perf kvm: Add kvm stat support on riscv
Subsystems (4)
cover perf kvm: Add kvm stat support on riscv
'perf kvm stat report/record' generates a statistical analysis of KVM events and can be used to analyze guest exit reasons. This patch tries to add stat support on riscv.
Map the return value of trace_kvm_exit() to the specific cause of the exception, and export it to userspace.
It records on two available KVM tracepoints for riscv: "kvm:kvm_entry" and "kvm:kvm_exit", and reports statistical data which includes events handles time, samples, and so on.
Simple tests go below:
# ./perf kvm record -e "kvm:kvm_entry" -e "kvm:kvm_exit" Lowering default frequency rate from 4000 to 2500. Please consider tweaking /proc/sys/kernel/perf_event_max_sample_rate. [ perf record: Woken up 18 times to write data ] [ perf record: Captured and wrote 5.433 MB perf.data.guest (62519 samples)
# ./perf kvm report 31K kvm:kvm_entry 31K kvm:kvm_exit
# ./perf kvm stat record -a [ perf record: Woken up 3 times to write data ] [ perf record: Captured and wrote 8.502 MB perf.data.guest (99338 samples) ]
# ./perf kvm stat report --event=vmexit Event name Samples Sample% Time (ns) Time% Max Time (ns) Min Time (ns) Mean Time (ns) STORE_GUEST_PAGE_FAULT 26968 54.00% 2003031800 40.00% 3361400 27600 74274 LOAD_GUEST_PAGE_FAULT 17645 35.00% 1153338100 23.00% 2513400 30800 65363 VIRTUAL_INST_FAULT 1247 2.00% 340820800 6.00% 1190800 43300 273312 INST_GUEST_PAGE_FAULT 1128 2.00% 340645800 6.00% 2123200 30200 301990 SUPERVISOR_SYSCALL 1019 2.00% 245989900 4.00% 1851500 29300 241403 LOAD_ACCESS 986 1.00% 671556200 13.00% 4180200 100700 681091 INST_ACCESS 655 1.00% 170054800 3.00% 1808300 54600 259625 HYPERVISOR_SYSCALL 21 0.00% 4276400 0.00% 716500 116000 203638
Shenlin Liang (2):
RISCV: KVM: add tracepoints for entry and exit events perf kvm/riscv: Port perf kvm stat to RISC-V
arch/riscv/kvm/trace_riscv.h | 60 ++++++++++++++ arch/riscv/kvm/vcpu.c | 7 ++ tools/perf/arch/riscv/Makefile | 1 + tools/perf/arch/riscv/util/Build | 1 + tools/perf/arch/riscv/util/kvm-stat.c | 78 +++++++++++++++++++ .../arch/riscv/util/riscv_exception_types.h | 41 ++++++++++ 6 files changed, 188 insertions(+) create mode 100644 arch/riscv/kvm/trace_riscv.h create mode 100644 tools/perf/arch/riscv/util/kvm-stat.c create mode 100644 tools/perf/arch/riscv/util/riscv_exception_types.h
Signature · 1 line
2.37.2
PATCH 1/2 RISCV: KVM: add tracepoints for entry and exit events
Like other architectures, RISCV KVM also needs to add these event tracepoints to count the number of times kvm guest entry/exit.
--- arch/riscv/kvm/trace_riscv.h | 60 ++++++++++++++++++++++++++++++++++++ arch/riscv/kvm/vcpu.c | 7 +++++ 2 files changed, 67 insertions(+) create mode 100644 arch/riscv/kvm/trace_riscv.h
arch/riscv/kvm/trace_riscv.h new file +60 -0
@@ -0,0 +1,60 @@ +// SPDX-License-Identifier: GPL-2.0 +/* +* Tracepoints for RISC-V KVM +* +* Copyright 2024 Beijing ESWIN Computing Technology Co., Ltd. +* +*/ +#if !defined(_TRACE_RSICV_KVM_H) || defined(TRACE_HEADER_MULTI_READ ) +#define _TRACE_RSICV_KVM_H + +#include <linux/tracepoint.h> + +#undef TRACE_SYSTEM +#define TRACE_SYSTEM kvm + +TRACE_EVENT (kvm_entry, + TP_PROTO(struct kvm_vcpu *vcpu), +TP_ARGS (vcpu), + +TP_STRUCT__entry ( + __field(unsigned long, pc) + ), + +TP_fast_assign ( + __entry->pc = vcpu->arch.guest_context.sepc; + ), + +TP_printk ("PC: 0x%016lx" , __entry->pc) +); + +TRACE_EVENT (kvm_exit, + TP_PROTO(struct kvm_vcpu *vcpu, unsigned long exit_reason, + unsigned long scause), +TP_ARGS (vcpu, exit_reason, scause), + +TP_STRUCT__entry ( + __field(unsigned long, pc) + __field(unsigned long, exit_reason) + __field(unsigned long, scause) + ), + +TP_fast_assign ( + __entry->pc = vcpu->arch.guest_context.sepc; + __entry->exit_reason = exit_reason; + __entry->scause = scause; + ), + +TP_printk ("EXIT_REASON:0x%lx,PC: 0x%016lx,SCAUSE:0x%lx" , + __entry->exit_reason, __entry->pc, __entry->scause) +); + +#endif /* _TRACE_RSICV_KVM_H */ + +#undef TRACE_INCLUDE_PATH +#define TRACE_INCLUDE_PATH . +#undef TRACE_INCLUDE_FILE +#define TRACE_INCLUDE_FILE trace_riscv + +/* This part must be outside protection */ +#include <trace/define_trace.h>
arch/riscv/kvm/vcpu.c +7 -0
@@ -21,6 +21,9 @@#include <asm/cacheflush.h> #include <asm/kvm_vcpu_vector.h> +#define CREATE_TRACE_POINTS +#include "trace_riscv.h" +const struct _kvm_stats_desc kvm_vcpu_stats_desc[] = {KVM_GENERIC_VCPU_STATS (),STATS_DESC_COUNTER (VCPU , ecall_exit_stat),
@@ -782,6 +785,8 @@ int kvm_arch_vcpu_ioctl_run(struct kvm_vcpu *vcpu)*/ kvm_riscv_local_tlb_sanitize (vcpu); +trace_kvm_entry (vcpu); +guest_timing_enter_irqoff ();kvm_riscv_vcpu_enter_exit (vcpu);
@@ -820,6 +825,8 @@ int kvm_arch_vcpu_ioctl_run(struct kvm_vcpu *vcpu)local_irq_enable (); +trace_kvm_exit (vcpu, run->exit_reason, trap.scause); +preempt_enable ();kvm_vcpu_srcu_read_lock (vcpu);
Signature · 1 line
2.37.2
Anup Patel
maintainerRe: [PATCH 1/2] RISCV: KVM: add tracepoints for entry and exit events
Quoted 17 lines · On Thu, Mar 28, 2024 at 8:49?AM Shenlin Liang…
Like other architectures, RISCV KVM also needs to add these event tracepoints to count the number of times kvm guest entry/exit.
Signed-off-by Shenlin Liang <[email protected]>--- arch/riscv/kvm/trace_riscv.h | 60 ++++++++++++++++++++++++++++++++++++ arch/riscv/kvm/vcpu.c | 7 +++++ 2 files changed, 67 insertions(+) create mode 100644 arch/riscv/kvm/trace_riscv.harch/riscv/kvm/trace_riscv.h new file +0
-0
arch/riscv/kvm/trace_riscv.h line 9
Quoted 9 lines
arch/riscv/kvm/trace_riscv.h new file +9
-0@@ -0,0 +1,60 @@ +// SPDX-License-Identifier: GPL-2.0 +/* +* Tracepoints for RISC-V KVM +* +* Copyright 2024 Beijing ESWIN Computing Technology Co., Ltd. +* +*/ +#if !defined(_TRACE_RSICV_KVM_H) || defined(TRACE_HEADER_MULTI_READ ) +#define _TRACE_RSICV_KVM_H
s/_RSICV_/_RISCV_/
Quoted 20 lines
arch/riscv/kvm/trace_riscv.h new file +20
-0+ +#include <linux/tracepoint.h> + +#undef TRACE_SYSTEM +#define TRACE_SYSTEM kvm + +TRACE_EVENT (kvm_entry, + TP_PROTO(struct kvm_vcpu *vcpu), +TP_ARGS (vcpu), + +TP_STRUCT__entry ( + __field(unsigned long, pc) + ), + +TP_fast_assign ( + __entry->pc = vcpu->arch.guest_context.sepc; + ), + +TP_printk ("PC: 0x%016lx" , __entry->pc) +);
arch/riscv/kvm/trace_riscv.h line 39
Quoted 10 lines
arch/riscv/kvm/trace_riscv.h new file +10
-0+ +TRACE_EVENT(kvm_exit, + TP_PROTO(struct kvm_vcpu *vcpu, unsigned long exit_reason, + unsigned long scause), + TP_ARGS(vcpu, exit_reason, scause), + + TP_STRUCT__entry( + __field(unsigned long, pc) + __field(unsigned long, exit_reason) + __field(unsigned long, scause)
This is not the right contents describing a KVM exit.
The fields over here should be aligned with "struct kvm_cpu_trap" so we should have following fields:
__field(unsigned long, sepc)
__field(unsigned long, scause)
__field(unsigned long, stval)
__field(unsigned long, htval)
__field(unsigned long, htinst)
Quoted 54 lines
arch/riscv/kvm/trace_riscv.h new file +21
-0+ ), + +TP_fast_assign ( + __entry->pc = vcpu->arch.guest_context.sepc; + __entry->exit_reason = exit_reason; + __entry->scause = scause; + ), + +TP_printk ("EXIT_REASON:0x%lx,PC: 0x%016lx,SCAUSE:0x%lx" , + __entry->exit_reason, __entry->pc, __entry->scause) +); + +#endif /* _TRACE_RSICV_KVM_H */ + +#undef TRACE_INCLUDE_PATH +#define TRACE_INCLUDE_PATH . +#undef TRACE_INCLUDE_FILE +#define TRACE_INCLUDE_FILE trace_riscv + +/* This part must be outside protection */ +#include <trace/define_trace.h> arch/riscv/kvm/vcpu.c +7
-1@@ -21,6 +21,9 @@#include <asm/cacheflush.h> #include <asm/kvm_vcpu_vector.h> +#define CREATE_TRACE_POINTS +#include "trace_riscv.h" +const struct _kvm_stats_desc kvm_vcpu_stats_desc[] = {KVM_GENERIC_VCPU_STATS (),STATS_DESC_COUNTER (VCPU , ecall_exit_stat),@@ -782,6 +785,8 @@ int kvm_arch_vcpu_ioctl_run(struct kvm_vcpu *vcpu)*/ kvm_riscv_local_tlb_sanitize (vcpu); +trace_kvm_entry (vcpu); +guest_timing_enter_irqoff ();kvm_riscv_vcpu_enter_exit (vcpu);@@ -820,6 +825,8 @@ int kvm_arch_vcpu_ioctl_run(struct kvm_vcpu *vcpu)local_irq_enable (); +trace_kvm_exit (vcpu, run->exit_reason, trap.scause); +preempt_enable ();kvm_vcpu_srcu_read_lock (vcpu); --
Quoted 2 lines
2.37.2
Regards,
Anup
Re: [PATCH 1/2] RISCV: KVM: add tracepoints for entry and exit events
Quoted 125 lines · Anup Patel
Quoted 17 lines · On Thu, Mar 28, 2024 at 8:49?AM Shenlin Liang…
Like other architectures, RISCV KVM also needs to add these event tracepoints to count the number of times kvm guest entry/exit.
Signed-off-by Shenlin Liang <[email protected]>--- arch/riscv/kvm/trace_riscv.h | 60 ++++++++++++++++++++++++++++++++++++ arch/riscv/kvm/vcpu.c | 7 +++++ 2 files changed, 67 insertions(+) create mode 100644 arch/riscv/kvm/trace_riscv.harch/riscv/kvm/trace_riscv.h new file +0
-0arch/riscv/kvm/trace_riscv.h line 9
Quoted 9 lines
arch/riscv/kvm/trace_riscv.h new file +9
-0@@ -0,0 +1,60 @@ +// SPDX-License-Identifier: GPL-2.0 +/* +* Tracepoints for RISC-V KVM +* +* Copyright 2024 Beijing ESWIN Computing Technology Co., Ltd. +* +*/ +#if !defined(_TRACE_RSICV_KVM_H) || defined(TRACE_HEADER_MULTI_READ ) +#define _TRACE_RSICV_KVM_Hs/_RSICV_/_RISCV_/
Quoted 20 lines
arch/riscv/kvm/trace_riscv.h new file +20
-0+ +#include <linux/tracepoint.h> + +#undef TRACE_SYSTEM +#define TRACE_SYSTEM kvm + +TRACE_EVENT (kvm_entry, + TP_PROTO(struct kvm_vcpu *vcpu), +TP_ARGS (vcpu), + +TP_STRUCT__entry ( + __field(unsigned long, pc) + ), + +TP_fast_assign ( + __entry->pc = vcpu->arch.guest_context.sepc; + ), + +TP_printk ("PC: 0x%016lx" , __entry->pc) +);arch/riscv/kvm/trace_riscv.h line 39
Quoted 10 lines
arch/riscv/kvm/trace_riscv.h new file +10
-0+ +TRACE_EVENT(kvm_exit, + TP_PROTO(struct kvm_vcpu *vcpu, unsigned long exit_reason, + unsigned long scause), + TP_ARGS(vcpu, exit_reason, scause), + + TP_STRUCT__entry( + __field(unsigned long, pc) + __field(unsigned long, exit_reason) + __field(unsigned long, scause)This is not the right contents describing a KVM exit.
The fields over here should be aligned with "struct
kvm_cpu_trap" so we should have following fields:
__field(unsigned long, sepc)
__field(unsigned long, scause)
__field(unsigned long, stval)
__field(unsigned long, htval)
__field(unsigned long, htinst)Quoted 52 lines
arch/riscv/kvm/trace_riscv.h new file +21
-0+ ), + +TP_fast_assign ( + __entry->pc = vcpu->arch.guest_context.sepc; + __entry->exit_reason = exit_reason; + __entry->scause = scause; + ), + +TP_printk ("EXIT_REASON:0x%lx,PC: 0x%016lx,SCAUSE:0x%lx" , + __entry->exit_reason, __entry->pc, __entry->scause) +); + +#endif /* _TRACE_RSICV_KVM_H */ + +#undef TRACE_INCLUDE_PATH +#define TRACE_INCLUDE_PATH . +#undef TRACE_INCLUDE_FILE +#define TRACE_INCLUDE_FILE trace_riscv + +/* This part must be outside protection */ +#include <trace/define_trace.h> arch/riscv/kvm/vcpu.c +7
-0@@ -21,6 +21,9 @@#include <asm/cacheflush.h> #include <asm/kvm_vcpu_vector.h> +#define CREATE_TRACE_POINTS +#include "trace_riscv.h" +const struct _kvm_stats_desc kvm_vcpu_stats_desc[] = {KVM_GENERIC_VCPU_STATS (),STATS_DESC_COUNTER (VCPU , ecall_exit_stat),@@ -782,6 +785,8 @@ int kvm_arch_vcpu_ioctl_run(struct kvm_vcpu *vcpu)*/ kvm_riscv_local_tlb_sanitize (vcpu); +trace_kvm_entry (vcpu); +guest_timing_enter_irqoff ();kvm_riscv_vcpu_enter_exit (vcpu);@@ -820,6 +825,8 @@ int kvm_arch_vcpu_ioctl_run(struct kvm_vcpu *vcpu)local_irq_enable (); +trace_kvm_exit (vcpu, run->exit_reason, trap.scause); +preempt_enable ();
Quoted 7 lines
Quoted 4 lines
arch/riscv/kvm/vcpu.c +0
-1kvm_vcpu_srcu_read_lock (vcpu); --2.37.2
Regards,
Anup
Thank you for the review. I will send the v2 version.
Re: [PATCH 1/2] RISCV: KVM: add tracepoints for entry and exit events
In reply to Shenlin LiangQuoted 6 lines · Shenlin Liang
Like other architectures, RISCV KVM also needs to add these event tracepoints to count the number of times kvm guest entry/exit.
Signed-off-by Shenlin Liang <[email protected]>--- arch/riscv/kvm/trace_riscv.h | 60 ++++++++++++++++++++++++++++++++++++
Conventionally, it should be named to trace.h, especially if only one. Refer to other arch's.
Re: [PATCH 1/2] RISCV: KVM: add tracepoints for entry and exit events
Quoted 11 lines · Eric Cheng
Quoted 6 lines · Shenlin Liang
Like other architectures, RISCV KVM also needs to add these event tracepoints to count the number of times kvm guest entry/exit.
Signed-off-by Shenlin Liang <[email protected]>--- arch/riscv/kvm/trace_riscv.h | 60 ++++++++++++++++++++++++++++++++++++Conventionally, it should be named to
trace.h, especially if only one. Refer to other arch's.
Thank you for the review. I will send the v2 version.
PATCH 2/2 perf kvm/riscv: Port perf kvm stat to RISC-V
In reply to Shenlin Liang'perf kvm stat report/record' generates a statistical analysis of KVM events and can be used to analyze guest exit reasons.
"report" reports statistical analysis of guest exit events.
To record kvm events on the host:
# perf kvm stat record -a
To report kvm VM EXIT events:
# perf kvm stat report --event=vmexit
--- tools/perf/arch/riscv/Makefile | 1 + tools/perf/arch/riscv/util/Build | 1 + tools/perf/arch/riscv/util/kvm-stat.c | 78 +++++++++++++++++++ .../arch/riscv/util/riscv_exception_types.h | 41 ++++++++++ 4 files changed, 121 insertions(+) create mode 100644 tools/perf/arch/riscv/util/kvm-stat.c create mode 100644 tools/perf/arch/riscv/util/riscv_exception_types.h
tools/perf/arch/riscv/Makefile +1 -0
@@ -3,3 +3,4 @@ PERF_HAVE_DWARF_REGS := 1 endif PERF_HAVE_ARCH_REGS_QUERY_REGISTER_OFFSET := 1 PERF_HAVE_JITDUMP := 1 +HAVE_KVM_STAT_SUPPORT := 1
tools/perf/arch/riscv/util/Build +1 -0
@@ -1,5 +1,6 @@ perf-y += perf_regs.o perf-y += header.o +perf-$(CONFIG_LIBTRACEEVENT) += kvm-stat.o perf-$(CONFIG_DWARF) += dwarf-regs.o perf-$(CONFIG_LIBDW_DWARF_UNWIND) += unwind-libdw.o
tools/perf/arch/riscv/util/kvm-stat.c new file +78 -0
78 lines not on this page · raw message
tools/perf/arch/riscv/util/riscv_exception_types.h new file +41 -0
41 lines not on this page · raw message
Signature · 1 line
2.37.2
Gentle ping...
Quoted 56 lines
'perf kvm stat report/record' generates a statistical analysis of KVM events and can be used to analyze guest exit reasons. This patch tries to add stat support on riscv.
Map the return value of
trace_kvm_exit()to the specific cause of the exception, and export it to userspace.It records on two available KVM tracepoints for riscv: "kvm:kvm_entry" and "kvm:kvm_exit", and reports statistical data which includes events handles time, samples, and so on.
Simple tests go below:
# ./perf kvm record -e "kvm:kvm_entry" -e "kvm:kvm_exit" Lowering default frequency rate from 4000 to 2500. Please consider tweaking /proc/sys/kernel/perf_event_max_sample_rate. [ perf record: Woken up 18 times to write data ] [ perf record: Captured and wrote 5.433 MB perf.data.guest (62519 samples)# ./perf kvm report 31K kvm:kvm_entry 31K kvm:kvm_exit# ./perf kvm stat record -a [ perf record: Woken up 3 times to write data ] [ perf record: Captured and wrote 8.502 MB perf.data.guest (99338 samples) ]# ./perf kvm stat report --event=vmexit Event name Samples Sample% Time (ns) Time% Max Time (ns) Min Time (ns) Mean Time (ns) STORE_GUEST_PAGE_FAULT 26968 54.00% 2003031800 40.00% 3361400 27600 74274 LOAD_GUEST_PAGE_FAULT 17645 35.00% 1153338100 23.00% 2513400 30800 65363 VIRTUAL_INST_FAULT 1247 2.00% 340820800 6.00% 1190800 43300 273312 INST_GUEST_PAGE_FAULT 1128 2.00% 340645800 6.00% 2123200 30200 301990 SUPERVISOR_SYSCALL 1019 2.00% 245989900 4.00% 1851500 29300 241403 LOAD_ACCESS 986 1.00% 671556200 13.00% 4180200 100700 681091 INST_ACCESS 655 1.00% 170054800 3.00% 1808300 54600 259625 HYPERVISOR_SYSCALL 21 0.00% 4276400 0.00% 716500 116000 203638Shenlin Liang (2):
RISCV: KVM: add tracepoints for entry and exit events perf kvm/riscv: Port perf kvm stat to RISC-Varch/riscv/kvm/trace_riscv.h | 60 ++++++++++++++ arch/riscv/kvm/vcpu.c | 7 ++ tools/perf/arch/riscv/Makefile | 1 + tools/perf/arch/riscv/util/Build | 1 + tools/perf/arch/riscv/util/kvm-stat.c | 78 +++++++++++++++++++ .../arch/riscv/util/riscv_exception_types.h | 41 ++++++++++ 6 files changed, 188 insertions(+) create mode 100644 arch/riscv/kvm/trace_riscv.h create mode 100644 tools/perf/arch/riscv/util/kvm-stat.c create mode 100644 tools/perf/arch/riscv/util/riscv_exception_types.hSignature · 1 line
2.37.2