[PATCH v10 00/11] Guest Last Branch Recording Enabling
From: Like Xu
Date: Thu Apr 23 2020 - 04:17:38 EST
Hi all,
Please help review your interesting parts in this stable version,
e.g. the first five patches involve the perf event subsystem
and the sixth patch concerns the KVM userspace interface.
v9->v10 Changelog:
- new patch (0002) to refactor hw->idx checks and cleanup for host perf;
- refine comments in guest LBR constraint patch and rename functions;
- still ack LBRS_FROZEN for guest LBR event in the intel_pmu_ack_status();
- add more checks before enabling LBR in the kvm_vm_ioctl_enable_cap();
- add pmu_ops->deliver_pmi to clear LBR enable bit for guest debugctl_msr;
- use vmx_supported_perf_capabilities() to expose PDCM via kvm_cpu_cap*();
You may check more details in each commit.
Previous:
https://lore.kernel.org/kvm/20200313021616.112322-1-like.xu@xxxxxxxxxxxxxxx/
---
The last branch recording (LBR) is a performance monitor unit (PMU)
feature on Intel processors that records a running trace of the most
recent branches taken by the processor in the LBR stack. This patch
series is going to enable this feature for plenty of KVM guests.
The userspace could configure whether it's enabled or not for each
guest via vm_ioctl KVM_CAP_X86_GUEST_LBR. As a first step, a guest
could only enable LBR feature if its cpu model is the same as the
host since the LBR feature is still one of model specific features.
If it's enabled on the guest, the guest LBR driver would accesses the
LBR MSR (including IA32_DEBUGCTLMSR and records MSRs) as host does.
The first guest access on the LBR related MSRs is always interceptible.
The KVM trap would create a special LBR event (called guest LBR event)
which enables the callstack mode and none of hardware counter is assigned.
The host perf would enable and schedule this event as usual.
Guest's first access to a LBR registers gets trapped to KVM, which
creates a guest LBR perf event. It's a regular LBR perf event which gets
the LBR facility assigned from the perf subsystem. Once that succeeds,
the LBR stack msrs are passed through to the guest for efficient accesses.
However, if another host LBR event comes in and takes over the LBR
facility, the LBR msrs will be made interceptible, and guest following
accesses to the LBR msrs will be trapped and meaningless.
Because saving/restoring tens of LBR MSRs (e.g. 32 LBR stack entries) in
VMX transition brings too excessive overhead to frequent vmx transition
itself, the guest LBR event would help save/restore the LBR stack msrs
during the context switching with the help of native LBR event callstack
mechanism, including LBR_SELECT msr.
If the guest no longer accesses the LBR-related MSRs within a scheduling
time slice and the LBR enable bit is unset, vPMU would release its guest
LBR event as a normal event of a unused vPMC and the pass-through
state of the LBR stack msrs would be canceled.
---
LBR testcase:
echo 1 > /proc/sys/kernel/watchdog
echo 25 > /proc/sys/kernel/perf_cpu_time_max_percent
echo 5000 > /proc/sys/kernel/perf_event_max_sample_rate
echo 0 > /proc/sys/kernel/perf_cpu_time_max_percent
./perf record -b ./br_instr a
- Perf report on the host:
Samples: 72K of event 'cycles', Event count (approx.): 72512
Overhead Command Source Shared Object Source Symbol Target Symbol Basic Block Cycles
12.12% br_instr br_instr [.] cmp_end [.] lfsr_cond 1
11.05% br_instr br_instr [.] lfsr_cond [.] cmp_end 5
8.81% br_instr br_instr [.] lfsr_cond [.] cmp_end 4
5.04% br_instr br_instr [.] cmp_end [.] lfsr_cond 20
4.92% br_instr br_instr [.] lfsr_cond [.] cmp_end 6
4.88% br_instr br_instr [.] cmp_end [.] lfsr_cond 6
4.58% br_instr br_instr [.] cmp_end [.] lfsr_cond 5
- Perf report on the guest:
Samples: 92K of event 'cycles', Event count (approx.): 92544
Overhead Command Source Shared Object Source Symbol Target Symbol Basic Block Cycles
12.03% br_instr br_instr [.] cmp_end [.] lfsr_cond 1
11.09% br_instr br_instr [.] lfsr_cond [.] cmp_end 5
8.57% br_instr br_instr [.] lfsr_cond [.] cmp_end 4
5.08% br_instr br_instr [.] lfsr_cond [.] cmp_end 6
5.06% br_instr br_instr [.] cmp_end [.] lfsr_cond 20
4.87% br_instr br_instr [.] cmp_end [.] lfsr_cond 6
4.70% br_instr br_instr [.] cmp_end [.] lfsr_cond 5
Conclusion: the profiling results on the guest are similar to that on the host.
Like Xu (8):
perf/x86/core: Refactor hw->idx checks and cleanup
perf/x86/lbr: Add interface to get basic information about LBR stack
perf/x86: Add constraint to create guest LBR event without hw counter
perf/x86: Keep LBR stack unchanged in host context for guest LBR event
KVM: x86: Add KVM_CAP_X86_GUEST_LBR to dis/enable LBR from user-space
KVM: x86/pmu: Add LBR feature emulation via guest LBR event
KVM: x86/pmu: Release guest LBR event via vPMU lazy release mechanism
KVM: x86: Expose MSR_IA32_PERF_CAPABILITIES for LBR record format
Wei Wang (3):
perf/x86: Fix variable type for LBR registers
KVM: x86/pmu: Tweak kvm_pmu_get_msr to pass 'struct msr_data' in
KVM: x86: Remove the common trap handler of the MSR_IA32_DEBUGCTLMSR
Documentation/virt/kvm/api.rst | 28 +++
arch/x86/events/core.c | 26 ++-
arch/x86/events/intel/core.c | 102 ++++++----
arch/x86/events/intel/lbr.c | 56 +++++-
arch/x86/events/perf_event.h | 18 +-
arch/x86/include/asm/kvm_host.h | 14 ++
arch/x86/include/asm/perf_event.h | 28 ++-
arch/x86/kvm/pmu.c | 27 ++-
arch/x86/kvm/pmu.h | 17 +-
arch/x86/kvm/svm/pmu.c | 7 +-
arch/x86/kvm/vmx/capabilities.h | 15 ++
arch/x86/kvm/vmx/pmu_intel.c | 301 ++++++++++++++++++++++++++++--
arch/x86/kvm/vmx/vmx.c | 11 +-
arch/x86/kvm/vmx/vmx.h | 2 +
arch/x86/kvm/x86.c | 34 ++--
include/linux/perf_event.h | 7 +
include/uapi/linux/kvm.h | 1 +
kernel/events/core.c | 7 -
18 files changed, 603 insertions(+), 98 deletions(-)
--
2.21.1