On 8/6/2021 1:03 AM, Jim Mattson wrote:
On Thu, Aug 5, 2021 at 8:38 AM Zeng Guang <guang.zeng@xxxxxxxxx> wrote:
Current IPI process in guest VM will virtualize the writing to interrupt
command register(ICR) of the local APIC which will cause VM-exit anyway
on source vCPU. Frequent VM-exit could induce much overhead accumulated
if running IPI intensive task.
IPI virtualization as a new VT-x feature targets to eliminate VM-exits
when issuing IPI on source vCPU. It introduces a new VM-execution
control - "IPI virtualization"(bit4) in the tertiary processor-based
VM-exection controls and a new data structure - "PID-pointer table
VM-execution
address" and "Last PID-pointer index" referenced by the VMCS. When "IPI
virtualization" is enabled, processor emulateds following kind of writes
emulates
to APIC registers that would send IPIs, moreover without causing VM-exits.
- Memory-mapped ICR writes
- MSR-mapped ICR writes
- SENDUIPI execution
This patch series implement IPI virtualization support in KVM.
implements
Patches 1-4 add tertiary processor-based VM-execution support
framework.
Patch 5 implement interrupt dispatch support in x2APIC mode with
implements
APIC-write VM exit. In previous platform, no CPU would produce
APIC-write VM exit with exit qulification 300H when the "virtual x2APIC
qualification
mode" VM-execution control was 1.
Patch 6 implement IPI virtualization related function including
feature enabling through tertiary processor-based VM-execution in
various scenario of VMCS configuration, PID table setup in vCPU creation
scenarios
and vCPU block consideration.
Document for IPI virtualization is now available at the latest "Intel
Architecture Instruction Set Extensions Programming Reference".
Document Link:
https://software.intel.com/content/www/us/en/develop/download/intel-architecture-instruction-set-extensions-programming-reference.html
We did experiment to measure average time sending IPI from source vCPU
to the target vCPU completing the IPI handling by kvm unittest w/ and
w/o IPI virtualization. When IPI virtualizatin enabled, it will reduce
virtualization
22.21% and 15.98% cycles consuming in xAPIC mode and x2APIC mode
respectly.
respectively
KMV unittest:vmexit/ipi, 2 vCPU, AP was modified to run in idle loop
KVM
Thanks for your all corrections.
instead of halt to ensure no VM exit impact on target vCPU.
Are you going to post the kvm-init-test changes?
We modified unit test common code specific for the IPIv test purpose,
currently no plan to post those changes to kvm-uint-tests.
Cycles of IPI
xAPIC mode x2APIC mode
test w/o IPIv w/ IPIv w/o IPIv w/ IPIv
1 6106 4816 4265 3768
2 6244 4656 4404 3546
3 6165 4658 4233 3474
4 5992 4710 4363 3430
5 6083 4741 4215 3551
6 6238 4904 4304 3547
7 6164 4617 4263 3709
8 5984 4763 4518 3779
9 5931 4712 4645 3667
10 5955 4530 4332 3724
11 5897 4673 4283 3569
12 6140 4794 4178 3598
13 6183 4728 4363 3628
14 5991 4994 4509 3842
15 5866 4665 4520 3739
16 6032 4654 4229 3701
17 6050 4653 4185 3726
18 6004 4792 4319 3746
19 5961 4626 4196 3392
20 6194 4576 4433 3760
Average cycles 6059 4713.1 4337.85 3644.8
%Reduction -22.21% -15.98%
--------------------------------------
IPI microbenchmark:
(https://lore.kernel.org/kvm/20171219085010.4081-1-ynorov@xxxxxxxxxxxxxxxxxx)
2 vCPUs, 1:1 pin vCPU to pCPU, guest VM runs with idle=poll, x2APIC mode
For dedicated CPUs, we would just disable the MONITOR/MWAIT intercepts
and expose MONITOR/MWAIT to the guest. We would not recommend that
anyone use idle=poll.
OK. We thought that straightforward way to make guest keep running in
idle without side
effect of VM exit and other factors, so used idle=poll for test.
Result with IPIv enabled:
Dry-run: 0, 272798 ns
Self-IPI: 5094123, 11114037 ns
Normal IPI: 131697087, 173321200 ns
Broadcast IPI: 0, 155649075 ns
Broadcast lock: 0, 161518031 ns
Result with IPIv disabled:
Dry-run: 0, 272766 ns
Self-IPI: 5091788, 11123699 ns
Normal IPI: 145215772, 174558920 ns
Broadcast IPI: 0, 175785384 ns
Broadcast lock: 0, 149076195 ns
As IPIv can benefit unicast IPI to other CPU, Noraml IPI test case gain
Normal
about 9.73% time saving on average out of 15 test runs when IPIv is
enabled.
Can you share the CDFs?
Normal IPI statistics:
test w/o IPIv (unit:ns) w/ IPIv
(unit:ns)
1 153346049 140907046
2 147218648 141660618
3 145215772 117890672
4 146621682 136430470
5 144821472 136199421
6 144704378 131676928
7 141403224 131697087
8 144775766 125476250
9 140658192 137263330
10 144768626 138593127
11 145166679 131946752
12 145020451 116852889
13 148161353 131406280
14 148378655 130174353
15 148903652 127969674
w/o IPIv w/ IPIv
Normal IPI: 145944306.6 ns 131742993.1 ns
%Reduction -9.73%
--------------------------------------
hackbench:
8 vCPUs, guest VM free run, x2APIC mode
./hackbench -p -l 100000
w/o IPIv w/ IPIv
Time: 91.887 74.605
%Reduction: -18.808%
96 vCPUs, guest VM free run, x2APIC mode
./hackbench -p -l 1000000
w/o IPIv w/ IPIv
Time: 287.504 235.185
%Reduction: -18.198%
--------------------------------------
v2 -> v3:
1. Misc change on tertiary execution control
definition and capability setup
2. Alternative to get tertiary execution
control configuration
v1 -> v2:
1. Refine the IPIv enabling logic for VM.
Remove ipiv_active definition per vCPU.
Gao Chao (1):
KVM: VMX: enable IPI virtualization
Robert Hoo (4):
x86/feat_ctl: Add new VMX feature, Tertiary VM-Execution control
KVM: VMX: Extend BUILD_CONTROLS_SHADOW macro to support 64-bit
variation
KVM: VMX: Detect Tertiary VM-Execution control when setup VMCS config
KVM: VMX: dump_vmcs() reports tertiary_exec_control field as well
Zeng Guang (1):
KVM: x86: Support interrupt dispatch in x2APIC mode with APIC-write VM
exit
arch/x86/include/asm/msr-index.h | 1 +
arch/x86/include/asm/vmx.h | 11 +++
arch/x86/include/asm/vmxfeatures.h | 5 +-
arch/x86/kernel/cpu/feat_ctl.c | 11 ++-
arch/x86/kvm/lapic.c | 9 ++-
arch/x86/kvm/vmx/capabilities.h | 14 ++++
arch/x86/kvm/vmx/evmcs.c | 2 +
arch/x86/kvm/vmx/evmcs.h | 1 +
arch/x86/kvm/vmx/posted_intr.c | 22 ++++--
arch/x86/kvm/vmx/vmcs.h | 1 +
arch/x86/kvm/vmx/vmx.c | 114 +++++++++++++++++++++++++++--
arch/x86/kvm/vmx/vmx.h | 27 ++++---
12 files changed, 193 insertions(+), 25 deletions(-)
--
2.25.1
