* [PATCH v2 00/22] KVM: ARM64: Add guest PMU support
@ 2015-09-11 8:54 ` Shannon Zhao
0 siblings, 0 replies; 113+ messages in thread
From: Shannon Zhao @ 2015-09-11 8:54 UTC (permalink / raw)
To: linux-arm-kernel
From: Shannon Zhao <shannon.zhao@linaro.org>
This patchset adds guest PMU support for KVM on ARM64. It takes
trap-and-emulate approach. When guest wants to monitor one event, it
will be trapped by KVM and KVM will call perf_event API to create a perf
event and call relevant perf_event APIs to get the count value of event.
Use perf to test this patchset in guest. When using "perf list", it
shows the list of the hardware events and hardware cache events perf
supports. Then use "perf stat -e EVENT" to monitor some event. For
example, use "perf stat -e cycles" to count cpu cycles and
"perf stat -e cache-misses" to count cache misses.
Below are the outputs of "perf stat -r 5 sleep 5" when running in host
and guest.
Host:
Performance counter stats for 'sleep 5' (5 runs):
0.551428 task-clock (msec) # 0.000 CPUs utilized ( +- 0.91% )
1 context-switches # 0.002 M/sec
0 cpu-migrations # 0.000 K/sec
48 page-faults # 0.088 M/sec ( +- 1.05% )
1150265 cycles # 2.086 GHz ( +- 0.92% )
<not supported> stalled-cycles-frontend
<not supported> stalled-cycles-backend
526398 instructions # 0.46 insns per cycle ( +- 0.89% )
<not supported> branches
9485 branch-misses # 17.201 M/sec ( +- 2.35% )
5.000831616 seconds time elapsed ( +- 0.00% )
Guest:
Performance counter stats for 'sleep 5' (5 runs):
0.730868 task-clock (msec) # 0.000 CPUs utilized ( +- 1.13% )
1 context-switches # 0.001 M/sec
0 cpu-migrations # 0.000 K/sec
48 page-faults # 0.065 M/sec ( +- 0.42% )
1642982 cycles # 2.248 GHz ( +- 1.04% )
<not supported> stalled-cycles-frontend
<not supported> stalled-cycles-backend
637964 instructions # 0.39 insns per cycle ( +- 0.65% )
<not supported> branches
10377 branch-misses # 14.198 M/sec ( +- 1.09% )
5.001289068 seconds time elapsed ( +- 0.00% )
This patchset can be fetched from [1] and the relevant QEMU version for
test can be fetched from [2].
Thanks,
Shannon
[1] https://git.linaro.org/people/shannon.zhao/linux-mainline.git KVM_ARM64_PMU_v2
[2] https://git.linaro.org/people/shannon.zhao/qemu.git PMU_v2
Shannon Zhao (22):
ARM64: Move PMU register related defines to asm/pmu.h
KVM: ARM64: Define PMU data structure for each vcpu
KVM: ARM64: Add offset defines for PMU registers
KVM: ARM64: Add reset and access handlers for PMCR_EL0 register
KVM: ARM64: Add a helper for CP15 registers reset to UNKNOWN
KVM: ARM64: Add reset and access handlers for PMSELR register
KVM: ARM64: Add reset and access handlers for PMCEID0 and PMCEID1
register
KVM: ARM64: PMU: Add perf event map and introduce perf event creating
function
KVM: ARM64: Add reset and access handlers for PMXEVTYPER register
KVM: ARM64: Add reset and access handlers for PMXEVCNTR register
KVM: ARM64: Add reset and access handlers for PMCCNTR register
KVM: ARM64: Add reset and access handlers for PMCNTENSET and
PMCNTENCLR register
KVM: ARM64: Add reset and access handlers for PMINTENSET and
PMINTENCLR register
KVM: ARM64: Add reset and access handlers for PMOVSSET and PMOVSCLR
register
KVM: ARM64: Add a helper for CP15 registers reset to specified value
KVM: ARM64: Add reset and access handlers for PMUSERENR register
KVM: ARM64: Add reset and access handlers for PMSWINC register
KVM: ARM64: Add access handlers for PMEVCNTRn and PMEVTYPERn register
KVM: ARM64: Add PMU overflow interrupt routing
KVM: ARM64: Reset PMU state when resetting vcpu
KVM: ARM64: Free perf event of PMU when destroying vcpu
KVM: ARM64: Add a new kvm ARM PMU device
Documentation/virtual/kvm/devices/arm-pmu.txt | 15 +
arch/arm/kvm/arm.c | 4 +
arch/arm64/include/asm/kvm_asm.h | 59 ++-
arch/arm64/include/asm/kvm_host.h | 2 +
arch/arm64/include/asm/pmu.h | 49 +++
arch/arm64/include/uapi/asm/kvm.h | 3 +
arch/arm64/kernel/perf_event.c | 35 --
arch/arm64/kvm/Kconfig | 8 +
arch/arm64/kvm/Makefile | 1 +
arch/arm64/kvm/reset.c | 3 +
arch/arm64/kvm/sys_regs.c | 490 +++++++++++++++++++++++--
arch/arm64/kvm/sys_regs.h | 16 +
include/kvm/arm_pmu.h | 66 ++++
include/linux/kvm_host.h | 1 +
include/uapi/linux/kvm.h | 2 +
virt/kvm/arm/pmu.c | 500 ++++++++++++++++++++++++++
virt/kvm/kvm_main.c | 4 +
17 files changed, 1188 insertions(+), 70 deletions(-)
create mode 100644 Documentation/virtual/kvm/devices/arm-pmu.txt
create mode 100644 include/kvm/arm_pmu.h
create mode 100644 virt/kvm/arm/pmu.c
--
2.0.4
^ permalink raw reply [flat|nested] 113+ messages in thread
* [PATCH v2 00/22] KVM: ARM64: Add guest PMU support
@ 2015-09-11 8:54 ` Shannon Zhao
0 siblings, 0 replies; 113+ messages in thread
From: Shannon Zhao @ 2015-09-11 8:54 UTC (permalink / raw)
To: kvmarm; +Cc: kvm, marc.zyngier, will.deacon, linux-arm-kernel, shannon.zhao
From: Shannon Zhao <shannon.zhao@linaro.org>
This patchset adds guest PMU support for KVM on ARM64. It takes
trap-and-emulate approach. When guest wants to monitor one event, it
will be trapped by KVM and KVM will call perf_event API to create a perf
event and call relevant perf_event APIs to get the count value of event.
Use perf to test this patchset in guest. When using "perf list", it
shows the list of the hardware events and hardware cache events perf
supports. Then use "perf stat -e EVENT" to monitor some event. For
example, use "perf stat -e cycles" to count cpu cycles and
"perf stat -e cache-misses" to count cache misses.
Below are the outputs of "perf stat -r 5 sleep 5" when running in host
and guest.
Host:
Performance counter stats for 'sleep 5' (5 runs):
0.551428 task-clock (msec) # 0.000 CPUs utilized ( +- 0.91% )
1 context-switches # 0.002 M/sec
0 cpu-migrations # 0.000 K/sec
48 page-faults # 0.088 M/sec ( +- 1.05% )
1150265 cycles # 2.086 GHz ( +- 0.92% )
<not supported> stalled-cycles-frontend
<not supported> stalled-cycles-backend
526398 instructions # 0.46 insns per cycle ( +- 0.89% )
<not supported> branches
9485 branch-misses # 17.201 M/sec ( +- 2.35% )
5.000831616 seconds time elapsed ( +- 0.00% )
Guest:
Performance counter stats for 'sleep 5' (5 runs):
0.730868 task-clock (msec) # 0.000 CPUs utilized ( +- 1.13% )
1 context-switches # 0.001 M/sec
0 cpu-migrations # 0.000 K/sec
48 page-faults # 0.065 M/sec ( +- 0.42% )
1642982 cycles # 2.248 GHz ( +- 1.04% )
<not supported> stalled-cycles-frontend
<not supported> stalled-cycles-backend
637964 instructions # 0.39 insns per cycle ( +- 0.65% )
<not supported> branches
10377 branch-misses # 14.198 M/sec ( +- 1.09% )
5.001289068 seconds time elapsed ( +- 0.00% )
This patchset can be fetched from [1] and the relevant QEMU version for
test can be fetched from [2].
Thanks,
Shannon
[1] https://git.linaro.org/people/shannon.zhao/linux-mainline.git KVM_ARM64_PMU_v2
[2] https://git.linaro.org/people/shannon.zhao/qemu.git PMU_v2
Shannon Zhao (22):
ARM64: Move PMU register related defines to asm/pmu.h
KVM: ARM64: Define PMU data structure for each vcpu
KVM: ARM64: Add offset defines for PMU registers
KVM: ARM64: Add reset and access handlers for PMCR_EL0 register
KVM: ARM64: Add a helper for CP15 registers reset to UNKNOWN
KVM: ARM64: Add reset and access handlers for PMSELR register
KVM: ARM64: Add reset and access handlers for PMCEID0 and PMCEID1
register
KVM: ARM64: PMU: Add perf event map and introduce perf event creating
function
KVM: ARM64: Add reset and access handlers for PMXEVTYPER register
KVM: ARM64: Add reset and access handlers for PMXEVCNTR register
KVM: ARM64: Add reset and access handlers for PMCCNTR register
KVM: ARM64: Add reset and access handlers for PMCNTENSET and
PMCNTENCLR register
KVM: ARM64: Add reset and access handlers for PMINTENSET and
PMINTENCLR register
KVM: ARM64: Add reset and access handlers for PMOVSSET and PMOVSCLR
register
KVM: ARM64: Add a helper for CP15 registers reset to specified value
KVM: ARM64: Add reset and access handlers for PMUSERENR register
KVM: ARM64: Add reset and access handlers for PMSWINC register
KVM: ARM64: Add access handlers for PMEVCNTRn and PMEVTYPERn register
KVM: ARM64: Add PMU overflow interrupt routing
KVM: ARM64: Reset PMU state when resetting vcpu
KVM: ARM64: Free perf event of PMU when destroying vcpu
KVM: ARM64: Add a new kvm ARM PMU device
Documentation/virtual/kvm/devices/arm-pmu.txt | 15 +
arch/arm/kvm/arm.c | 4 +
arch/arm64/include/asm/kvm_asm.h | 59 ++-
arch/arm64/include/asm/kvm_host.h | 2 +
arch/arm64/include/asm/pmu.h | 49 +++
arch/arm64/include/uapi/asm/kvm.h | 3 +
arch/arm64/kernel/perf_event.c | 35 --
arch/arm64/kvm/Kconfig | 8 +
arch/arm64/kvm/Makefile | 1 +
arch/arm64/kvm/reset.c | 3 +
arch/arm64/kvm/sys_regs.c | 490 +++++++++++++++++++++++--
arch/arm64/kvm/sys_regs.h | 16 +
include/kvm/arm_pmu.h | 66 ++++
include/linux/kvm_host.h | 1 +
include/uapi/linux/kvm.h | 2 +
virt/kvm/arm/pmu.c | 500 ++++++++++++++++++++++++++
virt/kvm/kvm_main.c | 4 +
17 files changed, 1188 insertions(+), 70 deletions(-)
create mode 100644 Documentation/virtual/kvm/devices/arm-pmu.txt
create mode 100644 include/kvm/arm_pmu.h
create mode 100644 virt/kvm/arm/pmu.c
--
2.0.4
^ permalink raw reply [flat|nested] 113+ messages in thread
* [PATCH v2 00/22] KVM: ARM64: Add guest PMU support
@ 2015-09-11 8:54 ` Shannon Zhao
0 siblings, 0 replies; 113+ messages in thread
From: Shannon Zhao @ 2015-09-11 8:54 UTC (permalink / raw)
To: kvmarm; +Cc: kvm, marc.zyngier, will.deacon, linux-arm-kernel, shannon.zhao
From: Shannon Zhao <shannon.zhao@linaro.org>
This patchset adds guest PMU support for KVM on ARM64. It takes
trap-and-emulate approach. When guest wants to monitor one event, it
will be trapped by KVM and KVM will call perf_event API to create a perf
event and call relevant perf_event APIs to get the count value of event.
Use perf to test this patchset in guest. When using "perf list", it
shows the list of the hardware events and hardware cache events perf
supports. Then use "perf stat -e EVENT" to monitor some event. For
example, use "perf stat -e cycles" to count cpu cycles and
"perf stat -e cache-misses" to count cache misses.
Below are the outputs of "perf stat -r 5 sleep 5" when running in host
and guest.
Host:
Performance counter stats for 'sleep 5' (5 runs):
0.551428 task-clock (msec) # 0.000 CPUs utilized ( +- 0.91% )
1 context-switches # 0.002 M/sec
0 cpu-migrations # 0.000 K/sec
48 page-faults # 0.088 M/sec ( +- 1.05% )
1150265 cycles # 2.086 GHz ( +- 0.92% )
<not supported> stalled-cycles-frontend
<not supported> stalled-cycles-backend
526398 instructions # 0.46 insns per cycle ( +- 0.89% )
<not supported> branches
9485 branch-misses # 17.201 M/sec ( +- 2.35% )
5.000831616 seconds time elapsed ( +- 0.00% )
Guest:
Performance counter stats for 'sleep 5' (5 runs):
0.730868 task-clock (msec) # 0.000 CPUs utilized ( +- 1.13% )
1 context-switches # 0.001 M/sec
0 cpu-migrations # 0.000 K/sec
48 page-faults # 0.065 M/sec ( +- 0.42% )
1642982 cycles # 2.248 GHz ( +- 1.04% )
<not supported> stalled-cycles-frontend
<not supported> stalled-cycles-backend
637964 instructions # 0.39 insns per cycle ( +- 0.65% )
<not supported> branches
10377 branch-misses # 14.198 M/sec ( +- 1.09% )
5.001289068 seconds time elapsed ( +- 0.00% )
This patchset can be fetched from [1] and the relevant QEMU version for
test can be fetched from [2].
Thanks,
Shannon
[1] https://git.linaro.org/people/shannon.zhao/linux-mainline.git KVM_ARM64_PMU_v2
[2] https://git.linaro.org/people/shannon.zhao/qemu.git PMU_v2
Shannon Zhao (22):
ARM64: Move PMU register related defines to asm/pmu.h
KVM: ARM64: Define PMU data structure for each vcpu
KVM: ARM64: Add offset defines for PMU registers
KVM: ARM64: Add reset and access handlers for PMCR_EL0 register
KVM: ARM64: Add a helper for CP15 registers reset to UNKNOWN
KVM: ARM64: Add reset and access handlers for PMSELR register
KVM: ARM64: Add reset and access handlers for PMCEID0 and PMCEID1
register
KVM: ARM64: PMU: Add perf event map and introduce perf event creating
function
KVM: ARM64: Add reset and access handlers for PMXEVTYPER register
KVM: ARM64: Add reset and access handlers for PMXEVCNTR register
KVM: ARM64: Add reset and access handlers for PMCCNTR register
KVM: ARM64: Add reset and access handlers for PMCNTENSET and
PMCNTENCLR register
KVM: ARM64: Add reset and access handlers for PMINTENSET and
PMINTENCLR register
KVM: ARM64: Add reset and access handlers for PMOVSSET and PMOVSCLR
register
KVM: ARM64: Add a helper for CP15 registers reset to specified value
KVM: ARM64: Add reset and access handlers for PMUSERENR register
KVM: ARM64: Add reset and access handlers for PMSWINC register
KVM: ARM64: Add access handlers for PMEVCNTRn and PMEVTYPERn register
KVM: ARM64: Add PMU overflow interrupt routing
KVM: ARM64: Reset PMU state when resetting vcpu
KVM: ARM64: Free perf event of PMU when destroying vcpu
KVM: ARM64: Add a new kvm ARM PMU device
Documentation/virtual/kvm/devices/arm-pmu.txt | 15 +
arch/arm/kvm/arm.c | 4 +
arch/arm64/include/asm/kvm_asm.h | 59 ++-
arch/arm64/include/asm/kvm_host.h | 2 +
arch/arm64/include/asm/pmu.h | 49 +++
arch/arm64/include/uapi/asm/kvm.h | 3 +
arch/arm64/kernel/perf_event.c | 35 --
arch/arm64/kvm/Kconfig | 8 +
arch/arm64/kvm/Makefile | 1 +
arch/arm64/kvm/reset.c | 3 +
arch/arm64/kvm/sys_regs.c | 490 +++++++++++++++++++++++--
arch/arm64/kvm/sys_regs.h | 16 +
include/kvm/arm_pmu.h | 66 ++++
include/linux/kvm_host.h | 1 +
include/uapi/linux/kvm.h | 2 +
virt/kvm/arm/pmu.c | 500 ++++++++++++++++++++++++++
virt/kvm/kvm_main.c | 4 +
17 files changed, 1188 insertions(+), 70 deletions(-)
create mode 100644 Documentation/virtual/kvm/devices/arm-pmu.txt
create mode 100644 include/kvm/arm_pmu.h
create mode 100644 virt/kvm/arm/pmu.c
--
2.0.4
^ permalink raw reply [flat|nested] 113+ messages in thread
* [PATCH v2 01/22] ARM64: Move PMU register related defines to asm/pmu.h
2015-09-11 8:54 ` Shannon Zhao
(?)
@ 2015-09-11 8:54 ` Shannon Zhao
-1 siblings, 0 replies; 113+ messages in thread
From: Shannon Zhao @ 2015-09-11 8:54 UTC (permalink / raw)
To: linux-arm-kernel
From: Shannon Zhao <shannon.zhao@linaro.org>
To use the ARMv8 PMU related register defines from the KVM code,
we move the relevant definitions to asm/pmu.h header file.
Signed-off-by: Anup Patel <anup.patel@linaro.org>
Signed-off-by: Shannon Zhao <shannon.zhao@linaro.org>
---
arch/arm64/include/asm/pmu.h | 45 ++++++++++++++++++++++++++++++++++++++++++
arch/arm64/kernel/perf_event.c | 35 --------------------------------
2 files changed, 45 insertions(+), 35 deletions(-)
diff --git a/arch/arm64/include/asm/pmu.h b/arch/arm64/include/asm/pmu.h
index b7710a5..b9f394a 100644
--- a/arch/arm64/include/asm/pmu.h
+++ b/arch/arm64/include/asm/pmu.h
@@ -19,6 +19,51 @@
#ifndef __ASM_PMU_H
#define __ASM_PMU_H
+#define ARMV8_MAX_COUNTERS 32
+#define ARMV8_COUNTER_MASK (ARMV8_MAX_COUNTERS - 1)
+
+/*
+ * Per-CPU PMCR: config reg
+ */
+#define ARMV8_PMCR_E (1 << 0) /* Enable all counters */
+#define ARMV8_PMCR_P (1 << 1) /* Reset all counters */
+#define ARMV8_PMCR_C (1 << 2) /* Cycle counter reset */
+#define ARMV8_PMCR_D (1 << 3) /* CCNT counts every 64th cpu cycle */
+#define ARMV8_PMCR_X (1 << 4) /* Export to ETM */
+#define ARMV8_PMCR_DP (1 << 5) /* Disable CCNT if non-invasive debug*/
+#define ARMV8_PMCR_N_SHIFT 11 /* Number of counters supported */
+#define ARMV8_PMCR_N_MASK 0x1f
+#define ARMV8_PMCR_MASK 0x3f /* Mask for writable bits */
+
+/*
+ * PMCNTEN: counters enable reg
+ */
+#define ARMV8_CNTEN_MASK 0xffffffff /* Mask for writable bits */
+
+/*
+ * PMINTEN: counters interrupt enable reg
+ */
+#define ARMV8_INTEN_MASK 0xffffffff /* Mask for writable bits */
+
+/*
+ * PMOVSR: counters overflow flag status reg
+ */
+#define ARMV8_OVSR_MASK 0xffffffff /* Mask for writable bits */
+#define ARMV8_OVERFLOWED_MASK ARMV8_OVSR_MASK
+
+/*
+ * PMXEVTYPER: Event selection reg
+ */
+#define ARMV8_EVTYPE_MASK 0xc80003ff /* Mask for writable bits */
+#define ARMV8_EVTYPE_EVENT 0x3ff /* Mask for EVENT bits */
+
+/*
+ * Event filters for PMUv3
+ */
+#define ARMV8_EXCLUDE_EL1 (1 << 31)
+#define ARMV8_EXCLUDE_EL0 (1 << 30)
+#define ARMV8_INCLUDE_EL2 (1 << 27)
+
#ifdef CONFIG_HW_PERF_EVENTS
/* The events for a given PMU register set. */
diff --git a/arch/arm64/kernel/perf_event.c b/arch/arm64/kernel/perf_event.c
index b31e9a4..96538d7 100644
--- a/arch/arm64/kernel/perf_event.c
+++ b/arch/arm64/kernel/perf_event.c
@@ -824,9 +824,6 @@ static const unsigned armv8_pmuv3_perf_cache_map[PERF_COUNT_HW_CACHE_MAX]
#define ARMV8_IDX_COUNTER0 1
#define ARMV8_IDX_COUNTER_LAST (ARMV8_IDX_CYCLE_COUNTER + cpu_pmu->num_events - 1)
-#define ARMV8_MAX_COUNTERS 32
-#define ARMV8_COUNTER_MASK (ARMV8_MAX_COUNTERS - 1)
-
/*
* ARMv8 low level PMU access
*/
@@ -837,38 +834,6 @@ static const unsigned armv8_pmuv3_perf_cache_map[PERF_COUNT_HW_CACHE_MAX]
#define ARMV8_IDX_TO_COUNTER(x) \
(((x) - ARMV8_IDX_COUNTER0) & ARMV8_COUNTER_MASK)
-/*
- * Per-CPU PMCR: config reg
- */
-#define ARMV8_PMCR_E (1 << 0) /* Enable all counters */
-#define ARMV8_PMCR_P (1 << 1) /* Reset all counters */
-#define ARMV8_PMCR_C (1 << 2) /* Cycle counter reset */
-#define ARMV8_PMCR_D (1 << 3) /* CCNT counts every 64th cpu cycle */
-#define ARMV8_PMCR_X (1 << 4) /* Export to ETM */
-#define ARMV8_PMCR_DP (1 << 5) /* Disable CCNT if non-invasive debug*/
-#define ARMV8_PMCR_N_SHIFT 11 /* Number of counters supported */
-#define ARMV8_PMCR_N_MASK 0x1f
-#define ARMV8_PMCR_MASK 0x3f /* Mask for writable bits */
-
-/*
- * PMOVSR: counters overflow flag status reg
- */
-#define ARMV8_OVSR_MASK 0xffffffff /* Mask for writable bits */
-#define ARMV8_OVERFLOWED_MASK ARMV8_OVSR_MASK
-
-/*
- * PMXEVTYPER: Event selection reg
- */
-#define ARMV8_EVTYPE_MASK 0xc80003ff /* Mask for writable bits */
-#define ARMV8_EVTYPE_EVENT 0x3ff /* Mask for EVENT bits */
-
-/*
- * Event filters for PMUv3
- */
-#define ARMV8_EXCLUDE_EL1 (1 << 31)
-#define ARMV8_EXCLUDE_EL0 (1 << 30)
-#define ARMV8_INCLUDE_EL2 (1 << 27)
-
static inline u32 armv8pmu_pmcr_read(void)
{
u32 val;
--
2.0.4
^ permalink raw reply related [flat|nested] 113+ messages in thread
* [PATCH v2 01/22] ARM64: Move PMU register related defines to asm/pmu.h
@ 2015-09-11 8:54 ` Shannon Zhao
0 siblings, 0 replies; 113+ messages in thread
From: Shannon Zhao @ 2015-09-11 8:54 UTC (permalink / raw)
To: kvmarm; +Cc: kvm, marc.zyngier, will.deacon, linux-arm-kernel, shannon.zhao
From: Shannon Zhao <shannon.zhao@linaro.org>
To use the ARMv8 PMU related register defines from the KVM code,
we move the relevant definitions to asm/pmu.h header file.
Signed-off-by: Anup Patel <anup.patel@linaro.org>
Signed-off-by: Shannon Zhao <shannon.zhao@linaro.org>
---
arch/arm64/include/asm/pmu.h | 45 ++++++++++++++++++++++++++++++++++++++++++
arch/arm64/kernel/perf_event.c | 35 --------------------------------
2 files changed, 45 insertions(+), 35 deletions(-)
diff --git a/arch/arm64/include/asm/pmu.h b/arch/arm64/include/asm/pmu.h
index b7710a5..b9f394a 100644
--- a/arch/arm64/include/asm/pmu.h
+++ b/arch/arm64/include/asm/pmu.h
@@ -19,6 +19,51 @@
#ifndef __ASM_PMU_H
#define __ASM_PMU_H
+#define ARMV8_MAX_COUNTERS 32
+#define ARMV8_COUNTER_MASK (ARMV8_MAX_COUNTERS - 1)
+
+/*
+ * Per-CPU PMCR: config reg
+ */
+#define ARMV8_PMCR_E (1 << 0) /* Enable all counters */
+#define ARMV8_PMCR_P (1 << 1) /* Reset all counters */
+#define ARMV8_PMCR_C (1 << 2) /* Cycle counter reset */
+#define ARMV8_PMCR_D (1 << 3) /* CCNT counts every 64th cpu cycle */
+#define ARMV8_PMCR_X (1 << 4) /* Export to ETM */
+#define ARMV8_PMCR_DP (1 << 5) /* Disable CCNT if non-invasive debug*/
+#define ARMV8_PMCR_N_SHIFT 11 /* Number of counters supported */
+#define ARMV8_PMCR_N_MASK 0x1f
+#define ARMV8_PMCR_MASK 0x3f /* Mask for writable bits */
+
+/*
+ * PMCNTEN: counters enable reg
+ */
+#define ARMV8_CNTEN_MASK 0xffffffff /* Mask for writable bits */
+
+/*
+ * PMINTEN: counters interrupt enable reg
+ */
+#define ARMV8_INTEN_MASK 0xffffffff /* Mask for writable bits */
+
+/*
+ * PMOVSR: counters overflow flag status reg
+ */
+#define ARMV8_OVSR_MASK 0xffffffff /* Mask for writable bits */
+#define ARMV8_OVERFLOWED_MASK ARMV8_OVSR_MASK
+
+/*
+ * PMXEVTYPER: Event selection reg
+ */
+#define ARMV8_EVTYPE_MASK 0xc80003ff /* Mask for writable bits */
+#define ARMV8_EVTYPE_EVENT 0x3ff /* Mask for EVENT bits */
+
+/*
+ * Event filters for PMUv3
+ */
+#define ARMV8_EXCLUDE_EL1 (1 << 31)
+#define ARMV8_EXCLUDE_EL0 (1 << 30)
+#define ARMV8_INCLUDE_EL2 (1 << 27)
+
#ifdef CONFIG_HW_PERF_EVENTS
/* The events for a given PMU register set. */
diff --git a/arch/arm64/kernel/perf_event.c b/arch/arm64/kernel/perf_event.c
index b31e9a4..96538d7 100644
--- a/arch/arm64/kernel/perf_event.c
+++ b/arch/arm64/kernel/perf_event.c
@@ -824,9 +824,6 @@ static const unsigned armv8_pmuv3_perf_cache_map[PERF_COUNT_HW_CACHE_MAX]
#define ARMV8_IDX_COUNTER0 1
#define ARMV8_IDX_COUNTER_LAST (ARMV8_IDX_CYCLE_COUNTER + cpu_pmu->num_events - 1)
-#define ARMV8_MAX_COUNTERS 32
-#define ARMV8_COUNTER_MASK (ARMV8_MAX_COUNTERS - 1)
-
/*
* ARMv8 low level PMU access
*/
@@ -837,38 +834,6 @@ static const unsigned armv8_pmuv3_perf_cache_map[PERF_COUNT_HW_CACHE_MAX]
#define ARMV8_IDX_TO_COUNTER(x) \
(((x) - ARMV8_IDX_COUNTER0) & ARMV8_COUNTER_MASK)
-/*
- * Per-CPU PMCR: config reg
- */
-#define ARMV8_PMCR_E (1 << 0) /* Enable all counters */
-#define ARMV8_PMCR_P (1 << 1) /* Reset all counters */
-#define ARMV8_PMCR_C (1 << 2) /* Cycle counter reset */
-#define ARMV8_PMCR_D (1 << 3) /* CCNT counts every 64th cpu cycle */
-#define ARMV8_PMCR_X (1 << 4) /* Export to ETM */
-#define ARMV8_PMCR_DP (1 << 5) /* Disable CCNT if non-invasive debug*/
-#define ARMV8_PMCR_N_SHIFT 11 /* Number of counters supported */
-#define ARMV8_PMCR_N_MASK 0x1f
-#define ARMV8_PMCR_MASK 0x3f /* Mask for writable bits */
-
-/*
- * PMOVSR: counters overflow flag status reg
- */
-#define ARMV8_OVSR_MASK 0xffffffff /* Mask for writable bits */
-#define ARMV8_OVERFLOWED_MASK ARMV8_OVSR_MASK
-
-/*
- * PMXEVTYPER: Event selection reg
- */
-#define ARMV8_EVTYPE_MASK 0xc80003ff /* Mask for writable bits */
-#define ARMV8_EVTYPE_EVENT 0x3ff /* Mask for EVENT bits */
-
-/*
- * Event filters for PMUv3
- */
-#define ARMV8_EXCLUDE_EL1 (1 << 31)
-#define ARMV8_EXCLUDE_EL0 (1 << 30)
-#define ARMV8_INCLUDE_EL2 (1 << 27)
-
static inline u32 armv8pmu_pmcr_read(void)
{
u32 val;
--
2.0.4
^ permalink raw reply related [flat|nested] 113+ messages in thread
* [PATCH v2 01/22] ARM64: Move PMU register related defines to asm/pmu.h
@ 2015-09-11 8:54 ` Shannon Zhao
0 siblings, 0 replies; 113+ messages in thread
From: Shannon Zhao @ 2015-09-11 8:54 UTC (permalink / raw)
To: kvmarm; +Cc: kvm, marc.zyngier, will.deacon, linux-arm-kernel, shannon.zhao
From: Shannon Zhao <shannon.zhao@linaro.org>
To use the ARMv8 PMU related register defines from the KVM code,
we move the relevant definitions to asm/pmu.h header file.
Signed-off-by: Anup Patel <anup.patel@linaro.org>
Signed-off-by: Shannon Zhao <shannon.zhao@linaro.org>
---
arch/arm64/include/asm/pmu.h | 45 ++++++++++++++++++++++++++++++++++++++++++
arch/arm64/kernel/perf_event.c | 35 --------------------------------
2 files changed, 45 insertions(+), 35 deletions(-)
diff --git a/arch/arm64/include/asm/pmu.h b/arch/arm64/include/asm/pmu.h
index b7710a5..b9f394a 100644
--- a/arch/arm64/include/asm/pmu.h
+++ b/arch/arm64/include/asm/pmu.h
@@ -19,6 +19,51 @@
#ifndef __ASM_PMU_H
#define __ASM_PMU_H
+#define ARMV8_MAX_COUNTERS 32
+#define ARMV8_COUNTER_MASK (ARMV8_MAX_COUNTERS - 1)
+
+/*
+ * Per-CPU PMCR: config reg
+ */
+#define ARMV8_PMCR_E (1 << 0) /* Enable all counters */
+#define ARMV8_PMCR_P (1 << 1) /* Reset all counters */
+#define ARMV8_PMCR_C (1 << 2) /* Cycle counter reset */
+#define ARMV8_PMCR_D (1 << 3) /* CCNT counts every 64th cpu cycle */
+#define ARMV8_PMCR_X (1 << 4) /* Export to ETM */
+#define ARMV8_PMCR_DP (1 << 5) /* Disable CCNT if non-invasive debug*/
+#define ARMV8_PMCR_N_SHIFT 11 /* Number of counters supported */
+#define ARMV8_PMCR_N_MASK 0x1f
+#define ARMV8_PMCR_MASK 0x3f /* Mask for writable bits */
+
+/*
+ * PMCNTEN: counters enable reg
+ */
+#define ARMV8_CNTEN_MASK 0xffffffff /* Mask for writable bits */
+
+/*
+ * PMINTEN: counters interrupt enable reg
+ */
+#define ARMV8_INTEN_MASK 0xffffffff /* Mask for writable bits */
+
+/*
+ * PMOVSR: counters overflow flag status reg
+ */
+#define ARMV8_OVSR_MASK 0xffffffff /* Mask for writable bits */
+#define ARMV8_OVERFLOWED_MASK ARMV8_OVSR_MASK
+
+/*
+ * PMXEVTYPER: Event selection reg
+ */
+#define ARMV8_EVTYPE_MASK 0xc80003ff /* Mask for writable bits */
+#define ARMV8_EVTYPE_EVENT 0x3ff /* Mask for EVENT bits */
+
+/*
+ * Event filters for PMUv3
+ */
+#define ARMV8_EXCLUDE_EL1 (1 << 31)
+#define ARMV8_EXCLUDE_EL0 (1 << 30)
+#define ARMV8_INCLUDE_EL2 (1 << 27)
+
#ifdef CONFIG_HW_PERF_EVENTS
/* The events for a given PMU register set. */
diff --git a/arch/arm64/kernel/perf_event.c b/arch/arm64/kernel/perf_event.c
index b31e9a4..96538d7 100644
--- a/arch/arm64/kernel/perf_event.c
+++ b/arch/arm64/kernel/perf_event.c
@@ -824,9 +824,6 @@ static const unsigned armv8_pmuv3_perf_cache_map[PERF_COUNT_HW_CACHE_MAX]
#define ARMV8_IDX_COUNTER0 1
#define ARMV8_IDX_COUNTER_LAST (ARMV8_IDX_CYCLE_COUNTER + cpu_pmu->num_events - 1)
-#define ARMV8_MAX_COUNTERS 32
-#define ARMV8_COUNTER_MASK (ARMV8_MAX_COUNTERS - 1)
-
/*
* ARMv8 low level PMU access
*/
@@ -837,38 +834,6 @@ static const unsigned armv8_pmuv3_perf_cache_map[PERF_COUNT_HW_CACHE_MAX]
#define ARMV8_IDX_TO_COUNTER(x) \
(((x) - ARMV8_IDX_COUNTER0) & ARMV8_COUNTER_MASK)
-/*
- * Per-CPU PMCR: config reg
- */
-#define ARMV8_PMCR_E (1 << 0) /* Enable all counters */
-#define ARMV8_PMCR_P (1 << 1) /* Reset all counters */
-#define ARMV8_PMCR_C (1 << 2) /* Cycle counter reset */
-#define ARMV8_PMCR_D (1 << 3) /* CCNT counts every 64th cpu cycle */
-#define ARMV8_PMCR_X (1 << 4) /* Export to ETM */
-#define ARMV8_PMCR_DP (1 << 5) /* Disable CCNT if non-invasive debug*/
-#define ARMV8_PMCR_N_SHIFT 11 /* Number of counters supported */
-#define ARMV8_PMCR_N_MASK 0x1f
-#define ARMV8_PMCR_MASK 0x3f /* Mask for writable bits */
-
-/*
- * PMOVSR: counters overflow flag status reg
- */
-#define ARMV8_OVSR_MASK 0xffffffff /* Mask for writable bits */
-#define ARMV8_OVERFLOWED_MASK ARMV8_OVSR_MASK
-
-/*
- * PMXEVTYPER: Event selection reg
- */
-#define ARMV8_EVTYPE_MASK 0xc80003ff /* Mask for writable bits */
-#define ARMV8_EVTYPE_EVENT 0x3ff /* Mask for EVENT bits */
-
-/*
- * Event filters for PMUv3
- */
-#define ARMV8_EXCLUDE_EL1 (1 << 31)
-#define ARMV8_EXCLUDE_EL0 (1 << 30)
-#define ARMV8_INCLUDE_EL2 (1 << 27)
-
static inline u32 armv8pmu_pmcr_read(void)
{
u32 val;
--
2.0.4
^ permalink raw reply related [flat|nested] 113+ messages in thread
* [PATCH v2 02/22] KVM: ARM64: Define PMU data structure for each vcpu
2015-09-11 8:54 ` Shannon Zhao
(?)
@ 2015-09-11 8:54 ` Shannon Zhao
-1 siblings, 0 replies; 113+ messages in thread
From: Shannon Zhao @ 2015-09-11 8:54 UTC (permalink / raw)
To: linux-arm-kernel
From: Shannon Zhao <shannon.zhao@linaro.org>
Here we plan to support virtual PMU for guest by full software
emulation, so define some basic structs and functions preparing for
futher steps. Define struct kvm_pmc for performance monitor counter and
struct kvm_pmu for performance monitor unit for each vcpu. According to
ARMv8 spec, the PMU contains at most 32(ARMV8_MAX_COUNTERS) counters.
Since this only supports ARM64 (or PMUv3), add a separate config symbol
for it.
Signed-off-by: Shannon Zhao <shannon.zhao@linaro.org>
---
arch/arm64/include/asm/kvm_host.h | 2 ++
arch/arm64/include/asm/pmu.h | 2 ++
arch/arm64/kvm/Kconfig | 8 ++++++++
include/kvm/arm_pmu.h | 39 +++++++++++++++++++++++++++++++++++++++
4 files changed, 51 insertions(+)
create mode 100644 include/kvm/arm_pmu.h
diff --git a/arch/arm64/include/asm/kvm_host.h b/arch/arm64/include/asm/kvm_host.h
index 2709db2..3c88873 100644
--- a/arch/arm64/include/asm/kvm_host.h
+++ b/arch/arm64/include/asm/kvm_host.h
@@ -42,6 +42,7 @@
#include <kvm/arm_vgic.h>
#include <kvm/arm_arch_timer.h>
+#include <kvm/arm_pmu.h>
#define KVM_VCPU_MAX_FEATURES 3
@@ -116,6 +117,7 @@ struct kvm_vcpu_arch {
/* VGIC state */
struct vgic_cpu vgic_cpu;
struct arch_timer_cpu timer_cpu;
+ struct kvm_pmu pmu;
/*
* Anything that is not used directly from assembly code goes
diff --git a/arch/arm64/include/asm/pmu.h b/arch/arm64/include/asm/pmu.h
index b9f394a..95681e6 100644
--- a/arch/arm64/include/asm/pmu.h
+++ b/arch/arm64/include/asm/pmu.h
@@ -19,6 +19,8 @@
#ifndef __ASM_PMU_H
#define __ASM_PMU_H
+#include <linux/perf_event.h>
+
#define ARMV8_MAX_COUNTERS 32
#define ARMV8_COUNTER_MASK (ARMV8_MAX_COUNTERS - 1)
diff --git a/arch/arm64/kvm/Kconfig b/arch/arm64/kvm/Kconfig
index bfffe8f..3c7c58d 100644
--- a/arch/arm64/kvm/Kconfig
+++ b/arch/arm64/kvm/Kconfig
@@ -31,6 +31,7 @@ config KVM
select KVM_VFIO
select HAVE_KVM_EVENTFD
select HAVE_KVM_IRQFD
+ select KVM_ARM_PMU
---help---
Support hosting virtualized guest machines.
@@ -52,4 +53,11 @@ config KVM_ARM_MAX_VCPUS
large, so only choose a reasonable number that you expect to
actually use.
+config KVM_ARM_PMU
+ bool
+ depends on KVM_ARM_HOST
+ ---help---
+ Adds support for a virtual Performance Monitoring Unit (PMU) in
+ virtual machines.
+
endif # VIRTUALIZATION
diff --git a/include/kvm/arm_pmu.h b/include/kvm/arm_pmu.h
new file mode 100644
index 0000000..64af88a
--- /dev/null
+++ b/include/kvm/arm_pmu.h
@@ -0,0 +1,39 @@
+/*
+ * Copyright (C) 2015 Linaro Ltd.
+ * Author: Shannon Zhao <shannon.zhao@linaro.org>
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program. If not, see <http://www.gnu.org/licenses/>.
+ */
+
+#ifndef __ASM_ARM_KVM_PMU_H
+#define __ASM_ARM_KVM_PMU_H
+
+#include <asm/pmu.h>
+
+struct kvm_pmc {
+ u8 idx;/* index into the pmu->pmc array */
+ struct perf_event *perf_event;
+ struct kvm_vcpu *vcpu;
+};
+
+struct kvm_pmu {
+#ifdef CONFIG_KVM_ARM_PMU
+ /* PMU IRQ Number per VCPU */
+ int irq_num;
+ /* IRQ pending flag */
+ bool irq_pending;
+ struct kvm_pmc pmc[ARMV8_MAX_COUNTERS];
+#endif
+};
+
+#endif
--
2.0.4
^ permalink raw reply related [flat|nested] 113+ messages in thread
* [PATCH v2 02/22] KVM: ARM64: Define PMU data structure for each vcpu
@ 2015-09-11 8:54 ` Shannon Zhao
0 siblings, 0 replies; 113+ messages in thread
From: Shannon Zhao @ 2015-09-11 8:54 UTC (permalink / raw)
To: kvmarm; +Cc: kvm, marc.zyngier, will.deacon, linux-arm-kernel, shannon.zhao
From: Shannon Zhao <shannon.zhao@linaro.org>
Here we plan to support virtual PMU for guest by full software
emulation, so define some basic structs and functions preparing for
futher steps. Define struct kvm_pmc for performance monitor counter and
struct kvm_pmu for performance monitor unit for each vcpu. According to
ARMv8 spec, the PMU contains at most 32(ARMV8_MAX_COUNTERS) counters.
Since this only supports ARM64 (or PMUv3), add a separate config symbol
for it.
Signed-off-by: Shannon Zhao <shannon.zhao@linaro.org>
---
arch/arm64/include/asm/kvm_host.h | 2 ++
arch/arm64/include/asm/pmu.h | 2 ++
arch/arm64/kvm/Kconfig | 8 ++++++++
include/kvm/arm_pmu.h | 39 +++++++++++++++++++++++++++++++++++++++
4 files changed, 51 insertions(+)
create mode 100644 include/kvm/arm_pmu.h
diff --git a/arch/arm64/include/asm/kvm_host.h b/arch/arm64/include/asm/kvm_host.h
index 2709db2..3c88873 100644
--- a/arch/arm64/include/asm/kvm_host.h
+++ b/arch/arm64/include/asm/kvm_host.h
@@ -42,6 +42,7 @@
#include <kvm/arm_vgic.h>
#include <kvm/arm_arch_timer.h>
+#include <kvm/arm_pmu.h>
#define KVM_VCPU_MAX_FEATURES 3
@@ -116,6 +117,7 @@ struct kvm_vcpu_arch {
/* VGIC state */
struct vgic_cpu vgic_cpu;
struct arch_timer_cpu timer_cpu;
+ struct kvm_pmu pmu;
/*
* Anything that is not used directly from assembly code goes
diff --git a/arch/arm64/include/asm/pmu.h b/arch/arm64/include/asm/pmu.h
index b9f394a..95681e6 100644
--- a/arch/arm64/include/asm/pmu.h
+++ b/arch/arm64/include/asm/pmu.h
@@ -19,6 +19,8 @@
#ifndef __ASM_PMU_H
#define __ASM_PMU_H
+#include <linux/perf_event.h>
+
#define ARMV8_MAX_COUNTERS 32
#define ARMV8_COUNTER_MASK (ARMV8_MAX_COUNTERS - 1)
diff --git a/arch/arm64/kvm/Kconfig b/arch/arm64/kvm/Kconfig
index bfffe8f..3c7c58d 100644
--- a/arch/arm64/kvm/Kconfig
+++ b/arch/arm64/kvm/Kconfig
@@ -31,6 +31,7 @@ config KVM
select KVM_VFIO
select HAVE_KVM_EVENTFD
select HAVE_KVM_IRQFD
+ select KVM_ARM_PMU
---help---
Support hosting virtualized guest machines.
@@ -52,4 +53,11 @@ config KVM_ARM_MAX_VCPUS
large, so only choose a reasonable number that you expect to
actually use.
+config KVM_ARM_PMU
+ bool
+ depends on KVM_ARM_HOST
+ ---help---
+ Adds support for a virtual Performance Monitoring Unit (PMU) in
+ virtual machines.
+
endif # VIRTUALIZATION
diff --git a/include/kvm/arm_pmu.h b/include/kvm/arm_pmu.h
new file mode 100644
index 0000000..64af88a
--- /dev/null
+++ b/include/kvm/arm_pmu.h
@@ -0,0 +1,39 @@
+/*
+ * Copyright (C) 2015 Linaro Ltd.
+ * Author: Shannon Zhao <shannon.zhao@linaro.org>
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program. If not, see <http://www.gnu.org/licenses/>.
+ */
+
+#ifndef __ASM_ARM_KVM_PMU_H
+#define __ASM_ARM_KVM_PMU_H
+
+#include <asm/pmu.h>
+
+struct kvm_pmc {
+ u8 idx;/* index into the pmu->pmc array */
+ struct perf_event *perf_event;
+ struct kvm_vcpu *vcpu;
+};
+
+struct kvm_pmu {
+#ifdef CONFIG_KVM_ARM_PMU
+ /* PMU IRQ Number per VCPU */
+ int irq_num;
+ /* IRQ pending flag */
+ bool irq_pending;
+ struct kvm_pmc pmc[ARMV8_MAX_COUNTERS];
+#endif
+};
+
+#endif
--
2.0.4
^ permalink raw reply related [flat|nested] 113+ messages in thread
* [PATCH v2 02/22] KVM: ARM64: Define PMU data structure for each vcpu
@ 2015-09-11 8:54 ` Shannon Zhao
0 siblings, 0 replies; 113+ messages in thread
From: Shannon Zhao @ 2015-09-11 8:54 UTC (permalink / raw)
To: kvmarm; +Cc: kvm, marc.zyngier, will.deacon, linux-arm-kernel, shannon.zhao
From: Shannon Zhao <shannon.zhao@linaro.org>
Here we plan to support virtual PMU for guest by full software
emulation, so define some basic structs and functions preparing for
futher steps. Define struct kvm_pmc for performance monitor counter and
struct kvm_pmu for performance monitor unit for each vcpu. According to
ARMv8 spec, the PMU contains at most 32(ARMV8_MAX_COUNTERS) counters.
Since this only supports ARM64 (or PMUv3), add a separate config symbol
for it.
Signed-off-by: Shannon Zhao <shannon.zhao@linaro.org>
---
arch/arm64/include/asm/kvm_host.h | 2 ++
arch/arm64/include/asm/pmu.h | 2 ++
arch/arm64/kvm/Kconfig | 8 ++++++++
include/kvm/arm_pmu.h | 39 +++++++++++++++++++++++++++++++++++++++
4 files changed, 51 insertions(+)
create mode 100644 include/kvm/arm_pmu.h
diff --git a/arch/arm64/include/asm/kvm_host.h b/arch/arm64/include/asm/kvm_host.h
index 2709db2..3c88873 100644
--- a/arch/arm64/include/asm/kvm_host.h
+++ b/arch/arm64/include/asm/kvm_host.h
@@ -42,6 +42,7 @@
#include <kvm/arm_vgic.h>
#include <kvm/arm_arch_timer.h>
+#include <kvm/arm_pmu.h>
#define KVM_VCPU_MAX_FEATURES 3
@@ -116,6 +117,7 @@ struct kvm_vcpu_arch {
/* VGIC state */
struct vgic_cpu vgic_cpu;
struct arch_timer_cpu timer_cpu;
+ struct kvm_pmu pmu;
/*
* Anything that is not used directly from assembly code goes
diff --git a/arch/arm64/include/asm/pmu.h b/arch/arm64/include/asm/pmu.h
index b9f394a..95681e6 100644
--- a/arch/arm64/include/asm/pmu.h
+++ b/arch/arm64/include/asm/pmu.h
@@ -19,6 +19,8 @@
#ifndef __ASM_PMU_H
#define __ASM_PMU_H
+#include <linux/perf_event.h>
+
#define ARMV8_MAX_COUNTERS 32
#define ARMV8_COUNTER_MASK (ARMV8_MAX_COUNTERS - 1)
diff --git a/arch/arm64/kvm/Kconfig b/arch/arm64/kvm/Kconfig
index bfffe8f..3c7c58d 100644
--- a/arch/arm64/kvm/Kconfig
+++ b/arch/arm64/kvm/Kconfig
@@ -31,6 +31,7 @@ config KVM
select KVM_VFIO
select HAVE_KVM_EVENTFD
select HAVE_KVM_IRQFD
+ select KVM_ARM_PMU
---help---
Support hosting virtualized guest machines.
@@ -52,4 +53,11 @@ config KVM_ARM_MAX_VCPUS
large, so only choose a reasonable number that you expect to
actually use.
+config KVM_ARM_PMU
+ bool
+ depends on KVM_ARM_HOST
+ ---help---
+ Adds support for a virtual Performance Monitoring Unit (PMU) in
+ virtual machines.
+
endif # VIRTUALIZATION
diff --git a/include/kvm/arm_pmu.h b/include/kvm/arm_pmu.h
new file mode 100644
index 0000000..64af88a
--- /dev/null
+++ b/include/kvm/arm_pmu.h
@@ -0,0 +1,39 @@
+/*
+ * Copyright (C) 2015 Linaro Ltd.
+ * Author: Shannon Zhao <shannon.zhao@linaro.org>
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program. If not, see <http://www.gnu.org/licenses/>.
+ */
+
+#ifndef __ASM_ARM_KVM_PMU_H
+#define __ASM_ARM_KVM_PMU_H
+
+#include <asm/pmu.h>
+
+struct kvm_pmc {
+ u8 idx;/* index into the pmu->pmc array */
+ struct perf_event *perf_event;
+ struct kvm_vcpu *vcpu;
+};
+
+struct kvm_pmu {
+#ifdef CONFIG_KVM_ARM_PMU
+ /* PMU IRQ Number per VCPU */
+ int irq_num;
+ /* IRQ pending flag */
+ bool irq_pending;
+ struct kvm_pmc pmc[ARMV8_MAX_COUNTERS];
+#endif
+};
+
+#endif
--
2.0.4
^ permalink raw reply related [flat|nested] 113+ messages in thread
* [PATCH v2 03/22] KVM: ARM64: Add offset defines for PMU registers
2015-09-11 8:54 ` Shannon Zhao
(?)
@ 2015-09-11 8:54 ` Shannon Zhao
-1 siblings, 0 replies; 113+ messages in thread
From: Shannon Zhao @ 2015-09-11 8:54 UTC (permalink / raw)
To: linux-arm-kernel
From: Shannon Zhao <shannon.zhao@linaro.org>
We are about to trap and emulate acccesses to each PMU register
individually. This adds the context offsets for the AArch64 PMU
registers and their AArch32 counterparts.
Signed-off-by: Shannon Zhao <shannon.zhao@linaro.org>
---
arch/arm64/include/asm/kvm_asm.h | 59 +++++++++++++++++++++++++++++++++++-----
1 file changed, 52 insertions(+), 7 deletions(-)
diff --git a/arch/arm64/include/asm/kvm_asm.h b/arch/arm64/include/asm/kvm_asm.h
index 3c5fe68..3a1df48 100644
--- a/arch/arm64/include/asm/kvm_asm.h
+++ b/arch/arm64/include/asm/kvm_asm.h
@@ -56,14 +56,36 @@
#define DBGWVR15_EL1 86
#define MDCCINT_EL1 87 /* Monitor Debug Comms Channel Interrupt Enable Reg */
+/* Performance Monitors Registers */
+#define PMCR_EL0 88 /* Control Register */
+#define PMOVSSET_EL0 89 /* Overflow Flag Status Set Register */
+#define PMOVSCLR_EL0 90 /* Overflow Flag Status Clear Register */
+#define PMSELR_EL0 91 /* Event Counter Selection Register */
+#define PMCEID0_EL0 92 /* Common Event Identification Register 0 */
+#define PMCEID1_EL0 93 /* Common Event Identification Register 1 */
+#define PMEVCNTR0_EL0 94 /* Event Counter Register (0-30) */
+#define PMEVCNTR30_EL0 124
+#define PMCCNTR_EL0 125 /* Cycle Counter Register */
+#define PMEVTYPER0_EL0 126 /* Event Type Register (0-30) */
+#define PMEVTYPER30_EL0 156
+#define PMCCFILTR_EL0 157 /* Cycle Count Filter Register */
+#define PMXEVCNTR_EL0 158 /* Selected Event Count Register */
+#define PMXEVTYPER_EL0 159 /* Selected Event Type Register */
+#define PMCNTENSET_EL0 160 /* Count Enable Set Register */
+#define PMCNTENCLR_EL0 161 /* Count Enable Clear Register */
+#define PMINTENSET_EL1 162 /* Interrupt Enable Set Register */
+#define PMINTENCLR_EL1 163 /* Interrupt Enable Clear Register */
+#define PMUSERENR_EL0 164 /* User Enable Register */
+#define PMSWINC_EL0 165 /* Software Increment Register */
+
/* 32bit specific registers. Keep them at the end of the range */
-#define DACR32_EL2 88 /* Domain Access Control Register */
-#define IFSR32_EL2 89 /* Instruction Fault Status Register */
-#define FPEXC32_EL2 90 /* Floating-Point Exception Control Register */
-#define DBGVCR32_EL2 91 /* Debug Vector Catch Register */
-#define TEECR32_EL1 92 /* ThumbEE Configuration Register */
-#define TEEHBR32_EL1 93 /* ThumbEE Handler Base Register */
-#define NR_SYS_REGS 94
+#define DACR32_EL2 166 /* Domain Access Control Register */
+#define IFSR32_EL2 167 /* Instruction Fault Status Register */
+#define FPEXC32_EL2 168 /* Floating-Point Exception Control Register */
+#define DBGVCR32_EL2 169 /* Debug Vector Catch Register */
+#define TEECR32_EL1 170 /* ThumbEE Configuration Register */
+#define TEEHBR32_EL1 171 /* ThumbEE Handler Base Register */
+#define NR_SYS_REGS 172
/* 32bit mapping */
#define c0_MPIDR (MPIDR_EL1 * 2) /* MultiProcessor ID Register */
@@ -85,6 +107,24 @@
#define c6_IFAR (c6_DFAR + 1) /* Instruction Fault Address Register */
#define c7_PAR (PAR_EL1 * 2) /* Physical Address Register */
#define c7_PAR_high (c7_PAR + 1) /* PAR top 32 bits */
+
+/* Performance Monitors*/
+#define c9_PMCR (PMCR_EL0 * 2)
+#define c9_PMOVSSET (PMOVSSET_EL0 * 2)
+#define c9_PMOVSCLR (PMOVSCLR_EL0 * 2)
+#define c9_PMCCNTR (PMCCNTR_EL0 * 2)
+#define c9_PMSELR (PMSELR_EL0 * 2)
+#define c9_PMCEID0 (PMCEID0_EL0 * 2)
+#define c9_PMCEID1 (PMCEID1_EL0 * 2)
+#define c9_PMXEVCNTR (PMXEVCNTR_EL0 * 2)
+#define c9_PMXEVTYPER (PMXEVTYPER_EL0 * 2)
+#define c9_PMCNTENSET (PMCNTENSET_EL0 * 2)
+#define c9_PMCNTENCLR (PMCNTENCLR_EL0 * 2)
+#define c9_PMINTENSET (PMINTENSET_EL1 * 2)
+#define c9_PMINTENCLR (PMINTENCLR_EL1 * 2)
+#define c9_PMUSERENR (PMUSERENR_EL0 * 2)
+#define c9_PMSWINC (PMSWINC_EL0 * 2)
+
#define c10_PRRR (MAIR_EL1 * 2) /* Primary Region Remap Register */
#define c10_NMRR (c10_PRRR + 1) /* Normal Memory Remap Register */
#define c12_VBAR (VBAR_EL1 * 2) /* Vector Base Address Register */
@@ -96,6 +136,11 @@
#define c10_AMAIR1 (c10_AMAIR0 + 1)/* Aux Memory Attr Indirection Reg */
#define c14_CNTKCTL (CNTKCTL_EL1 * 2) /* Timer Control Register (PL1) */
+/* Performance Monitors*/
+#define c14_PMEVCNTR0 (PMEVCNTR0_EL0 * 2)
+#define c14_PMEVTYPER0 (PMEVTYPER0_EL0 * 2)
+#define c14_PMCCFILTR (PMCCFILTR_EL0 * 2)
+
#define cp14_DBGDSCRext (MDSCR_EL1 * 2)
#define cp14_DBGBCR0 (DBGBCR0_EL1 * 2)
#define cp14_DBGBVR0 (DBGBVR0_EL1 * 2)
--
2.0.4
^ permalink raw reply related [flat|nested] 113+ messages in thread
* [PATCH v2 03/22] KVM: ARM64: Add offset defines for PMU registers
@ 2015-09-11 8:54 ` Shannon Zhao
0 siblings, 0 replies; 113+ messages in thread
From: Shannon Zhao @ 2015-09-11 8:54 UTC (permalink / raw)
To: kvmarm; +Cc: kvm, marc.zyngier, will.deacon, linux-arm-kernel, shannon.zhao
From: Shannon Zhao <shannon.zhao@linaro.org>
We are about to trap and emulate acccesses to each PMU register
individually. This adds the context offsets for the AArch64 PMU
registers and their AArch32 counterparts.
Signed-off-by: Shannon Zhao <shannon.zhao@linaro.org>
---
arch/arm64/include/asm/kvm_asm.h | 59 +++++++++++++++++++++++++++++++++++-----
1 file changed, 52 insertions(+), 7 deletions(-)
diff --git a/arch/arm64/include/asm/kvm_asm.h b/arch/arm64/include/asm/kvm_asm.h
index 3c5fe68..3a1df48 100644
--- a/arch/arm64/include/asm/kvm_asm.h
+++ b/arch/arm64/include/asm/kvm_asm.h
@@ -56,14 +56,36 @@
#define DBGWVR15_EL1 86
#define MDCCINT_EL1 87 /* Monitor Debug Comms Channel Interrupt Enable Reg */
+/* Performance Monitors Registers */
+#define PMCR_EL0 88 /* Control Register */
+#define PMOVSSET_EL0 89 /* Overflow Flag Status Set Register */
+#define PMOVSCLR_EL0 90 /* Overflow Flag Status Clear Register */
+#define PMSELR_EL0 91 /* Event Counter Selection Register */
+#define PMCEID0_EL0 92 /* Common Event Identification Register 0 */
+#define PMCEID1_EL0 93 /* Common Event Identification Register 1 */
+#define PMEVCNTR0_EL0 94 /* Event Counter Register (0-30) */
+#define PMEVCNTR30_EL0 124
+#define PMCCNTR_EL0 125 /* Cycle Counter Register */
+#define PMEVTYPER0_EL0 126 /* Event Type Register (0-30) */
+#define PMEVTYPER30_EL0 156
+#define PMCCFILTR_EL0 157 /* Cycle Count Filter Register */
+#define PMXEVCNTR_EL0 158 /* Selected Event Count Register */
+#define PMXEVTYPER_EL0 159 /* Selected Event Type Register */
+#define PMCNTENSET_EL0 160 /* Count Enable Set Register */
+#define PMCNTENCLR_EL0 161 /* Count Enable Clear Register */
+#define PMINTENSET_EL1 162 /* Interrupt Enable Set Register */
+#define PMINTENCLR_EL1 163 /* Interrupt Enable Clear Register */
+#define PMUSERENR_EL0 164 /* User Enable Register */
+#define PMSWINC_EL0 165 /* Software Increment Register */
+
/* 32bit specific registers. Keep them at the end of the range */
-#define DACR32_EL2 88 /* Domain Access Control Register */
-#define IFSR32_EL2 89 /* Instruction Fault Status Register */
-#define FPEXC32_EL2 90 /* Floating-Point Exception Control Register */
-#define DBGVCR32_EL2 91 /* Debug Vector Catch Register */
-#define TEECR32_EL1 92 /* ThumbEE Configuration Register */
-#define TEEHBR32_EL1 93 /* ThumbEE Handler Base Register */
-#define NR_SYS_REGS 94
+#define DACR32_EL2 166 /* Domain Access Control Register */
+#define IFSR32_EL2 167 /* Instruction Fault Status Register */
+#define FPEXC32_EL2 168 /* Floating-Point Exception Control Register */
+#define DBGVCR32_EL2 169 /* Debug Vector Catch Register */
+#define TEECR32_EL1 170 /* ThumbEE Configuration Register */
+#define TEEHBR32_EL1 171 /* ThumbEE Handler Base Register */
+#define NR_SYS_REGS 172
/* 32bit mapping */
#define c0_MPIDR (MPIDR_EL1 * 2) /* MultiProcessor ID Register */
@@ -85,6 +107,24 @@
#define c6_IFAR (c6_DFAR + 1) /* Instruction Fault Address Register */
#define c7_PAR (PAR_EL1 * 2) /* Physical Address Register */
#define c7_PAR_high (c7_PAR + 1) /* PAR top 32 bits */
+
+/* Performance Monitors*/
+#define c9_PMCR (PMCR_EL0 * 2)
+#define c9_PMOVSSET (PMOVSSET_EL0 * 2)
+#define c9_PMOVSCLR (PMOVSCLR_EL0 * 2)
+#define c9_PMCCNTR (PMCCNTR_EL0 * 2)
+#define c9_PMSELR (PMSELR_EL0 * 2)
+#define c9_PMCEID0 (PMCEID0_EL0 * 2)
+#define c9_PMCEID1 (PMCEID1_EL0 * 2)
+#define c9_PMXEVCNTR (PMXEVCNTR_EL0 * 2)
+#define c9_PMXEVTYPER (PMXEVTYPER_EL0 * 2)
+#define c9_PMCNTENSET (PMCNTENSET_EL0 * 2)
+#define c9_PMCNTENCLR (PMCNTENCLR_EL0 * 2)
+#define c9_PMINTENSET (PMINTENSET_EL1 * 2)
+#define c9_PMINTENCLR (PMINTENCLR_EL1 * 2)
+#define c9_PMUSERENR (PMUSERENR_EL0 * 2)
+#define c9_PMSWINC (PMSWINC_EL0 * 2)
+
#define c10_PRRR (MAIR_EL1 * 2) /* Primary Region Remap Register */
#define c10_NMRR (c10_PRRR + 1) /* Normal Memory Remap Register */
#define c12_VBAR (VBAR_EL1 * 2) /* Vector Base Address Register */
@@ -96,6 +136,11 @@
#define c10_AMAIR1 (c10_AMAIR0 + 1)/* Aux Memory Attr Indirection Reg */
#define c14_CNTKCTL (CNTKCTL_EL1 * 2) /* Timer Control Register (PL1) */
+/* Performance Monitors*/
+#define c14_PMEVCNTR0 (PMEVCNTR0_EL0 * 2)
+#define c14_PMEVTYPER0 (PMEVTYPER0_EL0 * 2)
+#define c14_PMCCFILTR (PMCCFILTR_EL0 * 2)
+
#define cp14_DBGDSCRext (MDSCR_EL1 * 2)
#define cp14_DBGBCR0 (DBGBCR0_EL1 * 2)
#define cp14_DBGBVR0 (DBGBVR0_EL1 * 2)
--
2.0.4
^ permalink raw reply related [flat|nested] 113+ messages in thread
* [PATCH v2 03/22] KVM: ARM64: Add offset defines for PMU registers
@ 2015-09-11 8:54 ` Shannon Zhao
0 siblings, 0 replies; 113+ messages in thread
From: Shannon Zhao @ 2015-09-11 8:54 UTC (permalink / raw)
To: kvmarm; +Cc: kvm, marc.zyngier, will.deacon, linux-arm-kernel, shannon.zhao
From: Shannon Zhao <shannon.zhao@linaro.org>
We are about to trap and emulate acccesses to each PMU register
individually. This adds the context offsets for the AArch64 PMU
registers and their AArch32 counterparts.
Signed-off-by: Shannon Zhao <shannon.zhao@linaro.org>
---
arch/arm64/include/asm/kvm_asm.h | 59 +++++++++++++++++++++++++++++++++++-----
1 file changed, 52 insertions(+), 7 deletions(-)
diff --git a/arch/arm64/include/asm/kvm_asm.h b/arch/arm64/include/asm/kvm_asm.h
index 3c5fe68..3a1df48 100644
--- a/arch/arm64/include/asm/kvm_asm.h
+++ b/arch/arm64/include/asm/kvm_asm.h
@@ -56,14 +56,36 @@
#define DBGWVR15_EL1 86
#define MDCCINT_EL1 87 /* Monitor Debug Comms Channel Interrupt Enable Reg */
+/* Performance Monitors Registers */
+#define PMCR_EL0 88 /* Control Register */
+#define PMOVSSET_EL0 89 /* Overflow Flag Status Set Register */
+#define PMOVSCLR_EL0 90 /* Overflow Flag Status Clear Register */
+#define PMSELR_EL0 91 /* Event Counter Selection Register */
+#define PMCEID0_EL0 92 /* Common Event Identification Register 0 */
+#define PMCEID1_EL0 93 /* Common Event Identification Register 1 */
+#define PMEVCNTR0_EL0 94 /* Event Counter Register (0-30) */
+#define PMEVCNTR30_EL0 124
+#define PMCCNTR_EL0 125 /* Cycle Counter Register */
+#define PMEVTYPER0_EL0 126 /* Event Type Register (0-30) */
+#define PMEVTYPER30_EL0 156
+#define PMCCFILTR_EL0 157 /* Cycle Count Filter Register */
+#define PMXEVCNTR_EL0 158 /* Selected Event Count Register */
+#define PMXEVTYPER_EL0 159 /* Selected Event Type Register */
+#define PMCNTENSET_EL0 160 /* Count Enable Set Register */
+#define PMCNTENCLR_EL0 161 /* Count Enable Clear Register */
+#define PMINTENSET_EL1 162 /* Interrupt Enable Set Register */
+#define PMINTENCLR_EL1 163 /* Interrupt Enable Clear Register */
+#define PMUSERENR_EL0 164 /* User Enable Register */
+#define PMSWINC_EL0 165 /* Software Increment Register */
+
/* 32bit specific registers. Keep them at the end of the range */
-#define DACR32_EL2 88 /* Domain Access Control Register */
-#define IFSR32_EL2 89 /* Instruction Fault Status Register */
-#define FPEXC32_EL2 90 /* Floating-Point Exception Control Register */
-#define DBGVCR32_EL2 91 /* Debug Vector Catch Register */
-#define TEECR32_EL1 92 /* ThumbEE Configuration Register */
-#define TEEHBR32_EL1 93 /* ThumbEE Handler Base Register */
-#define NR_SYS_REGS 94
+#define DACR32_EL2 166 /* Domain Access Control Register */
+#define IFSR32_EL2 167 /* Instruction Fault Status Register */
+#define FPEXC32_EL2 168 /* Floating-Point Exception Control Register */
+#define DBGVCR32_EL2 169 /* Debug Vector Catch Register */
+#define TEECR32_EL1 170 /* ThumbEE Configuration Register */
+#define TEEHBR32_EL1 171 /* ThumbEE Handler Base Register */
+#define NR_SYS_REGS 172
/* 32bit mapping */
#define c0_MPIDR (MPIDR_EL1 * 2) /* MultiProcessor ID Register */
@@ -85,6 +107,24 @@
#define c6_IFAR (c6_DFAR + 1) /* Instruction Fault Address Register */
#define c7_PAR (PAR_EL1 * 2) /* Physical Address Register */
#define c7_PAR_high (c7_PAR + 1) /* PAR top 32 bits */
+
+/* Performance Monitors*/
+#define c9_PMCR (PMCR_EL0 * 2)
+#define c9_PMOVSSET (PMOVSSET_EL0 * 2)
+#define c9_PMOVSCLR (PMOVSCLR_EL0 * 2)
+#define c9_PMCCNTR (PMCCNTR_EL0 * 2)
+#define c9_PMSELR (PMSELR_EL0 * 2)
+#define c9_PMCEID0 (PMCEID0_EL0 * 2)
+#define c9_PMCEID1 (PMCEID1_EL0 * 2)
+#define c9_PMXEVCNTR (PMXEVCNTR_EL0 * 2)
+#define c9_PMXEVTYPER (PMXEVTYPER_EL0 * 2)
+#define c9_PMCNTENSET (PMCNTENSET_EL0 * 2)
+#define c9_PMCNTENCLR (PMCNTENCLR_EL0 * 2)
+#define c9_PMINTENSET (PMINTENSET_EL1 * 2)
+#define c9_PMINTENCLR (PMINTENCLR_EL1 * 2)
+#define c9_PMUSERENR (PMUSERENR_EL0 * 2)
+#define c9_PMSWINC (PMSWINC_EL0 * 2)
+
#define c10_PRRR (MAIR_EL1 * 2) /* Primary Region Remap Register */
#define c10_NMRR (c10_PRRR + 1) /* Normal Memory Remap Register */
#define c12_VBAR (VBAR_EL1 * 2) /* Vector Base Address Register */
@@ -96,6 +136,11 @@
#define c10_AMAIR1 (c10_AMAIR0 + 1)/* Aux Memory Attr Indirection Reg */
#define c14_CNTKCTL (CNTKCTL_EL1 * 2) /* Timer Control Register (PL1) */
+/* Performance Monitors*/
+#define c14_PMEVCNTR0 (PMEVCNTR0_EL0 * 2)
+#define c14_PMEVTYPER0 (PMEVTYPER0_EL0 * 2)
+#define c14_PMCCFILTR (PMCCFILTR_EL0 * 2)
+
#define cp14_DBGDSCRext (MDSCR_EL1 * 2)
#define cp14_DBGBCR0 (DBGBCR0_EL1 * 2)
#define cp14_DBGBVR0 (DBGBVR0_EL1 * 2)
--
2.0.4
^ permalink raw reply related [flat|nested] 113+ messages in thread
* [PATCH v2 04/22] KVM: ARM64: Add reset and access handlers for PMCR_EL0 register
2015-09-11 8:54 ` Shannon Zhao
(?)
@ 2015-09-11 8:54 ` Shannon Zhao
-1 siblings, 0 replies; 113+ messages in thread
From: Shannon Zhao @ 2015-09-11 8:54 UTC (permalink / raw)
To: linux-arm-kernel
From: Shannon Zhao <shannon.zhao@linaro.org>
Add reset handler which gets host value of PMCR_EL0 and make writable
bits architecturally UNKNOWN. Add a common access handler for PMU
registers which emulates writing and reading register and add emulation
for PMCR.
Signed-off-by: Shannon Zhao <shannon.zhao@linaro.org>
---
arch/arm64/kvm/sys_regs.c | 76 +++++++++++++++++++++++++++++++++++++++++++++--
1 file changed, 74 insertions(+), 2 deletions(-)
diff --git a/arch/arm64/kvm/sys_regs.c b/arch/arm64/kvm/sys_regs.c
index c370b40..db1be44 100644
--- a/arch/arm64/kvm/sys_regs.c
+++ b/arch/arm64/kvm/sys_regs.c
@@ -33,6 +33,7 @@
#include <asm/kvm_emulate.h>
#include <asm/kvm_host.h>
#include <asm/kvm_mmu.h>
+#include <asm/pmu.h>
#include <trace/events/kvm.h>
@@ -236,6 +237,48 @@ static void reset_mpidr(struct kvm_vcpu *vcpu, const struct sys_reg_desc *r)
vcpu_sys_reg(vcpu, MPIDR_EL1) = (1ULL << 31) | mpidr;
}
+static void reset_pmcr(struct kvm_vcpu *vcpu, const struct sys_reg_desc *r)
+{
+ u32 pmcr;
+
+ asm volatile("mrs %0, pmcr_el0\n" : "=r" (pmcr));
+ /* Writable bits of PMCR_EL0 (ARMV8_PMCR_MASK) is reset to UNKNOWN*/
+ if (!vcpu_mode_is_32bit(vcpu))
+ vcpu_sys_reg(vcpu, r->reg) = (pmcr & ~ARMV8_PMCR_MASK)
+ | (ARMV8_PMCR_MASK & 0xdecafbad);
+ else
+ vcpu_cp15(vcpu, r->reg) = (pmcr & ~ARMV8_PMCR_MASK)
+ | (ARMV8_PMCR_MASK & 0xdecafbad);
+}
+
+/* PMU registers accessor. */
+static bool access_pmu_regs(struct kvm_vcpu *vcpu,
+ const struct sys_reg_params *p,
+ const struct sys_reg_desc *r)
+{
+ unsigned long val;
+
+ if (p->is_write) {
+ switch (r->reg) {
+ case PMCR_EL0: {
+ /* Only update writeable bits of PMCR */
+ val = vcpu_sys_reg(vcpu, r->reg);
+ val &= ~ARMV8_PMCR_MASK;
+ val |= *vcpu_reg(vcpu, p->Rt) & ARMV8_PMCR_MASK;
+ vcpu_sys_reg(vcpu, r->reg) = val;
+ break;
+ }
+ default:
+ vcpu_sys_reg(vcpu, r->reg) = *vcpu_reg(vcpu, p->Rt);
+ break;
+ }
+ } else {
+ *vcpu_reg(vcpu, p->Rt) = vcpu_sys_reg(vcpu, r->reg);
+ }
+
+ return true;
+}
+
/* Silly macro to expand the DBG{BCR,BVR,WVR,WCR}n_EL1 registers in one go */
#define DBG_BCR_BVR_WCR_WVR_EL1(n) \
/* DBGBVRn_EL1 */ \
@@ -427,7 +470,7 @@ static const struct sys_reg_desc sys_reg_descs[] = {
/* PMCR_EL0 */
{ Op0(0b11), Op1(0b011), CRn(0b1001), CRm(0b1100), Op2(0b000),
- trap_raz_wi },
+ access_pmu_regs, reset_pmcr, PMCR_EL0, },
/* PMCNTENSET_EL0 */
{ Op0(0b11), Op1(0b011), CRn(0b1001), CRm(0b1100), Op2(0b001),
trap_raz_wi },
@@ -632,6 +675,34 @@ static const struct sys_reg_desc cp14_64_regs[] = {
{ Op1( 0), CRm( 2), .access = trap_raz_wi },
};
+/* PMU CP15 registers accessor. */
+static bool access_pmu_cp15_regs(struct kvm_vcpu *vcpu,
+ const struct sys_reg_params *p,
+ const struct sys_reg_desc *r)
+{
+ unsigned long val;
+
+ if (p->is_write) {
+ switch (r->reg) {
+ case c9_PMCR: {
+ /* Only update writeable bits of PMCR */
+ val = vcpu_cp15(vcpu, r->reg);
+ val &= ~ARMV8_PMCR_MASK;
+ val |= *vcpu_reg(vcpu, p->Rt) & ARMV8_PMCR_MASK;
+ vcpu_cp15(vcpu, r->reg) = val;
+ break;
+ }
+ default:
+ vcpu_cp15(vcpu, r->reg) = *vcpu_reg(vcpu, p->Rt);
+ break;
+ }
+ } else {
+ *vcpu_reg(vcpu, p->Rt) = vcpu_cp15(vcpu, r->reg);
+ }
+
+ return true;
+}
+
/*
* Trapped cp15 registers. TTBR0/TTBR1 get a double encoding,
* depending on the way they are accessed (as a 32bit or a 64bit
@@ -660,7 +731,8 @@ static const struct sys_reg_desc cp15_regs[] = {
{ Op1( 0), CRn( 7), CRm(14), Op2( 2), access_dcsw },
/* PMU */
- { Op1( 0), CRn( 9), CRm(12), Op2( 0), trap_raz_wi },
+ { Op1( 0), CRn( 9), CRm(12), Op2( 0), access_pmu_cp15_regs,
+ reset_pmcr, c9_PMCR },
{ Op1( 0), CRn( 9), CRm(12), Op2( 1), trap_raz_wi },
{ Op1( 0), CRn( 9), CRm(12), Op2( 2), trap_raz_wi },
{ Op1( 0), CRn( 9), CRm(12), Op2( 3), trap_raz_wi },
--
2.0.4
^ permalink raw reply related [flat|nested] 113+ messages in thread
* [PATCH v2 04/22] KVM: ARM64: Add reset and access handlers for PMCR_EL0 register
@ 2015-09-11 8:54 ` Shannon Zhao
0 siblings, 0 replies; 113+ messages in thread
From: Shannon Zhao @ 2015-09-11 8:54 UTC (permalink / raw)
To: kvmarm; +Cc: kvm, marc.zyngier, will.deacon, linux-arm-kernel, shannon.zhao
From: Shannon Zhao <shannon.zhao@linaro.org>
Add reset handler which gets host value of PMCR_EL0 and make writable
bits architecturally UNKNOWN. Add a common access handler for PMU
registers which emulates writing and reading register and add emulation
for PMCR.
Signed-off-by: Shannon Zhao <shannon.zhao@linaro.org>
---
arch/arm64/kvm/sys_regs.c | 76 +++++++++++++++++++++++++++++++++++++++++++++--
1 file changed, 74 insertions(+), 2 deletions(-)
diff --git a/arch/arm64/kvm/sys_regs.c b/arch/arm64/kvm/sys_regs.c
index c370b40..db1be44 100644
--- a/arch/arm64/kvm/sys_regs.c
+++ b/arch/arm64/kvm/sys_regs.c
@@ -33,6 +33,7 @@
#include <asm/kvm_emulate.h>
#include <asm/kvm_host.h>
#include <asm/kvm_mmu.h>
+#include <asm/pmu.h>
#include <trace/events/kvm.h>
@@ -236,6 +237,48 @@ static void reset_mpidr(struct kvm_vcpu *vcpu, const struct sys_reg_desc *r)
vcpu_sys_reg(vcpu, MPIDR_EL1) = (1ULL << 31) | mpidr;
}
+static void reset_pmcr(struct kvm_vcpu *vcpu, const struct sys_reg_desc *r)
+{
+ u32 pmcr;
+
+ asm volatile("mrs %0, pmcr_el0\n" : "=r" (pmcr));
+ /* Writable bits of PMCR_EL0 (ARMV8_PMCR_MASK) is reset to UNKNOWN*/
+ if (!vcpu_mode_is_32bit(vcpu))
+ vcpu_sys_reg(vcpu, r->reg) = (pmcr & ~ARMV8_PMCR_MASK)
+ | (ARMV8_PMCR_MASK & 0xdecafbad);
+ else
+ vcpu_cp15(vcpu, r->reg) = (pmcr & ~ARMV8_PMCR_MASK)
+ | (ARMV8_PMCR_MASK & 0xdecafbad);
+}
+
+/* PMU registers accessor. */
+static bool access_pmu_regs(struct kvm_vcpu *vcpu,
+ const struct sys_reg_params *p,
+ const struct sys_reg_desc *r)
+{
+ unsigned long val;
+
+ if (p->is_write) {
+ switch (r->reg) {
+ case PMCR_EL0: {
+ /* Only update writeable bits of PMCR */
+ val = vcpu_sys_reg(vcpu, r->reg);
+ val &= ~ARMV8_PMCR_MASK;
+ val |= *vcpu_reg(vcpu, p->Rt) & ARMV8_PMCR_MASK;
+ vcpu_sys_reg(vcpu, r->reg) = val;
+ break;
+ }
+ default:
+ vcpu_sys_reg(vcpu, r->reg) = *vcpu_reg(vcpu, p->Rt);
+ break;
+ }
+ } else {
+ *vcpu_reg(vcpu, p->Rt) = vcpu_sys_reg(vcpu, r->reg);
+ }
+
+ return true;
+}
+
/* Silly macro to expand the DBG{BCR,BVR,WVR,WCR}n_EL1 registers in one go */
#define DBG_BCR_BVR_WCR_WVR_EL1(n) \
/* DBGBVRn_EL1 */ \
@@ -427,7 +470,7 @@ static const struct sys_reg_desc sys_reg_descs[] = {
/* PMCR_EL0 */
{ Op0(0b11), Op1(0b011), CRn(0b1001), CRm(0b1100), Op2(0b000),
- trap_raz_wi },
+ access_pmu_regs, reset_pmcr, PMCR_EL0, },
/* PMCNTENSET_EL0 */
{ Op0(0b11), Op1(0b011), CRn(0b1001), CRm(0b1100), Op2(0b001),
trap_raz_wi },
@@ -632,6 +675,34 @@ static const struct sys_reg_desc cp14_64_regs[] = {
{ Op1( 0), CRm( 2), .access = trap_raz_wi },
};
+/* PMU CP15 registers accessor. */
+static bool access_pmu_cp15_regs(struct kvm_vcpu *vcpu,
+ const struct sys_reg_params *p,
+ const struct sys_reg_desc *r)
+{
+ unsigned long val;
+
+ if (p->is_write) {
+ switch (r->reg) {
+ case c9_PMCR: {
+ /* Only update writeable bits of PMCR */
+ val = vcpu_cp15(vcpu, r->reg);
+ val &= ~ARMV8_PMCR_MASK;
+ val |= *vcpu_reg(vcpu, p->Rt) & ARMV8_PMCR_MASK;
+ vcpu_cp15(vcpu, r->reg) = val;
+ break;
+ }
+ default:
+ vcpu_cp15(vcpu, r->reg) = *vcpu_reg(vcpu, p->Rt);
+ break;
+ }
+ } else {
+ *vcpu_reg(vcpu, p->Rt) = vcpu_cp15(vcpu, r->reg);
+ }
+
+ return true;
+}
+
/*
* Trapped cp15 registers. TTBR0/TTBR1 get a double encoding,
* depending on the way they are accessed (as a 32bit or a 64bit
@@ -660,7 +731,8 @@ static const struct sys_reg_desc cp15_regs[] = {
{ Op1( 0), CRn( 7), CRm(14), Op2( 2), access_dcsw },
/* PMU */
- { Op1( 0), CRn( 9), CRm(12), Op2( 0), trap_raz_wi },
+ { Op1( 0), CRn( 9), CRm(12), Op2( 0), access_pmu_cp15_regs,
+ reset_pmcr, c9_PMCR },
{ Op1( 0), CRn( 9), CRm(12), Op2( 1), trap_raz_wi },
{ Op1( 0), CRn( 9), CRm(12), Op2( 2), trap_raz_wi },
{ Op1( 0), CRn( 9), CRm(12), Op2( 3), trap_raz_wi },
--
2.0.4
^ permalink raw reply related [flat|nested] 113+ messages in thread
* [PATCH v2 04/22] KVM: ARM64: Add reset and access handlers for PMCR_EL0 register
@ 2015-09-11 8:54 ` Shannon Zhao
0 siblings, 0 replies; 113+ messages in thread
From: Shannon Zhao @ 2015-09-11 8:54 UTC (permalink / raw)
To: kvmarm; +Cc: kvm, marc.zyngier, will.deacon, linux-arm-kernel, shannon.zhao
From: Shannon Zhao <shannon.zhao@linaro.org>
Add reset handler which gets host value of PMCR_EL0 and make writable
bits architecturally UNKNOWN. Add a common access handler for PMU
registers which emulates writing and reading register and add emulation
for PMCR.
Signed-off-by: Shannon Zhao <shannon.zhao@linaro.org>
---
arch/arm64/kvm/sys_regs.c | 76 +++++++++++++++++++++++++++++++++++++++++++++--
1 file changed, 74 insertions(+), 2 deletions(-)
diff --git a/arch/arm64/kvm/sys_regs.c b/arch/arm64/kvm/sys_regs.c
index c370b40..db1be44 100644
--- a/arch/arm64/kvm/sys_regs.c
+++ b/arch/arm64/kvm/sys_regs.c
@@ -33,6 +33,7 @@
#include <asm/kvm_emulate.h>
#include <asm/kvm_host.h>
#include <asm/kvm_mmu.h>
+#include <asm/pmu.h>
#include <trace/events/kvm.h>
@@ -236,6 +237,48 @@ static void reset_mpidr(struct kvm_vcpu *vcpu, const struct sys_reg_desc *r)
vcpu_sys_reg(vcpu, MPIDR_EL1) = (1ULL << 31) | mpidr;
}
+static void reset_pmcr(struct kvm_vcpu *vcpu, const struct sys_reg_desc *r)
+{
+ u32 pmcr;
+
+ asm volatile("mrs %0, pmcr_el0\n" : "=r" (pmcr));
+ /* Writable bits of PMCR_EL0 (ARMV8_PMCR_MASK) is reset to UNKNOWN*/
+ if (!vcpu_mode_is_32bit(vcpu))
+ vcpu_sys_reg(vcpu, r->reg) = (pmcr & ~ARMV8_PMCR_MASK)
+ | (ARMV8_PMCR_MASK & 0xdecafbad);
+ else
+ vcpu_cp15(vcpu, r->reg) = (pmcr & ~ARMV8_PMCR_MASK)
+ | (ARMV8_PMCR_MASK & 0xdecafbad);
+}
+
+/* PMU registers accessor. */
+static bool access_pmu_regs(struct kvm_vcpu *vcpu,
+ const struct sys_reg_params *p,
+ const struct sys_reg_desc *r)
+{
+ unsigned long val;
+
+ if (p->is_write) {
+ switch (r->reg) {
+ case PMCR_EL0: {
+ /* Only update writeable bits of PMCR */
+ val = vcpu_sys_reg(vcpu, r->reg);
+ val &= ~ARMV8_PMCR_MASK;
+ val |= *vcpu_reg(vcpu, p->Rt) & ARMV8_PMCR_MASK;
+ vcpu_sys_reg(vcpu, r->reg) = val;
+ break;
+ }
+ default:
+ vcpu_sys_reg(vcpu, r->reg) = *vcpu_reg(vcpu, p->Rt);
+ break;
+ }
+ } else {
+ *vcpu_reg(vcpu, p->Rt) = vcpu_sys_reg(vcpu, r->reg);
+ }
+
+ return true;
+}
+
/* Silly macro to expand the DBG{BCR,BVR,WVR,WCR}n_EL1 registers in one go */
#define DBG_BCR_BVR_WCR_WVR_EL1(n) \
/* DBGBVRn_EL1 */ \
@@ -427,7 +470,7 @@ static const struct sys_reg_desc sys_reg_descs[] = {
/* PMCR_EL0 */
{ Op0(0b11), Op1(0b011), CRn(0b1001), CRm(0b1100), Op2(0b000),
- trap_raz_wi },
+ access_pmu_regs, reset_pmcr, PMCR_EL0, },
/* PMCNTENSET_EL0 */
{ Op0(0b11), Op1(0b011), CRn(0b1001), CRm(0b1100), Op2(0b001),
trap_raz_wi },
@@ -632,6 +675,34 @@ static const struct sys_reg_desc cp14_64_regs[] = {
{ Op1( 0), CRm( 2), .access = trap_raz_wi },
};
+/* PMU CP15 registers accessor. */
+static bool access_pmu_cp15_regs(struct kvm_vcpu *vcpu,
+ const struct sys_reg_params *p,
+ const struct sys_reg_desc *r)
+{
+ unsigned long val;
+
+ if (p->is_write) {
+ switch (r->reg) {
+ case c9_PMCR: {
+ /* Only update writeable bits of PMCR */
+ val = vcpu_cp15(vcpu, r->reg);
+ val &= ~ARMV8_PMCR_MASK;
+ val |= *vcpu_reg(vcpu, p->Rt) & ARMV8_PMCR_MASK;
+ vcpu_cp15(vcpu, r->reg) = val;
+ break;
+ }
+ default:
+ vcpu_cp15(vcpu, r->reg) = *vcpu_reg(vcpu, p->Rt);
+ break;
+ }
+ } else {
+ *vcpu_reg(vcpu, p->Rt) = vcpu_cp15(vcpu, r->reg);
+ }
+
+ return true;
+}
+
/*
* Trapped cp15 registers. TTBR0/TTBR1 get a double encoding,
* depending on the way they are accessed (as a 32bit or a 64bit
@@ -660,7 +731,8 @@ static const struct sys_reg_desc cp15_regs[] = {
{ Op1( 0), CRn( 7), CRm(14), Op2( 2), access_dcsw },
/* PMU */
- { Op1( 0), CRn( 9), CRm(12), Op2( 0), trap_raz_wi },
+ { Op1( 0), CRn( 9), CRm(12), Op2( 0), access_pmu_cp15_regs,
+ reset_pmcr, c9_PMCR },
{ Op1( 0), CRn( 9), CRm(12), Op2( 1), trap_raz_wi },
{ Op1( 0), CRn( 9), CRm(12), Op2( 2), trap_raz_wi },
{ Op1( 0), CRn( 9), CRm(12), Op2( 3), trap_raz_wi },
--
2.0.4
^ permalink raw reply related [flat|nested] 113+ messages in thread
* [PATCH v2 05/22] KVM: ARM64: Add a helper for CP15 registers reset to UNKNOWN
2015-09-11 8:54 ` Shannon Zhao
(?)
@ 2015-09-11 8:54 ` Shannon Zhao
-1 siblings, 0 replies; 113+ messages in thread
From: Shannon Zhao @ 2015-09-11 8:54 UTC (permalink / raw)
To: linux-arm-kernel
From: Shannon Zhao <shannon.zhao@linaro.org>
Signed-off-by: Shannon Zhao <shannon.zhao@linaro.org>
---
arch/arm64/kvm/sys_regs.h | 8 ++++++++
1 file changed, 8 insertions(+)
diff --git a/arch/arm64/kvm/sys_regs.h b/arch/arm64/kvm/sys_regs.h
index d411e25..a0b3811 100644
--- a/arch/arm64/kvm/sys_regs.h
+++ b/arch/arm64/kvm/sys_regs.h
@@ -104,6 +104,14 @@ static inline void reset_unknown(struct kvm_vcpu *vcpu,
vcpu_sys_reg(vcpu, r->reg) = 0x1de7ec7edbadc0deULL;
}
+static inline void reset_unknown_cp15(struct kvm_vcpu *vcpu,
+ const struct sys_reg_desc *r)
+{
+ BUG_ON(!r->reg);
+ BUG_ON(r->reg >= NR_COPRO_REGS);
+ vcpu_cp15(vcpu, r->reg) = 0xdecafbad;
+}
+
static inline void reset_val(struct kvm_vcpu *vcpu, const struct sys_reg_desc *r)
{
BUG_ON(!r->reg);
--
2.0.4
^ permalink raw reply related [flat|nested] 113+ messages in thread
* [PATCH v2 05/22] KVM: ARM64: Add a helper for CP15 registers reset to UNKNOWN
@ 2015-09-11 8:54 ` Shannon Zhao
0 siblings, 0 replies; 113+ messages in thread
From: Shannon Zhao @ 2015-09-11 8:54 UTC (permalink / raw)
To: kvmarm; +Cc: kvm, marc.zyngier, will.deacon, linux-arm-kernel, shannon.zhao
From: Shannon Zhao <shannon.zhao@linaro.org>
Signed-off-by: Shannon Zhao <shannon.zhao@linaro.org>
---
arch/arm64/kvm/sys_regs.h | 8 ++++++++
1 file changed, 8 insertions(+)
diff --git a/arch/arm64/kvm/sys_regs.h b/arch/arm64/kvm/sys_regs.h
index d411e25..a0b3811 100644
--- a/arch/arm64/kvm/sys_regs.h
+++ b/arch/arm64/kvm/sys_regs.h
@@ -104,6 +104,14 @@ static inline void reset_unknown(struct kvm_vcpu *vcpu,
vcpu_sys_reg(vcpu, r->reg) = 0x1de7ec7edbadc0deULL;
}
+static inline void reset_unknown_cp15(struct kvm_vcpu *vcpu,
+ const struct sys_reg_desc *r)
+{
+ BUG_ON(!r->reg);
+ BUG_ON(r->reg >= NR_COPRO_REGS);
+ vcpu_cp15(vcpu, r->reg) = 0xdecafbad;
+}
+
static inline void reset_val(struct kvm_vcpu *vcpu, const struct sys_reg_desc *r)
{
BUG_ON(!r->reg);
--
2.0.4
^ permalink raw reply related [flat|nested] 113+ messages in thread
* [PATCH v2 05/22] KVM: ARM64: Add a helper for CP15 registers reset to UNKNOWN
@ 2015-09-11 8:54 ` Shannon Zhao
0 siblings, 0 replies; 113+ messages in thread
From: Shannon Zhao @ 2015-09-11 8:54 UTC (permalink / raw)
To: kvmarm; +Cc: kvm, marc.zyngier, will.deacon, linux-arm-kernel, shannon.zhao
From: Shannon Zhao <shannon.zhao@linaro.org>
Signed-off-by: Shannon Zhao <shannon.zhao@linaro.org>
---
arch/arm64/kvm/sys_regs.h | 8 ++++++++
1 file changed, 8 insertions(+)
diff --git a/arch/arm64/kvm/sys_regs.h b/arch/arm64/kvm/sys_regs.h
index d411e25..a0b3811 100644
--- a/arch/arm64/kvm/sys_regs.h
+++ b/arch/arm64/kvm/sys_regs.h
@@ -104,6 +104,14 @@ static inline void reset_unknown(struct kvm_vcpu *vcpu,
vcpu_sys_reg(vcpu, r->reg) = 0x1de7ec7edbadc0deULL;
}
+static inline void reset_unknown_cp15(struct kvm_vcpu *vcpu,
+ const struct sys_reg_desc *r)
+{
+ BUG_ON(!r->reg);
+ BUG_ON(r->reg >= NR_COPRO_REGS);
+ vcpu_cp15(vcpu, r->reg) = 0xdecafbad;
+}
+
static inline void reset_val(struct kvm_vcpu *vcpu, const struct sys_reg_desc *r)
{
BUG_ON(!r->reg);
--
2.0.4
^ permalink raw reply related [flat|nested] 113+ messages in thread
* [PATCH v2 06/22] KVM: ARM64: Add reset and access handlers for PMSELR register
2015-09-11 8:54 ` Shannon Zhao
(?)
@ 2015-09-11 8:54 ` Shannon Zhao
-1 siblings, 0 replies; 113+ messages in thread
From: Shannon Zhao @ 2015-09-11 8:54 UTC (permalink / raw)
To: linux-arm-kernel
From: Shannon Zhao <shannon.zhao@linaro.org>
Since the reset value of PMSELR_EL0 is UNKNOWN, use reset_unknown for
its reset handler. As it doesn't need to deal with the acsessing action
specially, it uses default case to emulate writing and reading PMSELR
register.
Signed-off-by: Shannon Zhao <shannon.zhao@linaro.org>
---
arch/arm64/kvm/sys_regs.c | 5 +++--
1 file changed, 3 insertions(+), 2 deletions(-)
diff --git a/arch/arm64/kvm/sys_regs.c b/arch/arm64/kvm/sys_regs.c
index db1be44..24b8972 100644
--- a/arch/arm64/kvm/sys_regs.c
+++ b/arch/arm64/kvm/sys_regs.c
@@ -485,7 +485,7 @@ static const struct sys_reg_desc sys_reg_descs[] = {
trap_raz_wi },
/* PMSELR_EL0 */
{ Op0(0b11), Op1(0b011), CRn(0b1001), CRm(0b1100), Op2(0b101),
- trap_raz_wi },
+ access_pmu_regs, reset_unknown, PMSELR_EL0 },
/* PMCEID0_EL0 */
{ Op0(0b11), Op1(0b011), CRn(0b1001), CRm(0b1100), Op2(0b110),
trap_raz_wi },
@@ -736,7 +736,8 @@ static const struct sys_reg_desc cp15_regs[] = {
{ Op1( 0), CRn( 9), CRm(12), Op2( 1), trap_raz_wi },
{ Op1( 0), CRn( 9), CRm(12), Op2( 2), trap_raz_wi },
{ Op1( 0), CRn( 9), CRm(12), Op2( 3), trap_raz_wi },
- { Op1( 0), CRn( 9), CRm(12), Op2( 5), trap_raz_wi },
+ { Op1( 0), CRn( 9), CRm(12), Op2( 5), access_pmu_cp15_regs,
+ reset_unknown_cp15, c9_PMSELR },
{ Op1( 0), CRn( 9), CRm(12), Op2( 6), trap_raz_wi },
{ Op1( 0), CRn( 9), CRm(12), Op2( 7), trap_raz_wi },
{ Op1( 0), CRn( 9), CRm(13), Op2( 0), trap_raz_wi },
--
2.0.4
^ permalink raw reply related [flat|nested] 113+ messages in thread
* [PATCH v2 06/22] KVM: ARM64: Add reset and access handlers for PMSELR register
@ 2015-09-11 8:54 ` Shannon Zhao
0 siblings, 0 replies; 113+ messages in thread
From: Shannon Zhao @ 2015-09-11 8:54 UTC (permalink / raw)
To: kvmarm; +Cc: kvm, marc.zyngier, will.deacon, linux-arm-kernel, shannon.zhao
From: Shannon Zhao <shannon.zhao@linaro.org>
Since the reset value of PMSELR_EL0 is UNKNOWN, use reset_unknown for
its reset handler. As it doesn't need to deal with the acsessing action
specially, it uses default case to emulate writing and reading PMSELR
register.
Signed-off-by: Shannon Zhao <shannon.zhao@linaro.org>
---
arch/arm64/kvm/sys_regs.c | 5 +++--
1 file changed, 3 insertions(+), 2 deletions(-)
diff --git a/arch/arm64/kvm/sys_regs.c b/arch/arm64/kvm/sys_regs.c
index db1be44..24b8972 100644
--- a/arch/arm64/kvm/sys_regs.c
+++ b/arch/arm64/kvm/sys_regs.c
@@ -485,7 +485,7 @@ static const struct sys_reg_desc sys_reg_descs[] = {
trap_raz_wi },
/* PMSELR_EL0 */
{ Op0(0b11), Op1(0b011), CRn(0b1001), CRm(0b1100), Op2(0b101),
- trap_raz_wi },
+ access_pmu_regs, reset_unknown, PMSELR_EL0 },
/* PMCEID0_EL0 */
{ Op0(0b11), Op1(0b011), CRn(0b1001), CRm(0b1100), Op2(0b110),
trap_raz_wi },
@@ -736,7 +736,8 @@ static const struct sys_reg_desc cp15_regs[] = {
{ Op1( 0), CRn( 9), CRm(12), Op2( 1), trap_raz_wi },
{ Op1( 0), CRn( 9), CRm(12), Op2( 2), trap_raz_wi },
{ Op1( 0), CRn( 9), CRm(12), Op2( 3), trap_raz_wi },
- { Op1( 0), CRn( 9), CRm(12), Op2( 5), trap_raz_wi },
+ { Op1( 0), CRn( 9), CRm(12), Op2( 5), access_pmu_cp15_regs,
+ reset_unknown_cp15, c9_PMSELR },
{ Op1( 0), CRn( 9), CRm(12), Op2( 6), trap_raz_wi },
{ Op1( 0), CRn( 9), CRm(12), Op2( 7), trap_raz_wi },
{ Op1( 0), CRn( 9), CRm(13), Op2( 0), trap_raz_wi },
--
2.0.4
^ permalink raw reply related [flat|nested] 113+ messages in thread
* [PATCH v2 06/22] KVM: ARM64: Add reset and access handlers for PMSELR register
@ 2015-09-11 8:54 ` Shannon Zhao
0 siblings, 0 replies; 113+ messages in thread
From: Shannon Zhao @ 2015-09-11 8:54 UTC (permalink / raw)
To: kvmarm; +Cc: kvm, marc.zyngier, will.deacon, linux-arm-kernel, shannon.zhao
From: Shannon Zhao <shannon.zhao@linaro.org>
Since the reset value of PMSELR_EL0 is UNKNOWN, use reset_unknown for
its reset handler. As it doesn't need to deal with the acsessing action
specially, it uses default case to emulate writing and reading PMSELR
register.
Signed-off-by: Shannon Zhao <shannon.zhao@linaro.org>
---
arch/arm64/kvm/sys_regs.c | 5 +++--
1 file changed, 3 insertions(+), 2 deletions(-)
diff --git a/arch/arm64/kvm/sys_regs.c b/arch/arm64/kvm/sys_regs.c
index db1be44..24b8972 100644
--- a/arch/arm64/kvm/sys_regs.c
+++ b/arch/arm64/kvm/sys_regs.c
@@ -485,7 +485,7 @@ static const struct sys_reg_desc sys_reg_descs[] = {
trap_raz_wi },
/* PMSELR_EL0 */
{ Op0(0b11), Op1(0b011), CRn(0b1001), CRm(0b1100), Op2(0b101),
- trap_raz_wi },
+ access_pmu_regs, reset_unknown, PMSELR_EL0 },
/* PMCEID0_EL0 */
{ Op0(0b11), Op1(0b011), CRn(0b1001), CRm(0b1100), Op2(0b110),
trap_raz_wi },
@@ -736,7 +736,8 @@ static const struct sys_reg_desc cp15_regs[] = {
{ Op1( 0), CRn( 9), CRm(12), Op2( 1), trap_raz_wi },
{ Op1( 0), CRn( 9), CRm(12), Op2( 2), trap_raz_wi },
{ Op1( 0), CRn( 9), CRm(12), Op2( 3), trap_raz_wi },
- { Op1( 0), CRn( 9), CRm(12), Op2( 5), trap_raz_wi },
+ { Op1( 0), CRn( 9), CRm(12), Op2( 5), access_pmu_cp15_regs,
+ reset_unknown_cp15, c9_PMSELR },
{ Op1( 0), CRn( 9), CRm(12), Op2( 6), trap_raz_wi },
{ Op1( 0), CRn( 9), CRm(12), Op2( 7), trap_raz_wi },
{ Op1( 0), CRn( 9), CRm(13), Op2( 0), trap_raz_wi },
--
2.0.4
^ permalink raw reply related [flat|nested] 113+ messages in thread
* [PATCH v2 07/22] KVM: ARM64: Add reset and access handlers for PMCEID0 and PMCEID1 register
2015-09-11 8:54 ` Shannon Zhao
(?)
@ 2015-09-11 8:55 ` Shannon Zhao
-1 siblings, 0 replies; 113+ messages in thread
From: Shannon Zhao @ 2015-09-11 8:55 UTC (permalink / raw)
To: linux-arm-kernel
From: Shannon Zhao <shannon.zhao@linaro.org>
Add reset handler which gets host value of PMCEID0 or PMCEID1. Since
write action to PMCEID0 or PMCEID1 is ignored, add a new case for this.
Signed-off-by: Shannon Zhao <shannon.zhao@linaro.org>
---
arch/arm64/kvm/sys_regs.c | 36 ++++++++++++++++++++++++++++++++----
1 file changed, 32 insertions(+), 4 deletions(-)
diff --git a/arch/arm64/kvm/sys_regs.c b/arch/arm64/kvm/sys_regs.c
index 24b8972..b3bc717 100644
--- a/arch/arm64/kvm/sys_regs.c
+++ b/arch/arm64/kvm/sys_regs.c
@@ -251,6 +251,26 @@ static void reset_pmcr(struct kvm_vcpu *vcpu, const struct sys_reg_desc *r)
| (ARMV8_PMCR_MASK & 0xdecafbad);
}
+static void reset_pmceid(struct kvm_vcpu *vcpu, const struct sys_reg_desc *r)
+{
+ u32 pmceid;
+
+ if (r->reg == PMCEID0_EL0 || r->reg == c9_PMCEID0) {
+ asm volatile("mrs %0, pmceid0_el0\n" : "=r" (pmceid));
+ if (!vcpu_mode_is_32bit(vcpu))
+ vcpu_sys_reg(vcpu, r->reg) = pmceid;
+ else
+ vcpu_cp15(vcpu, r->reg) = pmceid;
+ } else {
+ /* PMCEID1_EL0 or c9_PMCEID1 */
+ asm volatile("mrs %0, pmceid1_el0\n" : "=r" (pmceid));
+ if (!vcpu_mode_is_32bit(vcpu))
+ vcpu_sys_reg(vcpu, r->reg) = pmceid;
+ else
+ vcpu_cp15(vcpu, r->reg) = pmceid;
+ }
+}
+
/* PMU registers accessor. */
static bool access_pmu_regs(struct kvm_vcpu *vcpu,
const struct sys_reg_params *p,
@@ -268,6 +288,9 @@ static bool access_pmu_regs(struct kvm_vcpu *vcpu,
vcpu_sys_reg(vcpu, r->reg) = val;
break;
}
+ case PMCEID0_EL0:
+ case PMCEID1_EL0:
+ return ignore_write(vcpu, p);
default:
vcpu_sys_reg(vcpu, r->reg) = *vcpu_reg(vcpu, p->Rt);
break;
@@ -488,10 +511,10 @@ static const struct sys_reg_desc sys_reg_descs[] = {
access_pmu_regs, reset_unknown, PMSELR_EL0 },
/* PMCEID0_EL0 */
{ Op0(0b11), Op1(0b011), CRn(0b1001), CRm(0b1100), Op2(0b110),
- trap_raz_wi },
+ access_pmu_regs, reset_pmceid, PMCEID0_EL0 },
/* PMCEID1_EL0 */
{ Op0(0b11), Op1(0b011), CRn(0b1001), CRm(0b1100), Op2(0b111),
- trap_raz_wi },
+ access_pmu_regs, reset_pmceid, PMCEID1_EL0 },
/* PMCCNTR_EL0 */
{ Op0(0b11), Op1(0b011), CRn(0b1001), CRm(0b1101), Op2(0b000),
trap_raz_wi },
@@ -692,6 +715,9 @@ static bool access_pmu_cp15_regs(struct kvm_vcpu *vcpu,
vcpu_cp15(vcpu, r->reg) = val;
break;
}
+ case c9_PMCEID0:
+ case c9_PMCEID1:
+ return ignore_write(vcpu, p);
default:
vcpu_cp15(vcpu, r->reg) = *vcpu_reg(vcpu, p->Rt);
break;
@@ -738,8 +764,10 @@ static const struct sys_reg_desc cp15_regs[] = {
{ Op1( 0), CRn( 9), CRm(12), Op2( 3), trap_raz_wi },
{ Op1( 0), CRn( 9), CRm(12), Op2( 5), access_pmu_cp15_regs,
reset_unknown_cp15, c9_PMSELR },
- { Op1( 0), CRn( 9), CRm(12), Op2( 6), trap_raz_wi },
- { Op1( 0), CRn( 9), CRm(12), Op2( 7), trap_raz_wi },
+ { Op1( 0), CRn( 9), CRm(12), Op2( 6), access_pmu_cp15_regs,
+ reset_pmceid, c9_PMCEID0 },
+ { Op1( 0), CRn( 9), CRm(12), Op2( 7), access_pmu_cp15_regs,
+ reset_pmceid, c9_PMCEID1 },
{ Op1( 0), CRn( 9), CRm(13), Op2( 0), trap_raz_wi },
{ Op1( 0), CRn( 9), CRm(13), Op2( 1), trap_raz_wi },
{ Op1( 0), CRn( 9), CRm(13), Op2( 2), trap_raz_wi },
--
2.0.4
^ permalink raw reply related [flat|nested] 113+ messages in thread
* [PATCH v2 07/22] KVM: ARM64: Add reset and access handlers for PMCEID0 and PMCEID1 register
@ 2015-09-11 8:55 ` Shannon Zhao
0 siblings, 0 replies; 113+ messages in thread
From: Shannon Zhao @ 2015-09-11 8:55 UTC (permalink / raw)
To: kvmarm; +Cc: kvm, marc.zyngier, will.deacon, linux-arm-kernel, shannon.zhao
From: Shannon Zhao <shannon.zhao@linaro.org>
Add reset handler which gets host value of PMCEID0 or PMCEID1. Since
write action to PMCEID0 or PMCEID1 is ignored, add a new case for this.
Signed-off-by: Shannon Zhao <shannon.zhao@linaro.org>
---
arch/arm64/kvm/sys_regs.c | 36 ++++++++++++++++++++++++++++++++----
1 file changed, 32 insertions(+), 4 deletions(-)
diff --git a/arch/arm64/kvm/sys_regs.c b/arch/arm64/kvm/sys_regs.c
index 24b8972..b3bc717 100644
--- a/arch/arm64/kvm/sys_regs.c
+++ b/arch/arm64/kvm/sys_regs.c
@@ -251,6 +251,26 @@ static void reset_pmcr(struct kvm_vcpu *vcpu, const struct sys_reg_desc *r)
| (ARMV8_PMCR_MASK & 0xdecafbad);
}
+static void reset_pmceid(struct kvm_vcpu *vcpu, const struct sys_reg_desc *r)
+{
+ u32 pmceid;
+
+ if (r->reg == PMCEID0_EL0 || r->reg == c9_PMCEID0) {
+ asm volatile("mrs %0, pmceid0_el0\n" : "=r" (pmceid));
+ if (!vcpu_mode_is_32bit(vcpu))
+ vcpu_sys_reg(vcpu, r->reg) = pmceid;
+ else
+ vcpu_cp15(vcpu, r->reg) = pmceid;
+ } else {
+ /* PMCEID1_EL0 or c9_PMCEID1 */
+ asm volatile("mrs %0, pmceid1_el0\n" : "=r" (pmceid));
+ if (!vcpu_mode_is_32bit(vcpu))
+ vcpu_sys_reg(vcpu, r->reg) = pmceid;
+ else
+ vcpu_cp15(vcpu, r->reg) = pmceid;
+ }
+}
+
/* PMU registers accessor. */
static bool access_pmu_regs(struct kvm_vcpu *vcpu,
const struct sys_reg_params *p,
@@ -268,6 +288,9 @@ static bool access_pmu_regs(struct kvm_vcpu *vcpu,
vcpu_sys_reg(vcpu, r->reg) = val;
break;
}
+ case PMCEID0_EL0:
+ case PMCEID1_EL0:
+ return ignore_write(vcpu, p);
default:
vcpu_sys_reg(vcpu, r->reg) = *vcpu_reg(vcpu, p->Rt);
break;
@@ -488,10 +511,10 @@ static const struct sys_reg_desc sys_reg_descs[] = {
access_pmu_regs, reset_unknown, PMSELR_EL0 },
/* PMCEID0_EL0 */
{ Op0(0b11), Op1(0b011), CRn(0b1001), CRm(0b1100), Op2(0b110),
- trap_raz_wi },
+ access_pmu_regs, reset_pmceid, PMCEID0_EL0 },
/* PMCEID1_EL0 */
{ Op0(0b11), Op1(0b011), CRn(0b1001), CRm(0b1100), Op2(0b111),
- trap_raz_wi },
+ access_pmu_regs, reset_pmceid, PMCEID1_EL0 },
/* PMCCNTR_EL0 */
{ Op0(0b11), Op1(0b011), CRn(0b1001), CRm(0b1101), Op2(0b000),
trap_raz_wi },
@@ -692,6 +715,9 @@ static bool access_pmu_cp15_regs(struct kvm_vcpu *vcpu,
vcpu_cp15(vcpu, r->reg) = val;
break;
}
+ case c9_PMCEID0:
+ case c9_PMCEID1:
+ return ignore_write(vcpu, p);
default:
vcpu_cp15(vcpu, r->reg) = *vcpu_reg(vcpu, p->Rt);
break;
@@ -738,8 +764,10 @@ static const struct sys_reg_desc cp15_regs[] = {
{ Op1( 0), CRn( 9), CRm(12), Op2( 3), trap_raz_wi },
{ Op1( 0), CRn( 9), CRm(12), Op2( 5), access_pmu_cp15_regs,
reset_unknown_cp15, c9_PMSELR },
- { Op1( 0), CRn( 9), CRm(12), Op2( 6), trap_raz_wi },
- { Op1( 0), CRn( 9), CRm(12), Op2( 7), trap_raz_wi },
+ { Op1( 0), CRn( 9), CRm(12), Op2( 6), access_pmu_cp15_regs,
+ reset_pmceid, c9_PMCEID0 },
+ { Op1( 0), CRn( 9), CRm(12), Op2( 7), access_pmu_cp15_regs,
+ reset_pmceid, c9_PMCEID1 },
{ Op1( 0), CRn( 9), CRm(13), Op2( 0), trap_raz_wi },
{ Op1( 0), CRn( 9), CRm(13), Op2( 1), trap_raz_wi },
{ Op1( 0), CRn( 9), CRm(13), Op2( 2), trap_raz_wi },
--
2.0.4
^ permalink raw reply related [flat|nested] 113+ messages in thread
* [PATCH v2 07/22] KVM: ARM64: Add reset and access handlers for PMCEID0 and PMCEID1 register
@ 2015-09-11 8:55 ` Shannon Zhao
0 siblings, 0 replies; 113+ messages in thread
From: Shannon Zhao @ 2015-09-11 8:55 UTC (permalink / raw)
To: kvmarm; +Cc: kvm, marc.zyngier, will.deacon, linux-arm-kernel, shannon.zhao
From: Shannon Zhao <shannon.zhao@linaro.org>
Add reset handler which gets host value of PMCEID0 or PMCEID1. Since
write action to PMCEID0 or PMCEID1 is ignored, add a new case for this.
Signed-off-by: Shannon Zhao <shannon.zhao@linaro.org>
---
arch/arm64/kvm/sys_regs.c | 36 ++++++++++++++++++++++++++++++++----
1 file changed, 32 insertions(+), 4 deletions(-)
diff --git a/arch/arm64/kvm/sys_regs.c b/arch/arm64/kvm/sys_regs.c
index 24b8972..b3bc717 100644
--- a/arch/arm64/kvm/sys_regs.c
+++ b/arch/arm64/kvm/sys_regs.c
@@ -251,6 +251,26 @@ static void reset_pmcr(struct kvm_vcpu *vcpu, const struct sys_reg_desc *r)
| (ARMV8_PMCR_MASK & 0xdecafbad);
}
+static void reset_pmceid(struct kvm_vcpu *vcpu, const struct sys_reg_desc *r)
+{
+ u32 pmceid;
+
+ if (r->reg == PMCEID0_EL0 || r->reg == c9_PMCEID0) {
+ asm volatile("mrs %0, pmceid0_el0\n" : "=r" (pmceid));
+ if (!vcpu_mode_is_32bit(vcpu))
+ vcpu_sys_reg(vcpu, r->reg) = pmceid;
+ else
+ vcpu_cp15(vcpu, r->reg) = pmceid;
+ } else {
+ /* PMCEID1_EL0 or c9_PMCEID1 */
+ asm volatile("mrs %0, pmceid1_el0\n" : "=r" (pmceid));
+ if (!vcpu_mode_is_32bit(vcpu))
+ vcpu_sys_reg(vcpu, r->reg) = pmceid;
+ else
+ vcpu_cp15(vcpu, r->reg) = pmceid;
+ }
+}
+
/* PMU registers accessor. */
static bool access_pmu_regs(struct kvm_vcpu *vcpu,
const struct sys_reg_params *p,
@@ -268,6 +288,9 @@ static bool access_pmu_regs(struct kvm_vcpu *vcpu,
vcpu_sys_reg(vcpu, r->reg) = val;
break;
}
+ case PMCEID0_EL0:
+ case PMCEID1_EL0:
+ return ignore_write(vcpu, p);
default:
vcpu_sys_reg(vcpu, r->reg) = *vcpu_reg(vcpu, p->Rt);
break;
@@ -488,10 +511,10 @@ static const struct sys_reg_desc sys_reg_descs[] = {
access_pmu_regs, reset_unknown, PMSELR_EL0 },
/* PMCEID0_EL0 */
{ Op0(0b11), Op1(0b011), CRn(0b1001), CRm(0b1100), Op2(0b110),
- trap_raz_wi },
+ access_pmu_regs, reset_pmceid, PMCEID0_EL0 },
/* PMCEID1_EL0 */
{ Op0(0b11), Op1(0b011), CRn(0b1001), CRm(0b1100), Op2(0b111),
- trap_raz_wi },
+ access_pmu_regs, reset_pmceid, PMCEID1_EL0 },
/* PMCCNTR_EL0 */
{ Op0(0b11), Op1(0b011), CRn(0b1001), CRm(0b1101), Op2(0b000),
trap_raz_wi },
@@ -692,6 +715,9 @@ static bool access_pmu_cp15_regs(struct kvm_vcpu *vcpu,
vcpu_cp15(vcpu, r->reg) = val;
break;
}
+ case c9_PMCEID0:
+ case c9_PMCEID1:
+ return ignore_write(vcpu, p);
default:
vcpu_cp15(vcpu, r->reg) = *vcpu_reg(vcpu, p->Rt);
break;
@@ -738,8 +764,10 @@ static const struct sys_reg_desc cp15_regs[] = {
{ Op1( 0), CRn( 9), CRm(12), Op2( 3), trap_raz_wi },
{ Op1( 0), CRn( 9), CRm(12), Op2( 5), access_pmu_cp15_regs,
reset_unknown_cp15, c9_PMSELR },
- { Op1( 0), CRn( 9), CRm(12), Op2( 6), trap_raz_wi },
- { Op1( 0), CRn( 9), CRm(12), Op2( 7), trap_raz_wi },
+ { Op1( 0), CRn( 9), CRm(12), Op2( 6), access_pmu_cp15_regs,
+ reset_pmceid, c9_PMCEID0 },
+ { Op1( 0), CRn( 9), CRm(12), Op2( 7), access_pmu_cp15_regs,
+ reset_pmceid, c9_PMCEID1 },
{ Op1( 0), CRn( 9), CRm(13), Op2( 0), trap_raz_wi },
{ Op1( 0), CRn( 9), CRm(13), Op2( 1), trap_raz_wi },
{ Op1( 0), CRn( 9), CRm(13), Op2( 2), trap_raz_wi },
--
2.0.4
^ permalink raw reply related [flat|nested] 113+ messages in thread
* [PATCH v2 08/22] KVM: ARM64: PMU: Add perf event map and introduce perf event creating function
2015-09-11 8:54 ` Shannon Zhao
(?)
@ 2015-09-11 8:55 ` Shannon Zhao
-1 siblings, 0 replies; 113+ messages in thread
From: Shannon Zhao @ 2015-09-11 8:55 UTC (permalink / raw)
To: linux-arm-kernel
From: Shannon Zhao <shannon.zhao@linaro.org>
When we use tools like perf on host, perf passes the event type and the
id of this event type category to kernel, then kernel will map them to
hardware event number and write this number to PMU PMEVTYPER<n>_EL0
register. While we're trapping and emulating guest accesses to PMU
registers, we get the hardware event number and map it to the event type
and the id reversely. Then call perf_event kernel API to create an event
for it.
Signed-off-by: Shannon Zhao <shannon.zhao@linaro.org>
---
arch/arm64/include/asm/pmu.h | 2 +
arch/arm64/kvm/Makefile | 1 +
include/kvm/arm_pmu.h | 15 +++
virt/kvm/arm/pmu.c | 240 +++++++++++++++++++++++++++++++++++++++++++
4 files changed, 258 insertions(+)
create mode 100644 virt/kvm/arm/pmu.c
diff --git a/arch/arm64/include/asm/pmu.h b/arch/arm64/include/asm/pmu.h
index 95681e6..42e7093 100644
--- a/arch/arm64/include/asm/pmu.h
+++ b/arch/arm64/include/asm/pmu.h
@@ -33,6 +33,8 @@
#define ARMV8_PMCR_D (1 << 3) /* CCNT counts every 64th cpu cycle */
#define ARMV8_PMCR_X (1 << 4) /* Export to ETM */
#define ARMV8_PMCR_DP (1 << 5) /* Disable CCNT if non-invasive debug*/
+/* Determines which PMCCNTR_EL0 bit generates an overflow */
+#define ARMV8_PMCR_LC (1 << 6)
#define ARMV8_PMCR_N_SHIFT 11 /* Number of counters supported */
#define ARMV8_PMCR_N_MASK 0x1f
#define ARMV8_PMCR_MASK 0x3f /* Mask for writable bits */
diff --git a/arch/arm64/kvm/Makefile b/arch/arm64/kvm/Makefile
index f90f4aa..78db4ee 100644
--- a/arch/arm64/kvm/Makefile
+++ b/arch/arm64/kvm/Makefile
@@ -27,3 +27,4 @@ kvm-$(CONFIG_KVM_ARM_HOST) += $(KVM)/arm/vgic-v3.o
kvm-$(CONFIG_KVM_ARM_HOST) += $(KVM)/arm/vgic-v3-emul.o
kvm-$(CONFIG_KVM_ARM_HOST) += vgic-v3-switch.o
kvm-$(CONFIG_KVM_ARM_HOST) += $(KVM)/arm/arch_timer.o
+kvm-$(CONFIG_KVM_ARM_PMU) += $(KVM)/arm/pmu.o
diff --git a/include/kvm/arm_pmu.h b/include/kvm/arm_pmu.h
index 64af88a..387ec6f 100644
--- a/include/kvm/arm_pmu.h
+++ b/include/kvm/arm_pmu.h
@@ -36,4 +36,19 @@ struct kvm_pmu {
#endif
};
+#ifdef CONFIG_KVM_ARM_PMU
+unsigned long kvm_pmu_get_counter_value(struct kvm_vcpu *vcpu,
+ unsigned long select_idx);
+void kvm_pmu_set_counter_event_type(struct kvm_vcpu *vcpu, unsigned long data,
+ unsigned long select_idx);
+#else
+unsigned long kvm_pmu_get_counter_value(struct kvm_vcpu *vcpu,
+ unsigned long select_idx)
+{
+ return 0;
+}
+void kvm_pmu_set_counter_event_type(struct kvm_vcpu *vcpu, unsigned long data,
+ unsigned long select_idx) {}
+#endif
+
#endif
diff --git a/virt/kvm/arm/pmu.c b/virt/kvm/arm/pmu.c
new file mode 100644
index 0000000..0c7fe5c
--- /dev/null
+++ b/virt/kvm/arm/pmu.c
@@ -0,0 +1,240 @@
+/*
+ * Copyright (C) 2015 Linaro Ltd.
+ * Author: Shannon Zhao <shannon.zhao@linaro.org>
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program. If not, see <http://www.gnu.org/licenses/>.
+ */
+
+#include <linux/cpu.h>
+#include <linux/kvm.h>
+#include <linux/kvm_host.h>
+#include <linux/perf_event.h>
+#include <asm/kvm_emulate.h>
+#include <kvm/arm_pmu.h>
+
+/* PMU HW events mapping. */
+static struct kvm_pmu_hw_event_map {
+ unsigned eventsel;
+ unsigned event_type;
+} kvm_pmu_hw_events[] = {
+ [0] = { 0x11, PERF_COUNT_HW_CPU_CYCLES },
+ [1] = { 0x08, PERF_COUNT_HW_INSTRUCTIONS },
+ [2] = { 0x04, PERF_COUNT_HW_CACHE_REFERENCES },
+ [3] = { 0x03, PERF_COUNT_HW_CACHE_MISSES },
+ [4] = { 0x10, PERF_COUNT_HW_BRANCH_MISSES },
+};
+
+/* PMU HW cache events mapping. */
+static struct kvm_pmu_hw_cache_event_map {
+ unsigned eventsel;
+ unsigned cache_type;
+ unsigned cache_op;
+ unsigned cache_result;
+} kvm_pmu_hw_cache_events[] = {
+ [0] = { 0x12, PERF_COUNT_HW_CACHE_BPU, PERF_COUNT_HW_CACHE_OP_READ,
+ PERF_COUNT_HW_CACHE_RESULT_ACCESS },
+ [1] = { 0x12, PERF_COUNT_HW_CACHE_BPU, PERF_COUNT_HW_CACHE_OP_WRITE,
+ PERF_COUNT_HW_CACHE_RESULT_ACCESS },
+};
+
+static void kvm_pmu_set_evttyper(struct kvm_vcpu *vcpu, unsigned long idx,
+ unsigned long val)
+{
+ if (!vcpu_mode_is_32bit(vcpu))
+ vcpu_sys_reg(vcpu, PMEVTYPER0_EL0 + idx) = val;
+ else
+ vcpu_cp15(vcpu, c14_PMEVTYPER0 + idx) = val;
+}
+
+static unsigned long kvm_pmu_get_evttyper(struct kvm_vcpu *vcpu,
+ unsigned long idx)
+{
+ if (!vcpu_mode_is_32bit(vcpu))
+ return vcpu_sys_reg(vcpu, PMEVTYPER0_EL0 + idx)
+ & ARMV8_EVTYPE_EVENT;
+ else
+ return vcpu_cp15(vcpu, c14_PMEVTYPER0 + idx)
+ & ARMV8_EVTYPE_EVENT;
+}
+
+/**
+ * kvm_pmu_stop_counter - stop PMU counter for the selected counter
+ * @vcpu: The vcpu pointer
+ * @select_idx: The counter index
+ *
+ * If this counter has been configured to monitor some event, disable and
+ * release it.
+ */
+static void kvm_pmu_stop_counter(struct kvm_vcpu *vcpu,
+ unsigned long select_idx)
+{
+ struct kvm_pmu *pmu = &vcpu->arch.pmu;
+ struct kvm_pmc *pmc = &pmu->pmc[select_idx];
+
+ if (pmc->perf_event) {
+ perf_event_disable(pmc->perf_event);
+ perf_event_release_kernel(pmc->perf_event);
+ pmc->perf_event = NULL;
+ }
+ kvm_pmu_set_evttyper(vcpu, select_idx, ARMV8_EVTYPE_EVENT);
+}
+
+/**
+ * kvm_pmu_get_counter_value - get PMU counter value
+ * @vcpu: The vcpu pointer
+ * @select_idx: The counter index
+ */
+unsigned long kvm_pmu_get_counter_value(struct kvm_vcpu *vcpu,
+ unsigned long select_idx)
+{
+ u64 enabled, running;
+ struct kvm_pmu *pmu = &vcpu->arch.pmu;
+ struct kvm_pmc *pmc = &pmu->pmc[select_idx];
+ unsigned long counter;
+
+ if (!vcpu_mode_is_32bit(vcpu))
+ counter = vcpu_sys_reg(vcpu, PMEVCNTR0_EL0 + select_idx);
+ else
+ counter = vcpu_cp15(vcpu, c14_PMEVCNTR0 + select_idx);
+
+ if (pmc->perf_event) {
+ counter += perf_event_read_value(pmc->perf_event,
+ &enabled, &running);
+ }
+ return counter;
+}
+
+/**
+ * kvm_pmu_find_hw_event - find hardware event
+ * @pmu: The pmu pointer
+ * @event_select: The number of selected event type
+ *
+ * Based on the number of selected event type, find out whether it belongs to
+ * PERF_TYPE_HARDWARE. If so, return the corresponding event id.
+ */
+static unsigned kvm_pmu_find_hw_event(struct kvm_pmu *pmu,
+ unsigned long event_select)
+{
+ int i;
+
+ for (i = 0; i < ARRAY_SIZE(kvm_pmu_hw_events); i++)
+ if (kvm_pmu_hw_events[i].eventsel == event_select)
+ return kvm_pmu_hw_events[i].event_type;
+
+ return PERF_COUNT_HW_MAX;
+}
+
+/**
+ * kvm_pmu_find_hw_cache_event - find hardware cache event
+ * @pmu: The pmu pointer
+ * @event_select: The number of selected event type
+ *
+ * Based on the number of selected event type, find out whether it belongs to
+ * PERF_TYPE_HW_CACHE. If so, return the corresponding event id.
+ */
+static unsigned kvm_pmu_find_hw_cache_event(struct kvm_pmu *pmu,
+ unsigned long event_select)
+{
+ int i;
+ unsigned config;
+
+ for (i = 0; i < ARRAY_SIZE(kvm_pmu_hw_cache_events); i++)
+ if (kvm_pmu_hw_cache_events[i].eventsel == event_select) {
+ config = (kvm_pmu_hw_cache_events[i].cache_type & 0xff)
+ | ((kvm_pmu_hw_cache_events[i].cache_op & 0xff) << 8)
+ | ((kvm_pmu_hw_cache_events[i].cache_result & 0xff) << 16);
+ }
+
+ return PERF_COUNT_HW_CACHE_MAX;
+}
+
+/**
+ * kvm_pmu_set_counter_event_type - set selected counter to monitor some event
+ * @vcpu: The vcpu pointer
+ * @data: The data guest writes to PMXEVTYPER_EL0
+ * @select_idx: The number of selected counter
+ *
+ * When OS accesses PMXEVTYPER_EL0, that means it wants to set a PMC to count an
+ * event with given hardware event number. Here we call perf_event API to
+ * emulate this action and create a kernel perf event for it.
+ */
+void kvm_pmu_set_counter_event_type(struct kvm_vcpu *vcpu, unsigned long data,
+ unsigned long select_idx)
+{
+ struct kvm_pmu *pmu = &vcpu->arch.pmu;
+ struct kvm_pmc *pmc = &pmu->pmc[select_idx];
+ struct perf_event *event;
+ struct perf_event_attr attr;
+ unsigned config, type = PERF_TYPE_RAW;
+ unsigned int new_eventsel, old_eventsel;
+ u64 counter;
+ int overflow_bit, pmcr_lc;
+
+ old_eventsel = kvm_pmu_get_evttyper(vcpu, select_idx);
+ new_eventsel = data & ARMV8_EVTYPE_EVENT;
+ if (new_eventsel == old_eventsel) {
+ if (pmc->perf_event)
+ local64_set(&pmc->perf_event->count, 0);
+ return;
+ }
+
+ kvm_pmu_stop_counter(vcpu, select_idx);
+ kvm_pmu_set_evttyper(vcpu, select_idx, data);
+
+ config = kvm_pmu_find_hw_event(pmu, new_eventsel);
+ if (config != PERF_COUNT_HW_MAX) {
+ type = PERF_TYPE_HARDWARE;
+ } else {
+ config = kvm_pmu_find_hw_cache_event(pmu, new_eventsel);
+ if (config != PERF_COUNT_HW_CACHE_MAX)
+ type = PERF_TYPE_HW_CACHE;
+ }
+
+ if (type == PERF_TYPE_RAW)
+ config = new_eventsel;
+
+ memset(&attr, 0, sizeof(struct perf_event_attr));
+ attr.type = type;
+ attr.size = sizeof(attr);
+ attr.pinned = 1;
+ attr.disabled = 1;
+ attr.exclude_user = data & ARMV8_EXCLUDE_EL0 ? 1 : 0;
+ attr.exclude_kernel = data & ARMV8_EXCLUDE_EL1 ? 1 : 0;
+ attr.exclude_host = 1; /* Don't count host events */
+ attr.config = config;
+
+ overflow_bit = 31; /* Generic counters are 32-bit registers*/
+ if (new_eventsel == 0x11) {
+ /* Cycle counter overflow on increment that changes PMCCNTR[63]
+ * or PMCCNTR[31] from 1 to 0 according to the value of
+ * ARMV8_PMCR_LC
+ */
+ if (!vcpu_mode_is_32bit(vcpu))
+ pmcr_lc = vcpu_sys_reg(vcpu, PMCR_EL0) & ARMV8_PMCR_LC;
+ else
+ pmcr_lc = vcpu_cp15(vcpu, c9_PMCR) & ARMV8_PMCR_LC;
+
+ overflow_bit = pmcr_lc ? 63 : 31;
+ }
+ counter = kvm_pmu_get_counter_value(vcpu, select_idx);
+ /* The initial sample period (overflow count) of an event. */
+ attr.sample_period = (-counter) & (((u64)1 << overflow_bit) - 1);
+
+ event = perf_event_create_kernel_counter(&attr, -1, current, NULL, pmc);
+ if (IS_ERR(event)) {
+ printk_once("kvm: pmu event creation failed %ld\n",
+ PTR_ERR(event));
+ return;
+ }
+ pmc->perf_event = event;
+}
--
2.0.4
^ permalink raw reply related [flat|nested] 113+ messages in thread
* [PATCH v2 08/22] KVM: ARM64: PMU: Add perf event map and introduce perf event creating function
@ 2015-09-11 8:55 ` Shannon Zhao
0 siblings, 0 replies; 113+ messages in thread
From: Shannon Zhao @ 2015-09-11 8:55 UTC (permalink / raw)
To: kvmarm; +Cc: kvm, marc.zyngier, will.deacon, linux-arm-kernel, shannon.zhao
From: Shannon Zhao <shannon.zhao@linaro.org>
When we use tools like perf on host, perf passes the event type and the
id of this event type category to kernel, then kernel will map them to
hardware event number and write this number to PMU PMEVTYPER<n>_EL0
register. While we're trapping and emulating guest accesses to PMU
registers, we get the hardware event number and map it to the event type
and the id reversely. Then call perf_event kernel API to create an event
for it.
Signed-off-by: Shannon Zhao <shannon.zhao@linaro.org>
---
arch/arm64/include/asm/pmu.h | 2 +
arch/arm64/kvm/Makefile | 1 +
include/kvm/arm_pmu.h | 15 +++
virt/kvm/arm/pmu.c | 240 +++++++++++++++++++++++++++++++++++++++++++
4 files changed, 258 insertions(+)
create mode 100644 virt/kvm/arm/pmu.c
diff --git a/arch/arm64/include/asm/pmu.h b/arch/arm64/include/asm/pmu.h
index 95681e6..42e7093 100644
--- a/arch/arm64/include/asm/pmu.h
+++ b/arch/arm64/include/asm/pmu.h
@@ -33,6 +33,8 @@
#define ARMV8_PMCR_D (1 << 3) /* CCNT counts every 64th cpu cycle */
#define ARMV8_PMCR_X (1 << 4) /* Export to ETM */
#define ARMV8_PMCR_DP (1 << 5) /* Disable CCNT if non-invasive debug*/
+/* Determines which PMCCNTR_EL0 bit generates an overflow */
+#define ARMV8_PMCR_LC (1 << 6)
#define ARMV8_PMCR_N_SHIFT 11 /* Number of counters supported */
#define ARMV8_PMCR_N_MASK 0x1f
#define ARMV8_PMCR_MASK 0x3f /* Mask for writable bits */
diff --git a/arch/arm64/kvm/Makefile b/arch/arm64/kvm/Makefile
index f90f4aa..78db4ee 100644
--- a/arch/arm64/kvm/Makefile
+++ b/arch/arm64/kvm/Makefile
@@ -27,3 +27,4 @@ kvm-$(CONFIG_KVM_ARM_HOST) += $(KVM)/arm/vgic-v3.o
kvm-$(CONFIG_KVM_ARM_HOST) += $(KVM)/arm/vgic-v3-emul.o
kvm-$(CONFIG_KVM_ARM_HOST) += vgic-v3-switch.o
kvm-$(CONFIG_KVM_ARM_HOST) += $(KVM)/arm/arch_timer.o
+kvm-$(CONFIG_KVM_ARM_PMU) += $(KVM)/arm/pmu.o
diff --git a/include/kvm/arm_pmu.h b/include/kvm/arm_pmu.h
index 64af88a..387ec6f 100644
--- a/include/kvm/arm_pmu.h
+++ b/include/kvm/arm_pmu.h
@@ -36,4 +36,19 @@ struct kvm_pmu {
#endif
};
+#ifdef CONFIG_KVM_ARM_PMU
+unsigned long kvm_pmu_get_counter_value(struct kvm_vcpu *vcpu,
+ unsigned long select_idx);
+void kvm_pmu_set_counter_event_type(struct kvm_vcpu *vcpu, unsigned long data,
+ unsigned long select_idx);
+#else
+unsigned long kvm_pmu_get_counter_value(struct kvm_vcpu *vcpu,
+ unsigned long select_idx)
+{
+ return 0;
+}
+void kvm_pmu_set_counter_event_type(struct kvm_vcpu *vcpu, unsigned long data,
+ unsigned long select_idx) {}
+#endif
+
#endif
diff --git a/virt/kvm/arm/pmu.c b/virt/kvm/arm/pmu.c
new file mode 100644
index 0000000..0c7fe5c
--- /dev/null
+++ b/virt/kvm/arm/pmu.c
@@ -0,0 +1,240 @@
+/*
+ * Copyright (C) 2015 Linaro Ltd.
+ * Author: Shannon Zhao <shannon.zhao@linaro.org>
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program. If not, see <http://www.gnu.org/licenses/>.
+ */
+
+#include <linux/cpu.h>
+#include <linux/kvm.h>
+#include <linux/kvm_host.h>
+#include <linux/perf_event.h>
+#include <asm/kvm_emulate.h>
+#include <kvm/arm_pmu.h>
+
+/* PMU HW events mapping. */
+static struct kvm_pmu_hw_event_map {
+ unsigned eventsel;
+ unsigned event_type;
+} kvm_pmu_hw_events[] = {
+ [0] = { 0x11, PERF_COUNT_HW_CPU_CYCLES },
+ [1] = { 0x08, PERF_COUNT_HW_INSTRUCTIONS },
+ [2] = { 0x04, PERF_COUNT_HW_CACHE_REFERENCES },
+ [3] = { 0x03, PERF_COUNT_HW_CACHE_MISSES },
+ [4] = { 0x10, PERF_COUNT_HW_BRANCH_MISSES },
+};
+
+/* PMU HW cache events mapping. */
+static struct kvm_pmu_hw_cache_event_map {
+ unsigned eventsel;
+ unsigned cache_type;
+ unsigned cache_op;
+ unsigned cache_result;
+} kvm_pmu_hw_cache_events[] = {
+ [0] = { 0x12, PERF_COUNT_HW_CACHE_BPU, PERF_COUNT_HW_CACHE_OP_READ,
+ PERF_COUNT_HW_CACHE_RESULT_ACCESS },
+ [1] = { 0x12, PERF_COUNT_HW_CACHE_BPU, PERF_COUNT_HW_CACHE_OP_WRITE,
+ PERF_COUNT_HW_CACHE_RESULT_ACCESS },
+};
+
+static void kvm_pmu_set_evttyper(struct kvm_vcpu *vcpu, unsigned long idx,
+ unsigned long val)
+{
+ if (!vcpu_mode_is_32bit(vcpu))
+ vcpu_sys_reg(vcpu, PMEVTYPER0_EL0 + idx) = val;
+ else
+ vcpu_cp15(vcpu, c14_PMEVTYPER0 + idx) = val;
+}
+
+static unsigned long kvm_pmu_get_evttyper(struct kvm_vcpu *vcpu,
+ unsigned long idx)
+{
+ if (!vcpu_mode_is_32bit(vcpu))
+ return vcpu_sys_reg(vcpu, PMEVTYPER0_EL0 + idx)
+ & ARMV8_EVTYPE_EVENT;
+ else
+ return vcpu_cp15(vcpu, c14_PMEVTYPER0 + idx)
+ & ARMV8_EVTYPE_EVENT;
+}
+
+/**
+ * kvm_pmu_stop_counter - stop PMU counter for the selected counter
+ * @vcpu: The vcpu pointer
+ * @select_idx: The counter index
+ *
+ * If this counter has been configured to monitor some event, disable and
+ * release it.
+ */
+static void kvm_pmu_stop_counter(struct kvm_vcpu *vcpu,
+ unsigned long select_idx)
+{
+ struct kvm_pmu *pmu = &vcpu->arch.pmu;
+ struct kvm_pmc *pmc = &pmu->pmc[select_idx];
+
+ if (pmc->perf_event) {
+ perf_event_disable(pmc->perf_event);
+ perf_event_release_kernel(pmc->perf_event);
+ pmc->perf_event = NULL;
+ }
+ kvm_pmu_set_evttyper(vcpu, select_idx, ARMV8_EVTYPE_EVENT);
+}
+
+/**
+ * kvm_pmu_get_counter_value - get PMU counter value
+ * @vcpu: The vcpu pointer
+ * @select_idx: The counter index
+ */
+unsigned long kvm_pmu_get_counter_value(struct kvm_vcpu *vcpu,
+ unsigned long select_idx)
+{
+ u64 enabled, running;
+ struct kvm_pmu *pmu = &vcpu->arch.pmu;
+ struct kvm_pmc *pmc = &pmu->pmc[select_idx];
+ unsigned long counter;
+
+ if (!vcpu_mode_is_32bit(vcpu))
+ counter = vcpu_sys_reg(vcpu, PMEVCNTR0_EL0 + select_idx);
+ else
+ counter = vcpu_cp15(vcpu, c14_PMEVCNTR0 + select_idx);
+
+ if (pmc->perf_event) {
+ counter += perf_event_read_value(pmc->perf_event,
+ &enabled, &running);
+ }
+ return counter;
+}
+
+/**
+ * kvm_pmu_find_hw_event - find hardware event
+ * @pmu: The pmu pointer
+ * @event_select: The number of selected event type
+ *
+ * Based on the number of selected event type, find out whether it belongs to
+ * PERF_TYPE_HARDWARE. If so, return the corresponding event id.
+ */
+static unsigned kvm_pmu_find_hw_event(struct kvm_pmu *pmu,
+ unsigned long event_select)
+{
+ int i;
+
+ for (i = 0; i < ARRAY_SIZE(kvm_pmu_hw_events); i++)
+ if (kvm_pmu_hw_events[i].eventsel == event_select)
+ return kvm_pmu_hw_events[i].event_type;
+
+ return PERF_COUNT_HW_MAX;
+}
+
+/**
+ * kvm_pmu_find_hw_cache_event - find hardware cache event
+ * @pmu: The pmu pointer
+ * @event_select: The number of selected event type
+ *
+ * Based on the number of selected event type, find out whether it belongs to
+ * PERF_TYPE_HW_CACHE. If so, return the corresponding event id.
+ */
+static unsigned kvm_pmu_find_hw_cache_event(struct kvm_pmu *pmu,
+ unsigned long event_select)
+{
+ int i;
+ unsigned config;
+
+ for (i = 0; i < ARRAY_SIZE(kvm_pmu_hw_cache_events); i++)
+ if (kvm_pmu_hw_cache_events[i].eventsel == event_select) {
+ config = (kvm_pmu_hw_cache_events[i].cache_type & 0xff)
+ | ((kvm_pmu_hw_cache_events[i].cache_op & 0xff) << 8)
+ | ((kvm_pmu_hw_cache_events[i].cache_result & 0xff) << 16);
+ }
+
+ return PERF_COUNT_HW_CACHE_MAX;
+}
+
+/**
+ * kvm_pmu_set_counter_event_type - set selected counter to monitor some event
+ * @vcpu: The vcpu pointer
+ * @data: The data guest writes to PMXEVTYPER_EL0
+ * @select_idx: The number of selected counter
+ *
+ * When OS accesses PMXEVTYPER_EL0, that means it wants to set a PMC to count an
+ * event with given hardware event number. Here we call perf_event API to
+ * emulate this action and create a kernel perf event for it.
+ */
+void kvm_pmu_set_counter_event_type(struct kvm_vcpu *vcpu, unsigned long data,
+ unsigned long select_idx)
+{
+ struct kvm_pmu *pmu = &vcpu->arch.pmu;
+ struct kvm_pmc *pmc = &pmu->pmc[select_idx];
+ struct perf_event *event;
+ struct perf_event_attr attr;
+ unsigned config, type = PERF_TYPE_RAW;
+ unsigned int new_eventsel, old_eventsel;
+ u64 counter;
+ int overflow_bit, pmcr_lc;
+
+ old_eventsel = kvm_pmu_get_evttyper(vcpu, select_idx);
+ new_eventsel = data & ARMV8_EVTYPE_EVENT;
+ if (new_eventsel == old_eventsel) {
+ if (pmc->perf_event)
+ local64_set(&pmc->perf_event->count, 0);
+ return;
+ }
+
+ kvm_pmu_stop_counter(vcpu, select_idx);
+ kvm_pmu_set_evttyper(vcpu, select_idx, data);
+
+ config = kvm_pmu_find_hw_event(pmu, new_eventsel);
+ if (config != PERF_COUNT_HW_MAX) {
+ type = PERF_TYPE_HARDWARE;
+ } else {
+ config = kvm_pmu_find_hw_cache_event(pmu, new_eventsel);
+ if (config != PERF_COUNT_HW_CACHE_MAX)
+ type = PERF_TYPE_HW_CACHE;
+ }
+
+ if (type == PERF_TYPE_RAW)
+ config = new_eventsel;
+
+ memset(&attr, 0, sizeof(struct perf_event_attr));
+ attr.type = type;
+ attr.size = sizeof(attr);
+ attr.pinned = 1;
+ attr.disabled = 1;
+ attr.exclude_user = data & ARMV8_EXCLUDE_EL0 ? 1 : 0;
+ attr.exclude_kernel = data & ARMV8_EXCLUDE_EL1 ? 1 : 0;
+ attr.exclude_host = 1; /* Don't count host events */
+ attr.config = config;
+
+ overflow_bit = 31; /* Generic counters are 32-bit registers*/
+ if (new_eventsel == 0x11) {
+ /* Cycle counter overflow on increment that changes PMCCNTR[63]
+ * or PMCCNTR[31] from 1 to 0 according to the value of
+ * ARMV8_PMCR_LC
+ */
+ if (!vcpu_mode_is_32bit(vcpu))
+ pmcr_lc = vcpu_sys_reg(vcpu, PMCR_EL0) & ARMV8_PMCR_LC;
+ else
+ pmcr_lc = vcpu_cp15(vcpu, c9_PMCR) & ARMV8_PMCR_LC;
+
+ overflow_bit = pmcr_lc ? 63 : 31;
+ }
+ counter = kvm_pmu_get_counter_value(vcpu, select_idx);
+ /* The initial sample period (overflow count) of an event. */
+ attr.sample_period = (-counter) & (((u64)1 << overflow_bit) - 1);
+
+ event = perf_event_create_kernel_counter(&attr, -1, current, NULL, pmc);
+ if (IS_ERR(event)) {
+ printk_once("kvm: pmu event creation failed %ld\n",
+ PTR_ERR(event));
+ return;
+ }
+ pmc->perf_event = event;
+}
--
2.0.4
^ permalink raw reply related [flat|nested] 113+ messages in thread
* [PATCH v2 08/22] KVM: ARM64: PMU: Add perf event map and introduce perf event creating function
@ 2015-09-11 8:55 ` Shannon Zhao
0 siblings, 0 replies; 113+ messages in thread
From: Shannon Zhao @ 2015-09-11 8:55 UTC (permalink / raw)
To: kvmarm; +Cc: kvm, marc.zyngier, will.deacon, linux-arm-kernel, shannon.zhao
From: Shannon Zhao <shannon.zhao@linaro.org>
When we use tools like perf on host, perf passes the event type and the
id of this event type category to kernel, then kernel will map them to
hardware event number and write this number to PMU PMEVTYPER<n>_EL0
register. While we're trapping and emulating guest accesses to PMU
registers, we get the hardware event number and map it to the event type
and the id reversely. Then call perf_event kernel API to create an event
for it.
Signed-off-by: Shannon Zhao <shannon.zhao@linaro.org>
---
arch/arm64/include/asm/pmu.h | 2 +
arch/arm64/kvm/Makefile | 1 +
include/kvm/arm_pmu.h | 15 +++
virt/kvm/arm/pmu.c | 240 +++++++++++++++++++++++++++++++++++++++++++
4 files changed, 258 insertions(+)
create mode 100644 virt/kvm/arm/pmu.c
diff --git a/arch/arm64/include/asm/pmu.h b/arch/arm64/include/asm/pmu.h
index 95681e6..42e7093 100644
--- a/arch/arm64/include/asm/pmu.h
+++ b/arch/arm64/include/asm/pmu.h
@@ -33,6 +33,8 @@
#define ARMV8_PMCR_D (1 << 3) /* CCNT counts every 64th cpu cycle */
#define ARMV8_PMCR_X (1 << 4) /* Export to ETM */
#define ARMV8_PMCR_DP (1 << 5) /* Disable CCNT if non-invasive debug*/
+/* Determines which PMCCNTR_EL0 bit generates an overflow */
+#define ARMV8_PMCR_LC (1 << 6)
#define ARMV8_PMCR_N_SHIFT 11 /* Number of counters supported */
#define ARMV8_PMCR_N_MASK 0x1f
#define ARMV8_PMCR_MASK 0x3f /* Mask for writable bits */
diff --git a/arch/arm64/kvm/Makefile b/arch/arm64/kvm/Makefile
index f90f4aa..78db4ee 100644
--- a/arch/arm64/kvm/Makefile
+++ b/arch/arm64/kvm/Makefile
@@ -27,3 +27,4 @@ kvm-$(CONFIG_KVM_ARM_HOST) += $(KVM)/arm/vgic-v3.o
kvm-$(CONFIG_KVM_ARM_HOST) += $(KVM)/arm/vgic-v3-emul.o
kvm-$(CONFIG_KVM_ARM_HOST) += vgic-v3-switch.o
kvm-$(CONFIG_KVM_ARM_HOST) += $(KVM)/arm/arch_timer.o
+kvm-$(CONFIG_KVM_ARM_PMU) += $(KVM)/arm/pmu.o
diff --git a/include/kvm/arm_pmu.h b/include/kvm/arm_pmu.h
index 64af88a..387ec6f 100644
--- a/include/kvm/arm_pmu.h
+++ b/include/kvm/arm_pmu.h
@@ -36,4 +36,19 @@ struct kvm_pmu {
#endif
};
+#ifdef CONFIG_KVM_ARM_PMU
+unsigned long kvm_pmu_get_counter_value(struct kvm_vcpu *vcpu,
+ unsigned long select_idx);
+void kvm_pmu_set_counter_event_type(struct kvm_vcpu *vcpu, unsigned long data,
+ unsigned long select_idx);
+#else
+unsigned long kvm_pmu_get_counter_value(struct kvm_vcpu *vcpu,
+ unsigned long select_idx)
+{
+ return 0;
+}
+void kvm_pmu_set_counter_event_type(struct kvm_vcpu *vcpu, unsigned long data,
+ unsigned long select_idx) {}
+#endif
+
#endif
diff --git a/virt/kvm/arm/pmu.c b/virt/kvm/arm/pmu.c
new file mode 100644
index 0000000..0c7fe5c
--- /dev/null
+++ b/virt/kvm/arm/pmu.c
@@ -0,0 +1,240 @@
+/*
+ * Copyright (C) 2015 Linaro Ltd.
+ * Author: Shannon Zhao <shannon.zhao@linaro.org>
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program. If not, see <http://www.gnu.org/licenses/>.
+ */
+
+#include <linux/cpu.h>
+#include <linux/kvm.h>
+#include <linux/kvm_host.h>
+#include <linux/perf_event.h>
+#include <asm/kvm_emulate.h>
+#include <kvm/arm_pmu.h>
+
+/* PMU HW events mapping. */
+static struct kvm_pmu_hw_event_map {
+ unsigned eventsel;
+ unsigned event_type;
+} kvm_pmu_hw_events[] = {
+ [0] = { 0x11, PERF_COUNT_HW_CPU_CYCLES },
+ [1] = { 0x08, PERF_COUNT_HW_INSTRUCTIONS },
+ [2] = { 0x04, PERF_COUNT_HW_CACHE_REFERENCES },
+ [3] = { 0x03, PERF_COUNT_HW_CACHE_MISSES },
+ [4] = { 0x10, PERF_COUNT_HW_BRANCH_MISSES },
+};
+
+/* PMU HW cache events mapping. */
+static struct kvm_pmu_hw_cache_event_map {
+ unsigned eventsel;
+ unsigned cache_type;
+ unsigned cache_op;
+ unsigned cache_result;
+} kvm_pmu_hw_cache_events[] = {
+ [0] = { 0x12, PERF_COUNT_HW_CACHE_BPU, PERF_COUNT_HW_CACHE_OP_READ,
+ PERF_COUNT_HW_CACHE_RESULT_ACCESS },
+ [1] = { 0x12, PERF_COUNT_HW_CACHE_BPU, PERF_COUNT_HW_CACHE_OP_WRITE,
+ PERF_COUNT_HW_CACHE_RESULT_ACCESS },
+};
+
+static void kvm_pmu_set_evttyper(struct kvm_vcpu *vcpu, unsigned long idx,
+ unsigned long val)
+{
+ if (!vcpu_mode_is_32bit(vcpu))
+ vcpu_sys_reg(vcpu, PMEVTYPER0_EL0 + idx) = val;
+ else
+ vcpu_cp15(vcpu, c14_PMEVTYPER0 + idx) = val;
+}
+
+static unsigned long kvm_pmu_get_evttyper(struct kvm_vcpu *vcpu,
+ unsigned long idx)
+{
+ if (!vcpu_mode_is_32bit(vcpu))
+ return vcpu_sys_reg(vcpu, PMEVTYPER0_EL0 + idx)
+ & ARMV8_EVTYPE_EVENT;
+ else
+ return vcpu_cp15(vcpu, c14_PMEVTYPER0 + idx)
+ & ARMV8_EVTYPE_EVENT;
+}
+
+/**
+ * kvm_pmu_stop_counter - stop PMU counter for the selected counter
+ * @vcpu: The vcpu pointer
+ * @select_idx: The counter index
+ *
+ * If this counter has been configured to monitor some event, disable and
+ * release it.
+ */
+static void kvm_pmu_stop_counter(struct kvm_vcpu *vcpu,
+ unsigned long select_idx)
+{
+ struct kvm_pmu *pmu = &vcpu->arch.pmu;
+ struct kvm_pmc *pmc = &pmu->pmc[select_idx];
+
+ if (pmc->perf_event) {
+ perf_event_disable(pmc->perf_event);
+ perf_event_release_kernel(pmc->perf_event);
+ pmc->perf_event = NULL;
+ }
+ kvm_pmu_set_evttyper(vcpu, select_idx, ARMV8_EVTYPE_EVENT);
+}
+
+/**
+ * kvm_pmu_get_counter_value - get PMU counter value
+ * @vcpu: The vcpu pointer
+ * @select_idx: The counter index
+ */
+unsigned long kvm_pmu_get_counter_value(struct kvm_vcpu *vcpu,
+ unsigned long select_idx)
+{
+ u64 enabled, running;
+ struct kvm_pmu *pmu = &vcpu->arch.pmu;
+ struct kvm_pmc *pmc = &pmu->pmc[select_idx];
+ unsigned long counter;
+
+ if (!vcpu_mode_is_32bit(vcpu))
+ counter = vcpu_sys_reg(vcpu, PMEVCNTR0_EL0 + select_idx);
+ else
+ counter = vcpu_cp15(vcpu, c14_PMEVCNTR0 + select_idx);
+
+ if (pmc->perf_event) {
+ counter += perf_event_read_value(pmc->perf_event,
+ &enabled, &running);
+ }
+ return counter;
+}
+
+/**
+ * kvm_pmu_find_hw_event - find hardware event
+ * @pmu: The pmu pointer
+ * @event_select: The number of selected event type
+ *
+ * Based on the number of selected event type, find out whether it belongs to
+ * PERF_TYPE_HARDWARE. If so, return the corresponding event id.
+ */
+static unsigned kvm_pmu_find_hw_event(struct kvm_pmu *pmu,
+ unsigned long event_select)
+{
+ int i;
+
+ for (i = 0; i < ARRAY_SIZE(kvm_pmu_hw_events); i++)
+ if (kvm_pmu_hw_events[i].eventsel == event_select)
+ return kvm_pmu_hw_events[i].event_type;
+
+ return PERF_COUNT_HW_MAX;
+}
+
+/**
+ * kvm_pmu_find_hw_cache_event - find hardware cache event
+ * @pmu: The pmu pointer
+ * @event_select: The number of selected event type
+ *
+ * Based on the number of selected event type, find out whether it belongs to
+ * PERF_TYPE_HW_CACHE. If so, return the corresponding event id.
+ */
+static unsigned kvm_pmu_find_hw_cache_event(struct kvm_pmu *pmu,
+ unsigned long event_select)
+{
+ int i;
+ unsigned config;
+
+ for (i = 0; i < ARRAY_SIZE(kvm_pmu_hw_cache_events); i++)
+ if (kvm_pmu_hw_cache_events[i].eventsel == event_select) {
+ config = (kvm_pmu_hw_cache_events[i].cache_type & 0xff)
+ | ((kvm_pmu_hw_cache_events[i].cache_op & 0xff) << 8)
+ | ((kvm_pmu_hw_cache_events[i].cache_result & 0xff) << 16);
+ }
+
+ return PERF_COUNT_HW_CACHE_MAX;
+}
+
+/**
+ * kvm_pmu_set_counter_event_type - set selected counter to monitor some event
+ * @vcpu: The vcpu pointer
+ * @data: The data guest writes to PMXEVTYPER_EL0
+ * @select_idx: The number of selected counter
+ *
+ * When OS accesses PMXEVTYPER_EL0, that means it wants to set a PMC to count an
+ * event with given hardware event number. Here we call perf_event API to
+ * emulate this action and create a kernel perf event for it.
+ */
+void kvm_pmu_set_counter_event_type(struct kvm_vcpu *vcpu, unsigned long data,
+ unsigned long select_idx)
+{
+ struct kvm_pmu *pmu = &vcpu->arch.pmu;
+ struct kvm_pmc *pmc = &pmu->pmc[select_idx];
+ struct perf_event *event;
+ struct perf_event_attr attr;
+ unsigned config, type = PERF_TYPE_RAW;
+ unsigned int new_eventsel, old_eventsel;
+ u64 counter;
+ int overflow_bit, pmcr_lc;
+
+ old_eventsel = kvm_pmu_get_evttyper(vcpu, select_idx);
+ new_eventsel = data & ARMV8_EVTYPE_EVENT;
+ if (new_eventsel == old_eventsel) {
+ if (pmc->perf_event)
+ local64_set(&pmc->perf_event->count, 0);
+ return;
+ }
+
+ kvm_pmu_stop_counter(vcpu, select_idx);
+ kvm_pmu_set_evttyper(vcpu, select_idx, data);
+
+ config = kvm_pmu_find_hw_event(pmu, new_eventsel);
+ if (config != PERF_COUNT_HW_MAX) {
+ type = PERF_TYPE_HARDWARE;
+ } else {
+ config = kvm_pmu_find_hw_cache_event(pmu, new_eventsel);
+ if (config != PERF_COUNT_HW_CACHE_MAX)
+ type = PERF_TYPE_HW_CACHE;
+ }
+
+ if (type == PERF_TYPE_RAW)
+ config = new_eventsel;
+
+ memset(&attr, 0, sizeof(struct perf_event_attr));
+ attr.type = type;
+ attr.size = sizeof(attr);
+ attr.pinned = 1;
+ attr.disabled = 1;
+ attr.exclude_user = data & ARMV8_EXCLUDE_EL0 ? 1 : 0;
+ attr.exclude_kernel = data & ARMV8_EXCLUDE_EL1 ? 1 : 0;
+ attr.exclude_host = 1; /* Don't count host events */
+ attr.config = config;
+
+ overflow_bit = 31; /* Generic counters are 32-bit registers*/
+ if (new_eventsel == 0x11) {
+ /* Cycle counter overflow on increment that changes PMCCNTR[63]
+ * or PMCCNTR[31] from 1 to 0 according to the value of
+ * ARMV8_PMCR_LC
+ */
+ if (!vcpu_mode_is_32bit(vcpu))
+ pmcr_lc = vcpu_sys_reg(vcpu, PMCR_EL0) & ARMV8_PMCR_LC;
+ else
+ pmcr_lc = vcpu_cp15(vcpu, c9_PMCR) & ARMV8_PMCR_LC;
+
+ overflow_bit = pmcr_lc ? 63 : 31;
+ }
+ counter = kvm_pmu_get_counter_value(vcpu, select_idx);
+ /* The initial sample period (overflow count) of an event. */
+ attr.sample_period = (-counter) & (((u64)1 << overflow_bit) - 1);
+
+ event = perf_event_create_kernel_counter(&attr, -1, current, NULL, pmc);
+ if (IS_ERR(event)) {
+ printk_once("kvm: pmu event creation failed %ld\n",
+ PTR_ERR(event));
+ return;
+ }
+ pmc->perf_event = event;
+}
--
2.0.4
^ permalink raw reply related [flat|nested] 113+ messages in thread
* [PATCH v2 09/22] KVM: ARM64: Add reset and access handlers for PMXEVTYPER register
2015-09-11 8:54 ` Shannon Zhao
(?)
@ 2015-09-11 8:55 ` Shannon Zhao
-1 siblings, 0 replies; 113+ messages in thread
From: Shannon Zhao @ 2015-09-11 8:55 UTC (permalink / raw)
To: linux-arm-kernel
From: Shannon Zhao <shannon.zhao@linaro.org>
Since the reset value of PMXEVTYPER is UNKNOWN, use reset_unknown or
reset_unknown_cp15 for its reset handler. Add access handler which
emulates writing and reading PMXEVTYPER register. When writing to
PMXEVTYPER, call kvm_pmu_set_counter_event_type to create a perf_event
for the selected event type.
Signed-off-by: Shannon Zhao <shannon.zhao@linaro.org>
---
arch/arm64/kvm/sys_regs.c | 19 +++++++++++++++++--
1 file changed, 17 insertions(+), 2 deletions(-)
diff --git a/arch/arm64/kvm/sys_regs.c b/arch/arm64/kvm/sys_regs.c
index b3bc717..59b7bc9 100644
--- a/arch/arm64/kvm/sys_regs.c
+++ b/arch/arm64/kvm/sys_regs.c
@@ -280,6 +280,13 @@ static bool access_pmu_regs(struct kvm_vcpu *vcpu,
if (p->is_write) {
switch (r->reg) {
+ case PMXEVTYPER_EL0: {
+ val = vcpu_sys_reg(vcpu, PMSELR_EL0);
+ kvm_pmu_set_counter_event_type(vcpu,
+ *vcpu_reg(vcpu, p->Rt),
+ val);
+ break;
+ }
case PMCR_EL0: {
/* Only update writeable bits of PMCR */
val = vcpu_sys_reg(vcpu, r->reg);
@@ -520,7 +527,7 @@ static const struct sys_reg_desc sys_reg_descs[] = {
trap_raz_wi },
/* PMXEVTYPER_EL0 */
{ Op0(0b11), Op1(0b011), CRn(0b1001), CRm(0b1101), Op2(0b001),
- trap_raz_wi },
+ access_pmu_regs, reset_unknown, PMXEVTYPER_EL0 },
/* PMXEVCNTR_EL0 */
{ Op0(0b11), Op1(0b011), CRn(0b1001), CRm(0b1101), Op2(0b010),
trap_raz_wi },
@@ -707,6 +714,13 @@ static bool access_pmu_cp15_regs(struct kvm_vcpu *vcpu,
if (p->is_write) {
switch (r->reg) {
+ case c9_PMXEVTYPER: {
+ val = vcpu_cp15(vcpu, c9_PMSELR);
+ kvm_pmu_set_counter_event_type(vcpu,
+ *vcpu_reg(vcpu, p->Rt),
+ val);
+ break;
+ }
case c9_PMCR: {
/* Only update writeable bits of PMCR */
val = vcpu_cp15(vcpu, r->reg);
@@ -769,7 +783,8 @@ static const struct sys_reg_desc cp15_regs[] = {
{ Op1( 0), CRn( 9), CRm(12), Op2( 7), access_pmu_cp15_regs,
reset_pmceid, c9_PMCEID1 },
{ Op1( 0), CRn( 9), CRm(13), Op2( 0), trap_raz_wi },
- { Op1( 0), CRn( 9), CRm(13), Op2( 1), trap_raz_wi },
+ { Op1( 0), CRn( 9), CRm(13), Op2( 1), access_pmu_cp15_regs,
+ reset_unknown_cp15, c9_PMXEVTYPER },
{ Op1( 0), CRn( 9), CRm(13), Op2( 2), trap_raz_wi },
{ Op1( 0), CRn( 9), CRm(14), Op2( 0), trap_raz_wi },
{ Op1( 0), CRn( 9), CRm(14), Op2( 1), trap_raz_wi },
--
2.0.4
^ permalink raw reply related [flat|nested] 113+ messages in thread
* [PATCH v2 09/22] KVM: ARM64: Add reset and access handlers for PMXEVTYPER register
@ 2015-09-11 8:55 ` Shannon Zhao
0 siblings, 0 replies; 113+ messages in thread
From: Shannon Zhao @ 2015-09-11 8:55 UTC (permalink / raw)
To: kvmarm; +Cc: kvm, marc.zyngier, will.deacon, linux-arm-kernel, shannon.zhao
From: Shannon Zhao <shannon.zhao@linaro.org>
Since the reset value of PMXEVTYPER is UNKNOWN, use reset_unknown or
reset_unknown_cp15 for its reset handler. Add access handler which
emulates writing and reading PMXEVTYPER register. When writing to
PMXEVTYPER, call kvm_pmu_set_counter_event_type to create a perf_event
for the selected event type.
Signed-off-by: Shannon Zhao <shannon.zhao@linaro.org>
---
arch/arm64/kvm/sys_regs.c | 19 +++++++++++++++++--
1 file changed, 17 insertions(+), 2 deletions(-)
diff --git a/arch/arm64/kvm/sys_regs.c b/arch/arm64/kvm/sys_regs.c
index b3bc717..59b7bc9 100644
--- a/arch/arm64/kvm/sys_regs.c
+++ b/arch/arm64/kvm/sys_regs.c
@@ -280,6 +280,13 @@ static bool access_pmu_regs(struct kvm_vcpu *vcpu,
if (p->is_write) {
switch (r->reg) {
+ case PMXEVTYPER_EL0: {
+ val = vcpu_sys_reg(vcpu, PMSELR_EL0);
+ kvm_pmu_set_counter_event_type(vcpu,
+ *vcpu_reg(vcpu, p->Rt),
+ val);
+ break;
+ }
case PMCR_EL0: {
/* Only update writeable bits of PMCR */
val = vcpu_sys_reg(vcpu, r->reg);
@@ -520,7 +527,7 @@ static const struct sys_reg_desc sys_reg_descs[] = {
trap_raz_wi },
/* PMXEVTYPER_EL0 */
{ Op0(0b11), Op1(0b011), CRn(0b1001), CRm(0b1101), Op2(0b001),
- trap_raz_wi },
+ access_pmu_regs, reset_unknown, PMXEVTYPER_EL0 },
/* PMXEVCNTR_EL0 */
{ Op0(0b11), Op1(0b011), CRn(0b1001), CRm(0b1101), Op2(0b010),
trap_raz_wi },
@@ -707,6 +714,13 @@ static bool access_pmu_cp15_regs(struct kvm_vcpu *vcpu,
if (p->is_write) {
switch (r->reg) {
+ case c9_PMXEVTYPER: {
+ val = vcpu_cp15(vcpu, c9_PMSELR);
+ kvm_pmu_set_counter_event_type(vcpu,
+ *vcpu_reg(vcpu, p->Rt),
+ val);
+ break;
+ }
case c9_PMCR: {
/* Only update writeable bits of PMCR */
val = vcpu_cp15(vcpu, r->reg);
@@ -769,7 +783,8 @@ static const struct sys_reg_desc cp15_regs[] = {
{ Op1( 0), CRn( 9), CRm(12), Op2( 7), access_pmu_cp15_regs,
reset_pmceid, c9_PMCEID1 },
{ Op1( 0), CRn( 9), CRm(13), Op2( 0), trap_raz_wi },
- { Op1( 0), CRn( 9), CRm(13), Op2( 1), trap_raz_wi },
+ { Op1( 0), CRn( 9), CRm(13), Op2( 1), access_pmu_cp15_regs,
+ reset_unknown_cp15, c9_PMXEVTYPER },
{ Op1( 0), CRn( 9), CRm(13), Op2( 2), trap_raz_wi },
{ Op1( 0), CRn( 9), CRm(14), Op2( 0), trap_raz_wi },
{ Op1( 0), CRn( 9), CRm(14), Op2( 1), trap_raz_wi },
--
2.0.4
^ permalink raw reply related [flat|nested] 113+ messages in thread
* [PATCH v2 09/22] KVM: ARM64: Add reset and access handlers for PMXEVTYPER register
@ 2015-09-11 8:55 ` Shannon Zhao
0 siblings, 0 replies; 113+ messages in thread
From: Shannon Zhao @ 2015-09-11 8:55 UTC (permalink / raw)
To: kvmarm; +Cc: kvm, marc.zyngier, will.deacon, linux-arm-kernel, shannon.zhao
From: Shannon Zhao <shannon.zhao@linaro.org>
Since the reset value of PMXEVTYPER is UNKNOWN, use reset_unknown or
reset_unknown_cp15 for its reset handler. Add access handler which
emulates writing and reading PMXEVTYPER register. When writing to
PMXEVTYPER, call kvm_pmu_set_counter_event_type to create a perf_event
for the selected event type.
Signed-off-by: Shannon Zhao <shannon.zhao@linaro.org>
---
arch/arm64/kvm/sys_regs.c | 19 +++++++++++++++++--
1 file changed, 17 insertions(+), 2 deletions(-)
diff --git a/arch/arm64/kvm/sys_regs.c b/arch/arm64/kvm/sys_regs.c
index b3bc717..59b7bc9 100644
--- a/arch/arm64/kvm/sys_regs.c
+++ b/arch/arm64/kvm/sys_regs.c
@@ -280,6 +280,13 @@ static bool access_pmu_regs(struct kvm_vcpu *vcpu,
if (p->is_write) {
switch (r->reg) {
+ case PMXEVTYPER_EL0: {
+ val = vcpu_sys_reg(vcpu, PMSELR_EL0);
+ kvm_pmu_set_counter_event_type(vcpu,
+ *vcpu_reg(vcpu, p->Rt),
+ val);
+ break;
+ }
case PMCR_EL0: {
/* Only update writeable bits of PMCR */
val = vcpu_sys_reg(vcpu, r->reg);
@@ -520,7 +527,7 @@ static const struct sys_reg_desc sys_reg_descs[] = {
trap_raz_wi },
/* PMXEVTYPER_EL0 */
{ Op0(0b11), Op1(0b011), CRn(0b1001), CRm(0b1101), Op2(0b001),
- trap_raz_wi },
+ access_pmu_regs, reset_unknown, PMXEVTYPER_EL0 },
/* PMXEVCNTR_EL0 */
{ Op0(0b11), Op1(0b011), CRn(0b1001), CRm(0b1101), Op2(0b010),
trap_raz_wi },
@@ -707,6 +714,13 @@ static bool access_pmu_cp15_regs(struct kvm_vcpu *vcpu,
if (p->is_write) {
switch (r->reg) {
+ case c9_PMXEVTYPER: {
+ val = vcpu_cp15(vcpu, c9_PMSELR);
+ kvm_pmu_set_counter_event_type(vcpu,
+ *vcpu_reg(vcpu, p->Rt),
+ val);
+ break;
+ }
case c9_PMCR: {
/* Only update writeable bits of PMCR */
val = vcpu_cp15(vcpu, r->reg);
@@ -769,7 +783,8 @@ static const struct sys_reg_desc cp15_regs[] = {
{ Op1( 0), CRn( 9), CRm(12), Op2( 7), access_pmu_cp15_regs,
reset_pmceid, c9_PMCEID1 },
{ Op1( 0), CRn( 9), CRm(13), Op2( 0), trap_raz_wi },
- { Op1( 0), CRn( 9), CRm(13), Op2( 1), trap_raz_wi },
+ { Op1( 0), CRn( 9), CRm(13), Op2( 1), access_pmu_cp15_regs,
+ reset_unknown_cp15, c9_PMXEVTYPER },
{ Op1( 0), CRn( 9), CRm(13), Op2( 2), trap_raz_wi },
{ Op1( 0), CRn( 9), CRm(14), Op2( 0), trap_raz_wi },
{ Op1( 0), CRn( 9), CRm(14), Op2( 1), trap_raz_wi },
--
2.0.4
^ permalink raw reply related [flat|nested] 113+ messages in thread
* [PATCH v2 10/22] KVM: ARM64: Add reset and access handlers for PMXEVCNTR register
2015-09-11 8:54 ` Shannon Zhao
(?)
@ 2015-09-11 8:55 ` Shannon Zhao
-1 siblings, 0 replies; 113+ messages in thread
From: Shannon Zhao @ 2015-09-11 8:55 UTC (permalink / raw)
To: linux-arm-kernel
From: Shannon Zhao <shannon.zhao@linaro.org>
Since the reset value of PMXEVCNTR is UNKNOWN, use reset_unknown for
its reset handler. Add access handler which emulates writing and reading
PMXEVCNTR register. When reading PMXEVCNTR, call perf_event_read_value
to get the count value of the perf event.
Signed-off-by: Shannon Zhao <shannon.zhao@linaro.org>
---
arch/arm64/kvm/sys_regs.c | 41 +++++++++++++++++++++++++++++++++++++----
1 file changed, 37 insertions(+), 4 deletions(-)
diff --git a/arch/arm64/kvm/sys_regs.c b/arch/arm64/kvm/sys_regs.c
index 59b7bc9..a51c8de 100644
--- a/arch/arm64/kvm/sys_regs.c
+++ b/arch/arm64/kvm/sys_regs.c
@@ -280,6 +280,12 @@ static bool access_pmu_regs(struct kvm_vcpu *vcpu,
if (p->is_write) {
switch (r->reg) {
+ case PMXEVCNTR_EL0: {
+ val = PMEVCNTR0_EL0 + vcpu_sys_reg(vcpu, PMSELR_EL0);
+ vcpu_sys_reg(vcpu, val) =
+ *vcpu_reg(vcpu, p->Rt) & 0xffffffffUL;
+ break;
+ }
case PMXEVTYPER_EL0: {
val = vcpu_sys_reg(vcpu, PMSELR_EL0);
kvm_pmu_set_counter_event_type(vcpu,
@@ -303,7 +309,17 @@ static bool access_pmu_regs(struct kvm_vcpu *vcpu,
break;
}
} else {
- *vcpu_reg(vcpu, p->Rt) = vcpu_sys_reg(vcpu, r->reg);
+ switch (r->reg) {
+ case PMXEVCNTR_EL0: {
+ val = kvm_pmu_get_counter_value(vcpu,
+ vcpu_sys_reg(vcpu, PMSELR_EL0));
+ *vcpu_reg(vcpu, p->Rt) = val;
+ break;
+ }
+ default:
+ *vcpu_reg(vcpu, p->Rt) = vcpu_sys_reg(vcpu, r->reg);
+ break;
+ }
}
return true;
@@ -530,7 +546,7 @@ static const struct sys_reg_desc sys_reg_descs[] = {
access_pmu_regs, reset_unknown, PMXEVTYPER_EL0 },
/* PMXEVCNTR_EL0 */
{ Op0(0b11), Op1(0b011), CRn(0b1001), CRm(0b1101), Op2(0b010),
- trap_raz_wi },
+ access_pmu_regs, reset_unknown, PMXEVCNTR_EL0 },
/* PMUSERENR_EL0 */
{ Op0(0b11), Op1(0b011), CRn(0b1001), CRm(0b1110), Op2(0b000),
trap_raz_wi },
@@ -714,6 +730,12 @@ static bool access_pmu_cp15_regs(struct kvm_vcpu *vcpu,
if (p->is_write) {
switch (r->reg) {
+ case c9_PMXEVCNTR: {
+ val = c14_PMEVCNTR0 + vcpu_cp15(vcpu, c9_PMSELR);
+ vcpu_cp15(vcpu, val) =
+ *vcpu_reg(vcpu, p->Rt) & 0xffffffffUL;
+ break;
+ }
case c9_PMXEVTYPER: {
val = vcpu_cp15(vcpu, c9_PMSELR);
kvm_pmu_set_counter_event_type(vcpu,
@@ -737,7 +759,17 @@ static bool access_pmu_cp15_regs(struct kvm_vcpu *vcpu,
break;
}
} else {
- *vcpu_reg(vcpu, p->Rt) = vcpu_cp15(vcpu, r->reg);
+ switch (r->reg) {
+ case c9_PMXEVCNTR: {
+ val = kvm_pmu_get_counter_value(vcpu,
+ vcpu_cp15(vcpu, c9_PMSELR));
+ *vcpu_reg(vcpu, p->Rt) = val;
+ break;
+ }
+ default:
+ *vcpu_reg(vcpu, p->Rt) = vcpu_cp15(vcpu, r->reg);
+ break;
+ }
}
return true;
@@ -785,7 +817,8 @@ static const struct sys_reg_desc cp15_regs[] = {
{ Op1( 0), CRn( 9), CRm(13), Op2( 0), trap_raz_wi },
{ Op1( 0), CRn( 9), CRm(13), Op2( 1), access_pmu_cp15_regs,
reset_unknown_cp15, c9_PMXEVTYPER },
- { Op1( 0), CRn( 9), CRm(13), Op2( 2), trap_raz_wi },
+ { Op1( 0), CRn( 9), CRm(13), Op2( 2), access_pmu_cp15_regs,
+ reset_unknown_cp15, c9_PMXEVCNTR },
{ Op1( 0), CRn( 9), CRm(14), Op2( 0), trap_raz_wi },
{ Op1( 0), CRn( 9), CRm(14), Op2( 1), trap_raz_wi },
{ Op1( 0), CRn( 9), CRm(14), Op2( 2), trap_raz_wi },
--
2.0.4
^ permalink raw reply related [flat|nested] 113+ messages in thread
* [PATCH v2 10/22] KVM: ARM64: Add reset and access handlers for PMXEVCNTR register
@ 2015-09-11 8:55 ` Shannon Zhao
0 siblings, 0 replies; 113+ messages in thread
From: Shannon Zhao @ 2015-09-11 8:55 UTC (permalink / raw)
To: kvmarm; +Cc: kvm, marc.zyngier, will.deacon, linux-arm-kernel, shannon.zhao
From: Shannon Zhao <shannon.zhao@linaro.org>
Since the reset value of PMXEVCNTR is UNKNOWN, use reset_unknown for
its reset handler. Add access handler which emulates writing and reading
PMXEVCNTR register. When reading PMXEVCNTR, call perf_event_read_value
to get the count value of the perf event.
Signed-off-by: Shannon Zhao <shannon.zhao@linaro.org>
---
arch/arm64/kvm/sys_regs.c | 41 +++++++++++++++++++++++++++++++++++++----
1 file changed, 37 insertions(+), 4 deletions(-)
diff --git a/arch/arm64/kvm/sys_regs.c b/arch/arm64/kvm/sys_regs.c
index 59b7bc9..a51c8de 100644
--- a/arch/arm64/kvm/sys_regs.c
+++ b/arch/arm64/kvm/sys_regs.c
@@ -280,6 +280,12 @@ static bool access_pmu_regs(struct kvm_vcpu *vcpu,
if (p->is_write) {
switch (r->reg) {
+ case PMXEVCNTR_EL0: {
+ val = PMEVCNTR0_EL0 + vcpu_sys_reg(vcpu, PMSELR_EL0);
+ vcpu_sys_reg(vcpu, val) =
+ *vcpu_reg(vcpu, p->Rt) & 0xffffffffUL;
+ break;
+ }
case PMXEVTYPER_EL0: {
val = vcpu_sys_reg(vcpu, PMSELR_EL0);
kvm_pmu_set_counter_event_type(vcpu,
@@ -303,7 +309,17 @@ static bool access_pmu_regs(struct kvm_vcpu *vcpu,
break;
}
} else {
- *vcpu_reg(vcpu, p->Rt) = vcpu_sys_reg(vcpu, r->reg);
+ switch (r->reg) {
+ case PMXEVCNTR_EL0: {
+ val = kvm_pmu_get_counter_value(vcpu,
+ vcpu_sys_reg(vcpu, PMSELR_EL0));
+ *vcpu_reg(vcpu, p->Rt) = val;
+ break;
+ }
+ default:
+ *vcpu_reg(vcpu, p->Rt) = vcpu_sys_reg(vcpu, r->reg);
+ break;
+ }
}
return true;
@@ -530,7 +546,7 @@ static const struct sys_reg_desc sys_reg_descs[] = {
access_pmu_regs, reset_unknown, PMXEVTYPER_EL0 },
/* PMXEVCNTR_EL0 */
{ Op0(0b11), Op1(0b011), CRn(0b1001), CRm(0b1101), Op2(0b010),
- trap_raz_wi },
+ access_pmu_regs, reset_unknown, PMXEVCNTR_EL0 },
/* PMUSERENR_EL0 */
{ Op0(0b11), Op1(0b011), CRn(0b1001), CRm(0b1110), Op2(0b000),
trap_raz_wi },
@@ -714,6 +730,12 @@ static bool access_pmu_cp15_regs(struct kvm_vcpu *vcpu,
if (p->is_write) {
switch (r->reg) {
+ case c9_PMXEVCNTR: {
+ val = c14_PMEVCNTR0 + vcpu_cp15(vcpu, c9_PMSELR);
+ vcpu_cp15(vcpu, val) =
+ *vcpu_reg(vcpu, p->Rt) & 0xffffffffUL;
+ break;
+ }
case c9_PMXEVTYPER: {
val = vcpu_cp15(vcpu, c9_PMSELR);
kvm_pmu_set_counter_event_type(vcpu,
@@ -737,7 +759,17 @@ static bool access_pmu_cp15_regs(struct kvm_vcpu *vcpu,
break;
}
} else {
- *vcpu_reg(vcpu, p->Rt) = vcpu_cp15(vcpu, r->reg);
+ switch (r->reg) {
+ case c9_PMXEVCNTR: {
+ val = kvm_pmu_get_counter_value(vcpu,
+ vcpu_cp15(vcpu, c9_PMSELR));
+ *vcpu_reg(vcpu, p->Rt) = val;
+ break;
+ }
+ default:
+ *vcpu_reg(vcpu, p->Rt) = vcpu_cp15(vcpu, r->reg);
+ break;
+ }
}
return true;
@@ -785,7 +817,8 @@ static const struct sys_reg_desc cp15_regs[] = {
{ Op1( 0), CRn( 9), CRm(13), Op2( 0), trap_raz_wi },
{ Op1( 0), CRn( 9), CRm(13), Op2( 1), access_pmu_cp15_regs,
reset_unknown_cp15, c9_PMXEVTYPER },
- { Op1( 0), CRn( 9), CRm(13), Op2( 2), trap_raz_wi },
+ { Op1( 0), CRn( 9), CRm(13), Op2( 2), access_pmu_cp15_regs,
+ reset_unknown_cp15, c9_PMXEVCNTR },
{ Op1( 0), CRn( 9), CRm(14), Op2( 0), trap_raz_wi },
{ Op1( 0), CRn( 9), CRm(14), Op2( 1), trap_raz_wi },
{ Op1( 0), CRn( 9), CRm(14), Op2( 2), trap_raz_wi },
--
2.0.4
^ permalink raw reply related [flat|nested] 113+ messages in thread
* [PATCH v2 10/22] KVM: ARM64: Add reset and access handlers for PMXEVCNTR register
@ 2015-09-11 8:55 ` Shannon Zhao
0 siblings, 0 replies; 113+ messages in thread
From: Shannon Zhao @ 2015-09-11 8:55 UTC (permalink / raw)
To: kvmarm; +Cc: kvm, marc.zyngier, will.deacon, linux-arm-kernel, shannon.zhao
From: Shannon Zhao <shannon.zhao@linaro.org>
Since the reset value of PMXEVCNTR is UNKNOWN, use reset_unknown for
its reset handler. Add access handler which emulates writing and reading
PMXEVCNTR register. When reading PMXEVCNTR, call perf_event_read_value
to get the count value of the perf event.
Signed-off-by: Shannon Zhao <shannon.zhao@linaro.org>
---
arch/arm64/kvm/sys_regs.c | 41 +++++++++++++++++++++++++++++++++++++----
1 file changed, 37 insertions(+), 4 deletions(-)
diff --git a/arch/arm64/kvm/sys_regs.c b/arch/arm64/kvm/sys_regs.c
index 59b7bc9..a51c8de 100644
--- a/arch/arm64/kvm/sys_regs.c
+++ b/arch/arm64/kvm/sys_regs.c
@@ -280,6 +280,12 @@ static bool access_pmu_regs(struct kvm_vcpu *vcpu,
if (p->is_write) {
switch (r->reg) {
+ case PMXEVCNTR_EL0: {
+ val = PMEVCNTR0_EL0 + vcpu_sys_reg(vcpu, PMSELR_EL0);
+ vcpu_sys_reg(vcpu, val) =
+ *vcpu_reg(vcpu, p->Rt) & 0xffffffffUL;
+ break;
+ }
case PMXEVTYPER_EL0: {
val = vcpu_sys_reg(vcpu, PMSELR_EL0);
kvm_pmu_set_counter_event_type(vcpu,
@@ -303,7 +309,17 @@ static bool access_pmu_regs(struct kvm_vcpu *vcpu,
break;
}
} else {
- *vcpu_reg(vcpu, p->Rt) = vcpu_sys_reg(vcpu, r->reg);
+ switch (r->reg) {
+ case PMXEVCNTR_EL0: {
+ val = kvm_pmu_get_counter_value(vcpu,
+ vcpu_sys_reg(vcpu, PMSELR_EL0));
+ *vcpu_reg(vcpu, p->Rt) = val;
+ break;
+ }
+ default:
+ *vcpu_reg(vcpu, p->Rt) = vcpu_sys_reg(vcpu, r->reg);
+ break;
+ }
}
return true;
@@ -530,7 +546,7 @@ static const struct sys_reg_desc sys_reg_descs[] = {
access_pmu_regs, reset_unknown, PMXEVTYPER_EL0 },
/* PMXEVCNTR_EL0 */
{ Op0(0b11), Op1(0b011), CRn(0b1001), CRm(0b1101), Op2(0b010),
- trap_raz_wi },
+ access_pmu_regs, reset_unknown, PMXEVCNTR_EL0 },
/* PMUSERENR_EL0 */
{ Op0(0b11), Op1(0b011), CRn(0b1001), CRm(0b1110), Op2(0b000),
trap_raz_wi },
@@ -714,6 +730,12 @@ static bool access_pmu_cp15_regs(struct kvm_vcpu *vcpu,
if (p->is_write) {
switch (r->reg) {
+ case c9_PMXEVCNTR: {
+ val = c14_PMEVCNTR0 + vcpu_cp15(vcpu, c9_PMSELR);
+ vcpu_cp15(vcpu, val) =
+ *vcpu_reg(vcpu, p->Rt) & 0xffffffffUL;
+ break;
+ }
case c9_PMXEVTYPER: {
val = vcpu_cp15(vcpu, c9_PMSELR);
kvm_pmu_set_counter_event_type(vcpu,
@@ -737,7 +759,17 @@ static bool access_pmu_cp15_regs(struct kvm_vcpu *vcpu,
break;
}
} else {
- *vcpu_reg(vcpu, p->Rt) = vcpu_cp15(vcpu, r->reg);
+ switch (r->reg) {
+ case c9_PMXEVCNTR: {
+ val = kvm_pmu_get_counter_value(vcpu,
+ vcpu_cp15(vcpu, c9_PMSELR));
+ *vcpu_reg(vcpu, p->Rt) = val;
+ break;
+ }
+ default:
+ *vcpu_reg(vcpu, p->Rt) = vcpu_cp15(vcpu, r->reg);
+ break;
+ }
}
return true;
@@ -785,7 +817,8 @@ static const struct sys_reg_desc cp15_regs[] = {
{ Op1( 0), CRn( 9), CRm(13), Op2( 0), trap_raz_wi },
{ Op1( 0), CRn( 9), CRm(13), Op2( 1), access_pmu_cp15_regs,
reset_unknown_cp15, c9_PMXEVTYPER },
- { Op1( 0), CRn( 9), CRm(13), Op2( 2), trap_raz_wi },
+ { Op1( 0), CRn( 9), CRm(13), Op2( 2), access_pmu_cp15_regs,
+ reset_unknown_cp15, c9_PMXEVCNTR },
{ Op1( 0), CRn( 9), CRm(14), Op2( 0), trap_raz_wi },
{ Op1( 0), CRn( 9), CRm(14), Op2( 1), trap_raz_wi },
{ Op1( 0), CRn( 9), CRm(14), Op2( 2), trap_raz_wi },
--
2.0.4
^ permalink raw reply related [flat|nested] 113+ messages in thread
* [PATCH v2 11/22] KVM: ARM64: Add reset and access handlers for PMCCNTR register
2015-09-11 8:54 ` Shannon Zhao
(?)
@ 2015-09-11 8:55 ` Shannon Zhao
-1 siblings, 0 replies; 113+ messages in thread
From: Shannon Zhao @ 2015-09-11 8:55 UTC (permalink / raw)
To: linux-arm-kernel
From: Shannon Zhao <shannon.zhao@linaro.org>
Since the reset value of PMCCNTR is UNKNOWN, use reset_unknown for its
reset handler. Add a new case to emulate reading to PMCCNTR register.
Signed-off-by: Shannon Zhao <shannon.zhao@linaro.org>
---
arch/arm64/kvm/sys_regs.c | 17 +++++++++++++++--
1 file changed, 15 insertions(+), 2 deletions(-)
diff --git a/arch/arm64/kvm/sys_regs.c b/arch/arm64/kvm/sys_regs.c
index a51c8de..f8d7de0 100644
--- a/arch/arm64/kvm/sys_regs.c
+++ b/arch/arm64/kvm/sys_regs.c
@@ -310,6 +310,12 @@ static bool access_pmu_regs(struct kvm_vcpu *vcpu,
}
} else {
switch (r->reg) {
+ case PMCCNTR_EL0: {
+ val = kvm_pmu_get_counter_value(vcpu,
+ ARMV8_MAX_COUNTERS - 1);
+ *vcpu_reg(vcpu, p->Rt) = val;
+ break;
+ }
case PMXEVCNTR_EL0: {
val = kvm_pmu_get_counter_value(vcpu,
vcpu_sys_reg(vcpu, PMSELR_EL0));
@@ -540,7 +546,7 @@ static const struct sys_reg_desc sys_reg_descs[] = {
access_pmu_regs, reset_pmceid, PMCEID1_EL0 },
/* PMCCNTR_EL0 */
{ Op0(0b11), Op1(0b011), CRn(0b1001), CRm(0b1101), Op2(0b000),
- trap_raz_wi },
+ access_pmu_regs, reset_unknown, PMCCNTR_EL0 },
/* PMXEVTYPER_EL0 */
{ Op0(0b11), Op1(0b011), CRn(0b1001), CRm(0b1101), Op2(0b001),
access_pmu_regs, reset_unknown, PMXEVTYPER_EL0 },
@@ -760,6 +766,12 @@ static bool access_pmu_cp15_regs(struct kvm_vcpu *vcpu,
}
} else {
switch (r->reg) {
+ case c9_PMCCNTR: {
+ val = kvm_pmu_get_counter_value(vcpu,
+ ARMV8_MAX_COUNTERS - 1);
+ *vcpu_reg(vcpu, p->Rt) = val;
+ break;
+ }
case c9_PMXEVCNTR: {
val = kvm_pmu_get_counter_value(vcpu,
vcpu_cp15(vcpu, c9_PMSELR));
@@ -814,7 +826,8 @@ static const struct sys_reg_desc cp15_regs[] = {
reset_pmceid, c9_PMCEID0 },
{ Op1( 0), CRn( 9), CRm(12), Op2( 7), access_pmu_cp15_regs,
reset_pmceid, c9_PMCEID1 },
- { Op1( 0), CRn( 9), CRm(13), Op2( 0), trap_raz_wi },
+ { Op1( 0), CRn( 9), CRm(13), Op2( 0), access_pmu_cp15_regs,
+ reset_unknown_cp15, c9_PMCCNTR },
{ Op1( 0), CRn( 9), CRm(13), Op2( 1), access_pmu_cp15_regs,
reset_unknown_cp15, c9_PMXEVTYPER },
{ Op1( 0), CRn( 9), CRm(13), Op2( 2), access_pmu_cp15_regs,
--
2.0.4
^ permalink raw reply related [flat|nested] 113+ messages in thread
* [PATCH v2 11/22] KVM: ARM64: Add reset and access handlers for PMCCNTR register
@ 2015-09-11 8:55 ` Shannon Zhao
0 siblings, 0 replies; 113+ messages in thread
From: Shannon Zhao @ 2015-09-11 8:55 UTC (permalink / raw)
To: kvmarm; +Cc: kvm, marc.zyngier, will.deacon, linux-arm-kernel, shannon.zhao
From: Shannon Zhao <shannon.zhao@linaro.org>
Since the reset value of PMCCNTR is UNKNOWN, use reset_unknown for its
reset handler. Add a new case to emulate reading to PMCCNTR register.
Signed-off-by: Shannon Zhao <shannon.zhao@linaro.org>
---
arch/arm64/kvm/sys_regs.c | 17 +++++++++++++++--
1 file changed, 15 insertions(+), 2 deletions(-)
diff --git a/arch/arm64/kvm/sys_regs.c b/arch/arm64/kvm/sys_regs.c
index a51c8de..f8d7de0 100644
--- a/arch/arm64/kvm/sys_regs.c
+++ b/arch/arm64/kvm/sys_regs.c
@@ -310,6 +310,12 @@ static bool access_pmu_regs(struct kvm_vcpu *vcpu,
}
} else {
switch (r->reg) {
+ case PMCCNTR_EL0: {
+ val = kvm_pmu_get_counter_value(vcpu,
+ ARMV8_MAX_COUNTERS - 1);
+ *vcpu_reg(vcpu, p->Rt) = val;
+ break;
+ }
case PMXEVCNTR_EL0: {
val = kvm_pmu_get_counter_value(vcpu,
vcpu_sys_reg(vcpu, PMSELR_EL0));
@@ -540,7 +546,7 @@ static const struct sys_reg_desc sys_reg_descs[] = {
access_pmu_regs, reset_pmceid, PMCEID1_EL0 },
/* PMCCNTR_EL0 */
{ Op0(0b11), Op1(0b011), CRn(0b1001), CRm(0b1101), Op2(0b000),
- trap_raz_wi },
+ access_pmu_regs, reset_unknown, PMCCNTR_EL0 },
/* PMXEVTYPER_EL0 */
{ Op0(0b11), Op1(0b011), CRn(0b1001), CRm(0b1101), Op2(0b001),
access_pmu_regs, reset_unknown, PMXEVTYPER_EL0 },
@@ -760,6 +766,12 @@ static bool access_pmu_cp15_regs(struct kvm_vcpu *vcpu,
}
} else {
switch (r->reg) {
+ case c9_PMCCNTR: {
+ val = kvm_pmu_get_counter_value(vcpu,
+ ARMV8_MAX_COUNTERS - 1);
+ *vcpu_reg(vcpu, p->Rt) = val;
+ break;
+ }
case c9_PMXEVCNTR: {
val = kvm_pmu_get_counter_value(vcpu,
vcpu_cp15(vcpu, c9_PMSELR));
@@ -814,7 +826,8 @@ static const struct sys_reg_desc cp15_regs[] = {
reset_pmceid, c9_PMCEID0 },
{ Op1( 0), CRn( 9), CRm(12), Op2( 7), access_pmu_cp15_regs,
reset_pmceid, c9_PMCEID1 },
- { Op1( 0), CRn( 9), CRm(13), Op2( 0), trap_raz_wi },
+ { Op1( 0), CRn( 9), CRm(13), Op2( 0), access_pmu_cp15_regs,
+ reset_unknown_cp15, c9_PMCCNTR },
{ Op1( 0), CRn( 9), CRm(13), Op2( 1), access_pmu_cp15_regs,
reset_unknown_cp15, c9_PMXEVTYPER },
{ Op1( 0), CRn( 9), CRm(13), Op2( 2), access_pmu_cp15_regs,
--
2.0.4
^ permalink raw reply related [flat|nested] 113+ messages in thread
* [PATCH v2 11/22] KVM: ARM64: Add reset and access handlers for PMCCNTR register
@ 2015-09-11 8:55 ` Shannon Zhao
0 siblings, 0 replies; 113+ messages in thread
From: Shannon Zhao @ 2015-09-11 8:55 UTC (permalink / raw)
To: kvmarm; +Cc: kvm, marc.zyngier, will.deacon, linux-arm-kernel, shannon.zhao
From: Shannon Zhao <shannon.zhao@linaro.org>
Since the reset value of PMCCNTR is UNKNOWN, use reset_unknown for its
reset handler. Add a new case to emulate reading to PMCCNTR register.
Signed-off-by: Shannon Zhao <shannon.zhao@linaro.org>
---
arch/arm64/kvm/sys_regs.c | 17 +++++++++++++++--
1 file changed, 15 insertions(+), 2 deletions(-)
diff --git a/arch/arm64/kvm/sys_regs.c b/arch/arm64/kvm/sys_regs.c
index a51c8de..f8d7de0 100644
--- a/arch/arm64/kvm/sys_regs.c
+++ b/arch/arm64/kvm/sys_regs.c
@@ -310,6 +310,12 @@ static bool access_pmu_regs(struct kvm_vcpu *vcpu,
}
} else {
switch (r->reg) {
+ case PMCCNTR_EL0: {
+ val = kvm_pmu_get_counter_value(vcpu,
+ ARMV8_MAX_COUNTERS - 1);
+ *vcpu_reg(vcpu, p->Rt) = val;
+ break;
+ }
case PMXEVCNTR_EL0: {
val = kvm_pmu_get_counter_value(vcpu,
vcpu_sys_reg(vcpu, PMSELR_EL0));
@@ -540,7 +546,7 @@ static const struct sys_reg_desc sys_reg_descs[] = {
access_pmu_regs, reset_pmceid, PMCEID1_EL0 },
/* PMCCNTR_EL0 */
{ Op0(0b11), Op1(0b011), CRn(0b1001), CRm(0b1101), Op2(0b000),
- trap_raz_wi },
+ access_pmu_regs, reset_unknown, PMCCNTR_EL0 },
/* PMXEVTYPER_EL0 */
{ Op0(0b11), Op1(0b011), CRn(0b1001), CRm(0b1101), Op2(0b001),
access_pmu_regs, reset_unknown, PMXEVTYPER_EL0 },
@@ -760,6 +766,12 @@ static bool access_pmu_cp15_regs(struct kvm_vcpu *vcpu,
}
} else {
switch (r->reg) {
+ case c9_PMCCNTR: {
+ val = kvm_pmu_get_counter_value(vcpu,
+ ARMV8_MAX_COUNTERS - 1);
+ *vcpu_reg(vcpu, p->Rt) = val;
+ break;
+ }
case c9_PMXEVCNTR: {
val = kvm_pmu_get_counter_value(vcpu,
vcpu_cp15(vcpu, c9_PMSELR));
@@ -814,7 +826,8 @@ static const struct sys_reg_desc cp15_regs[] = {
reset_pmceid, c9_PMCEID0 },
{ Op1( 0), CRn( 9), CRm(12), Op2( 7), access_pmu_cp15_regs,
reset_pmceid, c9_PMCEID1 },
- { Op1( 0), CRn( 9), CRm(13), Op2( 0), trap_raz_wi },
+ { Op1( 0), CRn( 9), CRm(13), Op2( 0), access_pmu_cp15_regs,
+ reset_unknown_cp15, c9_PMCCNTR },
{ Op1( 0), CRn( 9), CRm(13), Op2( 1), access_pmu_cp15_regs,
reset_unknown_cp15, c9_PMXEVTYPER },
{ Op1( 0), CRn( 9), CRm(13), Op2( 2), access_pmu_cp15_regs,
--
2.0.4
^ permalink raw reply related [flat|nested] 113+ messages in thread
* [PATCH v2 12/22] KVM: ARM64: Add reset and access handlers for PMCNTENSET and PMCNTENCLR register
2015-09-11 8:54 ` Shannon Zhao
(?)
@ 2015-09-11 8:55 ` Shannon Zhao
-1 siblings, 0 replies; 113+ messages in thread
From: Shannon Zhao @ 2015-09-11 8:55 UTC (permalink / raw)
To: linux-arm-kernel
From: Shannon Zhao <shannon.zhao@linaro.org>
Since the reset value of PMCNTENSET and PMCNTENCLR is UNKNOWN, use
reset_unknown for its reset handler. Add a new case to emulate writing
PMCNTENSET or PMCNTENCLR register.
When writing to PMCNTENSET, call perf_event_enable to enable the perf
event. When writing to PMCNTENCLR, call perf_event_disable to disable
the perf event.
Signed-off-by: Shannon Zhao <shannon.zhao@linaro.org>
---
arch/arm64/kvm/sys_regs.c | 46 +++++++++++++++++++++++++++++++++++++++++----
include/kvm/arm_pmu.h | 4 ++++
virt/kvm/arm/pmu.c | 48 +++++++++++++++++++++++++++++++++++++++++++++++
3 files changed, 94 insertions(+), 4 deletions(-)
diff --git a/arch/arm64/kvm/sys_regs.c b/arch/arm64/kvm/sys_regs.c
index f8d7de0..8307189 100644
--- a/arch/arm64/kvm/sys_regs.c
+++ b/arch/arm64/kvm/sys_regs.c
@@ -293,6 +293,24 @@ static bool access_pmu_regs(struct kvm_vcpu *vcpu,
val);
break;
}
+ case PMCNTENSET_EL0: {
+ val = *vcpu_reg(vcpu, p->Rt);
+ kvm_pmu_enable_counter(vcpu, val);
+ /*Value 1 of PMCNTENSET_EL0 and PMCNTENCLR_EL0 means
+ * corresponding counter enabled */
+ vcpu_sys_reg(vcpu, r->reg) |= val;
+ vcpu_sys_reg(vcpu, PMCNTENCLR_EL0) |= val;
+ break;
+ }
+ case PMCNTENCLR_EL0: {
+ val = *vcpu_reg(vcpu, p->Rt);
+ kvm_pmu_disable_counter(vcpu, val);
+ /*Value 0 of PMCNTENSET_EL0 and PMCNTENCLR_EL0 means
+ * corresponding counter disabled */
+ vcpu_sys_reg(vcpu, r->reg) &= ~val;
+ vcpu_sys_reg(vcpu, PMCNTENSET_EL0) &= ~val;
+ break;
+ }
case PMCR_EL0: {
/* Only update writeable bits of PMCR */
val = vcpu_sys_reg(vcpu, r->reg);
@@ -525,10 +543,10 @@ static const struct sys_reg_desc sys_reg_descs[] = {
access_pmu_regs, reset_pmcr, PMCR_EL0, },
/* PMCNTENSET_EL0 */
{ Op0(0b11), Op1(0b011), CRn(0b1001), CRm(0b1100), Op2(0b001),
- trap_raz_wi },
+ access_pmu_regs, reset_unknown, PMCNTENSET_EL0 },
/* PMCNTENCLR_EL0 */
{ Op0(0b11), Op1(0b011), CRn(0b1001), CRm(0b1100), Op2(0b010),
- trap_raz_wi },
+ access_pmu_regs, reset_unknown, PMCNTENCLR_EL0 },
/* PMOVSCLR_EL0 */
{ Op0(0b11), Op1(0b011), CRn(0b1001), CRm(0b1100), Op2(0b011),
trap_raz_wi },
@@ -749,6 +767,24 @@ static bool access_pmu_cp15_regs(struct kvm_vcpu *vcpu,
val);
break;
}
+ case c9_PMCNTENSET: {
+ val = *vcpu_reg(vcpu, p->Rt);
+ kvm_pmu_enable_counter(vcpu, val);
+ /*Value 1 of PMCNTENSET_EL0 and PMCNTENCLR_EL0 means
+ * corresponding counter enabled */
+ vcpu_cp15(vcpu, r->reg) |= val;
+ vcpu_cp15(vcpu, c9_PMCNTENCLR) |= val;
+ break;
+ }
+ case c9_PMCNTENCLR: {
+ val = *vcpu_reg(vcpu, p->Rt);
+ kvm_pmu_disable_counter(vcpu, val);
+ /*Value 0 of PMCNTENSET_EL0 and PMCNTENCLR_EL0 means
+ * corresponding counter disabled */
+ vcpu_cp15(vcpu, r->reg) &= ~val;
+ vcpu_cp15(vcpu, c9_PMCNTENSET) &= ~val;
+ break;
+ }
case c9_PMCR: {
/* Only update writeable bits of PMCR */
val = vcpu_cp15(vcpu, r->reg);
@@ -817,8 +853,10 @@ static const struct sys_reg_desc cp15_regs[] = {
/* PMU */
{ Op1( 0), CRn( 9), CRm(12), Op2( 0), access_pmu_cp15_regs,
reset_pmcr, c9_PMCR },
- { Op1( 0), CRn( 9), CRm(12), Op2( 1), trap_raz_wi },
- { Op1( 0), CRn( 9), CRm(12), Op2( 2), trap_raz_wi },
+ { Op1( 0), CRn( 9), CRm(12), Op2( 1), access_pmu_cp15_regs,
+ reset_unknown_cp15, c9_PMCNTENSET },
+ { Op1( 0), CRn( 9), CRm(12), Op2( 2), access_pmu_cp15_regs,
+ reset_unknown_cp15, c9_PMCNTENCLR },
{ Op1( 0), CRn( 9), CRm(12), Op2( 3), trap_raz_wi },
{ Op1( 0), CRn( 9), CRm(12), Op2( 5), access_pmu_cp15_regs,
reset_unknown_cp15, c9_PMSELR },
diff --git a/include/kvm/arm_pmu.h b/include/kvm/arm_pmu.h
index 387ec6f..59e70af 100644
--- a/include/kvm/arm_pmu.h
+++ b/include/kvm/arm_pmu.h
@@ -39,6 +39,8 @@ struct kvm_pmu {
#ifdef CONFIG_KVM_ARM_PMU
unsigned long kvm_pmu_get_counter_value(struct kvm_vcpu *vcpu,
unsigned long select_idx);
+void kvm_pmu_disable_counter(struct kvm_vcpu *vcpu, unsigned long val);
+void kvm_pmu_enable_counter(struct kvm_vcpu *vcpu, unsigned long val);
void kvm_pmu_set_counter_event_type(struct kvm_vcpu *vcpu, unsigned long data,
unsigned long select_idx);
#else
@@ -47,6 +49,8 @@ unsigned long kvm_pmu_get_counter_value(struct kvm_vcpu *vcpu,
{
return 0;
}
+void kvm_pmu_disable_counter(struct kvm_vcpu *vcpu, unsigned long val) {}
+void kvm_pmu_enable_counter(struct kvm_vcpu *vcpu, unsigned long val) {}
void kvm_pmu_set_counter_event_type(struct kvm_vcpu *vcpu, unsigned long data,
unsigned long select_idx) {}
#endif
diff --git a/virt/kvm/arm/pmu.c b/virt/kvm/arm/pmu.c
index 0c7fe5c..c6cdc4e 100644
--- a/virt/kvm/arm/pmu.c
+++ b/virt/kvm/arm/pmu.c
@@ -115,6 +115,54 @@ unsigned long kvm_pmu_get_counter_value(struct kvm_vcpu *vcpu,
}
/**
+ * kvm_pmu_enable_counter - enable selected PMU counter
+ * @vcpu: The vcpu pointer
+ * @val: the value guest writes to PMCNTENSET register
+ *
+ * Call perf_event_enable to start counting the perf event
+ */
+void kvm_pmu_enable_counter(struct kvm_vcpu *vcpu, unsigned long val)
+{
+ int i;
+ struct kvm_pmu *pmu = &vcpu->arch.pmu;
+ struct kvm_pmc *pmc;
+
+ for (i = 0; i < 32; i++) {
+ if ((val >> i) & 0x1) {
+ pmc = &pmu->pmc[i];
+ if (pmc->perf_event) {
+ perf_event_enable(pmc->perf_event);
+ if (pmc->perf_event->state
+ != PERF_EVENT_STATE_ACTIVE)
+ kvm_debug("fail to enable event\n");
+ }
+ }
+ }
+}
+
+/**
+ * kvm_pmu_disable_counter - disable selected PMU counter
+ * @vcpu: The vcpu pointer
+ * @val: the value guest writes to PMCNTENCLR register
+ *
+ * Call perf_event_disable to stop counting the perf event
+ */
+void kvm_pmu_disable_counter(struct kvm_vcpu *vcpu, unsigned long val)
+{
+ int i;
+ struct kvm_pmu *pmu = &vcpu->arch.pmu;
+ struct kvm_pmc *pmc;
+
+ for (i = 0; i < 32; i++) {
+ if ((val >> i) & 0x1) {
+ pmc = &pmu->pmc[i];
+ if (pmc->perf_event)
+ perf_event_disable(pmc->perf_event);
+ }
+ }
+}
+
+/**
* kvm_pmu_find_hw_event - find hardware event
* @pmu: The pmu pointer
* @event_select: The number of selected event type
--
2.0.4
^ permalink raw reply related [flat|nested] 113+ messages in thread
* [PATCH v2 12/22] KVM: ARM64: Add reset and access handlers for PMCNTENSET and PMCNTENCLR register
@ 2015-09-11 8:55 ` Shannon Zhao
0 siblings, 0 replies; 113+ messages in thread
From: Shannon Zhao @ 2015-09-11 8:55 UTC (permalink / raw)
To: kvmarm; +Cc: kvm, marc.zyngier, will.deacon, linux-arm-kernel, shannon.zhao
From: Shannon Zhao <shannon.zhao@linaro.org>
Since the reset value of PMCNTENSET and PMCNTENCLR is UNKNOWN, use
reset_unknown for its reset handler. Add a new case to emulate writing
PMCNTENSET or PMCNTENCLR register.
When writing to PMCNTENSET, call perf_event_enable to enable the perf
event. When writing to PMCNTENCLR, call perf_event_disable to disable
the perf event.
Signed-off-by: Shannon Zhao <shannon.zhao@linaro.org>
---
arch/arm64/kvm/sys_regs.c | 46 +++++++++++++++++++++++++++++++++++++++++----
include/kvm/arm_pmu.h | 4 ++++
virt/kvm/arm/pmu.c | 48 +++++++++++++++++++++++++++++++++++++++++++++++
3 files changed, 94 insertions(+), 4 deletions(-)
diff --git a/arch/arm64/kvm/sys_regs.c b/arch/arm64/kvm/sys_regs.c
index f8d7de0..8307189 100644
--- a/arch/arm64/kvm/sys_regs.c
+++ b/arch/arm64/kvm/sys_regs.c
@@ -293,6 +293,24 @@ static bool access_pmu_regs(struct kvm_vcpu *vcpu,
val);
break;
}
+ case PMCNTENSET_EL0: {
+ val = *vcpu_reg(vcpu, p->Rt);
+ kvm_pmu_enable_counter(vcpu, val);
+ /*Value 1 of PMCNTENSET_EL0 and PMCNTENCLR_EL0 means
+ * corresponding counter enabled */
+ vcpu_sys_reg(vcpu, r->reg) |= val;
+ vcpu_sys_reg(vcpu, PMCNTENCLR_EL0) |= val;
+ break;
+ }
+ case PMCNTENCLR_EL0: {
+ val = *vcpu_reg(vcpu, p->Rt);
+ kvm_pmu_disable_counter(vcpu, val);
+ /*Value 0 of PMCNTENSET_EL0 and PMCNTENCLR_EL0 means
+ * corresponding counter disabled */
+ vcpu_sys_reg(vcpu, r->reg) &= ~val;
+ vcpu_sys_reg(vcpu, PMCNTENSET_EL0) &= ~val;
+ break;
+ }
case PMCR_EL0: {
/* Only update writeable bits of PMCR */
val = vcpu_sys_reg(vcpu, r->reg);
@@ -525,10 +543,10 @@ static const struct sys_reg_desc sys_reg_descs[] = {
access_pmu_regs, reset_pmcr, PMCR_EL0, },
/* PMCNTENSET_EL0 */
{ Op0(0b11), Op1(0b011), CRn(0b1001), CRm(0b1100), Op2(0b001),
- trap_raz_wi },
+ access_pmu_regs, reset_unknown, PMCNTENSET_EL0 },
/* PMCNTENCLR_EL0 */
{ Op0(0b11), Op1(0b011), CRn(0b1001), CRm(0b1100), Op2(0b010),
- trap_raz_wi },
+ access_pmu_regs, reset_unknown, PMCNTENCLR_EL0 },
/* PMOVSCLR_EL0 */
{ Op0(0b11), Op1(0b011), CRn(0b1001), CRm(0b1100), Op2(0b011),
trap_raz_wi },
@@ -749,6 +767,24 @@ static bool access_pmu_cp15_regs(struct kvm_vcpu *vcpu,
val);
break;
}
+ case c9_PMCNTENSET: {
+ val = *vcpu_reg(vcpu, p->Rt);
+ kvm_pmu_enable_counter(vcpu, val);
+ /*Value 1 of PMCNTENSET_EL0 and PMCNTENCLR_EL0 means
+ * corresponding counter enabled */
+ vcpu_cp15(vcpu, r->reg) |= val;
+ vcpu_cp15(vcpu, c9_PMCNTENCLR) |= val;
+ break;
+ }
+ case c9_PMCNTENCLR: {
+ val = *vcpu_reg(vcpu, p->Rt);
+ kvm_pmu_disable_counter(vcpu, val);
+ /*Value 0 of PMCNTENSET_EL0 and PMCNTENCLR_EL0 means
+ * corresponding counter disabled */
+ vcpu_cp15(vcpu, r->reg) &= ~val;
+ vcpu_cp15(vcpu, c9_PMCNTENSET) &= ~val;
+ break;
+ }
case c9_PMCR: {
/* Only update writeable bits of PMCR */
val = vcpu_cp15(vcpu, r->reg);
@@ -817,8 +853,10 @@ static const struct sys_reg_desc cp15_regs[] = {
/* PMU */
{ Op1( 0), CRn( 9), CRm(12), Op2( 0), access_pmu_cp15_regs,
reset_pmcr, c9_PMCR },
- { Op1( 0), CRn( 9), CRm(12), Op2( 1), trap_raz_wi },
- { Op1( 0), CRn( 9), CRm(12), Op2( 2), trap_raz_wi },
+ { Op1( 0), CRn( 9), CRm(12), Op2( 1), access_pmu_cp15_regs,
+ reset_unknown_cp15, c9_PMCNTENSET },
+ { Op1( 0), CRn( 9), CRm(12), Op2( 2), access_pmu_cp15_regs,
+ reset_unknown_cp15, c9_PMCNTENCLR },
{ Op1( 0), CRn( 9), CRm(12), Op2( 3), trap_raz_wi },
{ Op1( 0), CRn( 9), CRm(12), Op2( 5), access_pmu_cp15_regs,
reset_unknown_cp15, c9_PMSELR },
diff --git a/include/kvm/arm_pmu.h b/include/kvm/arm_pmu.h
index 387ec6f..59e70af 100644
--- a/include/kvm/arm_pmu.h
+++ b/include/kvm/arm_pmu.h
@@ -39,6 +39,8 @@ struct kvm_pmu {
#ifdef CONFIG_KVM_ARM_PMU
unsigned long kvm_pmu_get_counter_value(struct kvm_vcpu *vcpu,
unsigned long select_idx);
+void kvm_pmu_disable_counter(struct kvm_vcpu *vcpu, unsigned long val);
+void kvm_pmu_enable_counter(struct kvm_vcpu *vcpu, unsigned long val);
void kvm_pmu_set_counter_event_type(struct kvm_vcpu *vcpu, unsigned long data,
unsigned long select_idx);
#else
@@ -47,6 +49,8 @@ unsigned long kvm_pmu_get_counter_value(struct kvm_vcpu *vcpu,
{
return 0;
}
+void kvm_pmu_disable_counter(struct kvm_vcpu *vcpu, unsigned long val) {}
+void kvm_pmu_enable_counter(struct kvm_vcpu *vcpu, unsigned long val) {}
void kvm_pmu_set_counter_event_type(struct kvm_vcpu *vcpu, unsigned long data,
unsigned long select_idx) {}
#endif
diff --git a/virt/kvm/arm/pmu.c b/virt/kvm/arm/pmu.c
index 0c7fe5c..c6cdc4e 100644
--- a/virt/kvm/arm/pmu.c
+++ b/virt/kvm/arm/pmu.c
@@ -115,6 +115,54 @@ unsigned long kvm_pmu_get_counter_value(struct kvm_vcpu *vcpu,
}
/**
+ * kvm_pmu_enable_counter - enable selected PMU counter
+ * @vcpu: The vcpu pointer
+ * @val: the value guest writes to PMCNTENSET register
+ *
+ * Call perf_event_enable to start counting the perf event
+ */
+void kvm_pmu_enable_counter(struct kvm_vcpu *vcpu, unsigned long val)
+{
+ int i;
+ struct kvm_pmu *pmu = &vcpu->arch.pmu;
+ struct kvm_pmc *pmc;
+
+ for (i = 0; i < 32; i++) {
+ if ((val >> i) & 0x1) {
+ pmc = &pmu->pmc[i];
+ if (pmc->perf_event) {
+ perf_event_enable(pmc->perf_event);
+ if (pmc->perf_event->state
+ != PERF_EVENT_STATE_ACTIVE)
+ kvm_debug("fail to enable event\n");
+ }
+ }
+ }
+}
+
+/**
+ * kvm_pmu_disable_counter - disable selected PMU counter
+ * @vcpu: The vcpu pointer
+ * @val: the value guest writes to PMCNTENCLR register
+ *
+ * Call perf_event_disable to stop counting the perf event
+ */
+void kvm_pmu_disable_counter(struct kvm_vcpu *vcpu, unsigned long val)
+{
+ int i;
+ struct kvm_pmu *pmu = &vcpu->arch.pmu;
+ struct kvm_pmc *pmc;
+
+ for (i = 0; i < 32; i++) {
+ if ((val >> i) & 0x1) {
+ pmc = &pmu->pmc[i];
+ if (pmc->perf_event)
+ perf_event_disable(pmc->perf_event);
+ }
+ }
+}
+
+/**
* kvm_pmu_find_hw_event - find hardware event
* @pmu: The pmu pointer
* @event_select: The number of selected event type
--
2.0.4
^ permalink raw reply related [flat|nested] 113+ messages in thread
* [PATCH v2 12/22] KVM: ARM64: Add reset and access handlers for PMCNTENSET and PMCNTENCLR register
@ 2015-09-11 8:55 ` Shannon Zhao
0 siblings, 0 replies; 113+ messages in thread
From: Shannon Zhao @ 2015-09-11 8:55 UTC (permalink / raw)
To: kvmarm; +Cc: kvm, marc.zyngier, will.deacon, linux-arm-kernel, shannon.zhao
From: Shannon Zhao <shannon.zhao@linaro.org>
Since the reset value of PMCNTENSET and PMCNTENCLR is UNKNOWN, use
reset_unknown for its reset handler. Add a new case to emulate writing
PMCNTENSET or PMCNTENCLR register.
When writing to PMCNTENSET, call perf_event_enable to enable the perf
event. When writing to PMCNTENCLR, call perf_event_disable to disable
the perf event.
Signed-off-by: Shannon Zhao <shannon.zhao@linaro.org>
---
arch/arm64/kvm/sys_regs.c | 46 +++++++++++++++++++++++++++++++++++++++++----
include/kvm/arm_pmu.h | 4 ++++
virt/kvm/arm/pmu.c | 48 +++++++++++++++++++++++++++++++++++++++++++++++
3 files changed, 94 insertions(+), 4 deletions(-)
diff --git a/arch/arm64/kvm/sys_regs.c b/arch/arm64/kvm/sys_regs.c
index f8d7de0..8307189 100644
--- a/arch/arm64/kvm/sys_regs.c
+++ b/arch/arm64/kvm/sys_regs.c
@@ -293,6 +293,24 @@ static bool access_pmu_regs(struct kvm_vcpu *vcpu,
val);
break;
}
+ case PMCNTENSET_EL0: {
+ val = *vcpu_reg(vcpu, p->Rt);
+ kvm_pmu_enable_counter(vcpu, val);
+ /*Value 1 of PMCNTENSET_EL0 and PMCNTENCLR_EL0 means
+ * corresponding counter enabled */
+ vcpu_sys_reg(vcpu, r->reg) |= val;
+ vcpu_sys_reg(vcpu, PMCNTENCLR_EL0) |= val;
+ break;
+ }
+ case PMCNTENCLR_EL0: {
+ val = *vcpu_reg(vcpu, p->Rt);
+ kvm_pmu_disable_counter(vcpu, val);
+ /*Value 0 of PMCNTENSET_EL0 and PMCNTENCLR_EL0 means
+ * corresponding counter disabled */
+ vcpu_sys_reg(vcpu, r->reg) &= ~val;
+ vcpu_sys_reg(vcpu, PMCNTENSET_EL0) &= ~val;
+ break;
+ }
case PMCR_EL0: {
/* Only update writeable bits of PMCR */
val = vcpu_sys_reg(vcpu, r->reg);
@@ -525,10 +543,10 @@ static const struct sys_reg_desc sys_reg_descs[] = {
access_pmu_regs, reset_pmcr, PMCR_EL0, },
/* PMCNTENSET_EL0 */
{ Op0(0b11), Op1(0b011), CRn(0b1001), CRm(0b1100), Op2(0b001),
- trap_raz_wi },
+ access_pmu_regs, reset_unknown, PMCNTENSET_EL0 },
/* PMCNTENCLR_EL0 */
{ Op0(0b11), Op1(0b011), CRn(0b1001), CRm(0b1100), Op2(0b010),
- trap_raz_wi },
+ access_pmu_regs, reset_unknown, PMCNTENCLR_EL0 },
/* PMOVSCLR_EL0 */
{ Op0(0b11), Op1(0b011), CRn(0b1001), CRm(0b1100), Op2(0b011),
trap_raz_wi },
@@ -749,6 +767,24 @@ static bool access_pmu_cp15_regs(struct kvm_vcpu *vcpu,
val);
break;
}
+ case c9_PMCNTENSET: {
+ val = *vcpu_reg(vcpu, p->Rt);
+ kvm_pmu_enable_counter(vcpu, val);
+ /*Value 1 of PMCNTENSET_EL0 and PMCNTENCLR_EL0 means
+ * corresponding counter enabled */
+ vcpu_cp15(vcpu, r->reg) |= val;
+ vcpu_cp15(vcpu, c9_PMCNTENCLR) |= val;
+ break;
+ }
+ case c9_PMCNTENCLR: {
+ val = *vcpu_reg(vcpu, p->Rt);
+ kvm_pmu_disable_counter(vcpu, val);
+ /*Value 0 of PMCNTENSET_EL0 and PMCNTENCLR_EL0 means
+ * corresponding counter disabled */
+ vcpu_cp15(vcpu, r->reg) &= ~val;
+ vcpu_cp15(vcpu, c9_PMCNTENSET) &= ~val;
+ break;
+ }
case c9_PMCR: {
/* Only update writeable bits of PMCR */
val = vcpu_cp15(vcpu, r->reg);
@@ -817,8 +853,10 @@ static const struct sys_reg_desc cp15_regs[] = {
/* PMU */
{ Op1( 0), CRn( 9), CRm(12), Op2( 0), access_pmu_cp15_regs,
reset_pmcr, c9_PMCR },
- { Op1( 0), CRn( 9), CRm(12), Op2( 1), trap_raz_wi },
- { Op1( 0), CRn( 9), CRm(12), Op2( 2), trap_raz_wi },
+ { Op1( 0), CRn( 9), CRm(12), Op2( 1), access_pmu_cp15_regs,
+ reset_unknown_cp15, c9_PMCNTENSET },
+ { Op1( 0), CRn( 9), CRm(12), Op2( 2), access_pmu_cp15_regs,
+ reset_unknown_cp15, c9_PMCNTENCLR },
{ Op1( 0), CRn( 9), CRm(12), Op2( 3), trap_raz_wi },
{ Op1( 0), CRn( 9), CRm(12), Op2( 5), access_pmu_cp15_regs,
reset_unknown_cp15, c9_PMSELR },
diff --git a/include/kvm/arm_pmu.h b/include/kvm/arm_pmu.h
index 387ec6f..59e70af 100644
--- a/include/kvm/arm_pmu.h
+++ b/include/kvm/arm_pmu.h
@@ -39,6 +39,8 @@ struct kvm_pmu {
#ifdef CONFIG_KVM_ARM_PMU
unsigned long kvm_pmu_get_counter_value(struct kvm_vcpu *vcpu,
unsigned long select_idx);
+void kvm_pmu_disable_counter(struct kvm_vcpu *vcpu, unsigned long val);
+void kvm_pmu_enable_counter(struct kvm_vcpu *vcpu, unsigned long val);
void kvm_pmu_set_counter_event_type(struct kvm_vcpu *vcpu, unsigned long data,
unsigned long select_idx);
#else
@@ -47,6 +49,8 @@ unsigned long kvm_pmu_get_counter_value(struct kvm_vcpu *vcpu,
{
return 0;
}
+void kvm_pmu_disable_counter(struct kvm_vcpu *vcpu, unsigned long val) {}
+void kvm_pmu_enable_counter(struct kvm_vcpu *vcpu, unsigned long val) {}
void kvm_pmu_set_counter_event_type(struct kvm_vcpu *vcpu, unsigned long data,
unsigned long select_idx) {}
#endif
diff --git a/virt/kvm/arm/pmu.c b/virt/kvm/arm/pmu.c
index 0c7fe5c..c6cdc4e 100644
--- a/virt/kvm/arm/pmu.c
+++ b/virt/kvm/arm/pmu.c
@@ -115,6 +115,54 @@ unsigned long kvm_pmu_get_counter_value(struct kvm_vcpu *vcpu,
}
/**
+ * kvm_pmu_enable_counter - enable selected PMU counter
+ * @vcpu: The vcpu pointer
+ * @val: the value guest writes to PMCNTENSET register
+ *
+ * Call perf_event_enable to start counting the perf event
+ */
+void kvm_pmu_enable_counter(struct kvm_vcpu *vcpu, unsigned long val)
+{
+ int i;
+ struct kvm_pmu *pmu = &vcpu->arch.pmu;
+ struct kvm_pmc *pmc;
+
+ for (i = 0; i < 32; i++) {
+ if ((val >> i) & 0x1) {
+ pmc = &pmu->pmc[i];
+ if (pmc->perf_event) {
+ perf_event_enable(pmc->perf_event);
+ if (pmc->perf_event->state
+ != PERF_EVENT_STATE_ACTIVE)
+ kvm_debug("fail to enable event\n");
+ }
+ }
+ }
+}
+
+/**
+ * kvm_pmu_disable_counter - disable selected PMU counter
+ * @vcpu: The vcpu pointer
+ * @val: the value guest writes to PMCNTENCLR register
+ *
+ * Call perf_event_disable to stop counting the perf event
+ */
+void kvm_pmu_disable_counter(struct kvm_vcpu *vcpu, unsigned long val)
+{
+ int i;
+ struct kvm_pmu *pmu = &vcpu->arch.pmu;
+ struct kvm_pmc *pmc;
+
+ for (i = 0; i < 32; i++) {
+ if ((val >> i) & 0x1) {
+ pmc = &pmu->pmc[i];
+ if (pmc->perf_event)
+ perf_event_disable(pmc->perf_event);
+ }
+ }
+}
+
+/**
* kvm_pmu_find_hw_event - find hardware event
* @pmu: The pmu pointer
* @event_select: The number of selected event type
--
2.0.4
^ permalink raw reply related [flat|nested] 113+ messages in thread
* [PATCH v2 13/22] KVM: ARM64: Add reset and access handlers for PMINTENSET and PMINTENCLR register
2015-09-11 8:54 ` Shannon Zhao
(?)
@ 2015-09-11 8:55 ` Shannon Zhao
-1 siblings, 0 replies; 113+ messages in thread
From: Shannon Zhao @ 2015-09-11 8:55 UTC (permalink / raw)
To: linux-arm-kernel
From: Shannon Zhao <shannon.zhao@linaro.org>
Since the reset value of PMINTENSET and PMINTENCLR is UNKNOWN, use
reset_unknown for its reset handler. Add a new case to emulate writing
PMINTENSET or PMINTENCLR register.
Signed-off-by: Shannon Zhao <shannon.zhao@linaro.org>
---
arch/arm64/kvm/sys_regs.c | 34 ++++++++++++++++++++++++++++++----
1 file changed, 30 insertions(+), 4 deletions(-)
diff --git a/arch/arm64/kvm/sys_regs.c b/arch/arm64/kvm/sys_regs.c
index 8307189..c3d4fb5 100644
--- a/arch/arm64/kvm/sys_regs.c
+++ b/arch/arm64/kvm/sys_regs.c
@@ -311,6 +311,18 @@ static bool access_pmu_regs(struct kvm_vcpu *vcpu,
vcpu_sys_reg(vcpu, PMCNTENSET_EL0) &= ~val;
break;
}
+ case PMINTENSET_EL1: {
+ val = *vcpu_reg(vcpu, p->Rt);
+ vcpu_sys_reg(vcpu, r->reg) |= val;
+ vcpu_sys_reg(vcpu, PMINTENCLR_EL1) |= val;
+ break;
+ }
+ case PMINTENCLR_EL1: {
+ val = *vcpu_reg(vcpu, p->Rt);
+ vcpu_sys_reg(vcpu, r->reg) &= ~val;
+ vcpu_sys_reg(vcpu, PMINTENSET_EL1) &= ~val;
+ break;
+ }
case PMCR_EL0: {
/* Only update writeable bits of PMCR */
val = vcpu_sys_reg(vcpu, r->reg);
@@ -500,10 +512,10 @@ static const struct sys_reg_desc sys_reg_descs[] = {
/* PMINTENSET_EL1 */
{ Op0(0b11), Op1(0b000), CRn(0b1001), CRm(0b1110), Op2(0b001),
- trap_raz_wi },
+ access_pmu_regs, reset_unknown, PMINTENSET_EL1 },
/* PMINTENCLR_EL1 */
{ Op0(0b11), Op1(0b000), CRn(0b1001), CRm(0b1110), Op2(0b010),
- trap_raz_wi },
+ access_pmu_regs, reset_unknown, PMINTENCLR_EL1 },
/* MAIR_EL1 */
{ Op0(0b11), Op1(0b000), CRn(0b1010), CRm(0b0010), Op2(0b000),
@@ -785,6 +797,18 @@ static bool access_pmu_cp15_regs(struct kvm_vcpu *vcpu,
vcpu_cp15(vcpu, c9_PMCNTENSET) &= ~val;
break;
}
+ case c9_PMINTENSET: {
+ val = *vcpu_reg(vcpu, p->Rt);
+ vcpu_cp15(vcpu, r->reg) |= val;
+ vcpu_cp15(vcpu, c9_PMINTENCLR) |= val;
+ break;
+ }
+ case c9_PMINTENCLR: {
+ val = *vcpu_reg(vcpu, p->Rt);
+ vcpu_cp15(vcpu, r->reg) &= ~val;
+ vcpu_cp15(vcpu, c9_PMINTENSET) &= ~val;
+ break;
+ }
case c9_PMCR: {
/* Only update writeable bits of PMCR */
val = vcpu_cp15(vcpu, r->reg);
@@ -871,8 +895,10 @@ static const struct sys_reg_desc cp15_regs[] = {
{ Op1( 0), CRn( 9), CRm(13), Op2( 2), access_pmu_cp15_regs,
reset_unknown_cp15, c9_PMXEVCNTR },
{ Op1( 0), CRn( 9), CRm(14), Op2( 0), trap_raz_wi },
- { Op1( 0), CRn( 9), CRm(14), Op2( 1), trap_raz_wi },
- { Op1( 0), CRn( 9), CRm(14), Op2( 2), trap_raz_wi },
+ { Op1( 0), CRn( 9), CRm(14), Op2( 1), access_pmu_cp15_regs,
+ reset_unknown_cp15, c9_PMINTENSET },
+ { Op1( 0), CRn( 9), CRm(14), Op2( 2), access_pmu_cp15_regs,
+ reset_unknown_cp15, c9_PMINTENCLR },
{ Op1( 0), CRn(10), CRm( 2), Op2( 0), access_vm_reg, NULL, c10_PRRR },
{ Op1( 0), CRn(10), CRm( 2), Op2( 1), access_vm_reg, NULL, c10_NMRR },
--
2.0.4
^ permalink raw reply related [flat|nested] 113+ messages in thread
* [PATCH v2 13/22] KVM: ARM64: Add reset and access handlers for PMINTENSET and PMINTENCLR register
@ 2015-09-11 8:55 ` Shannon Zhao
0 siblings, 0 replies; 113+ messages in thread
From: Shannon Zhao @ 2015-09-11 8:55 UTC (permalink / raw)
To: kvmarm; +Cc: kvm, marc.zyngier, will.deacon, linux-arm-kernel, shannon.zhao
From: Shannon Zhao <shannon.zhao@linaro.org>
Since the reset value of PMINTENSET and PMINTENCLR is UNKNOWN, use
reset_unknown for its reset handler. Add a new case to emulate writing
PMINTENSET or PMINTENCLR register.
Signed-off-by: Shannon Zhao <shannon.zhao@linaro.org>
---
arch/arm64/kvm/sys_regs.c | 34 ++++++++++++++++++++++++++++++----
1 file changed, 30 insertions(+), 4 deletions(-)
diff --git a/arch/arm64/kvm/sys_regs.c b/arch/arm64/kvm/sys_regs.c
index 8307189..c3d4fb5 100644
--- a/arch/arm64/kvm/sys_regs.c
+++ b/arch/arm64/kvm/sys_regs.c
@@ -311,6 +311,18 @@ static bool access_pmu_regs(struct kvm_vcpu *vcpu,
vcpu_sys_reg(vcpu, PMCNTENSET_EL0) &= ~val;
break;
}
+ case PMINTENSET_EL1: {
+ val = *vcpu_reg(vcpu, p->Rt);
+ vcpu_sys_reg(vcpu, r->reg) |= val;
+ vcpu_sys_reg(vcpu, PMINTENCLR_EL1) |= val;
+ break;
+ }
+ case PMINTENCLR_EL1: {
+ val = *vcpu_reg(vcpu, p->Rt);
+ vcpu_sys_reg(vcpu, r->reg) &= ~val;
+ vcpu_sys_reg(vcpu, PMINTENSET_EL1) &= ~val;
+ break;
+ }
case PMCR_EL0: {
/* Only update writeable bits of PMCR */
val = vcpu_sys_reg(vcpu, r->reg);
@@ -500,10 +512,10 @@ static const struct sys_reg_desc sys_reg_descs[] = {
/* PMINTENSET_EL1 */
{ Op0(0b11), Op1(0b000), CRn(0b1001), CRm(0b1110), Op2(0b001),
- trap_raz_wi },
+ access_pmu_regs, reset_unknown, PMINTENSET_EL1 },
/* PMINTENCLR_EL1 */
{ Op0(0b11), Op1(0b000), CRn(0b1001), CRm(0b1110), Op2(0b010),
- trap_raz_wi },
+ access_pmu_regs, reset_unknown, PMINTENCLR_EL1 },
/* MAIR_EL1 */
{ Op0(0b11), Op1(0b000), CRn(0b1010), CRm(0b0010), Op2(0b000),
@@ -785,6 +797,18 @@ static bool access_pmu_cp15_regs(struct kvm_vcpu *vcpu,
vcpu_cp15(vcpu, c9_PMCNTENSET) &= ~val;
break;
}
+ case c9_PMINTENSET: {
+ val = *vcpu_reg(vcpu, p->Rt);
+ vcpu_cp15(vcpu, r->reg) |= val;
+ vcpu_cp15(vcpu, c9_PMINTENCLR) |= val;
+ break;
+ }
+ case c9_PMINTENCLR: {
+ val = *vcpu_reg(vcpu, p->Rt);
+ vcpu_cp15(vcpu, r->reg) &= ~val;
+ vcpu_cp15(vcpu, c9_PMINTENSET) &= ~val;
+ break;
+ }
case c9_PMCR: {
/* Only update writeable bits of PMCR */
val = vcpu_cp15(vcpu, r->reg);
@@ -871,8 +895,10 @@ static const struct sys_reg_desc cp15_regs[] = {
{ Op1( 0), CRn( 9), CRm(13), Op2( 2), access_pmu_cp15_regs,
reset_unknown_cp15, c9_PMXEVCNTR },
{ Op1( 0), CRn( 9), CRm(14), Op2( 0), trap_raz_wi },
- { Op1( 0), CRn( 9), CRm(14), Op2( 1), trap_raz_wi },
- { Op1( 0), CRn( 9), CRm(14), Op2( 2), trap_raz_wi },
+ { Op1( 0), CRn( 9), CRm(14), Op2( 1), access_pmu_cp15_regs,
+ reset_unknown_cp15, c9_PMINTENSET },
+ { Op1( 0), CRn( 9), CRm(14), Op2( 2), access_pmu_cp15_regs,
+ reset_unknown_cp15, c9_PMINTENCLR },
{ Op1( 0), CRn(10), CRm( 2), Op2( 0), access_vm_reg, NULL, c10_PRRR },
{ Op1( 0), CRn(10), CRm( 2), Op2( 1), access_vm_reg, NULL, c10_NMRR },
--
2.0.4
^ permalink raw reply related [flat|nested] 113+ messages in thread
* [PATCH v2 13/22] KVM: ARM64: Add reset and access handlers for PMINTENSET and PMINTENCLR register
@ 2015-09-11 8:55 ` Shannon Zhao
0 siblings, 0 replies; 113+ messages in thread
From: Shannon Zhao @ 2015-09-11 8:55 UTC (permalink / raw)
To: kvmarm; +Cc: kvm, marc.zyngier, will.deacon, linux-arm-kernel, shannon.zhao
From: Shannon Zhao <shannon.zhao@linaro.org>
Since the reset value of PMINTENSET and PMINTENCLR is UNKNOWN, use
reset_unknown for its reset handler. Add a new case to emulate writing
PMINTENSET or PMINTENCLR register.
Signed-off-by: Shannon Zhao <shannon.zhao@linaro.org>
---
arch/arm64/kvm/sys_regs.c | 34 ++++++++++++++++++++++++++++++----
1 file changed, 30 insertions(+), 4 deletions(-)
diff --git a/arch/arm64/kvm/sys_regs.c b/arch/arm64/kvm/sys_regs.c
index 8307189..c3d4fb5 100644
--- a/arch/arm64/kvm/sys_regs.c
+++ b/arch/arm64/kvm/sys_regs.c
@@ -311,6 +311,18 @@ static bool access_pmu_regs(struct kvm_vcpu *vcpu,
vcpu_sys_reg(vcpu, PMCNTENSET_EL0) &= ~val;
break;
}
+ case PMINTENSET_EL1: {
+ val = *vcpu_reg(vcpu, p->Rt);
+ vcpu_sys_reg(vcpu, r->reg) |= val;
+ vcpu_sys_reg(vcpu, PMINTENCLR_EL1) |= val;
+ break;
+ }
+ case PMINTENCLR_EL1: {
+ val = *vcpu_reg(vcpu, p->Rt);
+ vcpu_sys_reg(vcpu, r->reg) &= ~val;
+ vcpu_sys_reg(vcpu, PMINTENSET_EL1) &= ~val;
+ break;
+ }
case PMCR_EL0: {
/* Only update writeable bits of PMCR */
val = vcpu_sys_reg(vcpu, r->reg);
@@ -500,10 +512,10 @@ static const struct sys_reg_desc sys_reg_descs[] = {
/* PMINTENSET_EL1 */
{ Op0(0b11), Op1(0b000), CRn(0b1001), CRm(0b1110), Op2(0b001),
- trap_raz_wi },
+ access_pmu_regs, reset_unknown, PMINTENSET_EL1 },
/* PMINTENCLR_EL1 */
{ Op0(0b11), Op1(0b000), CRn(0b1001), CRm(0b1110), Op2(0b010),
- trap_raz_wi },
+ access_pmu_regs, reset_unknown, PMINTENCLR_EL1 },
/* MAIR_EL1 */
{ Op0(0b11), Op1(0b000), CRn(0b1010), CRm(0b0010), Op2(0b000),
@@ -785,6 +797,18 @@ static bool access_pmu_cp15_regs(struct kvm_vcpu *vcpu,
vcpu_cp15(vcpu, c9_PMCNTENSET) &= ~val;
break;
}
+ case c9_PMINTENSET: {
+ val = *vcpu_reg(vcpu, p->Rt);
+ vcpu_cp15(vcpu, r->reg) |= val;
+ vcpu_cp15(vcpu, c9_PMINTENCLR) |= val;
+ break;
+ }
+ case c9_PMINTENCLR: {
+ val = *vcpu_reg(vcpu, p->Rt);
+ vcpu_cp15(vcpu, r->reg) &= ~val;
+ vcpu_cp15(vcpu, c9_PMINTENSET) &= ~val;
+ break;
+ }
case c9_PMCR: {
/* Only update writeable bits of PMCR */
val = vcpu_cp15(vcpu, r->reg);
@@ -871,8 +895,10 @@ static const struct sys_reg_desc cp15_regs[] = {
{ Op1( 0), CRn( 9), CRm(13), Op2( 2), access_pmu_cp15_regs,
reset_unknown_cp15, c9_PMXEVCNTR },
{ Op1( 0), CRn( 9), CRm(14), Op2( 0), trap_raz_wi },
- { Op1( 0), CRn( 9), CRm(14), Op2( 1), trap_raz_wi },
- { Op1( 0), CRn( 9), CRm(14), Op2( 2), trap_raz_wi },
+ { Op1( 0), CRn( 9), CRm(14), Op2( 1), access_pmu_cp15_regs,
+ reset_unknown_cp15, c9_PMINTENSET },
+ { Op1( 0), CRn( 9), CRm(14), Op2( 2), access_pmu_cp15_regs,
+ reset_unknown_cp15, c9_PMINTENCLR },
{ Op1( 0), CRn(10), CRm( 2), Op2( 0), access_vm_reg, NULL, c10_PRRR },
{ Op1( 0), CRn(10), CRm( 2), Op2( 1), access_vm_reg, NULL, c10_NMRR },
--
2.0.4
^ permalink raw reply related [flat|nested] 113+ messages in thread
* [PATCH v2 14/22] KVM: ARM64: Add reset and access handlers for PMOVSSET and PMOVSCLR register
2015-09-11 8:54 ` Shannon Zhao
(?)
@ 2015-09-11 8:55 ` Shannon Zhao
-1 siblings, 0 replies; 113+ messages in thread
From: Shannon Zhao @ 2015-09-11 8:55 UTC (permalink / raw)
To: linux-arm-kernel
From: Shannon Zhao <shannon.zhao@linaro.org>
Since the reset value of PMOVSSET and PMOVSCLR is UNKNOWN, use
reset_unknown for its reset handler. Add a new case to emulate writing
PMOVSSET or PMOVSCLR register.
Signed-off-by: Shannon Zhao <shannon.zhao@linaro.org>
---
arch/arm64/kvm/sys_regs.c | 33 ++++++++++++++++++++++++++++++---
1 file changed, 30 insertions(+), 3 deletions(-)
diff --git a/arch/arm64/kvm/sys_regs.c b/arch/arm64/kvm/sys_regs.c
index c3d4fb5..11fc183 100644
--- a/arch/arm64/kvm/sys_regs.c
+++ b/arch/arm64/kvm/sys_regs.c
@@ -323,6 +323,18 @@ static bool access_pmu_regs(struct kvm_vcpu *vcpu,
vcpu_sys_reg(vcpu, PMINTENSET_EL1) &= ~val;
break;
}
+ case PMOVSSET_EL0: {
+ val = *vcpu_reg(vcpu, p->Rt);
+ vcpu_sys_reg(vcpu, r->reg) |= val;
+ vcpu_sys_reg(vcpu, PMOVSCLR_EL0) |= val;
+ break;
+ }
+ case PMOVSCLR_EL0: {
+ val = *vcpu_reg(vcpu, p->Rt);
+ vcpu_sys_reg(vcpu, r->reg) &= ~val;
+ vcpu_sys_reg(vcpu, PMOVSSET_EL0) &= ~val;
+ break;
+ }
case PMCR_EL0: {
/* Only update writeable bits of PMCR */
val = vcpu_sys_reg(vcpu, r->reg);
@@ -561,7 +573,7 @@ static const struct sys_reg_desc sys_reg_descs[] = {
access_pmu_regs, reset_unknown, PMCNTENCLR_EL0 },
/* PMOVSCLR_EL0 */
{ Op0(0b11), Op1(0b011), CRn(0b1001), CRm(0b1100), Op2(0b011),
- trap_raz_wi },
+ access_pmu_regs, reset_unknown, PMOVSCLR_EL0 },
/* PMSWINC_EL0 */
{ Op0(0b11), Op1(0b011), CRn(0b1001), CRm(0b1100), Op2(0b100),
trap_raz_wi },
@@ -588,7 +600,7 @@ static const struct sys_reg_desc sys_reg_descs[] = {
trap_raz_wi },
/* PMOVSSET_EL0 */
{ Op0(0b11), Op1(0b011), CRn(0b1001), CRm(0b1110), Op2(0b011),
- trap_raz_wi },
+ access_pmu_regs, reset_unknown, PMOVSSET_EL0 },
/* TPIDR_EL0 */
{ Op0(0b11), Op1(0b011), CRn(0b1101), CRm(0b0000), Op2(0b010),
@@ -809,6 +821,18 @@ static bool access_pmu_cp15_regs(struct kvm_vcpu *vcpu,
vcpu_cp15(vcpu, c9_PMINTENSET) &= ~val;
break;
}
+ case c9_PMOVSSET: {
+ val = *vcpu_reg(vcpu, p->Rt);
+ vcpu_cp15(vcpu, r->reg) |= val;
+ vcpu_cp15(vcpu, c9_PMOVSCLR) |= val;
+ break;
+ }
+ case c9_PMOVSCLR: {
+ val = *vcpu_reg(vcpu, p->Rt);
+ vcpu_cp15(vcpu, r->reg) &= ~val;
+ vcpu_cp15(vcpu, c9_PMOVSSET) &= ~val;
+ break;
+ }
case c9_PMCR: {
/* Only update writeable bits of PMCR */
val = vcpu_cp15(vcpu, r->reg);
@@ -881,7 +905,8 @@ static const struct sys_reg_desc cp15_regs[] = {
reset_unknown_cp15, c9_PMCNTENSET },
{ Op1( 0), CRn( 9), CRm(12), Op2( 2), access_pmu_cp15_regs,
reset_unknown_cp15, c9_PMCNTENCLR },
- { Op1( 0), CRn( 9), CRm(12), Op2( 3), trap_raz_wi },
+ { Op1( 0), CRn( 9), CRm(12), Op2( 3), access_pmu_cp15_regs,
+ reset_unknown_cp15, c9_PMOVSCLR },
{ Op1( 0), CRn( 9), CRm(12), Op2( 5), access_pmu_cp15_regs,
reset_unknown_cp15, c9_PMSELR },
{ Op1( 0), CRn( 9), CRm(12), Op2( 6), access_pmu_cp15_regs,
@@ -899,6 +924,8 @@ static const struct sys_reg_desc cp15_regs[] = {
reset_unknown_cp15, c9_PMINTENSET },
{ Op1( 0), CRn( 9), CRm(14), Op2( 2), access_pmu_cp15_regs,
reset_unknown_cp15, c9_PMINTENCLR },
+ { Op1( 0), CRn( 9), CRm(14), Op2( 3), access_pmu_cp15_regs,
+ reset_unknown_cp15, c9_PMOVSSET },
{ Op1( 0), CRn(10), CRm( 2), Op2( 0), access_vm_reg, NULL, c10_PRRR },
{ Op1( 0), CRn(10), CRm( 2), Op2( 1), access_vm_reg, NULL, c10_NMRR },
--
2.0.4
^ permalink raw reply related [flat|nested] 113+ messages in thread
* [PATCH v2 14/22] KVM: ARM64: Add reset and access handlers for PMOVSSET and PMOVSCLR register
@ 2015-09-11 8:55 ` Shannon Zhao
0 siblings, 0 replies; 113+ messages in thread
From: Shannon Zhao @ 2015-09-11 8:55 UTC (permalink / raw)
To: kvmarm; +Cc: kvm, marc.zyngier, will.deacon, linux-arm-kernel, shannon.zhao
From: Shannon Zhao <shannon.zhao@linaro.org>
Since the reset value of PMOVSSET and PMOVSCLR is UNKNOWN, use
reset_unknown for its reset handler. Add a new case to emulate writing
PMOVSSET or PMOVSCLR register.
Signed-off-by: Shannon Zhao <shannon.zhao@linaro.org>
---
arch/arm64/kvm/sys_regs.c | 33 ++++++++++++++++++++++++++++++---
1 file changed, 30 insertions(+), 3 deletions(-)
diff --git a/arch/arm64/kvm/sys_regs.c b/arch/arm64/kvm/sys_regs.c
index c3d4fb5..11fc183 100644
--- a/arch/arm64/kvm/sys_regs.c
+++ b/arch/arm64/kvm/sys_regs.c
@@ -323,6 +323,18 @@ static bool access_pmu_regs(struct kvm_vcpu *vcpu,
vcpu_sys_reg(vcpu, PMINTENSET_EL1) &= ~val;
break;
}
+ case PMOVSSET_EL0: {
+ val = *vcpu_reg(vcpu, p->Rt);
+ vcpu_sys_reg(vcpu, r->reg) |= val;
+ vcpu_sys_reg(vcpu, PMOVSCLR_EL0) |= val;
+ break;
+ }
+ case PMOVSCLR_EL0: {
+ val = *vcpu_reg(vcpu, p->Rt);
+ vcpu_sys_reg(vcpu, r->reg) &= ~val;
+ vcpu_sys_reg(vcpu, PMOVSSET_EL0) &= ~val;
+ break;
+ }
case PMCR_EL0: {
/* Only update writeable bits of PMCR */
val = vcpu_sys_reg(vcpu, r->reg);
@@ -561,7 +573,7 @@ static const struct sys_reg_desc sys_reg_descs[] = {
access_pmu_regs, reset_unknown, PMCNTENCLR_EL0 },
/* PMOVSCLR_EL0 */
{ Op0(0b11), Op1(0b011), CRn(0b1001), CRm(0b1100), Op2(0b011),
- trap_raz_wi },
+ access_pmu_regs, reset_unknown, PMOVSCLR_EL0 },
/* PMSWINC_EL0 */
{ Op0(0b11), Op1(0b011), CRn(0b1001), CRm(0b1100), Op2(0b100),
trap_raz_wi },
@@ -588,7 +600,7 @@ static const struct sys_reg_desc sys_reg_descs[] = {
trap_raz_wi },
/* PMOVSSET_EL0 */
{ Op0(0b11), Op1(0b011), CRn(0b1001), CRm(0b1110), Op2(0b011),
- trap_raz_wi },
+ access_pmu_regs, reset_unknown, PMOVSSET_EL0 },
/* TPIDR_EL0 */
{ Op0(0b11), Op1(0b011), CRn(0b1101), CRm(0b0000), Op2(0b010),
@@ -809,6 +821,18 @@ static bool access_pmu_cp15_regs(struct kvm_vcpu *vcpu,
vcpu_cp15(vcpu, c9_PMINTENSET) &= ~val;
break;
}
+ case c9_PMOVSSET: {
+ val = *vcpu_reg(vcpu, p->Rt);
+ vcpu_cp15(vcpu, r->reg) |= val;
+ vcpu_cp15(vcpu, c9_PMOVSCLR) |= val;
+ break;
+ }
+ case c9_PMOVSCLR: {
+ val = *vcpu_reg(vcpu, p->Rt);
+ vcpu_cp15(vcpu, r->reg) &= ~val;
+ vcpu_cp15(vcpu, c9_PMOVSSET) &= ~val;
+ break;
+ }
case c9_PMCR: {
/* Only update writeable bits of PMCR */
val = vcpu_cp15(vcpu, r->reg);
@@ -881,7 +905,8 @@ static const struct sys_reg_desc cp15_regs[] = {
reset_unknown_cp15, c9_PMCNTENSET },
{ Op1( 0), CRn( 9), CRm(12), Op2( 2), access_pmu_cp15_regs,
reset_unknown_cp15, c9_PMCNTENCLR },
- { Op1( 0), CRn( 9), CRm(12), Op2( 3), trap_raz_wi },
+ { Op1( 0), CRn( 9), CRm(12), Op2( 3), access_pmu_cp15_regs,
+ reset_unknown_cp15, c9_PMOVSCLR },
{ Op1( 0), CRn( 9), CRm(12), Op2( 5), access_pmu_cp15_regs,
reset_unknown_cp15, c9_PMSELR },
{ Op1( 0), CRn( 9), CRm(12), Op2( 6), access_pmu_cp15_regs,
@@ -899,6 +924,8 @@ static const struct sys_reg_desc cp15_regs[] = {
reset_unknown_cp15, c9_PMINTENSET },
{ Op1( 0), CRn( 9), CRm(14), Op2( 2), access_pmu_cp15_regs,
reset_unknown_cp15, c9_PMINTENCLR },
+ { Op1( 0), CRn( 9), CRm(14), Op2( 3), access_pmu_cp15_regs,
+ reset_unknown_cp15, c9_PMOVSSET },
{ Op1( 0), CRn(10), CRm( 2), Op2( 0), access_vm_reg, NULL, c10_PRRR },
{ Op1( 0), CRn(10), CRm( 2), Op2( 1), access_vm_reg, NULL, c10_NMRR },
--
2.0.4
^ permalink raw reply related [flat|nested] 113+ messages in thread
* [PATCH v2 14/22] KVM: ARM64: Add reset and access handlers for PMOVSSET and PMOVSCLR register
@ 2015-09-11 8:55 ` Shannon Zhao
0 siblings, 0 replies; 113+ messages in thread
From: Shannon Zhao @ 2015-09-11 8:55 UTC (permalink / raw)
To: kvmarm; +Cc: kvm, marc.zyngier, will.deacon, linux-arm-kernel, shannon.zhao
From: Shannon Zhao <shannon.zhao@linaro.org>
Since the reset value of PMOVSSET and PMOVSCLR is UNKNOWN, use
reset_unknown for its reset handler. Add a new case to emulate writing
PMOVSSET or PMOVSCLR register.
Signed-off-by: Shannon Zhao <shannon.zhao@linaro.org>
---
arch/arm64/kvm/sys_regs.c | 33 ++++++++++++++++++++++++++++++---
1 file changed, 30 insertions(+), 3 deletions(-)
diff --git a/arch/arm64/kvm/sys_regs.c b/arch/arm64/kvm/sys_regs.c
index c3d4fb5..11fc183 100644
--- a/arch/arm64/kvm/sys_regs.c
+++ b/arch/arm64/kvm/sys_regs.c
@@ -323,6 +323,18 @@ static bool access_pmu_regs(struct kvm_vcpu *vcpu,
vcpu_sys_reg(vcpu, PMINTENSET_EL1) &= ~val;
break;
}
+ case PMOVSSET_EL0: {
+ val = *vcpu_reg(vcpu, p->Rt);
+ vcpu_sys_reg(vcpu, r->reg) |= val;
+ vcpu_sys_reg(vcpu, PMOVSCLR_EL0) |= val;
+ break;
+ }
+ case PMOVSCLR_EL0: {
+ val = *vcpu_reg(vcpu, p->Rt);
+ vcpu_sys_reg(vcpu, r->reg) &= ~val;
+ vcpu_sys_reg(vcpu, PMOVSSET_EL0) &= ~val;
+ break;
+ }
case PMCR_EL0: {
/* Only update writeable bits of PMCR */
val = vcpu_sys_reg(vcpu, r->reg);
@@ -561,7 +573,7 @@ static const struct sys_reg_desc sys_reg_descs[] = {
access_pmu_regs, reset_unknown, PMCNTENCLR_EL0 },
/* PMOVSCLR_EL0 */
{ Op0(0b11), Op1(0b011), CRn(0b1001), CRm(0b1100), Op2(0b011),
- trap_raz_wi },
+ access_pmu_regs, reset_unknown, PMOVSCLR_EL0 },
/* PMSWINC_EL0 */
{ Op0(0b11), Op1(0b011), CRn(0b1001), CRm(0b1100), Op2(0b100),
trap_raz_wi },
@@ -588,7 +600,7 @@ static const struct sys_reg_desc sys_reg_descs[] = {
trap_raz_wi },
/* PMOVSSET_EL0 */
{ Op0(0b11), Op1(0b011), CRn(0b1001), CRm(0b1110), Op2(0b011),
- trap_raz_wi },
+ access_pmu_regs, reset_unknown, PMOVSSET_EL0 },
/* TPIDR_EL0 */
{ Op0(0b11), Op1(0b011), CRn(0b1101), CRm(0b0000), Op2(0b010),
@@ -809,6 +821,18 @@ static bool access_pmu_cp15_regs(struct kvm_vcpu *vcpu,
vcpu_cp15(vcpu, c9_PMINTENSET) &= ~val;
break;
}
+ case c9_PMOVSSET: {
+ val = *vcpu_reg(vcpu, p->Rt);
+ vcpu_cp15(vcpu, r->reg) |= val;
+ vcpu_cp15(vcpu, c9_PMOVSCLR) |= val;
+ break;
+ }
+ case c9_PMOVSCLR: {
+ val = *vcpu_reg(vcpu, p->Rt);
+ vcpu_cp15(vcpu, r->reg) &= ~val;
+ vcpu_cp15(vcpu, c9_PMOVSSET) &= ~val;
+ break;
+ }
case c9_PMCR: {
/* Only update writeable bits of PMCR */
val = vcpu_cp15(vcpu, r->reg);
@@ -881,7 +905,8 @@ static const struct sys_reg_desc cp15_regs[] = {
reset_unknown_cp15, c9_PMCNTENSET },
{ Op1( 0), CRn( 9), CRm(12), Op2( 2), access_pmu_cp15_regs,
reset_unknown_cp15, c9_PMCNTENCLR },
- { Op1( 0), CRn( 9), CRm(12), Op2( 3), trap_raz_wi },
+ { Op1( 0), CRn( 9), CRm(12), Op2( 3), access_pmu_cp15_regs,
+ reset_unknown_cp15, c9_PMOVSCLR },
{ Op1( 0), CRn( 9), CRm(12), Op2( 5), access_pmu_cp15_regs,
reset_unknown_cp15, c9_PMSELR },
{ Op1( 0), CRn( 9), CRm(12), Op2( 6), access_pmu_cp15_regs,
@@ -899,6 +924,8 @@ static const struct sys_reg_desc cp15_regs[] = {
reset_unknown_cp15, c9_PMINTENSET },
{ Op1( 0), CRn( 9), CRm(14), Op2( 2), access_pmu_cp15_regs,
reset_unknown_cp15, c9_PMINTENCLR },
+ { Op1( 0), CRn( 9), CRm(14), Op2( 3), access_pmu_cp15_regs,
+ reset_unknown_cp15, c9_PMOVSSET },
{ Op1( 0), CRn(10), CRm( 2), Op2( 0), access_vm_reg, NULL, c10_PRRR },
{ Op1( 0), CRn(10), CRm( 2), Op2( 1), access_vm_reg, NULL, c10_NMRR },
--
2.0.4
^ permalink raw reply related [flat|nested] 113+ messages in thread
* [PATCH v2 15/22] KVM: ARM64: Add a helper for CP15 registers reset to specified value
2015-09-11 8:54 ` Shannon Zhao
(?)
@ 2015-09-11 8:55 ` Shannon Zhao
-1 siblings, 0 replies; 113+ messages in thread
From: Shannon Zhao @ 2015-09-11 8:55 UTC (permalink / raw)
To: linux-arm-kernel
From: Shannon Zhao <shannon.zhao@linaro.org>
Signed-off-by: Shannon Zhao <shannon.zhao@linaro.org>
---
arch/arm64/kvm/sys_regs.h | 8 ++++++++
1 file changed, 8 insertions(+)
diff --git a/arch/arm64/kvm/sys_regs.h b/arch/arm64/kvm/sys_regs.h
index a0b3811..a476b1b 100644
--- a/arch/arm64/kvm/sys_regs.h
+++ b/arch/arm64/kvm/sys_regs.h
@@ -119,6 +119,14 @@ static inline void reset_val(struct kvm_vcpu *vcpu, const struct sys_reg_desc *r
vcpu_sys_reg(vcpu, r->reg) = r->val;
}
+static inline void reset_val_cp15(struct kvm_vcpu *vcpu,
+ const struct sys_reg_desc *r)
+{
+ BUG_ON(!r->reg);
+ BUG_ON(r->reg >= NR_SYS_REGS);
+ vcpu_cp15(vcpu, r->reg) = r->val;
+}
+
static inline int cmp_sys_reg(const struct sys_reg_desc *i1,
const struct sys_reg_desc *i2)
{
--
2.0.4
^ permalink raw reply related [flat|nested] 113+ messages in thread
* [PATCH v2 15/22] KVM: ARM64: Add a helper for CP15 registers reset to specified value
@ 2015-09-11 8:55 ` Shannon Zhao
0 siblings, 0 replies; 113+ messages in thread
From: Shannon Zhao @ 2015-09-11 8:55 UTC (permalink / raw)
To: kvmarm; +Cc: kvm, marc.zyngier, will.deacon, linux-arm-kernel, shannon.zhao
From: Shannon Zhao <shannon.zhao@linaro.org>
Signed-off-by: Shannon Zhao <shannon.zhao@linaro.org>
---
arch/arm64/kvm/sys_regs.h | 8 ++++++++
1 file changed, 8 insertions(+)
diff --git a/arch/arm64/kvm/sys_regs.h b/arch/arm64/kvm/sys_regs.h
index a0b3811..a476b1b 100644
--- a/arch/arm64/kvm/sys_regs.h
+++ b/arch/arm64/kvm/sys_regs.h
@@ -119,6 +119,14 @@ static inline void reset_val(struct kvm_vcpu *vcpu, const struct sys_reg_desc *r
vcpu_sys_reg(vcpu, r->reg) = r->val;
}
+static inline void reset_val_cp15(struct kvm_vcpu *vcpu,
+ const struct sys_reg_desc *r)
+{
+ BUG_ON(!r->reg);
+ BUG_ON(r->reg >= NR_SYS_REGS);
+ vcpu_cp15(vcpu, r->reg) = r->val;
+}
+
static inline int cmp_sys_reg(const struct sys_reg_desc *i1,
const struct sys_reg_desc *i2)
{
--
2.0.4
^ permalink raw reply related [flat|nested] 113+ messages in thread
* [PATCH v2 15/22] KVM: ARM64: Add a helper for CP15 registers reset to specified value
@ 2015-09-11 8:55 ` Shannon Zhao
0 siblings, 0 replies; 113+ messages in thread
From: Shannon Zhao @ 2015-09-11 8:55 UTC (permalink / raw)
To: kvmarm; +Cc: kvm, marc.zyngier, will.deacon, linux-arm-kernel, shannon.zhao
From: Shannon Zhao <shannon.zhao@linaro.org>
Signed-off-by: Shannon Zhao <shannon.zhao@linaro.org>
---
arch/arm64/kvm/sys_regs.h | 8 ++++++++
1 file changed, 8 insertions(+)
diff --git a/arch/arm64/kvm/sys_regs.h b/arch/arm64/kvm/sys_regs.h
index a0b3811..a476b1b 100644
--- a/arch/arm64/kvm/sys_regs.h
+++ b/arch/arm64/kvm/sys_regs.h
@@ -119,6 +119,14 @@ static inline void reset_val(struct kvm_vcpu *vcpu, const struct sys_reg_desc *r
vcpu_sys_reg(vcpu, r->reg) = r->val;
}
+static inline void reset_val_cp15(struct kvm_vcpu *vcpu,
+ const struct sys_reg_desc *r)
+{
+ BUG_ON(!r->reg);
+ BUG_ON(r->reg >= NR_SYS_REGS);
+ vcpu_cp15(vcpu, r->reg) = r->val;
+}
+
static inline int cmp_sys_reg(const struct sys_reg_desc *i1,
const struct sys_reg_desc *i2)
{
--
2.0.4
^ permalink raw reply related [flat|nested] 113+ messages in thread
* [PATCH v2 16/22] KVM: ARM64: Add reset and access handlers for PMUSERENR register
2015-09-11 8:54 ` Shannon Zhao
(?)
@ 2015-09-11 8:55 ` Shannon Zhao
-1 siblings, 0 replies; 113+ messages in thread
From: Shannon Zhao @ 2015-09-11 8:55 UTC (permalink / raw)
To: linux-arm-kernel
From: Shannon Zhao <shannon.zhao@linaro.org>
Since the reset value of PMUSERENR is zero, use reset_val(_cp15) with
zero for its reset handler.
Signed-off-by: Shannon Zhao <shannon.zhao@linaro.org>
---
arch/arm64/kvm/sys_regs.c | 5 +++--
1 file changed, 3 insertions(+), 2 deletions(-)
diff --git a/arch/arm64/kvm/sys_regs.c b/arch/arm64/kvm/sys_regs.c
index 11fc183..b47cd0b 100644
--- a/arch/arm64/kvm/sys_regs.c
+++ b/arch/arm64/kvm/sys_regs.c
@@ -597,7 +597,7 @@ static const struct sys_reg_desc sys_reg_descs[] = {
access_pmu_regs, reset_unknown, PMXEVCNTR_EL0 },
/* PMUSERENR_EL0 */
{ Op0(0b11), Op1(0b011), CRn(0b1001), CRm(0b1110), Op2(0b000),
- trap_raz_wi },
+ access_pmu_regs, reset_unknown, PMUSERENR_EL0 },
/* PMOVSSET_EL0 */
{ Op0(0b11), Op1(0b011), CRn(0b1001), CRm(0b1110), Op2(0b011),
access_pmu_regs, reset_unknown, PMOVSSET_EL0 },
@@ -919,7 +919,8 @@ static const struct sys_reg_desc cp15_regs[] = {
reset_unknown_cp15, c9_PMXEVTYPER },
{ Op1( 0), CRn( 9), CRm(13), Op2( 2), access_pmu_cp15_regs,
reset_unknown_cp15, c9_PMXEVCNTR },
- { Op1( 0), CRn( 9), CRm(14), Op2( 0), trap_raz_wi },
+ { Op1( 0), CRn( 9), CRm(14), Op2( 0), access_pmu_cp15_regs,
+ reset_val_cp15, c9_PMUSERENR, 0 },
{ Op1( 0), CRn( 9), CRm(14), Op2( 1), access_pmu_cp15_regs,
reset_unknown_cp15, c9_PMINTENSET },
{ Op1( 0), CRn( 9), CRm(14), Op2( 2), access_pmu_cp15_regs,
--
2.0.4
^ permalink raw reply related [flat|nested] 113+ messages in thread
* [PATCH v2 16/22] KVM: ARM64: Add reset and access handlers for PMUSERENR register
@ 2015-09-11 8:55 ` Shannon Zhao
0 siblings, 0 replies; 113+ messages in thread
From: Shannon Zhao @ 2015-09-11 8:55 UTC (permalink / raw)
To: kvmarm; +Cc: kvm, marc.zyngier, will.deacon, linux-arm-kernel, shannon.zhao
From: Shannon Zhao <shannon.zhao@linaro.org>
Since the reset value of PMUSERENR is zero, use reset_val(_cp15) with
zero for its reset handler.
Signed-off-by: Shannon Zhao <shannon.zhao@linaro.org>
---
arch/arm64/kvm/sys_regs.c | 5 +++--
1 file changed, 3 insertions(+), 2 deletions(-)
diff --git a/arch/arm64/kvm/sys_regs.c b/arch/arm64/kvm/sys_regs.c
index 11fc183..b47cd0b 100644
--- a/arch/arm64/kvm/sys_regs.c
+++ b/arch/arm64/kvm/sys_regs.c
@@ -597,7 +597,7 @@ static const struct sys_reg_desc sys_reg_descs[] = {
access_pmu_regs, reset_unknown, PMXEVCNTR_EL0 },
/* PMUSERENR_EL0 */
{ Op0(0b11), Op1(0b011), CRn(0b1001), CRm(0b1110), Op2(0b000),
- trap_raz_wi },
+ access_pmu_regs, reset_unknown, PMUSERENR_EL0 },
/* PMOVSSET_EL0 */
{ Op0(0b11), Op1(0b011), CRn(0b1001), CRm(0b1110), Op2(0b011),
access_pmu_regs, reset_unknown, PMOVSSET_EL0 },
@@ -919,7 +919,8 @@ static const struct sys_reg_desc cp15_regs[] = {
reset_unknown_cp15, c9_PMXEVTYPER },
{ Op1( 0), CRn( 9), CRm(13), Op2( 2), access_pmu_cp15_regs,
reset_unknown_cp15, c9_PMXEVCNTR },
- { Op1( 0), CRn( 9), CRm(14), Op2( 0), trap_raz_wi },
+ { Op1( 0), CRn( 9), CRm(14), Op2( 0), access_pmu_cp15_regs,
+ reset_val_cp15, c9_PMUSERENR, 0 },
{ Op1( 0), CRn( 9), CRm(14), Op2( 1), access_pmu_cp15_regs,
reset_unknown_cp15, c9_PMINTENSET },
{ Op1( 0), CRn( 9), CRm(14), Op2( 2), access_pmu_cp15_regs,
--
2.0.4
^ permalink raw reply related [flat|nested] 113+ messages in thread
* [PATCH v2 16/22] KVM: ARM64: Add reset and access handlers for PMUSERENR register
@ 2015-09-11 8:55 ` Shannon Zhao
0 siblings, 0 replies; 113+ messages in thread
From: Shannon Zhao @ 2015-09-11 8:55 UTC (permalink / raw)
To: kvmarm; +Cc: kvm, marc.zyngier, will.deacon, linux-arm-kernel, shannon.zhao
From: Shannon Zhao <shannon.zhao@linaro.org>
Since the reset value of PMUSERENR is zero, use reset_val(_cp15) with
zero for its reset handler.
Signed-off-by: Shannon Zhao <shannon.zhao@linaro.org>
---
arch/arm64/kvm/sys_regs.c | 5 +++--
1 file changed, 3 insertions(+), 2 deletions(-)
diff --git a/arch/arm64/kvm/sys_regs.c b/arch/arm64/kvm/sys_regs.c
index 11fc183..b47cd0b 100644
--- a/arch/arm64/kvm/sys_regs.c
+++ b/arch/arm64/kvm/sys_regs.c
@@ -597,7 +597,7 @@ static const struct sys_reg_desc sys_reg_descs[] = {
access_pmu_regs, reset_unknown, PMXEVCNTR_EL0 },
/* PMUSERENR_EL0 */
{ Op0(0b11), Op1(0b011), CRn(0b1001), CRm(0b1110), Op2(0b000),
- trap_raz_wi },
+ access_pmu_regs, reset_unknown, PMUSERENR_EL0 },
/* PMOVSSET_EL0 */
{ Op0(0b11), Op1(0b011), CRn(0b1001), CRm(0b1110), Op2(0b011),
access_pmu_regs, reset_unknown, PMOVSSET_EL0 },
@@ -919,7 +919,8 @@ static const struct sys_reg_desc cp15_regs[] = {
reset_unknown_cp15, c9_PMXEVTYPER },
{ Op1( 0), CRn( 9), CRm(13), Op2( 2), access_pmu_cp15_regs,
reset_unknown_cp15, c9_PMXEVCNTR },
- { Op1( 0), CRn( 9), CRm(14), Op2( 0), trap_raz_wi },
+ { Op1( 0), CRn( 9), CRm(14), Op2( 0), access_pmu_cp15_regs,
+ reset_val_cp15, c9_PMUSERENR, 0 },
{ Op1( 0), CRn( 9), CRm(14), Op2( 1), access_pmu_cp15_regs,
reset_unknown_cp15, c9_PMINTENSET },
{ Op1( 0), CRn( 9), CRm(14), Op2( 2), access_pmu_cp15_regs,
--
2.0.4
^ permalink raw reply related [flat|nested] 113+ messages in thread
* [PATCH v2 17/22] KVM: ARM64: Add reset and access handlers for PMSWINC register
2015-09-11 8:54 ` Shannon Zhao
(?)
@ 2015-09-11 8:55 ` Shannon Zhao
-1 siblings, 0 replies; 113+ messages in thread
From: Shannon Zhao @ 2015-09-11 8:55 UTC (permalink / raw)
To: linux-arm-kernel
From: Shannon Zhao <shannon.zhao@linaro.org>
Add access handler which emulates writing and reading PMSWINC
register and add support for creating software increment event.
Signed-off-by: Shannon Zhao <shannon.zhao@linaro.org>
---
arch/arm64/kvm/sys_regs.c | 18 +++++++++++++++++-
include/kvm/arm_pmu.h | 2 ++
virt/kvm/arm/pmu.c | 33 +++++++++++++++++++++++++++++++++
3 files changed, 52 insertions(+), 1 deletion(-)
diff --git a/arch/arm64/kvm/sys_regs.c b/arch/arm64/kvm/sys_regs.c
index b47cd0b..24d00a0 100644
--- a/arch/arm64/kvm/sys_regs.c
+++ b/arch/arm64/kvm/sys_regs.c
@@ -335,6 +335,11 @@ static bool access_pmu_regs(struct kvm_vcpu *vcpu,
vcpu_sys_reg(vcpu, PMOVSSET_EL0) &= ~val;
break;
}
+ case PMSWINC_EL0: {
+ val = *vcpu_reg(vcpu, p->Rt);
+ kvm_pmu_software_increment(vcpu, val);
+ break;
+ }
case PMCR_EL0: {
/* Only update writeable bits of PMCR */
val = vcpu_sys_reg(vcpu, r->reg);
@@ -364,6 +369,8 @@ static bool access_pmu_regs(struct kvm_vcpu *vcpu,
*vcpu_reg(vcpu, p->Rt) = val;
break;
}
+ case PMSWINC_EL0:
+ return read_zero(vcpu, p);
default:
*vcpu_reg(vcpu, p->Rt) = vcpu_sys_reg(vcpu, r->reg);
break;
@@ -576,7 +583,7 @@ static const struct sys_reg_desc sys_reg_descs[] = {
access_pmu_regs, reset_unknown, PMOVSCLR_EL0 },
/* PMSWINC_EL0 */
{ Op0(0b11), Op1(0b011), CRn(0b1001), CRm(0b1100), Op2(0b100),
- trap_raz_wi },
+ access_pmu_regs, reset_unknown, PMSWINC_EL0 },
/* PMSELR_EL0 */
{ Op0(0b11), Op1(0b011), CRn(0b1001), CRm(0b1100), Op2(0b101),
access_pmu_regs, reset_unknown, PMSELR_EL0 },
@@ -833,6 +840,11 @@ static bool access_pmu_cp15_regs(struct kvm_vcpu *vcpu,
vcpu_cp15(vcpu, c9_PMOVSSET) &= ~val;
break;
}
+ case c9_PMSWINC: {
+ val = *vcpu_reg(vcpu, p->Rt);
+ kvm_pmu_software_increment(vcpu, val);
+ break;
+ }
case c9_PMCR: {
/* Only update writeable bits of PMCR */
val = vcpu_cp15(vcpu, r->reg);
@@ -862,6 +874,8 @@ static bool access_pmu_cp15_regs(struct kvm_vcpu *vcpu,
*vcpu_reg(vcpu, p->Rt) = val;
break;
}
+ case c9_PMSWINC:
+ return read_zero(vcpu, p);
default:
*vcpu_reg(vcpu, p->Rt) = vcpu_cp15(vcpu, r->reg);
break;
@@ -907,6 +921,8 @@ static const struct sys_reg_desc cp15_regs[] = {
reset_unknown_cp15, c9_PMCNTENCLR },
{ Op1( 0), CRn( 9), CRm(12), Op2( 3), access_pmu_cp15_regs,
reset_unknown_cp15, c9_PMOVSCLR },
+ { Op1( 0), CRn( 9), CRm(12), Op2( 4), access_pmu_cp15_regs,
+ reset_unknown_cp15, c9_PMSWINC },
{ Op1( 0), CRn( 9), CRm(12), Op2( 5), access_pmu_cp15_regs,
reset_unknown_cp15, c9_PMSELR },
{ Op1( 0), CRn( 9), CRm(12), Op2( 6), access_pmu_cp15_regs,
diff --git a/include/kvm/arm_pmu.h b/include/kvm/arm_pmu.h
index 59e70af..1a27357 100644
--- a/include/kvm/arm_pmu.h
+++ b/include/kvm/arm_pmu.h
@@ -41,6 +41,7 @@ unsigned long kvm_pmu_get_counter_value(struct kvm_vcpu *vcpu,
unsigned long select_idx);
void kvm_pmu_disable_counter(struct kvm_vcpu *vcpu, unsigned long val);
void kvm_pmu_enable_counter(struct kvm_vcpu *vcpu, unsigned long val);
+void kvm_pmu_software_increment(struct kvm_vcpu *vcpu, unsigned long val);
void kvm_pmu_set_counter_event_type(struct kvm_vcpu *vcpu, unsigned long data,
unsigned long select_idx);
#else
@@ -51,6 +52,7 @@ unsigned long kvm_pmu_get_counter_value(struct kvm_vcpu *vcpu,
}
void kvm_pmu_disable_counter(struct kvm_vcpu *vcpu, unsigned long val) {}
void kvm_pmu_enable_counter(struct kvm_vcpu *vcpu, unsigned long val) {}
+void kvm_pmu_software_increment(struct kvm_vcpu *vcpu, unsigned long val) {}
void kvm_pmu_set_counter_event_type(struct kvm_vcpu *vcpu, unsigned long data,
unsigned long select_idx) {}
#endif
diff --git a/virt/kvm/arm/pmu.c b/virt/kvm/arm/pmu.c
index c6cdc4e..5f5a483 100644
--- a/virt/kvm/arm/pmu.c
+++ b/virt/kvm/arm/pmu.c
@@ -163,6 +163,35 @@ void kvm_pmu_disable_counter(struct kvm_vcpu *vcpu, unsigned long val)
}
/**
+ * kvm_pmu_software_increment - do software increment
+ * @vcpu: The vcpu pointer
+ * @val: the value guest writes to PMSWINC register
+ */
+void kvm_pmu_software_increment(struct kvm_vcpu *vcpu, unsigned long val)
+{
+ int i;
+ unsigned int type, enable;
+
+ for (i = 0; i < 32; i++) {
+ if ((val >> i) & 0x1) {
+ if (!vcpu_mode_is_32bit(vcpu)) {
+ type = vcpu_sys_reg(vcpu, PMEVTYPER0_EL0 + i)
+ & ARMV8_EVTYPE_EVENT;
+ enable = vcpu_sys_reg(vcpu, PMCNTENSET_EL0);
+ if ((type == 0) && ((enable >> i) & 0x1))
+ vcpu_sys_reg(vcpu, PMEVCNTR0_EL0 + i)++;
+ } else {
+ type = vcpu_cp15(vcpu, c14_PMEVTYPER0 + i)
+ & ARMV8_EVTYPE_EVENT;
+ enable = vcpu_cp15(vcpu, c9_PMCNTENSET);
+ if ((type == 0) && ((enable >> i) & 0x1))
+ vcpu_cp15(vcpu, c14_PMEVCNTR0 + i)++;
+ }
+ }
+ }
+}
+
+/**
* kvm_pmu_find_hw_event - find hardware event
* @pmu: The pmu pointer
* @event_select: The number of selected event type
@@ -239,6 +268,10 @@ void kvm_pmu_set_counter_event_type(struct kvm_vcpu *vcpu, unsigned long data,
kvm_pmu_stop_counter(vcpu, select_idx);
kvm_pmu_set_evttyper(vcpu, select_idx, data);
+ /* For software increment event we don't need to create perf event */
+ if (new_eventsel == 0)
+ return;
+
config = kvm_pmu_find_hw_event(pmu, new_eventsel);
if (config != PERF_COUNT_HW_MAX) {
type = PERF_TYPE_HARDWARE;
--
2.0.4
^ permalink raw reply related [flat|nested] 113+ messages in thread
* [PATCH v2 17/22] KVM: ARM64: Add reset and access handlers for PMSWINC register
@ 2015-09-11 8:55 ` Shannon Zhao
0 siblings, 0 replies; 113+ messages in thread
From: Shannon Zhao @ 2015-09-11 8:55 UTC (permalink / raw)
To: kvmarm; +Cc: kvm, marc.zyngier, will.deacon, linux-arm-kernel, shannon.zhao
From: Shannon Zhao <shannon.zhao@linaro.org>
Add access handler which emulates writing and reading PMSWINC
register and add support for creating software increment event.
Signed-off-by: Shannon Zhao <shannon.zhao@linaro.org>
---
arch/arm64/kvm/sys_regs.c | 18 +++++++++++++++++-
include/kvm/arm_pmu.h | 2 ++
virt/kvm/arm/pmu.c | 33 +++++++++++++++++++++++++++++++++
3 files changed, 52 insertions(+), 1 deletion(-)
diff --git a/arch/arm64/kvm/sys_regs.c b/arch/arm64/kvm/sys_regs.c
index b47cd0b..24d00a0 100644
--- a/arch/arm64/kvm/sys_regs.c
+++ b/arch/arm64/kvm/sys_regs.c
@@ -335,6 +335,11 @@ static bool access_pmu_regs(struct kvm_vcpu *vcpu,
vcpu_sys_reg(vcpu, PMOVSSET_EL0) &= ~val;
break;
}
+ case PMSWINC_EL0: {
+ val = *vcpu_reg(vcpu, p->Rt);
+ kvm_pmu_software_increment(vcpu, val);
+ break;
+ }
case PMCR_EL0: {
/* Only update writeable bits of PMCR */
val = vcpu_sys_reg(vcpu, r->reg);
@@ -364,6 +369,8 @@ static bool access_pmu_regs(struct kvm_vcpu *vcpu,
*vcpu_reg(vcpu, p->Rt) = val;
break;
}
+ case PMSWINC_EL0:
+ return read_zero(vcpu, p);
default:
*vcpu_reg(vcpu, p->Rt) = vcpu_sys_reg(vcpu, r->reg);
break;
@@ -576,7 +583,7 @@ static const struct sys_reg_desc sys_reg_descs[] = {
access_pmu_regs, reset_unknown, PMOVSCLR_EL0 },
/* PMSWINC_EL0 */
{ Op0(0b11), Op1(0b011), CRn(0b1001), CRm(0b1100), Op2(0b100),
- trap_raz_wi },
+ access_pmu_regs, reset_unknown, PMSWINC_EL0 },
/* PMSELR_EL0 */
{ Op0(0b11), Op1(0b011), CRn(0b1001), CRm(0b1100), Op2(0b101),
access_pmu_regs, reset_unknown, PMSELR_EL0 },
@@ -833,6 +840,11 @@ static bool access_pmu_cp15_regs(struct kvm_vcpu *vcpu,
vcpu_cp15(vcpu, c9_PMOVSSET) &= ~val;
break;
}
+ case c9_PMSWINC: {
+ val = *vcpu_reg(vcpu, p->Rt);
+ kvm_pmu_software_increment(vcpu, val);
+ break;
+ }
case c9_PMCR: {
/* Only update writeable bits of PMCR */
val = vcpu_cp15(vcpu, r->reg);
@@ -862,6 +874,8 @@ static bool access_pmu_cp15_regs(struct kvm_vcpu *vcpu,
*vcpu_reg(vcpu, p->Rt) = val;
break;
}
+ case c9_PMSWINC:
+ return read_zero(vcpu, p);
default:
*vcpu_reg(vcpu, p->Rt) = vcpu_cp15(vcpu, r->reg);
break;
@@ -907,6 +921,8 @@ static const struct sys_reg_desc cp15_regs[] = {
reset_unknown_cp15, c9_PMCNTENCLR },
{ Op1( 0), CRn( 9), CRm(12), Op2( 3), access_pmu_cp15_regs,
reset_unknown_cp15, c9_PMOVSCLR },
+ { Op1( 0), CRn( 9), CRm(12), Op2( 4), access_pmu_cp15_regs,
+ reset_unknown_cp15, c9_PMSWINC },
{ Op1( 0), CRn( 9), CRm(12), Op2( 5), access_pmu_cp15_regs,
reset_unknown_cp15, c9_PMSELR },
{ Op1( 0), CRn( 9), CRm(12), Op2( 6), access_pmu_cp15_regs,
diff --git a/include/kvm/arm_pmu.h b/include/kvm/arm_pmu.h
index 59e70af..1a27357 100644
--- a/include/kvm/arm_pmu.h
+++ b/include/kvm/arm_pmu.h
@@ -41,6 +41,7 @@ unsigned long kvm_pmu_get_counter_value(struct kvm_vcpu *vcpu,
unsigned long select_idx);
void kvm_pmu_disable_counter(struct kvm_vcpu *vcpu, unsigned long val);
void kvm_pmu_enable_counter(struct kvm_vcpu *vcpu, unsigned long val);
+void kvm_pmu_software_increment(struct kvm_vcpu *vcpu, unsigned long val);
void kvm_pmu_set_counter_event_type(struct kvm_vcpu *vcpu, unsigned long data,
unsigned long select_idx);
#else
@@ -51,6 +52,7 @@ unsigned long kvm_pmu_get_counter_value(struct kvm_vcpu *vcpu,
}
void kvm_pmu_disable_counter(struct kvm_vcpu *vcpu, unsigned long val) {}
void kvm_pmu_enable_counter(struct kvm_vcpu *vcpu, unsigned long val) {}
+void kvm_pmu_software_increment(struct kvm_vcpu *vcpu, unsigned long val) {}
void kvm_pmu_set_counter_event_type(struct kvm_vcpu *vcpu, unsigned long data,
unsigned long select_idx) {}
#endif
diff --git a/virt/kvm/arm/pmu.c b/virt/kvm/arm/pmu.c
index c6cdc4e..5f5a483 100644
--- a/virt/kvm/arm/pmu.c
+++ b/virt/kvm/arm/pmu.c
@@ -163,6 +163,35 @@ void kvm_pmu_disable_counter(struct kvm_vcpu *vcpu, unsigned long val)
}
/**
+ * kvm_pmu_software_increment - do software increment
+ * @vcpu: The vcpu pointer
+ * @val: the value guest writes to PMSWINC register
+ */
+void kvm_pmu_software_increment(struct kvm_vcpu *vcpu, unsigned long val)
+{
+ int i;
+ unsigned int type, enable;
+
+ for (i = 0; i < 32; i++) {
+ if ((val >> i) & 0x1) {
+ if (!vcpu_mode_is_32bit(vcpu)) {
+ type = vcpu_sys_reg(vcpu, PMEVTYPER0_EL0 + i)
+ & ARMV8_EVTYPE_EVENT;
+ enable = vcpu_sys_reg(vcpu, PMCNTENSET_EL0);
+ if ((type == 0) && ((enable >> i) & 0x1))
+ vcpu_sys_reg(vcpu, PMEVCNTR0_EL0 + i)++;
+ } else {
+ type = vcpu_cp15(vcpu, c14_PMEVTYPER0 + i)
+ & ARMV8_EVTYPE_EVENT;
+ enable = vcpu_cp15(vcpu, c9_PMCNTENSET);
+ if ((type == 0) && ((enable >> i) & 0x1))
+ vcpu_cp15(vcpu, c14_PMEVCNTR0 + i)++;
+ }
+ }
+ }
+}
+
+/**
* kvm_pmu_find_hw_event - find hardware event
* @pmu: The pmu pointer
* @event_select: The number of selected event type
@@ -239,6 +268,10 @@ void kvm_pmu_set_counter_event_type(struct kvm_vcpu *vcpu, unsigned long data,
kvm_pmu_stop_counter(vcpu, select_idx);
kvm_pmu_set_evttyper(vcpu, select_idx, data);
+ /* For software increment event we don't need to create perf event */
+ if (new_eventsel == 0)
+ return;
+
config = kvm_pmu_find_hw_event(pmu, new_eventsel);
if (config != PERF_COUNT_HW_MAX) {
type = PERF_TYPE_HARDWARE;
--
2.0.4
^ permalink raw reply related [flat|nested] 113+ messages in thread
* [PATCH v2 17/22] KVM: ARM64: Add reset and access handlers for PMSWINC register
@ 2015-09-11 8:55 ` Shannon Zhao
0 siblings, 0 replies; 113+ messages in thread
From: Shannon Zhao @ 2015-09-11 8:55 UTC (permalink / raw)
To: kvmarm; +Cc: kvm, marc.zyngier, will.deacon, linux-arm-kernel, shannon.zhao
From: Shannon Zhao <shannon.zhao@linaro.org>
Add access handler which emulates writing and reading PMSWINC
register and add support for creating software increment event.
Signed-off-by: Shannon Zhao <shannon.zhao@linaro.org>
---
arch/arm64/kvm/sys_regs.c | 18 +++++++++++++++++-
include/kvm/arm_pmu.h | 2 ++
virt/kvm/arm/pmu.c | 33 +++++++++++++++++++++++++++++++++
3 files changed, 52 insertions(+), 1 deletion(-)
diff --git a/arch/arm64/kvm/sys_regs.c b/arch/arm64/kvm/sys_regs.c
index b47cd0b..24d00a0 100644
--- a/arch/arm64/kvm/sys_regs.c
+++ b/arch/arm64/kvm/sys_regs.c
@@ -335,6 +335,11 @@ static bool access_pmu_regs(struct kvm_vcpu *vcpu,
vcpu_sys_reg(vcpu, PMOVSSET_EL0) &= ~val;
break;
}
+ case PMSWINC_EL0: {
+ val = *vcpu_reg(vcpu, p->Rt);
+ kvm_pmu_software_increment(vcpu, val);
+ break;
+ }
case PMCR_EL0: {
/* Only update writeable bits of PMCR */
val = vcpu_sys_reg(vcpu, r->reg);
@@ -364,6 +369,8 @@ static bool access_pmu_regs(struct kvm_vcpu *vcpu,
*vcpu_reg(vcpu, p->Rt) = val;
break;
}
+ case PMSWINC_EL0:
+ return read_zero(vcpu, p);
default:
*vcpu_reg(vcpu, p->Rt) = vcpu_sys_reg(vcpu, r->reg);
break;
@@ -576,7 +583,7 @@ static const struct sys_reg_desc sys_reg_descs[] = {
access_pmu_regs, reset_unknown, PMOVSCLR_EL0 },
/* PMSWINC_EL0 */
{ Op0(0b11), Op1(0b011), CRn(0b1001), CRm(0b1100), Op2(0b100),
- trap_raz_wi },
+ access_pmu_regs, reset_unknown, PMSWINC_EL0 },
/* PMSELR_EL0 */
{ Op0(0b11), Op1(0b011), CRn(0b1001), CRm(0b1100), Op2(0b101),
access_pmu_regs, reset_unknown, PMSELR_EL0 },
@@ -833,6 +840,11 @@ static bool access_pmu_cp15_regs(struct kvm_vcpu *vcpu,
vcpu_cp15(vcpu, c9_PMOVSSET) &= ~val;
break;
}
+ case c9_PMSWINC: {
+ val = *vcpu_reg(vcpu, p->Rt);
+ kvm_pmu_software_increment(vcpu, val);
+ break;
+ }
case c9_PMCR: {
/* Only update writeable bits of PMCR */
val = vcpu_cp15(vcpu, r->reg);
@@ -862,6 +874,8 @@ static bool access_pmu_cp15_regs(struct kvm_vcpu *vcpu,
*vcpu_reg(vcpu, p->Rt) = val;
break;
}
+ case c9_PMSWINC:
+ return read_zero(vcpu, p);
default:
*vcpu_reg(vcpu, p->Rt) = vcpu_cp15(vcpu, r->reg);
break;
@@ -907,6 +921,8 @@ static const struct sys_reg_desc cp15_regs[] = {
reset_unknown_cp15, c9_PMCNTENCLR },
{ Op1( 0), CRn( 9), CRm(12), Op2( 3), access_pmu_cp15_regs,
reset_unknown_cp15, c9_PMOVSCLR },
+ { Op1( 0), CRn( 9), CRm(12), Op2( 4), access_pmu_cp15_regs,
+ reset_unknown_cp15, c9_PMSWINC },
{ Op1( 0), CRn( 9), CRm(12), Op2( 5), access_pmu_cp15_regs,
reset_unknown_cp15, c9_PMSELR },
{ Op1( 0), CRn( 9), CRm(12), Op2( 6), access_pmu_cp15_regs,
diff --git a/include/kvm/arm_pmu.h b/include/kvm/arm_pmu.h
index 59e70af..1a27357 100644
--- a/include/kvm/arm_pmu.h
+++ b/include/kvm/arm_pmu.h
@@ -41,6 +41,7 @@ unsigned long kvm_pmu_get_counter_value(struct kvm_vcpu *vcpu,
unsigned long select_idx);
void kvm_pmu_disable_counter(struct kvm_vcpu *vcpu, unsigned long val);
void kvm_pmu_enable_counter(struct kvm_vcpu *vcpu, unsigned long val);
+void kvm_pmu_software_increment(struct kvm_vcpu *vcpu, unsigned long val);
void kvm_pmu_set_counter_event_type(struct kvm_vcpu *vcpu, unsigned long data,
unsigned long select_idx);
#else
@@ -51,6 +52,7 @@ unsigned long kvm_pmu_get_counter_value(struct kvm_vcpu *vcpu,
}
void kvm_pmu_disable_counter(struct kvm_vcpu *vcpu, unsigned long val) {}
void kvm_pmu_enable_counter(struct kvm_vcpu *vcpu, unsigned long val) {}
+void kvm_pmu_software_increment(struct kvm_vcpu *vcpu, unsigned long val) {}
void kvm_pmu_set_counter_event_type(struct kvm_vcpu *vcpu, unsigned long data,
unsigned long select_idx) {}
#endif
diff --git a/virt/kvm/arm/pmu.c b/virt/kvm/arm/pmu.c
index c6cdc4e..5f5a483 100644
--- a/virt/kvm/arm/pmu.c
+++ b/virt/kvm/arm/pmu.c
@@ -163,6 +163,35 @@ void kvm_pmu_disable_counter(struct kvm_vcpu *vcpu, unsigned long val)
}
/**
+ * kvm_pmu_software_increment - do software increment
+ * @vcpu: The vcpu pointer
+ * @val: the value guest writes to PMSWINC register
+ */
+void kvm_pmu_software_increment(struct kvm_vcpu *vcpu, unsigned long val)
+{
+ int i;
+ unsigned int type, enable;
+
+ for (i = 0; i < 32; i++) {
+ if ((val >> i) & 0x1) {
+ if (!vcpu_mode_is_32bit(vcpu)) {
+ type = vcpu_sys_reg(vcpu, PMEVTYPER0_EL0 + i)
+ & ARMV8_EVTYPE_EVENT;
+ enable = vcpu_sys_reg(vcpu, PMCNTENSET_EL0);
+ if ((type == 0) && ((enable >> i) & 0x1))
+ vcpu_sys_reg(vcpu, PMEVCNTR0_EL0 + i)++;
+ } else {
+ type = vcpu_cp15(vcpu, c14_PMEVTYPER0 + i)
+ & ARMV8_EVTYPE_EVENT;
+ enable = vcpu_cp15(vcpu, c9_PMCNTENSET);
+ if ((type == 0) && ((enable >> i) & 0x1))
+ vcpu_cp15(vcpu, c14_PMEVCNTR0 + i)++;
+ }
+ }
+ }
+}
+
+/**
* kvm_pmu_find_hw_event - find hardware event
* @pmu: The pmu pointer
* @event_select: The number of selected event type
@@ -239,6 +268,10 @@ void kvm_pmu_set_counter_event_type(struct kvm_vcpu *vcpu, unsigned long data,
kvm_pmu_stop_counter(vcpu, select_idx);
kvm_pmu_set_evttyper(vcpu, select_idx, data);
+ /* For software increment event we don't need to create perf event */
+ if (new_eventsel == 0)
+ return;
+
config = kvm_pmu_find_hw_event(pmu, new_eventsel);
if (config != PERF_COUNT_HW_MAX) {
type = PERF_TYPE_HARDWARE;
--
2.0.4
^ permalink raw reply related [flat|nested] 113+ messages in thread
* [PATCH v2 18/22] KVM: ARM64: Add access handlers for PMEVCNTRn and PMEVTYPERn register
2015-09-11 8:54 ` Shannon Zhao
(?)
@ 2015-09-11 8:55 ` Shannon Zhao
-1 siblings, 0 replies; 113+ messages in thread
From: Shannon Zhao @ 2015-09-11 8:55 UTC (permalink / raw)
To: linux-arm-kernel
From: Shannon Zhao <shannon.zhao@linaro.org>
Add access handler which emulates writing and reading PMEVCNTRn and
PMEVTYPERn.
Signed-off-by: Shannon Zhao <shannon.zhao@linaro.org>
---
arch/arm64/kvm/sys_regs.c | 164 ++++++++++++++++++++++++++++++++++++++++++++++
1 file changed, 164 insertions(+)
diff --git a/arch/arm64/kvm/sys_regs.c b/arch/arm64/kvm/sys_regs.c
index 24d00a0..cf2cb47 100644
--- a/arch/arm64/kvm/sys_regs.c
+++ b/arch/arm64/kvm/sys_regs.c
@@ -395,6 +395,20 @@ static bool access_pmu_regs(struct kvm_vcpu *vcpu,
{ Op0(0b10), Op1(0b000), CRn(0b0000), CRm((n)), Op2(0b111), \
trap_debug_regs, reset_val, (DBGWCR0_EL1 + (n)), 0 }
+/* Macro to expand the PMEVCNTRn_EL0 register */
+#define PMU_PMEVCNTR_EL0(n) \
+ /* PMEVCNTRn_EL0 */ \
+ { Op0(0b11), Op1(0b011), CRn(0b1110), \
+ CRm((0b1000 | (((n) >> 3) & 0x3))), Op2(((n) & 0x7)), \
+ access_pmu_regs, reset_unknown, (PMEVCNTR0_EL0 + n), }
+
+/* Macro to expand the PMEVTYPERn_EL0 register */
+#define PMU_PMEVTYPER_EL0(n) \
+ /* PMEVTYPERn_EL0 */ \
+ { Op0(0b11), Op1(0b011), CRn(0b1110), \
+ CRm((0b1100 | (((n) >> 3) & 0x3))), Op2(((n) & 0x7)), \
+ access_pmu_regs, reset_unknown, (PMEVTYPER0_EL0 + n), }
+
/*
* Architected system registers.
* Important: Must be sorted ascending by Op0, Op1, CRn, CRm, Op2
@@ -616,6 +630,74 @@ static const struct sys_reg_desc sys_reg_descs[] = {
{ Op0(0b11), Op1(0b011), CRn(0b1101), CRm(0b0000), Op2(0b011),
NULL, reset_unknown, TPIDRRO_EL0 },
+ /* PMEVCNTRn_EL0 */
+ PMU_PMEVCNTR_EL0(0),
+ PMU_PMEVCNTR_EL0(1),
+ PMU_PMEVCNTR_EL0(2),
+ PMU_PMEVCNTR_EL0(3),
+ PMU_PMEVCNTR_EL0(4),
+ PMU_PMEVCNTR_EL0(5),
+ PMU_PMEVCNTR_EL0(6),
+ PMU_PMEVCNTR_EL0(7),
+ PMU_PMEVCNTR_EL0(8),
+ PMU_PMEVCNTR_EL0(9),
+ PMU_PMEVCNTR_EL0(10),
+ PMU_PMEVCNTR_EL0(11),
+ PMU_PMEVCNTR_EL0(12),
+ PMU_PMEVCNTR_EL0(13),
+ PMU_PMEVCNTR_EL0(14),
+ PMU_PMEVCNTR_EL0(15),
+ PMU_PMEVCNTR_EL0(16),
+ PMU_PMEVCNTR_EL0(17),
+ PMU_PMEVCNTR_EL0(18),
+ PMU_PMEVCNTR_EL0(19),
+ PMU_PMEVCNTR_EL0(20),
+ PMU_PMEVCNTR_EL0(21),
+ PMU_PMEVCNTR_EL0(22),
+ PMU_PMEVCNTR_EL0(23),
+ PMU_PMEVCNTR_EL0(24),
+ PMU_PMEVCNTR_EL0(25),
+ PMU_PMEVCNTR_EL0(26),
+ PMU_PMEVCNTR_EL0(27),
+ PMU_PMEVCNTR_EL0(28),
+ PMU_PMEVCNTR_EL0(29),
+ PMU_PMEVCNTR_EL0(30),
+ /* PMEVTYPERn_EL0 */
+ PMU_PMEVTYPER_EL0(0),
+ PMU_PMEVTYPER_EL0(1),
+ PMU_PMEVTYPER_EL0(2),
+ PMU_PMEVTYPER_EL0(3),
+ PMU_PMEVTYPER_EL0(4),
+ PMU_PMEVTYPER_EL0(5),
+ PMU_PMEVTYPER_EL0(6),
+ PMU_PMEVTYPER_EL0(7),
+ PMU_PMEVTYPER_EL0(8),
+ PMU_PMEVTYPER_EL0(9),
+ PMU_PMEVTYPER_EL0(10),
+ PMU_PMEVTYPER_EL0(11),
+ PMU_PMEVTYPER_EL0(12),
+ PMU_PMEVTYPER_EL0(13),
+ PMU_PMEVTYPER_EL0(14),
+ PMU_PMEVTYPER_EL0(15),
+ PMU_PMEVTYPER_EL0(16),
+ PMU_PMEVTYPER_EL0(17),
+ PMU_PMEVTYPER_EL0(18),
+ PMU_PMEVTYPER_EL0(19),
+ PMU_PMEVTYPER_EL0(20),
+ PMU_PMEVTYPER_EL0(21),
+ PMU_PMEVTYPER_EL0(22),
+ PMU_PMEVTYPER_EL0(23),
+ PMU_PMEVTYPER_EL0(24),
+ PMU_PMEVTYPER_EL0(25),
+ PMU_PMEVTYPER_EL0(26),
+ PMU_PMEVTYPER_EL0(27),
+ PMU_PMEVTYPER_EL0(28),
+ PMU_PMEVTYPER_EL0(29),
+ PMU_PMEVTYPER_EL0(30),
+ /* PMCCFILTR_EL0 */
+ { Op0(0b11), Op1(0b011), CRn(0b1110), CRm(0b1111), Op2(0b111),
+ access_pmu_regs, reset_unknown, PMCCFILTR_EL0, },
+
/* DACR32_EL2 */
{ Op0(0b11), Op1(0b100), CRn(0b0011), CRm(0b0000), Op2(0b000),
NULL, reset_unknown, DACR32_EL2 },
@@ -885,6 +967,20 @@ static bool access_pmu_cp15_regs(struct kvm_vcpu *vcpu,
return true;
}
+/* Macro to expand the PMEVCNTRn register */
+#define PMU_PMEVCNTR(n) \
+ /* PMEVCNTRn */ \
+ { Op1(0), CRn(0b1110), \
+ CRm((0b1000 | (((n) >> 3) & 0x3))), Op2(((n) & 0x7)), \
+ access_pmu_cp15_regs, reset_unknown_cp15, (c14_PMEVCNTR0 + n), }
+
+/* Macro to expand the PMEVTYPERn register */
+#define PMU_PMEVTYPER(n) \
+ /* PMEVTYPERn */ \
+ { Op1(0), CRn(0b1110), \
+ CRm((0b1100 | (((n) >> 3) & 0x3))), Op2(((n) & 0x7)), \
+ access_pmu_cp15_regs, reset_unknown_cp15, (c14_PMEVTYPER0 + n), }
+
/*
* Trapped cp15 registers. TTBR0/TTBR1 get a double encoding,
* depending on the way they are accessed (as a 32bit or a 64bit
@@ -953,6 +1049,74 @@ static const struct sys_reg_desc cp15_regs[] = {
{ Op1( 0), CRn(12), CRm(12), Op2( 5), trap_raz_wi },
{ Op1( 0), CRn(13), CRm( 0), Op2( 1), access_vm_reg, NULL, c13_CID },
+
+ /* PMEVCNTRn */
+ PMU_PMEVCNTR(0),
+ PMU_PMEVCNTR(1),
+ PMU_PMEVCNTR(2),
+ PMU_PMEVCNTR(3),
+ PMU_PMEVCNTR(4),
+ PMU_PMEVCNTR(5),
+ PMU_PMEVCNTR(6),
+ PMU_PMEVCNTR(7),
+ PMU_PMEVCNTR(8),
+ PMU_PMEVCNTR(9),
+ PMU_PMEVCNTR(10),
+ PMU_PMEVCNTR(11),
+ PMU_PMEVCNTR(12),
+ PMU_PMEVCNTR(13),
+ PMU_PMEVCNTR(14),
+ PMU_PMEVCNTR(15),
+ PMU_PMEVCNTR(16),
+ PMU_PMEVCNTR(17),
+ PMU_PMEVCNTR(18),
+ PMU_PMEVCNTR(19),
+ PMU_PMEVCNTR(20),
+ PMU_PMEVCNTR(21),
+ PMU_PMEVCNTR(22),
+ PMU_PMEVCNTR(23),
+ PMU_PMEVCNTR(24),
+ PMU_PMEVCNTR(25),
+ PMU_PMEVCNTR(26),
+ PMU_PMEVCNTR(27),
+ PMU_PMEVCNTR(28),
+ PMU_PMEVCNTR(29),
+ PMU_PMEVCNTR(30),
+ /* PMEVTYPERn */
+ PMU_PMEVTYPER(0),
+ PMU_PMEVTYPER(1),
+ PMU_PMEVTYPER(2),
+ PMU_PMEVTYPER(3),
+ PMU_PMEVTYPER(4),
+ PMU_PMEVTYPER(5),
+ PMU_PMEVTYPER(6),
+ PMU_PMEVTYPER(7),
+ PMU_PMEVTYPER(8),
+ PMU_PMEVTYPER(9),
+ PMU_PMEVTYPER(10),
+ PMU_PMEVTYPER(11),
+ PMU_PMEVTYPER(12),
+ PMU_PMEVTYPER(13),
+ PMU_PMEVTYPER(14),
+ PMU_PMEVTYPER(15),
+ PMU_PMEVTYPER(16),
+ PMU_PMEVTYPER(17),
+ PMU_PMEVTYPER(18),
+ PMU_PMEVTYPER(19),
+ PMU_PMEVTYPER(20),
+ PMU_PMEVTYPER(21),
+ PMU_PMEVTYPER(22),
+ PMU_PMEVTYPER(23),
+ PMU_PMEVTYPER(24),
+ PMU_PMEVTYPER(25),
+ PMU_PMEVTYPER(26),
+ PMU_PMEVTYPER(27),
+ PMU_PMEVTYPER(28),
+ PMU_PMEVTYPER(29),
+ PMU_PMEVTYPER(30),
+ /* PMCCFILTR */
+ { Op1(0), CRn(14), CRm(15), Op2(7), access_pmu_cp15_regs,
+ reset_val_cp15, c14_PMCCFILTR, 0 },
};
static const struct sys_reg_desc cp15_64_regs[] = {
--
2.0.4
^ permalink raw reply related [flat|nested] 113+ messages in thread
* [PATCH v2 18/22] KVM: ARM64: Add access handlers for PMEVCNTRn and PMEVTYPERn register
@ 2015-09-11 8:55 ` Shannon Zhao
0 siblings, 0 replies; 113+ messages in thread
From: Shannon Zhao @ 2015-09-11 8:55 UTC (permalink / raw)
To: kvmarm; +Cc: kvm, marc.zyngier, will.deacon, linux-arm-kernel, shannon.zhao
From: Shannon Zhao <shannon.zhao@linaro.org>
Add access handler which emulates writing and reading PMEVCNTRn and
PMEVTYPERn.
Signed-off-by: Shannon Zhao <shannon.zhao@linaro.org>
---
arch/arm64/kvm/sys_regs.c | 164 ++++++++++++++++++++++++++++++++++++++++++++++
1 file changed, 164 insertions(+)
diff --git a/arch/arm64/kvm/sys_regs.c b/arch/arm64/kvm/sys_regs.c
index 24d00a0..cf2cb47 100644
--- a/arch/arm64/kvm/sys_regs.c
+++ b/arch/arm64/kvm/sys_regs.c
@@ -395,6 +395,20 @@ static bool access_pmu_regs(struct kvm_vcpu *vcpu,
{ Op0(0b10), Op1(0b000), CRn(0b0000), CRm((n)), Op2(0b111), \
trap_debug_regs, reset_val, (DBGWCR0_EL1 + (n)), 0 }
+/* Macro to expand the PMEVCNTRn_EL0 register */
+#define PMU_PMEVCNTR_EL0(n) \
+ /* PMEVCNTRn_EL0 */ \
+ { Op0(0b11), Op1(0b011), CRn(0b1110), \
+ CRm((0b1000 | (((n) >> 3) & 0x3))), Op2(((n) & 0x7)), \
+ access_pmu_regs, reset_unknown, (PMEVCNTR0_EL0 + n), }
+
+/* Macro to expand the PMEVTYPERn_EL0 register */
+#define PMU_PMEVTYPER_EL0(n) \
+ /* PMEVTYPERn_EL0 */ \
+ { Op0(0b11), Op1(0b011), CRn(0b1110), \
+ CRm((0b1100 | (((n) >> 3) & 0x3))), Op2(((n) & 0x7)), \
+ access_pmu_regs, reset_unknown, (PMEVTYPER0_EL0 + n), }
+
/*
* Architected system registers.
* Important: Must be sorted ascending by Op0, Op1, CRn, CRm, Op2
@@ -616,6 +630,74 @@ static const struct sys_reg_desc sys_reg_descs[] = {
{ Op0(0b11), Op1(0b011), CRn(0b1101), CRm(0b0000), Op2(0b011),
NULL, reset_unknown, TPIDRRO_EL0 },
+ /* PMEVCNTRn_EL0 */
+ PMU_PMEVCNTR_EL0(0),
+ PMU_PMEVCNTR_EL0(1),
+ PMU_PMEVCNTR_EL0(2),
+ PMU_PMEVCNTR_EL0(3),
+ PMU_PMEVCNTR_EL0(4),
+ PMU_PMEVCNTR_EL0(5),
+ PMU_PMEVCNTR_EL0(6),
+ PMU_PMEVCNTR_EL0(7),
+ PMU_PMEVCNTR_EL0(8),
+ PMU_PMEVCNTR_EL0(9),
+ PMU_PMEVCNTR_EL0(10),
+ PMU_PMEVCNTR_EL0(11),
+ PMU_PMEVCNTR_EL0(12),
+ PMU_PMEVCNTR_EL0(13),
+ PMU_PMEVCNTR_EL0(14),
+ PMU_PMEVCNTR_EL0(15),
+ PMU_PMEVCNTR_EL0(16),
+ PMU_PMEVCNTR_EL0(17),
+ PMU_PMEVCNTR_EL0(18),
+ PMU_PMEVCNTR_EL0(19),
+ PMU_PMEVCNTR_EL0(20),
+ PMU_PMEVCNTR_EL0(21),
+ PMU_PMEVCNTR_EL0(22),
+ PMU_PMEVCNTR_EL0(23),
+ PMU_PMEVCNTR_EL0(24),
+ PMU_PMEVCNTR_EL0(25),
+ PMU_PMEVCNTR_EL0(26),
+ PMU_PMEVCNTR_EL0(27),
+ PMU_PMEVCNTR_EL0(28),
+ PMU_PMEVCNTR_EL0(29),
+ PMU_PMEVCNTR_EL0(30),
+ /* PMEVTYPERn_EL0 */
+ PMU_PMEVTYPER_EL0(0),
+ PMU_PMEVTYPER_EL0(1),
+ PMU_PMEVTYPER_EL0(2),
+ PMU_PMEVTYPER_EL0(3),
+ PMU_PMEVTYPER_EL0(4),
+ PMU_PMEVTYPER_EL0(5),
+ PMU_PMEVTYPER_EL0(6),
+ PMU_PMEVTYPER_EL0(7),
+ PMU_PMEVTYPER_EL0(8),
+ PMU_PMEVTYPER_EL0(9),
+ PMU_PMEVTYPER_EL0(10),
+ PMU_PMEVTYPER_EL0(11),
+ PMU_PMEVTYPER_EL0(12),
+ PMU_PMEVTYPER_EL0(13),
+ PMU_PMEVTYPER_EL0(14),
+ PMU_PMEVTYPER_EL0(15),
+ PMU_PMEVTYPER_EL0(16),
+ PMU_PMEVTYPER_EL0(17),
+ PMU_PMEVTYPER_EL0(18),
+ PMU_PMEVTYPER_EL0(19),
+ PMU_PMEVTYPER_EL0(20),
+ PMU_PMEVTYPER_EL0(21),
+ PMU_PMEVTYPER_EL0(22),
+ PMU_PMEVTYPER_EL0(23),
+ PMU_PMEVTYPER_EL0(24),
+ PMU_PMEVTYPER_EL0(25),
+ PMU_PMEVTYPER_EL0(26),
+ PMU_PMEVTYPER_EL0(27),
+ PMU_PMEVTYPER_EL0(28),
+ PMU_PMEVTYPER_EL0(29),
+ PMU_PMEVTYPER_EL0(30),
+ /* PMCCFILTR_EL0 */
+ { Op0(0b11), Op1(0b011), CRn(0b1110), CRm(0b1111), Op2(0b111),
+ access_pmu_regs, reset_unknown, PMCCFILTR_EL0, },
+
/* DACR32_EL2 */
{ Op0(0b11), Op1(0b100), CRn(0b0011), CRm(0b0000), Op2(0b000),
NULL, reset_unknown, DACR32_EL2 },
@@ -885,6 +967,20 @@ static bool access_pmu_cp15_regs(struct kvm_vcpu *vcpu,
return true;
}
+/* Macro to expand the PMEVCNTRn register */
+#define PMU_PMEVCNTR(n) \
+ /* PMEVCNTRn */ \
+ { Op1(0), CRn(0b1110), \
+ CRm((0b1000 | (((n) >> 3) & 0x3))), Op2(((n) & 0x7)), \
+ access_pmu_cp15_regs, reset_unknown_cp15, (c14_PMEVCNTR0 + n), }
+
+/* Macro to expand the PMEVTYPERn register */
+#define PMU_PMEVTYPER(n) \
+ /* PMEVTYPERn */ \
+ { Op1(0), CRn(0b1110), \
+ CRm((0b1100 | (((n) >> 3) & 0x3))), Op2(((n) & 0x7)), \
+ access_pmu_cp15_regs, reset_unknown_cp15, (c14_PMEVTYPER0 + n), }
+
/*
* Trapped cp15 registers. TTBR0/TTBR1 get a double encoding,
* depending on the way they are accessed (as a 32bit or a 64bit
@@ -953,6 +1049,74 @@ static const struct sys_reg_desc cp15_regs[] = {
{ Op1( 0), CRn(12), CRm(12), Op2( 5), trap_raz_wi },
{ Op1( 0), CRn(13), CRm( 0), Op2( 1), access_vm_reg, NULL, c13_CID },
+
+ /* PMEVCNTRn */
+ PMU_PMEVCNTR(0),
+ PMU_PMEVCNTR(1),
+ PMU_PMEVCNTR(2),
+ PMU_PMEVCNTR(3),
+ PMU_PMEVCNTR(4),
+ PMU_PMEVCNTR(5),
+ PMU_PMEVCNTR(6),
+ PMU_PMEVCNTR(7),
+ PMU_PMEVCNTR(8),
+ PMU_PMEVCNTR(9),
+ PMU_PMEVCNTR(10),
+ PMU_PMEVCNTR(11),
+ PMU_PMEVCNTR(12),
+ PMU_PMEVCNTR(13),
+ PMU_PMEVCNTR(14),
+ PMU_PMEVCNTR(15),
+ PMU_PMEVCNTR(16),
+ PMU_PMEVCNTR(17),
+ PMU_PMEVCNTR(18),
+ PMU_PMEVCNTR(19),
+ PMU_PMEVCNTR(20),
+ PMU_PMEVCNTR(21),
+ PMU_PMEVCNTR(22),
+ PMU_PMEVCNTR(23),
+ PMU_PMEVCNTR(24),
+ PMU_PMEVCNTR(25),
+ PMU_PMEVCNTR(26),
+ PMU_PMEVCNTR(27),
+ PMU_PMEVCNTR(28),
+ PMU_PMEVCNTR(29),
+ PMU_PMEVCNTR(30),
+ /* PMEVTYPERn */
+ PMU_PMEVTYPER(0),
+ PMU_PMEVTYPER(1),
+ PMU_PMEVTYPER(2),
+ PMU_PMEVTYPER(3),
+ PMU_PMEVTYPER(4),
+ PMU_PMEVTYPER(5),
+ PMU_PMEVTYPER(6),
+ PMU_PMEVTYPER(7),
+ PMU_PMEVTYPER(8),
+ PMU_PMEVTYPER(9),
+ PMU_PMEVTYPER(10),
+ PMU_PMEVTYPER(11),
+ PMU_PMEVTYPER(12),
+ PMU_PMEVTYPER(13),
+ PMU_PMEVTYPER(14),
+ PMU_PMEVTYPER(15),
+ PMU_PMEVTYPER(16),
+ PMU_PMEVTYPER(17),
+ PMU_PMEVTYPER(18),
+ PMU_PMEVTYPER(19),
+ PMU_PMEVTYPER(20),
+ PMU_PMEVTYPER(21),
+ PMU_PMEVTYPER(22),
+ PMU_PMEVTYPER(23),
+ PMU_PMEVTYPER(24),
+ PMU_PMEVTYPER(25),
+ PMU_PMEVTYPER(26),
+ PMU_PMEVTYPER(27),
+ PMU_PMEVTYPER(28),
+ PMU_PMEVTYPER(29),
+ PMU_PMEVTYPER(30),
+ /* PMCCFILTR */
+ { Op1(0), CRn(14), CRm(15), Op2(7), access_pmu_cp15_regs,
+ reset_val_cp15, c14_PMCCFILTR, 0 },
};
static const struct sys_reg_desc cp15_64_regs[] = {
--
2.0.4
^ permalink raw reply related [flat|nested] 113+ messages in thread
* [PATCH v2 18/22] KVM: ARM64: Add access handlers for PMEVCNTRn and PMEVTYPERn register
@ 2015-09-11 8:55 ` Shannon Zhao
0 siblings, 0 replies; 113+ messages in thread
From: Shannon Zhao @ 2015-09-11 8:55 UTC (permalink / raw)
To: kvmarm; +Cc: kvm, marc.zyngier, will.deacon, linux-arm-kernel, shannon.zhao
From: Shannon Zhao <shannon.zhao@linaro.org>
Add access handler which emulates writing and reading PMEVCNTRn and
PMEVTYPERn.
Signed-off-by: Shannon Zhao <shannon.zhao@linaro.org>
---
arch/arm64/kvm/sys_regs.c | 164 ++++++++++++++++++++++++++++++++++++++++++++++
1 file changed, 164 insertions(+)
diff --git a/arch/arm64/kvm/sys_regs.c b/arch/arm64/kvm/sys_regs.c
index 24d00a0..cf2cb47 100644
--- a/arch/arm64/kvm/sys_regs.c
+++ b/arch/arm64/kvm/sys_regs.c
@@ -395,6 +395,20 @@ static bool access_pmu_regs(struct kvm_vcpu *vcpu,
{ Op0(0b10), Op1(0b000), CRn(0b0000), CRm((n)), Op2(0b111), \
trap_debug_regs, reset_val, (DBGWCR0_EL1 + (n)), 0 }
+/* Macro to expand the PMEVCNTRn_EL0 register */
+#define PMU_PMEVCNTR_EL0(n) \
+ /* PMEVCNTRn_EL0 */ \
+ { Op0(0b11), Op1(0b011), CRn(0b1110), \
+ CRm((0b1000 | (((n) >> 3) & 0x3))), Op2(((n) & 0x7)), \
+ access_pmu_regs, reset_unknown, (PMEVCNTR0_EL0 + n), }
+
+/* Macro to expand the PMEVTYPERn_EL0 register */
+#define PMU_PMEVTYPER_EL0(n) \
+ /* PMEVTYPERn_EL0 */ \
+ { Op0(0b11), Op1(0b011), CRn(0b1110), \
+ CRm((0b1100 | (((n) >> 3) & 0x3))), Op2(((n) & 0x7)), \
+ access_pmu_regs, reset_unknown, (PMEVTYPER0_EL0 + n), }
+
/*
* Architected system registers.
* Important: Must be sorted ascending by Op0, Op1, CRn, CRm, Op2
@@ -616,6 +630,74 @@ static const struct sys_reg_desc sys_reg_descs[] = {
{ Op0(0b11), Op1(0b011), CRn(0b1101), CRm(0b0000), Op2(0b011),
NULL, reset_unknown, TPIDRRO_EL0 },
+ /* PMEVCNTRn_EL0 */
+ PMU_PMEVCNTR_EL0(0),
+ PMU_PMEVCNTR_EL0(1),
+ PMU_PMEVCNTR_EL0(2),
+ PMU_PMEVCNTR_EL0(3),
+ PMU_PMEVCNTR_EL0(4),
+ PMU_PMEVCNTR_EL0(5),
+ PMU_PMEVCNTR_EL0(6),
+ PMU_PMEVCNTR_EL0(7),
+ PMU_PMEVCNTR_EL0(8),
+ PMU_PMEVCNTR_EL0(9),
+ PMU_PMEVCNTR_EL0(10),
+ PMU_PMEVCNTR_EL0(11),
+ PMU_PMEVCNTR_EL0(12),
+ PMU_PMEVCNTR_EL0(13),
+ PMU_PMEVCNTR_EL0(14),
+ PMU_PMEVCNTR_EL0(15),
+ PMU_PMEVCNTR_EL0(16),
+ PMU_PMEVCNTR_EL0(17),
+ PMU_PMEVCNTR_EL0(18),
+ PMU_PMEVCNTR_EL0(19),
+ PMU_PMEVCNTR_EL0(20),
+ PMU_PMEVCNTR_EL0(21),
+ PMU_PMEVCNTR_EL0(22),
+ PMU_PMEVCNTR_EL0(23),
+ PMU_PMEVCNTR_EL0(24),
+ PMU_PMEVCNTR_EL0(25),
+ PMU_PMEVCNTR_EL0(26),
+ PMU_PMEVCNTR_EL0(27),
+ PMU_PMEVCNTR_EL0(28),
+ PMU_PMEVCNTR_EL0(29),
+ PMU_PMEVCNTR_EL0(30),
+ /* PMEVTYPERn_EL0 */
+ PMU_PMEVTYPER_EL0(0),
+ PMU_PMEVTYPER_EL0(1),
+ PMU_PMEVTYPER_EL0(2),
+ PMU_PMEVTYPER_EL0(3),
+ PMU_PMEVTYPER_EL0(4),
+ PMU_PMEVTYPER_EL0(5),
+ PMU_PMEVTYPER_EL0(6),
+ PMU_PMEVTYPER_EL0(7),
+ PMU_PMEVTYPER_EL0(8),
+ PMU_PMEVTYPER_EL0(9),
+ PMU_PMEVTYPER_EL0(10),
+ PMU_PMEVTYPER_EL0(11),
+ PMU_PMEVTYPER_EL0(12),
+ PMU_PMEVTYPER_EL0(13),
+ PMU_PMEVTYPER_EL0(14),
+ PMU_PMEVTYPER_EL0(15),
+ PMU_PMEVTYPER_EL0(16),
+ PMU_PMEVTYPER_EL0(17),
+ PMU_PMEVTYPER_EL0(18),
+ PMU_PMEVTYPER_EL0(19),
+ PMU_PMEVTYPER_EL0(20),
+ PMU_PMEVTYPER_EL0(21),
+ PMU_PMEVTYPER_EL0(22),
+ PMU_PMEVTYPER_EL0(23),
+ PMU_PMEVTYPER_EL0(24),
+ PMU_PMEVTYPER_EL0(25),
+ PMU_PMEVTYPER_EL0(26),
+ PMU_PMEVTYPER_EL0(27),
+ PMU_PMEVTYPER_EL0(28),
+ PMU_PMEVTYPER_EL0(29),
+ PMU_PMEVTYPER_EL0(30),
+ /* PMCCFILTR_EL0 */
+ { Op0(0b11), Op1(0b011), CRn(0b1110), CRm(0b1111), Op2(0b111),
+ access_pmu_regs, reset_unknown, PMCCFILTR_EL0, },
+
/* DACR32_EL2 */
{ Op0(0b11), Op1(0b100), CRn(0b0011), CRm(0b0000), Op2(0b000),
NULL, reset_unknown, DACR32_EL2 },
@@ -885,6 +967,20 @@ static bool access_pmu_cp15_regs(struct kvm_vcpu *vcpu,
return true;
}
+/* Macro to expand the PMEVCNTRn register */
+#define PMU_PMEVCNTR(n) \
+ /* PMEVCNTRn */ \
+ { Op1(0), CRn(0b1110), \
+ CRm((0b1000 | (((n) >> 3) & 0x3))), Op2(((n) & 0x7)), \
+ access_pmu_cp15_regs, reset_unknown_cp15, (c14_PMEVCNTR0 + n), }
+
+/* Macro to expand the PMEVTYPERn register */
+#define PMU_PMEVTYPER(n) \
+ /* PMEVTYPERn */ \
+ { Op1(0), CRn(0b1110), \
+ CRm((0b1100 | (((n) >> 3) & 0x3))), Op2(((n) & 0x7)), \
+ access_pmu_cp15_regs, reset_unknown_cp15, (c14_PMEVTYPER0 + n), }
+
/*
* Trapped cp15 registers. TTBR0/TTBR1 get a double encoding,
* depending on the way they are accessed (as a 32bit or a 64bit
@@ -953,6 +1049,74 @@ static const struct sys_reg_desc cp15_regs[] = {
{ Op1( 0), CRn(12), CRm(12), Op2( 5), trap_raz_wi },
{ Op1( 0), CRn(13), CRm( 0), Op2( 1), access_vm_reg, NULL, c13_CID },
+
+ /* PMEVCNTRn */
+ PMU_PMEVCNTR(0),
+ PMU_PMEVCNTR(1),
+ PMU_PMEVCNTR(2),
+ PMU_PMEVCNTR(3),
+ PMU_PMEVCNTR(4),
+ PMU_PMEVCNTR(5),
+ PMU_PMEVCNTR(6),
+ PMU_PMEVCNTR(7),
+ PMU_PMEVCNTR(8),
+ PMU_PMEVCNTR(9),
+ PMU_PMEVCNTR(10),
+ PMU_PMEVCNTR(11),
+ PMU_PMEVCNTR(12),
+ PMU_PMEVCNTR(13),
+ PMU_PMEVCNTR(14),
+ PMU_PMEVCNTR(15),
+ PMU_PMEVCNTR(16),
+ PMU_PMEVCNTR(17),
+ PMU_PMEVCNTR(18),
+ PMU_PMEVCNTR(19),
+ PMU_PMEVCNTR(20),
+ PMU_PMEVCNTR(21),
+ PMU_PMEVCNTR(22),
+ PMU_PMEVCNTR(23),
+ PMU_PMEVCNTR(24),
+ PMU_PMEVCNTR(25),
+ PMU_PMEVCNTR(26),
+ PMU_PMEVCNTR(27),
+ PMU_PMEVCNTR(28),
+ PMU_PMEVCNTR(29),
+ PMU_PMEVCNTR(30),
+ /* PMEVTYPERn */
+ PMU_PMEVTYPER(0),
+ PMU_PMEVTYPER(1),
+ PMU_PMEVTYPER(2),
+ PMU_PMEVTYPER(3),
+ PMU_PMEVTYPER(4),
+ PMU_PMEVTYPER(5),
+ PMU_PMEVTYPER(6),
+ PMU_PMEVTYPER(7),
+ PMU_PMEVTYPER(8),
+ PMU_PMEVTYPER(9),
+ PMU_PMEVTYPER(10),
+ PMU_PMEVTYPER(11),
+ PMU_PMEVTYPER(12),
+ PMU_PMEVTYPER(13),
+ PMU_PMEVTYPER(14),
+ PMU_PMEVTYPER(15),
+ PMU_PMEVTYPER(16),
+ PMU_PMEVTYPER(17),
+ PMU_PMEVTYPER(18),
+ PMU_PMEVTYPER(19),
+ PMU_PMEVTYPER(20),
+ PMU_PMEVTYPER(21),
+ PMU_PMEVTYPER(22),
+ PMU_PMEVTYPER(23),
+ PMU_PMEVTYPER(24),
+ PMU_PMEVTYPER(25),
+ PMU_PMEVTYPER(26),
+ PMU_PMEVTYPER(27),
+ PMU_PMEVTYPER(28),
+ PMU_PMEVTYPER(29),
+ PMU_PMEVTYPER(30),
+ /* PMCCFILTR */
+ { Op1(0), CRn(14), CRm(15), Op2(7), access_pmu_cp15_regs,
+ reset_val_cp15, c14_PMCCFILTR, 0 },
};
static const struct sys_reg_desc cp15_64_regs[] = {
--
2.0.4
^ permalink raw reply related [flat|nested] 113+ messages in thread
* [PATCH v2 19/22] KVM: ARM64: Add PMU overflow interrupt routing
2015-09-11 8:54 ` Shannon Zhao
(?)
@ 2015-09-11 8:55 ` Shannon Zhao
-1 siblings, 0 replies; 113+ messages in thread
From: Shannon Zhao @ 2015-09-11 8:55 UTC (permalink / raw)
To: linux-arm-kernel
From: Shannon Zhao <shannon.zhao@linaro.org>
When calling perf_event_create_kernel_counter to create perf_event,
assign a overflow handler. Then when perf event overflows, set
irq_pending and call kvm_vcpu_kick() to sync the interrupt.
Signed-off-by: Shannon Zhao <shannon.zhao@linaro.org>
---
arch/arm/kvm/arm.c | 3 +++
include/kvm/arm_pmu.h | 2 ++
virt/kvm/arm/pmu.c | 54 ++++++++++++++++++++++++++++++++++++++++++++++++++-
3 files changed, 58 insertions(+), 1 deletion(-)
diff --git a/arch/arm/kvm/arm.c b/arch/arm/kvm/arm.c
index bc738d2..ee5d667 100644
--- a/arch/arm/kvm/arm.c
+++ b/arch/arm/kvm/arm.c
@@ -28,6 +28,7 @@
#include <linux/sched.h>
#include <linux/kvm.h>
#include <trace/events/kvm.h>
+#include <kvm/arm_pmu.h>
#define CREATE_TRACE_POINTS
#include "trace.h"
@@ -547,6 +548,7 @@ int kvm_arch_vcpu_ioctl_run(struct kvm_vcpu *vcpu, struct kvm_run *run)
preempt_enable();
kvm_timer_sync_hwstate(vcpu);
kvm_vgic_sync_hwstate(vcpu);
+ kvm_pmu_sync_hwstate(vcpu);
continue;
}
@@ -591,6 +593,7 @@ int kvm_arch_vcpu_ioctl_run(struct kvm_vcpu *vcpu, struct kvm_run *run)
kvm_timer_sync_hwstate(vcpu);
kvm_vgic_sync_hwstate(vcpu);
+ kvm_pmu_sync_hwstate(vcpu);
ret = handle_exit(vcpu, run, ret);
}
diff --git a/include/kvm/arm_pmu.h b/include/kvm/arm_pmu.h
index 1a27357..5f1ea2b 100644
--- a/include/kvm/arm_pmu.h
+++ b/include/kvm/arm_pmu.h
@@ -37,6 +37,7 @@ struct kvm_pmu {
};
#ifdef CONFIG_KVM_ARM_PMU
+void kvm_pmu_sync_hwstate(struct kvm_vcpu *vcpu);
unsigned long kvm_pmu_get_counter_value(struct kvm_vcpu *vcpu,
unsigned long select_idx);
void kvm_pmu_disable_counter(struct kvm_vcpu *vcpu, unsigned long val);
@@ -45,6 +46,7 @@ void kvm_pmu_software_increment(struct kvm_vcpu *vcpu, unsigned long val);
void kvm_pmu_set_counter_event_type(struct kvm_vcpu *vcpu, unsigned long data,
unsigned long select_idx);
#else
+void kvm_pmu_sync_hwstate(struct kvm_vcpu *vcpu) {}
unsigned long kvm_pmu_get_counter_value(struct kvm_vcpu *vcpu,
unsigned long select_idx)
{
diff --git a/virt/kvm/arm/pmu.c b/virt/kvm/arm/pmu.c
index 5f5a483..9e75372 100644
--- a/virt/kvm/arm/pmu.c
+++ b/virt/kvm/arm/pmu.c
@@ -21,6 +21,7 @@
#include <linux/perf_event.h>
#include <asm/kvm_emulate.h>
#include <kvm/arm_pmu.h>
+#include <kvm/arm_vgic.h>
/* PMU HW events mapping. */
static struct kvm_pmu_hw_event_map {
@@ -90,6 +91,56 @@ static void kvm_pmu_stop_counter(struct kvm_vcpu *vcpu,
}
/**
+ * kvm_pmu_sync_hwstate - sync pmu state for cpu
+ * @vcpu: The vcpu pointer
+ *
+ * Inject virtual PMU IRQ if IRQ is pending for this cpu.
+ */
+void kvm_pmu_sync_hwstate(struct kvm_vcpu *vcpu)
+{
+ struct kvm_pmu *pmu = &vcpu->arch.pmu;
+
+ if (pmu->irq_pending && (pmu->irq_num != -1)) {
+ kvm_vgic_inject_irq(vcpu->kvm, vcpu->vcpu_id, pmu->irq_num, 1);
+ pmu->irq_pending = false;
+ }
+}
+
+/**
+ * When perf event overflows, set irq_pending and call kvm_vcpu_kick() to inject
+ * the interrupt.
+ */
+static void kvm_pmu_perf_overflow(struct perf_event *perf_event,
+ struct perf_sample_data *data,
+ struct pt_regs *regs)
+{
+ struct kvm_pmc *pmc = perf_event->overflow_handler_context;
+ struct kvm_vcpu *vcpu = pmc->vcpu;
+ struct kvm_pmu *pmu = &vcpu->arch.pmu;
+ int idx = pmc->idx;
+
+ if (!vcpu_mode_is_32bit(vcpu)) {
+ if ((vcpu_sys_reg(vcpu, PMINTENSET_EL1) >> idx) & 0x1) {
+ __set_bit(idx,
+ (unsigned long *)&vcpu_sys_reg(vcpu, PMOVSSET_EL0));
+ __set_bit(idx,
+ (unsigned long *)&vcpu_sys_reg(vcpu, PMOVSCLR_EL0));
+ pmu->irq_pending = true;
+ kvm_vcpu_kick(vcpu);
+ }
+ } else {
+ if ((vcpu_cp15(vcpu, c9_PMINTENSET) >> idx) & 0x1) {
+ __set_bit(idx,
+ (unsigned long *)&vcpu_cp15(vcpu, c9_PMOVSSET));
+ __set_bit(idx,
+ (unsigned long *)&vcpu_cp15(vcpu, c9_PMOVSCLR));
+ pmu->irq_pending = true;
+ kvm_vcpu_kick(vcpu);
+ }
+ }
+}
+
+/**
* kvm_pmu_get_counter_value - get PMU counter value
* @vcpu: The vcpu pointer
* @select_idx: The counter index
@@ -311,7 +362,8 @@ void kvm_pmu_set_counter_event_type(struct kvm_vcpu *vcpu, unsigned long data,
/* The initial sample period (overflow count) of an event. */
attr.sample_period = (-counter) & (((u64)1 << overflow_bit) - 1);
- event = perf_event_create_kernel_counter(&attr, -1, current, NULL, pmc);
+ event = perf_event_create_kernel_counter(&attr, -1, current,
+ kvm_pmu_perf_overflow, pmc);
if (IS_ERR(event)) {
printk_once("kvm: pmu event creation failed %ld\n",
PTR_ERR(event));
--
2.0.4
^ permalink raw reply related [flat|nested] 113+ messages in thread
* [PATCH v2 19/22] KVM: ARM64: Add PMU overflow interrupt routing
@ 2015-09-11 8:55 ` Shannon Zhao
0 siblings, 0 replies; 113+ messages in thread
From: Shannon Zhao @ 2015-09-11 8:55 UTC (permalink / raw)
To: kvmarm; +Cc: kvm, marc.zyngier, will.deacon, linux-arm-kernel, shannon.zhao
From: Shannon Zhao <shannon.zhao@linaro.org>
When calling perf_event_create_kernel_counter to create perf_event,
assign a overflow handler. Then when perf event overflows, set
irq_pending and call kvm_vcpu_kick() to sync the interrupt.
Signed-off-by: Shannon Zhao <shannon.zhao@linaro.org>
---
arch/arm/kvm/arm.c | 3 +++
include/kvm/arm_pmu.h | 2 ++
virt/kvm/arm/pmu.c | 54 ++++++++++++++++++++++++++++++++++++++++++++++++++-
3 files changed, 58 insertions(+), 1 deletion(-)
diff --git a/arch/arm/kvm/arm.c b/arch/arm/kvm/arm.c
index bc738d2..ee5d667 100644
--- a/arch/arm/kvm/arm.c
+++ b/arch/arm/kvm/arm.c
@@ -28,6 +28,7 @@
#include <linux/sched.h>
#include <linux/kvm.h>
#include <trace/events/kvm.h>
+#include <kvm/arm_pmu.h>
#define CREATE_TRACE_POINTS
#include "trace.h"
@@ -547,6 +548,7 @@ int kvm_arch_vcpu_ioctl_run(struct kvm_vcpu *vcpu, struct kvm_run *run)
preempt_enable();
kvm_timer_sync_hwstate(vcpu);
kvm_vgic_sync_hwstate(vcpu);
+ kvm_pmu_sync_hwstate(vcpu);
continue;
}
@@ -591,6 +593,7 @@ int kvm_arch_vcpu_ioctl_run(struct kvm_vcpu *vcpu, struct kvm_run *run)
kvm_timer_sync_hwstate(vcpu);
kvm_vgic_sync_hwstate(vcpu);
+ kvm_pmu_sync_hwstate(vcpu);
ret = handle_exit(vcpu, run, ret);
}
diff --git a/include/kvm/arm_pmu.h b/include/kvm/arm_pmu.h
index 1a27357..5f1ea2b 100644
--- a/include/kvm/arm_pmu.h
+++ b/include/kvm/arm_pmu.h
@@ -37,6 +37,7 @@ struct kvm_pmu {
};
#ifdef CONFIG_KVM_ARM_PMU
+void kvm_pmu_sync_hwstate(struct kvm_vcpu *vcpu);
unsigned long kvm_pmu_get_counter_value(struct kvm_vcpu *vcpu,
unsigned long select_idx);
void kvm_pmu_disable_counter(struct kvm_vcpu *vcpu, unsigned long val);
@@ -45,6 +46,7 @@ void kvm_pmu_software_increment(struct kvm_vcpu *vcpu, unsigned long val);
void kvm_pmu_set_counter_event_type(struct kvm_vcpu *vcpu, unsigned long data,
unsigned long select_idx);
#else
+void kvm_pmu_sync_hwstate(struct kvm_vcpu *vcpu) {}
unsigned long kvm_pmu_get_counter_value(struct kvm_vcpu *vcpu,
unsigned long select_idx)
{
diff --git a/virt/kvm/arm/pmu.c b/virt/kvm/arm/pmu.c
index 5f5a483..9e75372 100644
--- a/virt/kvm/arm/pmu.c
+++ b/virt/kvm/arm/pmu.c
@@ -21,6 +21,7 @@
#include <linux/perf_event.h>
#include <asm/kvm_emulate.h>
#include <kvm/arm_pmu.h>
+#include <kvm/arm_vgic.h>
/* PMU HW events mapping. */
static struct kvm_pmu_hw_event_map {
@@ -90,6 +91,56 @@ static void kvm_pmu_stop_counter(struct kvm_vcpu *vcpu,
}
/**
+ * kvm_pmu_sync_hwstate - sync pmu state for cpu
+ * @vcpu: The vcpu pointer
+ *
+ * Inject virtual PMU IRQ if IRQ is pending for this cpu.
+ */
+void kvm_pmu_sync_hwstate(struct kvm_vcpu *vcpu)
+{
+ struct kvm_pmu *pmu = &vcpu->arch.pmu;
+
+ if (pmu->irq_pending && (pmu->irq_num != -1)) {
+ kvm_vgic_inject_irq(vcpu->kvm, vcpu->vcpu_id, pmu->irq_num, 1);
+ pmu->irq_pending = false;
+ }
+}
+
+/**
+ * When perf event overflows, set irq_pending and call kvm_vcpu_kick() to inject
+ * the interrupt.
+ */
+static void kvm_pmu_perf_overflow(struct perf_event *perf_event,
+ struct perf_sample_data *data,
+ struct pt_regs *regs)
+{
+ struct kvm_pmc *pmc = perf_event->overflow_handler_context;
+ struct kvm_vcpu *vcpu = pmc->vcpu;
+ struct kvm_pmu *pmu = &vcpu->arch.pmu;
+ int idx = pmc->idx;
+
+ if (!vcpu_mode_is_32bit(vcpu)) {
+ if ((vcpu_sys_reg(vcpu, PMINTENSET_EL1) >> idx) & 0x1) {
+ __set_bit(idx,
+ (unsigned long *)&vcpu_sys_reg(vcpu, PMOVSSET_EL0));
+ __set_bit(idx,
+ (unsigned long *)&vcpu_sys_reg(vcpu, PMOVSCLR_EL0));
+ pmu->irq_pending = true;
+ kvm_vcpu_kick(vcpu);
+ }
+ } else {
+ if ((vcpu_cp15(vcpu, c9_PMINTENSET) >> idx) & 0x1) {
+ __set_bit(idx,
+ (unsigned long *)&vcpu_cp15(vcpu, c9_PMOVSSET));
+ __set_bit(idx,
+ (unsigned long *)&vcpu_cp15(vcpu, c9_PMOVSCLR));
+ pmu->irq_pending = true;
+ kvm_vcpu_kick(vcpu);
+ }
+ }
+}
+
+/**
* kvm_pmu_get_counter_value - get PMU counter value
* @vcpu: The vcpu pointer
* @select_idx: The counter index
@@ -311,7 +362,8 @@ void kvm_pmu_set_counter_event_type(struct kvm_vcpu *vcpu, unsigned long data,
/* The initial sample period (overflow count) of an event. */
attr.sample_period = (-counter) & (((u64)1 << overflow_bit) - 1);
- event = perf_event_create_kernel_counter(&attr, -1, current, NULL, pmc);
+ event = perf_event_create_kernel_counter(&attr, -1, current,
+ kvm_pmu_perf_overflow, pmc);
if (IS_ERR(event)) {
printk_once("kvm: pmu event creation failed %ld\n",
PTR_ERR(event));
--
2.0.4
^ permalink raw reply related [flat|nested] 113+ messages in thread
* [PATCH v2 19/22] KVM: ARM64: Add PMU overflow interrupt routing
@ 2015-09-11 8:55 ` Shannon Zhao
0 siblings, 0 replies; 113+ messages in thread
From: Shannon Zhao @ 2015-09-11 8:55 UTC (permalink / raw)
To: kvmarm; +Cc: kvm, marc.zyngier, will.deacon, linux-arm-kernel, shannon.zhao
From: Shannon Zhao <shannon.zhao@linaro.org>
When calling perf_event_create_kernel_counter to create perf_event,
assign a overflow handler. Then when perf event overflows, set
irq_pending and call kvm_vcpu_kick() to sync the interrupt.
Signed-off-by: Shannon Zhao <shannon.zhao@linaro.org>
---
arch/arm/kvm/arm.c | 3 +++
include/kvm/arm_pmu.h | 2 ++
virt/kvm/arm/pmu.c | 54 ++++++++++++++++++++++++++++++++++++++++++++++++++-
3 files changed, 58 insertions(+), 1 deletion(-)
diff --git a/arch/arm/kvm/arm.c b/arch/arm/kvm/arm.c
index bc738d2..ee5d667 100644
--- a/arch/arm/kvm/arm.c
+++ b/arch/arm/kvm/arm.c
@@ -28,6 +28,7 @@
#include <linux/sched.h>
#include <linux/kvm.h>
#include <trace/events/kvm.h>
+#include <kvm/arm_pmu.h>
#define CREATE_TRACE_POINTS
#include "trace.h"
@@ -547,6 +548,7 @@ int kvm_arch_vcpu_ioctl_run(struct kvm_vcpu *vcpu, struct kvm_run *run)
preempt_enable();
kvm_timer_sync_hwstate(vcpu);
kvm_vgic_sync_hwstate(vcpu);
+ kvm_pmu_sync_hwstate(vcpu);
continue;
}
@@ -591,6 +593,7 @@ int kvm_arch_vcpu_ioctl_run(struct kvm_vcpu *vcpu, struct kvm_run *run)
kvm_timer_sync_hwstate(vcpu);
kvm_vgic_sync_hwstate(vcpu);
+ kvm_pmu_sync_hwstate(vcpu);
ret = handle_exit(vcpu, run, ret);
}
diff --git a/include/kvm/arm_pmu.h b/include/kvm/arm_pmu.h
index 1a27357..5f1ea2b 100644
--- a/include/kvm/arm_pmu.h
+++ b/include/kvm/arm_pmu.h
@@ -37,6 +37,7 @@ struct kvm_pmu {
};
#ifdef CONFIG_KVM_ARM_PMU
+void kvm_pmu_sync_hwstate(struct kvm_vcpu *vcpu);
unsigned long kvm_pmu_get_counter_value(struct kvm_vcpu *vcpu,
unsigned long select_idx);
void kvm_pmu_disable_counter(struct kvm_vcpu *vcpu, unsigned long val);
@@ -45,6 +46,7 @@ void kvm_pmu_software_increment(struct kvm_vcpu *vcpu, unsigned long val);
void kvm_pmu_set_counter_event_type(struct kvm_vcpu *vcpu, unsigned long data,
unsigned long select_idx);
#else
+void kvm_pmu_sync_hwstate(struct kvm_vcpu *vcpu) {}
unsigned long kvm_pmu_get_counter_value(struct kvm_vcpu *vcpu,
unsigned long select_idx)
{
diff --git a/virt/kvm/arm/pmu.c b/virt/kvm/arm/pmu.c
index 5f5a483..9e75372 100644
--- a/virt/kvm/arm/pmu.c
+++ b/virt/kvm/arm/pmu.c
@@ -21,6 +21,7 @@
#include <linux/perf_event.h>
#include <asm/kvm_emulate.h>
#include <kvm/arm_pmu.h>
+#include <kvm/arm_vgic.h>
/* PMU HW events mapping. */
static struct kvm_pmu_hw_event_map {
@@ -90,6 +91,56 @@ static void kvm_pmu_stop_counter(struct kvm_vcpu *vcpu,
}
/**
+ * kvm_pmu_sync_hwstate - sync pmu state for cpu
+ * @vcpu: The vcpu pointer
+ *
+ * Inject virtual PMU IRQ if IRQ is pending for this cpu.
+ */
+void kvm_pmu_sync_hwstate(struct kvm_vcpu *vcpu)
+{
+ struct kvm_pmu *pmu = &vcpu->arch.pmu;
+
+ if (pmu->irq_pending && (pmu->irq_num != -1)) {
+ kvm_vgic_inject_irq(vcpu->kvm, vcpu->vcpu_id, pmu->irq_num, 1);
+ pmu->irq_pending = false;
+ }
+}
+
+/**
+ * When perf event overflows, set irq_pending and call kvm_vcpu_kick() to inject
+ * the interrupt.
+ */
+static void kvm_pmu_perf_overflow(struct perf_event *perf_event,
+ struct perf_sample_data *data,
+ struct pt_regs *regs)
+{
+ struct kvm_pmc *pmc = perf_event->overflow_handler_context;
+ struct kvm_vcpu *vcpu = pmc->vcpu;
+ struct kvm_pmu *pmu = &vcpu->arch.pmu;
+ int idx = pmc->idx;
+
+ if (!vcpu_mode_is_32bit(vcpu)) {
+ if ((vcpu_sys_reg(vcpu, PMINTENSET_EL1) >> idx) & 0x1) {
+ __set_bit(idx,
+ (unsigned long *)&vcpu_sys_reg(vcpu, PMOVSSET_EL0));
+ __set_bit(idx,
+ (unsigned long *)&vcpu_sys_reg(vcpu, PMOVSCLR_EL0));
+ pmu->irq_pending = true;
+ kvm_vcpu_kick(vcpu);
+ }
+ } else {
+ if ((vcpu_cp15(vcpu, c9_PMINTENSET) >> idx) & 0x1) {
+ __set_bit(idx,
+ (unsigned long *)&vcpu_cp15(vcpu, c9_PMOVSSET));
+ __set_bit(idx,
+ (unsigned long *)&vcpu_cp15(vcpu, c9_PMOVSCLR));
+ pmu->irq_pending = true;
+ kvm_vcpu_kick(vcpu);
+ }
+ }
+}
+
+/**
* kvm_pmu_get_counter_value - get PMU counter value
* @vcpu: The vcpu pointer
* @select_idx: The counter index
@@ -311,7 +362,8 @@ void kvm_pmu_set_counter_event_type(struct kvm_vcpu *vcpu, unsigned long data,
/* The initial sample period (overflow count) of an event. */
attr.sample_period = (-counter) & (((u64)1 << overflow_bit) - 1);
- event = perf_event_create_kernel_counter(&attr, -1, current, NULL, pmc);
+ event = perf_event_create_kernel_counter(&attr, -1, current,
+ kvm_pmu_perf_overflow, pmc);
if (IS_ERR(event)) {
printk_once("kvm: pmu event creation failed %ld\n",
PTR_ERR(event));
--
2.0.4
^ permalink raw reply related [flat|nested] 113+ messages in thread
* [PATCH v2 20/22] KVM: ARM64: Reset PMU state when resetting vcpu
2015-09-11 8:54 ` Shannon Zhao
(?)
@ 2015-09-11 8:55 ` Shannon Zhao
-1 siblings, 0 replies; 113+ messages in thread
From: Shannon Zhao @ 2015-09-11 8:55 UTC (permalink / raw)
To: linux-arm-kernel
From: Shannon Zhao <shannon.zhao@linaro.org>
Signed-off-by: Shannon Zhao <shannon.zhao@linaro.org>
---
arch/arm64/kvm/reset.c | 3 +++
include/kvm/arm_pmu.h | 2 ++
virt/kvm/arm/pmu.c | 18 ++++++++++++++++++
3 files changed, 23 insertions(+)
diff --git a/arch/arm64/kvm/reset.c b/arch/arm64/kvm/reset.c
index 0b43265..ee2c2e9 100644
--- a/arch/arm64/kvm/reset.c
+++ b/arch/arm64/kvm/reset.c
@@ -107,5 +107,8 @@ int kvm_reset_vcpu(struct kvm_vcpu *vcpu)
/* Reset timer */
kvm_timer_vcpu_reset(vcpu, cpu_vtimer_irq);
+ /* Reset PMU */
+ kvm_pmu_vcpu_reset(vcpu);
+
return 0;
}
diff --git a/include/kvm/arm_pmu.h b/include/kvm/arm_pmu.h
index 5f1ea2b..d74e183 100644
--- a/include/kvm/arm_pmu.h
+++ b/include/kvm/arm_pmu.h
@@ -37,6 +37,7 @@ struct kvm_pmu {
};
#ifdef CONFIG_KVM_ARM_PMU
+void kvm_pmu_vcpu_reset(struct kvm_vcpu *vcpu);
void kvm_pmu_sync_hwstate(struct kvm_vcpu *vcpu);
unsigned long kvm_pmu_get_counter_value(struct kvm_vcpu *vcpu,
unsigned long select_idx);
@@ -46,6 +47,7 @@ void kvm_pmu_software_increment(struct kvm_vcpu *vcpu, unsigned long val);
void kvm_pmu_set_counter_event_type(struct kvm_vcpu *vcpu, unsigned long data,
unsigned long select_idx);
#else
+void kvm_pmu_vcpu_reset(struct kvm_vcpu *vcpu) {}
void kvm_pmu_sync_hwstate(struct kvm_vcpu *vcpu) {}
unsigned long kvm_pmu_get_counter_value(struct kvm_vcpu *vcpu,
unsigned long select_idx)
diff --git a/virt/kvm/arm/pmu.c b/virt/kvm/arm/pmu.c
index 9e75372..8476d5e 100644
--- a/virt/kvm/arm/pmu.c
+++ b/virt/kvm/arm/pmu.c
@@ -91,6 +91,24 @@ static void kvm_pmu_stop_counter(struct kvm_vcpu *vcpu,
}
/**
+ * kvm_pmu_vcpu_reset - reset pmu state for cpu
+ * @vcpu: The vcpu pointer
+ *
+ */
+void kvm_pmu_vcpu_reset(struct kvm_vcpu *vcpu)
+{
+ int i;
+ struct kvm_pmu *pmu = &vcpu->arch.pmu;
+
+ for (i = 0; i < ARMV8_MAX_COUNTERS; i++) {
+ kvm_pmu_stop_counter(vcpu, i);
+ pmu->pmc[i].idx = i;
+ pmu->pmc[i].vcpu = vcpu;
+ }
+ pmu->irq_pending = false;
+}
+
+/**
* kvm_pmu_sync_hwstate - sync pmu state for cpu
* @vcpu: The vcpu pointer
*
--
2.0.4
^ permalink raw reply related [flat|nested] 113+ messages in thread
* [PATCH v2 20/22] KVM: ARM64: Reset PMU state when resetting vcpu
@ 2015-09-11 8:55 ` Shannon Zhao
0 siblings, 0 replies; 113+ messages in thread
From: Shannon Zhao @ 2015-09-11 8:55 UTC (permalink / raw)
To: kvmarm; +Cc: kvm, marc.zyngier, will.deacon, linux-arm-kernel, shannon.zhao
From: Shannon Zhao <shannon.zhao@linaro.org>
Signed-off-by: Shannon Zhao <shannon.zhao@linaro.org>
---
arch/arm64/kvm/reset.c | 3 +++
include/kvm/arm_pmu.h | 2 ++
virt/kvm/arm/pmu.c | 18 ++++++++++++++++++
3 files changed, 23 insertions(+)
diff --git a/arch/arm64/kvm/reset.c b/arch/arm64/kvm/reset.c
index 0b43265..ee2c2e9 100644
--- a/arch/arm64/kvm/reset.c
+++ b/arch/arm64/kvm/reset.c
@@ -107,5 +107,8 @@ int kvm_reset_vcpu(struct kvm_vcpu *vcpu)
/* Reset timer */
kvm_timer_vcpu_reset(vcpu, cpu_vtimer_irq);
+ /* Reset PMU */
+ kvm_pmu_vcpu_reset(vcpu);
+
return 0;
}
diff --git a/include/kvm/arm_pmu.h b/include/kvm/arm_pmu.h
index 5f1ea2b..d74e183 100644
--- a/include/kvm/arm_pmu.h
+++ b/include/kvm/arm_pmu.h
@@ -37,6 +37,7 @@ struct kvm_pmu {
};
#ifdef CONFIG_KVM_ARM_PMU
+void kvm_pmu_vcpu_reset(struct kvm_vcpu *vcpu);
void kvm_pmu_sync_hwstate(struct kvm_vcpu *vcpu);
unsigned long kvm_pmu_get_counter_value(struct kvm_vcpu *vcpu,
unsigned long select_idx);
@@ -46,6 +47,7 @@ void kvm_pmu_software_increment(struct kvm_vcpu *vcpu, unsigned long val);
void kvm_pmu_set_counter_event_type(struct kvm_vcpu *vcpu, unsigned long data,
unsigned long select_idx);
#else
+void kvm_pmu_vcpu_reset(struct kvm_vcpu *vcpu) {}
void kvm_pmu_sync_hwstate(struct kvm_vcpu *vcpu) {}
unsigned long kvm_pmu_get_counter_value(struct kvm_vcpu *vcpu,
unsigned long select_idx)
diff --git a/virt/kvm/arm/pmu.c b/virt/kvm/arm/pmu.c
index 9e75372..8476d5e 100644
--- a/virt/kvm/arm/pmu.c
+++ b/virt/kvm/arm/pmu.c
@@ -91,6 +91,24 @@ static void kvm_pmu_stop_counter(struct kvm_vcpu *vcpu,
}
/**
+ * kvm_pmu_vcpu_reset - reset pmu state for cpu
+ * @vcpu: The vcpu pointer
+ *
+ */
+void kvm_pmu_vcpu_reset(struct kvm_vcpu *vcpu)
+{
+ int i;
+ struct kvm_pmu *pmu = &vcpu->arch.pmu;
+
+ for (i = 0; i < ARMV8_MAX_COUNTERS; i++) {
+ kvm_pmu_stop_counter(vcpu, i);
+ pmu->pmc[i].idx = i;
+ pmu->pmc[i].vcpu = vcpu;
+ }
+ pmu->irq_pending = false;
+}
+
+/**
* kvm_pmu_sync_hwstate - sync pmu state for cpu
* @vcpu: The vcpu pointer
*
--
2.0.4
^ permalink raw reply related [flat|nested] 113+ messages in thread
* [PATCH v2 20/22] KVM: ARM64: Reset PMU state when resetting vcpu
@ 2015-09-11 8:55 ` Shannon Zhao
0 siblings, 0 replies; 113+ messages in thread
From: Shannon Zhao @ 2015-09-11 8:55 UTC (permalink / raw)
To: kvmarm; +Cc: kvm, marc.zyngier, will.deacon, linux-arm-kernel, shannon.zhao
From: Shannon Zhao <shannon.zhao@linaro.org>
Signed-off-by: Shannon Zhao <shannon.zhao@linaro.org>
---
arch/arm64/kvm/reset.c | 3 +++
include/kvm/arm_pmu.h | 2 ++
virt/kvm/arm/pmu.c | 18 ++++++++++++++++++
3 files changed, 23 insertions(+)
diff --git a/arch/arm64/kvm/reset.c b/arch/arm64/kvm/reset.c
index 0b43265..ee2c2e9 100644
--- a/arch/arm64/kvm/reset.c
+++ b/arch/arm64/kvm/reset.c
@@ -107,5 +107,8 @@ int kvm_reset_vcpu(struct kvm_vcpu *vcpu)
/* Reset timer */
kvm_timer_vcpu_reset(vcpu, cpu_vtimer_irq);
+ /* Reset PMU */
+ kvm_pmu_vcpu_reset(vcpu);
+
return 0;
}
diff --git a/include/kvm/arm_pmu.h b/include/kvm/arm_pmu.h
index 5f1ea2b..d74e183 100644
--- a/include/kvm/arm_pmu.h
+++ b/include/kvm/arm_pmu.h
@@ -37,6 +37,7 @@ struct kvm_pmu {
};
#ifdef CONFIG_KVM_ARM_PMU
+void kvm_pmu_vcpu_reset(struct kvm_vcpu *vcpu);
void kvm_pmu_sync_hwstate(struct kvm_vcpu *vcpu);
unsigned long kvm_pmu_get_counter_value(struct kvm_vcpu *vcpu,
unsigned long select_idx);
@@ -46,6 +47,7 @@ void kvm_pmu_software_increment(struct kvm_vcpu *vcpu, unsigned long val);
void kvm_pmu_set_counter_event_type(struct kvm_vcpu *vcpu, unsigned long data,
unsigned long select_idx);
#else
+void kvm_pmu_vcpu_reset(struct kvm_vcpu *vcpu) {}
void kvm_pmu_sync_hwstate(struct kvm_vcpu *vcpu) {}
unsigned long kvm_pmu_get_counter_value(struct kvm_vcpu *vcpu,
unsigned long select_idx)
diff --git a/virt/kvm/arm/pmu.c b/virt/kvm/arm/pmu.c
index 9e75372..8476d5e 100644
--- a/virt/kvm/arm/pmu.c
+++ b/virt/kvm/arm/pmu.c
@@ -91,6 +91,24 @@ static void kvm_pmu_stop_counter(struct kvm_vcpu *vcpu,
}
/**
+ * kvm_pmu_vcpu_reset - reset pmu state for cpu
+ * @vcpu: The vcpu pointer
+ *
+ */
+void kvm_pmu_vcpu_reset(struct kvm_vcpu *vcpu)
+{
+ int i;
+ struct kvm_pmu *pmu = &vcpu->arch.pmu;
+
+ for (i = 0; i < ARMV8_MAX_COUNTERS; i++) {
+ kvm_pmu_stop_counter(vcpu, i);
+ pmu->pmc[i].idx = i;
+ pmu->pmc[i].vcpu = vcpu;
+ }
+ pmu->irq_pending = false;
+}
+
+/**
* kvm_pmu_sync_hwstate - sync pmu state for cpu
* @vcpu: The vcpu pointer
*
--
2.0.4
^ permalink raw reply related [flat|nested] 113+ messages in thread
* [PATCH v2 21/22] KVM: ARM64: Free perf event of PMU when destroying vcpu
2015-09-11 8:54 ` Shannon Zhao
(?)
@ 2015-09-11 8:55 ` Shannon Zhao
-1 siblings, 0 replies; 113+ messages in thread
From: Shannon Zhao @ 2015-09-11 8:55 UTC (permalink / raw)
To: linux-arm-kernel
From: Shannon Zhao <shannon.zhao@linaro.org>
When KVM frees VCPU, it needs to free the perf_event of PMU.
Signed-off-by: Shannon Zhao <shannon.zhao@linaro.org>
---
arch/arm/kvm/arm.c | 1 +
include/kvm/arm_pmu.h | 2 ++
virt/kvm/arm/pmu.c | 21 +++++++++++++++++++++
3 files changed, 24 insertions(+)
diff --git a/arch/arm/kvm/arm.c b/arch/arm/kvm/arm.c
index ee5d667..a7215c0 100644
--- a/arch/arm/kvm/arm.c
+++ b/arch/arm/kvm/arm.c
@@ -257,6 +257,7 @@ void kvm_arch_vcpu_free(struct kvm_vcpu *vcpu)
kvm_mmu_free_memory_caches(vcpu);
kvm_timer_vcpu_terminate(vcpu);
kvm_vgic_vcpu_destroy(vcpu);
+ kvm_pmu_vcpu_destroy(vcpu);
kmem_cache_free(kvm_vcpu_cache, vcpu);
}
diff --git a/include/kvm/arm_pmu.h b/include/kvm/arm_pmu.h
index d74e183..2b6fcb1 100644
--- a/include/kvm/arm_pmu.h
+++ b/include/kvm/arm_pmu.h
@@ -38,6 +38,7 @@ struct kvm_pmu {
#ifdef CONFIG_KVM_ARM_PMU
void kvm_pmu_vcpu_reset(struct kvm_vcpu *vcpu);
+void kvm_pmu_vcpu_destroy(struct kvm_vcpu *vcpu);
void kvm_pmu_sync_hwstate(struct kvm_vcpu *vcpu);
unsigned long kvm_pmu_get_counter_value(struct kvm_vcpu *vcpu,
unsigned long select_idx);
@@ -48,6 +49,7 @@ void kvm_pmu_set_counter_event_type(struct kvm_vcpu *vcpu, unsigned long data,
unsigned long select_idx);
#else
void kvm_pmu_vcpu_reset(struct kvm_vcpu *vcpu) {}
+void kvm_pmu_vcpu_destroy(struct kvm_vcpu *vcpu) {}
void kvm_pmu_sync_hwstate(struct kvm_vcpu *vcpu) {}
unsigned long kvm_pmu_get_counter_value(struct kvm_vcpu *vcpu,
unsigned long select_idx)
diff --git a/virt/kvm/arm/pmu.c b/virt/kvm/arm/pmu.c
index 8476d5e..8eccac9 100644
--- a/virt/kvm/arm/pmu.c
+++ b/virt/kvm/arm/pmu.c
@@ -109,6 +109,27 @@ void kvm_pmu_vcpu_reset(struct kvm_vcpu *vcpu)
}
/**
+ * kvm_pmu_vcpu_destroy - free perf event of PMU for cpu
+ * @vcpu: The vcpu pointer
+ *
+ */
+void kvm_pmu_vcpu_destroy(struct kvm_vcpu *vcpu)
+{
+ int i;
+ struct kvm_pmu *pmu = &vcpu->arch.pmu;
+
+ for (i = 0; i < ARMV8_MAX_COUNTERS; i++) {
+ struct kvm_pmc *pmc = &pmu->pmc[i];
+
+ if (pmc->perf_event) {
+ perf_event_disable(pmc->perf_event);
+ perf_event_release_kernel(pmc->perf_event);
+ pmc->perf_event = NULL;
+ }
+ }
+}
+
+/**
* kvm_pmu_sync_hwstate - sync pmu state for cpu
* @vcpu: The vcpu pointer
*
--
2.0.4
^ permalink raw reply related [flat|nested] 113+ messages in thread
* [PATCH v2 21/22] KVM: ARM64: Free perf event of PMU when destroying vcpu
@ 2015-09-11 8:55 ` Shannon Zhao
0 siblings, 0 replies; 113+ messages in thread
From: Shannon Zhao @ 2015-09-11 8:55 UTC (permalink / raw)
To: kvmarm; +Cc: kvm, marc.zyngier, will.deacon, linux-arm-kernel, shannon.zhao
From: Shannon Zhao <shannon.zhao@linaro.org>
When KVM frees VCPU, it needs to free the perf_event of PMU.
Signed-off-by: Shannon Zhao <shannon.zhao@linaro.org>
---
arch/arm/kvm/arm.c | 1 +
include/kvm/arm_pmu.h | 2 ++
virt/kvm/arm/pmu.c | 21 +++++++++++++++++++++
3 files changed, 24 insertions(+)
diff --git a/arch/arm/kvm/arm.c b/arch/arm/kvm/arm.c
index ee5d667..a7215c0 100644
--- a/arch/arm/kvm/arm.c
+++ b/arch/arm/kvm/arm.c
@@ -257,6 +257,7 @@ void kvm_arch_vcpu_free(struct kvm_vcpu *vcpu)
kvm_mmu_free_memory_caches(vcpu);
kvm_timer_vcpu_terminate(vcpu);
kvm_vgic_vcpu_destroy(vcpu);
+ kvm_pmu_vcpu_destroy(vcpu);
kmem_cache_free(kvm_vcpu_cache, vcpu);
}
diff --git a/include/kvm/arm_pmu.h b/include/kvm/arm_pmu.h
index d74e183..2b6fcb1 100644
--- a/include/kvm/arm_pmu.h
+++ b/include/kvm/arm_pmu.h
@@ -38,6 +38,7 @@ struct kvm_pmu {
#ifdef CONFIG_KVM_ARM_PMU
void kvm_pmu_vcpu_reset(struct kvm_vcpu *vcpu);
+void kvm_pmu_vcpu_destroy(struct kvm_vcpu *vcpu);
void kvm_pmu_sync_hwstate(struct kvm_vcpu *vcpu);
unsigned long kvm_pmu_get_counter_value(struct kvm_vcpu *vcpu,
unsigned long select_idx);
@@ -48,6 +49,7 @@ void kvm_pmu_set_counter_event_type(struct kvm_vcpu *vcpu, unsigned long data,
unsigned long select_idx);
#else
void kvm_pmu_vcpu_reset(struct kvm_vcpu *vcpu) {}
+void kvm_pmu_vcpu_destroy(struct kvm_vcpu *vcpu) {}
void kvm_pmu_sync_hwstate(struct kvm_vcpu *vcpu) {}
unsigned long kvm_pmu_get_counter_value(struct kvm_vcpu *vcpu,
unsigned long select_idx)
diff --git a/virt/kvm/arm/pmu.c b/virt/kvm/arm/pmu.c
index 8476d5e..8eccac9 100644
--- a/virt/kvm/arm/pmu.c
+++ b/virt/kvm/arm/pmu.c
@@ -109,6 +109,27 @@ void kvm_pmu_vcpu_reset(struct kvm_vcpu *vcpu)
}
/**
+ * kvm_pmu_vcpu_destroy - free perf event of PMU for cpu
+ * @vcpu: The vcpu pointer
+ *
+ */
+void kvm_pmu_vcpu_destroy(struct kvm_vcpu *vcpu)
+{
+ int i;
+ struct kvm_pmu *pmu = &vcpu->arch.pmu;
+
+ for (i = 0; i < ARMV8_MAX_COUNTERS; i++) {
+ struct kvm_pmc *pmc = &pmu->pmc[i];
+
+ if (pmc->perf_event) {
+ perf_event_disable(pmc->perf_event);
+ perf_event_release_kernel(pmc->perf_event);
+ pmc->perf_event = NULL;
+ }
+ }
+}
+
+/**
* kvm_pmu_sync_hwstate - sync pmu state for cpu
* @vcpu: The vcpu pointer
*
--
2.0.4
^ permalink raw reply related [flat|nested] 113+ messages in thread
* [PATCH v2 21/22] KVM: ARM64: Free perf event of PMU when destroying vcpu
@ 2015-09-11 8:55 ` Shannon Zhao
0 siblings, 0 replies; 113+ messages in thread
From: Shannon Zhao @ 2015-09-11 8:55 UTC (permalink / raw)
To: kvmarm; +Cc: kvm, marc.zyngier, will.deacon, linux-arm-kernel, shannon.zhao
From: Shannon Zhao <shannon.zhao@linaro.org>
When KVM frees VCPU, it needs to free the perf_event of PMU.
Signed-off-by: Shannon Zhao <shannon.zhao@linaro.org>
---
arch/arm/kvm/arm.c | 1 +
include/kvm/arm_pmu.h | 2 ++
virt/kvm/arm/pmu.c | 21 +++++++++++++++++++++
3 files changed, 24 insertions(+)
diff --git a/arch/arm/kvm/arm.c b/arch/arm/kvm/arm.c
index ee5d667..a7215c0 100644
--- a/arch/arm/kvm/arm.c
+++ b/arch/arm/kvm/arm.c
@@ -257,6 +257,7 @@ void kvm_arch_vcpu_free(struct kvm_vcpu *vcpu)
kvm_mmu_free_memory_caches(vcpu);
kvm_timer_vcpu_terminate(vcpu);
kvm_vgic_vcpu_destroy(vcpu);
+ kvm_pmu_vcpu_destroy(vcpu);
kmem_cache_free(kvm_vcpu_cache, vcpu);
}
diff --git a/include/kvm/arm_pmu.h b/include/kvm/arm_pmu.h
index d74e183..2b6fcb1 100644
--- a/include/kvm/arm_pmu.h
+++ b/include/kvm/arm_pmu.h
@@ -38,6 +38,7 @@ struct kvm_pmu {
#ifdef CONFIG_KVM_ARM_PMU
void kvm_pmu_vcpu_reset(struct kvm_vcpu *vcpu);
+void kvm_pmu_vcpu_destroy(struct kvm_vcpu *vcpu);
void kvm_pmu_sync_hwstate(struct kvm_vcpu *vcpu);
unsigned long kvm_pmu_get_counter_value(struct kvm_vcpu *vcpu,
unsigned long select_idx);
@@ -48,6 +49,7 @@ void kvm_pmu_set_counter_event_type(struct kvm_vcpu *vcpu, unsigned long data,
unsigned long select_idx);
#else
void kvm_pmu_vcpu_reset(struct kvm_vcpu *vcpu) {}
+void kvm_pmu_vcpu_destroy(struct kvm_vcpu *vcpu) {}
void kvm_pmu_sync_hwstate(struct kvm_vcpu *vcpu) {}
unsigned long kvm_pmu_get_counter_value(struct kvm_vcpu *vcpu,
unsigned long select_idx)
diff --git a/virt/kvm/arm/pmu.c b/virt/kvm/arm/pmu.c
index 8476d5e..8eccac9 100644
--- a/virt/kvm/arm/pmu.c
+++ b/virt/kvm/arm/pmu.c
@@ -109,6 +109,27 @@ void kvm_pmu_vcpu_reset(struct kvm_vcpu *vcpu)
}
/**
+ * kvm_pmu_vcpu_destroy - free perf event of PMU for cpu
+ * @vcpu: The vcpu pointer
+ *
+ */
+void kvm_pmu_vcpu_destroy(struct kvm_vcpu *vcpu)
+{
+ int i;
+ struct kvm_pmu *pmu = &vcpu->arch.pmu;
+
+ for (i = 0; i < ARMV8_MAX_COUNTERS; i++) {
+ struct kvm_pmc *pmc = &pmu->pmc[i];
+
+ if (pmc->perf_event) {
+ perf_event_disable(pmc->perf_event);
+ perf_event_release_kernel(pmc->perf_event);
+ pmc->perf_event = NULL;
+ }
+ }
+}
+
+/**
* kvm_pmu_sync_hwstate - sync pmu state for cpu
* @vcpu: The vcpu pointer
*
--
2.0.4
^ permalink raw reply related [flat|nested] 113+ messages in thread
* [PATCH v2 22/22] KVM: ARM64: Add a new kvm ARM PMU device
2015-09-11 8:54 ` Shannon Zhao
(?)
@ 2015-09-11 8:55 ` Shannon Zhao
-1 siblings, 0 replies; 113+ messages in thread
From: Shannon Zhao @ 2015-09-11 8:55 UTC (permalink / raw)
To: linux-arm-kernel
From: Shannon Zhao <shannon.zhao@linaro.org>
Add a new kvm device type KVM_DEV_TYPE_ARM_PMU_V3 for ARM PMU. Implement
the kvm_device_ops for it.
Signed-off-by: Shannon Zhao <shannon.zhao@linaro.org>
---
Documentation/virtual/kvm/devices/arm-pmu.txt | 15 +++++
arch/arm64/include/uapi/asm/kvm.h | 3 +
include/linux/kvm_host.h | 1 +
include/uapi/linux/kvm.h | 2 +
virt/kvm/arm/pmu.c | 88 +++++++++++++++++++++++++++
virt/kvm/kvm_main.c | 4 ++
6 files changed, 113 insertions(+)
create mode 100644 Documentation/virtual/kvm/devices/arm-pmu.txt
diff --git a/Documentation/virtual/kvm/devices/arm-pmu.txt b/Documentation/virtual/kvm/devices/arm-pmu.txt
new file mode 100644
index 0000000..49481c4
--- /dev/null
+++ b/Documentation/virtual/kvm/devices/arm-pmu.txt
@@ -0,0 +1,15 @@
+ARM Virtual Performance Monitor Unit (vPMU)
+===========================================
+
+Device types supported:
+ KVM_DEV_TYPE_ARM_PMU_V3 ARM Performance Monitor Unit v3
+
+Instantiate one PMU instance for per VCPU through this API.
+
+Groups:
+ KVM_DEV_ARM_PMU_GRP_IRQ
+ Attributes:
+ A value describing the interrupt number of PMU overflow interrupt.
+
+ Errors:
+ -EINVAL: Value set is out of the expected range
diff --git a/arch/arm64/include/uapi/asm/kvm.h b/arch/arm64/include/uapi/asm/kvm.h
index d268320..9960889 100644
--- a/arch/arm64/include/uapi/asm/kvm.h
+++ b/arch/arm64/include/uapi/asm/kvm.h
@@ -171,6 +171,9 @@ struct kvm_arch_memory_slot {
#define KVM_DEV_ARM_VGIC_GRP_CTRL 4
#define KVM_DEV_ARM_VGIC_CTRL_INIT 0
+/* Device Control API: ARM PMU */
+#define KVM_DEV_ARM_PMU_GRP_IRQ 0
+
/* KVM_IRQ_LINE irq field index values */
#define KVM_ARM_IRQ_TYPE_SHIFT 24
#define KVM_ARM_IRQ_TYPE_MASK 0xff
diff --git a/include/linux/kvm_host.h b/include/linux/kvm_host.h
index 05e99b8..9136dec 100644
--- a/include/linux/kvm_host.h
+++ b/include/linux/kvm_host.h
@@ -1129,6 +1129,7 @@ extern struct kvm_device_ops kvm_mpic_ops;
extern struct kvm_device_ops kvm_xics_ops;
extern struct kvm_device_ops kvm_arm_vgic_v2_ops;
extern struct kvm_device_ops kvm_arm_vgic_v3_ops;
+extern struct kvm_device_ops kvm_arm_pmu_ops;
#ifdef CONFIG_HAVE_KVM_CPU_RELAX_INTERCEPT
diff --git a/include/uapi/linux/kvm.h b/include/uapi/linux/kvm.h
index 716ad4a..591d185 100644
--- a/include/uapi/linux/kvm.h
+++ b/include/uapi/linux/kvm.h
@@ -1018,6 +1018,8 @@ enum kvm_device_type {
#define KVM_DEV_TYPE_FLIC KVM_DEV_TYPE_FLIC
KVM_DEV_TYPE_ARM_VGIC_V3,
#define KVM_DEV_TYPE_ARM_VGIC_V3 KVM_DEV_TYPE_ARM_VGIC_V3
+ KVM_DEV_TYPE_ARM_PMU_V3,
+#define KVM_DEV_TYPE_ARM_PMU_V3 KVM_DEV_TYPE_ARM_PMU_V3
KVM_DEV_TYPE_MAX,
};
diff --git a/virt/kvm/arm/pmu.c b/virt/kvm/arm/pmu.c
index 8eccac9..d9f2e5b 100644
--- a/virt/kvm/arm/pmu.c
+++ b/virt/kvm/arm/pmu.c
@@ -19,6 +19,7 @@
#include <linux/kvm.h>
#include <linux/kvm_host.h>
#include <linux/perf_event.h>
+#include <linux/uaccess.h>
#include <asm/kvm_emulate.h>
#include <kvm/arm_pmu.h>
#include <kvm/arm_vgic.h>
@@ -410,3 +411,90 @@ void kvm_pmu_set_counter_event_type(struct kvm_vcpu *vcpu, unsigned long data,
}
pmc->perf_event = event;
}
+
+static int kvm_arm_pmu_set_irq(struct kvm *kvm, int irq)
+{
+ int j;
+ struct kvm_vcpu *vcpu;
+
+ kvm_for_each_vcpu(j, vcpu, kvm) {
+ struct kvm_pmu *pmu = &vcpu->arch.pmu;
+
+ kvm_debug("Set kvm ARM PMU irq: %d\n", irq);
+ pmu->irq_num = irq;
+ }
+
+ return 0;
+}
+
+static int kvm_arm_pmu_create(struct kvm_device *dev, u32 type)
+{
+ int i, j;
+ struct kvm_vcpu *vcpu;
+ struct kvm *kvm = dev->kvm;
+
+ kvm_for_each_vcpu(j, vcpu, kvm) {
+ struct kvm_pmu *pmu = &vcpu->arch.pmu;
+
+ memset(pmu, 0, sizeof(*pmu));
+ for (i = 0; i < ARMV8_MAX_COUNTERS; i++) {
+ pmu->pmc[i].idx = i;
+ pmu->pmc[i].vcpu = vcpu;
+ }
+ pmu->irq_num = -1;
+ }
+
+ return 0;
+}
+
+static void kvm_arm_pmu_destroy(struct kvm_device *dev)
+{
+ kfree(dev);
+}
+
+static int kvm_arm_pmu_set_attr(struct kvm_device *dev,
+ struct kvm_device_attr *attr)
+{
+ switch (attr->group) {
+ case KVM_DEV_ARM_PMU_GRP_IRQ: {
+ int __user *uaddr = (int __user *)(long)attr->addr;
+ int reg;
+
+ if (get_user(reg, uaddr))
+ return -EFAULT;
+
+ if (reg < VGIC_NR_SGIS || reg > dev->kvm->arch.vgic.nr_irqs)
+ return -EINVAL;
+
+ return kvm_arm_pmu_set_irq(dev->kvm, reg);
+ }
+ }
+
+ return -ENXIO;
+}
+
+static int kvm_arm_pmu_get_attr(struct kvm_device *dev,
+ struct kvm_device_attr *attr)
+{
+ return 0;
+}
+
+static int kvm_arm_pmu_has_attr(struct kvm_device *dev,
+ struct kvm_device_attr *attr)
+{
+ switch (attr->group) {
+ case KVM_DEV_ARM_PMU_GRP_IRQ:
+ return 0;
+ }
+
+ return -ENXIO;
+}
+
+struct kvm_device_ops kvm_arm_pmu_ops = {
+ .name = "kvm-arm-pmu",
+ .create = kvm_arm_pmu_create,
+ .destroy = kvm_arm_pmu_destroy,
+ .set_attr = kvm_arm_pmu_set_attr,
+ .get_attr = kvm_arm_pmu_get_attr,
+ .has_attr = kvm_arm_pmu_has_attr,
+};
diff --git a/virt/kvm/kvm_main.c b/virt/kvm/kvm_main.c
index 8b8a444..31aa245 100644
--- a/virt/kvm/kvm_main.c
+++ b/virt/kvm/kvm_main.c
@@ -2549,6 +2549,10 @@ static struct kvm_device_ops *kvm_device_ops_table[KVM_DEV_TYPE_MAX] = {
#ifdef CONFIG_KVM_XICS
[KVM_DEV_TYPE_XICS] = &kvm_xics_ops,
#endif
+
+#ifdef CONFIG_KVM_ARM_PMU
+ [KVM_DEV_TYPE_ARM_PMU_V3] = &kvm_arm_pmu_ops,
+#endif
};
int kvm_register_device_ops(struct kvm_device_ops *ops, u32 type)
--
2.0.4
^ permalink raw reply related [flat|nested] 113+ messages in thread
* [PATCH v2 22/22] KVM: ARM64: Add a new kvm ARM PMU device
@ 2015-09-11 8:55 ` Shannon Zhao
0 siblings, 0 replies; 113+ messages in thread
From: Shannon Zhao @ 2015-09-11 8:55 UTC (permalink / raw)
To: kvmarm; +Cc: kvm, marc.zyngier, will.deacon, linux-arm-kernel, shannon.zhao
From: Shannon Zhao <shannon.zhao@linaro.org>
Add a new kvm device type KVM_DEV_TYPE_ARM_PMU_V3 for ARM PMU. Implement
the kvm_device_ops for it.
Signed-off-by: Shannon Zhao <shannon.zhao@linaro.org>
---
Documentation/virtual/kvm/devices/arm-pmu.txt | 15 +++++
arch/arm64/include/uapi/asm/kvm.h | 3 +
include/linux/kvm_host.h | 1 +
include/uapi/linux/kvm.h | 2 +
virt/kvm/arm/pmu.c | 88 +++++++++++++++++++++++++++
virt/kvm/kvm_main.c | 4 ++
6 files changed, 113 insertions(+)
create mode 100644 Documentation/virtual/kvm/devices/arm-pmu.txt
diff --git a/Documentation/virtual/kvm/devices/arm-pmu.txt b/Documentation/virtual/kvm/devices/arm-pmu.txt
new file mode 100644
index 0000000..49481c4
--- /dev/null
+++ b/Documentation/virtual/kvm/devices/arm-pmu.txt
@@ -0,0 +1,15 @@
+ARM Virtual Performance Monitor Unit (vPMU)
+===========================================
+
+Device types supported:
+ KVM_DEV_TYPE_ARM_PMU_V3 ARM Performance Monitor Unit v3
+
+Instantiate one PMU instance for per VCPU through this API.
+
+Groups:
+ KVM_DEV_ARM_PMU_GRP_IRQ
+ Attributes:
+ A value describing the interrupt number of PMU overflow interrupt.
+
+ Errors:
+ -EINVAL: Value set is out of the expected range
diff --git a/arch/arm64/include/uapi/asm/kvm.h b/arch/arm64/include/uapi/asm/kvm.h
index d268320..9960889 100644
--- a/arch/arm64/include/uapi/asm/kvm.h
+++ b/arch/arm64/include/uapi/asm/kvm.h
@@ -171,6 +171,9 @@ struct kvm_arch_memory_slot {
#define KVM_DEV_ARM_VGIC_GRP_CTRL 4
#define KVM_DEV_ARM_VGIC_CTRL_INIT 0
+/* Device Control API: ARM PMU */
+#define KVM_DEV_ARM_PMU_GRP_IRQ 0
+
/* KVM_IRQ_LINE irq field index values */
#define KVM_ARM_IRQ_TYPE_SHIFT 24
#define KVM_ARM_IRQ_TYPE_MASK 0xff
diff --git a/include/linux/kvm_host.h b/include/linux/kvm_host.h
index 05e99b8..9136dec 100644
--- a/include/linux/kvm_host.h
+++ b/include/linux/kvm_host.h
@@ -1129,6 +1129,7 @@ extern struct kvm_device_ops kvm_mpic_ops;
extern struct kvm_device_ops kvm_xics_ops;
extern struct kvm_device_ops kvm_arm_vgic_v2_ops;
extern struct kvm_device_ops kvm_arm_vgic_v3_ops;
+extern struct kvm_device_ops kvm_arm_pmu_ops;
#ifdef CONFIG_HAVE_KVM_CPU_RELAX_INTERCEPT
diff --git a/include/uapi/linux/kvm.h b/include/uapi/linux/kvm.h
index 716ad4a..591d185 100644
--- a/include/uapi/linux/kvm.h
+++ b/include/uapi/linux/kvm.h
@@ -1018,6 +1018,8 @@ enum kvm_device_type {
#define KVM_DEV_TYPE_FLIC KVM_DEV_TYPE_FLIC
KVM_DEV_TYPE_ARM_VGIC_V3,
#define KVM_DEV_TYPE_ARM_VGIC_V3 KVM_DEV_TYPE_ARM_VGIC_V3
+ KVM_DEV_TYPE_ARM_PMU_V3,
+#define KVM_DEV_TYPE_ARM_PMU_V3 KVM_DEV_TYPE_ARM_PMU_V3
KVM_DEV_TYPE_MAX,
};
diff --git a/virt/kvm/arm/pmu.c b/virt/kvm/arm/pmu.c
index 8eccac9..d9f2e5b 100644
--- a/virt/kvm/arm/pmu.c
+++ b/virt/kvm/arm/pmu.c
@@ -19,6 +19,7 @@
#include <linux/kvm.h>
#include <linux/kvm_host.h>
#include <linux/perf_event.h>
+#include <linux/uaccess.h>
#include <asm/kvm_emulate.h>
#include <kvm/arm_pmu.h>
#include <kvm/arm_vgic.h>
@@ -410,3 +411,90 @@ void kvm_pmu_set_counter_event_type(struct kvm_vcpu *vcpu, unsigned long data,
}
pmc->perf_event = event;
}
+
+static int kvm_arm_pmu_set_irq(struct kvm *kvm, int irq)
+{
+ int j;
+ struct kvm_vcpu *vcpu;
+
+ kvm_for_each_vcpu(j, vcpu, kvm) {
+ struct kvm_pmu *pmu = &vcpu->arch.pmu;
+
+ kvm_debug("Set kvm ARM PMU irq: %d\n", irq);
+ pmu->irq_num = irq;
+ }
+
+ return 0;
+}
+
+static int kvm_arm_pmu_create(struct kvm_device *dev, u32 type)
+{
+ int i, j;
+ struct kvm_vcpu *vcpu;
+ struct kvm *kvm = dev->kvm;
+
+ kvm_for_each_vcpu(j, vcpu, kvm) {
+ struct kvm_pmu *pmu = &vcpu->arch.pmu;
+
+ memset(pmu, 0, sizeof(*pmu));
+ for (i = 0; i < ARMV8_MAX_COUNTERS; i++) {
+ pmu->pmc[i].idx = i;
+ pmu->pmc[i].vcpu = vcpu;
+ }
+ pmu->irq_num = -1;
+ }
+
+ return 0;
+}
+
+static void kvm_arm_pmu_destroy(struct kvm_device *dev)
+{
+ kfree(dev);
+}
+
+static int kvm_arm_pmu_set_attr(struct kvm_device *dev,
+ struct kvm_device_attr *attr)
+{
+ switch (attr->group) {
+ case KVM_DEV_ARM_PMU_GRP_IRQ: {
+ int __user *uaddr = (int __user *)(long)attr->addr;
+ int reg;
+
+ if (get_user(reg, uaddr))
+ return -EFAULT;
+
+ if (reg < VGIC_NR_SGIS || reg > dev->kvm->arch.vgic.nr_irqs)
+ return -EINVAL;
+
+ return kvm_arm_pmu_set_irq(dev->kvm, reg);
+ }
+ }
+
+ return -ENXIO;
+}
+
+static int kvm_arm_pmu_get_attr(struct kvm_device *dev,
+ struct kvm_device_attr *attr)
+{
+ return 0;
+}
+
+static int kvm_arm_pmu_has_attr(struct kvm_device *dev,
+ struct kvm_device_attr *attr)
+{
+ switch (attr->group) {
+ case KVM_DEV_ARM_PMU_GRP_IRQ:
+ return 0;
+ }
+
+ return -ENXIO;
+}
+
+struct kvm_device_ops kvm_arm_pmu_ops = {
+ .name = "kvm-arm-pmu",
+ .create = kvm_arm_pmu_create,
+ .destroy = kvm_arm_pmu_destroy,
+ .set_attr = kvm_arm_pmu_set_attr,
+ .get_attr = kvm_arm_pmu_get_attr,
+ .has_attr = kvm_arm_pmu_has_attr,
+};
diff --git a/virt/kvm/kvm_main.c b/virt/kvm/kvm_main.c
index 8b8a444..31aa245 100644
--- a/virt/kvm/kvm_main.c
+++ b/virt/kvm/kvm_main.c
@@ -2549,6 +2549,10 @@ static struct kvm_device_ops *kvm_device_ops_table[KVM_DEV_TYPE_MAX] = {
#ifdef CONFIG_KVM_XICS
[KVM_DEV_TYPE_XICS] = &kvm_xics_ops,
#endif
+
+#ifdef CONFIG_KVM_ARM_PMU
+ [KVM_DEV_TYPE_ARM_PMU_V3] = &kvm_arm_pmu_ops,
+#endif
};
int kvm_register_device_ops(struct kvm_device_ops *ops, u32 type)
--
2.0.4
^ permalink raw reply related [flat|nested] 113+ messages in thread
* [PATCH v2 22/22] KVM: ARM64: Add a new kvm ARM PMU device
@ 2015-09-11 8:55 ` Shannon Zhao
0 siblings, 0 replies; 113+ messages in thread
From: Shannon Zhao @ 2015-09-11 8:55 UTC (permalink / raw)
To: kvmarm; +Cc: kvm, marc.zyngier, will.deacon, linux-arm-kernel, shannon.zhao
From: Shannon Zhao <shannon.zhao@linaro.org>
Add a new kvm device type KVM_DEV_TYPE_ARM_PMU_V3 for ARM PMU. Implement
the kvm_device_ops for it.
Signed-off-by: Shannon Zhao <shannon.zhao@linaro.org>
---
Documentation/virtual/kvm/devices/arm-pmu.txt | 15 +++++
arch/arm64/include/uapi/asm/kvm.h | 3 +
include/linux/kvm_host.h | 1 +
include/uapi/linux/kvm.h | 2 +
virt/kvm/arm/pmu.c | 88 +++++++++++++++++++++++++++
virt/kvm/kvm_main.c | 4 ++
6 files changed, 113 insertions(+)
create mode 100644 Documentation/virtual/kvm/devices/arm-pmu.txt
diff --git a/Documentation/virtual/kvm/devices/arm-pmu.txt b/Documentation/virtual/kvm/devices/arm-pmu.txt
new file mode 100644
index 0000000..49481c4
--- /dev/null
+++ b/Documentation/virtual/kvm/devices/arm-pmu.txt
@@ -0,0 +1,15 @@
+ARM Virtual Performance Monitor Unit (vPMU)
+===========================================
+
+Device types supported:
+ KVM_DEV_TYPE_ARM_PMU_V3 ARM Performance Monitor Unit v3
+
+Instantiate one PMU instance for per VCPU through this API.
+
+Groups:
+ KVM_DEV_ARM_PMU_GRP_IRQ
+ Attributes:
+ A value describing the interrupt number of PMU overflow interrupt.
+
+ Errors:
+ -EINVAL: Value set is out of the expected range
diff --git a/arch/arm64/include/uapi/asm/kvm.h b/arch/arm64/include/uapi/asm/kvm.h
index d268320..9960889 100644
--- a/arch/arm64/include/uapi/asm/kvm.h
+++ b/arch/arm64/include/uapi/asm/kvm.h
@@ -171,6 +171,9 @@ struct kvm_arch_memory_slot {
#define KVM_DEV_ARM_VGIC_GRP_CTRL 4
#define KVM_DEV_ARM_VGIC_CTRL_INIT 0
+/* Device Control API: ARM PMU */
+#define KVM_DEV_ARM_PMU_GRP_IRQ 0
+
/* KVM_IRQ_LINE irq field index values */
#define KVM_ARM_IRQ_TYPE_SHIFT 24
#define KVM_ARM_IRQ_TYPE_MASK 0xff
diff --git a/include/linux/kvm_host.h b/include/linux/kvm_host.h
index 05e99b8..9136dec 100644
--- a/include/linux/kvm_host.h
+++ b/include/linux/kvm_host.h
@@ -1129,6 +1129,7 @@ extern struct kvm_device_ops kvm_mpic_ops;
extern struct kvm_device_ops kvm_xics_ops;
extern struct kvm_device_ops kvm_arm_vgic_v2_ops;
extern struct kvm_device_ops kvm_arm_vgic_v3_ops;
+extern struct kvm_device_ops kvm_arm_pmu_ops;
#ifdef CONFIG_HAVE_KVM_CPU_RELAX_INTERCEPT
diff --git a/include/uapi/linux/kvm.h b/include/uapi/linux/kvm.h
index 716ad4a..591d185 100644
--- a/include/uapi/linux/kvm.h
+++ b/include/uapi/linux/kvm.h
@@ -1018,6 +1018,8 @@ enum kvm_device_type {
#define KVM_DEV_TYPE_FLIC KVM_DEV_TYPE_FLIC
KVM_DEV_TYPE_ARM_VGIC_V3,
#define KVM_DEV_TYPE_ARM_VGIC_V3 KVM_DEV_TYPE_ARM_VGIC_V3
+ KVM_DEV_TYPE_ARM_PMU_V3,
+#define KVM_DEV_TYPE_ARM_PMU_V3 KVM_DEV_TYPE_ARM_PMU_V3
KVM_DEV_TYPE_MAX,
};
diff --git a/virt/kvm/arm/pmu.c b/virt/kvm/arm/pmu.c
index 8eccac9..d9f2e5b 100644
--- a/virt/kvm/arm/pmu.c
+++ b/virt/kvm/arm/pmu.c
@@ -19,6 +19,7 @@
#include <linux/kvm.h>
#include <linux/kvm_host.h>
#include <linux/perf_event.h>
+#include <linux/uaccess.h>
#include <asm/kvm_emulate.h>
#include <kvm/arm_pmu.h>
#include <kvm/arm_vgic.h>
@@ -410,3 +411,90 @@ void kvm_pmu_set_counter_event_type(struct kvm_vcpu *vcpu, unsigned long data,
}
pmc->perf_event = event;
}
+
+static int kvm_arm_pmu_set_irq(struct kvm *kvm, int irq)
+{
+ int j;
+ struct kvm_vcpu *vcpu;
+
+ kvm_for_each_vcpu(j, vcpu, kvm) {
+ struct kvm_pmu *pmu = &vcpu->arch.pmu;
+
+ kvm_debug("Set kvm ARM PMU irq: %d\n", irq);
+ pmu->irq_num = irq;
+ }
+
+ return 0;
+}
+
+static int kvm_arm_pmu_create(struct kvm_device *dev, u32 type)
+{
+ int i, j;
+ struct kvm_vcpu *vcpu;
+ struct kvm *kvm = dev->kvm;
+
+ kvm_for_each_vcpu(j, vcpu, kvm) {
+ struct kvm_pmu *pmu = &vcpu->arch.pmu;
+
+ memset(pmu, 0, sizeof(*pmu));
+ for (i = 0; i < ARMV8_MAX_COUNTERS; i++) {
+ pmu->pmc[i].idx = i;
+ pmu->pmc[i].vcpu = vcpu;
+ }
+ pmu->irq_num = -1;
+ }
+
+ return 0;
+}
+
+static void kvm_arm_pmu_destroy(struct kvm_device *dev)
+{
+ kfree(dev);
+}
+
+static int kvm_arm_pmu_set_attr(struct kvm_device *dev,
+ struct kvm_device_attr *attr)
+{
+ switch (attr->group) {
+ case KVM_DEV_ARM_PMU_GRP_IRQ: {
+ int __user *uaddr = (int __user *)(long)attr->addr;
+ int reg;
+
+ if (get_user(reg, uaddr))
+ return -EFAULT;
+
+ if (reg < VGIC_NR_SGIS || reg > dev->kvm->arch.vgic.nr_irqs)
+ return -EINVAL;
+
+ return kvm_arm_pmu_set_irq(dev->kvm, reg);
+ }
+ }
+
+ return -ENXIO;
+}
+
+static int kvm_arm_pmu_get_attr(struct kvm_device *dev,
+ struct kvm_device_attr *attr)
+{
+ return 0;
+}
+
+static int kvm_arm_pmu_has_attr(struct kvm_device *dev,
+ struct kvm_device_attr *attr)
+{
+ switch (attr->group) {
+ case KVM_DEV_ARM_PMU_GRP_IRQ:
+ return 0;
+ }
+
+ return -ENXIO;
+}
+
+struct kvm_device_ops kvm_arm_pmu_ops = {
+ .name = "kvm-arm-pmu",
+ .create = kvm_arm_pmu_create,
+ .destroy = kvm_arm_pmu_destroy,
+ .set_attr = kvm_arm_pmu_set_attr,
+ .get_attr = kvm_arm_pmu_get_attr,
+ .has_attr = kvm_arm_pmu_has_attr,
+};
diff --git a/virt/kvm/kvm_main.c b/virt/kvm/kvm_main.c
index 8b8a444..31aa245 100644
--- a/virt/kvm/kvm_main.c
+++ b/virt/kvm/kvm_main.c
@@ -2549,6 +2549,10 @@ static struct kvm_device_ops *kvm_device_ops_table[KVM_DEV_TYPE_MAX] = {
#ifdef CONFIG_KVM_XICS
[KVM_DEV_TYPE_XICS] = &kvm_xics_ops,
#endif
+
+#ifdef CONFIG_KVM_ARM_PMU
+ [KVM_DEV_TYPE_ARM_PMU_V3] = &kvm_arm_pmu_ops,
+#endif
};
int kvm_register_device_ops(struct kvm_device_ops *ops, u32 type)
--
2.0.4
^ permalink raw reply related [flat|nested] 113+ messages in thread
* [PATCH v2 02/22] KVM: ARM64: Define PMU data structure for each vcpu
2015-09-11 8:54 ` Shannon Zhao
@ 2015-09-11 9:10 ` Marc Zyngier
-1 siblings, 0 replies; 113+ messages in thread
From: Marc Zyngier @ 2015-09-11 9:10 UTC (permalink / raw)
To: linux-arm-kernel
On 11/09/15 09:54, Shannon Zhao wrote:
> From: Shannon Zhao <shannon.zhao@linaro.org>
>
> Here we plan to support virtual PMU for guest by full software
> emulation, so define some basic structs and functions preparing for
> futher steps. Define struct kvm_pmc for performance monitor counter and
> struct kvm_pmu for performance monitor unit for each vcpu. According to
> ARMv8 spec, the PMU contains at most 32(ARMV8_MAX_COUNTERS) counters.
>
> Since this only supports ARM64 (or PMUv3), add a separate config symbol
> for it.
>
> Signed-off-by: Shannon Zhao <shannon.zhao@linaro.org>
> ---
> arch/arm64/include/asm/kvm_host.h | 2 ++
> arch/arm64/include/asm/pmu.h | 2 ++
> arch/arm64/kvm/Kconfig | 8 ++++++++
> include/kvm/arm_pmu.h | 39 +++++++++++++++++++++++++++++++++++++++
> 4 files changed, 51 insertions(+)
> create mode 100644 include/kvm/arm_pmu.h
>
> diff --git a/arch/arm64/include/asm/kvm_host.h b/arch/arm64/include/asm/kvm_host.h
> index 2709db2..3c88873 100644
> --- a/arch/arm64/include/asm/kvm_host.h
> +++ b/arch/arm64/include/asm/kvm_host.h
> @@ -42,6 +42,7 @@
>
> #include <kvm/arm_vgic.h>
> #include <kvm/arm_arch_timer.h>
> +#include <kvm/arm_pmu.h>
>
> #define KVM_VCPU_MAX_FEATURES 3
>
> @@ -116,6 +117,7 @@ struct kvm_vcpu_arch {
> /* VGIC state */
> struct vgic_cpu vgic_cpu;
> struct arch_timer_cpu timer_cpu;
> + struct kvm_pmu pmu;
>
> /*
> * Anything that is not used directly from assembly code goes
> diff --git a/arch/arm64/include/asm/pmu.h b/arch/arm64/include/asm/pmu.h
> index b9f394a..95681e6 100644
> --- a/arch/arm64/include/asm/pmu.h
> +++ b/arch/arm64/include/asm/pmu.h
> @@ -19,6 +19,8 @@
> #ifndef __ASM_PMU_H
> #define __ASM_PMU_H
>
> +#include <linux/perf_event.h>
> +
This include file seems out of place here, as I don't think anything in
this file is directly PMU related.
> #define ARMV8_MAX_COUNTERS 32
> #define ARMV8_COUNTER_MASK (ARMV8_MAX_COUNTERS - 1)
>
> diff --git a/arch/arm64/kvm/Kconfig b/arch/arm64/kvm/Kconfig
> index bfffe8f..3c7c58d 100644
> --- a/arch/arm64/kvm/Kconfig
> +++ b/arch/arm64/kvm/Kconfig
> @@ -31,6 +31,7 @@ config KVM
> select KVM_VFIO
> select HAVE_KVM_EVENTFD
> select HAVE_KVM_IRQFD
> + select KVM_ARM_PMU
> ---help---
> Support hosting virtualized guest machines.
>
> @@ -52,4 +53,11 @@ config KVM_ARM_MAX_VCPUS
> large, so only choose a reasonable number that you expect to
> actually use.
>
> +config KVM_ARM_PMU
> + bool
> + depends on KVM_ARM_HOST
> + ---help---
> + Adds support for a virtual Performance Monitoring Unit (PMU) in
> + virtual machines.
> +
> endif # VIRTUALIZATION
> diff --git a/include/kvm/arm_pmu.h b/include/kvm/arm_pmu.h
> new file mode 100644
> index 0000000..64af88a
> --- /dev/null
> +++ b/include/kvm/arm_pmu.h
> @@ -0,0 +1,39 @@
> +/*
> + * Copyright (C) 2015 Linaro Ltd.
> + * Author: Shannon Zhao <shannon.zhao@linaro.org>
> + *
> + * This program is free software; you can redistribute it and/or modify
> + * it under the terms of the GNU General Public License version 2 as
> + * published by the Free Software Foundation.
> + *
> + * This program is distributed in the hope that it will be useful,
> + * but WITHOUT ANY WARRANTY; without even the implied warranty of
> + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
> + * GNU General Public License for more details.
> + *
> + * You should have received a copy of the GNU General Public License
> + * along with this program. If not, see <http://www.gnu.org/licenses/>.
> + */
> +
> +#ifndef __ASM_ARM_KVM_PMU_H
> +#define __ASM_ARM_KVM_PMU_H
> +
> +#include <asm/pmu.h>
> +
> +struct kvm_pmc {
> + u8 idx;/* index into the pmu->pmc array */
> + struct perf_event *perf_event;
> + struct kvm_vcpu *vcpu;
> +};
> +
Since you have a struct perf_event here, it makes more sense to #include
linux/perf.h here.
> +struct kvm_pmu {
> +#ifdef CONFIG_KVM_ARM_PMU
> + /* PMU IRQ Number per VCPU */
> + int irq_num;
> + /* IRQ pending flag */
> + bool irq_pending;
> + struct kvm_pmc pmc[ARMV8_MAX_COUNTERS];
> +#endif
> +};
> +
> +#endif
>
Thanks,
M.
--
Jazz is not dead. It just smells funny...
^ permalink raw reply [flat|nested] 113+ messages in thread
* Re: [PATCH v2 02/22] KVM: ARM64: Define PMU data structure for each vcpu
@ 2015-09-11 9:10 ` Marc Zyngier
0 siblings, 0 replies; 113+ messages in thread
From: Marc Zyngier @ 2015-09-11 9:10 UTC (permalink / raw)
To: Shannon Zhao, kvmarm
Cc: linux-arm-kernel, kvm, christoffer.dall, will.deacon, alex.bennee,
wei, shannon.zhao, peter.huangpeng
On 11/09/15 09:54, Shannon Zhao wrote:
> From: Shannon Zhao <shannon.zhao@linaro.org>
>
> Here we plan to support virtual PMU for guest by full software
> emulation, so define some basic structs and functions preparing for
> futher steps. Define struct kvm_pmc for performance monitor counter and
> struct kvm_pmu for performance monitor unit for each vcpu. According to
> ARMv8 spec, the PMU contains at most 32(ARMV8_MAX_COUNTERS) counters.
>
> Since this only supports ARM64 (or PMUv3), add a separate config symbol
> for it.
>
> Signed-off-by: Shannon Zhao <shannon.zhao@linaro.org>
> ---
> arch/arm64/include/asm/kvm_host.h | 2 ++
> arch/arm64/include/asm/pmu.h | 2 ++
> arch/arm64/kvm/Kconfig | 8 ++++++++
> include/kvm/arm_pmu.h | 39 +++++++++++++++++++++++++++++++++++++++
> 4 files changed, 51 insertions(+)
> create mode 100644 include/kvm/arm_pmu.h
>
> diff --git a/arch/arm64/include/asm/kvm_host.h b/arch/arm64/include/asm/kvm_host.h
> index 2709db2..3c88873 100644
> --- a/arch/arm64/include/asm/kvm_host.h
> +++ b/arch/arm64/include/asm/kvm_host.h
> @@ -42,6 +42,7 @@
>
> #include <kvm/arm_vgic.h>
> #include <kvm/arm_arch_timer.h>
> +#include <kvm/arm_pmu.h>
>
> #define KVM_VCPU_MAX_FEATURES 3
>
> @@ -116,6 +117,7 @@ struct kvm_vcpu_arch {
> /* VGIC state */
> struct vgic_cpu vgic_cpu;
> struct arch_timer_cpu timer_cpu;
> + struct kvm_pmu pmu;
>
> /*
> * Anything that is not used directly from assembly code goes
> diff --git a/arch/arm64/include/asm/pmu.h b/arch/arm64/include/asm/pmu.h
> index b9f394a..95681e6 100644
> --- a/arch/arm64/include/asm/pmu.h
> +++ b/arch/arm64/include/asm/pmu.h
> @@ -19,6 +19,8 @@
> #ifndef __ASM_PMU_H
> #define __ASM_PMU_H
>
> +#include <linux/perf_event.h>
> +
This include file seems out of place here, as I don't think anything in
this file is directly PMU related.
> #define ARMV8_MAX_COUNTERS 32
> #define ARMV8_COUNTER_MASK (ARMV8_MAX_COUNTERS - 1)
>
> diff --git a/arch/arm64/kvm/Kconfig b/arch/arm64/kvm/Kconfig
> index bfffe8f..3c7c58d 100644
> --- a/arch/arm64/kvm/Kconfig
> +++ b/arch/arm64/kvm/Kconfig
> @@ -31,6 +31,7 @@ config KVM
> select KVM_VFIO
> select HAVE_KVM_EVENTFD
> select HAVE_KVM_IRQFD
> + select KVM_ARM_PMU
> ---help---
> Support hosting virtualized guest machines.
>
> @@ -52,4 +53,11 @@ config KVM_ARM_MAX_VCPUS
> large, so only choose a reasonable number that you expect to
> actually use.
>
> +config KVM_ARM_PMU
> + bool
> + depends on KVM_ARM_HOST
> + ---help---
> + Adds support for a virtual Performance Monitoring Unit (PMU) in
> + virtual machines.
> +
> endif # VIRTUALIZATION
> diff --git a/include/kvm/arm_pmu.h b/include/kvm/arm_pmu.h
> new file mode 100644
> index 0000000..64af88a
> --- /dev/null
> +++ b/include/kvm/arm_pmu.h
> @@ -0,0 +1,39 @@
> +/*
> + * Copyright (C) 2015 Linaro Ltd.
> + * Author: Shannon Zhao <shannon.zhao@linaro.org>
> + *
> + * This program is free software; you can redistribute it and/or modify
> + * it under the terms of the GNU General Public License version 2 as
> + * published by the Free Software Foundation.
> + *
> + * This program is distributed in the hope that it will be useful,
> + * but WITHOUT ANY WARRANTY; without even the implied warranty of
> + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
> + * GNU General Public License for more details.
> + *
> + * You should have received a copy of the GNU General Public License
> + * along with this program. If not, see <http://www.gnu.org/licenses/>.
> + */
> +
> +#ifndef __ASM_ARM_KVM_PMU_H
> +#define __ASM_ARM_KVM_PMU_H
> +
> +#include <asm/pmu.h>
> +
> +struct kvm_pmc {
> + u8 idx;/* index into the pmu->pmc array */
> + struct perf_event *perf_event;
> + struct kvm_vcpu *vcpu;
> +};
> +
Since you have a struct perf_event here, it makes more sense to #include
linux/perf.h here.
> +struct kvm_pmu {
> +#ifdef CONFIG_KVM_ARM_PMU
> + /* PMU IRQ Number per VCPU */
> + int irq_num;
> + /* IRQ pending flag */
> + bool irq_pending;
> + struct kvm_pmc pmc[ARMV8_MAX_COUNTERS];
> +#endif
> +};
> +
> +#endif
>
Thanks,
M.
--
Jazz is not dead. It just smells funny...
^ permalink raw reply [flat|nested] 113+ messages in thread
* [PATCH v2 02/22] KVM: ARM64: Define PMU data structure for each vcpu
2015-09-11 9:10 ` Marc Zyngier
(?)
@ 2015-09-11 9:58 ` Shannon Zhao
-1 siblings, 0 replies; 113+ messages in thread
From: Shannon Zhao @ 2015-09-11 9:58 UTC (permalink / raw)
To: linux-arm-kernel
On 2015/9/11 17:10, Marc Zyngier wrote:
> On 11/09/15 09:54, Shannon Zhao wrote:
>> From: Shannon Zhao <shannon.zhao@linaro.org>
>>
>> Here we plan to support virtual PMU for guest by full software
>> emulation, so define some basic structs and functions preparing for
>> futher steps. Define struct kvm_pmc for performance monitor counter and
>> struct kvm_pmu for performance monitor unit for each vcpu. According to
>> ARMv8 spec, the PMU contains at most 32(ARMV8_MAX_COUNTERS) counters.
>>
>> Since this only supports ARM64 (or PMUv3), add a separate config symbol
>> for it.
>>
>> Signed-off-by: Shannon Zhao <shannon.zhao@linaro.org>
>> ---
>> arch/arm64/include/asm/kvm_host.h | 2 ++
>> arch/arm64/include/asm/pmu.h | 2 ++
>> arch/arm64/kvm/Kconfig | 8 ++++++++
>> include/kvm/arm_pmu.h | 39 +++++++++++++++++++++++++++++++++++++++
>> 4 files changed, 51 insertions(+)
>> create mode 100644 include/kvm/arm_pmu.h
>>
>> diff --git a/arch/arm64/include/asm/kvm_host.h b/arch/arm64/include/asm/kvm_host.h
>> index 2709db2..3c88873 100644
>> --- a/arch/arm64/include/asm/kvm_host.h
>> +++ b/arch/arm64/include/asm/kvm_host.h
>> @@ -42,6 +42,7 @@
>>
>> #include <kvm/arm_vgic.h>
>> #include <kvm/arm_arch_timer.h>
>> +#include <kvm/arm_pmu.h>
>>
>> #define KVM_VCPU_MAX_FEATURES 3
>>
>> @@ -116,6 +117,7 @@ struct kvm_vcpu_arch {
>> /* VGIC state */
>> struct vgic_cpu vgic_cpu;
>> struct arch_timer_cpu timer_cpu;
>> + struct kvm_pmu pmu;
>>
>> /*
>> * Anything that is not used directly from assembly code goes
>> diff --git a/arch/arm64/include/asm/pmu.h b/arch/arm64/include/asm/pmu.h
>> index b9f394a..95681e6 100644
>> --- a/arch/arm64/include/asm/pmu.h
>> +++ b/arch/arm64/include/asm/pmu.h
>> @@ -19,6 +19,8 @@
>> #ifndef __ASM_PMU_H
>> #define __ASM_PMU_H
>>
>> +#include <linux/perf_event.h>
>> +
>
> This include file seems out of place here, as I don't think anything in
> this file is directly PMU related.
>
>> #define ARMV8_MAX_COUNTERS 32
>> #define ARMV8_COUNTER_MASK (ARMV8_MAX_COUNTERS - 1)
>>
>> diff --git a/arch/arm64/kvm/Kconfig b/arch/arm64/kvm/Kconfig
>> index bfffe8f..3c7c58d 100644
>> --- a/arch/arm64/kvm/Kconfig
>> +++ b/arch/arm64/kvm/Kconfig
>> @@ -31,6 +31,7 @@ config KVM
>> select KVM_VFIO
>> select HAVE_KVM_EVENTFD
>> select HAVE_KVM_IRQFD
>> + select KVM_ARM_PMU
>> ---help---
>> Support hosting virtualized guest machines.
>>
>> @@ -52,4 +53,11 @@ config KVM_ARM_MAX_VCPUS
>> large, so only choose a reasonable number that you expect to
>> actually use.
>>
>> +config KVM_ARM_PMU
>> + bool
>> + depends on KVM_ARM_HOST
>> + ---help---
>> + Adds support for a virtual Performance Monitoring Unit (PMU) in
>> + virtual machines.
>> +
>> endif # VIRTUALIZATION
>> diff --git a/include/kvm/arm_pmu.h b/include/kvm/arm_pmu.h
>> new file mode 100644
>> index 0000000..64af88a
>> --- /dev/null
>> +++ b/include/kvm/arm_pmu.h
>> @@ -0,0 +1,39 @@
>> +/*
>> + * Copyright (C) 2015 Linaro Ltd.
>> + * Author: Shannon Zhao <shannon.zhao@linaro.org>
>> + *
>> + * This program is free software; you can redistribute it and/or modify
>> + * it under the terms of the GNU General Public License version 2 as
>> + * published by the Free Software Foundation.
>> + *
>> + * This program is distributed in the hope that it will be useful,
>> + * but WITHOUT ANY WARRANTY; without even the implied warranty of
>> + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
>> + * GNU General Public License for more details.
>> + *
>> + * You should have received a copy of the GNU General Public License
>> + * along with this program. If not, see <http://www.gnu.org/licenses/>.
>> + */
>> +
>> +#ifndef __ASM_ARM_KVM_PMU_H
>> +#define __ASM_ARM_KVM_PMU_H
>> +
>> +#include <asm/pmu.h>
>> +
>> +struct kvm_pmc {
>> + u8 idx;/* index into the pmu->pmc array */
>> + struct perf_event *perf_event;
>> + struct kvm_vcpu *vcpu;
>> +};
>> +
>
> Since you have a struct perf_event here, it makes more sense to #include
> linux/perf.h here.
>
Ok, will fix this.
>> +struct kvm_pmu {
>> +#ifdef CONFIG_KVM_ARM_PMU
>> + /* PMU IRQ Number per VCPU */
>> + int irq_num;
>> + /* IRQ pending flag */
>> + bool irq_pending;
>> + struct kvm_pmc pmc[ARMV8_MAX_COUNTERS];
>> +#endif
>> +};
>> +
>> +#endif
>>
>
> Thanks,
>
> M.
>
--
Shannon
^ permalink raw reply [flat|nested] 113+ messages in thread
* Re: [PATCH v2 02/22] KVM: ARM64: Define PMU data structure for each vcpu
@ 2015-09-11 9:58 ` Shannon Zhao
0 siblings, 0 replies; 113+ messages in thread
From: Shannon Zhao @ 2015-09-11 9:58 UTC (permalink / raw)
To: Marc Zyngier, kvmarm; +Cc: kvm, will.deacon, linux-arm-kernel, shannon.zhao
On 2015/9/11 17:10, Marc Zyngier wrote:
> On 11/09/15 09:54, Shannon Zhao wrote:
>> From: Shannon Zhao <shannon.zhao@linaro.org>
>>
>> Here we plan to support virtual PMU for guest by full software
>> emulation, so define some basic structs and functions preparing for
>> futher steps. Define struct kvm_pmc for performance monitor counter and
>> struct kvm_pmu for performance monitor unit for each vcpu. According to
>> ARMv8 spec, the PMU contains at most 32(ARMV8_MAX_COUNTERS) counters.
>>
>> Since this only supports ARM64 (or PMUv3), add a separate config symbol
>> for it.
>>
>> Signed-off-by: Shannon Zhao <shannon.zhao@linaro.org>
>> ---
>> arch/arm64/include/asm/kvm_host.h | 2 ++
>> arch/arm64/include/asm/pmu.h | 2 ++
>> arch/arm64/kvm/Kconfig | 8 ++++++++
>> include/kvm/arm_pmu.h | 39 +++++++++++++++++++++++++++++++++++++++
>> 4 files changed, 51 insertions(+)
>> create mode 100644 include/kvm/arm_pmu.h
>>
>> diff --git a/arch/arm64/include/asm/kvm_host.h b/arch/arm64/include/asm/kvm_host.h
>> index 2709db2..3c88873 100644
>> --- a/arch/arm64/include/asm/kvm_host.h
>> +++ b/arch/arm64/include/asm/kvm_host.h
>> @@ -42,6 +42,7 @@
>>
>> #include <kvm/arm_vgic.h>
>> #include <kvm/arm_arch_timer.h>
>> +#include <kvm/arm_pmu.h>
>>
>> #define KVM_VCPU_MAX_FEATURES 3
>>
>> @@ -116,6 +117,7 @@ struct kvm_vcpu_arch {
>> /* VGIC state */
>> struct vgic_cpu vgic_cpu;
>> struct arch_timer_cpu timer_cpu;
>> + struct kvm_pmu pmu;
>>
>> /*
>> * Anything that is not used directly from assembly code goes
>> diff --git a/arch/arm64/include/asm/pmu.h b/arch/arm64/include/asm/pmu.h
>> index b9f394a..95681e6 100644
>> --- a/arch/arm64/include/asm/pmu.h
>> +++ b/arch/arm64/include/asm/pmu.h
>> @@ -19,6 +19,8 @@
>> #ifndef __ASM_PMU_H
>> #define __ASM_PMU_H
>>
>> +#include <linux/perf_event.h>
>> +
>
> This include file seems out of place here, as I don't think anything in
> this file is directly PMU related.
>
>> #define ARMV8_MAX_COUNTERS 32
>> #define ARMV8_COUNTER_MASK (ARMV8_MAX_COUNTERS - 1)
>>
>> diff --git a/arch/arm64/kvm/Kconfig b/arch/arm64/kvm/Kconfig
>> index bfffe8f..3c7c58d 100644
>> --- a/arch/arm64/kvm/Kconfig
>> +++ b/arch/arm64/kvm/Kconfig
>> @@ -31,6 +31,7 @@ config KVM
>> select KVM_VFIO
>> select HAVE_KVM_EVENTFD
>> select HAVE_KVM_IRQFD
>> + select KVM_ARM_PMU
>> ---help---
>> Support hosting virtualized guest machines.
>>
>> @@ -52,4 +53,11 @@ config KVM_ARM_MAX_VCPUS
>> large, so only choose a reasonable number that you expect to
>> actually use.
>>
>> +config KVM_ARM_PMU
>> + bool
>> + depends on KVM_ARM_HOST
>> + ---help---
>> + Adds support for a virtual Performance Monitoring Unit (PMU) in
>> + virtual machines.
>> +
>> endif # VIRTUALIZATION
>> diff --git a/include/kvm/arm_pmu.h b/include/kvm/arm_pmu.h
>> new file mode 100644
>> index 0000000..64af88a
>> --- /dev/null
>> +++ b/include/kvm/arm_pmu.h
>> @@ -0,0 +1,39 @@
>> +/*
>> + * Copyright (C) 2015 Linaro Ltd.
>> + * Author: Shannon Zhao <shannon.zhao@linaro.org>
>> + *
>> + * This program is free software; you can redistribute it and/or modify
>> + * it under the terms of the GNU General Public License version 2 as
>> + * published by the Free Software Foundation.
>> + *
>> + * This program is distributed in the hope that it will be useful,
>> + * but WITHOUT ANY WARRANTY; without even the implied warranty of
>> + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
>> + * GNU General Public License for more details.
>> + *
>> + * You should have received a copy of the GNU General Public License
>> + * along with this program. If not, see <http://www.gnu.org/licenses/>.
>> + */
>> +
>> +#ifndef __ASM_ARM_KVM_PMU_H
>> +#define __ASM_ARM_KVM_PMU_H
>> +
>> +#include <asm/pmu.h>
>> +
>> +struct kvm_pmc {
>> + u8 idx;/* index into the pmu->pmc array */
>> + struct perf_event *perf_event;
>> + struct kvm_vcpu *vcpu;
>> +};
>> +
>
> Since you have a struct perf_event here, it makes more sense to #include
> linux/perf.h here.
>
Ok, will fix this.
>> +struct kvm_pmu {
>> +#ifdef CONFIG_KVM_ARM_PMU
>> + /* PMU IRQ Number per VCPU */
>> + int irq_num;
>> + /* IRQ pending flag */
>> + bool irq_pending;
>> + struct kvm_pmc pmc[ARMV8_MAX_COUNTERS];
>> +#endif
>> +};
>> +
>> +#endif
>>
>
> Thanks,
>
> M.
>
--
Shannon
^ permalink raw reply [flat|nested] 113+ messages in thread
* Re: [PATCH v2 02/22] KVM: ARM64: Define PMU data structure for each vcpu
@ 2015-09-11 9:58 ` Shannon Zhao
0 siblings, 0 replies; 113+ messages in thread
From: Shannon Zhao @ 2015-09-11 9:58 UTC (permalink / raw)
To: Marc Zyngier, kvmarm; +Cc: kvm, will.deacon, linux-arm-kernel, shannon.zhao
On 2015/9/11 17:10, Marc Zyngier wrote:
> On 11/09/15 09:54, Shannon Zhao wrote:
>> From: Shannon Zhao <shannon.zhao@linaro.org>
>>
>> Here we plan to support virtual PMU for guest by full software
>> emulation, so define some basic structs and functions preparing for
>> futher steps. Define struct kvm_pmc for performance monitor counter and
>> struct kvm_pmu for performance monitor unit for each vcpu. According to
>> ARMv8 spec, the PMU contains at most 32(ARMV8_MAX_COUNTERS) counters.
>>
>> Since this only supports ARM64 (or PMUv3), add a separate config symbol
>> for it.
>>
>> Signed-off-by: Shannon Zhao <shannon.zhao@linaro.org>
>> ---
>> arch/arm64/include/asm/kvm_host.h | 2 ++
>> arch/arm64/include/asm/pmu.h | 2 ++
>> arch/arm64/kvm/Kconfig | 8 ++++++++
>> include/kvm/arm_pmu.h | 39 +++++++++++++++++++++++++++++++++++++++
>> 4 files changed, 51 insertions(+)
>> create mode 100644 include/kvm/arm_pmu.h
>>
>> diff --git a/arch/arm64/include/asm/kvm_host.h b/arch/arm64/include/asm/kvm_host.h
>> index 2709db2..3c88873 100644
>> --- a/arch/arm64/include/asm/kvm_host.h
>> +++ b/arch/arm64/include/asm/kvm_host.h
>> @@ -42,6 +42,7 @@
>>
>> #include <kvm/arm_vgic.h>
>> #include <kvm/arm_arch_timer.h>
>> +#include <kvm/arm_pmu.h>
>>
>> #define KVM_VCPU_MAX_FEATURES 3
>>
>> @@ -116,6 +117,7 @@ struct kvm_vcpu_arch {
>> /* VGIC state */
>> struct vgic_cpu vgic_cpu;
>> struct arch_timer_cpu timer_cpu;
>> + struct kvm_pmu pmu;
>>
>> /*
>> * Anything that is not used directly from assembly code goes
>> diff --git a/arch/arm64/include/asm/pmu.h b/arch/arm64/include/asm/pmu.h
>> index b9f394a..95681e6 100644
>> --- a/arch/arm64/include/asm/pmu.h
>> +++ b/arch/arm64/include/asm/pmu.h
>> @@ -19,6 +19,8 @@
>> #ifndef __ASM_PMU_H
>> #define __ASM_PMU_H
>>
>> +#include <linux/perf_event.h>
>> +
>
> This include file seems out of place here, as I don't think anything in
> this file is directly PMU related.
>
>> #define ARMV8_MAX_COUNTERS 32
>> #define ARMV8_COUNTER_MASK (ARMV8_MAX_COUNTERS - 1)
>>
>> diff --git a/arch/arm64/kvm/Kconfig b/arch/arm64/kvm/Kconfig
>> index bfffe8f..3c7c58d 100644
>> --- a/arch/arm64/kvm/Kconfig
>> +++ b/arch/arm64/kvm/Kconfig
>> @@ -31,6 +31,7 @@ config KVM
>> select KVM_VFIO
>> select HAVE_KVM_EVENTFD
>> select HAVE_KVM_IRQFD
>> + select KVM_ARM_PMU
>> ---help---
>> Support hosting virtualized guest machines.
>>
>> @@ -52,4 +53,11 @@ config KVM_ARM_MAX_VCPUS
>> large, so only choose a reasonable number that you expect to
>> actually use.
>>
>> +config KVM_ARM_PMU
>> + bool
>> + depends on KVM_ARM_HOST
>> + ---help---
>> + Adds support for a virtual Performance Monitoring Unit (PMU) in
>> + virtual machines.
>> +
>> endif # VIRTUALIZATION
>> diff --git a/include/kvm/arm_pmu.h b/include/kvm/arm_pmu.h
>> new file mode 100644
>> index 0000000..64af88a
>> --- /dev/null
>> +++ b/include/kvm/arm_pmu.h
>> @@ -0,0 +1,39 @@
>> +/*
>> + * Copyright (C) 2015 Linaro Ltd.
>> + * Author: Shannon Zhao <shannon.zhao@linaro.org>
>> + *
>> + * This program is free software; you can redistribute it and/or modify
>> + * it under the terms of the GNU General Public License version 2 as
>> + * published by the Free Software Foundation.
>> + *
>> + * This program is distributed in the hope that it will be useful,
>> + * but WITHOUT ANY WARRANTY; without even the implied warranty of
>> + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
>> + * GNU General Public License for more details.
>> + *
>> + * You should have received a copy of the GNU General Public License
>> + * along with this program. If not, see <http://www.gnu.org/licenses/>.
>> + */
>> +
>> +#ifndef __ASM_ARM_KVM_PMU_H
>> +#define __ASM_ARM_KVM_PMU_H
>> +
>> +#include <asm/pmu.h>
>> +
>> +struct kvm_pmc {
>> + u8 idx;/* index into the pmu->pmc array */
>> + struct perf_event *perf_event;
>> + struct kvm_vcpu *vcpu;
>> +};
>> +
>
> Since you have a struct perf_event here, it makes more sense to #include
> linux/perf.h here.
>
Ok, will fix this.
>> +struct kvm_pmu {
>> +#ifdef CONFIG_KVM_ARM_PMU
>> + /* PMU IRQ Number per VCPU */
>> + int irq_num;
>> + /* IRQ pending flag */
>> + bool irq_pending;
>> + struct kvm_pmc pmc[ARMV8_MAX_COUNTERS];
>> +#endif
>> +};
>> +
>> +#endif
>>
>
> Thanks,
>
> M.
>
--
Shannon
^ permalink raw reply [flat|nested] 113+ messages in thread
* [PATCH v2 04/22] KVM: ARM64: Add reset and access handlers for PMCR_EL0 register
2015-09-11 8:54 ` Shannon Zhao
@ 2015-09-11 10:07 ` Marc Zyngier
-1 siblings, 0 replies; 113+ messages in thread
From: Marc Zyngier @ 2015-09-11 10:07 UTC (permalink / raw)
To: linux-arm-kernel
On 11/09/15 09:54, Shannon Zhao wrote:
> From: Shannon Zhao <shannon.zhao@linaro.org>
>
> Add reset handler which gets host value of PMCR_EL0 and make writable
> bits architecturally UNKNOWN. Add a common access handler for PMU
> registers which emulates writing and reading register and add emulation
> for PMCR.
>
> Signed-off-by: Shannon Zhao <shannon.zhao@linaro.org>
> ---
> arch/arm64/kvm/sys_regs.c | 76 +++++++++++++++++++++++++++++++++++++++++++++--
> 1 file changed, 74 insertions(+), 2 deletions(-)
>
> diff --git a/arch/arm64/kvm/sys_regs.c b/arch/arm64/kvm/sys_regs.c
> index c370b40..db1be44 100644
> --- a/arch/arm64/kvm/sys_regs.c
> +++ b/arch/arm64/kvm/sys_regs.c
> @@ -33,6 +33,7 @@
> #include <asm/kvm_emulate.h>
> #include <asm/kvm_host.h>
> #include <asm/kvm_mmu.h>
> +#include <asm/pmu.h>
>
> #include <trace/events/kvm.h>
>
> @@ -236,6 +237,48 @@ static void reset_mpidr(struct kvm_vcpu *vcpu, const struct sys_reg_desc *r)
> vcpu_sys_reg(vcpu, MPIDR_EL1) = (1ULL << 31) | mpidr;
> }
>
> +static void reset_pmcr(struct kvm_vcpu *vcpu, const struct sys_reg_desc *r)
> +{
> + u32 pmcr;
> +
> + asm volatile("mrs %0, pmcr_el0\n" : "=r" (pmcr));
> + /* Writable bits of PMCR_EL0 (ARMV8_PMCR_MASK) is reset to UNKNOWN*/
> + if (!vcpu_mode_is_32bit(vcpu))
> + vcpu_sys_reg(vcpu, r->reg) = (pmcr & ~ARMV8_PMCR_MASK)
> + | (ARMV8_PMCR_MASK & 0xdecafbad);
> + else
> + vcpu_cp15(vcpu, r->reg) = (pmcr & ~ARMV8_PMCR_MASK)
> + | (ARMV8_PMCR_MASK & 0xdecafbad);
I have some concerns about blindly reusing the top bits of the host's
PMCR_EL0 register, specially when it comes to the PMCR_EL0.N. Given that
we're fully emulating the PMU, shouldn't we simply define how many
counters we're emulating?
> +}
> +
> +/* PMU registers accessor. */
> +static bool access_pmu_regs(struct kvm_vcpu *vcpu,
> + const struct sys_reg_params *p,
> + const struct sys_reg_desc *r)
> +{
> + unsigned long val;
> +
> + if (p->is_write) {
> + switch (r->reg) {
> + case PMCR_EL0: {
> + /* Only update writeable bits of PMCR */
> + val = vcpu_sys_reg(vcpu, r->reg);
> + val &= ~ARMV8_PMCR_MASK;
> + val |= *vcpu_reg(vcpu, p->Rt) & ARMV8_PMCR_MASK;
> + vcpu_sys_reg(vcpu, r->reg) = val;
> + break;
> + }
> + default:
> + vcpu_sys_reg(vcpu, r->reg) = *vcpu_reg(vcpu, p->Rt);
> + break;
> + }
> + } else {
> + *vcpu_reg(vcpu, p->Rt) = vcpu_sys_reg(vcpu, r->reg);
> + }
> +
> + return true;
> +}
> +
> /* Silly macro to expand the DBG{BCR,BVR,WVR,WCR}n_EL1 registers in one go */
> #define DBG_BCR_BVR_WCR_WVR_EL1(n) \
> /* DBGBVRn_EL1 */ \
> @@ -427,7 +470,7 @@ static const struct sys_reg_desc sys_reg_descs[] = {
>
> /* PMCR_EL0 */
> { Op0(0b11), Op1(0b011), CRn(0b1001), CRm(0b1100), Op2(0b000),
> - trap_raz_wi },
> + access_pmu_regs, reset_pmcr, PMCR_EL0, },
> /* PMCNTENSET_EL0 */
> { Op0(0b11), Op1(0b011), CRn(0b1001), CRm(0b1100), Op2(0b001),
> trap_raz_wi },
> @@ -632,6 +675,34 @@ static const struct sys_reg_desc cp14_64_regs[] = {
> { Op1( 0), CRm( 2), .access = trap_raz_wi },
> };
>
> +/* PMU CP15 registers accessor. */
> +static bool access_pmu_cp15_regs(struct kvm_vcpu *vcpu,
> + const struct sys_reg_params *p,
> + const struct sys_reg_desc *r)
> +{
> + unsigned long val;
> +
> + if (p->is_write) {
> + switch (r->reg) {
> + case c9_PMCR: {
> + /* Only update writeable bits of PMCR */
> + val = vcpu_cp15(vcpu, r->reg);
> + val &= ~ARMV8_PMCR_MASK;
> + val |= *vcpu_reg(vcpu, p->Rt) & ARMV8_PMCR_MASK;
> + vcpu_cp15(vcpu, r->reg) = val;
> + break;
> + }
> + default:
> + vcpu_cp15(vcpu, r->reg) = *vcpu_reg(vcpu, p->Rt);
> + break;
> + }
> + } else {
> + *vcpu_reg(vcpu, p->Rt) = vcpu_cp15(vcpu, r->reg);
> + }
> +
> + return true;
> +}
> +
> /*
> * Trapped cp15 registers. TTBR0/TTBR1 get a double encoding,
> * depending on the way they are accessed (as a 32bit or a 64bit
> @@ -660,7 +731,8 @@ static const struct sys_reg_desc cp15_regs[] = {
> { Op1( 0), CRn( 7), CRm(14), Op2( 2), access_dcsw },
>
> /* PMU */
> - { Op1( 0), CRn( 9), CRm(12), Op2( 0), trap_raz_wi },
> + { Op1( 0), CRn( 9), CRm(12), Op2( 0), access_pmu_cp15_regs,
> + reset_pmcr, c9_PMCR },
> { Op1( 0), CRn( 9), CRm(12), Op2( 1), trap_raz_wi },
> { Op1( 0), CRn( 9), CRm(12), Op2( 2), trap_raz_wi },
> { Op1( 0), CRn( 9), CRm(12), Op2( 3), trap_raz_wi },
>
--
Jazz is not dead. It just smells funny...
^ permalink raw reply [flat|nested] 113+ messages in thread
* Re: [PATCH v2 04/22] KVM: ARM64: Add reset and access handlers for PMCR_EL0 register
@ 2015-09-11 10:07 ` Marc Zyngier
0 siblings, 0 replies; 113+ messages in thread
From: Marc Zyngier @ 2015-09-11 10:07 UTC (permalink / raw)
To: Shannon Zhao, kvmarm
Cc: linux-arm-kernel, kvm, christoffer.dall, will.deacon, alex.bennee,
wei, shannon.zhao, peter.huangpeng
On 11/09/15 09:54, Shannon Zhao wrote:
> From: Shannon Zhao <shannon.zhao@linaro.org>
>
> Add reset handler which gets host value of PMCR_EL0 and make writable
> bits architecturally UNKNOWN. Add a common access handler for PMU
> registers which emulates writing and reading register and add emulation
> for PMCR.
>
> Signed-off-by: Shannon Zhao <shannon.zhao@linaro.org>
> ---
> arch/arm64/kvm/sys_regs.c | 76 +++++++++++++++++++++++++++++++++++++++++++++--
> 1 file changed, 74 insertions(+), 2 deletions(-)
>
> diff --git a/arch/arm64/kvm/sys_regs.c b/arch/arm64/kvm/sys_regs.c
> index c370b40..db1be44 100644
> --- a/arch/arm64/kvm/sys_regs.c
> +++ b/arch/arm64/kvm/sys_regs.c
> @@ -33,6 +33,7 @@
> #include <asm/kvm_emulate.h>
> #include <asm/kvm_host.h>
> #include <asm/kvm_mmu.h>
> +#include <asm/pmu.h>
>
> #include <trace/events/kvm.h>
>
> @@ -236,6 +237,48 @@ static void reset_mpidr(struct kvm_vcpu *vcpu, const struct sys_reg_desc *r)
> vcpu_sys_reg(vcpu, MPIDR_EL1) = (1ULL << 31) | mpidr;
> }
>
> +static void reset_pmcr(struct kvm_vcpu *vcpu, const struct sys_reg_desc *r)
> +{
> + u32 pmcr;
> +
> + asm volatile("mrs %0, pmcr_el0\n" : "=r" (pmcr));
> + /* Writable bits of PMCR_EL0 (ARMV8_PMCR_MASK) is reset to UNKNOWN*/
> + if (!vcpu_mode_is_32bit(vcpu))
> + vcpu_sys_reg(vcpu, r->reg) = (pmcr & ~ARMV8_PMCR_MASK)
> + | (ARMV8_PMCR_MASK & 0xdecafbad);
> + else
> + vcpu_cp15(vcpu, r->reg) = (pmcr & ~ARMV8_PMCR_MASK)
> + | (ARMV8_PMCR_MASK & 0xdecafbad);
I have some concerns about blindly reusing the top bits of the host's
PMCR_EL0 register, specially when it comes to the PMCR_EL0.N. Given that
we're fully emulating the PMU, shouldn't we simply define how many
counters we're emulating?
> +}
> +
> +/* PMU registers accessor. */
> +static bool access_pmu_regs(struct kvm_vcpu *vcpu,
> + const struct sys_reg_params *p,
> + const struct sys_reg_desc *r)
> +{
> + unsigned long val;
> +
> + if (p->is_write) {
> + switch (r->reg) {
> + case PMCR_EL0: {
> + /* Only update writeable bits of PMCR */
> + val = vcpu_sys_reg(vcpu, r->reg);
> + val &= ~ARMV8_PMCR_MASK;
> + val |= *vcpu_reg(vcpu, p->Rt) & ARMV8_PMCR_MASK;
> + vcpu_sys_reg(vcpu, r->reg) = val;
> + break;
> + }
> + default:
> + vcpu_sys_reg(vcpu, r->reg) = *vcpu_reg(vcpu, p->Rt);
> + break;
> + }
> + } else {
> + *vcpu_reg(vcpu, p->Rt) = vcpu_sys_reg(vcpu, r->reg);
> + }
> +
> + return true;
> +}
> +
> /* Silly macro to expand the DBG{BCR,BVR,WVR,WCR}n_EL1 registers in one go */
> #define DBG_BCR_BVR_WCR_WVR_EL1(n) \
> /* DBGBVRn_EL1 */ \
> @@ -427,7 +470,7 @@ static const struct sys_reg_desc sys_reg_descs[] = {
>
> /* PMCR_EL0 */
> { Op0(0b11), Op1(0b011), CRn(0b1001), CRm(0b1100), Op2(0b000),
> - trap_raz_wi },
> + access_pmu_regs, reset_pmcr, PMCR_EL0, },
> /* PMCNTENSET_EL0 */
> { Op0(0b11), Op1(0b011), CRn(0b1001), CRm(0b1100), Op2(0b001),
> trap_raz_wi },
> @@ -632,6 +675,34 @@ static const struct sys_reg_desc cp14_64_regs[] = {
> { Op1( 0), CRm( 2), .access = trap_raz_wi },
> };
>
> +/* PMU CP15 registers accessor. */
> +static bool access_pmu_cp15_regs(struct kvm_vcpu *vcpu,
> + const struct sys_reg_params *p,
> + const struct sys_reg_desc *r)
> +{
> + unsigned long val;
> +
> + if (p->is_write) {
> + switch (r->reg) {
> + case c9_PMCR: {
> + /* Only update writeable bits of PMCR */
> + val = vcpu_cp15(vcpu, r->reg);
> + val &= ~ARMV8_PMCR_MASK;
> + val |= *vcpu_reg(vcpu, p->Rt) & ARMV8_PMCR_MASK;
> + vcpu_cp15(vcpu, r->reg) = val;
> + break;
> + }
> + default:
> + vcpu_cp15(vcpu, r->reg) = *vcpu_reg(vcpu, p->Rt);
> + break;
> + }
> + } else {
> + *vcpu_reg(vcpu, p->Rt) = vcpu_cp15(vcpu, r->reg);
> + }
> +
> + return true;
> +}
> +
> /*
> * Trapped cp15 registers. TTBR0/TTBR1 get a double encoding,
> * depending on the way they are accessed (as a 32bit or a 64bit
> @@ -660,7 +731,8 @@ static const struct sys_reg_desc cp15_regs[] = {
> { Op1( 0), CRn( 7), CRm(14), Op2( 2), access_dcsw },
>
> /* PMU */
> - { Op1( 0), CRn( 9), CRm(12), Op2( 0), trap_raz_wi },
> + { Op1( 0), CRn( 9), CRm(12), Op2( 0), access_pmu_cp15_regs,
> + reset_pmcr, c9_PMCR },
> { Op1( 0), CRn( 9), CRm(12), Op2( 1), trap_raz_wi },
> { Op1( 0), CRn( 9), CRm(12), Op2( 2), trap_raz_wi },
> { Op1( 0), CRn( 9), CRm(12), Op2( 3), trap_raz_wi },
>
--
Jazz is not dead. It just smells funny...
^ permalink raw reply [flat|nested] 113+ messages in thread
* [PATCH v2 05/22] KVM: ARM64: Add a helper for CP15 registers reset to UNKNOWN
2015-09-11 8:54 ` Shannon Zhao
@ 2015-09-11 10:16 ` Marc Zyngier
-1 siblings, 0 replies; 113+ messages in thread
From: Marc Zyngier @ 2015-09-11 10:16 UTC (permalink / raw)
To: linux-arm-kernel
On 11/09/15 09:54, Shannon Zhao wrote:
> From: Shannon Zhao <shannon.zhao@linaro.org>
>
> Signed-off-by: Shannon Zhao <shannon.zhao@linaro.org>
> ---
> arch/arm64/kvm/sys_regs.h | 8 ++++++++
> 1 file changed, 8 insertions(+)
>
> diff --git a/arch/arm64/kvm/sys_regs.h b/arch/arm64/kvm/sys_regs.h
> index d411e25..a0b3811 100644
> --- a/arch/arm64/kvm/sys_regs.h
> +++ b/arch/arm64/kvm/sys_regs.h
> @@ -104,6 +104,14 @@ static inline void reset_unknown(struct kvm_vcpu *vcpu,
> vcpu_sys_reg(vcpu, r->reg) = 0x1de7ec7edbadc0deULL;
> }
>
> +static inline void reset_unknown_cp15(struct kvm_vcpu *vcpu,
> + const struct sys_reg_desc *r)
> +{
> + BUG_ON(!r->reg);
> + BUG_ON(r->reg >= NR_COPRO_REGS);
> + vcpu_cp15(vcpu, r->reg) = 0xdecafbad;
Given how many times we use this constant, maybe we should consider
having a #define for it?
Thanks,
M.
--
Jazz is not dead. It just smells funny...
^ permalink raw reply [flat|nested] 113+ messages in thread
* Re: [PATCH v2 05/22] KVM: ARM64: Add a helper for CP15 registers reset to UNKNOWN
@ 2015-09-11 10:16 ` Marc Zyngier
0 siblings, 0 replies; 113+ messages in thread
From: Marc Zyngier @ 2015-09-11 10:16 UTC (permalink / raw)
To: Shannon Zhao, kvmarm
Cc: linux-arm-kernel, kvm, christoffer.dall, will.deacon, alex.bennee,
wei, shannon.zhao, peter.huangpeng
On 11/09/15 09:54, Shannon Zhao wrote:
> From: Shannon Zhao <shannon.zhao@linaro.org>
>
> Signed-off-by: Shannon Zhao <shannon.zhao@linaro.org>
> ---
> arch/arm64/kvm/sys_regs.h | 8 ++++++++
> 1 file changed, 8 insertions(+)
>
> diff --git a/arch/arm64/kvm/sys_regs.h b/arch/arm64/kvm/sys_regs.h
> index d411e25..a0b3811 100644
> --- a/arch/arm64/kvm/sys_regs.h
> +++ b/arch/arm64/kvm/sys_regs.h
> @@ -104,6 +104,14 @@ static inline void reset_unknown(struct kvm_vcpu *vcpu,
> vcpu_sys_reg(vcpu, r->reg) = 0x1de7ec7edbadc0deULL;
> }
>
> +static inline void reset_unknown_cp15(struct kvm_vcpu *vcpu,
> + const struct sys_reg_desc *r)
> +{
> + BUG_ON(!r->reg);
> + BUG_ON(r->reg >= NR_COPRO_REGS);
> + vcpu_cp15(vcpu, r->reg) = 0xdecafbad;
Given how many times we use this constant, maybe we should consider
having a #define for it?
Thanks,
M.
--
Jazz is not dead. It just smells funny...
^ permalink raw reply [flat|nested] 113+ messages in thread
* [PATCH v2 05/22] KVM: ARM64: Add a helper for CP15 registers reset to UNKNOWN
2015-09-11 8:54 ` Shannon Zhao
@ 2015-09-11 10:17 ` Marc Zyngier
-1 siblings, 0 replies; 113+ messages in thread
From: Marc Zyngier @ 2015-09-11 10:17 UTC (permalink / raw)
To: linux-arm-kernel
On 11/09/15 09:54, Shannon Zhao wrote:
> From: Shannon Zhao <shannon.zhao@linaro.org>
>
> Signed-off-by: Shannon Zhao <shannon.zhao@linaro.org>
> ---
> arch/arm64/kvm/sys_regs.h | 8 ++++++++
> 1 file changed, 8 insertions(+)
>
> diff --git a/arch/arm64/kvm/sys_regs.h b/arch/arm64/kvm/sys_regs.h
> index d411e25..a0b3811 100644
> --- a/arch/arm64/kvm/sys_regs.h
> +++ b/arch/arm64/kvm/sys_regs.h
> @@ -104,6 +104,14 @@ static inline void reset_unknown(struct kvm_vcpu *vcpu,
> vcpu_sys_reg(vcpu, r->reg) = 0x1de7ec7edbadc0deULL;
> }
>
> +static inline void reset_unknown_cp15(struct kvm_vcpu *vcpu,
> + const struct sys_reg_desc *r)
> +{
Oh, and given that you are using this as a function pointer in the next
patch, why is it an inline function?
Please merge this with patch #6.
M.
--
Jazz is not dead. It just smells funny...
^ permalink raw reply [flat|nested] 113+ messages in thread
* Re: [PATCH v2 05/22] KVM: ARM64: Add a helper for CP15 registers reset to UNKNOWN
@ 2015-09-11 10:17 ` Marc Zyngier
0 siblings, 0 replies; 113+ messages in thread
From: Marc Zyngier @ 2015-09-11 10:17 UTC (permalink / raw)
To: Shannon Zhao, kvmarm
Cc: linux-arm-kernel, kvm, christoffer.dall, will.deacon, alex.bennee,
wei, shannon.zhao, peter.huangpeng
On 11/09/15 09:54, Shannon Zhao wrote:
> From: Shannon Zhao <shannon.zhao@linaro.org>
>
> Signed-off-by: Shannon Zhao <shannon.zhao@linaro.org>
> ---
> arch/arm64/kvm/sys_regs.h | 8 ++++++++
> 1 file changed, 8 insertions(+)
>
> diff --git a/arch/arm64/kvm/sys_regs.h b/arch/arm64/kvm/sys_regs.h
> index d411e25..a0b3811 100644
> --- a/arch/arm64/kvm/sys_regs.h
> +++ b/arch/arm64/kvm/sys_regs.h
> @@ -104,6 +104,14 @@ static inline void reset_unknown(struct kvm_vcpu *vcpu,
> vcpu_sys_reg(vcpu, r->reg) = 0x1de7ec7edbadc0deULL;
> }
>
> +static inline void reset_unknown_cp15(struct kvm_vcpu *vcpu,
> + const struct sys_reg_desc *r)
> +{
Oh, and given that you are using this as a function pointer in the next
patch, why is it an inline function?
Please merge this with patch #6.
M.
--
Jazz is not dead. It just smells funny...
^ permalink raw reply [flat|nested] 113+ messages in thread
* [PATCH v2 07/22] KVM: ARM64: Add reset and access handlers for PMCEID0 and PMCEID1 register
2015-09-11 8:55 ` Shannon Zhao
@ 2015-09-11 10:27 ` Marc Zyngier
-1 siblings, 0 replies; 113+ messages in thread
From: Marc Zyngier @ 2015-09-11 10:27 UTC (permalink / raw)
To: linux-arm-kernel
On 11/09/15 09:55, Shannon Zhao wrote:
> From: Shannon Zhao <shannon.zhao@linaro.org>
>
> Add reset handler which gets host value of PMCEID0 or PMCEID1. Since
> write action to PMCEID0 or PMCEID1 is ignored, add a new case for this.
>
> Signed-off-by: Shannon Zhao <shannon.zhao@linaro.org>
> ---
> arch/arm64/kvm/sys_regs.c | 36 ++++++++++++++++++++++++++++++++----
> 1 file changed, 32 insertions(+), 4 deletions(-)
>
> diff --git a/arch/arm64/kvm/sys_regs.c b/arch/arm64/kvm/sys_regs.c
> index 24b8972..b3bc717 100644
> --- a/arch/arm64/kvm/sys_regs.c
> +++ b/arch/arm64/kvm/sys_regs.c
> @@ -251,6 +251,26 @@ static void reset_pmcr(struct kvm_vcpu *vcpu, const struct sys_reg_desc *r)
> | (ARMV8_PMCR_MASK & 0xdecafbad);
> }
>
> +static void reset_pmceid(struct kvm_vcpu *vcpu, const struct sys_reg_desc *r)
> +{
> + u32 pmceid;
> +
> + if (r->reg == PMCEID0_EL0 || r->reg == c9_PMCEID0) {
> + asm volatile("mrs %0, pmceid0_el0\n" : "=r" (pmceid));
Careful here. mrs always acts on a 64bit quantity, even if the register is
internally 32bit. I'd rather you use a u64 variable.
> + if (!vcpu_mode_is_32bit(vcpu))
> + vcpu_sys_reg(vcpu, r->reg) = pmceid;
> + else
> + vcpu_cp15(vcpu, r->reg) = pmceid;
> + } else {
> + /* PMCEID1_EL0 or c9_PMCEID1 */
> + asm volatile("mrs %0, pmceid1_el0\n" : "=r" (pmceid));
> + if (!vcpu_mode_is_32bit(vcpu))
> + vcpu_sys_reg(vcpu, r->reg) = pmceid;
> + else
> + vcpu_cp15(vcpu, r->reg) = pmceid;
Maybe we could have a helper for this kind of sequence:
static void vcpu_sysreg_write(vcpu, const struct sys_reg_desc *r, u64 val)
{
if (!vcpu_mode_is_32_bit(vcpu))
vcpu_sys_reg(vcpu, r->reg) = val;
else
vcpu_cp15(vcpu, r->reg) = lower_32_bit(val);
}
> + }
> +}
> +
> /* PMU registers accessor. */
> static bool access_pmu_regs(struct kvm_vcpu *vcpu,
> const struct sys_reg_params *p,
> @@ -268,6 +288,9 @@ static bool access_pmu_regs(struct kvm_vcpu *vcpu,
> vcpu_sys_reg(vcpu, r->reg) = val;
> break;
> }
> + case PMCEID0_EL0:
> + case PMCEID1_EL0:
> + return ignore_write(vcpu, p);
> default:
> vcpu_sys_reg(vcpu, r->reg) = *vcpu_reg(vcpu, p->Rt);
> break;
> @@ -488,10 +511,10 @@ static const struct sys_reg_desc sys_reg_descs[] = {
> access_pmu_regs, reset_unknown, PMSELR_EL0 },
> /* PMCEID0_EL0 */
> { Op0(0b11), Op1(0b011), CRn(0b1001), CRm(0b1100), Op2(0b110),
> - trap_raz_wi },
> + access_pmu_regs, reset_pmceid, PMCEID0_EL0 },
> /* PMCEID1_EL0 */
> { Op0(0b11), Op1(0b011), CRn(0b1001), CRm(0b1100), Op2(0b111),
> - trap_raz_wi },
> + access_pmu_regs, reset_pmceid, PMCEID1_EL0 },
> /* PMCCNTR_EL0 */
> { Op0(0b11), Op1(0b011), CRn(0b1001), CRm(0b1101), Op2(0b000),
> trap_raz_wi },
> @@ -692,6 +715,9 @@ static bool access_pmu_cp15_regs(struct kvm_vcpu *vcpu,
> vcpu_cp15(vcpu, r->reg) = val;
> break;
> }
> + case c9_PMCEID0:
> + case c9_PMCEID1:
> + return ignore_write(vcpu, p);
> default:
> vcpu_cp15(vcpu, r->reg) = *vcpu_reg(vcpu, p->Rt);
> break;
> @@ -738,8 +764,10 @@ static const struct sys_reg_desc cp15_regs[] = {
> { Op1( 0), CRn( 9), CRm(12), Op2( 3), trap_raz_wi },
> { Op1( 0), CRn( 9), CRm(12), Op2( 5), access_pmu_cp15_regs,
> reset_unknown_cp15, c9_PMSELR },
> - { Op1( 0), CRn( 9), CRm(12), Op2( 6), trap_raz_wi },
> - { Op1( 0), CRn( 9), CRm(12), Op2( 7), trap_raz_wi },
> + { Op1( 0), CRn( 9), CRm(12), Op2( 6), access_pmu_cp15_regs,
> + reset_pmceid, c9_PMCEID0 },
> + { Op1( 0), CRn( 9), CRm(12), Op2( 7), access_pmu_cp15_regs,
> + reset_pmceid, c9_PMCEID1 },
> { Op1( 0), CRn( 9), CRm(13), Op2( 0), trap_raz_wi },
> { Op1( 0), CRn( 9), CRm(13), Op2( 1), trap_raz_wi },
> { Op1( 0), CRn( 9), CRm(13), Op2( 2), trap_raz_wi },
>
--
Jazz is not dead. It just smells funny...
^ permalink raw reply [flat|nested] 113+ messages in thread
* Re: [PATCH v2 07/22] KVM: ARM64: Add reset and access handlers for PMCEID0 and PMCEID1 register
@ 2015-09-11 10:27 ` Marc Zyngier
0 siblings, 0 replies; 113+ messages in thread
From: Marc Zyngier @ 2015-09-11 10:27 UTC (permalink / raw)
To: Shannon Zhao, kvmarm
Cc: linux-arm-kernel, kvm, christoffer.dall, will.deacon, alex.bennee,
wei, shannon.zhao, peter.huangpeng
On 11/09/15 09:55, Shannon Zhao wrote:
> From: Shannon Zhao <shannon.zhao@linaro.org>
>
> Add reset handler which gets host value of PMCEID0 or PMCEID1. Since
> write action to PMCEID0 or PMCEID1 is ignored, add a new case for this.
>
> Signed-off-by: Shannon Zhao <shannon.zhao@linaro.org>
> ---
> arch/arm64/kvm/sys_regs.c | 36 ++++++++++++++++++++++++++++++++----
> 1 file changed, 32 insertions(+), 4 deletions(-)
>
> diff --git a/arch/arm64/kvm/sys_regs.c b/arch/arm64/kvm/sys_regs.c
> index 24b8972..b3bc717 100644
> --- a/arch/arm64/kvm/sys_regs.c
> +++ b/arch/arm64/kvm/sys_regs.c
> @@ -251,6 +251,26 @@ static void reset_pmcr(struct kvm_vcpu *vcpu, const struct sys_reg_desc *r)
> | (ARMV8_PMCR_MASK & 0xdecafbad);
> }
>
> +static void reset_pmceid(struct kvm_vcpu *vcpu, const struct sys_reg_desc *r)
> +{
> + u32 pmceid;
> +
> + if (r->reg == PMCEID0_EL0 || r->reg == c9_PMCEID0) {
> + asm volatile("mrs %0, pmceid0_el0\n" : "=r" (pmceid));
Careful here. mrs always acts on a 64bit quantity, even if the register is
internally 32bit. I'd rather you use a u64 variable.
> + if (!vcpu_mode_is_32bit(vcpu))
> + vcpu_sys_reg(vcpu, r->reg) = pmceid;
> + else
> + vcpu_cp15(vcpu, r->reg) = pmceid;
> + } else {
> + /* PMCEID1_EL0 or c9_PMCEID1 */
> + asm volatile("mrs %0, pmceid1_el0\n" : "=r" (pmceid));
> + if (!vcpu_mode_is_32bit(vcpu))
> + vcpu_sys_reg(vcpu, r->reg) = pmceid;
> + else
> + vcpu_cp15(vcpu, r->reg) = pmceid;
Maybe we could have a helper for this kind of sequence:
static void vcpu_sysreg_write(vcpu, const struct sys_reg_desc *r, u64 val)
{
if (!vcpu_mode_is_32_bit(vcpu))
vcpu_sys_reg(vcpu, r->reg) = val;
else
vcpu_cp15(vcpu, r->reg) = lower_32_bit(val);
}
> + }
> +}
> +
> /* PMU registers accessor. */
> static bool access_pmu_regs(struct kvm_vcpu *vcpu,
> const struct sys_reg_params *p,
> @@ -268,6 +288,9 @@ static bool access_pmu_regs(struct kvm_vcpu *vcpu,
> vcpu_sys_reg(vcpu, r->reg) = val;
> break;
> }
> + case PMCEID0_EL0:
> + case PMCEID1_EL0:
> + return ignore_write(vcpu, p);
> default:
> vcpu_sys_reg(vcpu, r->reg) = *vcpu_reg(vcpu, p->Rt);
> break;
> @@ -488,10 +511,10 @@ static const struct sys_reg_desc sys_reg_descs[] = {
> access_pmu_regs, reset_unknown, PMSELR_EL0 },
> /* PMCEID0_EL0 */
> { Op0(0b11), Op1(0b011), CRn(0b1001), CRm(0b1100), Op2(0b110),
> - trap_raz_wi },
> + access_pmu_regs, reset_pmceid, PMCEID0_EL0 },
> /* PMCEID1_EL0 */
> { Op0(0b11), Op1(0b011), CRn(0b1001), CRm(0b1100), Op2(0b111),
> - trap_raz_wi },
> + access_pmu_regs, reset_pmceid, PMCEID1_EL0 },
> /* PMCCNTR_EL0 */
> { Op0(0b11), Op1(0b011), CRn(0b1001), CRm(0b1101), Op2(0b000),
> trap_raz_wi },
> @@ -692,6 +715,9 @@ static bool access_pmu_cp15_regs(struct kvm_vcpu *vcpu,
> vcpu_cp15(vcpu, r->reg) = val;
> break;
> }
> + case c9_PMCEID0:
> + case c9_PMCEID1:
> + return ignore_write(vcpu, p);
> default:
> vcpu_cp15(vcpu, r->reg) = *vcpu_reg(vcpu, p->Rt);
> break;
> @@ -738,8 +764,10 @@ static const struct sys_reg_desc cp15_regs[] = {
> { Op1( 0), CRn( 9), CRm(12), Op2( 3), trap_raz_wi },
> { Op1( 0), CRn( 9), CRm(12), Op2( 5), access_pmu_cp15_regs,
> reset_unknown_cp15, c9_PMSELR },
> - { Op1( 0), CRn( 9), CRm(12), Op2( 6), trap_raz_wi },
> - { Op1( 0), CRn( 9), CRm(12), Op2( 7), trap_raz_wi },
> + { Op1( 0), CRn( 9), CRm(12), Op2( 6), access_pmu_cp15_regs,
> + reset_pmceid, c9_PMCEID0 },
> + { Op1( 0), CRn( 9), CRm(12), Op2( 7), access_pmu_cp15_regs,
> + reset_pmceid, c9_PMCEID1 },
> { Op1( 0), CRn( 9), CRm(13), Op2( 0), trap_raz_wi },
> { Op1( 0), CRn( 9), CRm(13), Op2( 1), trap_raz_wi },
> { Op1( 0), CRn( 9), CRm(13), Op2( 2), trap_raz_wi },
>
--
Jazz is not dead. It just smells funny...
^ permalink raw reply [flat|nested] 113+ messages in thread
* [PATCH v2 08/22] KVM: ARM64: PMU: Add perf event map and introduce perf event creating function
2015-09-11 8:55 ` Shannon Zhao
@ 2015-09-11 11:04 ` Marc Zyngier
-1 siblings, 0 replies; 113+ messages in thread
From: Marc Zyngier @ 2015-09-11 11:04 UTC (permalink / raw)
To: linux-arm-kernel
On 11/09/15 09:55, Shannon Zhao wrote:
> From: Shannon Zhao <shannon.zhao@linaro.org>
>
> When we use tools like perf on host, perf passes the event type and the
> id of this event type category to kernel, then kernel will map them to
> hardware event number and write this number to PMU PMEVTYPER<n>_EL0
> register. While we're trapping and emulating guest accesses to PMU
> registers, we get the hardware event number and map it to the event type
> and the id reversely. Then call perf_event kernel API to create an event
> for it.
>
> Signed-off-by: Shannon Zhao <shannon.zhao@linaro.org>
> ---
> arch/arm64/include/asm/pmu.h | 2 +
> arch/arm64/kvm/Makefile | 1 +
> include/kvm/arm_pmu.h | 15 +++
> virt/kvm/arm/pmu.c | 240 +++++++++++++++++++++++++++++++++++++++++++
> 4 files changed, 258 insertions(+)
> create mode 100644 virt/kvm/arm/pmu.c
>
> diff --git a/arch/arm64/include/asm/pmu.h b/arch/arm64/include/asm/pmu.h
> index 95681e6..42e7093 100644
> --- a/arch/arm64/include/asm/pmu.h
> +++ b/arch/arm64/include/asm/pmu.h
> @@ -33,6 +33,8 @@
> #define ARMV8_PMCR_D (1 << 3) /* CCNT counts every 64th cpu cycle */
> #define ARMV8_PMCR_X (1 << 4) /* Export to ETM */
> #define ARMV8_PMCR_DP (1 << 5) /* Disable CCNT if non-invasive debug*/
> +/* Determines which PMCCNTR_EL0 bit generates an overflow */
> +#define ARMV8_PMCR_LC (1 << 6)
> #define ARMV8_PMCR_N_SHIFT 11 /* Number of counters supported */
> #define ARMV8_PMCR_N_MASK 0x1f
> #define ARMV8_PMCR_MASK 0x3f /* Mask for writable bits */
> diff --git a/arch/arm64/kvm/Makefile b/arch/arm64/kvm/Makefile
> index f90f4aa..78db4ee 100644
> --- a/arch/arm64/kvm/Makefile
> +++ b/arch/arm64/kvm/Makefile
> @@ -27,3 +27,4 @@ kvm-$(CONFIG_KVM_ARM_HOST) += $(KVM)/arm/vgic-v3.o
> kvm-$(CONFIG_KVM_ARM_HOST) += $(KVM)/arm/vgic-v3-emul.o
> kvm-$(CONFIG_KVM_ARM_HOST) += vgic-v3-switch.o
> kvm-$(CONFIG_KVM_ARM_HOST) += $(KVM)/arm/arch_timer.o
> +kvm-$(CONFIG_KVM_ARM_PMU) += $(KVM)/arm/pmu.o
> diff --git a/include/kvm/arm_pmu.h b/include/kvm/arm_pmu.h
> index 64af88a..387ec6f 100644
> --- a/include/kvm/arm_pmu.h
> +++ b/include/kvm/arm_pmu.h
> @@ -36,4 +36,19 @@ struct kvm_pmu {
> #endif
> };
>
> +#ifdef CONFIG_KVM_ARM_PMU
> +unsigned long kvm_pmu_get_counter_value(struct kvm_vcpu *vcpu,
> + unsigned long select_idx);
> +void kvm_pmu_set_counter_event_type(struct kvm_vcpu *vcpu, unsigned long data,
> + unsigned long select_idx);
> +#else
> +unsigned long kvm_pmu_get_counter_value(struct kvm_vcpu *vcpu,
> + unsigned long select_idx)
> +{
> + return 0;
> +}
> +void kvm_pmu_set_counter_event_type(struct kvm_vcpu *vcpu, unsigned long data,
> + unsigned long select_idx) {}
> +#endif
> +
> #endif
> diff --git a/virt/kvm/arm/pmu.c b/virt/kvm/arm/pmu.c
> new file mode 100644
> index 0000000..0c7fe5c
> --- /dev/null
> +++ b/virt/kvm/arm/pmu.c
> @@ -0,0 +1,240 @@
> +/*
> + * Copyright (C) 2015 Linaro Ltd.
> + * Author: Shannon Zhao <shannon.zhao@linaro.org>
> + *
> + * This program is free software; you can redistribute it and/or modify
> + * it under the terms of the GNU General Public License version 2 as
> + * published by the Free Software Foundation.
> + *
> + * This program is distributed in the hope that it will be useful,
> + * but WITHOUT ANY WARRANTY; without even the implied warranty of
> + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
> + * GNU General Public License for more details.
> + *
> + * You should have received a copy of the GNU General Public License
> + * along with this program. If not, see <http://www.gnu.org/licenses/>.
> + */
> +
> +#include <linux/cpu.h>
> +#include <linux/kvm.h>
> +#include <linux/kvm_host.h>
> +#include <linux/perf_event.h>
> +#include <asm/kvm_emulate.h>
> +#include <kvm/arm_pmu.h>
> +
> +/* PMU HW events mapping. */
> +static struct kvm_pmu_hw_event_map {
> + unsigned eventsel;
> + unsigned event_type;
> +} kvm_pmu_hw_events[] = {
> + [0] = { 0x11, PERF_COUNT_HW_CPU_CYCLES },
> + [1] = { 0x08, PERF_COUNT_HW_INSTRUCTIONS },
> + [2] = { 0x04, PERF_COUNT_HW_CACHE_REFERENCES },
> + [3] = { 0x03, PERF_COUNT_HW_CACHE_MISSES },
> + [4] = { 0x10, PERF_COUNT_HW_BRANCH_MISSES },
How about using enum armv8_pmuv3_perf_types here?
> +};
> +
> +/* PMU HW cache events mapping. */
> +static struct kvm_pmu_hw_cache_event_map {
> + unsigned eventsel;
> + unsigned cache_type;
> + unsigned cache_op;
> + unsigned cache_result;
> +} kvm_pmu_hw_cache_events[] = {
> + [0] = { 0x12, PERF_COUNT_HW_CACHE_BPU, PERF_COUNT_HW_CACHE_OP_READ,
> + PERF_COUNT_HW_CACHE_RESULT_ACCESS },
> + [1] = { 0x12, PERF_COUNT_HW_CACHE_BPU, PERF_COUNT_HW_CACHE_OP_WRITE,
> + PERF_COUNT_HW_CACHE_RESULT_ACCESS },
> +};
> +
> +static void kvm_pmu_set_evttyper(struct kvm_vcpu *vcpu, unsigned long idx,
> + unsigned long val)
> +{
> + if (!vcpu_mode_is_32bit(vcpu))
> + vcpu_sys_reg(vcpu, PMEVTYPER0_EL0 + idx) = val;
> + else
> + vcpu_cp15(vcpu, c14_PMEVTYPER0 + idx) = val;
> +}
> +
> +static unsigned long kvm_pmu_get_evttyper(struct kvm_vcpu *vcpu,
> + unsigned long idx)
> +{
> + if (!vcpu_mode_is_32bit(vcpu))
> + return vcpu_sys_reg(vcpu, PMEVTYPER0_EL0 + idx)
> + & ARMV8_EVTYPE_EVENT;
> + else
> + return vcpu_cp15(vcpu, c14_PMEVTYPER0 + idx)
> + & ARMV8_EVTYPE_EVENT;
> +}
> +
> +/**
> + * kvm_pmu_stop_counter - stop PMU counter for the selected counter
> + * @vcpu: The vcpu pointer
> + * @select_idx: The counter index
> + *
> + * If this counter has been configured to monitor some event, disable and
> + * release it.
> + */
> +static void kvm_pmu_stop_counter(struct kvm_vcpu *vcpu,
> + unsigned long select_idx)
> +{
> + struct kvm_pmu *pmu = &vcpu->arch.pmu;
> + struct kvm_pmc *pmc = &pmu->pmc[select_idx];
> +
> + if (pmc->perf_event) {
> + perf_event_disable(pmc->perf_event);
> + perf_event_release_kernel(pmc->perf_event);
> + pmc->perf_event = NULL;
> + }
> + kvm_pmu_set_evttyper(vcpu, select_idx, ARMV8_EVTYPE_EVENT);
> +}
> +
> +/**
> + * kvm_pmu_get_counter_value - get PMU counter value
> + * @vcpu: The vcpu pointer
> + * @select_idx: The counter index
> + */
> +unsigned long kvm_pmu_get_counter_value(struct kvm_vcpu *vcpu,
> + unsigned long select_idx)
> +{
> + u64 enabled, running;
> + struct kvm_pmu *pmu = &vcpu->arch.pmu;
> + struct kvm_pmc *pmc = &pmu->pmc[select_idx];
> + unsigned long counter;
> +
> + if (!vcpu_mode_is_32bit(vcpu))
> + counter = vcpu_sys_reg(vcpu, PMEVCNTR0_EL0 + select_idx);
> + else
> + counter = vcpu_cp15(vcpu, c14_PMEVCNTR0 + select_idx);
> +
> + if (pmc->perf_event) {
> + counter += perf_event_read_value(pmc->perf_event,
> + &enabled, &running);
> + }
> + return counter;
> +}
> +
> +/**
> + * kvm_pmu_find_hw_event - find hardware event
> + * @pmu: The pmu pointer
> + * @event_select: The number of selected event type
> + *
> + * Based on the number of selected event type, find out whether it belongs to
> + * PERF_TYPE_HARDWARE. If so, return the corresponding event id.
> + */
> +static unsigned kvm_pmu_find_hw_event(struct kvm_pmu *pmu,
> + unsigned long event_select)
> +{
> + int i;
> +
> + for (i = 0; i < ARRAY_SIZE(kvm_pmu_hw_events); i++)
> + if (kvm_pmu_hw_events[i].eventsel == event_select)
> + return kvm_pmu_hw_events[i].event_type;
> +
> + return PERF_COUNT_HW_MAX;
> +}
> +
> +/**
> + * kvm_pmu_find_hw_cache_event - find hardware cache event
> + * @pmu: The pmu pointer
> + * @event_select: The number of selected event type
> + *
> + * Based on the number of selected event type, find out whether it belongs to
> + * PERF_TYPE_HW_CACHE. If so, return the corresponding event id.
> + */
> +static unsigned kvm_pmu_find_hw_cache_event(struct kvm_pmu *pmu,
> + unsigned long event_select)
> +{
> + int i;
> + unsigned config;
Please use an explicitely sized type (u32, u64).
> +
> + for (i = 0; i < ARRAY_SIZE(kvm_pmu_hw_cache_events); i++)
> + if (kvm_pmu_hw_cache_events[i].eventsel == event_select) {
> + config = (kvm_pmu_hw_cache_events[i].cache_type & 0xff)
> + | ((kvm_pmu_hw_cache_events[i].cache_op & 0xff) << 8)
> + | ((kvm_pmu_hw_cache_events[i].cache_result & 0xff) << 16);
I don't understand what this does. You only update a local variable?
> + }
> +
> + return PERF_COUNT_HW_CACHE_MAX;
> +}
> +
> +/**
> + * kvm_pmu_set_counter_event_type - set selected counter to monitor some event
> + * @vcpu: The vcpu pointer
> + * @data: The data guest writes to PMXEVTYPER_EL0
> + * @select_idx: The number of selected counter
> + *
> + * When OS accesses PMXEVTYPER_EL0, that means it wants to set a PMC to count an
> + * event with given hardware event number. Here we call perf_event API to
> + * emulate this action and create a kernel perf event for it.
> + */
> +void kvm_pmu_set_counter_event_type(struct kvm_vcpu *vcpu, unsigned long data,
> + unsigned long select_idx)
> +{
> + struct kvm_pmu *pmu = &vcpu->arch.pmu;
> + struct kvm_pmc *pmc = &pmu->pmc[select_idx];
> + struct perf_event *event;
> + struct perf_event_attr attr;
> + unsigned config, type = PERF_TYPE_RAW;
> + unsigned int new_eventsel, old_eventsel;
> + u64 counter;
> + int overflow_bit, pmcr_lc;
> +
> + old_eventsel = kvm_pmu_get_evttyper(vcpu, select_idx);
> + new_eventsel = data & ARMV8_EVTYPE_EVENT;
> + if (new_eventsel == old_eventsel) {
> + if (pmc->perf_event)
> + local64_set(&pmc->perf_event->count, 0);
> + return;
> + }
> +
> + kvm_pmu_stop_counter(vcpu, select_idx);
> + kvm_pmu_set_evttyper(vcpu, select_idx, data);
> +
> + config = kvm_pmu_find_hw_event(pmu, new_eventsel);
> + if (config != PERF_COUNT_HW_MAX) {
> + type = PERF_TYPE_HARDWARE;
> + } else {
> + config = kvm_pmu_find_hw_cache_event(pmu, new_eventsel);
> + if (config != PERF_COUNT_HW_CACHE_MAX)
> + type = PERF_TYPE_HW_CACHE;
> + }
> +
> + if (type == PERF_TYPE_RAW)
> + config = new_eventsel;
> +
> + memset(&attr, 0, sizeof(struct perf_event_attr));
> + attr.type = type;
> + attr.size = sizeof(attr);
> + attr.pinned = 1;
> + attr.disabled = 1;
> + attr.exclude_user = data & ARMV8_EXCLUDE_EL0 ? 1 : 0;
> + attr.exclude_kernel = data & ARMV8_EXCLUDE_EL1 ? 1 : 0;
> + attr.exclude_host = 1; /* Don't count host events */
> + attr.config = config;
> +
> + overflow_bit = 31; /* Generic counters are 32-bit registers*/
> + if (new_eventsel == 0x11) {
> + /* Cycle counter overflow on increment that changes PMCCNTR[63]
> + * or PMCCNTR[31] from 1 to 0 according to the value of
> + * ARMV8_PMCR_LC
> + */
> + if (!vcpu_mode_is_32bit(vcpu))
> + pmcr_lc = vcpu_sys_reg(vcpu, PMCR_EL0) & ARMV8_PMCR_LC;
> + else
> + pmcr_lc = vcpu_cp15(vcpu, c9_PMCR) & ARMV8_PMCR_LC;
> +
> + overflow_bit = pmcr_lc ? 63 : 31;
> + }
> + counter = kvm_pmu_get_counter_value(vcpu, select_idx);
> + /* The initial sample period (overflow count) of an event. */
> + attr.sample_period = (-counter) & (((u64)1 << overflow_bit) - 1);
> +
> + event = perf_event_create_kernel_counter(&attr, -1, current, NULL, pmc);
> + if (IS_ERR(event)) {
> + printk_once("kvm: pmu event creation failed %ld\n",
> + PTR_ERR(event));
> + return;
> + }
> + pmc->perf_event = event;
> +}
>
Having had a chat with Will, it appears that a much better solution
would be to ask perf to use raw events instead of trying to map things
to perf events (which the guest has already done).
See drivers/oprofile/oprofile_perf.c::op_perf_setup().
Thoughts?
M.
--
Jazz is not dead. It just smells funny...
^ permalink raw reply [flat|nested] 113+ messages in thread
* Re: [PATCH v2 08/22] KVM: ARM64: PMU: Add perf event map and introduce perf event creating function
@ 2015-09-11 11:04 ` Marc Zyngier
0 siblings, 0 replies; 113+ messages in thread
From: Marc Zyngier @ 2015-09-11 11:04 UTC (permalink / raw)
To: Shannon Zhao, kvmarm; +Cc: kvm, will.deacon, linux-arm-kernel, shannon.zhao
On 11/09/15 09:55, Shannon Zhao wrote:
> From: Shannon Zhao <shannon.zhao@linaro.org>
>
> When we use tools like perf on host, perf passes the event type and the
> id of this event type category to kernel, then kernel will map them to
> hardware event number and write this number to PMU PMEVTYPER<n>_EL0
> register. While we're trapping and emulating guest accesses to PMU
> registers, we get the hardware event number and map it to the event type
> and the id reversely. Then call perf_event kernel API to create an event
> for it.
>
> Signed-off-by: Shannon Zhao <shannon.zhao@linaro.org>
> ---
> arch/arm64/include/asm/pmu.h | 2 +
> arch/arm64/kvm/Makefile | 1 +
> include/kvm/arm_pmu.h | 15 +++
> virt/kvm/arm/pmu.c | 240 +++++++++++++++++++++++++++++++++++++++++++
> 4 files changed, 258 insertions(+)
> create mode 100644 virt/kvm/arm/pmu.c
>
> diff --git a/arch/arm64/include/asm/pmu.h b/arch/arm64/include/asm/pmu.h
> index 95681e6..42e7093 100644
> --- a/arch/arm64/include/asm/pmu.h
> +++ b/arch/arm64/include/asm/pmu.h
> @@ -33,6 +33,8 @@
> #define ARMV8_PMCR_D (1 << 3) /* CCNT counts every 64th cpu cycle */
> #define ARMV8_PMCR_X (1 << 4) /* Export to ETM */
> #define ARMV8_PMCR_DP (1 << 5) /* Disable CCNT if non-invasive debug*/
> +/* Determines which PMCCNTR_EL0 bit generates an overflow */
> +#define ARMV8_PMCR_LC (1 << 6)
> #define ARMV8_PMCR_N_SHIFT 11 /* Number of counters supported */
> #define ARMV8_PMCR_N_MASK 0x1f
> #define ARMV8_PMCR_MASK 0x3f /* Mask for writable bits */
> diff --git a/arch/arm64/kvm/Makefile b/arch/arm64/kvm/Makefile
> index f90f4aa..78db4ee 100644
> --- a/arch/arm64/kvm/Makefile
> +++ b/arch/arm64/kvm/Makefile
> @@ -27,3 +27,4 @@ kvm-$(CONFIG_KVM_ARM_HOST) += $(KVM)/arm/vgic-v3.o
> kvm-$(CONFIG_KVM_ARM_HOST) += $(KVM)/arm/vgic-v3-emul.o
> kvm-$(CONFIG_KVM_ARM_HOST) += vgic-v3-switch.o
> kvm-$(CONFIG_KVM_ARM_HOST) += $(KVM)/arm/arch_timer.o
> +kvm-$(CONFIG_KVM_ARM_PMU) += $(KVM)/arm/pmu.o
> diff --git a/include/kvm/arm_pmu.h b/include/kvm/arm_pmu.h
> index 64af88a..387ec6f 100644
> --- a/include/kvm/arm_pmu.h
> +++ b/include/kvm/arm_pmu.h
> @@ -36,4 +36,19 @@ struct kvm_pmu {
> #endif
> };
>
> +#ifdef CONFIG_KVM_ARM_PMU
> +unsigned long kvm_pmu_get_counter_value(struct kvm_vcpu *vcpu,
> + unsigned long select_idx);
> +void kvm_pmu_set_counter_event_type(struct kvm_vcpu *vcpu, unsigned long data,
> + unsigned long select_idx);
> +#else
> +unsigned long kvm_pmu_get_counter_value(struct kvm_vcpu *vcpu,
> + unsigned long select_idx)
> +{
> + return 0;
> +}
> +void kvm_pmu_set_counter_event_type(struct kvm_vcpu *vcpu, unsigned long data,
> + unsigned long select_idx) {}
> +#endif
> +
> #endif
> diff --git a/virt/kvm/arm/pmu.c b/virt/kvm/arm/pmu.c
> new file mode 100644
> index 0000000..0c7fe5c
> --- /dev/null
> +++ b/virt/kvm/arm/pmu.c
> @@ -0,0 +1,240 @@
> +/*
> + * Copyright (C) 2015 Linaro Ltd.
> + * Author: Shannon Zhao <shannon.zhao@linaro.org>
> + *
> + * This program is free software; you can redistribute it and/or modify
> + * it under the terms of the GNU General Public License version 2 as
> + * published by the Free Software Foundation.
> + *
> + * This program is distributed in the hope that it will be useful,
> + * but WITHOUT ANY WARRANTY; without even the implied warranty of
> + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
> + * GNU General Public License for more details.
> + *
> + * You should have received a copy of the GNU General Public License
> + * along with this program. If not, see <http://www.gnu.org/licenses/>.
> + */
> +
> +#include <linux/cpu.h>
> +#include <linux/kvm.h>
> +#include <linux/kvm_host.h>
> +#include <linux/perf_event.h>
> +#include <asm/kvm_emulate.h>
> +#include <kvm/arm_pmu.h>
> +
> +/* PMU HW events mapping. */
> +static struct kvm_pmu_hw_event_map {
> + unsigned eventsel;
> + unsigned event_type;
> +} kvm_pmu_hw_events[] = {
> + [0] = { 0x11, PERF_COUNT_HW_CPU_CYCLES },
> + [1] = { 0x08, PERF_COUNT_HW_INSTRUCTIONS },
> + [2] = { 0x04, PERF_COUNT_HW_CACHE_REFERENCES },
> + [3] = { 0x03, PERF_COUNT_HW_CACHE_MISSES },
> + [4] = { 0x10, PERF_COUNT_HW_BRANCH_MISSES },
How about using enum armv8_pmuv3_perf_types here?
> +};
> +
> +/* PMU HW cache events mapping. */
> +static struct kvm_pmu_hw_cache_event_map {
> + unsigned eventsel;
> + unsigned cache_type;
> + unsigned cache_op;
> + unsigned cache_result;
> +} kvm_pmu_hw_cache_events[] = {
> + [0] = { 0x12, PERF_COUNT_HW_CACHE_BPU, PERF_COUNT_HW_CACHE_OP_READ,
> + PERF_COUNT_HW_CACHE_RESULT_ACCESS },
> + [1] = { 0x12, PERF_COUNT_HW_CACHE_BPU, PERF_COUNT_HW_CACHE_OP_WRITE,
> + PERF_COUNT_HW_CACHE_RESULT_ACCESS },
> +};
> +
> +static void kvm_pmu_set_evttyper(struct kvm_vcpu *vcpu, unsigned long idx,
> + unsigned long val)
> +{
> + if (!vcpu_mode_is_32bit(vcpu))
> + vcpu_sys_reg(vcpu, PMEVTYPER0_EL0 + idx) = val;
> + else
> + vcpu_cp15(vcpu, c14_PMEVTYPER0 + idx) = val;
> +}
> +
> +static unsigned long kvm_pmu_get_evttyper(struct kvm_vcpu *vcpu,
> + unsigned long idx)
> +{
> + if (!vcpu_mode_is_32bit(vcpu))
> + return vcpu_sys_reg(vcpu, PMEVTYPER0_EL0 + idx)
> + & ARMV8_EVTYPE_EVENT;
> + else
> + return vcpu_cp15(vcpu, c14_PMEVTYPER0 + idx)
> + & ARMV8_EVTYPE_EVENT;
> +}
> +
> +/**
> + * kvm_pmu_stop_counter - stop PMU counter for the selected counter
> + * @vcpu: The vcpu pointer
> + * @select_idx: The counter index
> + *
> + * If this counter has been configured to monitor some event, disable and
> + * release it.
> + */
> +static void kvm_pmu_stop_counter(struct kvm_vcpu *vcpu,
> + unsigned long select_idx)
> +{
> + struct kvm_pmu *pmu = &vcpu->arch.pmu;
> + struct kvm_pmc *pmc = &pmu->pmc[select_idx];
> +
> + if (pmc->perf_event) {
> + perf_event_disable(pmc->perf_event);
> + perf_event_release_kernel(pmc->perf_event);
> + pmc->perf_event = NULL;
> + }
> + kvm_pmu_set_evttyper(vcpu, select_idx, ARMV8_EVTYPE_EVENT);
> +}
> +
> +/**
> + * kvm_pmu_get_counter_value - get PMU counter value
> + * @vcpu: The vcpu pointer
> + * @select_idx: The counter index
> + */
> +unsigned long kvm_pmu_get_counter_value(struct kvm_vcpu *vcpu,
> + unsigned long select_idx)
> +{
> + u64 enabled, running;
> + struct kvm_pmu *pmu = &vcpu->arch.pmu;
> + struct kvm_pmc *pmc = &pmu->pmc[select_idx];
> + unsigned long counter;
> +
> + if (!vcpu_mode_is_32bit(vcpu))
> + counter = vcpu_sys_reg(vcpu, PMEVCNTR0_EL0 + select_idx);
> + else
> + counter = vcpu_cp15(vcpu, c14_PMEVCNTR0 + select_idx);
> +
> + if (pmc->perf_event) {
> + counter += perf_event_read_value(pmc->perf_event,
> + &enabled, &running);
> + }
> + return counter;
> +}
> +
> +/**
> + * kvm_pmu_find_hw_event - find hardware event
> + * @pmu: The pmu pointer
> + * @event_select: The number of selected event type
> + *
> + * Based on the number of selected event type, find out whether it belongs to
> + * PERF_TYPE_HARDWARE. If so, return the corresponding event id.
> + */
> +static unsigned kvm_pmu_find_hw_event(struct kvm_pmu *pmu,
> + unsigned long event_select)
> +{
> + int i;
> +
> + for (i = 0; i < ARRAY_SIZE(kvm_pmu_hw_events); i++)
> + if (kvm_pmu_hw_events[i].eventsel == event_select)
> + return kvm_pmu_hw_events[i].event_type;
> +
> + return PERF_COUNT_HW_MAX;
> +}
> +
> +/**
> + * kvm_pmu_find_hw_cache_event - find hardware cache event
> + * @pmu: The pmu pointer
> + * @event_select: The number of selected event type
> + *
> + * Based on the number of selected event type, find out whether it belongs to
> + * PERF_TYPE_HW_CACHE. If so, return the corresponding event id.
> + */
> +static unsigned kvm_pmu_find_hw_cache_event(struct kvm_pmu *pmu,
> + unsigned long event_select)
> +{
> + int i;
> + unsigned config;
Please use an explicitely sized type (u32, u64).
> +
> + for (i = 0; i < ARRAY_SIZE(kvm_pmu_hw_cache_events); i++)
> + if (kvm_pmu_hw_cache_events[i].eventsel == event_select) {
> + config = (kvm_pmu_hw_cache_events[i].cache_type & 0xff)
> + | ((kvm_pmu_hw_cache_events[i].cache_op & 0xff) << 8)
> + | ((kvm_pmu_hw_cache_events[i].cache_result & 0xff) << 16);
I don't understand what this does. You only update a local variable?
> + }
> +
> + return PERF_COUNT_HW_CACHE_MAX;
> +}
> +
> +/**
> + * kvm_pmu_set_counter_event_type - set selected counter to monitor some event
> + * @vcpu: The vcpu pointer
> + * @data: The data guest writes to PMXEVTYPER_EL0
> + * @select_idx: The number of selected counter
> + *
> + * When OS accesses PMXEVTYPER_EL0, that means it wants to set a PMC to count an
> + * event with given hardware event number. Here we call perf_event API to
> + * emulate this action and create a kernel perf event for it.
> + */
> +void kvm_pmu_set_counter_event_type(struct kvm_vcpu *vcpu, unsigned long data,
> + unsigned long select_idx)
> +{
> + struct kvm_pmu *pmu = &vcpu->arch.pmu;
> + struct kvm_pmc *pmc = &pmu->pmc[select_idx];
> + struct perf_event *event;
> + struct perf_event_attr attr;
> + unsigned config, type = PERF_TYPE_RAW;
> + unsigned int new_eventsel, old_eventsel;
> + u64 counter;
> + int overflow_bit, pmcr_lc;
> +
> + old_eventsel = kvm_pmu_get_evttyper(vcpu, select_idx);
> + new_eventsel = data & ARMV8_EVTYPE_EVENT;
> + if (new_eventsel == old_eventsel) {
> + if (pmc->perf_event)
> + local64_set(&pmc->perf_event->count, 0);
> + return;
> + }
> +
> + kvm_pmu_stop_counter(vcpu, select_idx);
> + kvm_pmu_set_evttyper(vcpu, select_idx, data);
> +
> + config = kvm_pmu_find_hw_event(pmu, new_eventsel);
> + if (config != PERF_COUNT_HW_MAX) {
> + type = PERF_TYPE_HARDWARE;
> + } else {
> + config = kvm_pmu_find_hw_cache_event(pmu, new_eventsel);
> + if (config != PERF_COUNT_HW_CACHE_MAX)
> + type = PERF_TYPE_HW_CACHE;
> + }
> +
> + if (type == PERF_TYPE_RAW)
> + config = new_eventsel;
> +
> + memset(&attr, 0, sizeof(struct perf_event_attr));
> + attr.type = type;
> + attr.size = sizeof(attr);
> + attr.pinned = 1;
> + attr.disabled = 1;
> + attr.exclude_user = data & ARMV8_EXCLUDE_EL0 ? 1 : 0;
> + attr.exclude_kernel = data & ARMV8_EXCLUDE_EL1 ? 1 : 0;
> + attr.exclude_host = 1; /* Don't count host events */
> + attr.config = config;
> +
> + overflow_bit = 31; /* Generic counters are 32-bit registers*/
> + if (new_eventsel == 0x11) {
> + /* Cycle counter overflow on increment that changes PMCCNTR[63]
> + * or PMCCNTR[31] from 1 to 0 according to the value of
> + * ARMV8_PMCR_LC
> + */
> + if (!vcpu_mode_is_32bit(vcpu))
> + pmcr_lc = vcpu_sys_reg(vcpu, PMCR_EL0) & ARMV8_PMCR_LC;
> + else
> + pmcr_lc = vcpu_cp15(vcpu, c9_PMCR) & ARMV8_PMCR_LC;
> +
> + overflow_bit = pmcr_lc ? 63 : 31;
> + }
> + counter = kvm_pmu_get_counter_value(vcpu, select_idx);
> + /* The initial sample period (overflow count) of an event. */
> + attr.sample_period = (-counter) & (((u64)1 << overflow_bit) - 1);
> +
> + event = perf_event_create_kernel_counter(&attr, -1, current, NULL, pmc);
> + if (IS_ERR(event)) {
> + printk_once("kvm: pmu event creation failed %ld\n",
> + PTR_ERR(event));
> + return;
> + }
> + pmc->perf_event = event;
> +}
>
Having had a chat with Will, it appears that a much better solution
would be to ask perf to use raw events instead of trying to map things
to perf events (which the guest has already done).
See drivers/oprofile/oprofile_perf.c::op_perf_setup().
Thoughts?
M.
--
Jazz is not dead. It just smells funny...
^ permalink raw reply [flat|nested] 113+ messages in thread
* [PATCH v2 08/22] KVM: ARM64: PMU: Add perf event map and introduce perf event creating function
2015-09-11 11:04 ` Marc Zyngier
@ 2015-09-11 13:35 ` Shannon Zhao
-1 siblings, 0 replies; 113+ messages in thread
From: Shannon Zhao @ 2015-09-11 13:35 UTC (permalink / raw)
To: linux-arm-kernel
On 2015/9/11 19:04, Marc Zyngier wrote:
> On 11/09/15 09:55, Shannon Zhao wrote:
>> From: Shannon Zhao <shannon.zhao@linaro.org>
>>
>> When we use tools like perf on host, perf passes the event type and the
>> id of this event type category to kernel, then kernel will map them to
>> hardware event number and write this number to PMU PMEVTYPER<n>_EL0
>> register. While we're trapping and emulating guest accesses to PMU
>> registers, we get the hardware event number and map it to the event type
>> and the id reversely. Then call perf_event kernel API to create an event
>> for it.
>>
>> Signed-off-by: Shannon Zhao <shannon.zhao@linaro.org>
>> ---
>> arch/arm64/include/asm/pmu.h | 2 +
>> arch/arm64/kvm/Makefile | 1 +
>> include/kvm/arm_pmu.h | 15 +++
>> virt/kvm/arm/pmu.c | 240 +++++++++++++++++++++++++++++++++++++++++++
>> 4 files changed, 258 insertions(+)
>> create mode 100644 virt/kvm/arm/pmu.c
>>
>> diff --git a/arch/arm64/include/asm/pmu.h b/arch/arm64/include/asm/pmu.h
>> index 95681e6..42e7093 100644
>> --- a/arch/arm64/include/asm/pmu.h
>> +++ b/arch/arm64/include/asm/pmu.h
>> @@ -33,6 +33,8 @@
>> #define ARMV8_PMCR_D (1 << 3) /* CCNT counts every 64th cpu cycle */
>> #define ARMV8_PMCR_X (1 << 4) /* Export to ETM */
>> #define ARMV8_PMCR_DP (1 << 5) /* Disable CCNT if non-invasive debug*/
>> +/* Determines which PMCCNTR_EL0 bit generates an overflow */
>> +#define ARMV8_PMCR_LC (1 << 6)
>> #define ARMV8_PMCR_N_SHIFT 11 /* Number of counters supported */
>> #define ARMV8_PMCR_N_MASK 0x1f
>> #define ARMV8_PMCR_MASK 0x3f /* Mask for writable bits */
>> diff --git a/arch/arm64/kvm/Makefile b/arch/arm64/kvm/Makefile
>> index f90f4aa..78db4ee 100644
>> --- a/arch/arm64/kvm/Makefile
>> +++ b/arch/arm64/kvm/Makefile
>> @@ -27,3 +27,4 @@ kvm-$(CONFIG_KVM_ARM_HOST) += $(KVM)/arm/vgic-v3.o
>> kvm-$(CONFIG_KVM_ARM_HOST) += $(KVM)/arm/vgic-v3-emul.o
>> kvm-$(CONFIG_KVM_ARM_HOST) += vgic-v3-switch.o
>> kvm-$(CONFIG_KVM_ARM_HOST) += $(KVM)/arm/arch_timer.o
>> +kvm-$(CONFIG_KVM_ARM_PMU) += $(KVM)/arm/pmu.o
>> diff --git a/include/kvm/arm_pmu.h b/include/kvm/arm_pmu.h
>> index 64af88a..387ec6f 100644
>> --- a/include/kvm/arm_pmu.h
>> +++ b/include/kvm/arm_pmu.h
>> @@ -36,4 +36,19 @@ struct kvm_pmu {
>> #endif
>> };
>>
>> +#ifdef CONFIG_KVM_ARM_PMU
>> +unsigned long kvm_pmu_get_counter_value(struct kvm_vcpu *vcpu,
>> + unsigned long select_idx);
>> +void kvm_pmu_set_counter_event_type(struct kvm_vcpu *vcpu, unsigned long data,
>> + unsigned long select_idx);
>> +#else
>> +unsigned long kvm_pmu_get_counter_value(struct kvm_vcpu *vcpu,
>> + unsigned long select_idx)
>> +{
>> + return 0;
>> +}
>> +void kvm_pmu_set_counter_event_type(struct kvm_vcpu *vcpu, unsigned long data,
>> + unsigned long select_idx) {}
>> +#endif
>> +
>> #endif
>> diff --git a/virt/kvm/arm/pmu.c b/virt/kvm/arm/pmu.c
>> new file mode 100644
>> index 0000000..0c7fe5c
>> --- /dev/null
>> +++ b/virt/kvm/arm/pmu.c
>> @@ -0,0 +1,240 @@
>> +/*
>> + * Copyright (C) 2015 Linaro Ltd.
>> + * Author: Shannon Zhao <shannon.zhao@linaro.org>
>> + *
>> + * This program is free software; you can redistribute it and/or modify
>> + * it under the terms of the GNU General Public License version 2 as
>> + * published by the Free Software Foundation.
>> + *
>> + * This program is distributed in the hope that it will be useful,
>> + * but WITHOUT ANY WARRANTY; without even the implied warranty of
>> + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
>> + * GNU General Public License for more details.
>> + *
>> + * You should have received a copy of the GNU General Public License
>> + * along with this program. If not, see <http://www.gnu.org/licenses/>.
>> + */
>> +
>> +#include <linux/cpu.h>
>> +#include <linux/kvm.h>
>> +#include <linux/kvm_host.h>
>> +#include <linux/perf_event.h>
>> +#include <asm/kvm_emulate.h>
>> +#include <kvm/arm_pmu.h>
>> +
>> +/* PMU HW events mapping. */
>> +static struct kvm_pmu_hw_event_map {
>> + unsigned eventsel;
>> + unsigned event_type;
>> +} kvm_pmu_hw_events[] = {
>> + [0] = { 0x11, PERF_COUNT_HW_CPU_CYCLES },
>> + [1] = { 0x08, PERF_COUNT_HW_INSTRUCTIONS },
>> + [2] = { 0x04, PERF_COUNT_HW_CACHE_REFERENCES },
>> + [3] = { 0x03, PERF_COUNT_HW_CACHE_MISSES },
>> + [4] = { 0x10, PERF_COUNT_HW_BRANCH_MISSES },
>
> How about using enum armv8_pmuv3_perf_types here?
>
>> +};
>> +
>> +/* PMU HW cache events mapping. */
>> +static struct kvm_pmu_hw_cache_event_map {
>> + unsigned eventsel;
>> + unsigned cache_type;
>> + unsigned cache_op;
>> + unsigned cache_result;
>> +} kvm_pmu_hw_cache_events[] = {
>> + [0] = { 0x12, PERF_COUNT_HW_CACHE_BPU, PERF_COUNT_HW_CACHE_OP_READ,
>> + PERF_COUNT_HW_CACHE_RESULT_ACCESS },
>> + [1] = { 0x12, PERF_COUNT_HW_CACHE_BPU, PERF_COUNT_HW_CACHE_OP_WRITE,
>> + PERF_COUNT_HW_CACHE_RESULT_ACCESS },
>> +};
>> +
>> +static void kvm_pmu_set_evttyper(struct kvm_vcpu *vcpu, unsigned long idx,
>> + unsigned long val)
>> +{
>> + if (!vcpu_mode_is_32bit(vcpu))
>> + vcpu_sys_reg(vcpu, PMEVTYPER0_EL0 + idx) = val;
>> + else
>> + vcpu_cp15(vcpu, c14_PMEVTYPER0 + idx) = val;
>> +}
>> +
>> +static unsigned long kvm_pmu_get_evttyper(struct kvm_vcpu *vcpu,
>> + unsigned long idx)
>> +{
>> + if (!vcpu_mode_is_32bit(vcpu))
>> + return vcpu_sys_reg(vcpu, PMEVTYPER0_EL0 + idx)
>> + & ARMV8_EVTYPE_EVENT;
>> + else
>> + return vcpu_cp15(vcpu, c14_PMEVTYPER0 + idx)
>> + & ARMV8_EVTYPE_EVENT;
>> +}
>> +
>> +/**
>> + * kvm_pmu_stop_counter - stop PMU counter for the selected counter
>> + * @vcpu: The vcpu pointer
>> + * @select_idx: The counter index
>> + *
>> + * If this counter has been configured to monitor some event, disable and
>> + * release it.
>> + */
>> +static void kvm_pmu_stop_counter(struct kvm_vcpu *vcpu,
>> + unsigned long select_idx)
>> +{
>> + struct kvm_pmu *pmu = &vcpu->arch.pmu;
>> + struct kvm_pmc *pmc = &pmu->pmc[select_idx];
>> +
>> + if (pmc->perf_event) {
>> + perf_event_disable(pmc->perf_event);
>> + perf_event_release_kernel(pmc->perf_event);
>> + pmc->perf_event = NULL;
>> + }
>> + kvm_pmu_set_evttyper(vcpu, select_idx, ARMV8_EVTYPE_EVENT);
>> +}
>> +
>> +/**
>> + * kvm_pmu_get_counter_value - get PMU counter value
>> + * @vcpu: The vcpu pointer
>> + * @select_idx: The counter index
>> + */
>> +unsigned long kvm_pmu_get_counter_value(struct kvm_vcpu *vcpu,
>> + unsigned long select_idx)
>> +{
>> + u64 enabled, running;
>> + struct kvm_pmu *pmu = &vcpu->arch.pmu;
>> + struct kvm_pmc *pmc = &pmu->pmc[select_idx];
>> + unsigned long counter;
>> +
>> + if (!vcpu_mode_is_32bit(vcpu))
>> + counter = vcpu_sys_reg(vcpu, PMEVCNTR0_EL0 + select_idx);
>> + else
>> + counter = vcpu_cp15(vcpu, c14_PMEVCNTR0 + select_idx);
>> +
>> + if (pmc->perf_event) {
>> + counter += perf_event_read_value(pmc->perf_event,
>> + &enabled, &running);
>> + }
>> + return counter;
>> +}
>> +
>> +/**
>> + * kvm_pmu_find_hw_event - find hardware event
>> + * @pmu: The pmu pointer
>> + * @event_select: The number of selected event type
>> + *
>> + * Based on the number of selected event type, find out whether it belongs to
>> + * PERF_TYPE_HARDWARE. If so, return the corresponding event id.
>> + */
>> +static unsigned kvm_pmu_find_hw_event(struct kvm_pmu *pmu,
>> + unsigned long event_select)
>> +{
>> + int i;
>> +
>> + for (i = 0; i < ARRAY_SIZE(kvm_pmu_hw_events); i++)
>> + if (kvm_pmu_hw_events[i].eventsel == event_select)
>> + return kvm_pmu_hw_events[i].event_type;
>> +
>> + return PERF_COUNT_HW_MAX;
>> +}
>> +
>> +/**
>> + * kvm_pmu_find_hw_cache_event - find hardware cache event
>> + * @pmu: The pmu pointer
>> + * @event_select: The number of selected event type
>> + *
>> + * Based on the number of selected event type, find out whether it belongs to
>> + * PERF_TYPE_HW_CACHE. If so, return the corresponding event id.
>> + */
>> +static unsigned kvm_pmu_find_hw_cache_event(struct kvm_pmu *pmu,
>> + unsigned long event_select)
>> +{
>> + int i;
>> + unsigned config;
>
> Please use an explicitely sized type (u32, u64).
>
ok.
>> +
>> + for (i = 0; i < ARRAY_SIZE(kvm_pmu_hw_cache_events); i++)
>> + if (kvm_pmu_hw_cache_events[i].eventsel == event_select) {
>> + config = (kvm_pmu_hw_cache_events[i].cache_type & 0xff)
>> + | ((kvm_pmu_hw_cache_events[i].cache_op & 0xff) << 8)
>> + | ((kvm_pmu_hw_cache_events[i].cache_result & 0xff) << 16);
>
> I don't understand what this does. You only update a local variable?
>
Oh, sorry, forgot "return config".
>> + }
>> +
>> + return PERF_COUNT_HW_CACHE_MAX;
>> +}
>> +
>> +/**
>> + * kvm_pmu_set_counter_event_type - set selected counter to monitor some event
>> + * @vcpu: The vcpu pointer
>> + * @data: The data guest writes to PMXEVTYPER_EL0
>> + * @select_idx: The number of selected counter
>> + *
>> + * When OS accesses PMXEVTYPER_EL0, that means it wants to set a PMC to count an
>> + * event with given hardware event number. Here we call perf_event API to
>> + * emulate this action and create a kernel perf event for it.
>> + */
>> +void kvm_pmu_set_counter_event_type(struct kvm_vcpu *vcpu, unsigned long data,
>> + unsigned long select_idx)
>> +{
>> + struct kvm_pmu *pmu = &vcpu->arch.pmu;
>> + struct kvm_pmc *pmc = &pmu->pmc[select_idx];
>> + struct perf_event *event;
>> + struct perf_event_attr attr;
>> + unsigned config, type = PERF_TYPE_RAW;
>> + unsigned int new_eventsel, old_eventsel;
>> + u64 counter;
>> + int overflow_bit, pmcr_lc;
>> +
>> + old_eventsel = kvm_pmu_get_evttyper(vcpu, select_idx);
>> + new_eventsel = data & ARMV8_EVTYPE_EVENT;
>> + if (new_eventsel == old_eventsel) {
>> + if (pmc->perf_event)
>> + local64_set(&pmc->perf_event->count, 0);
>> + return;
>> + }
>> +
>> + kvm_pmu_stop_counter(vcpu, select_idx);
>> + kvm_pmu_set_evttyper(vcpu, select_idx, data);
>> +
>> + config = kvm_pmu_find_hw_event(pmu, new_eventsel);
>> + if (config != PERF_COUNT_HW_MAX) {
>> + type = PERF_TYPE_HARDWARE;
>> + } else {
>> + config = kvm_pmu_find_hw_cache_event(pmu, new_eventsel);
>> + if (config != PERF_COUNT_HW_CACHE_MAX)
>> + type = PERF_TYPE_HW_CACHE;
>> + }
>> +
>> + if (type == PERF_TYPE_RAW)
>> + config = new_eventsel;
>> +
>> + memset(&attr, 0, sizeof(struct perf_event_attr));
>> + attr.type = type;
>> + attr.size = sizeof(attr);
>> + attr.pinned = 1;
>> + attr.disabled = 1;
>> + attr.exclude_user = data & ARMV8_EXCLUDE_EL0 ? 1 : 0;
>> + attr.exclude_kernel = data & ARMV8_EXCLUDE_EL1 ? 1 : 0;
>> + attr.exclude_host = 1; /* Don't count host events */
>> + attr.config = config;
>> +
>> + overflow_bit = 31; /* Generic counters are 32-bit registers*/
>> + if (new_eventsel == 0x11) {
>> + /* Cycle counter overflow on increment that changes PMCCNTR[63]
>> + * or PMCCNTR[31] from 1 to 0 according to the value of
>> + * ARMV8_PMCR_LC
>> + */
>> + if (!vcpu_mode_is_32bit(vcpu))
>> + pmcr_lc = vcpu_sys_reg(vcpu, PMCR_EL0) & ARMV8_PMCR_LC;
>> + else
>> + pmcr_lc = vcpu_cp15(vcpu, c9_PMCR) & ARMV8_PMCR_LC;
>> +
>> + overflow_bit = pmcr_lc ? 63 : 31;
>> + }
>> + counter = kvm_pmu_get_counter_value(vcpu, select_idx);
>> + /* The initial sample period (overflow count) of an event. */
>> + attr.sample_period = (-counter) & (((u64)1 << overflow_bit) - 1);
>> +
>> + event = perf_event_create_kernel_counter(&attr, -1, current, NULL, pmc);
>> + if (IS_ERR(event)) {
>> + printk_once("kvm: pmu event creation failed %ld\n",
>> + PTR_ERR(event));
>> + return;
>> + }
>> + pmc->perf_event = event;
>> +}
>>
>
> Having had a chat with Will, it appears that a much better solution
> would be to ask perf to use raw events instead of trying to map things
> to perf events (which the guest has already done).
>
> See drivers/oprofile/oprofile_perf.c::op_perf_setup().
>
> Thoughts?
>
Yeah, directly using PERF_TYPE_RAW event looks simpler. But should we
check whether the value of event number written to PMXEVTYPER_EL0 is
valid? Or That is guaranteed by guest?
--
Shannon
^ permalink raw reply [flat|nested] 113+ messages in thread
* Re: [PATCH v2 08/22] KVM: ARM64: PMU: Add perf event map and introduce perf event creating function
@ 2015-09-11 13:35 ` Shannon Zhao
0 siblings, 0 replies; 113+ messages in thread
From: Shannon Zhao @ 2015-09-11 13:35 UTC (permalink / raw)
To: Marc Zyngier, Shannon Zhao, kvmarm
Cc: linux-arm-kernel, kvm, christoffer.dall, will.deacon, alex.bennee,
wei, peter.huangpeng
On 2015/9/11 19:04, Marc Zyngier wrote:
> On 11/09/15 09:55, Shannon Zhao wrote:
>> From: Shannon Zhao <shannon.zhao@linaro.org>
>>
>> When we use tools like perf on host, perf passes the event type and the
>> id of this event type category to kernel, then kernel will map them to
>> hardware event number and write this number to PMU PMEVTYPER<n>_EL0
>> register. While we're trapping and emulating guest accesses to PMU
>> registers, we get the hardware event number and map it to the event type
>> and the id reversely. Then call perf_event kernel API to create an event
>> for it.
>>
>> Signed-off-by: Shannon Zhao <shannon.zhao@linaro.org>
>> ---
>> arch/arm64/include/asm/pmu.h | 2 +
>> arch/arm64/kvm/Makefile | 1 +
>> include/kvm/arm_pmu.h | 15 +++
>> virt/kvm/arm/pmu.c | 240 +++++++++++++++++++++++++++++++++++++++++++
>> 4 files changed, 258 insertions(+)
>> create mode 100644 virt/kvm/arm/pmu.c
>>
>> diff --git a/arch/arm64/include/asm/pmu.h b/arch/arm64/include/asm/pmu.h
>> index 95681e6..42e7093 100644
>> --- a/arch/arm64/include/asm/pmu.h
>> +++ b/arch/arm64/include/asm/pmu.h
>> @@ -33,6 +33,8 @@
>> #define ARMV8_PMCR_D (1 << 3) /* CCNT counts every 64th cpu cycle */
>> #define ARMV8_PMCR_X (1 << 4) /* Export to ETM */
>> #define ARMV8_PMCR_DP (1 << 5) /* Disable CCNT if non-invasive debug*/
>> +/* Determines which PMCCNTR_EL0 bit generates an overflow */
>> +#define ARMV8_PMCR_LC (1 << 6)
>> #define ARMV8_PMCR_N_SHIFT 11 /* Number of counters supported */
>> #define ARMV8_PMCR_N_MASK 0x1f
>> #define ARMV8_PMCR_MASK 0x3f /* Mask for writable bits */
>> diff --git a/arch/arm64/kvm/Makefile b/arch/arm64/kvm/Makefile
>> index f90f4aa..78db4ee 100644
>> --- a/arch/arm64/kvm/Makefile
>> +++ b/arch/arm64/kvm/Makefile
>> @@ -27,3 +27,4 @@ kvm-$(CONFIG_KVM_ARM_HOST) += $(KVM)/arm/vgic-v3.o
>> kvm-$(CONFIG_KVM_ARM_HOST) += $(KVM)/arm/vgic-v3-emul.o
>> kvm-$(CONFIG_KVM_ARM_HOST) += vgic-v3-switch.o
>> kvm-$(CONFIG_KVM_ARM_HOST) += $(KVM)/arm/arch_timer.o
>> +kvm-$(CONFIG_KVM_ARM_PMU) += $(KVM)/arm/pmu.o
>> diff --git a/include/kvm/arm_pmu.h b/include/kvm/arm_pmu.h
>> index 64af88a..387ec6f 100644
>> --- a/include/kvm/arm_pmu.h
>> +++ b/include/kvm/arm_pmu.h
>> @@ -36,4 +36,19 @@ struct kvm_pmu {
>> #endif
>> };
>>
>> +#ifdef CONFIG_KVM_ARM_PMU
>> +unsigned long kvm_pmu_get_counter_value(struct kvm_vcpu *vcpu,
>> + unsigned long select_idx);
>> +void kvm_pmu_set_counter_event_type(struct kvm_vcpu *vcpu, unsigned long data,
>> + unsigned long select_idx);
>> +#else
>> +unsigned long kvm_pmu_get_counter_value(struct kvm_vcpu *vcpu,
>> + unsigned long select_idx)
>> +{
>> + return 0;
>> +}
>> +void kvm_pmu_set_counter_event_type(struct kvm_vcpu *vcpu, unsigned long data,
>> + unsigned long select_idx) {}
>> +#endif
>> +
>> #endif
>> diff --git a/virt/kvm/arm/pmu.c b/virt/kvm/arm/pmu.c
>> new file mode 100644
>> index 0000000..0c7fe5c
>> --- /dev/null
>> +++ b/virt/kvm/arm/pmu.c
>> @@ -0,0 +1,240 @@
>> +/*
>> + * Copyright (C) 2015 Linaro Ltd.
>> + * Author: Shannon Zhao <shannon.zhao@linaro.org>
>> + *
>> + * This program is free software; you can redistribute it and/or modify
>> + * it under the terms of the GNU General Public License version 2 as
>> + * published by the Free Software Foundation.
>> + *
>> + * This program is distributed in the hope that it will be useful,
>> + * but WITHOUT ANY WARRANTY; without even the implied warranty of
>> + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
>> + * GNU General Public License for more details.
>> + *
>> + * You should have received a copy of the GNU General Public License
>> + * along with this program. If not, see <http://www.gnu.org/licenses/>.
>> + */
>> +
>> +#include <linux/cpu.h>
>> +#include <linux/kvm.h>
>> +#include <linux/kvm_host.h>
>> +#include <linux/perf_event.h>
>> +#include <asm/kvm_emulate.h>
>> +#include <kvm/arm_pmu.h>
>> +
>> +/* PMU HW events mapping. */
>> +static struct kvm_pmu_hw_event_map {
>> + unsigned eventsel;
>> + unsigned event_type;
>> +} kvm_pmu_hw_events[] = {
>> + [0] = { 0x11, PERF_COUNT_HW_CPU_CYCLES },
>> + [1] = { 0x08, PERF_COUNT_HW_INSTRUCTIONS },
>> + [2] = { 0x04, PERF_COUNT_HW_CACHE_REFERENCES },
>> + [3] = { 0x03, PERF_COUNT_HW_CACHE_MISSES },
>> + [4] = { 0x10, PERF_COUNT_HW_BRANCH_MISSES },
>
> How about using enum armv8_pmuv3_perf_types here?
>
>> +};
>> +
>> +/* PMU HW cache events mapping. */
>> +static struct kvm_pmu_hw_cache_event_map {
>> + unsigned eventsel;
>> + unsigned cache_type;
>> + unsigned cache_op;
>> + unsigned cache_result;
>> +} kvm_pmu_hw_cache_events[] = {
>> + [0] = { 0x12, PERF_COUNT_HW_CACHE_BPU, PERF_COUNT_HW_CACHE_OP_READ,
>> + PERF_COUNT_HW_CACHE_RESULT_ACCESS },
>> + [1] = { 0x12, PERF_COUNT_HW_CACHE_BPU, PERF_COUNT_HW_CACHE_OP_WRITE,
>> + PERF_COUNT_HW_CACHE_RESULT_ACCESS },
>> +};
>> +
>> +static void kvm_pmu_set_evttyper(struct kvm_vcpu *vcpu, unsigned long idx,
>> + unsigned long val)
>> +{
>> + if (!vcpu_mode_is_32bit(vcpu))
>> + vcpu_sys_reg(vcpu, PMEVTYPER0_EL0 + idx) = val;
>> + else
>> + vcpu_cp15(vcpu, c14_PMEVTYPER0 + idx) = val;
>> +}
>> +
>> +static unsigned long kvm_pmu_get_evttyper(struct kvm_vcpu *vcpu,
>> + unsigned long idx)
>> +{
>> + if (!vcpu_mode_is_32bit(vcpu))
>> + return vcpu_sys_reg(vcpu, PMEVTYPER0_EL0 + idx)
>> + & ARMV8_EVTYPE_EVENT;
>> + else
>> + return vcpu_cp15(vcpu, c14_PMEVTYPER0 + idx)
>> + & ARMV8_EVTYPE_EVENT;
>> +}
>> +
>> +/**
>> + * kvm_pmu_stop_counter - stop PMU counter for the selected counter
>> + * @vcpu: The vcpu pointer
>> + * @select_idx: The counter index
>> + *
>> + * If this counter has been configured to monitor some event, disable and
>> + * release it.
>> + */
>> +static void kvm_pmu_stop_counter(struct kvm_vcpu *vcpu,
>> + unsigned long select_idx)
>> +{
>> + struct kvm_pmu *pmu = &vcpu->arch.pmu;
>> + struct kvm_pmc *pmc = &pmu->pmc[select_idx];
>> +
>> + if (pmc->perf_event) {
>> + perf_event_disable(pmc->perf_event);
>> + perf_event_release_kernel(pmc->perf_event);
>> + pmc->perf_event = NULL;
>> + }
>> + kvm_pmu_set_evttyper(vcpu, select_idx, ARMV8_EVTYPE_EVENT);
>> +}
>> +
>> +/**
>> + * kvm_pmu_get_counter_value - get PMU counter value
>> + * @vcpu: The vcpu pointer
>> + * @select_idx: The counter index
>> + */
>> +unsigned long kvm_pmu_get_counter_value(struct kvm_vcpu *vcpu,
>> + unsigned long select_idx)
>> +{
>> + u64 enabled, running;
>> + struct kvm_pmu *pmu = &vcpu->arch.pmu;
>> + struct kvm_pmc *pmc = &pmu->pmc[select_idx];
>> + unsigned long counter;
>> +
>> + if (!vcpu_mode_is_32bit(vcpu))
>> + counter = vcpu_sys_reg(vcpu, PMEVCNTR0_EL0 + select_idx);
>> + else
>> + counter = vcpu_cp15(vcpu, c14_PMEVCNTR0 + select_idx);
>> +
>> + if (pmc->perf_event) {
>> + counter += perf_event_read_value(pmc->perf_event,
>> + &enabled, &running);
>> + }
>> + return counter;
>> +}
>> +
>> +/**
>> + * kvm_pmu_find_hw_event - find hardware event
>> + * @pmu: The pmu pointer
>> + * @event_select: The number of selected event type
>> + *
>> + * Based on the number of selected event type, find out whether it belongs to
>> + * PERF_TYPE_HARDWARE. If so, return the corresponding event id.
>> + */
>> +static unsigned kvm_pmu_find_hw_event(struct kvm_pmu *pmu,
>> + unsigned long event_select)
>> +{
>> + int i;
>> +
>> + for (i = 0; i < ARRAY_SIZE(kvm_pmu_hw_events); i++)
>> + if (kvm_pmu_hw_events[i].eventsel == event_select)
>> + return kvm_pmu_hw_events[i].event_type;
>> +
>> + return PERF_COUNT_HW_MAX;
>> +}
>> +
>> +/**
>> + * kvm_pmu_find_hw_cache_event - find hardware cache event
>> + * @pmu: The pmu pointer
>> + * @event_select: The number of selected event type
>> + *
>> + * Based on the number of selected event type, find out whether it belongs to
>> + * PERF_TYPE_HW_CACHE. If so, return the corresponding event id.
>> + */
>> +static unsigned kvm_pmu_find_hw_cache_event(struct kvm_pmu *pmu,
>> + unsigned long event_select)
>> +{
>> + int i;
>> + unsigned config;
>
> Please use an explicitely sized type (u32, u64).
>
ok.
>> +
>> + for (i = 0; i < ARRAY_SIZE(kvm_pmu_hw_cache_events); i++)
>> + if (kvm_pmu_hw_cache_events[i].eventsel == event_select) {
>> + config = (kvm_pmu_hw_cache_events[i].cache_type & 0xff)
>> + | ((kvm_pmu_hw_cache_events[i].cache_op & 0xff) << 8)
>> + | ((kvm_pmu_hw_cache_events[i].cache_result & 0xff) << 16);
>
> I don't understand what this does. You only update a local variable?
>
Oh, sorry, forgot "return config".
>> + }
>> +
>> + return PERF_COUNT_HW_CACHE_MAX;
>> +}
>> +
>> +/**
>> + * kvm_pmu_set_counter_event_type - set selected counter to monitor some event
>> + * @vcpu: The vcpu pointer
>> + * @data: The data guest writes to PMXEVTYPER_EL0
>> + * @select_idx: The number of selected counter
>> + *
>> + * When OS accesses PMXEVTYPER_EL0, that means it wants to set a PMC to count an
>> + * event with given hardware event number. Here we call perf_event API to
>> + * emulate this action and create a kernel perf event for it.
>> + */
>> +void kvm_pmu_set_counter_event_type(struct kvm_vcpu *vcpu, unsigned long data,
>> + unsigned long select_idx)
>> +{
>> + struct kvm_pmu *pmu = &vcpu->arch.pmu;
>> + struct kvm_pmc *pmc = &pmu->pmc[select_idx];
>> + struct perf_event *event;
>> + struct perf_event_attr attr;
>> + unsigned config, type = PERF_TYPE_RAW;
>> + unsigned int new_eventsel, old_eventsel;
>> + u64 counter;
>> + int overflow_bit, pmcr_lc;
>> +
>> + old_eventsel = kvm_pmu_get_evttyper(vcpu, select_idx);
>> + new_eventsel = data & ARMV8_EVTYPE_EVENT;
>> + if (new_eventsel == old_eventsel) {
>> + if (pmc->perf_event)
>> + local64_set(&pmc->perf_event->count, 0);
>> + return;
>> + }
>> +
>> + kvm_pmu_stop_counter(vcpu, select_idx);
>> + kvm_pmu_set_evttyper(vcpu, select_idx, data);
>> +
>> + config = kvm_pmu_find_hw_event(pmu, new_eventsel);
>> + if (config != PERF_COUNT_HW_MAX) {
>> + type = PERF_TYPE_HARDWARE;
>> + } else {
>> + config = kvm_pmu_find_hw_cache_event(pmu, new_eventsel);
>> + if (config != PERF_COUNT_HW_CACHE_MAX)
>> + type = PERF_TYPE_HW_CACHE;
>> + }
>> +
>> + if (type == PERF_TYPE_RAW)
>> + config = new_eventsel;
>> +
>> + memset(&attr, 0, sizeof(struct perf_event_attr));
>> + attr.type = type;
>> + attr.size = sizeof(attr);
>> + attr.pinned = 1;
>> + attr.disabled = 1;
>> + attr.exclude_user = data & ARMV8_EXCLUDE_EL0 ? 1 : 0;
>> + attr.exclude_kernel = data & ARMV8_EXCLUDE_EL1 ? 1 : 0;
>> + attr.exclude_host = 1; /* Don't count host events */
>> + attr.config = config;
>> +
>> + overflow_bit = 31; /* Generic counters are 32-bit registers*/
>> + if (new_eventsel == 0x11) {
>> + /* Cycle counter overflow on increment that changes PMCCNTR[63]
>> + * or PMCCNTR[31] from 1 to 0 according to the value of
>> + * ARMV8_PMCR_LC
>> + */
>> + if (!vcpu_mode_is_32bit(vcpu))
>> + pmcr_lc = vcpu_sys_reg(vcpu, PMCR_EL0) & ARMV8_PMCR_LC;
>> + else
>> + pmcr_lc = vcpu_cp15(vcpu, c9_PMCR) & ARMV8_PMCR_LC;
>> +
>> + overflow_bit = pmcr_lc ? 63 : 31;
>> + }
>> + counter = kvm_pmu_get_counter_value(vcpu, select_idx);
>> + /* The initial sample period (overflow count) of an event. */
>> + attr.sample_period = (-counter) & (((u64)1 << overflow_bit) - 1);
>> +
>> + event = perf_event_create_kernel_counter(&attr, -1, current, NULL, pmc);
>> + if (IS_ERR(event)) {
>> + printk_once("kvm: pmu event creation failed %ld\n",
>> + PTR_ERR(event));
>> + return;
>> + }
>> + pmc->perf_event = event;
>> +}
>>
>
> Having had a chat with Will, it appears that a much better solution
> would be to ask perf to use raw events instead of trying to map things
> to perf events (which the guest has already done).
>
> See drivers/oprofile/oprofile_perf.c::op_perf_setup().
>
> Thoughts?
>
Yeah, directly using PERF_TYPE_RAW event looks simpler. But should we
check whether the value of event number written to PMXEVTYPER_EL0 is
valid? Or That is guaranteed by guest?
--
Shannon
^ permalink raw reply [flat|nested] 113+ messages in thread
* [PATCH v2 08/22] KVM: ARM64: PMU: Add perf event map and introduce perf event creating function
2015-09-11 13:35 ` Shannon Zhao
@ 2015-09-11 14:14 ` Marc Zyngier
-1 siblings, 0 replies; 113+ messages in thread
From: Marc Zyngier @ 2015-09-11 14:14 UTC (permalink / raw)
To: linux-arm-kernel
On 11/09/15 14:35, Shannon Zhao wrote:
>
>
> On 2015/9/11 19:04, Marc Zyngier wrote:
[...]
>> Having had a chat with Will, it appears that a much better solution
>> would be to ask perf to use raw events instead of trying to map things
>> to perf events (which the guest has already done).
>>
>> See drivers/oprofile/oprofile_perf.c::op_perf_setup().
>>
>> Thoughts?
>>
>
> Yeah, directly using PERF_TYPE_RAW event looks simpler. But should we
> check whether the value of event number written to PMXEVTYPER_EL0 is
> valid? Or That is guaranteed by guest?
I don't think we need to check for anything. Userspace is allowed to
request any event and the PMU will count it if this event exists, or
won't if it doesn't.
We shouldn't be doing more validation than that, if only because most
events are microarchitectural, and they are mostly undocumented.
Thanks,
M.
--
Jazz is not dead. It just smells funny...
^ permalink raw reply [flat|nested] 113+ messages in thread
* Re: [PATCH v2 08/22] KVM: ARM64: PMU: Add perf event map and introduce perf event creating function
@ 2015-09-11 14:14 ` Marc Zyngier
0 siblings, 0 replies; 113+ messages in thread
From: Marc Zyngier @ 2015-09-11 14:14 UTC (permalink / raw)
To: Shannon Zhao, Shannon Zhao, kvmarm
Cc: linux-arm-kernel, kvm, christoffer.dall, will.deacon, alex.bennee,
wei, peter.huangpeng
On 11/09/15 14:35, Shannon Zhao wrote:
>
>
> On 2015/9/11 19:04, Marc Zyngier wrote:
[...]
>> Having had a chat with Will, it appears that a much better solution
>> would be to ask perf to use raw events instead of trying to map things
>> to perf events (which the guest has already done).
>>
>> See drivers/oprofile/oprofile_perf.c::op_perf_setup().
>>
>> Thoughts?
>>
>
> Yeah, directly using PERF_TYPE_RAW event looks simpler. But should we
> check whether the value of event number written to PMXEVTYPER_EL0 is
> valid? Or That is guaranteed by guest?
I don't think we need to check for anything. Userspace is allowed to
request any event and the PMU will count it if this event exists, or
won't if it doesn't.
We shouldn't be doing more validation than that, if only because most
events are microarchitectural, and they are mostly undocumented.
Thanks,
M.
--
Jazz is not dead. It just smells funny...
^ permalink raw reply [flat|nested] 113+ messages in thread
* [PATCH v2 04/22] KVM: ARM64: Add reset and access handlers for PMCR_EL0 register
2015-09-11 10:07 ` Marc Zyngier
(?)
@ 2015-09-14 3:14 ` Shannon Zhao
-1 siblings, 0 replies; 113+ messages in thread
From: Shannon Zhao @ 2015-09-14 3:14 UTC (permalink / raw)
To: linux-arm-kernel
On 2015/9/11 18:07, Marc Zyngier wrote:
> On 11/09/15 09:54, Shannon Zhao wrote:
>> > From: Shannon Zhao <shannon.zhao@linaro.org>
>> >
>> > Add reset handler which gets host value of PMCR_EL0 and make writable
>> > bits architecturally UNKNOWN. Add a common access handler for PMU
>> > registers which emulates writing and reading register and add emulation
>> > for PMCR.
>> >
>> > Signed-off-by: Shannon Zhao <shannon.zhao@linaro.org>
>> > ---
>> > arch/arm64/kvm/sys_regs.c | 76 +++++++++++++++++++++++++++++++++++++++++++++--
>> > 1 file changed, 74 insertions(+), 2 deletions(-)
>> >
>> > diff --git a/arch/arm64/kvm/sys_regs.c b/arch/arm64/kvm/sys_regs.c
>> > index c370b40..db1be44 100644
>> > --- a/arch/arm64/kvm/sys_regs.c
>> > +++ b/arch/arm64/kvm/sys_regs.c
>> > @@ -33,6 +33,7 @@
>> > #include <asm/kvm_emulate.h>
>> > #include <asm/kvm_host.h>
>> > #include <asm/kvm_mmu.h>
>> > +#include <asm/pmu.h>
>> >
>> > #include <trace/events/kvm.h>
>> >
>> > @@ -236,6 +237,48 @@ static void reset_mpidr(struct kvm_vcpu *vcpu, const struct sys_reg_desc *r)
>> > vcpu_sys_reg(vcpu, MPIDR_EL1) = (1ULL << 31) | mpidr;
>> > }
>> >
>> > +static void reset_pmcr(struct kvm_vcpu *vcpu, const struct sys_reg_desc *r)
>> > +{
>> > + u32 pmcr;
>> > +
>> > + asm volatile("mrs %0, pmcr_el0\n" : "=r" (pmcr));
>> > + /* Writable bits of PMCR_EL0 (ARMV8_PMCR_MASK) is reset to UNKNOWN*/
>> > + if (!vcpu_mode_is_32bit(vcpu))
>> > + vcpu_sys_reg(vcpu, r->reg) = (pmcr & ~ARMV8_PMCR_MASK)
>> > + | (ARMV8_PMCR_MASK & 0xdecafbad);
>> > + else
>> > + vcpu_cp15(vcpu, r->reg) = (pmcr & ~ARMV8_PMCR_MASK)
>> > + | (ARMV8_PMCR_MASK & 0xdecafbad);
> I have some concerns about blindly reusing the top bits of the host's
> PMCR_EL0 register, specially when it comes to the PMCR_EL0.N. Given that
> we're fully emulating the PMU, shouldn't we simply define how many
> counters we're emulating?
>
But how many counters should we define? And what does this definition
based on? The only gist I think is the number of counters on host. And
what's the reason to define less or more than PMCR_EL0.N? I didn't find
one. So I choose to be consistent with host.
Thanks,
--
Shannon
^ permalink raw reply [flat|nested] 113+ messages in thread
* Re: [PATCH v2 04/22] KVM: ARM64: Add reset and access handlers for PMCR_EL0 register
@ 2015-09-14 3:14 ` Shannon Zhao
0 siblings, 0 replies; 113+ messages in thread
From: Shannon Zhao @ 2015-09-14 3:14 UTC (permalink / raw)
To: Marc Zyngier, kvmarm; +Cc: kvm, will.deacon, linux-arm-kernel, shannon.zhao
On 2015/9/11 18:07, Marc Zyngier wrote:
> On 11/09/15 09:54, Shannon Zhao wrote:
>> > From: Shannon Zhao <shannon.zhao@linaro.org>
>> >
>> > Add reset handler which gets host value of PMCR_EL0 and make writable
>> > bits architecturally UNKNOWN. Add a common access handler for PMU
>> > registers which emulates writing and reading register and add emulation
>> > for PMCR.
>> >
>> > Signed-off-by: Shannon Zhao <shannon.zhao@linaro.org>
>> > ---
>> > arch/arm64/kvm/sys_regs.c | 76 +++++++++++++++++++++++++++++++++++++++++++++--
>> > 1 file changed, 74 insertions(+), 2 deletions(-)
>> >
>> > diff --git a/arch/arm64/kvm/sys_regs.c b/arch/arm64/kvm/sys_regs.c
>> > index c370b40..db1be44 100644
>> > --- a/arch/arm64/kvm/sys_regs.c
>> > +++ b/arch/arm64/kvm/sys_regs.c
>> > @@ -33,6 +33,7 @@
>> > #include <asm/kvm_emulate.h>
>> > #include <asm/kvm_host.h>
>> > #include <asm/kvm_mmu.h>
>> > +#include <asm/pmu.h>
>> >
>> > #include <trace/events/kvm.h>
>> >
>> > @@ -236,6 +237,48 @@ static void reset_mpidr(struct kvm_vcpu *vcpu, const struct sys_reg_desc *r)
>> > vcpu_sys_reg(vcpu, MPIDR_EL1) = (1ULL << 31) | mpidr;
>> > }
>> >
>> > +static void reset_pmcr(struct kvm_vcpu *vcpu, const struct sys_reg_desc *r)
>> > +{
>> > + u32 pmcr;
>> > +
>> > + asm volatile("mrs %0, pmcr_el0\n" : "=r" (pmcr));
>> > + /* Writable bits of PMCR_EL0 (ARMV8_PMCR_MASK) is reset to UNKNOWN*/
>> > + if (!vcpu_mode_is_32bit(vcpu))
>> > + vcpu_sys_reg(vcpu, r->reg) = (pmcr & ~ARMV8_PMCR_MASK)
>> > + | (ARMV8_PMCR_MASK & 0xdecafbad);
>> > + else
>> > + vcpu_cp15(vcpu, r->reg) = (pmcr & ~ARMV8_PMCR_MASK)
>> > + | (ARMV8_PMCR_MASK & 0xdecafbad);
> I have some concerns about blindly reusing the top bits of the host's
> PMCR_EL0 register, specially when it comes to the PMCR_EL0.N. Given that
> we're fully emulating the PMU, shouldn't we simply define how many
> counters we're emulating?
>
But how many counters should we define? And what does this definition
based on? The only gist I think is the number of counters on host. And
what's the reason to define less or more than PMCR_EL0.N? I didn't find
one. So I choose to be consistent with host.
Thanks,
--
Shannon
^ permalink raw reply [flat|nested] 113+ messages in thread
* Re: [PATCH v2 04/22] KVM: ARM64: Add reset and access handlers for PMCR_EL0 register
@ 2015-09-14 3:14 ` Shannon Zhao
0 siblings, 0 replies; 113+ messages in thread
From: Shannon Zhao @ 2015-09-14 3:14 UTC (permalink / raw)
To: Marc Zyngier, kvmarm; +Cc: kvm, will.deacon, linux-arm-kernel, shannon.zhao
On 2015/9/11 18:07, Marc Zyngier wrote:
> On 11/09/15 09:54, Shannon Zhao wrote:
>> > From: Shannon Zhao <shannon.zhao@linaro.org>
>> >
>> > Add reset handler which gets host value of PMCR_EL0 and make writable
>> > bits architecturally UNKNOWN. Add a common access handler for PMU
>> > registers which emulates writing and reading register and add emulation
>> > for PMCR.
>> >
>> > Signed-off-by: Shannon Zhao <shannon.zhao@linaro.org>
>> > ---
>> > arch/arm64/kvm/sys_regs.c | 76 +++++++++++++++++++++++++++++++++++++++++++++--
>> > 1 file changed, 74 insertions(+), 2 deletions(-)
>> >
>> > diff --git a/arch/arm64/kvm/sys_regs.c b/arch/arm64/kvm/sys_regs.c
>> > index c370b40..db1be44 100644
>> > --- a/arch/arm64/kvm/sys_regs.c
>> > +++ b/arch/arm64/kvm/sys_regs.c
>> > @@ -33,6 +33,7 @@
>> > #include <asm/kvm_emulate.h>
>> > #include <asm/kvm_host.h>
>> > #include <asm/kvm_mmu.h>
>> > +#include <asm/pmu.h>
>> >
>> > #include <trace/events/kvm.h>
>> >
>> > @@ -236,6 +237,48 @@ static void reset_mpidr(struct kvm_vcpu *vcpu, const struct sys_reg_desc *r)
>> > vcpu_sys_reg(vcpu, MPIDR_EL1) = (1ULL << 31) | mpidr;
>> > }
>> >
>> > +static void reset_pmcr(struct kvm_vcpu *vcpu, const struct sys_reg_desc *r)
>> > +{
>> > + u32 pmcr;
>> > +
>> > + asm volatile("mrs %0, pmcr_el0\n" : "=r" (pmcr));
>> > + /* Writable bits of PMCR_EL0 (ARMV8_PMCR_MASK) is reset to UNKNOWN*/
>> > + if (!vcpu_mode_is_32bit(vcpu))
>> > + vcpu_sys_reg(vcpu, r->reg) = (pmcr & ~ARMV8_PMCR_MASK)
>> > + | (ARMV8_PMCR_MASK & 0xdecafbad);
>> > + else
>> > + vcpu_cp15(vcpu, r->reg) = (pmcr & ~ARMV8_PMCR_MASK)
>> > + | (ARMV8_PMCR_MASK & 0xdecafbad);
> I have some concerns about blindly reusing the top bits of the host's
> PMCR_EL0 register, specially when it comes to the PMCR_EL0.N. Given that
> we're fully emulating the PMU, shouldn't we simply define how many
> counters we're emulating?
>
But how many counters should we define? And what does this definition
based on? The only gist I think is the number of counters on host. And
what's the reason to define less or more than PMCR_EL0.N? I didn't find
one. So I choose to be consistent with host.
Thanks,
--
Shannon
^ permalink raw reply [flat|nested] 113+ messages in thread
* [PATCH v2 00/22] KVM: ARM64: Add guest PMU support
2015-09-11 8:54 ` Shannon Zhao
@ 2015-09-14 11:53 ` Christoffer Dall
-1 siblings, 0 replies; 113+ messages in thread
From: Christoffer Dall @ 2015-09-14 11:53 UTC (permalink / raw)
To: linux-arm-kernel
Hi Shannon,
On Fri, Sep 11, 2015 at 04:54:53PM +0800, Shannon Zhao wrote:
> From: Shannon Zhao <shannon.zhao@linaro.org>
>
> This patchset adds guest PMU support for KVM on ARM64. It takes
> trap-and-emulate approach. When guest wants to monitor one event, it
> will be trapped by KVM and KVM will call perf_event API to create a perf
> event and call relevant perf_event APIs to get the count value of event.
>
> Use perf to test this patchset in guest. When using "perf list", it
> shows the list of the hardware events and hardware cache events perf
> supports. Then use "perf stat -e EVENT" to monitor some event. For
> example, use "perf stat -e cycles" to count cpu cycles and
> "perf stat -e cache-misses" to count cache misses.
>
> Below are the outputs of "perf stat -r 5 sleep 5" when running in host
> and guest.
>
> Host:
> Performance counter stats for 'sleep 5' (5 runs):
>
> 0.551428 task-clock (msec) # 0.000 CPUs utilized ( +- 0.91% )
> 1 context-switches # 0.002 M/sec
> 0 cpu-migrations # 0.000 K/sec
> 48 page-faults # 0.088 M/sec ( +- 1.05% )
> 1150265 cycles # 2.086 GHz ( +- 0.92% )
> <not supported> stalled-cycles-frontend
> <not supported> stalled-cycles-backend
> 526398 instructions # 0.46 insns per cycle ( +- 0.89% )
> <not supported> branches
> 9485 branch-misses # 17.201 M/sec ( +- 2.35% )
>
> 5.000831616 seconds time elapsed ( +- 0.00% )
>
> Guest:
> Performance counter stats for 'sleep 5' (5 runs):
>
> 0.730868 task-clock (msec) # 0.000 CPUs utilized ( +- 1.13% )
> 1 context-switches # 0.001 M/sec
> 0 cpu-migrations # 0.000 K/sec
> 48 page-faults # 0.065 M/sec ( +- 0.42% )
> 1642982 cycles # 2.248 GHz ( +- 1.04% )
> <not supported> stalled-cycles-frontend
> <not supported> stalled-cycles-backend
> 637964 instructions # 0.39 insns per cycle ( +- 0.65% )
> <not supported> branches
> 10377 branch-misses # 14.198 M/sec ( +- 1.09% )
>
> 5.001289068 seconds time elapsed ( +- 0.00% )
>
> This patchset can be fetched from [1] and the relevant QEMU version for
> test can be fetched from [2].
>
Is there a changelog for this patch set somewhere?
If not, can you include this in v3 addressing Marc's comments?
Thanks,
-Christoffer
^ permalink raw reply [flat|nested] 113+ messages in thread
* Re: [PATCH v2 00/22] KVM: ARM64: Add guest PMU support
@ 2015-09-14 11:53 ` Christoffer Dall
0 siblings, 0 replies; 113+ messages in thread
From: Christoffer Dall @ 2015-09-14 11:53 UTC (permalink / raw)
To: Shannon Zhao
Cc: kvm, marc.zyngier, will.deacon, shannon.zhao, kvmarm,
linux-arm-kernel
Hi Shannon,
On Fri, Sep 11, 2015 at 04:54:53PM +0800, Shannon Zhao wrote:
> From: Shannon Zhao <shannon.zhao@linaro.org>
>
> This patchset adds guest PMU support for KVM on ARM64. It takes
> trap-and-emulate approach. When guest wants to monitor one event, it
> will be trapped by KVM and KVM will call perf_event API to create a perf
> event and call relevant perf_event APIs to get the count value of event.
>
> Use perf to test this patchset in guest. When using "perf list", it
> shows the list of the hardware events and hardware cache events perf
> supports. Then use "perf stat -e EVENT" to monitor some event. For
> example, use "perf stat -e cycles" to count cpu cycles and
> "perf stat -e cache-misses" to count cache misses.
>
> Below are the outputs of "perf stat -r 5 sleep 5" when running in host
> and guest.
>
> Host:
> Performance counter stats for 'sleep 5' (5 runs):
>
> 0.551428 task-clock (msec) # 0.000 CPUs utilized ( +- 0.91% )
> 1 context-switches # 0.002 M/sec
> 0 cpu-migrations # 0.000 K/sec
> 48 page-faults # 0.088 M/sec ( +- 1.05% )
> 1150265 cycles # 2.086 GHz ( +- 0.92% )
> <not supported> stalled-cycles-frontend
> <not supported> stalled-cycles-backend
> 526398 instructions # 0.46 insns per cycle ( +- 0.89% )
> <not supported> branches
> 9485 branch-misses # 17.201 M/sec ( +- 2.35% )
>
> 5.000831616 seconds time elapsed ( +- 0.00% )
>
> Guest:
> Performance counter stats for 'sleep 5' (5 runs):
>
> 0.730868 task-clock (msec) # 0.000 CPUs utilized ( +- 1.13% )
> 1 context-switches # 0.001 M/sec
> 0 cpu-migrations # 0.000 K/sec
> 48 page-faults # 0.065 M/sec ( +- 0.42% )
> 1642982 cycles # 2.248 GHz ( +- 1.04% )
> <not supported> stalled-cycles-frontend
> <not supported> stalled-cycles-backend
> 637964 instructions # 0.39 insns per cycle ( +- 0.65% )
> <not supported> branches
> 10377 branch-misses # 14.198 M/sec ( +- 1.09% )
>
> 5.001289068 seconds time elapsed ( +- 0.00% )
>
> This patchset can be fetched from [1] and the relevant QEMU version for
> test can be fetched from [2].
>
Is there a changelog for this patch set somewhere?
If not, can you include this in v3 addressing Marc's comments?
Thanks,
-Christoffer
^ permalink raw reply [flat|nested] 113+ messages in thread
* [PATCH v2 04/22] KVM: ARM64: Add reset and access handlers for PMCR_EL0 register
2015-09-14 3:14 ` Shannon Zhao
@ 2015-09-14 12:11 ` Marc Zyngier
-1 siblings, 0 replies; 113+ messages in thread
From: Marc Zyngier @ 2015-09-14 12:11 UTC (permalink / raw)
To: linux-arm-kernel
On 14/09/15 04:14, Shannon Zhao wrote:
>
>
> On 2015/9/11 18:07, Marc Zyngier wrote:
>> On 11/09/15 09:54, Shannon Zhao wrote:
>>>> From: Shannon Zhao <shannon.zhao@linaro.org>
>>>>
>>>> Add reset handler which gets host value of PMCR_EL0 and make writable
>>>> bits architecturally UNKNOWN. Add a common access handler for PMU
>>>> registers which emulates writing and reading register and add emulation
>>>> for PMCR.
>>>>
>>>> Signed-off-by: Shannon Zhao <shannon.zhao@linaro.org>
>>>> ---
>>>> arch/arm64/kvm/sys_regs.c | 76 +++++++++++++++++++++++++++++++++++++++++++++--
>>>> 1 file changed, 74 insertions(+), 2 deletions(-)
>>>>
>>>> diff --git a/arch/arm64/kvm/sys_regs.c b/arch/arm64/kvm/sys_regs.c
>>>> index c370b40..db1be44 100644
>>>> --- a/arch/arm64/kvm/sys_regs.c
>>>> +++ b/arch/arm64/kvm/sys_regs.c
>>>> @@ -33,6 +33,7 @@
>>>> #include <asm/kvm_emulate.h>
>>>> #include <asm/kvm_host.h>
>>>> #include <asm/kvm_mmu.h>
>>>> +#include <asm/pmu.h>
>>>>
>>>> #include <trace/events/kvm.h>
>>>>
>>>> @@ -236,6 +237,48 @@ static void reset_mpidr(struct kvm_vcpu *vcpu, const struct sys_reg_desc *r)
>>>> vcpu_sys_reg(vcpu, MPIDR_EL1) = (1ULL << 31) | mpidr;
>>>> }
>>>>
>>>> +static void reset_pmcr(struct kvm_vcpu *vcpu, const struct sys_reg_desc *r)
>>>> +{
>>>> + u32 pmcr;
>>>> +
>>>> + asm volatile("mrs %0, pmcr_el0\n" : "=r" (pmcr));
>>>> + /* Writable bits of PMCR_EL0 (ARMV8_PMCR_MASK) is reset to UNKNOWN*/
>>>> + if (!vcpu_mode_is_32bit(vcpu))
>>>> + vcpu_sys_reg(vcpu, r->reg) = (pmcr & ~ARMV8_PMCR_MASK)
>>>> + | (ARMV8_PMCR_MASK & 0xdecafbad);
>>>> + else
>>>> + vcpu_cp15(vcpu, r->reg) = (pmcr & ~ARMV8_PMCR_MASK)
>>>> + | (ARMV8_PMCR_MASK & 0xdecafbad);
>> I have some concerns about blindly reusing the top bits of the host's
>> PMCR_EL0 register, specially when it comes to the PMCR_EL0.N. Given that
>> we're fully emulating the PMU, shouldn't we simply define how many
>> counters we're emulating?
>>
>
> But how many counters should we define? And what does this definition
> based on? The only gist I think is the number of counters on host. And
> what's the reason to define less or more than PMCR_EL0.N? I didn't find
> one. So I choose to be consistent with host.
The problem is that choosing the host value may be just the wrong thing
once you migrate it. It is not a big deal, but it is worth keeping it in
mind. it should anyway be possible to size the PMU from userspace,
overriding the value you've selected at reset.
Thanks,
M.
--
Jazz is not dead. It just smells funny...
^ permalink raw reply [flat|nested] 113+ messages in thread
* Re: [PATCH v2 04/22] KVM: ARM64: Add reset and access handlers for PMCR_EL0 register
@ 2015-09-14 12:11 ` Marc Zyngier
0 siblings, 0 replies; 113+ messages in thread
From: Marc Zyngier @ 2015-09-14 12:11 UTC (permalink / raw)
To: Shannon Zhao, kvmarm
Cc: linux-arm-kernel, kvm, christoffer.dall, will.deacon, alex.bennee,
wei, shannon.zhao, peter.huangpeng
On 14/09/15 04:14, Shannon Zhao wrote:
>
>
> On 2015/9/11 18:07, Marc Zyngier wrote:
>> On 11/09/15 09:54, Shannon Zhao wrote:
>>>> From: Shannon Zhao <shannon.zhao@linaro.org>
>>>>
>>>> Add reset handler which gets host value of PMCR_EL0 and make writable
>>>> bits architecturally UNKNOWN. Add a common access handler for PMU
>>>> registers which emulates writing and reading register and add emulation
>>>> for PMCR.
>>>>
>>>> Signed-off-by: Shannon Zhao <shannon.zhao@linaro.org>
>>>> ---
>>>> arch/arm64/kvm/sys_regs.c | 76 +++++++++++++++++++++++++++++++++++++++++++++--
>>>> 1 file changed, 74 insertions(+), 2 deletions(-)
>>>>
>>>> diff --git a/arch/arm64/kvm/sys_regs.c b/arch/arm64/kvm/sys_regs.c
>>>> index c370b40..db1be44 100644
>>>> --- a/arch/arm64/kvm/sys_regs.c
>>>> +++ b/arch/arm64/kvm/sys_regs.c
>>>> @@ -33,6 +33,7 @@
>>>> #include <asm/kvm_emulate.h>
>>>> #include <asm/kvm_host.h>
>>>> #include <asm/kvm_mmu.h>
>>>> +#include <asm/pmu.h>
>>>>
>>>> #include <trace/events/kvm.h>
>>>>
>>>> @@ -236,6 +237,48 @@ static void reset_mpidr(struct kvm_vcpu *vcpu, const struct sys_reg_desc *r)
>>>> vcpu_sys_reg(vcpu, MPIDR_EL1) = (1ULL << 31) | mpidr;
>>>> }
>>>>
>>>> +static void reset_pmcr(struct kvm_vcpu *vcpu, const struct sys_reg_desc *r)
>>>> +{
>>>> + u32 pmcr;
>>>> +
>>>> + asm volatile("mrs %0, pmcr_el0\n" : "=r" (pmcr));
>>>> + /* Writable bits of PMCR_EL0 (ARMV8_PMCR_MASK) is reset to UNKNOWN*/
>>>> + if (!vcpu_mode_is_32bit(vcpu))
>>>> + vcpu_sys_reg(vcpu, r->reg) = (pmcr & ~ARMV8_PMCR_MASK)
>>>> + | (ARMV8_PMCR_MASK & 0xdecafbad);
>>>> + else
>>>> + vcpu_cp15(vcpu, r->reg) = (pmcr & ~ARMV8_PMCR_MASK)
>>>> + | (ARMV8_PMCR_MASK & 0xdecafbad);
>> I have some concerns about blindly reusing the top bits of the host's
>> PMCR_EL0 register, specially when it comes to the PMCR_EL0.N. Given that
>> we're fully emulating the PMU, shouldn't we simply define how many
>> counters we're emulating?
>>
>
> But how many counters should we define? And what does this definition
> based on? The only gist I think is the number of counters on host. And
> what's the reason to define less or more than PMCR_EL0.N? I didn't find
> one. So I choose to be consistent with host.
The problem is that choosing the host value may be just the wrong thing
once you migrate it. It is not a big deal, but it is worth keeping it in
mind. it should anyway be possible to size the PMU from userspace,
overriding the value you've selected at reset.
Thanks,
M.
--
Jazz is not dead. It just smells funny...
^ permalink raw reply [flat|nested] 113+ messages in thread
* [PATCH v2 00/22] KVM: ARM64: Add guest PMU support
2015-09-14 11:53 ` Christoffer Dall
@ 2015-09-14 12:58 ` Shannon Zhao
-1 siblings, 0 replies; 113+ messages in thread
From: Shannon Zhao @ 2015-09-14 12:58 UTC (permalink / raw)
To: linux-arm-kernel
On 2015/9/14 19:53, Christoffer Dall wrote:
> Hi Shannon,
>
> On Fri, Sep 11, 2015 at 04:54:53PM +0800, Shannon Zhao wrote:
>> From: Shannon Zhao <shannon.zhao@linaro.org>
>>
>> This patchset adds guest PMU support for KVM on ARM64. It takes
>> trap-and-emulate approach. When guest wants to monitor one event, it
>> will be trapped by KVM and KVM will call perf_event API to create a perf
>> event and call relevant perf_event APIs to get the count value of event.
>>
>> Use perf to test this patchset in guest. When using "perf list", it
>> shows the list of the hardware events and hardware cache events perf
>> supports. Then use "perf stat -e EVENT" to monitor some event. For
>> example, use "perf stat -e cycles" to count cpu cycles and
>> "perf stat -e cache-misses" to count cache misses.
>>
>> Below are the outputs of "perf stat -r 5 sleep 5" when running in host
>> and guest.
>>
>> Host:
>> Performance counter stats for 'sleep 5' (5 runs):
>>
>> 0.551428 task-clock (msec) # 0.000 CPUs utilized ( +- 0.91% )
>> 1 context-switches # 0.002 M/sec
>> 0 cpu-migrations # 0.000 K/sec
>> 48 page-faults # 0.088 M/sec ( +- 1.05% )
>> 1150265 cycles # 2.086 GHz ( +- 0.92% )
>> <not supported> stalled-cycles-frontend
>> <not supported> stalled-cycles-backend
>> 526398 instructions # 0.46 insns per cycle ( +- 0.89% )
>> <not supported> branches
>> 9485 branch-misses # 17.201 M/sec ( +- 2.35% )
>>
>> 5.000831616 seconds time elapsed ( +- 0.00% )
>>
>> Guest:
>> Performance counter stats for 'sleep 5' (5 runs):
>>
>> 0.730868 task-clock (msec) # 0.000 CPUs utilized ( +- 1.13% )
>> 1 context-switches # 0.001 M/sec
>> 0 cpu-migrations # 0.000 K/sec
>> 48 page-faults # 0.065 M/sec ( +- 0.42% )
>> 1642982 cycles # 2.248 GHz ( +- 1.04% )
>> <not supported> stalled-cycles-frontend
>> <not supported> stalled-cycles-backend
>> 637964 instructions # 0.39 insns per cycle ( +- 0.65% )
>> <not supported> branches
>> 10377 branch-misses # 14.198 M/sec ( +- 1.09% )
>>
>> 5.001289068 seconds time elapsed ( +- 0.00% )
>>
>> This patchset can be fetched from [1] and the relevant QEMU version for
>> test can be fetched from [2].
>>
> Is there a changelog for this patch set somewhere?
>
I forgot this, while I realized it after sending this patchset out.
> If not, can you include this in v3 addressing Marc's comments?
>
> Thanks,
> -Christoffer
>
--
Shannon
^ permalink raw reply [flat|nested] 113+ messages in thread
* Re: [PATCH v2 00/22] KVM: ARM64: Add guest PMU support
@ 2015-09-14 12:58 ` Shannon Zhao
0 siblings, 0 replies; 113+ messages in thread
From: Shannon Zhao @ 2015-09-14 12:58 UTC (permalink / raw)
To: Christoffer Dall, Shannon Zhao
Cc: kvm, marc.zyngier, will.deacon, kvmarm, linux-arm-kernel
On 2015/9/14 19:53, Christoffer Dall wrote:
> Hi Shannon,
>
> On Fri, Sep 11, 2015 at 04:54:53PM +0800, Shannon Zhao wrote:
>> From: Shannon Zhao <shannon.zhao@linaro.org>
>>
>> This patchset adds guest PMU support for KVM on ARM64. It takes
>> trap-and-emulate approach. When guest wants to monitor one event, it
>> will be trapped by KVM and KVM will call perf_event API to create a perf
>> event and call relevant perf_event APIs to get the count value of event.
>>
>> Use perf to test this patchset in guest. When using "perf list", it
>> shows the list of the hardware events and hardware cache events perf
>> supports. Then use "perf stat -e EVENT" to monitor some event. For
>> example, use "perf stat -e cycles" to count cpu cycles and
>> "perf stat -e cache-misses" to count cache misses.
>>
>> Below are the outputs of "perf stat -r 5 sleep 5" when running in host
>> and guest.
>>
>> Host:
>> Performance counter stats for 'sleep 5' (5 runs):
>>
>> 0.551428 task-clock (msec) # 0.000 CPUs utilized ( +- 0.91% )
>> 1 context-switches # 0.002 M/sec
>> 0 cpu-migrations # 0.000 K/sec
>> 48 page-faults # 0.088 M/sec ( +- 1.05% )
>> 1150265 cycles # 2.086 GHz ( +- 0.92% )
>> <not supported> stalled-cycles-frontend
>> <not supported> stalled-cycles-backend
>> 526398 instructions # 0.46 insns per cycle ( +- 0.89% )
>> <not supported> branches
>> 9485 branch-misses # 17.201 M/sec ( +- 2.35% )
>>
>> 5.000831616 seconds time elapsed ( +- 0.00% )
>>
>> Guest:
>> Performance counter stats for 'sleep 5' (5 runs):
>>
>> 0.730868 task-clock (msec) # 0.000 CPUs utilized ( +- 1.13% )
>> 1 context-switches # 0.001 M/sec
>> 0 cpu-migrations # 0.000 K/sec
>> 48 page-faults # 0.065 M/sec ( +- 0.42% )
>> 1642982 cycles # 2.248 GHz ( +- 1.04% )
>> <not supported> stalled-cycles-frontend
>> <not supported> stalled-cycles-backend
>> 637964 instructions # 0.39 insns per cycle ( +- 0.65% )
>> <not supported> branches
>> 10377 branch-misses # 14.198 M/sec ( +- 1.09% )
>>
>> 5.001289068 seconds time elapsed ( +- 0.00% )
>>
>> This patchset can be fetched from [1] and the relevant QEMU version for
>> test can be fetched from [2].
>>
> Is there a changelog for this patch set somewhere?
>
I forgot this, while I realized it after sending this patchset out.
> If not, can you include this in v3 addressing Marc's comments?
>
> Thanks,
> -Christoffer
>
--
Shannon
^ permalink raw reply [flat|nested] 113+ messages in thread
* [PATCH v2 00/22] KVM: ARM64: Add guest PMU support
2015-09-11 8:54 ` Shannon Zhao
@ 2015-09-14 13:24 ` Shannon Zhao
-1 siblings, 0 replies; 113+ messages in thread
From: Shannon Zhao @ 2015-09-14 13:24 UTC (permalink / raw)
To: linux-arm-kernel
Sorry, I forgot the changlog. Below are the main changes of this v2
patchset.
Changes since v1->v2:
* Use switch...case for registers access handler instead of adding
alone handler for each register
* Try to use the sys_regs to store the register value instead of adding
new variables in struct kvm_pmc
* Fix the handle of cp15 regs
* Create a new kvm device vPMU, then userspace could choose whether to
create PMU
* Fix the handle of PMU overflow interrupt
On 2015/9/11 16:54, Shannon Zhao wrote:
> From: Shannon Zhao <shannon.zhao@linaro.org>
>
> This patchset adds guest PMU support for KVM on ARM64. It takes
> trap-and-emulate approach. When guest wants to monitor one event, it
> will be trapped by KVM and KVM will call perf_event API to create a perf
> event and call relevant perf_event APIs to get the count value of event.
>
> Use perf to test this patchset in guest. When using "perf list", it
> shows the list of the hardware events and hardware cache events perf
> supports. Then use "perf stat -e EVENT" to monitor some event. For
> example, use "perf stat -e cycles" to count cpu cycles and
> "perf stat -e cache-misses" to count cache misses.
>
> Below are the outputs of "perf stat -r 5 sleep 5" when running in host
> and guest.
>
> Host:
> Performance counter stats for 'sleep 5' (5 runs):
>
> 0.551428 task-clock (msec) # 0.000 CPUs utilized ( +- 0.91% )
> 1 context-switches # 0.002 M/sec
> 0 cpu-migrations # 0.000 K/sec
> 48 page-faults # 0.088 M/sec ( +- 1.05% )
> 1150265 cycles # 2.086 GHz ( +- 0.92% )
> <not supported> stalled-cycles-frontend
> <not supported> stalled-cycles-backend
> 526398 instructions # 0.46 insns per cycle ( +- 0.89% )
> <not supported> branches
> 9485 branch-misses # 17.201 M/sec ( +- 2.35% )
>
> 5.000831616 seconds time elapsed ( +- 0.00% )
>
> Guest:
> Performance counter stats for 'sleep 5' (5 runs):
>
> 0.730868 task-clock (msec) # 0.000 CPUs utilized ( +- 1.13% )
> 1 context-switches # 0.001 M/sec
> 0 cpu-migrations # 0.000 K/sec
> 48 page-faults # 0.065 M/sec ( +- 0.42% )
> 1642982 cycles # 2.248 GHz ( +- 1.04% )
> <not supported> stalled-cycles-frontend
> <not supported> stalled-cycles-backend
> 637964 instructions # 0.39 insns per cycle ( +- 0.65% )
> <not supported> branches
> 10377 branch-misses # 14.198 M/sec ( +- 1.09% )
>
> 5.001289068 seconds time elapsed ( +- 0.00% )
>
> This patchset can be fetched from [1] and the relevant QEMU version for
> test can be fetched from [2].
>
> Thanks,
> Shannon
>
> [1] https://git.linaro.org/people/shannon.zhao/linux-mainline.git KVM_ARM64_PMU_v2
> [2] https://git.linaro.org/people/shannon.zhao/qemu.git PMU_v2
>
> Shannon Zhao (22):
> ARM64: Move PMU register related defines to asm/pmu.h
> KVM: ARM64: Define PMU data structure for each vcpu
> KVM: ARM64: Add offset defines for PMU registers
> KVM: ARM64: Add reset and access handlers for PMCR_EL0 register
> KVM: ARM64: Add a helper for CP15 registers reset to UNKNOWN
> KVM: ARM64: Add reset and access handlers for PMSELR register
> KVM: ARM64: Add reset and access handlers for PMCEID0 and PMCEID1
> register
> KVM: ARM64: PMU: Add perf event map and introduce perf event creating
> function
> KVM: ARM64: Add reset and access handlers for PMXEVTYPER register
> KVM: ARM64: Add reset and access handlers for PMXEVCNTR register
> KVM: ARM64: Add reset and access handlers for PMCCNTR register
> KVM: ARM64: Add reset and access handlers for PMCNTENSET and
> PMCNTENCLR register
> KVM: ARM64: Add reset and access handlers for PMINTENSET and
> PMINTENCLR register
> KVM: ARM64: Add reset and access handlers for PMOVSSET and PMOVSCLR
> register
> KVM: ARM64: Add a helper for CP15 registers reset to specified value
> KVM: ARM64: Add reset and access handlers for PMUSERENR register
> KVM: ARM64: Add reset and access handlers for PMSWINC register
> KVM: ARM64: Add access handlers for PMEVCNTRn and PMEVTYPERn register
> KVM: ARM64: Add PMU overflow interrupt routing
> KVM: ARM64: Reset PMU state when resetting vcpu
> KVM: ARM64: Free perf event of PMU when destroying vcpu
> KVM: ARM64: Add a new kvm ARM PMU device
>
> Documentation/virtual/kvm/devices/arm-pmu.txt | 15 +
> arch/arm/kvm/arm.c | 4 +
> arch/arm64/include/asm/kvm_asm.h | 59 ++-
> arch/arm64/include/asm/kvm_host.h | 2 +
> arch/arm64/include/asm/pmu.h | 49 +++
> arch/arm64/include/uapi/asm/kvm.h | 3 +
> arch/arm64/kernel/perf_event.c | 35 --
> arch/arm64/kvm/Kconfig | 8 +
> arch/arm64/kvm/Makefile | 1 +
> arch/arm64/kvm/reset.c | 3 +
> arch/arm64/kvm/sys_regs.c | 490 +++++++++++++++++++++++--
> arch/arm64/kvm/sys_regs.h | 16 +
> include/kvm/arm_pmu.h | 66 ++++
> include/linux/kvm_host.h | 1 +
> include/uapi/linux/kvm.h | 2 +
> virt/kvm/arm/pmu.c | 500 ++++++++++++++++++++++++++
> virt/kvm/kvm_main.c | 4 +
> 17 files changed, 1188 insertions(+), 70 deletions(-)
> create mode 100644 Documentation/virtual/kvm/devices/arm-pmu.txt
> create mode 100644 include/kvm/arm_pmu.h
> create mode 100644 virt/kvm/arm/pmu.c
>
--
Shannon
^ permalink raw reply [flat|nested] 113+ messages in thread
* Re: [PATCH v2 00/22] KVM: ARM64: Add guest PMU support
@ 2015-09-14 13:24 ` Shannon Zhao
0 siblings, 0 replies; 113+ messages in thread
From: Shannon Zhao @ 2015-09-14 13:24 UTC (permalink / raw)
To: Shannon Zhao, kvmarm
Cc: linux-arm-kernel, kvm, marc.zyngier, christoffer.dall,
will.deacon, alex.bennee, wei, peter.huangpeng
Sorry, I forgot the changlog. Below are the main changes of this v2
patchset.
Changes since v1->v2:
* Use switch...case for registers access handler instead of adding
alone handler for each register
* Try to use the sys_regs to store the register value instead of adding
new variables in struct kvm_pmc
* Fix the handle of cp15 regs
* Create a new kvm device vPMU, then userspace could choose whether to
create PMU
* Fix the handle of PMU overflow interrupt
On 2015/9/11 16:54, Shannon Zhao wrote:
> From: Shannon Zhao <shannon.zhao@linaro.org>
>
> This patchset adds guest PMU support for KVM on ARM64. It takes
> trap-and-emulate approach. When guest wants to monitor one event, it
> will be trapped by KVM and KVM will call perf_event API to create a perf
> event and call relevant perf_event APIs to get the count value of event.
>
> Use perf to test this patchset in guest. When using "perf list", it
> shows the list of the hardware events and hardware cache events perf
> supports. Then use "perf stat -e EVENT" to monitor some event. For
> example, use "perf stat -e cycles" to count cpu cycles and
> "perf stat -e cache-misses" to count cache misses.
>
> Below are the outputs of "perf stat -r 5 sleep 5" when running in host
> and guest.
>
> Host:
> Performance counter stats for 'sleep 5' (5 runs):
>
> 0.551428 task-clock (msec) # 0.000 CPUs utilized ( +- 0.91% )
> 1 context-switches # 0.002 M/sec
> 0 cpu-migrations # 0.000 K/sec
> 48 page-faults # 0.088 M/sec ( +- 1.05% )
> 1150265 cycles # 2.086 GHz ( +- 0.92% )
> <not supported> stalled-cycles-frontend
> <not supported> stalled-cycles-backend
> 526398 instructions # 0.46 insns per cycle ( +- 0.89% )
> <not supported> branches
> 9485 branch-misses # 17.201 M/sec ( +- 2.35% )
>
> 5.000831616 seconds time elapsed ( +- 0.00% )
>
> Guest:
> Performance counter stats for 'sleep 5' (5 runs):
>
> 0.730868 task-clock (msec) # 0.000 CPUs utilized ( +- 1.13% )
> 1 context-switches # 0.001 M/sec
> 0 cpu-migrations # 0.000 K/sec
> 48 page-faults # 0.065 M/sec ( +- 0.42% )
> 1642982 cycles # 2.248 GHz ( +- 1.04% )
> <not supported> stalled-cycles-frontend
> <not supported> stalled-cycles-backend
> 637964 instructions # 0.39 insns per cycle ( +- 0.65% )
> <not supported> branches
> 10377 branch-misses # 14.198 M/sec ( +- 1.09% )
>
> 5.001289068 seconds time elapsed ( +- 0.00% )
>
> This patchset can be fetched from [1] and the relevant QEMU version for
> test can be fetched from [2].
>
> Thanks,
> Shannon
>
> [1] https://git.linaro.org/people/shannon.zhao/linux-mainline.git KVM_ARM64_PMU_v2
> [2] https://git.linaro.org/people/shannon.zhao/qemu.git PMU_v2
>
> Shannon Zhao (22):
> ARM64: Move PMU register related defines to asm/pmu.h
> KVM: ARM64: Define PMU data structure for each vcpu
> KVM: ARM64: Add offset defines for PMU registers
> KVM: ARM64: Add reset and access handlers for PMCR_EL0 register
> KVM: ARM64: Add a helper for CP15 registers reset to UNKNOWN
> KVM: ARM64: Add reset and access handlers for PMSELR register
> KVM: ARM64: Add reset and access handlers for PMCEID0 and PMCEID1
> register
> KVM: ARM64: PMU: Add perf event map and introduce perf event creating
> function
> KVM: ARM64: Add reset and access handlers for PMXEVTYPER register
> KVM: ARM64: Add reset and access handlers for PMXEVCNTR register
> KVM: ARM64: Add reset and access handlers for PMCCNTR register
> KVM: ARM64: Add reset and access handlers for PMCNTENSET and
> PMCNTENCLR register
> KVM: ARM64: Add reset and access handlers for PMINTENSET and
> PMINTENCLR register
> KVM: ARM64: Add reset and access handlers for PMOVSSET and PMOVSCLR
> register
> KVM: ARM64: Add a helper for CP15 registers reset to specified value
> KVM: ARM64: Add reset and access handlers for PMUSERENR register
> KVM: ARM64: Add reset and access handlers for PMSWINC register
> KVM: ARM64: Add access handlers for PMEVCNTRn and PMEVTYPERn register
> KVM: ARM64: Add PMU overflow interrupt routing
> KVM: ARM64: Reset PMU state when resetting vcpu
> KVM: ARM64: Free perf event of PMU when destroying vcpu
> KVM: ARM64: Add a new kvm ARM PMU device
>
> Documentation/virtual/kvm/devices/arm-pmu.txt | 15 +
> arch/arm/kvm/arm.c | 4 +
> arch/arm64/include/asm/kvm_asm.h | 59 ++-
> arch/arm64/include/asm/kvm_host.h | 2 +
> arch/arm64/include/asm/pmu.h | 49 +++
> arch/arm64/include/uapi/asm/kvm.h | 3 +
> arch/arm64/kernel/perf_event.c | 35 --
> arch/arm64/kvm/Kconfig | 8 +
> arch/arm64/kvm/Makefile | 1 +
> arch/arm64/kvm/reset.c | 3 +
> arch/arm64/kvm/sys_regs.c | 490 +++++++++++++++++++++++--
> arch/arm64/kvm/sys_regs.h | 16 +
> include/kvm/arm_pmu.h | 66 ++++
> include/linux/kvm_host.h | 1 +
> include/uapi/linux/kvm.h | 2 +
> virt/kvm/arm/pmu.c | 500 ++++++++++++++++++++++++++
> virt/kvm/kvm_main.c | 4 +
> 17 files changed, 1188 insertions(+), 70 deletions(-)
> create mode 100644 Documentation/virtual/kvm/devices/arm-pmu.txt
> create mode 100644 include/kvm/arm_pmu.h
> create mode 100644 virt/kvm/arm/pmu.c
>
--
Shannon
^ permalink raw reply [flat|nested] 113+ messages in thread
* [PATCH v2 00/22] KVM: ARM64: Add guest PMU support
2015-09-11 8:54 ` Shannon Zhao
@ 2015-09-16 21:07 ` Wei Huang
-1 siblings, 0 replies; 113+ messages in thread
From: Wei Huang @ 2015-09-16 21:07 UTC (permalink / raw)
To: linux-arm-kernel
On 09/11/2015 03:54 AM, Shannon Zhao wrote:
> From: Shannon Zhao <shannon.zhao@linaro.org>
>
> This patchset adds guest PMU support for KVM on ARM64. It takes
> trap-and-emulate approach. When guest wants to monitor one event, it
> will be trapped by KVM and KVM will call perf_event API to create a perf
> event and call relevant perf_event APIs to get the count value of event.
>
> Use perf to test this patchset in guest. When using "perf list", it
> shows the list of the hardware events and hardware cache events perf
> supports. Then use "perf stat -e EVENT" to monitor some event. For
> example, use "perf stat -e cycles" to count cpu cycles and
> "perf stat -e cache-misses" to count cache misses.
>
> Below are the outputs of "perf stat -r 5 sleep 5" when running in host
> and guest.
>
> Host:
> Performance counter stats for 'sleep 5' (5 runs):
>
> 0.551428 task-clock (msec) # 0.000 CPUs utilized ( +- 0.91% )
> 1 context-switches # 0.002 M/sec
> 0 cpu-migrations # 0.000 K/sec
> 48 page-faults # 0.088 M/sec ( +- 1.05% )
> 1150265 cycles # 2.086 GHz ( +- 0.92% )
> <not supported> stalled-cycles-frontend
> <not supported> stalled-cycles-backend
> 526398 instructions # 0.46 insns per cycle ( +- 0.89% )
> <not supported> branches
> 9485 branch-misses # 17.201 M/sec ( +- 2.35% )
>
> 5.000831616 seconds time elapsed ( +- 0.00% )
>
> Guest:
> Performance counter stats for 'sleep 5' (5 runs):
>
> 0.730868 task-clock (msec) # 0.000 CPUs utilized ( +- 1.13% )
> 1 context-switches # 0.001 M/sec
> 0 cpu-migrations # 0.000 K/sec
> 48 page-faults # 0.065 M/sec ( +- 0.42% )
> 1642982 cycles # 2.248 GHz ( +- 1.04% )
> <not supported> stalled-cycles-frontend
> <not supported> stalled-cycles-backend
> 637964 instructions # 0.39 insns per cycle ( +- 0.65% )
> <not supported> branches
> 10377 branch-misses # 14.198 M/sec ( +- 1.09% )
>
> 5.001289068 seconds time elapsed ( +- 0.00% )
>
> This patchset can be fetched from [1] and the relevant QEMU version for
> test can be fetched from [2].
>
> Thanks,
> Shannon
>
> [1] https://git.linaro.org/people/shannon.zhao/linux-mainline.git KVM_ARM64_PMU_v2
> [2] https://git.linaro.org/people/shannon.zhao/qemu.git PMU_v2
I am testing this series. The first question is: do you plan to add ACPI
support in QEMU? My in-house kernel uses ACPI for device probing. I had
to force "acpi=off" when I test this patch series.
>
> Shannon Zhao (22):
> ARM64: Move PMU register related defines to asm/pmu.h
> KVM: ARM64: Define PMU data structure for each vcpu
> KVM: ARM64: Add offset defines for PMU registers
> KVM: ARM64: Add reset and access handlers for PMCR_EL0 register
> KVM: ARM64: Add a helper for CP15 registers reset to UNKNOWN
> KVM: ARM64: Add reset and access handlers for PMSELR register
> KVM: ARM64: Add reset and access handlers for PMCEID0 and PMCEID1
> register
> KVM: ARM64: PMU: Add perf event map and introduce perf event creating
> function
> KVM: ARM64: Add reset and access handlers for PMXEVTYPER register
> KVM: ARM64: Add reset and access handlers for PMXEVCNTR register
> KVM: ARM64: Add reset and access handlers for PMCCNTR register
> KVM: ARM64: Add reset and access handlers for PMCNTENSET and
> PMCNTENCLR register
> KVM: ARM64: Add reset and access handlers for PMINTENSET and
> PMINTENCLR register
> KVM: ARM64: Add reset and access handlers for PMOVSSET and PMOVSCLR
> register
> KVM: ARM64: Add a helper for CP15 registers reset to specified value
> KVM: ARM64: Add reset and access handlers for PMUSERENR register
> KVM: ARM64: Add reset and access handlers for PMSWINC register
> KVM: ARM64: Add access handlers for PMEVCNTRn and PMEVTYPERn register
> KVM: ARM64: Add PMU overflow interrupt routing
> KVM: ARM64: Reset PMU state when resetting vcpu
> KVM: ARM64: Free perf event of PMU when destroying vcpu
> KVM: ARM64: Add a new kvm ARM PMU device
>
> Documentation/virtual/kvm/devices/arm-pmu.txt | 15 +
> arch/arm/kvm/arm.c | 4 +
> arch/arm64/include/asm/kvm_asm.h | 59 ++-
> arch/arm64/include/asm/kvm_host.h | 2 +
> arch/arm64/include/asm/pmu.h | 49 +++
> arch/arm64/include/uapi/asm/kvm.h | 3 +
> arch/arm64/kernel/perf_event.c | 35 --
> arch/arm64/kvm/Kconfig | 8 +
> arch/arm64/kvm/Makefile | 1 +
> arch/arm64/kvm/reset.c | 3 +
> arch/arm64/kvm/sys_regs.c | 490 +++++++++++++++++++++++--
> arch/arm64/kvm/sys_regs.h | 16 +
> include/kvm/arm_pmu.h | 66 ++++
> include/linux/kvm_host.h | 1 +
> include/uapi/linux/kvm.h | 2 +
> virt/kvm/arm/pmu.c | 500 ++++++++++++++++++++++++++
> virt/kvm/kvm_main.c | 4 +
> 17 files changed, 1188 insertions(+), 70 deletions(-)
> create mode 100644 Documentation/virtual/kvm/devices/arm-pmu.txt
> create mode 100644 include/kvm/arm_pmu.h
> create mode 100644 virt/kvm/arm/pmu.c
>
^ permalink raw reply [flat|nested] 113+ messages in thread
* Re: [PATCH v2 00/22] KVM: ARM64: Add guest PMU support
@ 2015-09-16 21:07 ` Wei Huang
0 siblings, 0 replies; 113+ messages in thread
From: Wei Huang @ 2015-09-16 21:07 UTC (permalink / raw)
To: Shannon Zhao, kvmarm
Cc: linux-arm-kernel, kvm, marc.zyngier, christoffer.dall,
will.deacon, alex.bennee, shannon.zhao, peter.huangpeng
On 09/11/2015 03:54 AM, Shannon Zhao wrote:
> From: Shannon Zhao <shannon.zhao@linaro.org>
>
> This patchset adds guest PMU support for KVM on ARM64. It takes
> trap-and-emulate approach. When guest wants to monitor one event, it
> will be trapped by KVM and KVM will call perf_event API to create a perf
> event and call relevant perf_event APIs to get the count value of event.
>
> Use perf to test this patchset in guest. When using "perf list", it
> shows the list of the hardware events and hardware cache events perf
> supports. Then use "perf stat -e EVENT" to monitor some event. For
> example, use "perf stat -e cycles" to count cpu cycles and
> "perf stat -e cache-misses" to count cache misses.
>
> Below are the outputs of "perf stat -r 5 sleep 5" when running in host
> and guest.
>
> Host:
> Performance counter stats for 'sleep 5' (5 runs):
>
> 0.551428 task-clock (msec) # 0.000 CPUs utilized ( +- 0.91% )
> 1 context-switches # 0.002 M/sec
> 0 cpu-migrations # 0.000 K/sec
> 48 page-faults # 0.088 M/sec ( +- 1.05% )
> 1150265 cycles # 2.086 GHz ( +- 0.92% )
> <not supported> stalled-cycles-frontend
> <not supported> stalled-cycles-backend
> 526398 instructions # 0.46 insns per cycle ( +- 0.89% )
> <not supported> branches
> 9485 branch-misses # 17.201 M/sec ( +- 2.35% )
>
> 5.000831616 seconds time elapsed ( +- 0.00% )
>
> Guest:
> Performance counter stats for 'sleep 5' (5 runs):
>
> 0.730868 task-clock (msec) # 0.000 CPUs utilized ( +- 1.13% )
> 1 context-switches # 0.001 M/sec
> 0 cpu-migrations # 0.000 K/sec
> 48 page-faults # 0.065 M/sec ( +- 0.42% )
> 1642982 cycles # 2.248 GHz ( +- 1.04% )
> <not supported> stalled-cycles-frontend
> <not supported> stalled-cycles-backend
> 637964 instructions # 0.39 insns per cycle ( +- 0.65% )
> <not supported> branches
> 10377 branch-misses # 14.198 M/sec ( +- 1.09% )
>
> 5.001289068 seconds time elapsed ( +- 0.00% )
>
> This patchset can be fetched from [1] and the relevant QEMU version for
> test can be fetched from [2].
>
> Thanks,
> Shannon
>
> [1] https://git.linaro.org/people/shannon.zhao/linux-mainline.git KVM_ARM64_PMU_v2
> [2] https://git.linaro.org/people/shannon.zhao/qemu.git PMU_v2
I am testing this series. The first question is: do you plan to add ACPI
support in QEMU? My in-house kernel uses ACPI for device probing. I had
to force "acpi=off" when I test this patch series.
>
> Shannon Zhao (22):
> ARM64: Move PMU register related defines to asm/pmu.h
> KVM: ARM64: Define PMU data structure for each vcpu
> KVM: ARM64: Add offset defines for PMU registers
> KVM: ARM64: Add reset and access handlers for PMCR_EL0 register
> KVM: ARM64: Add a helper for CP15 registers reset to UNKNOWN
> KVM: ARM64: Add reset and access handlers for PMSELR register
> KVM: ARM64: Add reset and access handlers for PMCEID0 and PMCEID1
> register
> KVM: ARM64: PMU: Add perf event map and introduce perf event creating
> function
> KVM: ARM64: Add reset and access handlers for PMXEVTYPER register
> KVM: ARM64: Add reset and access handlers for PMXEVCNTR register
> KVM: ARM64: Add reset and access handlers for PMCCNTR register
> KVM: ARM64: Add reset and access handlers for PMCNTENSET and
> PMCNTENCLR register
> KVM: ARM64: Add reset and access handlers for PMINTENSET and
> PMINTENCLR register
> KVM: ARM64: Add reset and access handlers for PMOVSSET and PMOVSCLR
> register
> KVM: ARM64: Add a helper for CP15 registers reset to specified value
> KVM: ARM64: Add reset and access handlers for PMUSERENR register
> KVM: ARM64: Add reset and access handlers for PMSWINC register
> KVM: ARM64: Add access handlers for PMEVCNTRn and PMEVTYPERn register
> KVM: ARM64: Add PMU overflow interrupt routing
> KVM: ARM64: Reset PMU state when resetting vcpu
> KVM: ARM64: Free perf event of PMU when destroying vcpu
> KVM: ARM64: Add a new kvm ARM PMU device
>
> Documentation/virtual/kvm/devices/arm-pmu.txt | 15 +
> arch/arm/kvm/arm.c | 4 +
> arch/arm64/include/asm/kvm_asm.h | 59 ++-
> arch/arm64/include/asm/kvm_host.h | 2 +
> arch/arm64/include/asm/pmu.h | 49 +++
> arch/arm64/include/uapi/asm/kvm.h | 3 +
> arch/arm64/kernel/perf_event.c | 35 --
> arch/arm64/kvm/Kconfig | 8 +
> arch/arm64/kvm/Makefile | 1 +
> arch/arm64/kvm/reset.c | 3 +
> arch/arm64/kvm/sys_regs.c | 490 +++++++++++++++++++++++--
> arch/arm64/kvm/sys_regs.h | 16 +
> include/kvm/arm_pmu.h | 66 ++++
> include/linux/kvm_host.h | 1 +
> include/uapi/linux/kvm.h | 2 +
> virt/kvm/arm/pmu.c | 500 ++++++++++++++++++++++++++
> virt/kvm/kvm_main.c | 4 +
> 17 files changed, 1188 insertions(+), 70 deletions(-)
> create mode 100644 Documentation/virtual/kvm/devices/arm-pmu.txt
> create mode 100644 include/kvm/arm_pmu.h
> create mode 100644 virt/kvm/arm/pmu.c
>
^ permalink raw reply [flat|nested] 113+ messages in thread
* [PATCH v2 00/22] KVM: ARM64: Add guest PMU support
2015-09-16 21:07 ` Wei Huang
@ 2015-09-17 1:32 ` Shannon Zhao
-1 siblings, 0 replies; 113+ messages in thread
From: Shannon Zhao @ 2015-09-17 1:32 UTC (permalink / raw)
To: linux-arm-kernel
Hi Wei,
On 2015/9/17 5:07, Wei Huang wrote:
> I am testing this series.
Thanks for your time and help.
> The first question is: do you plan to add ACPI
> support in QEMU?
To the completeness, this should be added. Maybe this could be added at
v3. But I have a look at the kernel PMU driver, it doesn't support
probing through ACPI, although there are some patches[1] out-of-tree.
> My in-house kernel uses ACPI for device probing. I had
> to force "acpi=off" when I test this patch series.
Guest kernel only boots with ACPI when you add "acpi=force". No need to
add "acpi=off".
Thanks,
[1] http://marc.info/?l=linaro-acpi&m=137949337925645&w=2
--
Shannon
^ permalink raw reply [flat|nested] 113+ messages in thread
* Re: [PATCH v2 00/22] KVM: ARM64: Add guest PMU support
@ 2015-09-17 1:32 ` Shannon Zhao
0 siblings, 0 replies; 113+ messages in thread
From: Shannon Zhao @ 2015-09-17 1:32 UTC (permalink / raw)
To: Wei Huang, Shannon Zhao, kvmarm
Cc: linux-arm-kernel, kvm, marc.zyngier, christoffer.dall,
will.deacon, alex.bennee, peter.huangpeng
Hi Wei,
On 2015/9/17 5:07, Wei Huang wrote:
> I am testing this series.
Thanks for your time and help.
> The first question is: do you plan to add ACPI
> support in QEMU?
To the completeness, this should be added. Maybe this could be added at
v3. But I have a look at the kernel PMU driver, it doesn't support
probing through ACPI, although there are some patches[1] out-of-tree.
> My in-house kernel uses ACPI for device probing. I had
> to force "acpi=off" when I test this patch series.
Guest kernel only boots with ACPI when you add "acpi=force". No need to
add "acpi=off".
Thanks,
[1] http://marc.info/?l=linaro-acpi&m=137949337925645&w=2
--
Shannon
^ permalink raw reply [flat|nested] 113+ messages in thread
* [PATCH v2 00/22] KVM: ARM64: Add guest PMU support
2015-09-17 1:32 ` Shannon Zhao
@ 2015-09-17 5:56 ` Wei Huang
-1 siblings, 0 replies; 113+ messages in thread
From: Wei Huang @ 2015-09-17 5:56 UTC (permalink / raw)
To: linux-arm-kernel
On 09/16/2015 08:32 PM, Shannon Zhao wrote:
> Hi Wei,
>
> On 2015/9/17 5:07, Wei Huang wrote:
>> I am testing this series.
> Thanks for your time and help.
>
>> The first question is: do you plan to add ACPI
>> support in QEMU?
I saw "KVM_{SET/GET}_DEVICE_ATTR failed: Invalid argument" while using
your QEMU tree (PMU_v2 branch). A quick debugging:
(1) dmesg on host kernel didn't show any vPMU initialization errors. So
I suspect the problem is related to QEMU.
(2) Commit 58771bc2a78 worked fine. So probably the problem was
introduced by new PMU code.
Have you seen it before?
Thanks,
-Wei
> To the completeness, this should be added. Maybe this could be added at
> v3. But I have a look at the kernel PMU driver, it doesn't support
> probing through ACPI, although there are some patches[1] out-of-tree.
>
>> My in-house kernel uses ACPI for device probing. I had
>> to force "acpi=off" when I test this patch series.
> Guest kernel only boots with ACPI when you add "acpi=force". No need to
> add "acpi=off".
>
> Thanks,
>
> [1] http://marc.info/?l=linaro-acpi&m=137949337925645&w=2
>
^ permalink raw reply [flat|nested] 113+ messages in thread
* Re: [PATCH v2 00/22] KVM: ARM64: Add guest PMU support
@ 2015-09-17 5:56 ` Wei Huang
0 siblings, 0 replies; 113+ messages in thread
From: Wei Huang @ 2015-09-17 5:56 UTC (permalink / raw)
To: Shannon Zhao, Shannon Zhao, kvmarm
Cc: linux-arm-kernel, kvm, marc.zyngier, christoffer.dall,
will.deacon, alex.bennee, peter.huangpeng
On 09/16/2015 08:32 PM, Shannon Zhao wrote:
> Hi Wei,
>
> On 2015/9/17 5:07, Wei Huang wrote:
>> I am testing this series.
> Thanks for your time and help.
>
>> The first question is: do you plan to add ACPI
>> support in QEMU?
I saw "KVM_{SET/GET}_DEVICE_ATTR failed: Invalid argument" while using
your QEMU tree (PMU_v2 branch). A quick debugging:
(1) dmesg on host kernel didn't show any vPMU initialization errors. So
I suspect the problem is related to QEMU.
(2) Commit 58771bc2a78 worked fine. So probably the problem was
introduced by new PMU code.
Have you seen it before?
Thanks,
-Wei
> To the completeness, this should be added. Maybe this could be added at
> v3. But I have a look at the kernel PMU driver, it doesn't support
> probing through ACPI, although there are some patches[1] out-of-tree.
>
>> My in-house kernel uses ACPI for device probing. I had
>> to force "acpi=off" when I test this patch series.
> Guest kernel only boots with ACPI when you add "acpi=force". No need to
> add "acpi=off".
>
> Thanks,
>
> [1] http://marc.info/?l=linaro-acpi&m=137949337925645&w=2
>
^ permalink raw reply [flat|nested] 113+ messages in thread
* [PATCH v2 00/22] KVM: ARM64: Add guest PMU support
2015-09-17 5:56 ` Wei Huang
(?)
@ 2015-09-17 6:47 ` Shannon Zhao
-1 siblings, 0 replies; 113+ messages in thread
From: Shannon Zhao @ 2015-09-17 6:47 UTC (permalink / raw)
To: linux-arm-kernel
Hi Wei,
On 2015/9/17 13:56, Wei Huang wrote:
>
>
> On 09/16/2015 08:32 PM, Shannon Zhao wrote:
>> Hi Wei,
>>
>> On 2015/9/17 5:07, Wei Huang wrote:
>>> I am testing this series.
>> Thanks for your time and help.
>>
>>> The first question is: do you plan to add ACPI
>>> support in QEMU?
>
> I saw "KVM_{SET/GET}_DEVICE_ATTR failed: Invalid argument" while using
> your QEMU tree (PMU_v2 branch). A quick debugging:
>
>From this log, it might fail at below check:
+ if (reg < VGIC_NR_SGIS || reg > dev->kvm->arch.vgic.nr_irqs)
+ return -EINVAL;
> (1) dmesg on host kernel didn't show any vPMU initialization errors. So
> I suspect the problem is related to QEMU.
> (2) Commit 58771bc2a78 worked fine. So probably the problem was
> introduced by new PMU code.
>
> Have you seen it before?
>
Oh, I didn't see this. And I checkout the code on git.linaro.org, it's
same with my local code.
Could you add some print in kvm_arm_pmu_set_irq of hw/misc/arm_pmu_kvm.c
and kvm_arm_pmu_set_attr, kvm_arm_pmu_set_irq of virt/kvm/arm/pmu.c.
Thanks,
--
Shannon
^ permalink raw reply [flat|nested] 113+ messages in thread
* Re: [PATCH v2 00/22] KVM: ARM64: Add guest PMU support
@ 2015-09-17 6:47 ` Shannon Zhao
0 siblings, 0 replies; 113+ messages in thread
From: Shannon Zhao @ 2015-09-17 6:47 UTC (permalink / raw)
To: Wei Huang, Shannon Zhao, kvmarm
Cc: kvm, marc.zyngier, will.deacon, linux-arm-kernel
Hi Wei,
On 2015/9/17 13:56, Wei Huang wrote:
>
>
> On 09/16/2015 08:32 PM, Shannon Zhao wrote:
>> Hi Wei,
>>
>> On 2015/9/17 5:07, Wei Huang wrote:
>>> I am testing this series.
>> Thanks for your time and help.
>>
>>> The first question is: do you plan to add ACPI
>>> support in QEMU?
>
> I saw "KVM_{SET/GET}_DEVICE_ATTR failed: Invalid argument" while using
> your QEMU tree (PMU_v2 branch). A quick debugging:
>
>From this log, it might fail at below check:
+ if (reg < VGIC_NR_SGIS || reg > dev->kvm->arch.vgic.nr_irqs)
+ return -EINVAL;
> (1) dmesg on host kernel didn't show any vPMU initialization errors. So
> I suspect the problem is related to QEMU.
> (2) Commit 58771bc2a78 worked fine. So probably the problem was
> introduced by new PMU code.
>
> Have you seen it before?
>
Oh, I didn't see this. And I checkout the code on git.linaro.org, it's
same with my local code.
Could you add some print in kvm_arm_pmu_set_irq of hw/misc/arm_pmu_kvm.c
and kvm_arm_pmu_set_attr, kvm_arm_pmu_set_irq of virt/kvm/arm/pmu.c.
Thanks,
--
Shannon
^ permalink raw reply [flat|nested] 113+ messages in thread
* Re: [PATCH v2 00/22] KVM: ARM64: Add guest PMU support
@ 2015-09-17 6:47 ` Shannon Zhao
0 siblings, 0 replies; 113+ messages in thread
From: Shannon Zhao @ 2015-09-17 6:47 UTC (permalink / raw)
To: Wei Huang, Shannon Zhao, kvmarm
Cc: kvm, marc.zyngier, will.deacon, linux-arm-kernel
Hi Wei,
On 2015/9/17 13:56, Wei Huang wrote:
>
>
> On 09/16/2015 08:32 PM, Shannon Zhao wrote:
>> Hi Wei,
>>
>> On 2015/9/17 5:07, Wei Huang wrote:
>>> I am testing this series.
>> Thanks for your time and help.
>>
>>> The first question is: do you plan to add ACPI
>>> support in QEMU?
>
> I saw "KVM_{SET/GET}_DEVICE_ATTR failed: Invalid argument" while using
> your QEMU tree (PMU_v2 branch). A quick debugging:
>
>From this log, it might fail at below check:
+ if (reg < VGIC_NR_SGIS || reg > dev->kvm->arch.vgic.nr_irqs)
+ return -EINVAL;
> (1) dmesg on host kernel didn't show any vPMU initialization errors. So
> I suspect the problem is related to QEMU.
> (2) Commit 58771bc2a78 worked fine. So probably the problem was
> introduced by new PMU code.
>
> Have you seen it before?
>
Oh, I didn't see this. And I checkout the code on git.linaro.org, it's
same with my local code.
Could you add some print in kvm_arm_pmu_set_irq of hw/misc/arm_pmu_kvm.c
and kvm_arm_pmu_set_attr, kvm_arm_pmu_set_irq of virt/kvm/arm/pmu.c.
Thanks,
--
Shannon
^ permalink raw reply [flat|nested] 113+ messages in thread
* [PATCH v2 00/22] KVM: ARM64: Add guest PMU support
2015-09-17 1:32 ` Shannon Zhao
@ 2015-09-17 9:30 ` Andrew Jones
-1 siblings, 0 replies; 113+ messages in thread
From: Andrew Jones @ 2015-09-17 9:30 UTC (permalink / raw)
To: linux-arm-kernel
On Thu, Sep 17, 2015 at 09:32:34AM +0800, Shannon Zhao wrote:
> Hi Wei,
>
> On 2015/9/17 5:07, Wei Huang wrote:
> > I am testing this series.
> Thanks for your time and help.
>
> > The first question is: do you plan to add ACPI
> > support in QEMU?
> To the completeness, this should be added. Maybe this could be added at
> v3. But I have a look at the kernel PMU driver, it doesn't support
> probing through ACPI, although there are some patches[1] out-of-tree.
>
> > My in-house kernel uses ACPI for device probing. I had
> > to force "acpi=off" when I test this patch series.
> Guest kernel only boots with ACPI when you add "acpi=force". No need to
> add "acpi=off".
The Red Hat kernel uses ACPI by default, when tables are present (Thanks
again for your QEMU table generation patches!)
drew
>
> Thanks,
>
> [1] http://marc.info/?l=linaro-acpi&m=137949337925645&w=2
> --
> Shannon
> _______________________________________________
> kvmarm mailing list
> kvmarm at lists.cs.columbia.edu
> https://lists.cs.columbia.edu/mailman/listinfo/kvmarm
^ permalink raw reply [flat|nested] 113+ messages in thread
* Re: [PATCH v2 00/22] KVM: ARM64: Add guest PMU support
@ 2015-09-17 9:30 ` Andrew Jones
0 siblings, 0 replies; 113+ messages in thread
From: Andrew Jones @ 2015-09-17 9:30 UTC (permalink / raw)
To: Shannon Zhao
Cc: Wei Huang, Shannon Zhao, kvmarm, kvm, marc.zyngier, will.deacon,
linux-arm-kernel
On Thu, Sep 17, 2015 at 09:32:34AM +0800, Shannon Zhao wrote:
> Hi Wei,
>
> On 2015/9/17 5:07, Wei Huang wrote:
> > I am testing this series.
> Thanks for your time and help.
>
> > The first question is: do you plan to add ACPI
> > support in QEMU?
> To the completeness, this should be added. Maybe this could be added at
> v3. But I have a look at the kernel PMU driver, it doesn't support
> probing through ACPI, although there are some patches[1] out-of-tree.
>
> > My in-house kernel uses ACPI for device probing. I had
> > to force "acpi=off" when I test this patch series.
> Guest kernel only boots with ACPI when you add "acpi=force". No need to
> add "acpi=off".
The Red Hat kernel uses ACPI by default, when tables are present (Thanks
again for your QEMU table generation patches!)
drew
>
> Thanks,
>
> [1] http://marc.info/?l=linaro-acpi&m=137949337925645&w=2
> --
> Shannon
> _______________________________________________
> kvmarm mailing list
> kvmarm@lists.cs.columbia.edu
> https://lists.cs.columbia.edu/mailman/listinfo/kvmarm
^ permalink raw reply [flat|nested] 113+ messages in thread
* [PATCH v2 00/22] KVM: ARM64: Add guest PMU support
2015-09-17 9:30 ` Andrew Jones
(?)
@ 2015-09-17 9:35 ` Shannon Zhao
-1 siblings, 0 replies; 113+ messages in thread
From: Shannon Zhao @ 2015-09-17 9:35 UTC (permalink / raw)
To: linux-arm-kernel
Hi Andrew,
On 2015/9/17 17:30, Andrew Jones wrote:
> On Thu, Sep 17, 2015 at 09:32:34AM +0800, Shannon Zhao wrote:
>> > Hi Wei,
>> >
>> > On 2015/9/17 5:07, Wei Huang wrote:
>>> > > I am testing this series.
>> > Thanks for your time and help.
>> >
>>> > > The first question is: do you plan to add ACPI
>>> > > support in QEMU?
>> > To the completeness, this should be added. Maybe this could be added at
>> > v3. But I have a look at the kernel PMU driver, it doesn't support
>> > probing through ACPI, although there are some patches[1] out-of-tree.
>> >
>>> > > My in-house kernel uses ACPI for device probing. I had
>>> > > to force "acpi=off" when I test this patch series.
>> > Guest kernel only boots with ACPI when you add "acpi=force". No need to
>> > add "acpi=off".
> The Red Hat kernel uses ACPI by default, when tables are present (Thanks
> again for your QEMU table generation patches!)
>
Oh, I see. Thanks for your explanation.
--
Shannon
^ permalink raw reply [flat|nested] 113+ messages in thread
* Re: [PATCH v2 00/22] KVM: ARM64: Add guest PMU support
@ 2015-09-17 9:35 ` Shannon Zhao
0 siblings, 0 replies; 113+ messages in thread
From: Shannon Zhao @ 2015-09-17 9:35 UTC (permalink / raw)
To: Andrew Jones, Shannon Zhao
Cc: Wei Huang, kvmarm, kvm, marc.zyngier, will.deacon,
linux-arm-kernel
Hi Andrew,
On 2015/9/17 17:30, Andrew Jones wrote:
> On Thu, Sep 17, 2015 at 09:32:34AM +0800, Shannon Zhao wrote:
>> > Hi Wei,
>> >
>> > On 2015/9/17 5:07, Wei Huang wrote:
>>> > > I am testing this series.
>> > Thanks for your time and help.
>> >
>>> > > The first question is: do you plan to add ACPI
>>> > > support in QEMU?
>> > To the completeness, this should be added. Maybe this could be added at
>> > v3. But I have a look at the kernel PMU driver, it doesn't support
>> > probing through ACPI, although there are some patches[1] out-of-tree.
>> >
>>> > > My in-house kernel uses ACPI for device probing. I had
>>> > > to force "acpi=off" when I test this patch series.
>> > Guest kernel only boots with ACPI when you add "acpi=force". No need to
>> > add "acpi=off".
> The Red Hat kernel uses ACPI by default, when tables are present (Thanks
> again for your QEMU table generation patches!)
>
Oh, I see. Thanks for your explanation.
--
Shannon
^ permalink raw reply [flat|nested] 113+ messages in thread
* Re: [PATCH v2 00/22] KVM: ARM64: Add guest PMU support
@ 2015-09-17 9:35 ` Shannon Zhao
0 siblings, 0 replies; 113+ messages in thread
From: Shannon Zhao @ 2015-09-17 9:35 UTC (permalink / raw)
To: Andrew Jones, Shannon Zhao
Cc: Wei Huang, kvmarm, kvm, marc.zyngier, will.deacon,
linux-arm-kernel
Hi Andrew,
On 2015/9/17 17:30, Andrew Jones wrote:
> On Thu, Sep 17, 2015 at 09:32:34AM +0800, Shannon Zhao wrote:
>> > Hi Wei,
>> >
>> > On 2015/9/17 5:07, Wei Huang wrote:
>>> > > I am testing this series.
>> > Thanks for your time and help.
>> >
>>> > > The first question is: do you plan to add ACPI
>>> > > support in QEMU?
>> > To the completeness, this should be added. Maybe this could be added at
>> > v3. But I have a look at the kernel PMU driver, it doesn't support
>> > probing through ACPI, although there are some patches[1] out-of-tree.
>> >
>>> > > My in-house kernel uses ACPI for device probing. I had
>>> > > to force "acpi=off" when I test this patch series.
>> > Guest kernel only boots with ACPI when you add "acpi=force". No need to
>> > add "acpi=off".
> The Red Hat kernel uses ACPI by default, when tables are present (Thanks
> again for your QEMU table generation patches!)
>
Oh, I see. Thanks for your explanation.
--
Shannon
^ permalink raw reply [flat|nested] 113+ messages in thread
end of thread, other threads:[~2015-09-17 9:39 UTC | newest]
Thread overview: 113+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2015-09-11 8:54 [PATCH v2 00/22] KVM: ARM64: Add guest PMU support Shannon Zhao
2015-09-11 8:54 ` Shannon Zhao
2015-09-11 8:54 ` Shannon Zhao
2015-09-11 8:54 ` [PATCH v2 01/22] ARM64: Move PMU register related defines to asm/pmu.h Shannon Zhao
2015-09-11 8:54 ` Shannon Zhao
2015-09-11 8:54 ` Shannon Zhao
2015-09-11 8:54 ` [PATCH v2 02/22] KVM: ARM64: Define PMU data structure for each vcpu Shannon Zhao
2015-09-11 8:54 ` Shannon Zhao
2015-09-11 8:54 ` Shannon Zhao
2015-09-11 9:10 ` Marc Zyngier
2015-09-11 9:10 ` Marc Zyngier
2015-09-11 9:58 ` Shannon Zhao
2015-09-11 9:58 ` Shannon Zhao
2015-09-11 9:58 ` Shannon Zhao
2015-09-11 8:54 ` [PATCH v2 03/22] KVM: ARM64: Add offset defines for PMU registers Shannon Zhao
2015-09-11 8:54 ` Shannon Zhao
2015-09-11 8:54 ` Shannon Zhao
2015-09-11 8:54 ` [PATCH v2 04/22] KVM: ARM64: Add reset and access handlers for PMCR_EL0 register Shannon Zhao
2015-09-11 8:54 ` Shannon Zhao
2015-09-11 8:54 ` Shannon Zhao
2015-09-11 10:07 ` Marc Zyngier
2015-09-11 10:07 ` Marc Zyngier
2015-09-14 3:14 ` Shannon Zhao
2015-09-14 3:14 ` Shannon Zhao
2015-09-14 3:14 ` Shannon Zhao
2015-09-14 12:11 ` Marc Zyngier
2015-09-14 12:11 ` Marc Zyngier
2015-09-11 8:54 ` [PATCH v2 05/22] KVM: ARM64: Add a helper for CP15 registers reset to UNKNOWN Shannon Zhao
2015-09-11 8:54 ` Shannon Zhao
2015-09-11 8:54 ` Shannon Zhao
2015-09-11 10:16 ` Marc Zyngier
2015-09-11 10:16 ` Marc Zyngier
2015-09-11 10:17 ` Marc Zyngier
2015-09-11 10:17 ` Marc Zyngier
2015-09-11 8:54 ` [PATCH v2 06/22] KVM: ARM64: Add reset and access handlers for PMSELR register Shannon Zhao
2015-09-11 8:54 ` Shannon Zhao
2015-09-11 8:54 ` Shannon Zhao
2015-09-11 8:55 ` [PATCH v2 07/22] KVM: ARM64: Add reset and access handlers for PMCEID0 and PMCEID1 register Shannon Zhao
2015-09-11 8:55 ` Shannon Zhao
2015-09-11 8:55 ` Shannon Zhao
2015-09-11 10:27 ` Marc Zyngier
2015-09-11 10:27 ` Marc Zyngier
2015-09-11 8:55 ` [PATCH v2 08/22] KVM: ARM64: PMU: Add perf event map and introduce perf event creating function Shannon Zhao
2015-09-11 8:55 ` Shannon Zhao
2015-09-11 8:55 ` Shannon Zhao
2015-09-11 11:04 ` Marc Zyngier
2015-09-11 11:04 ` Marc Zyngier
2015-09-11 13:35 ` Shannon Zhao
2015-09-11 13:35 ` Shannon Zhao
2015-09-11 14:14 ` Marc Zyngier
2015-09-11 14:14 ` Marc Zyngier
2015-09-11 8:55 ` [PATCH v2 09/22] KVM: ARM64: Add reset and access handlers for PMXEVTYPER register Shannon Zhao
2015-09-11 8:55 ` Shannon Zhao
2015-09-11 8:55 ` Shannon Zhao
2015-09-11 8:55 ` [PATCH v2 10/22] KVM: ARM64: Add reset and access handlers for PMXEVCNTR register Shannon Zhao
2015-09-11 8:55 ` Shannon Zhao
2015-09-11 8:55 ` Shannon Zhao
2015-09-11 8:55 ` [PATCH v2 11/22] KVM: ARM64: Add reset and access handlers for PMCCNTR register Shannon Zhao
2015-09-11 8:55 ` Shannon Zhao
2015-09-11 8:55 ` Shannon Zhao
2015-09-11 8:55 ` [PATCH v2 12/22] KVM: ARM64: Add reset and access handlers for PMCNTENSET and PMCNTENCLR register Shannon Zhao
2015-09-11 8:55 ` Shannon Zhao
2015-09-11 8:55 ` Shannon Zhao
2015-09-11 8:55 ` [PATCH v2 13/22] KVM: ARM64: Add reset and access handlers for PMINTENSET and PMINTENCLR register Shannon Zhao
2015-09-11 8:55 ` Shannon Zhao
2015-09-11 8:55 ` Shannon Zhao
2015-09-11 8:55 ` [PATCH v2 14/22] KVM: ARM64: Add reset and access handlers for PMOVSSET and PMOVSCLR register Shannon Zhao
2015-09-11 8:55 ` Shannon Zhao
2015-09-11 8:55 ` Shannon Zhao
2015-09-11 8:55 ` [PATCH v2 15/22] KVM: ARM64: Add a helper for CP15 registers reset to specified value Shannon Zhao
2015-09-11 8:55 ` Shannon Zhao
2015-09-11 8:55 ` Shannon Zhao
2015-09-11 8:55 ` [PATCH v2 16/22] KVM: ARM64: Add reset and access handlers for PMUSERENR register Shannon Zhao
2015-09-11 8:55 ` Shannon Zhao
2015-09-11 8:55 ` Shannon Zhao
2015-09-11 8:55 ` [PATCH v2 17/22] KVM: ARM64: Add reset and access handlers for PMSWINC register Shannon Zhao
2015-09-11 8:55 ` Shannon Zhao
2015-09-11 8:55 ` Shannon Zhao
2015-09-11 8:55 ` [PATCH v2 18/22] KVM: ARM64: Add access handlers for PMEVCNTRn and PMEVTYPERn register Shannon Zhao
2015-09-11 8:55 ` Shannon Zhao
2015-09-11 8:55 ` Shannon Zhao
2015-09-11 8:55 ` [PATCH v2 19/22] KVM: ARM64: Add PMU overflow interrupt routing Shannon Zhao
2015-09-11 8:55 ` Shannon Zhao
2015-09-11 8:55 ` Shannon Zhao
2015-09-11 8:55 ` [PATCH v2 20/22] KVM: ARM64: Reset PMU state when resetting vcpu Shannon Zhao
2015-09-11 8:55 ` Shannon Zhao
2015-09-11 8:55 ` Shannon Zhao
2015-09-11 8:55 ` [PATCH v2 21/22] KVM: ARM64: Free perf event of PMU when destroying vcpu Shannon Zhao
2015-09-11 8:55 ` Shannon Zhao
2015-09-11 8:55 ` Shannon Zhao
2015-09-11 8:55 ` [PATCH v2 22/22] KVM: ARM64: Add a new kvm ARM PMU device Shannon Zhao
2015-09-11 8:55 ` Shannon Zhao
2015-09-11 8:55 ` Shannon Zhao
2015-09-14 11:53 ` [PATCH v2 00/22] KVM: ARM64: Add guest PMU support Christoffer Dall
2015-09-14 11:53 ` Christoffer Dall
2015-09-14 12:58 ` Shannon Zhao
2015-09-14 12:58 ` Shannon Zhao
2015-09-14 13:24 ` Shannon Zhao
2015-09-14 13:24 ` Shannon Zhao
2015-09-16 21:07 ` Wei Huang
2015-09-16 21:07 ` Wei Huang
2015-09-17 1:32 ` Shannon Zhao
2015-09-17 1:32 ` Shannon Zhao
2015-09-17 5:56 ` Wei Huang
2015-09-17 5:56 ` Wei Huang
2015-09-17 6:47 ` Shannon Zhao
2015-09-17 6:47 ` Shannon Zhao
2015-09-17 6:47 ` Shannon Zhao
2015-09-17 9:30 ` Andrew Jones
2015-09-17 9:30 ` Andrew Jones
2015-09-17 9:35 ` Shannon Zhao
2015-09-17 9:35 ` Shannon Zhao
2015-09-17 9:35 ` Shannon Zhao
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.