triggering. The only impact on irqfd process is the resamplefd is not triggered anymore for level sensitive IRQs since deactivation is not trapped by KVM. More backgroung on ARM IRQ forwarding in the text below and at http://www.linux-kvm.org/images/a/a8/01x04-ARMdevice.pdf. This was tested on: - Calxeda Midway, assigning the xgmac main IRQ. - AMD Seattle, where all XGBE interrupts are assigned: - 2 level sensitive and 4 DMA edge sensitive IRQs applies on top of: KVM: arm/arm64: leave the LR active state on GICD_ICENABLERn access All those pieces can be found at: https://git.linaro.org/people/eric.auger/linux.git/shortlog/refs/heads/v4.4-rc1-forward-v4 History: v3 -> v4: - dependencies now are fixed (IRQ bypass manager, guest synchronous halt/resume, VGIC architected timer) - only applies on top of: KVM: arm/arm64: leave the LR active state on GICD_ICENABLERn access - Rebase on vgic series related to forwarded shared interrupts, mainly [PATCH v2 0/8] Rework architected timer and forwarded IRQs handling & [PATCH 0/2] arm/arm64: KVM: Fix architected timer issues - took into account Alex's comment on VFIO part: - vfio_platform_set_forwarded replacing vfio_platform_set_automasked - automasked flag not updated - vfio_platform_is_active renamed into vfio_platform_irq_is_active - handle returned value for vfio_platform_irq_is_active - remove WARN_ON on irq_bypass_register_producer failure - took into account Christoffer's comments on vgic part - clarify comments - clarify edge-sensitive handling and validate_injection of edge mapped unshared IRQ with level == 0 - use actual vcpu_id instead of VCPU 0 - revisit unset_forward implement using unqueue - moved KVM: arm/arm64: select IRQ_BYPASS_MANAGER in this series - add support for clear active and clear pending v2 (RFC) -> v3(PATCH): - all dependencies now have a patch status - we dropped the producer active boolean exchanged between the VFIO producer and irqfd arm consumer. As a consequence, on unforward, if the IRQ is active, this latter is deactivated without VFIO-masking it. So we do not exactly come back to the exact state where we would be in unforwarded state. A new physical IRQ can hit while the previous virtual IRQ is under treatment. Its injection in the guest may be rejected thanks to the VGIC state machine. This IRQ will be lost but I don't think this is a severe issue. In case no new IRQ hits, the guest deactivation of the virtual IRQ will trigger the resamplefd which will VFIO unmask the non-masked IRQ. This also has no consequence. - VFIO platform driver consumer_add now can fail. It rejects the transition for forwarding state in case the IRQ is active - the series is rebased on new irq_vcpu_affinity series - no dependency anymore on "chip/vgic adaptations for forwarded irq" which was partially integrated into Marc's series. A fix is still needed through. - Guest synchronous halt/resume patch re-integrated into this series - integrate a new patch file coming mixing [PATCH v4 11/11] KVM: arm/arm64: vgic: Allow HW interrupts for non-shared devices & [RFC v2 2/4] KVM: arm: vgic: fix state machine for forwarded IRQ v1 -> v2: - irq bypass manager and irqfd consumer moved in a separate patch - kvm_arm_[halt,resume]_guest moved in a separate patch - remove VFIO external functions since we do not need them anymore - apply container_of strategy advised by Paolo. Only active field remains and discussions will tell whether we get rid of it. - renamed kvm_arch functions - kvm-vfio v6 -> RFC v1 based on IRQ bypass manager see previous history in https://lkml.org/lkml/2015/4/13/353). Best Regards Eric Eric Auger (12): KVM: arm/arm64: select IRQ_BYPASS_MANAGER VFIO: platform: registration of a dummy IRQ bypass producer VFIO: platform: test forwarded state when selecting the IRQ handler VFIO: platform: single handler using function pointer VFIO: platform: add vfio_platform_set_forwarded VFIO: platform: add vfio_platform_irq_is_active VFIO: platform: add irq bypass producer management KVM: arm/arm64: vgic: support irqfd injection of a mapped IRQ KVM: arm/arm64: vgic: forwarding control KVM: arm/arm64: vgic: implement clear active for non shared mapped IRQ KVM: arm/arm64: vgic: implement clear pending for non shared mapped IRQ KVM: arm/arm64: implement IRQ bypass consumer functions Marc Zyngier (1): KVM: arm/arm64: vgic: adapt state machine for non shared mapped interrupts arch/arm/kvm/Kconfig | 3 + arch/arm/kvm/arm.c | 37 +++++ arch/arm64/kvm/Kconfig | 3 + drivers/vfio/platform/vfio_platform_irq.c | 120 +++++++++++++- drivers/vfio/platform/vfio_platform_private.h | 4 + include/kvm/arm_vgic.h | 8 +- virt/kvm/arm/arch_timer.c | 2 +- virt/kvm/arm/vgic.c | 221 ++++++++++++++++++++++++-- 8 files changed, 376 insertions(+), 22 deletions(-) -- 1.9.1 -- 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
