| Message ID | 20200525144125.143875-1-vkuznets@redhat.com (mailing list archive) |
|---|---|
| Headers | show |
| Series | KVM: x86: Interrupt-based mechanism for async_pf 'page present' notifications | expand |
On 25/05/20 16:41, Vitaly Kuznetsov wrote: > Concerns were expressed around (ab)using #PF for KVM's async_pf mechanism, > it seems that re-using #PF exception for a PV mechanism wasn't a great > idea after all. The Grand Plan is to switch to using e.g. #VE for 'page > not present' events and normal APIC interrupts for 'page ready' events. > This series does the later. > > Changes since v1: > - struct kvm_vcpu_pv_apf_data's fields renamed to 'flags' and 'token', > comments added [Vivek Goyal] > - 'type1/2' names for APF events dropped from everywhere [Vivek Goyal] > - kvm_arch_can_inject_async_page_present() renamed to > kvm_arch_can_dequeue_async_page_present [Vivek Goyal] > - 'KVM: x86: deprecate KVM_ASYNC_PF_SEND_ALWAYS' patch added. > > v1: https://lore.kernel.org/kvm/20200511164752.2158645-1-vkuznets@redhat.com/ > QEMU patches for testing: https://github.com/vittyvk/qemu.git (async_pf2_v2 branch) I'll do another round of review and queue patches 1-7; 8-9 will be queued later and separately due to the conflicts with the interrupt entry rework, but it's my job and you don't need to do anything else. Thanks, Paolo
On Thu, May 28, 2020 at 01:04:55PM +0200, Paolo Bonzini wrote: > On 25/05/20 16:41, Vitaly Kuznetsov wrote: > > Concerns were expressed around (ab)using #PF for KVM's async_pf mechanism, > > it seems that re-using #PF exception for a PV mechanism wasn't a great > > idea after all. The Grand Plan is to switch to using e.g. #VE for 'page > > not present' events and normal APIC interrupts for 'page ready' events. > > This series does the later. > > > > Changes since v1: > > - struct kvm_vcpu_pv_apf_data's fields renamed to 'flags' and 'token', > > comments added [Vivek Goyal] > > - 'type1/2' names for APF events dropped from everywhere [Vivek Goyal] > > - kvm_arch_can_inject_async_page_present() renamed to > > kvm_arch_can_dequeue_async_page_present [Vivek Goyal] > > - 'KVM: x86: deprecate KVM_ASYNC_PF_SEND_ALWAYS' patch added. > > > > v1: https://lore.kernel.org/kvm/20200511164752.2158645-1-vkuznets@redhat.com/ > > QEMU patches for testing: https://github.com/vittyvk/qemu.git (async_pf2_v2 branch) > > I'll do another round of review and queue patches 1-7; 8-9 will be > queued later and separately due to the conflicts with the interrupt > entry rework, but it's my job and you don't need to do anything else. Hi Paolo, I seee 1-7 got merged for 5.8. When you say patch 8-9 will be queue later, you mean later in 5.8 or it will held till 5.9 merge window opens. Thanks Vivek
On 04/06/20 19:45, Vivek Goyal wrote: >> I'll do another round of review and queue patches 1-7; 8-9 will be >> queued later and separately due to the conflicts with the interrupt >> entry rework, but it's my job and you don't need to do anything else. > Hi Paolo, > > I seee 1-7 got merged for 5.8. When you say patch 8-9 will be queue later, > you mean later in 5.8 or it will held till 5.9 merge window opens. I hope to get them in 5.8. They have some pretty nasty conflicts that are too much for Linus to resolve. So my plan is to put 8-9 in a topic branch and do the merge myself. Whether this works out depends on the timing of the tip pull request. Paolo
Concerns were expressed around (ab)using #PF for KVM's async_pf mechanism, it seems that re-using #PF exception for a PV mechanism wasn't a great idea after all. The Grand Plan is to switch to using e.g. #VE for 'page not present' events and normal APIC interrupts for 'page ready' events. This series does the later. Changes since v1: - struct kvm_vcpu_pv_apf_data's fields renamed to 'flags' and 'token', comments added [Vivek Goyal] - 'type1/2' names for APF events dropped from everywhere [Vivek Goyal] - kvm_arch_can_inject_async_page_present() renamed to kvm_arch_can_dequeue_async_page_present [Vivek Goyal] - 'KVM: x86: deprecate KVM_ASYNC_PF_SEND_ALWAYS' patch added. v1: https://lore.kernel.org/kvm/20200511164752.2158645-1-vkuznets@redhat.com/ QEMU patches for testing: https://github.com/vittyvk/qemu.git (async_pf2_v2 branch) Vitaly Kuznetsov (10): Revert "KVM: async_pf: Fix #DF due to inject "Page not Present" and "Page Ready" exceptions simultaneously" KVM: x86: extend struct kvm_vcpu_pv_apf_data with token info KVM: rename kvm_arch_can_inject_async_page_present() to kvm_arch_can_dequeue_async_page_present() KVM: introduce kvm_read_guest_offset_cached() KVM: x86: interrupt based APF 'page ready' event delivery KVM: x86: acknowledgment mechanism for async pf page ready notifications KVM: x86: announce KVM_FEATURE_ASYNC_PF_INT KVM: x86: Switch KVM guest to using interrupts for page ready APF delivery KVM: x86: drop KVM_PV_REASON_PAGE_READY case from kvm_handle_page_fault() KVM: x86: deprecate KVM_ASYNC_PF_SEND_ALWAYS Documentation/virt/kvm/cpuid.rst | 6 ++ Documentation/virt/kvm/msr.rst | 120 +++++++++++++++------ arch/s390/include/asm/kvm_host.h | 4 +- arch/s390/kvm/kvm-s390.c | 2 +- arch/x86/entry/entry_32.S | 5 + arch/x86/entry/entry_64.S | 5 + arch/x86/include/asm/hardirq.h | 3 + arch/x86/include/asm/irq_vectors.h | 6 +- arch/x86/include/asm/kvm_host.h | 12 ++- arch/x86/include/asm/kvm_para.h | 10 +- arch/x86/include/uapi/asm/kvm_para.h | 19 +++- arch/x86/kernel/irq.c | 9 ++ arch/x86/kernel/kvm.c | 62 +++++++---- arch/x86/kvm/cpuid.c | 3 +- arch/x86/kvm/mmu/mmu.c | 19 ++-- arch/x86/kvm/svm/nested.c | 2 +- arch/x86/kvm/svm/svm.c | 3 +- arch/x86/kvm/vmx/nested.c | 2 +- arch/x86/kvm/vmx/vmx.c | 5 +- arch/x86/kvm/x86.c | 149 ++++++++++++++++++--------- include/linux/kvm_host.h | 3 + include/uapi/linux/kvm.h | 1 + virt/kvm/async_pf.c | 12 ++- virt/kvm/kvm_main.c | 19 +++- 24 files changed, 344 insertions(+), 137 deletions(-)