[kvm-unit-tests PATCH v3 00/12] KVM: arm64: PMUv3 Event Counter Tests

[kvm-unit-tests PATCH v3 00/12] KVM: arm64: PMUv3 Event Counter Tests

[Eric Auger]

This series implements tests exercising the PMUv3 event counters.
It tests both the 32-bit and 64-bit versions. Overflow interrupts
also are checked. Those tests only are written for arm64.

It allowed to reveal some issues related to SW_INCR implementation
(esp. related to 64-bit implementation), some problems related to
32-bit <-> 64-bit transitions and consistency of enabled states
of odd and event counters. Fixes now are upstream.

Drew kindly provided "arm64: Provide read/write_sysreg_s".

All PMU tests can be launched with:
./run_tests.sh -g pmu
Tests also can be launched individually. For example:
./arm-run arm/pmu.flat -append 'pmu-chained-sw-incr'

With KVM:
- On TX2, I have some random failures due to MEM_ACCESS event
  measured with a great disparity. This is not observed on
  other machines I have access to.

The series can be found at:
https://github.com/eauger/kut/tree/pmu_event_counters_v3

History:

v2 -> v3:
- Took into account numerous comments from Drew
- Applied Andre's R-b when code has not changed too much
- Don't check PMCR.IMP anymore
- better handling of version
- put simple SW_INCR tests for easier TCG testing

v1 -> v2:
- Took into account Andre's comments except I did not
  use cnbz in the mem_access_loop() and I did not use
  @loop directly. Those changes had side effects I
  cannot explain on the tests. Anyway I think this can
  be improved later on.
- removed [kvm-unit-tests PATCH 09/10] arm/arm64: gic:
  Introduce setup_irq() helper

RFC -> v1:
- Use new report() proto
- Style cleanup
- do not warn about ARM spec recommendations
- add a comment about PMCEID0/1 splits

Andrew Jones (1):
  arm64: Provide read/write_sysreg_s

Eric Auger (11):
  arm: pmu: Let pmu tests take a sub-test parameter
  arm: pmu: Don't check PMCR.IMP anymore
  arm: pmu: Add a pmu struct
  arm: pmu: Introduce defines for PMU versions
  arm: pmu: Check Required Event Support
  arm: pmu: Basic event counter Tests
  arm: pmu: Test SW_INCR event count
  arm: pmu: Test chained counters
  arm: pmu: test 32-bit <-> 64-bit transitions
  arm: gic: Introduce gic_irq_set_clr_enable() helper
  arm: pmu: Test overflow interrupts

 arm/pmu.c              | 839 +++++++++++++++++++++++++++++++++++++++--
 arm/unittests.cfg      |  61 ++-
 lib/arm/asm/gic.h      |   4 +
 lib/arm/gic.c          |  31 ++
 lib/arm64/asm/sysreg.h |  17 +
 lib/bitops.h           |   3 +
 6 files changed, 918 insertions(+), 37 deletions(-)

-- 
2.20.1

[kvm-unit-tests PATCH v3 00/12] KVM: arm64: PMUv3 Event Counter Tests

[Eric Auger]

This series implements tests exercising the PMUv3 event counters.
It tests both the 32-bit and 64-bit versions. Overflow interrupts
also are checked. Those tests only are written for arm64.

It allowed to reveal some issues related to SW_INCR implementation
(esp. related to 64-bit implementation), some problems related to
32-bit <-> 64-bit transitions and consistency of enabled states
of odd and event counters. Fixes now are upstream.

Drew kindly provided "arm64: Provide read/write_sysreg_s".

All PMU tests can be launched with:
./run_tests.sh -g pmu
Tests also can be launched individually. For example:
./arm-run arm/pmu.flat -append 'pmu-chained-sw-incr'

With KVM:
- On TX2, I have some random failures due to MEM_ACCESS event
  measured with a great disparity. This is not observed on
  other machines I have access to.

The series can be found at:
https://github.com/eauger/kut/tree/pmu_event_counters_v3

History:

v2 -> v3:
- Took into account numerous comments from Drew
- Applied Andre's R-b when code has not changed too much
- Don't check PMCR.IMP anymore
- better handling of version
- put simple SW_INCR tests for easier TCG testing

v1 -> v2:
- Took into account Andre's comments except I did not
  use cnbz in the mem_access_loop() and I did not use
  @loop directly. Those changes had side effects I
  cannot explain on the tests. Anyway I think this can
  be improved later on.
- removed [kvm-unit-tests PATCH 09/10] arm/arm64: gic:
  Introduce setup_irq() helper

RFC -> v1:
- Use new report() proto
- Style cleanup
- do not warn about ARM spec recommendations
- add a comment about PMCEID0/1 splits

Andrew Jones (1):
  arm64: Provide read/write_sysreg_s

Eric Auger (11):
  arm: pmu: Let pmu tests take a sub-test parameter
  arm: pmu: Don't check PMCR.IMP anymore
  arm: pmu: Add a pmu struct
  arm: pmu: Introduce defines for PMU versions
  arm: pmu: Check Required Event Support
  arm: pmu: Basic event counter Tests
  arm: pmu: Test SW_INCR event count
  arm: pmu: Test chained counters
  arm: pmu: test 32-bit <-> 64-bit transitions
  arm: gic: Introduce gic_irq_set_clr_enable() helper
  arm: pmu: Test overflow interrupts

 arm/pmu.c              | 839 +++++++++++++++++++++++++++++++++++++++--
 arm/unittests.cfg      |  61 ++-
 lib/arm/asm/gic.h      |   4 +
 lib/arm/gic.c          |  31 ++
 lib/arm64/asm/sysreg.h |  17 +
 lib/bitops.h           |   3 +
 6 files changed, 918 insertions(+), 37 deletions(-)

-- 
2.20.1

[kvm-unit-tests PATCH v3 00/12] KVM: arm64: PMUv3 Event Counter Tests

[Eric Auger]

This series implements tests exercising the PMUv3 event counters.
It tests both the 32-bit and 64-bit versions. Overflow interrupts
also are checked. Those tests only are written for arm64.

It allowed to reveal some issues related to SW_INCR implementation
(esp. related to 64-bit implementation), some problems related to
32-bit <-> 64-bit transitions and consistency of enabled states
of odd and event counters. Fixes now are upstream.

Drew kindly provided "arm64: Provide read/write_sysreg_s".

All PMU tests can be launched with:
./run_tests.sh -g pmu
Tests also can be launched individually. For example:
./arm-run arm/pmu.flat -append 'pmu-chained-sw-incr'

With KVM:
- On TX2, I have some random failures due to MEM_ACCESS event
  measured with a great disparity. This is not observed on
  other machines I have access to.

The series can be found at:
https://github.com/eauger/kut/tree/pmu_event_counters_v3

History:

v2 -> v3:
- Took into account numerous comments from Drew
- Applied Andre's R-b when code has not changed too much
- Don't check PMCR.IMP anymore
- better handling of version
- put simple SW_INCR tests for easier TCG testing

v1 -> v2:
- Took into account Andre's comments except I did not
  use cnbz in the mem_access_loop() and I did not use
  @loop directly. Those changes had side effects I
  cannot explain on the tests. Anyway I think this can
  be improved later on.
- removed [kvm-unit-tests PATCH 09/10] arm/arm64: gic:
  Introduce setup_irq() helper

RFC -> v1:
- Use new report() proto
- Style cleanup
- do not warn about ARM spec recommendations
- add a comment about PMCEID0/1 splits

Andrew Jones (1):
  arm64: Provide read/write_sysreg_s

Eric Auger (11):
  arm: pmu: Let pmu tests take a sub-test parameter
  arm: pmu: Don't check PMCR.IMP anymore
  arm: pmu: Add a pmu struct
  arm: pmu: Introduce defines for PMU versions
  arm: pmu: Check Required Event Support
  arm: pmu: Basic event counter Tests
  arm: pmu: Test SW_INCR event count
  arm: pmu: Test chained counters
  arm: pmu: test 32-bit <-> 64-bit transitions
  arm: gic: Introduce gic_irq_set_clr_enable() helper
  arm: pmu: Test overflow interrupts

 arm/pmu.c              | 839 +++++++++++++++++++++++++++++++++++++++--
 arm/unittests.cfg      |  61 ++-
 lib/arm/asm/gic.h      |   4 +
 lib/arm/gic.c          |  31 ++
 lib/arm64/asm/sysreg.h |  17 +
 lib/bitops.h           |   3 +
 6 files changed, 918 insertions(+), 37 deletions(-)

-- 
2.20.1

_______________________________________________
kvmarm mailing list
kvmarm@lists.cs.columbia.edu
https://lists.cs.columbia.edu/mailman/listinfo/kvmarm

[kvm-unit-tests PATCH v3 01/12] arm64: Provide read/write_sysreg_s

[Eric Auger]

From: Andrew Jones <drjones@redhat.com>

Sometimes we need to test access to system registers which are
missing assembler mnemonics.

Signed-off-by: Andrew Jones <drjones@redhat.com>
Reviewed-by: Alexandru Elisei <alexandru.elisei@arm.com>
---
 lib/arm64/asm/sysreg.h | 11 +++++++++++
 1 file changed, 11 insertions(+)

diff --git a/lib/arm64/asm/sysreg.h b/lib/arm64/asm/sysreg.h
index a03830b..a45eebd 100644
--- a/lib/arm64/asm/sysreg.h
+++ b/lib/arm64/asm/sysreg.h
@@ -38,6 +38,17 @@
 	asm volatile("msr " xstr(r) ", %x0" : : "rZ" (__val));	\
 } while (0)
 
+#define read_sysreg_s(r) ({					\
+	u64 __val;						\
+	asm volatile("mrs_s %0, " xstr(r) : "=r" (__val));	\
+	__val;							\
+})
+
+#define write_sysreg_s(v, r) do {				\
+	u64 __val = (u64)v;					\
+	asm volatile("msr_s " xstr(r) ", %x0" : : "rZ" (__val));\
+} while (0)
+
 asm(
 "	.irp	num,0,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30\n"
 "	.equ	.L__reg_num_x\\num, \\num\n"
-- 
2.20.1

[kvm-unit-tests PATCH v3 01/12] arm64: Provide read/write_sysreg_s

[Eric Auger]

From: Andrew Jones <drjones@redhat.com>

Sometimes we need to test access to system registers which are
missing assembler mnemonics.

Signed-off-by: Andrew Jones <drjones@redhat.com>
Reviewed-by: Alexandru Elisei <alexandru.elisei@arm.com>
---
 lib/arm64/asm/sysreg.h | 11 +++++++++++
 1 file changed, 11 insertions(+)

diff --git a/lib/arm64/asm/sysreg.h b/lib/arm64/asm/sysreg.h
index a03830b..a45eebd 100644
--- a/lib/arm64/asm/sysreg.h
+++ b/lib/arm64/asm/sysreg.h
@@ -38,6 +38,17 @@
 	asm volatile("msr " xstr(r) ", %x0" : : "rZ" (__val));	\
 } while (0)
 
+#define read_sysreg_s(r) ({					\
+	u64 __val;						\
+	asm volatile("mrs_s %0, " xstr(r) : "=r" (__val));	\
+	__val;							\
+})
+
+#define write_sysreg_s(v, r) do {				\
+	u64 __val = (u64)v;					\
+	asm volatile("msr_s " xstr(r) ", %x0" : : "rZ" (__val));\
+} while (0)
+
 asm(
 "	.irp	num,0,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30\n"
 "	.equ	.L__reg_num_x\\num, \\num\n"
-- 
2.20.1

[kvm-unit-tests PATCH v3 01/12] arm64: Provide read/write_sysreg_s

[Eric Auger]

From: Andrew Jones <drjones@redhat.com>

Sometimes we need to test access to system registers which are
missing assembler mnemonics.

Signed-off-by: Andrew Jones <drjones@redhat.com>
Reviewed-by: Alexandru Elisei <alexandru.elisei@arm.com>
---
 lib/arm64/asm/sysreg.h | 11 +++++++++++
 1 file changed, 11 insertions(+)

diff --git a/lib/arm64/asm/sysreg.h b/lib/arm64/asm/sysreg.h
index a03830b..a45eebd 100644
--- a/lib/arm64/asm/sysreg.h
+++ b/lib/arm64/asm/sysreg.h
@@ -38,6 +38,17 @@
 	asm volatile("msr " xstr(r) ", %x0" : : "rZ" (__val));	\
 } while (0)
 
+#define read_sysreg_s(r) ({					\
+	u64 __val;						\
+	asm volatile("mrs_s %0, " xstr(r) : "=r" (__val));	\
+	__val;							\
+})
+
+#define write_sysreg_s(v, r) do {				\
+	u64 __val = (u64)v;					\
+	asm volatile("msr_s " xstr(r) ", %x0" : : "rZ" (__val));\
+} while (0)
+
 asm(
 "	.irp	num,0,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30\n"
 "	.equ	.L__reg_num_x\\num, \\num\n"
-- 
2.20.1

_______________________________________________
kvmarm mailing list
kvmarm@lists.cs.columbia.edu
https://lists.cs.columbia.edu/mailman/listinfo/kvmarm

[kvm-unit-tests PATCH v3 02/12] arm: pmu: Let pmu tests take a sub-test parameter

[Eric Auger]

As we intend to introduce more PMU tests, let's add
a sub-test parameter that will allow to categorize
them. Existing tests are in the cycle-counter category.

Signed-off-by: Eric Auger <eric.auger@redhat.com>
Reviewed-by: Andre Przywara <andre.przywara@arm.com>

---

v2 -> v3:
- added report_prefix_pop()
---
 arm/pmu.c         | 25 ++++++++++++++++---------
 arm/unittests.cfg |  7 ++++---
 2 files changed, 20 insertions(+), 12 deletions(-)

diff --git a/arm/pmu.c b/arm/pmu.c
index d5a03a6..0122f0a 100644
--- a/arm/pmu.c
+++ b/arm/pmu.c
@@ -287,22 +287,29 @@ int main(int argc, char *argv[])
 {
 	int cpi = 0;
 
-	if (argc > 1)
-		cpi = atol(argv[1]);
-
 	if (!pmu_probe()) {
 		printf("No PMU found, test skipped...\n");
 		return report_summary();
 	}
 
+	if (argc < 2)
+		report_abort("no test specified");
+
 	report_prefix_push("pmu");
 
-	report(check_pmcr(), "Control register");
-	report(check_cycles_increase(),
-	       "Monotonically increasing cycle count");
-	report(check_cpi(cpi), "Cycle/instruction ratio");
-
-	pmccntr64_test();
+	if (strcmp(argv[1], "cycle-counter") == 0) {
+		report_prefix_push(argv[1]);
+		if (argc > 2)
+			cpi = atol(argv[2]);
+		report(check_pmcr(), "Control register");
+		report(check_cycles_increase(),
+		       "Monotonically increasing cycle count");
+		report(check_cpi(cpi), "Cycle/instruction ratio");
+		pmccntr64_test();
+		report_prefix_pop();
+	} else {
+		report_abort("Unknown sub-test '%s'", argv[1]);
+	}
 
 	return report_summary();
 }
diff --git a/arm/unittests.cfg b/arm/unittests.cfg
index 017958d..fe6515c 100644
--- a/arm/unittests.cfg
+++ b/arm/unittests.cfg
@@ -61,21 +61,22 @@ file = pci-test.flat
 groups = pci
 
 # Test PMU support
-[pmu]
+[pmu-cycle-counter]
 file = pmu.flat
 groups = pmu
+extra_params = -append 'cycle-counter 0'
 
 # Test PMU support (TCG) with -icount IPC=1
 #[pmu-tcg-icount-1]
 #file = pmu.flat
-#extra_params = -icount 0 -append '1'
+#extra_params = -icount 0 -append 'cycle-counter 1'
 #groups = pmu
 #accel = tcg
 
 # Test PMU support (TCG) with -icount IPC=256
 #[pmu-tcg-icount-256]
 #file = pmu.flat
-#extra_params = -icount 8 -append '256'
+#extra_params = -icount 8 -append 'cycle-counter 256'
 #groups = pmu
 #accel = tcg
 
-- 
2.20.1

[kvm-unit-tests PATCH v3 02/12] arm: pmu: Let pmu tests take a sub-test parameter

[Eric Auger]

As we intend to introduce more PMU tests, let's add
a sub-test parameter that will allow to categorize
them. Existing tests are in the cycle-counter category.

Signed-off-by: Eric Auger <eric.auger@redhat.com>
Reviewed-by: Andre Przywara <andre.przywara@arm.com>

---

v2 -> v3:
- added report_prefix_pop()
---
 arm/pmu.c         | 25 ++++++++++++++++---------
 arm/unittests.cfg |  7 ++++---
 2 files changed, 20 insertions(+), 12 deletions(-)

diff --git a/arm/pmu.c b/arm/pmu.c
index d5a03a6..0122f0a 100644
--- a/arm/pmu.c
+++ b/arm/pmu.c
@@ -287,22 +287,29 @@ int main(int argc, char *argv[])
 {
 	int cpi = 0;
 
-	if (argc > 1)
-		cpi = atol(argv[1]);
-
 	if (!pmu_probe()) {
 		printf("No PMU found, test skipped...\n");
 		return report_summary();
 	}
 
+	if (argc < 2)
+		report_abort("no test specified");
+
 	report_prefix_push("pmu");
 
-	report(check_pmcr(), "Control register");
-	report(check_cycles_increase(),
-	       "Monotonically increasing cycle count");
-	report(check_cpi(cpi), "Cycle/instruction ratio");
-
-	pmccntr64_test();
+	if (strcmp(argv[1], "cycle-counter") == 0) {
+		report_prefix_push(argv[1]);
+		if (argc > 2)
+			cpi = atol(argv[2]);
+		report(check_pmcr(), "Control register");
+		report(check_cycles_increase(),
+		       "Monotonically increasing cycle count");
+		report(check_cpi(cpi), "Cycle/instruction ratio");
+		pmccntr64_test();
+		report_prefix_pop();
+	} else {
+		report_abort("Unknown sub-test '%s'", argv[1]);
+	}
 
 	return report_summary();
 }
diff --git a/arm/unittests.cfg b/arm/unittests.cfg
index 017958d..fe6515c 100644
--- a/arm/unittests.cfg
+++ b/arm/unittests.cfg
@@ -61,21 +61,22 @@ file = pci-test.flat
 groups = pci
 
 # Test PMU support
-[pmu]
+[pmu-cycle-counter]
 file = pmu.flat
 groups = pmu
+extra_params = -append 'cycle-counter 0'
 
 # Test PMU support (TCG) with -icount IPC=1
 #[pmu-tcg-icount-1]
 #file = pmu.flat
-#extra_params = -icount 0 -append '1'
+#extra_params = -icount 0 -append 'cycle-counter 1'
 #groups = pmu
 #accel = tcg
 
 # Test PMU support (TCG) with -icount IPC=256
 #[pmu-tcg-icount-256]
 #file = pmu.flat
-#extra_params = -icount 8 -append '256'
+#extra_params = -icount 8 -append 'cycle-counter 256'
 #groups = pmu
 #accel = tcg
 
-- 
2.20.1

[kvm-unit-tests PATCH v3 02/12] arm: pmu: Let pmu tests take a sub-test parameter

[Eric Auger]

As we intend to introduce more PMU tests, let's add
a sub-test parameter that will allow to categorize
them. Existing tests are in the cycle-counter category.

Signed-off-by: Eric Auger <eric.auger@redhat.com>
Reviewed-by: Andre Przywara <andre.przywara@arm.com>

---

v2 -> v3:
- added report_prefix_pop()
---
 arm/pmu.c         | 25 ++++++++++++++++---------
 arm/unittests.cfg |  7 ++++---
 2 files changed, 20 insertions(+), 12 deletions(-)

diff --git a/arm/pmu.c b/arm/pmu.c
index d5a03a6..0122f0a 100644
--- a/arm/pmu.c
+++ b/arm/pmu.c
@@ -287,22 +287,29 @@ int main(int argc, char *argv[])
 {
 	int cpi = 0;
 
-	if (argc > 1)
-		cpi = atol(argv[1]);
-
 	if (!pmu_probe()) {
 		printf("No PMU found, test skipped...\n");
 		return report_summary();
 	}
 
+	if (argc < 2)
+		report_abort("no test specified");
+
 	report_prefix_push("pmu");
 
-	report(check_pmcr(), "Control register");
-	report(check_cycles_increase(),
-	       "Monotonically increasing cycle count");
-	report(check_cpi(cpi), "Cycle/instruction ratio");
-
-	pmccntr64_test();
+	if (strcmp(argv[1], "cycle-counter") == 0) {
+		report_prefix_push(argv[1]);
+		if (argc > 2)
+			cpi = atol(argv[2]);
+		report(check_pmcr(), "Control register");
+		report(check_cycles_increase(),
+		       "Monotonically increasing cycle count");
+		report(check_cpi(cpi), "Cycle/instruction ratio");
+		pmccntr64_test();
+		report_prefix_pop();
+	} else {
+		report_abort("Unknown sub-test '%s'", argv[1]);
+	}
 
 	return report_summary();
 }
diff --git a/arm/unittests.cfg b/arm/unittests.cfg
index 017958d..fe6515c 100644
--- a/arm/unittests.cfg
+++ b/arm/unittests.cfg
@@ -61,21 +61,22 @@ file = pci-test.flat
 groups = pci
 
 # Test PMU support
-[pmu]
+[pmu-cycle-counter]
 file = pmu.flat
 groups = pmu
+extra_params = -append 'cycle-counter 0'
 
 # Test PMU support (TCG) with -icount IPC=1
 #[pmu-tcg-icount-1]
 #file = pmu.flat
-#extra_params = -icount 0 -append '1'
+#extra_params = -icount 0 -append 'cycle-counter 1'
 #groups = pmu
 #accel = tcg
 
 # Test PMU support (TCG) with -icount IPC=256
 #[pmu-tcg-icount-256]
 #file = pmu.flat
-#extra_params = -icount 8 -append '256'
+#extra_params = -icount 8 -append 'cycle-counter 256'
 #groups = pmu
 #accel = tcg
 
-- 
2.20.1

_______________________________________________
kvmarm mailing list
kvmarm@lists.cs.columbia.edu
https://lists.cs.columbia.edu/mailman/listinfo/kvmarm

[kvm-unit-tests PATCH v3 03/12] arm: pmu: Don't check PMCR.IMP anymore

[Eric Auger]

check_pmcr() checks the IMP field is different than 0.
However A zero IMP field is permitted by the architecture,
meaning the MIDR_EL1 should be looked at instead. This
causes TCG to fail this test on '-cpu max' because in
that case PMCR.IMP is set equal to MIDR_EL1.Implementer
which is 0.

So let's remove the check_pmcr() test and just print PMCR
info in the pmu_probe() function.

Signed-off-by: Eric Auger <eric.auger@redhat.com>
Reported-by: Peter Maydell <peter.maydell@linaro.org>
---
 arm/pmu.c | 39 ++++++++++++++-------------------------
 1 file changed, 14 insertions(+), 25 deletions(-)

diff --git a/arm/pmu.c b/arm/pmu.c
index 0122f0a..44f3543 100644
--- a/arm/pmu.c
+++ b/arm/pmu.c
@@ -134,29 +134,6 @@ static inline void precise_instrs_loop(int loop, uint32_t pmcr)
 }
 #endif
 
-/*
- * As a simple sanity check on the PMCR_EL0, ensure the implementer field isn't
- * null. Also print out a couple other interesting fields for diagnostic
- * purposes. For example, as of fall 2016, QEMU TCG mode doesn't implement
- * event counters and therefore reports zero event counters, but hopefully
- * support for at least the instructions event will be added in the future and
- * the reported number of event counters will become nonzero.
- */
-static bool check_pmcr(void)
-{
-	uint32_t pmcr;
-
-	pmcr = get_pmcr();
-
-	report_info("PMU implementer/ID code/counters: %#x(\"%c\")/%#x/%d",
-		    (pmcr >> PMU_PMCR_IMP_SHIFT) & PMU_PMCR_IMP_MASK,
-		    ((pmcr >> PMU_PMCR_IMP_SHIFT) & PMU_PMCR_IMP_MASK) ? : ' ',
-		    (pmcr >> PMU_PMCR_ID_SHIFT) & PMU_PMCR_ID_MASK,
-		    (pmcr >> PMU_PMCR_N_SHIFT) & PMU_PMCR_N_MASK);
-
-	return ((pmcr >> PMU_PMCR_IMP_SHIFT) & PMU_PMCR_IMP_MASK) != 0;
-}
-
 /*
  * Ensure that the cycle counter progresses between back-to-back reads.
  */
@@ -278,9 +255,22 @@ static void pmccntr64_test(void)
 /* Return FALSE if no PMU found, otherwise return TRUE */
 static bool pmu_probe(void)
 {
+	uint32_t pmcr;
+
 	pmu_version = get_pmu_version();
+	if (pmu_version == 0 || pmu_version == 0xf)
+		return false;
+
 	report_info("PMU version: %d", pmu_version);
-	return pmu_version != 0 && pmu_version != 0xf;
+
+	pmcr = get_pmcr();
+	report_info("PMU implementer/ID code/counters: %#x(\"%c\")/%#x/%d",
+		    (pmcr >> PMU_PMCR_IMP_SHIFT) & PMU_PMCR_IMP_MASK,
+		    ((pmcr >> PMU_PMCR_IMP_SHIFT) & PMU_PMCR_IMP_MASK) ? : ' ',
+		    (pmcr >> PMU_PMCR_ID_SHIFT) & PMU_PMCR_ID_MASK,
+		    (pmcr >> PMU_PMCR_N_SHIFT) & PMU_PMCR_N_MASK);
+
+	return true;
 }
 
 int main(int argc, char *argv[])
@@ -301,7 +291,6 @@ int main(int argc, char *argv[])
 		report_prefix_push(argv[1]);
 		if (argc > 2)
 			cpi = atol(argv[2]);
-		report(check_pmcr(), "Control register");
 		report(check_cycles_increase(),
 		       "Monotonically increasing cycle count");
 		report(check_cpi(cpi), "Cycle/instruction ratio");
-- 
2.20.1

[kvm-unit-tests PATCH v3 03/12] arm: pmu: Don't check PMCR.IMP anymore

[Eric Auger]

check_pmcr() checks the IMP field is different than 0.
However A zero IMP field is permitted by the architecture,
meaning the MIDR_EL1 should be looked at instead. This
causes TCG to fail this test on '-cpu max' because in
that case PMCR.IMP is set equal to MIDR_EL1.Implementer
which is 0.

So let's remove the check_pmcr() test and just print PMCR
info in the pmu_probe() function.

Signed-off-by: Eric Auger <eric.auger@redhat.com>
Reported-by: Peter Maydell <peter.maydell@linaro.org>
---
 arm/pmu.c | 39 ++++++++++++++-------------------------
 1 file changed, 14 insertions(+), 25 deletions(-)

diff --git a/arm/pmu.c b/arm/pmu.c
index 0122f0a..44f3543 100644
--- a/arm/pmu.c
+++ b/arm/pmu.c
@@ -134,29 +134,6 @@ static inline void precise_instrs_loop(int loop, uint32_t pmcr)
 }
 #endif
 
-/*
- * As a simple sanity check on the PMCR_EL0, ensure the implementer field isn't
- * null. Also print out a couple other interesting fields for diagnostic
- * purposes. For example, as of fall 2016, QEMU TCG mode doesn't implement
- * event counters and therefore reports zero event counters, but hopefully
- * support for at least the instructions event will be added in the future and
- * the reported number of event counters will become nonzero.
- */
-static bool check_pmcr(void)
-{
-	uint32_t pmcr;
-
-	pmcr = get_pmcr();
-
-	report_info("PMU implementer/ID code/counters: %#x(\"%c\")/%#x/%d",
-		    (pmcr >> PMU_PMCR_IMP_SHIFT) & PMU_PMCR_IMP_MASK,
-		    ((pmcr >> PMU_PMCR_IMP_SHIFT) & PMU_PMCR_IMP_MASK) ? : ' ',
-		    (pmcr >> PMU_PMCR_ID_SHIFT) & PMU_PMCR_ID_MASK,
-		    (pmcr >> PMU_PMCR_N_SHIFT) & PMU_PMCR_N_MASK);
-
-	return ((pmcr >> PMU_PMCR_IMP_SHIFT) & PMU_PMCR_IMP_MASK) != 0;
-}
-
 /*
  * Ensure that the cycle counter progresses between back-to-back reads.
  */
@@ -278,9 +255,22 @@ static void pmccntr64_test(void)
 /* Return FALSE if no PMU found, otherwise return TRUE */
 static bool pmu_probe(void)
 {
+	uint32_t pmcr;
+
 	pmu_version = get_pmu_version();
+	if (pmu_version == 0 || pmu_version == 0xf)
+		return false;
+
 	report_info("PMU version: %d", pmu_version);
-	return pmu_version != 0 && pmu_version != 0xf;
+
+	pmcr = get_pmcr();
+	report_info("PMU implementer/ID code/counters: %#x(\"%c\")/%#x/%d",
+		    (pmcr >> PMU_PMCR_IMP_SHIFT) & PMU_PMCR_IMP_MASK,
+		    ((pmcr >> PMU_PMCR_IMP_SHIFT) & PMU_PMCR_IMP_MASK) ? : ' ',
+		    (pmcr >> PMU_PMCR_ID_SHIFT) & PMU_PMCR_ID_MASK,
+		    (pmcr >> PMU_PMCR_N_SHIFT) & PMU_PMCR_N_MASK);
+
+	return true;
 }
 
 int main(int argc, char *argv[])
@@ -301,7 +291,6 @@ int main(int argc, char *argv[])
 		report_prefix_push(argv[1]);
 		if (argc > 2)
 			cpi = atol(argv[2]);
-		report(check_pmcr(), "Control register");
 		report(check_cycles_increase(),
 		       "Monotonically increasing cycle count");
 		report(check_cpi(cpi), "Cycle/instruction ratio");
-- 
2.20.1

[kvm-unit-tests PATCH v3 03/12] arm: pmu: Don't check PMCR.IMP anymore

[Eric Auger]

check_pmcr() checks the IMP field is different than 0.
However A zero IMP field is permitted by the architecture,
meaning the MIDR_EL1 should be looked at instead. This
causes TCG to fail this test on '-cpu max' because in
that case PMCR.IMP is set equal to MIDR_EL1.Implementer
which is 0.

So let's remove the check_pmcr() test and just print PMCR
info in the pmu_probe() function.

Signed-off-by: Eric Auger <eric.auger@redhat.com>
Reported-by: Peter Maydell <peter.maydell@linaro.org>
---
 arm/pmu.c | 39 ++++++++++++++-------------------------
 1 file changed, 14 insertions(+), 25 deletions(-)

diff --git a/arm/pmu.c b/arm/pmu.c
index 0122f0a..44f3543 100644
--- a/arm/pmu.c
+++ b/arm/pmu.c
@@ -134,29 +134,6 @@ static inline void precise_instrs_loop(int loop, uint32_t pmcr)
 }
 #endif
 
-/*
- * As a simple sanity check on the PMCR_EL0, ensure the implementer field isn't
- * null. Also print out a couple other interesting fields for diagnostic
- * purposes. For example, as of fall 2016, QEMU TCG mode doesn't implement
- * event counters and therefore reports zero event counters, but hopefully
- * support for at least the instructions event will be added in the future and
- * the reported number of event counters will become nonzero.
- */
-static bool check_pmcr(void)
-{
-	uint32_t pmcr;
-
-	pmcr = get_pmcr();
-
-	report_info("PMU implementer/ID code/counters: %#x(\"%c\")/%#x/%d",
-		    (pmcr >> PMU_PMCR_IMP_SHIFT) & PMU_PMCR_IMP_MASK,
-		    ((pmcr >> PMU_PMCR_IMP_SHIFT) & PMU_PMCR_IMP_MASK) ? : ' ',
-		    (pmcr >> PMU_PMCR_ID_SHIFT) & PMU_PMCR_ID_MASK,
-		    (pmcr >> PMU_PMCR_N_SHIFT) & PMU_PMCR_N_MASK);
-
-	return ((pmcr >> PMU_PMCR_IMP_SHIFT) & PMU_PMCR_IMP_MASK) != 0;
-}
-
 /*
  * Ensure that the cycle counter progresses between back-to-back reads.
  */
@@ -278,9 +255,22 @@ static void pmccntr64_test(void)
 /* Return FALSE if no PMU found, otherwise return TRUE */
 static bool pmu_probe(void)
 {
+	uint32_t pmcr;
+
 	pmu_version = get_pmu_version();
+	if (pmu_version == 0 || pmu_version == 0xf)
+		return false;
+
 	report_info("PMU version: %d", pmu_version);
-	return pmu_version != 0 && pmu_version != 0xf;
+
+	pmcr = get_pmcr();
+	report_info("PMU implementer/ID code/counters: %#x(\"%c\")/%#x/%d",
+		    (pmcr >> PMU_PMCR_IMP_SHIFT) & PMU_PMCR_IMP_MASK,
+		    ((pmcr >> PMU_PMCR_IMP_SHIFT) & PMU_PMCR_IMP_MASK) ? : ' ',
+		    (pmcr >> PMU_PMCR_ID_SHIFT) & PMU_PMCR_ID_MASK,
+		    (pmcr >> PMU_PMCR_N_SHIFT) & PMU_PMCR_N_MASK);
+
+	return true;
 }
 
 int main(int argc, char *argv[])
@@ -301,7 +291,6 @@ int main(int argc, char *argv[])
 		report_prefix_push(argv[1]);
 		if (argc > 2)
 			cpi = atol(argv[2]);
-		report(check_pmcr(), "Control register");
 		report(check_cycles_increase(),
 		       "Monotonically increasing cycle count");
 		report(check_cpi(cpi), "Cycle/instruction ratio");
-- 
2.20.1

_______________________________________________
kvmarm mailing list
kvmarm@lists.cs.columbia.edu
https://lists.cs.columbia.edu/mailman/listinfo/kvmarm

[kvm-unit-tests PATCH v3 04/12] arm: pmu: Add a pmu struct

[Eric Auger]

This struct aims at storing information potentially used by
all tests such as the pmu version, the read-only part of the
PMCR, the number of implemented event counters, ...

Signed-off-by: Eric Auger <eric.auger@redhat.com>
Reviewed-by: Andre Przywara <andre.przywara@arm.com>

---

v2 -> v3:
- Fix pmcr_ro and add a comment
---
 arm/pmu.c | 29 +++++++++++++++++++++--------
 1 file changed, 21 insertions(+), 8 deletions(-)

diff --git a/arm/pmu.c b/arm/pmu.c
index 44f3543..d827e82 100644
--- a/arm/pmu.c
+++ b/arm/pmu.c
@@ -33,7 +33,14 @@
 
 #define NR_SAMPLES 10
 
-static unsigned int pmu_version;
+struct pmu {
+	unsigned int version;
+	unsigned int nb_implemented_counters;
+	uint32_t pmcr_ro;
+};
+
+static struct pmu pmu;
+
 #if defined(__arm__)
 #define ID_DFR0_PERFMON_SHIFT 24
 #define ID_DFR0_PERFMON_MASK  0xf
@@ -242,7 +249,7 @@ static bool check_cpi(int cpi)
 static void pmccntr64_test(void)
 {
 #ifdef __arm__
-	if (pmu_version == 0x3) {
+	if (pmu.version == 0x3) {
 		if (ERRATA(9e3f7a296940)) {
 			write_sysreg(0xdead, PMCCNTR64);
 			report(read_sysreg(PMCCNTR64) == 0xdead, "pmccntr64");
@@ -257,18 +264,24 @@ static bool pmu_probe(void)
 {
 	uint32_t pmcr;
 
-	pmu_version = get_pmu_version();
-	if (pmu_version == 0 || pmu_version == 0xf)
+	pmu.version = get_pmu_version();
+	if (pmu.version == 0 || pmu.version == 0xf)
 		return false;
 
-	report_info("PMU version: %d", pmu_version);
+	report_info("PMU version: %d", pmu.version);
 
 	pmcr = get_pmcr();
-	report_info("PMU implementer/ID code/counters: %#x(\"%c\")/%#x/%d",
+	report_info("PMU implementer/ID code: %#x(\"%c\")/%#x",
 		    (pmcr >> PMU_PMCR_IMP_SHIFT) & PMU_PMCR_IMP_MASK,
 		    ((pmcr >> PMU_PMCR_IMP_SHIFT) & PMU_PMCR_IMP_MASK) ? : ' ',
-		    (pmcr >> PMU_PMCR_ID_SHIFT) & PMU_PMCR_ID_MASK,
-		    (pmcr >> PMU_PMCR_N_SHIFT) & PMU_PMCR_N_MASK);
+		    (pmcr >> PMU_PMCR_ID_SHIFT) & PMU_PMCR_ID_MASK);
+
+	/* store read-only and RES0 fields of the PMCR bottom-half*/
+	pmu.pmcr_ro = pmcr & 0xFFFFFF00;
+	pmu.nb_implemented_counters =
+		(pmcr >> PMU_PMCR_N_SHIFT) & PMU_PMCR_N_MASK;
+	report_info("Implements %d event counters",
+		    pmu.nb_implemented_counters);
 
 	return true;
 }
-- 
2.20.1

[kvm-unit-tests PATCH v3 04/12] arm: pmu: Add a pmu struct

[Eric Auger]

This struct aims at storing information potentially used by
all tests such as the pmu version, the read-only part of the
PMCR, the number of implemented event counters, ...

Signed-off-by: Eric Auger <eric.auger@redhat.com>
Reviewed-by: Andre Przywara <andre.przywara@arm.com>

---

v2 -> v3:
- Fix pmcr_ro and add a comment
---
 arm/pmu.c | 29 +++++++++++++++++++++--------
 1 file changed, 21 insertions(+), 8 deletions(-)

diff --git a/arm/pmu.c b/arm/pmu.c
index 44f3543..d827e82 100644
--- a/arm/pmu.c
+++ b/arm/pmu.c
@@ -33,7 +33,14 @@
 
 #define NR_SAMPLES 10
 
-static unsigned int pmu_version;
+struct pmu {
+	unsigned int version;
+	unsigned int nb_implemented_counters;
+	uint32_t pmcr_ro;
+};
+
+static struct pmu pmu;
+
 #if defined(__arm__)
 #define ID_DFR0_PERFMON_SHIFT 24
 #define ID_DFR0_PERFMON_MASK  0xf
@@ -242,7 +249,7 @@ static bool check_cpi(int cpi)
 static void pmccntr64_test(void)
 {
 #ifdef __arm__
-	if (pmu_version == 0x3) {
+	if (pmu.version == 0x3) {
 		if (ERRATA(9e3f7a296940)) {
 			write_sysreg(0xdead, PMCCNTR64);
 			report(read_sysreg(PMCCNTR64) == 0xdead, "pmccntr64");
@@ -257,18 +264,24 @@ static bool pmu_probe(void)
 {
 	uint32_t pmcr;
 
-	pmu_version = get_pmu_version();
-	if (pmu_version == 0 || pmu_version == 0xf)
+	pmu.version = get_pmu_version();
+	if (pmu.version == 0 || pmu.version == 0xf)
 		return false;
 
-	report_info("PMU version: %d", pmu_version);
+	report_info("PMU version: %d", pmu.version);
 
 	pmcr = get_pmcr();
-	report_info("PMU implementer/ID code/counters: %#x(\"%c\")/%#x/%d",
+	report_info("PMU implementer/ID code: %#x(\"%c\")/%#x",
 		    (pmcr >> PMU_PMCR_IMP_SHIFT) & PMU_PMCR_IMP_MASK,
 		    ((pmcr >> PMU_PMCR_IMP_SHIFT) & PMU_PMCR_IMP_MASK) ? : ' ',
-		    (pmcr >> PMU_PMCR_ID_SHIFT) & PMU_PMCR_ID_MASK,
-		    (pmcr >> PMU_PMCR_N_SHIFT) & PMU_PMCR_N_MASK);
+		    (pmcr >> PMU_PMCR_ID_SHIFT) & PMU_PMCR_ID_MASK);
+
+	/* store read-only and RES0 fields of the PMCR bottom-half*/
+	pmu.pmcr_ro = pmcr & 0xFFFFFF00;
+	pmu.nb_implemented_counters =
+		(pmcr >> PMU_PMCR_N_SHIFT) & PMU_PMCR_N_MASK;
+	report_info("Implements %d event counters",
+		    pmu.nb_implemented_counters);
 
 	return true;
 }
-- 
2.20.1

[kvm-unit-tests PATCH v3 04/12] arm: pmu: Add a pmu struct

[Eric Auger]

This struct aims at storing information potentially used by
all tests such as the pmu version, the read-only part of the
PMCR, the number of implemented event counters, ...

Signed-off-by: Eric Auger <eric.auger@redhat.com>
Reviewed-by: Andre Przywara <andre.przywara@arm.com>

---

v2 -> v3:
- Fix pmcr_ro and add a comment
---
 arm/pmu.c | 29 +++++++++++++++++++++--------
 1 file changed, 21 insertions(+), 8 deletions(-)

diff --git a/arm/pmu.c b/arm/pmu.c
index 44f3543..d827e82 100644
--- a/arm/pmu.c
+++ b/arm/pmu.c
@@ -33,7 +33,14 @@
 
 #define NR_SAMPLES 10
 
-static unsigned int pmu_version;
+struct pmu {
+	unsigned int version;
+	unsigned int nb_implemented_counters;
+	uint32_t pmcr_ro;
+};
+
+static struct pmu pmu;
+
 #if defined(__arm__)
 #define ID_DFR0_PERFMON_SHIFT 24
 #define ID_DFR0_PERFMON_MASK  0xf
@@ -242,7 +249,7 @@ static bool check_cpi(int cpi)
 static void pmccntr64_test(void)
 {
 #ifdef __arm__
-	if (pmu_version == 0x3) {
+	if (pmu.version == 0x3) {
 		if (ERRATA(9e3f7a296940)) {
 			write_sysreg(0xdead, PMCCNTR64);
 			report(read_sysreg(PMCCNTR64) == 0xdead, "pmccntr64");
@@ -257,18 +264,24 @@ static bool pmu_probe(void)
 {
 	uint32_t pmcr;
 
-	pmu_version = get_pmu_version();
-	if (pmu_version == 0 || pmu_version == 0xf)
+	pmu.version = get_pmu_version();
+	if (pmu.version == 0 || pmu.version == 0xf)
 		return false;
 
-	report_info("PMU version: %d", pmu_version);
+	report_info("PMU version: %d", pmu.version);
 
 	pmcr = get_pmcr();
-	report_info("PMU implementer/ID code/counters: %#x(\"%c\")/%#x/%d",
+	report_info("PMU implementer/ID code: %#x(\"%c\")/%#x",
 		    (pmcr >> PMU_PMCR_IMP_SHIFT) & PMU_PMCR_IMP_MASK,
 		    ((pmcr >> PMU_PMCR_IMP_SHIFT) & PMU_PMCR_IMP_MASK) ? : ' ',
-		    (pmcr >> PMU_PMCR_ID_SHIFT) & PMU_PMCR_ID_MASK,
-		    (pmcr >> PMU_PMCR_N_SHIFT) & PMU_PMCR_N_MASK);
+		    (pmcr >> PMU_PMCR_ID_SHIFT) & PMU_PMCR_ID_MASK);
+
+	/* store read-only and RES0 fields of the PMCR bottom-half*/
+	pmu.pmcr_ro = pmcr & 0xFFFFFF00;
+	pmu.nb_implemented_counters =
+		(pmcr >> PMU_PMCR_N_SHIFT) & PMU_PMCR_N_MASK;
+	report_info("Implements %d event counters",
+		    pmu.nb_implemented_counters);
 
 	return true;
 }
-- 
2.20.1

_______________________________________________
kvmarm mailing list
kvmarm@lists.cs.columbia.edu
https://lists.cs.columbia.edu/mailman/listinfo/kvmarm

[kvm-unit-tests PATCH v3 05/12] arm: pmu: Introduce defines for PMU versions

[Eric Auger]

Introduce some defines encoding the different PMU versions.
v3 is encoded differently in 32 and 64 bits.

Signed-off-by: Eric Auger <eric.auger@redhat.com>
---
 arm/pmu.c | 26 +++++++++++++++++++++-----
 1 file changed, 21 insertions(+), 5 deletions(-)

diff --git a/arm/pmu.c b/arm/pmu.c
index d827e82..a04588a 100644
--- a/arm/pmu.c
+++ b/arm/pmu.c
@@ -45,6 +45,15 @@ static struct pmu pmu;
 #define ID_DFR0_PERFMON_SHIFT 24
 #define ID_DFR0_PERFMON_MASK  0xf
 
+#define ID_DFR0_PMU_NOTIMPL	0b0000
+#define ID_DFR0_PMU_V1		0b0001
+#define ID_DFR0_PMU_V2		0b0010
+#define ID_DFR0_PMU_V3		0b0011
+#define ID_DFR0_PMU_V3_8_1	0b0100
+#define ID_DFR0_PMU_V3_8_4	0b0101
+#define ID_DFR0_PMU_V3_8_5	0b0110
+#define ID_DFR0_PMU_IMPDEF	0b1111
+
 #define PMCR         __ACCESS_CP15(c9, 0, c12, 0)
 #define ID_DFR0      __ACCESS_CP15(c0, 0, c1, 2)
 #define PMSELR       __ACCESS_CP15(c9, 0, c12, 5)
@@ -105,6 +114,13 @@ static inline void precise_instrs_loop(int loop, uint32_t pmcr)
 #define ID_AA64DFR0_PERFMON_SHIFT 8
 #define ID_AA64DFR0_PERFMON_MASK  0xf
 
+#define ID_DFR0_PMU_NOTIMPL	0b0000
+#define ID_DFR0_PMU_V3		0b0001
+#define ID_DFR0_PMU_V3_8_1	0b0100
+#define ID_DFR0_PMU_V3_8_4	0b0101
+#define ID_DFR0_PMU_V3_8_5	0b0110
+#define ID_DFR0_PMU_IMPDEF	0b1111
+
 static inline uint32_t get_id_aa64dfr0(void) { return read_sysreg(id_aa64dfr0_el1); }
 static inline uint32_t get_pmcr(void) { return read_sysreg(pmcr_el0); }
 static inline void set_pmcr(uint32_t v) { write_sysreg(v, pmcr_el0); }
@@ -116,7 +132,7 @@ static inline void set_pmccfiltr(uint32_t v) { write_sysreg(v, pmccfiltr_el0); }
 static inline uint8_t get_pmu_version(void)
 {
 	uint8_t ver = (get_id_aa64dfr0() >> ID_AA64DFR0_PERFMON_SHIFT) & ID_AA64DFR0_PERFMON_MASK;
-	return ver == 1 ? 3 : ver;
+	return ver;
 }
 
 /*
@@ -249,7 +265,7 @@ static bool check_cpi(int cpi)
 static void pmccntr64_test(void)
 {
 #ifdef __arm__
-	if (pmu.version == 0x3) {
+	if (pmu.version == ID_DFR0_PMU_V3) {
 		if (ERRATA(9e3f7a296940)) {
 			write_sysreg(0xdead, PMCCNTR64);
 			report(read_sysreg(PMCCNTR64) == 0xdead, "pmccntr64");
@@ -262,13 +278,13 @@ static void pmccntr64_test(void)
 /* Return FALSE if no PMU found, otherwise return TRUE */
 static bool pmu_probe(void)
 {
-	uint32_t pmcr;
+	uint32_t pmcr = get_pmcr();
 
 	pmu.version = get_pmu_version();
-	if (pmu.version == 0 || pmu.version == 0xf)
+	if (pmu.version == ID_DFR0_PMU_NOTIMPL || pmu.version == ID_DFR0_PMU_IMPDEF)
 		return false;
 
-	report_info("PMU version: %d", pmu.version);
+	report_info("PMU version: 0x%x", pmu.version);
 
 	pmcr = get_pmcr();
 	report_info("PMU implementer/ID code: %#x(\"%c\")/%#x",
-- 
2.20.1

[kvm-unit-tests PATCH v3 05/12] arm: pmu: Introduce defines for PMU versions

[Eric Auger]

Introduce some defines encoding the different PMU versions.
v3 is encoded differently in 32 and 64 bits.

Signed-off-by: Eric Auger <eric.auger@redhat.com>
---
 arm/pmu.c | 26 +++++++++++++++++++++-----
 1 file changed, 21 insertions(+), 5 deletions(-)

diff --git a/arm/pmu.c b/arm/pmu.c
index d827e82..a04588a 100644
--- a/arm/pmu.c
+++ b/arm/pmu.c
@@ -45,6 +45,15 @@ static struct pmu pmu;
 #define ID_DFR0_PERFMON_SHIFT 24
 #define ID_DFR0_PERFMON_MASK  0xf
 
+#define ID_DFR0_PMU_NOTIMPL	0b0000
+#define ID_DFR0_PMU_V1		0b0001
+#define ID_DFR0_PMU_V2		0b0010
+#define ID_DFR0_PMU_V3		0b0011
+#define ID_DFR0_PMU_V3_8_1	0b0100
+#define ID_DFR0_PMU_V3_8_4	0b0101
+#define ID_DFR0_PMU_V3_8_5	0b0110
+#define ID_DFR0_PMU_IMPDEF	0b1111
+
 #define PMCR         __ACCESS_CP15(c9, 0, c12, 0)
 #define ID_DFR0      __ACCESS_CP15(c0, 0, c1, 2)
 #define PMSELR       __ACCESS_CP15(c9, 0, c12, 5)
@@ -105,6 +114,13 @@ static inline void precise_instrs_loop(int loop, uint32_t pmcr)
 #define ID_AA64DFR0_PERFMON_SHIFT 8
 #define ID_AA64DFR0_PERFMON_MASK  0xf
 
+#define ID_DFR0_PMU_NOTIMPL	0b0000
+#define ID_DFR0_PMU_V3		0b0001
+#define ID_DFR0_PMU_V3_8_1	0b0100
+#define ID_DFR0_PMU_V3_8_4	0b0101
+#define ID_DFR0_PMU_V3_8_5	0b0110
+#define ID_DFR0_PMU_IMPDEF	0b1111
+
 static inline uint32_t get_id_aa64dfr0(void) { return read_sysreg(id_aa64dfr0_el1); }
 static inline uint32_t get_pmcr(void) { return read_sysreg(pmcr_el0); }
 static inline void set_pmcr(uint32_t v) { write_sysreg(v, pmcr_el0); }
@@ -116,7 +132,7 @@ static inline void set_pmccfiltr(uint32_t v) { write_sysreg(v, pmccfiltr_el0); }
 static inline uint8_t get_pmu_version(void)
 {
 	uint8_t ver = (get_id_aa64dfr0() >> ID_AA64DFR0_PERFMON_SHIFT) & ID_AA64DFR0_PERFMON_MASK;
-	return ver == 1 ? 3 : ver;
+	return ver;
 }
 
 /*
@@ -249,7 +265,7 @@ static bool check_cpi(int cpi)
 static void pmccntr64_test(void)
 {
 #ifdef __arm__
-	if (pmu.version == 0x3) {
+	if (pmu.version == ID_DFR0_PMU_V3) {
 		if (ERRATA(9e3f7a296940)) {
 			write_sysreg(0xdead, PMCCNTR64);
 			report(read_sysreg(PMCCNTR64) == 0xdead, "pmccntr64");
@@ -262,13 +278,13 @@ static void pmccntr64_test(void)
 /* Return FALSE if no PMU found, otherwise return TRUE */
 static bool pmu_probe(void)
 {
-	uint32_t pmcr;
+	uint32_t pmcr = get_pmcr();
 
 	pmu.version = get_pmu_version();
-	if (pmu.version == 0 || pmu.version == 0xf)
+	if (pmu.version == ID_DFR0_PMU_NOTIMPL || pmu.version == ID_DFR0_PMU_IMPDEF)
 		return false;
 
-	report_info("PMU version: %d", pmu.version);
+	report_info("PMU version: 0x%x", pmu.version);
 
 	pmcr = get_pmcr();
 	report_info("PMU implementer/ID code: %#x(\"%c\")/%#x",
-- 
2.20.1

[kvm-unit-tests PATCH v3 05/12] arm: pmu: Introduce defines for PMU versions

[Eric Auger]

Introduce some defines encoding the different PMU versions.
v3 is encoded differently in 32 and 64 bits.

Signed-off-by: Eric Auger <eric.auger@redhat.com>
---
 arm/pmu.c | 26 +++++++++++++++++++++-----
 1 file changed, 21 insertions(+), 5 deletions(-)

diff --git a/arm/pmu.c b/arm/pmu.c
index d827e82..a04588a 100644
--- a/arm/pmu.c
+++ b/arm/pmu.c
@@ -45,6 +45,15 @@ static struct pmu pmu;
 #define ID_DFR0_PERFMON_SHIFT 24
 #define ID_DFR0_PERFMON_MASK  0xf
 
+#define ID_DFR0_PMU_NOTIMPL	0b0000
+#define ID_DFR0_PMU_V1		0b0001
+#define ID_DFR0_PMU_V2		0b0010
+#define ID_DFR0_PMU_V3		0b0011
+#define ID_DFR0_PMU_V3_8_1	0b0100
+#define ID_DFR0_PMU_V3_8_4	0b0101
+#define ID_DFR0_PMU_V3_8_5	0b0110
+#define ID_DFR0_PMU_IMPDEF	0b1111
+
 #define PMCR         __ACCESS_CP15(c9, 0, c12, 0)
 #define ID_DFR0      __ACCESS_CP15(c0, 0, c1, 2)
 #define PMSELR       __ACCESS_CP15(c9, 0, c12, 5)
@@ -105,6 +114,13 @@ static inline void precise_instrs_loop(int loop, uint32_t pmcr)
 #define ID_AA64DFR0_PERFMON_SHIFT 8
 #define ID_AA64DFR0_PERFMON_MASK  0xf
 
+#define ID_DFR0_PMU_NOTIMPL	0b0000
+#define ID_DFR0_PMU_V3		0b0001
+#define ID_DFR0_PMU_V3_8_1	0b0100
+#define ID_DFR0_PMU_V3_8_4	0b0101
+#define ID_DFR0_PMU_V3_8_5	0b0110
+#define ID_DFR0_PMU_IMPDEF	0b1111
+
 static inline uint32_t get_id_aa64dfr0(void) { return read_sysreg(id_aa64dfr0_el1); }
 static inline uint32_t get_pmcr(void) { return read_sysreg(pmcr_el0); }
 static inline void set_pmcr(uint32_t v) { write_sysreg(v, pmcr_el0); }
@@ -116,7 +132,7 @@ static inline void set_pmccfiltr(uint32_t v) { write_sysreg(v, pmccfiltr_el0); }
 static inline uint8_t get_pmu_version(void)
 {
 	uint8_t ver = (get_id_aa64dfr0() >> ID_AA64DFR0_PERFMON_SHIFT) & ID_AA64DFR0_PERFMON_MASK;
-	return ver == 1 ? 3 : ver;
+	return ver;
 }
 
 /*
@@ -249,7 +265,7 @@ static bool check_cpi(int cpi)
 static void pmccntr64_test(void)
 {
 #ifdef __arm__
-	if (pmu.version == 0x3) {
+	if (pmu.version == ID_DFR0_PMU_V3) {
 		if (ERRATA(9e3f7a296940)) {
 			write_sysreg(0xdead, PMCCNTR64);
 			report(read_sysreg(PMCCNTR64) == 0xdead, "pmccntr64");
@@ -262,13 +278,13 @@ static void pmccntr64_test(void)
 /* Return FALSE if no PMU found, otherwise return TRUE */
 static bool pmu_probe(void)
 {
-	uint32_t pmcr;
+	uint32_t pmcr = get_pmcr();
 
 	pmu.version = get_pmu_version();
-	if (pmu.version == 0 || pmu.version == 0xf)
+	if (pmu.version == ID_DFR0_PMU_NOTIMPL || pmu.version == ID_DFR0_PMU_IMPDEF)
 		return false;
 
-	report_info("PMU version: %d", pmu.version);
+	report_info("PMU version: 0x%x", pmu.version);
 
 	pmcr = get_pmcr();
 	report_info("PMU implementer/ID code: %#x(\"%c\")/%#x",
-- 
2.20.1

_______________________________________________
kvmarm mailing list
kvmarm@lists.cs.columbia.edu
https://lists.cs.columbia.edu/mailman/listinfo/kvmarm

[kvm-unit-tests PATCH v3 06/12] arm: pmu: Check Required Event Support

[Eric Auger]

If event counters are implemented check the common events
required by the PMUv3 are implemented.

Some are unconditionally required (SW_INCR, CPU_CYCLES,
either INST_RETIRED or INST_SPEC). Some others only are
required if the implementation implements some other features.

Check those wich are unconditionally required.

This test currently fails on TCG as neither INST_RETIRED
or INST_SPEC are supported.

Signed-off-by: Eric Auger <eric.auger@redhat.com>

---

v2 -> v3:
- use 0x%x instead %d in trace
- pmu.version >= ID_DFR0_PMU_v3_8_1
- added prefix pop
- assert instead of abort, inverse assert and test
- add defines for used events and common events
- given the changes I did not apply Andre's R-b
- introduce and use upper_32_bits()/lower_32_bits()
- added pmu prefix to the test name

v1 -> v2:
- fix is_event_supported()
- fix boolean condition for PMU v4
- fix PMCEID0 definition

RFC ->v1:
- add a comment to explain the PMCEID0/1 splits
---
 arm/pmu.c         | 77 +++++++++++++++++++++++++++++++++++++++++++++++
 arm/unittests.cfg |  6 ++++
 lib/bitops.h      |  3 ++
 3 files changed, 86 insertions(+)

diff --git a/arm/pmu.c b/arm/pmu.c
index a04588a..8c49e50 100644
--- a/arm/pmu.c
+++ b/arm/pmu.c
@@ -18,6 +18,7 @@
 #include "asm/barrier.h"
 #include "asm/sysreg.h"
 #include "asm/processor.h"
+#include <bitops.h>
 
 #define PMU_PMCR_E         (1 << 0)
 #define PMU_PMCR_C         (1 << 2)
@@ -33,6 +34,19 @@
 
 #define NR_SAMPLES 10
 
+/* Some PMU events */
+#define SW_INCR			0x0
+#define INST_RETIRED		0x8
+#define CPU_CYCLES		0x11
+#define INST_PREC		0x1B
+#define STALL_FRONTEND		0x23
+#define STALL_BACKEND		0x24
+
+#define COMMON_EVENTS_LOW	0x0
+#define COMMON_EVENTS_HIGH	0x3F
+#define EXT_COMMON_EVENTS_LOW	0x4000
+#define EXT_COMMON_EVENTS_HIGH	0x403F
+
 struct pmu {
 	unsigned int version;
 	unsigned int nb_implemented_counters;
@@ -110,6 +124,10 @@ static inline void precise_instrs_loop(int loop, uint32_t pmcr)
 	: [pmcr] "r" (pmcr), [z] "r" (0)
 	: "cc");
 }
+
+/* event counter tests only implemented for aarch64 */
+static void test_event_introspection(void) {}
+
 #elif defined(__aarch64__)
 #define ID_AA64DFR0_PERFMON_SHIFT 8
 #define ID_AA64DFR0_PERFMON_MASK  0xf
@@ -155,6 +173,61 @@ static inline void precise_instrs_loop(int loop, uint32_t pmcr)
 	: [pmcr] "r" (pmcr)
 	: "cc");
 }
+
+#define PMCEID1_EL0 sys_reg(3, 3, 9, 12, 7)
+
+static bool is_event_supported(uint32_t n, bool warn)
+{
+	uint64_t pmceid0 = read_sysreg(pmceid0_el0);
+	uint64_t pmceid1 = read_sysreg_s(PMCEID1_EL0);
+	bool supported;
+	uint64_t reg;
+
+	/*
+	 * The low 32-bits of PMCEID0/1 respectively describe
+	 * event support for events 0-31/32-63. Their High
+	 * 32-bits describe support for extended events
+	 * starting at 0x4000, using the same split.
+	 */
+	assert((n >= COMMON_EVENTS_LOW  && n <= COMMON_EVENTS_HIGH) ||
+	       (n >= EXT_COMMON_EVENTS_LOW && n <= EXT_COMMON_EVENTS_HIGH));
+
+	if (n <= COMMON_EVENTS_HIGH)
+		reg = lower_32_bits(pmceid0) | ((u64)lower_32_bits(pmceid1) << 32);
+	else
+		reg = upper_32_bits(pmceid0) | ((u64)upper_32_bits(pmceid1) << 32);
+
+	supported =  reg & (1UL << (n & 0x3F));
+
+	if (!supported && warn)
+		report_info("event 0x%x is not supported", n);
+	return supported;
+}
+
+static void test_event_introspection(void)
+{
+	bool required_events;
+
+	if (!pmu.nb_implemented_counters) {
+		report_skip("No event counter, skip ...");
+		return;
+	}
+
+	/* PMUv3 requires an implementation includes some common events */
+	required_events = is_event_supported(SW_INCR, true) &&
+			  is_event_supported(CPU_CYCLES, true) &&
+			  (is_event_supported(INST_RETIRED, true) ||
+			   is_event_supported(INST_PREC, true));
+
+	if (pmu.version >= ID_DFR0_PMU_V3_8_1) {
+		required_events = required_events &&
+				  is_event_supported(STALL_FRONTEND, true) &&
+				  is_event_supported(STALL_BACKEND, true);
+	}
+
+	report(required_events, "Check required events are implemented");
+}
+
 #endif
 
 /*
@@ -325,6 +398,10 @@ int main(int argc, char *argv[])
 		report(check_cpi(cpi), "Cycle/instruction ratio");
 		pmccntr64_test();
 		report_prefix_pop();
+	} else if (strcmp(argv[1], "pmu-event-introspection") == 0) {
+		report_prefix_push(argv[1]);
+		test_event_introspection();
+		report_prefix_pop();
 	} else {
 		report_abort("Unknown sub-test '%s'", argv[1]);
 	}
diff --git a/arm/unittests.cfg b/arm/unittests.cfg
index fe6515c..f993548 100644
--- a/arm/unittests.cfg
+++ b/arm/unittests.cfg
@@ -66,6 +66,12 @@ file = pmu.flat
 groups = pmu
 extra_params = -append 'cycle-counter 0'
 
+[pmu-event-introspection]
+file = pmu.flat
+groups = pmu
+arch = arm64
+extra_params = -append 'pmu-event-introspection'
+
 # Test PMU support (TCG) with -icount IPC=1
 #[pmu-tcg-icount-1]
 #file = pmu.flat
diff --git a/lib/bitops.h b/lib/bitops.h
index 636064c..b310a22 100644
--- a/lib/bitops.h
+++ b/lib/bitops.h
@@ -33,6 +33,9 @@
 #define GENMASK_ULL(h, l) \
 	(((~0ULL) << (l)) & (~0ULL >> (BITS_PER_LONG_LONG - 1 - (h))))
 
+#define upper_32_bits(n) ((u32)(((n) >> 16) >> 16))
+#define lower_32_bits(n) ((u32)(n))
+
 #ifndef HAVE_BUILTIN_FLS
 static inline unsigned long fls(unsigned long word)
 {
-- 
2.20.1

[kvm-unit-tests PATCH v3 06/12] arm: pmu: Check Required Event Support

[Eric Auger]

If event counters are implemented check the common events
required by the PMUv3 are implemented.

Some are unconditionally required (SW_INCR, CPU_CYCLES,
either INST_RETIRED or INST_SPEC). Some others only are
required if the implementation implements some other features.

Check those wich are unconditionally required.

This test currently fails on TCG as neither INST_RETIRED
or INST_SPEC are supported.

Signed-off-by: Eric Auger <eric.auger@redhat.com>

---

v2 -> v3:
- use 0x%x instead %d in trace
- pmu.version >= ID_DFR0_PMU_v3_8_1
- added prefix pop
- assert instead of abort, inverse assert and test
- add defines for used events and common events
- given the changes I did not apply Andre's R-b
- introduce and use upper_32_bits()/lower_32_bits()
- added pmu prefix to the test name

v1 -> v2:
- fix is_event_supported()
- fix boolean condition for PMU v4
- fix PMCEID0 definition

RFC ->v1:
- add a comment to explain the PMCEID0/1 splits
---
 arm/pmu.c         | 77 +++++++++++++++++++++++++++++++++++++++++++++++
 arm/unittests.cfg |  6 ++++
 lib/bitops.h      |  3 ++
 3 files changed, 86 insertions(+)

diff --git a/arm/pmu.c b/arm/pmu.c
index a04588a..8c49e50 100644
--- a/arm/pmu.c
+++ b/arm/pmu.c
@@ -18,6 +18,7 @@
 #include "asm/barrier.h"
 #include "asm/sysreg.h"
 #include "asm/processor.h"
+#include <bitops.h>
 
 #define PMU_PMCR_E         (1 << 0)
 #define PMU_PMCR_C         (1 << 2)
@@ -33,6 +34,19 @@
 
 #define NR_SAMPLES 10
 
+/* Some PMU events */
+#define SW_INCR			0x0
+#define INST_RETIRED		0x8
+#define CPU_CYCLES		0x11
+#define INST_PREC		0x1B
+#define STALL_FRONTEND		0x23
+#define STALL_BACKEND		0x24
+
+#define COMMON_EVENTS_LOW	0x0
+#define COMMON_EVENTS_HIGH	0x3F
+#define EXT_COMMON_EVENTS_LOW	0x4000
+#define EXT_COMMON_EVENTS_HIGH	0x403F
+
 struct pmu {
 	unsigned int version;
 	unsigned int nb_implemented_counters;
@@ -110,6 +124,10 @@ static inline void precise_instrs_loop(int loop, uint32_t pmcr)
 	: [pmcr] "r" (pmcr), [z] "r" (0)
 	: "cc");
 }
+
+/* event counter tests only implemented for aarch64 */
+static void test_event_introspection(void) {}
+
 #elif defined(__aarch64__)
 #define ID_AA64DFR0_PERFMON_SHIFT 8
 #define ID_AA64DFR0_PERFMON_MASK  0xf
@@ -155,6 +173,61 @@ static inline void precise_instrs_loop(int loop, uint32_t pmcr)
 	: [pmcr] "r" (pmcr)
 	: "cc");
 }
+
+#define PMCEID1_EL0 sys_reg(3, 3, 9, 12, 7)
+
+static bool is_event_supported(uint32_t n, bool warn)
+{
+	uint64_t pmceid0 = read_sysreg(pmceid0_el0);
+	uint64_t pmceid1 = read_sysreg_s(PMCEID1_EL0);
+	bool supported;
+	uint64_t reg;
+
+	/*
+	 * The low 32-bits of PMCEID0/1 respectively describe
+	 * event support for events 0-31/32-63. Their High
+	 * 32-bits describe support for extended events
+	 * starting at 0x4000, using the same split.
+	 */
+	assert((n >= COMMON_EVENTS_LOW  && n <= COMMON_EVENTS_HIGH) ||
+	       (n >= EXT_COMMON_EVENTS_LOW && n <= EXT_COMMON_EVENTS_HIGH));
+
+	if (n <= COMMON_EVENTS_HIGH)
+		reg = lower_32_bits(pmceid0) | ((u64)lower_32_bits(pmceid1) << 32);
+	else
+		reg = upper_32_bits(pmceid0) | ((u64)upper_32_bits(pmceid1) << 32);
+
+	supported =  reg & (1UL << (n & 0x3F));
+
+	if (!supported && warn)
+		report_info("event 0x%x is not supported", n);
+	return supported;
+}
+
+static void test_event_introspection(void)
+{
+	bool required_events;
+
+	if (!pmu.nb_implemented_counters) {
+		report_skip("No event counter, skip ...");
+		return;
+	}
+
+	/* PMUv3 requires an implementation includes some common events */
+	required_events = is_event_supported(SW_INCR, true) &&
+			  is_event_supported(CPU_CYCLES, true) &&
+			  (is_event_supported(INST_RETIRED, true) ||
+			   is_event_supported(INST_PREC, true));
+
+	if (pmu.version >= ID_DFR0_PMU_V3_8_1) {
+		required_events = required_events &&
+				  is_event_supported(STALL_FRONTEND, true) &&
+				  is_event_supported(STALL_BACKEND, true);
+	}
+
+	report(required_events, "Check required events are implemented");
+}
+
 #endif
 
 /*
@@ -325,6 +398,10 @@ int main(int argc, char *argv[])
 		report(check_cpi(cpi), "Cycle/instruction ratio");
 		pmccntr64_test();
 		report_prefix_pop();
+	} else if (strcmp(argv[1], "pmu-event-introspection") == 0) {
+		report_prefix_push(argv[1]);
+		test_event_introspection();
+		report_prefix_pop();
 	} else {
 		report_abort("Unknown sub-test '%s'", argv[1]);
 	}
diff --git a/arm/unittests.cfg b/arm/unittests.cfg
index fe6515c..f993548 100644
--- a/arm/unittests.cfg
+++ b/arm/unittests.cfg
@@ -66,6 +66,12 @@ file = pmu.flat
 groups = pmu
 extra_params = -append 'cycle-counter 0'
 
+[pmu-event-introspection]
+file = pmu.flat
+groups = pmu
+arch = arm64
+extra_params = -append 'pmu-event-introspection'
+
 # Test PMU support (TCG) with -icount IPC=1
 #[pmu-tcg-icount-1]
 #file = pmu.flat
diff --git a/lib/bitops.h b/lib/bitops.h
index 636064c..b310a22 100644
--- a/lib/bitops.h
+++ b/lib/bitops.h
@@ -33,6 +33,9 @@
 #define GENMASK_ULL(h, l) \
 	(((~0ULL) << (l)) & (~0ULL >> (BITS_PER_LONG_LONG - 1 - (h))))
 
+#define upper_32_bits(n) ((u32)(((n) >> 16) >> 16))
+#define lower_32_bits(n) ((u32)(n))
+
 #ifndef HAVE_BUILTIN_FLS
 static inline unsigned long fls(unsigned long word)
 {
-- 
2.20.1

[kvm-unit-tests PATCH v3 06/12] arm: pmu: Check Required Event Support

[Eric Auger]

If event counters are implemented check the common events
required by the PMUv3 are implemented.

Some are unconditionally required (SW_INCR, CPU_CYCLES,
either INST_RETIRED or INST_SPEC). Some others only are
required if the implementation implements some other features.

Check those wich are unconditionally required.

This test currently fails on TCG as neither INST_RETIRED
or INST_SPEC are supported.

Signed-off-by: Eric Auger <eric.auger@redhat.com>

---

v2 -> v3:
- use 0x%x instead %d in trace
- pmu.version >= ID_DFR0_PMU_v3_8_1
- added prefix pop
- assert instead of abort, inverse assert and test
- add defines for used events and common events
- given the changes I did not apply Andre's R-b
- introduce and use upper_32_bits()/lower_32_bits()
- added pmu prefix to the test name

v1 -> v2:
- fix is_event_supported()
- fix boolean condition for PMU v4
- fix PMCEID0 definition

RFC ->v1:
- add a comment to explain the PMCEID0/1 splits
---
 arm/pmu.c         | 77 +++++++++++++++++++++++++++++++++++++++++++++++
 arm/unittests.cfg |  6 ++++
 lib/bitops.h      |  3 ++
 3 files changed, 86 insertions(+)

diff --git a/arm/pmu.c b/arm/pmu.c
index a04588a..8c49e50 100644
--- a/arm/pmu.c
+++ b/arm/pmu.c
@@ -18,6 +18,7 @@
 #include "asm/barrier.h"
 #include "asm/sysreg.h"
 #include "asm/processor.h"
+#include <bitops.h>
 
 #define PMU_PMCR_E         (1 << 0)
 #define PMU_PMCR_C         (1 << 2)
@@ -33,6 +34,19 @@
 
 #define NR_SAMPLES 10
 
+/* Some PMU events */
+#define SW_INCR			0x0
+#define INST_RETIRED		0x8
+#define CPU_CYCLES		0x11
+#define INST_PREC		0x1B
+#define STALL_FRONTEND		0x23
+#define STALL_BACKEND		0x24
+
+#define COMMON_EVENTS_LOW	0x0
+#define COMMON_EVENTS_HIGH	0x3F
+#define EXT_COMMON_EVENTS_LOW	0x4000
+#define EXT_COMMON_EVENTS_HIGH	0x403F
+
 struct pmu {
 	unsigned int version;
 	unsigned int nb_implemented_counters;
@@ -110,6 +124,10 @@ static inline void precise_instrs_loop(int loop, uint32_t pmcr)
 	: [pmcr] "r" (pmcr), [z] "r" (0)
 	: "cc");
 }
+
+/* event counter tests only implemented for aarch64 */
+static void test_event_introspection(void) {}
+
 #elif defined(__aarch64__)
 #define ID_AA64DFR0_PERFMON_SHIFT 8
 #define ID_AA64DFR0_PERFMON_MASK  0xf
@@ -155,6 +173,61 @@ static inline void precise_instrs_loop(int loop, uint32_t pmcr)
 	: [pmcr] "r" (pmcr)
 	: "cc");
 }
+
+#define PMCEID1_EL0 sys_reg(3, 3, 9, 12, 7)
+
+static bool is_event_supported(uint32_t n, bool warn)
+{
+	uint64_t pmceid0 = read_sysreg(pmceid0_el0);
+	uint64_t pmceid1 = read_sysreg_s(PMCEID1_EL0);
+	bool supported;
+	uint64_t reg;
+
+	/*
+	 * The low 32-bits of PMCEID0/1 respectively describe
+	 * event support for events 0-31/32-63. Their High
+	 * 32-bits describe support for extended events
+	 * starting at 0x4000, using the same split.
+	 */
+	assert((n >= COMMON_EVENTS_LOW  && n <= COMMON_EVENTS_HIGH) ||
+	       (n >= EXT_COMMON_EVENTS_LOW && n <= EXT_COMMON_EVENTS_HIGH));
+
+	if (n <= COMMON_EVENTS_HIGH)
+		reg = lower_32_bits(pmceid0) | ((u64)lower_32_bits(pmceid1) << 32);
+	else
+		reg = upper_32_bits(pmceid0) | ((u64)upper_32_bits(pmceid1) << 32);
+
+	supported =  reg & (1UL << (n & 0x3F));
+
+	if (!supported && warn)
+		report_info("event 0x%x is not supported", n);
+	return supported;
+}
+
+static void test_event_introspection(void)
+{
+	bool required_events;
+
+	if (!pmu.nb_implemented_counters) {
+		report_skip("No event counter, skip ...");
+		return;
+	}
+
+	/* PMUv3 requires an implementation includes some common events */
+	required_events = is_event_supported(SW_INCR, true) &&
+			  is_event_supported(CPU_CYCLES, true) &&
+			  (is_event_supported(INST_RETIRED, true) ||
+			   is_event_supported(INST_PREC, true));
+
+	if (pmu.version >= ID_DFR0_PMU_V3_8_1) {
+		required_events = required_events &&
+				  is_event_supported(STALL_FRONTEND, true) &&
+				  is_event_supported(STALL_BACKEND, true);
+	}
+
+	report(required_events, "Check required events are implemented");
+}
+
 #endif
 
 /*
@@ -325,6 +398,10 @@ int main(int argc, char *argv[])
 		report(check_cpi(cpi), "Cycle/instruction ratio");
 		pmccntr64_test();
 		report_prefix_pop();
+	} else if (strcmp(argv[1], "pmu-event-introspection") == 0) {
+		report_prefix_push(argv[1]);
+		test_event_introspection();
+		report_prefix_pop();
 	} else {
 		report_abort("Unknown sub-test '%s'", argv[1]);
 	}
diff --git a/arm/unittests.cfg b/arm/unittests.cfg
index fe6515c..f993548 100644
--- a/arm/unittests.cfg
+++ b/arm/unittests.cfg
@@ -66,6 +66,12 @@ file = pmu.flat
 groups = pmu
 extra_params = -append 'cycle-counter 0'
 
+[pmu-event-introspection]
+file = pmu.flat
+groups = pmu
+arch = arm64
+extra_params = -append 'pmu-event-introspection'
+
 # Test PMU support (TCG) with -icount IPC=1
 #[pmu-tcg-icount-1]
 #file = pmu.flat
diff --git a/lib/bitops.h b/lib/bitops.h
index 636064c..b310a22 100644
--- a/lib/bitops.h
+++ b/lib/bitops.h
@@ -33,6 +33,9 @@
 #define GENMASK_ULL(h, l) \
 	(((~0ULL) << (l)) & (~0ULL >> (BITS_PER_LONG_LONG - 1 - (h))))
 
+#define upper_32_bits(n) ((u32)(((n) >> 16) >> 16))
+#define lower_32_bits(n) ((u32)(n))
+
 #ifndef HAVE_BUILTIN_FLS
 static inline unsigned long fls(unsigned long word)
 {
-- 
2.20.1

_______________________________________________
kvmarm mailing list
kvmarm@lists.cs.columbia.edu
https://lists.cs.columbia.edu/mailman/listinfo/kvmarm

[kvm-unit-tests PATCH v3 07/12] arm: pmu: Basic event counter Tests

[Eric Auger]

Adds the following tests:
- event-counter-config: test event counter configuration
- basic-event-count:
  - programs counters #0 and #1 to count 2 required events
  (resp. CPU_CYCLES and INST_RETIRED). Counter #0 is preset
  to a value close enough to the 32b
  overflow limit so that we check the overflow bit is set
  after the execution of the asm loop.
- mem-access: counts MEM_ACCESS event on counters #0 and #1
  with and without 32-bit overflow.

Signed-off-by: Eric Auger <eric.auger@redhat.com>
Reviewed-by: Andre Przywara <andre.przywara@arm.com>

---

v2 -> v3:
- used BIT() for PMEVTYPER_EXCLUDE_EL1/0
- print the event id in hexa (Peter)
- added prefix pop
- s/write_regn/write_regn_el0
- s/read_regn/read_regn_el0
- remove print_pmevtyper
- moved write/read_regn_el0 into lib/arm64/asm/sysreg.h
- added pmu prefix to the test names
- Kept Andre's R-b as functional code has not changed

v1 -> v2:
- fix PMCNTENSET_EL0 and PMCNTENCLR_EL0 op0
- print PMEVTYPER SH
- properly clobber used regs and add "cc"
- simplify mem_access_loop
---
 arm/pmu.c              | 248 +++++++++++++++++++++++++++++++++++++++++
 arm/unittests.cfg      |  18 +++
 lib/arm64/asm/sysreg.h |   6 +
 3 files changed, 272 insertions(+)

diff --git a/arm/pmu.c b/arm/pmu.c
index 8c49e50..45dccf7 100644
--- a/arm/pmu.c
+++ b/arm/pmu.c
@@ -19,9 +19,14 @@
 #include "asm/sysreg.h"
 #include "asm/processor.h"
 #include <bitops.h>
+#include <asm/gic.h>
 
 #define PMU_PMCR_E         (1 << 0)
+#define PMU_PMCR_P         (1 << 1)
 #define PMU_PMCR_C         (1 << 2)
+#define PMU_PMCR_D         (1 << 3)
+#define PMU_PMCR_X         (1 << 4)
+#define PMU_PMCR_DP        (1 << 5)
 #define PMU_PMCR_LC        (1 << 6)
 #define PMU_PMCR_N_SHIFT   11
 #define PMU_PMCR_N_MASK    0x1f
@@ -38,6 +43,7 @@
 #define SW_INCR			0x0
 #define INST_RETIRED		0x8
 #define CPU_CYCLES		0x11
+#define MEM_ACCESS		0x13
 #define INST_PREC		0x1B
 #define STALL_FRONTEND		0x23
 #define STALL_BACKEND		0x24
@@ -47,6 +53,10 @@
 #define EXT_COMMON_EVENTS_LOW	0x4000
 #define EXT_COMMON_EVENTS_HIGH	0x403F
 
+#define ALL_SET			0xFFFFFFFF
+#define ALL_CLEAR		0x0
+#define PRE_OVERFLOW		0xFFFFFFF0
+
 struct pmu {
 	unsigned int version;
 	unsigned int nb_implemented_counters;
@@ -127,6 +137,9 @@ static inline void precise_instrs_loop(int loop, uint32_t pmcr)
 
 /* event counter tests only implemented for aarch64 */
 static void test_event_introspection(void) {}
+static void test_event_counter_config(void) {}
+static void test_basic_event_count(void) {}
+static void test_mem_access(void) {}
 
 #elif defined(__aarch64__)
 #define ID_AA64DFR0_PERFMON_SHIFT 8
@@ -175,6 +188,11 @@ static inline void precise_instrs_loop(int loop, uint32_t pmcr)
 }
 
 #define PMCEID1_EL0 sys_reg(3, 3, 9, 12, 7)
+#define PMCNTENSET_EL0 sys_reg(3, 3, 9, 12, 1)
+#define PMCNTENCLR_EL0 sys_reg(3, 3, 9, 12, 2)
+
+#define PMEVTYPER_EXCLUDE_EL1 BIT(31)
+#define PMEVTYPER_EXCLUDE_EL0 BIT(30)
 
 static bool is_event_supported(uint32_t n, bool warn)
 {
@@ -228,6 +246,224 @@ static void test_event_introspection(void)
 	report(required_events, "Check required events are implemented");
 }
 
+/*
+ * Extra instructions inserted by the compiler would be difficult to compensate
+ * for, so hand assemble everything between, and including, the PMCR accesses
+ * to start and stop counting. isb instructions are inserted to make sure
+ * pmccntr read after this function returns the exact instructions executed
+ * in the controlled block. Loads @loop times the data at @address into x9.
+ */
+static void mem_access_loop(void *addr, int loop, uint32_t pmcr)
+{
+asm volatile(
+	"       msr     pmcr_el0, %[pmcr]\n"
+	"       isb\n"
+	"       mov     x10, %[loop]\n"
+	"1:     sub     x10, x10, #1\n"
+	"       ldr	x9, [%[addr]]\n"
+	"       cmp     x10, #0x0\n"
+	"       b.gt    1b\n"
+	"       msr     pmcr_el0, xzr\n"
+	"       isb\n"
+	:
+	: [addr] "r" (addr), [pmcr] "r" (pmcr), [loop] "r" (loop)
+	: "x9", "x10", "cc");
+}
+
+static void pmu_reset(void)
+{
+	/* reset all counters, counting disabled at PMCR level*/
+	set_pmcr(pmu.pmcr_ro | PMU_PMCR_LC | PMU_PMCR_C | PMU_PMCR_P);
+	/* Disable all counters */
+	write_sysreg_s(ALL_SET, PMCNTENCLR_EL0);
+	/* clear overflow reg */
+	write_sysreg(ALL_SET, pmovsclr_el0);
+	/* disable overflow interrupts on all counters */
+	write_sysreg(ALL_SET, pmintenclr_el1);
+	isb();
+}
+
+static void test_event_counter_config(void)
+{
+	int i;
+
+	if (!pmu.nb_implemented_counters) {
+		report_skip("No event counter, skip ...");
+		return;
+	}
+
+	pmu_reset();
+
+	/*
+	 * Test setting through PMESELR/PMXEVTYPER and PMEVTYPERn read,
+	 * select counter 0
+	 */
+	write_sysreg(1, PMSELR_EL0);
+	/* program this counter to count unsupported event */
+	write_sysreg(0xEA, PMXEVTYPER_EL0);
+	write_sysreg(0xdeadbeef, PMXEVCNTR_EL0);
+	report((read_regn_el0(pmevtyper, 1) & 0xFFF) == 0xEA,
+		"PMESELR/PMXEVTYPER/PMEVTYPERn");
+	report((read_regn_el0(pmevcntr, 1) == 0xdeadbeef),
+		"PMESELR/PMXEVCNTR/PMEVCNTRn");
+
+	/* try to configure an unsupported event within the range [0x0, 0x3F] */
+	for (i = 0; i <= 0x3F; i++) {
+		if (!is_event_supported(i, false))
+			break;
+	}
+	if (i > 0x3F) {
+		report_skip("pmevtyper: all events within [0x0, 0x3F] are supported");
+		return;
+	}
+
+	/* select counter 0 */
+	write_sysreg(0, PMSELR_EL0);
+	/* program this counter to count unsupported event */
+	write_sysreg(i, PMXEVCNTR_EL0);
+	/* read the counter value */
+	read_sysreg(PMXEVCNTR_EL0);
+	report(read_sysreg(PMXEVCNTR_EL0) == i,
+		"read of a counter programmed with unsupported event");
+
+}
+
+static bool satisfy_prerequisites(uint32_t *events, unsigned int nb_events)
+{
+	int i;
+
+	if (pmu.nb_implemented_counters < nb_events) {
+		report_skip("Skip test as number of counters is too small (%d)",
+			    pmu.nb_implemented_counters);
+		return false;
+	}
+
+	for (i = 0; i < nb_events; i++) {
+		if (!is_event_supported(events[i], false)) {
+			report_skip("Skip test as event 0x%x is not supported",
+				    events[i]);
+			return false;
+		}
+	}
+	return true;
+}
+
+static void test_basic_event_count(void)
+{
+	uint32_t implemented_counter_mask, non_implemented_counter_mask;
+	uint32_t counter_mask;
+	uint32_t events[] = {CPU_CYCLES, INST_RETIRED};
+
+	if (!satisfy_prerequisites(events, ARRAY_SIZE(events)))
+		return;
+
+	implemented_counter_mask = BIT(pmu.nb_implemented_counters) - 1;
+	non_implemented_counter_mask = ~(BIT(31) | implemented_counter_mask);
+	counter_mask = implemented_counter_mask | non_implemented_counter_mask;
+
+	write_regn_el0(pmevtyper, 0, CPU_CYCLES | PMEVTYPER_EXCLUDE_EL0);
+	write_regn_el0(pmevtyper, 1, INST_RETIRED | PMEVTYPER_EXCLUDE_EL0);
+
+	/* disable all counters */
+	write_sysreg_s(ALL_SET, PMCNTENCLR_EL0);
+	report(!read_sysreg_s(PMCNTENCLR_EL0) && !read_sysreg_s(PMCNTENSET_EL0),
+		"pmcntenclr: disable all counters");
+
+	/*
+	 * clear cycle and all event counters and allow counter enablement
+	 * through PMCNTENSET. LC is RES1.
+	 */
+	set_pmcr(pmu.pmcr_ro | PMU_PMCR_LC | PMU_PMCR_C | PMU_PMCR_P);
+	isb();
+	report(get_pmcr() == (pmu.pmcr_ro | PMU_PMCR_LC), "pmcr: reset counters");
+
+	/* Preset counter #0 to pre overflow value to trigger an overflow */
+	write_regn_el0(pmevcntr, 0, PRE_OVERFLOW);
+	report(read_regn_el0(pmevcntr, 0) == PRE_OVERFLOW,
+		"counter #0 preset to pre-overflow value");
+	report(!read_regn_el0(pmevcntr, 1), "counter #1 is 0");
+
+	/*
+	 * Enable all implemented counters and also attempt to enable
+	 * not supported counters. Counting still is disabled by !PMCR.E
+	 */
+	write_sysreg_s(counter_mask, PMCNTENSET_EL0);
+
+	/* check only those implemented are enabled */
+	report((read_sysreg_s(PMCNTENSET_EL0) == read_sysreg_s(PMCNTENCLR_EL0)) &&
+		(read_sysreg_s(PMCNTENSET_EL0) == implemented_counter_mask),
+		"pmcntenset: enabled implemented_counters");
+
+	/* Disable all counters but counters #0 and #1 */
+	write_sysreg_s(~0x3, PMCNTENCLR_EL0);
+	report((read_sysreg_s(PMCNTENSET_EL0) == read_sysreg_s(PMCNTENCLR_EL0)) &&
+		(read_sysreg_s(PMCNTENSET_EL0) == 0x3),
+		"pmcntenset: just enabled #0 and #1");
+
+	/* clear overflow register */
+	write_sysreg(ALL_SET, pmovsclr_el0);
+	report(!read_sysreg(pmovsclr_el0), "check overflow reg is 0");
+
+	/* disable overflow interrupts on all counters*/
+	write_sysreg(ALL_SET, pmintenclr_el1);
+	report(!read_sysreg(pmintenclr_el1),
+		"pmintenclr_el1=0, all interrupts disabled");
+
+	/* enable overflow interrupts on all event counters */
+	write_sysreg(implemented_counter_mask | non_implemented_counter_mask,
+		     pmintenset_el1);
+	report(read_sysreg(pmintenset_el1) == implemented_counter_mask,
+		"overflow interrupts enabled on all implemented counters");
+
+	/* Set PMCR.E, execute asm code and unset PMCR.E */
+	precise_instrs_loop(20, pmu.pmcr_ro | PMU_PMCR_E);
+
+	report_info("counter #0 is 0x%lx (CPU_CYCLES)",
+		    read_regn_el0(pmevcntr, 0));
+	report_info("counter #1 is 0x%lx (INST_RETIRED)",
+		    read_regn_el0(pmevcntr, 1));
+
+	report_info("overflow reg = 0x%lx", read_sysreg(pmovsclr_el0));
+	report(read_sysreg(pmovsclr_el0) & 0x1,
+		"check overflow happened on #0 only");
+}
+
+static void test_mem_access(void)
+{
+	void *addr = malloc(PAGE_SIZE);
+	uint32_t events[] = {MEM_ACCESS, MEM_ACCESS};
+
+	if (!satisfy_prerequisites(events, ARRAY_SIZE(events)))
+		return;
+
+	pmu_reset();
+
+	write_regn_el0(pmevtyper, 0, MEM_ACCESS | PMEVTYPER_EXCLUDE_EL0);
+	write_regn_el0(pmevtyper, 1, MEM_ACCESS | PMEVTYPER_EXCLUDE_EL0);
+	write_sysreg_s(0x3, PMCNTENSET_EL0);
+	isb();
+	mem_access_loop(addr, 20, pmu.pmcr_ro | PMU_PMCR_E);
+	report_info("counter #0 is %ld (MEM_ACCESS)", read_regn_el0(pmevcntr, 0));
+	report_info("counter #1 is %ld (MEM_ACCESS)", read_regn_el0(pmevcntr, 1));
+	/* We may measure more than 20 mem access depending on the core */
+	report((read_regn_el0(pmevcntr, 0) == read_regn_el0(pmevcntr, 1)) &&
+	       (read_regn_el0(pmevcntr, 0) >= 20) && !read_sysreg(pmovsclr_el0),
+	       "Ran 20 mem accesses");
+
+	pmu_reset();
+
+	write_regn_el0(pmevcntr, 0, PRE_OVERFLOW);
+	write_regn_el0(pmevcntr, 1, PRE_OVERFLOW);
+	write_sysreg_s(0x3, PMCNTENSET_EL0);
+	isb();
+	mem_access_loop(addr, 20, pmu.pmcr_ro | PMU_PMCR_E);
+	report(read_sysreg(pmovsclr_el0) == 0x3,
+	       "Ran 20 mem accesses with expected overflows on both counters");
+	report_info("cnt#0 = %ld cnt#1=%ld overflow=0x%lx",
+			read_regn_el0(pmevcntr, 0), read_regn_el0(pmevcntr, 1),
+			read_sysreg(pmovsclr_el0));
+}
+
 #endif
 
 /*
@@ -402,6 +638,18 @@ int main(int argc, char *argv[])
 		report_prefix_push(argv[1]);
 		test_event_introspection();
 		report_prefix_pop();
+	} else if (strcmp(argv[1], "pmu-event-counter-config") == 0) {
+		report_prefix_push(argv[1]);
+		test_event_counter_config();
+		report_prefix_pop();
+	} else if (strcmp(argv[1], "pmu-basic-event-count") == 0) {
+		report_prefix_push(argv[1]);
+		test_basic_event_count();
+		report_prefix_pop();
+	} else if (strcmp(argv[1], "pmu-mem-access") == 0) {
+		report_prefix_push(argv[1]);
+		test_mem_access();
+		report_prefix_pop();
 	} else {
 		report_abort("Unknown sub-test '%s'", argv[1]);
 	}
diff --git a/arm/unittests.cfg b/arm/unittests.cfg
index f993548..32ab8c6 100644
--- a/arm/unittests.cfg
+++ b/arm/unittests.cfg
@@ -72,6 +72,24 @@ groups = pmu
 arch = arm64
 extra_params = -append 'pmu-event-introspection'
 
+[pmu-event-counter-config]
+file = pmu.flat
+groups = pmu
+arch = arm64
+extra_params = -append 'pmu-event-counter-config'
+
+[pmu-basic-event-count]
+file = pmu.flat
+groups = pmu
+arch = arm64
+extra_params = -append 'pmu-basic-event-count'
+
+[pmu-mem-access]
+file = pmu.flat
+groups = pmu
+arch = arm64
+extra_params = -append 'pmu-mem-access'
+
 # Test PMU support (TCG) with -icount IPC=1
 #[pmu-tcg-icount-1]
 #file = pmu.flat
diff --git a/lib/arm64/asm/sysreg.h b/lib/arm64/asm/sysreg.h
index a45eebd..378bf7e 100644
--- a/lib/arm64/asm/sysreg.h
+++ b/lib/arm64/asm/sysreg.h
@@ -49,6 +49,12 @@
 	asm volatile("msr_s " xstr(r) ", %x0" : : "rZ" (__val));\
 } while (0)
 
+#define write_regn_el0(__reg, __n, __val) \
+	write_sysreg((__val), __reg ## __n ## _el0)
+
+#define read_regn_el0(__reg, __n) \
+	read_sysreg(__reg ## __n ## _el0)
+
 asm(
 "	.irp	num,0,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30\n"
 "	.equ	.L__reg_num_x\\num, \\num\n"
-- 
2.20.1

[kvm-unit-tests PATCH v3 07/12] arm: pmu: Basic event counter Tests

[Eric Auger]

Adds the following tests:
- event-counter-config: test event counter configuration
- basic-event-count:
  - programs counters #0 and #1 to count 2 required events
  (resp. CPU_CYCLES and INST_RETIRED). Counter #0 is preset
  to a value close enough to the 32b
  overflow limit so that we check the overflow bit is set
  after the execution of the asm loop.
- mem-access: counts MEM_ACCESS event on counters #0 and #1
  with and without 32-bit overflow.

Signed-off-by: Eric Auger <eric.auger@redhat.com>
Reviewed-by: Andre Przywara <andre.przywara@arm.com>

---

v2 -> v3:
- used BIT() for PMEVTYPER_EXCLUDE_EL1/0
- print the event id in hexa (Peter)
- added prefix pop
- s/write_regn/write_regn_el0
- s/read_regn/read_regn_el0
- remove print_pmevtyper
- moved write/read_regn_el0 into lib/arm64/asm/sysreg.h
- added pmu prefix to the test names
- Kept Andre's R-b as functional code has not changed

v1 -> v2:
- fix PMCNTENSET_EL0 and PMCNTENCLR_EL0 op0
- print PMEVTYPER SH
- properly clobber used regs and add "cc"
- simplify mem_access_loop
---
 arm/pmu.c              | 248 +++++++++++++++++++++++++++++++++++++++++
 arm/unittests.cfg      |  18 +++
 lib/arm64/asm/sysreg.h |   6 +
 3 files changed, 272 insertions(+)

diff --git a/arm/pmu.c b/arm/pmu.c
index 8c49e50..45dccf7 100644
--- a/arm/pmu.c
+++ b/arm/pmu.c
@@ -19,9 +19,14 @@
 #include "asm/sysreg.h"
 #include "asm/processor.h"
 #include <bitops.h>
+#include <asm/gic.h>
 
 #define PMU_PMCR_E         (1 << 0)
+#define PMU_PMCR_P         (1 << 1)
 #define PMU_PMCR_C         (1 << 2)
+#define PMU_PMCR_D         (1 << 3)
+#define PMU_PMCR_X         (1 << 4)
+#define PMU_PMCR_DP        (1 << 5)
 #define PMU_PMCR_LC        (1 << 6)
 #define PMU_PMCR_N_SHIFT   11
 #define PMU_PMCR_N_MASK    0x1f
@@ -38,6 +43,7 @@
 #define SW_INCR			0x0
 #define INST_RETIRED		0x8
 #define CPU_CYCLES		0x11
+#define MEM_ACCESS		0x13
 #define INST_PREC		0x1B
 #define STALL_FRONTEND		0x23
 #define STALL_BACKEND		0x24
@@ -47,6 +53,10 @@
 #define EXT_COMMON_EVENTS_LOW	0x4000
 #define EXT_COMMON_EVENTS_HIGH	0x403F
 
+#define ALL_SET			0xFFFFFFFF
+#define ALL_CLEAR		0x0
+#define PRE_OVERFLOW		0xFFFFFFF0
+
 struct pmu {
 	unsigned int version;
 	unsigned int nb_implemented_counters;
@@ -127,6 +137,9 @@ static inline void precise_instrs_loop(int loop, uint32_t pmcr)
 
 /* event counter tests only implemented for aarch64 */
 static void test_event_introspection(void) {}
+static void test_event_counter_config(void) {}
+static void test_basic_event_count(void) {}
+static void test_mem_access(void) {}
 
 #elif defined(__aarch64__)
 #define ID_AA64DFR0_PERFMON_SHIFT 8
@@ -175,6 +188,11 @@ static inline void precise_instrs_loop(int loop, uint32_t pmcr)
 }
 
 #define PMCEID1_EL0 sys_reg(3, 3, 9, 12, 7)
+#define PMCNTENSET_EL0 sys_reg(3, 3, 9, 12, 1)
+#define PMCNTENCLR_EL0 sys_reg(3, 3, 9, 12, 2)
+
+#define PMEVTYPER_EXCLUDE_EL1 BIT(31)
+#define PMEVTYPER_EXCLUDE_EL0 BIT(30)
 
 static bool is_event_supported(uint32_t n, bool warn)
 {
@@ -228,6 +246,224 @@ static void test_event_introspection(void)
 	report(required_events, "Check required events are implemented");
 }
 
+/*
+ * Extra instructions inserted by the compiler would be difficult to compensate
+ * for, so hand assemble everything between, and including, the PMCR accesses
+ * to start and stop counting. isb instructions are inserted to make sure
+ * pmccntr read after this function returns the exact instructions executed
+ * in the controlled block. Loads @loop times the data at @address into x9.
+ */
+static void mem_access_loop(void *addr, int loop, uint32_t pmcr)
+{
+asm volatile(
+	"       msr     pmcr_el0, %[pmcr]\n"
+	"       isb\n"
+	"       mov     x10, %[loop]\n"
+	"1:     sub     x10, x10, #1\n"
+	"       ldr	x9, [%[addr]]\n"
+	"       cmp     x10, #0x0\n"
+	"       b.gt    1b\n"
+	"       msr     pmcr_el0, xzr\n"
+	"       isb\n"
+	:
+	: [addr] "r" (addr), [pmcr] "r" (pmcr), [loop] "r" (loop)
+	: "x9", "x10", "cc");
+}
+
+static void pmu_reset(void)
+{
+	/* reset all counters, counting disabled at PMCR level*/
+	set_pmcr(pmu.pmcr_ro | PMU_PMCR_LC | PMU_PMCR_C | PMU_PMCR_P);
+	/* Disable all counters */
+	write_sysreg_s(ALL_SET, PMCNTENCLR_EL0);
+	/* clear overflow reg */
+	write_sysreg(ALL_SET, pmovsclr_el0);
+	/* disable overflow interrupts on all counters */
+	write_sysreg(ALL_SET, pmintenclr_el1);
+	isb();
+}
+
+static void test_event_counter_config(void)
+{
+	int i;
+
+	if (!pmu.nb_implemented_counters) {
+		report_skip("No event counter, skip ...");
+		return;
+	}
+
+	pmu_reset();
+
+	/*
+	 * Test setting through PMESELR/PMXEVTYPER and PMEVTYPERn read,
+	 * select counter 0
+	 */
+	write_sysreg(1, PMSELR_EL0);
+	/* program this counter to count unsupported event */
+	write_sysreg(0xEA, PMXEVTYPER_EL0);
+	write_sysreg(0xdeadbeef, PMXEVCNTR_EL0);
+	report((read_regn_el0(pmevtyper, 1) & 0xFFF) == 0xEA,
+		"PMESELR/PMXEVTYPER/PMEVTYPERn");
+	report((read_regn_el0(pmevcntr, 1) == 0xdeadbeef),
+		"PMESELR/PMXEVCNTR/PMEVCNTRn");
+
+	/* try to configure an unsupported event within the range [0x0, 0x3F] */
+	for (i = 0; i <= 0x3F; i++) {
+		if (!is_event_supported(i, false))
+			break;
+	}
+	if (i > 0x3F) {
+		report_skip("pmevtyper: all events within [0x0, 0x3F] are supported");
+		return;
+	}
+
+	/* select counter 0 */
+	write_sysreg(0, PMSELR_EL0);
+	/* program this counter to count unsupported event */
+	write_sysreg(i, PMXEVCNTR_EL0);
+	/* read the counter value */
+	read_sysreg(PMXEVCNTR_EL0);
+	report(read_sysreg(PMXEVCNTR_EL0) == i,
+		"read of a counter programmed with unsupported event");
+
+}
+
+static bool satisfy_prerequisites(uint32_t *events, unsigned int nb_events)
+{
+	int i;
+
+	if (pmu.nb_implemented_counters < nb_events) {
+		report_skip("Skip test as number of counters is too small (%d)",
+			    pmu.nb_implemented_counters);
+		return false;
+	}
+
+	for (i = 0; i < nb_events; i++) {
+		if (!is_event_supported(events[i], false)) {
+			report_skip("Skip test as event 0x%x is not supported",
+				    events[i]);
+			return false;
+		}
+	}
+	return true;
+}
+
+static void test_basic_event_count(void)
+{
+	uint32_t implemented_counter_mask, non_implemented_counter_mask;
+	uint32_t counter_mask;
+	uint32_t events[] = {CPU_CYCLES, INST_RETIRED};
+
+	if (!satisfy_prerequisites(events, ARRAY_SIZE(events)))
+		return;
+
+	implemented_counter_mask = BIT(pmu.nb_implemented_counters) - 1;
+	non_implemented_counter_mask = ~(BIT(31) | implemented_counter_mask);
+	counter_mask = implemented_counter_mask | non_implemented_counter_mask;
+
+	write_regn_el0(pmevtyper, 0, CPU_CYCLES | PMEVTYPER_EXCLUDE_EL0);
+	write_regn_el0(pmevtyper, 1, INST_RETIRED | PMEVTYPER_EXCLUDE_EL0);
+
+	/* disable all counters */
+	write_sysreg_s(ALL_SET, PMCNTENCLR_EL0);
+	report(!read_sysreg_s(PMCNTENCLR_EL0) && !read_sysreg_s(PMCNTENSET_EL0),
+		"pmcntenclr: disable all counters");
+
+	/*
+	 * clear cycle and all event counters and allow counter enablement
+	 * through PMCNTENSET. LC is RES1.
+	 */
+	set_pmcr(pmu.pmcr_ro | PMU_PMCR_LC | PMU_PMCR_C | PMU_PMCR_P);
+	isb();
+	report(get_pmcr() == (pmu.pmcr_ro | PMU_PMCR_LC), "pmcr: reset counters");
+
+	/* Preset counter #0 to pre overflow value to trigger an overflow */
+	write_regn_el0(pmevcntr, 0, PRE_OVERFLOW);
+	report(read_regn_el0(pmevcntr, 0) == PRE_OVERFLOW,
+		"counter #0 preset to pre-overflow value");
+	report(!read_regn_el0(pmevcntr, 1), "counter #1 is 0");
+
+	/*
+	 * Enable all implemented counters and also attempt to enable
+	 * not supported counters. Counting still is disabled by !PMCR.E
+	 */
+	write_sysreg_s(counter_mask, PMCNTENSET_EL0);
+
+	/* check only those implemented are enabled */
+	report((read_sysreg_s(PMCNTENSET_EL0) == read_sysreg_s(PMCNTENCLR_EL0)) &&
+		(read_sysreg_s(PMCNTENSET_EL0) == implemented_counter_mask),
+		"pmcntenset: enabled implemented_counters");
+
+	/* Disable all counters but counters #0 and #1 */
+	write_sysreg_s(~0x3, PMCNTENCLR_EL0);
+	report((read_sysreg_s(PMCNTENSET_EL0) == read_sysreg_s(PMCNTENCLR_EL0)) &&
+		(read_sysreg_s(PMCNTENSET_EL0) == 0x3),
+		"pmcntenset: just enabled #0 and #1");
+
+	/* clear overflow register */
+	write_sysreg(ALL_SET, pmovsclr_el0);
+	report(!read_sysreg(pmovsclr_el0), "check overflow reg is 0");
+
+	/* disable overflow interrupts on all counters*/
+	write_sysreg(ALL_SET, pmintenclr_el1);
+	report(!read_sysreg(pmintenclr_el1),
+		"pmintenclr_el1=0, all interrupts disabled");
+
+	/* enable overflow interrupts on all event counters */
+	write_sysreg(implemented_counter_mask | non_implemented_counter_mask,
+		     pmintenset_el1);
+	report(read_sysreg(pmintenset_el1) == implemented_counter_mask,
+		"overflow interrupts enabled on all implemented counters");
+
+	/* Set PMCR.E, execute asm code and unset PMCR.E */
+	precise_instrs_loop(20, pmu.pmcr_ro | PMU_PMCR_E);
+
+	report_info("counter #0 is 0x%lx (CPU_CYCLES)",
+		    read_regn_el0(pmevcntr, 0));
+	report_info("counter #1 is 0x%lx (INST_RETIRED)",
+		    read_regn_el0(pmevcntr, 1));
+
+	report_info("overflow reg = 0x%lx", read_sysreg(pmovsclr_el0));
+	report(read_sysreg(pmovsclr_el0) & 0x1,
+		"check overflow happened on #0 only");
+}
+
+static void test_mem_access(void)
+{
+	void *addr = malloc(PAGE_SIZE);
+	uint32_t events[] = {MEM_ACCESS, MEM_ACCESS};
+
+	if (!satisfy_prerequisites(events, ARRAY_SIZE(events)))
+		return;
+
+	pmu_reset();
+
+	write_regn_el0(pmevtyper, 0, MEM_ACCESS | PMEVTYPER_EXCLUDE_EL0);
+	write_regn_el0(pmevtyper, 1, MEM_ACCESS | PMEVTYPER_EXCLUDE_EL0);
+	write_sysreg_s(0x3, PMCNTENSET_EL0);
+	isb();
+	mem_access_loop(addr, 20, pmu.pmcr_ro | PMU_PMCR_E);
+	report_info("counter #0 is %ld (MEM_ACCESS)", read_regn_el0(pmevcntr, 0));
+	report_info("counter #1 is %ld (MEM_ACCESS)", read_regn_el0(pmevcntr, 1));
+	/* We may measure more than 20 mem access depending on the core */
+	report((read_regn_el0(pmevcntr, 0) == read_regn_el0(pmevcntr, 1)) &&
+	       (read_regn_el0(pmevcntr, 0) >= 20) && !read_sysreg(pmovsclr_el0),
+	       "Ran 20 mem accesses");
+
+	pmu_reset();
+
+	write_regn_el0(pmevcntr, 0, PRE_OVERFLOW);
+	write_regn_el0(pmevcntr, 1, PRE_OVERFLOW);
+	write_sysreg_s(0x3, PMCNTENSET_EL0);
+	isb();
+	mem_access_loop(addr, 20, pmu.pmcr_ro | PMU_PMCR_E);
+	report(read_sysreg(pmovsclr_el0) == 0x3,
+	       "Ran 20 mem accesses with expected overflows on both counters");
+	report_info("cnt#0 = %ld cnt#1=%ld overflow=0x%lx",
+			read_regn_el0(pmevcntr, 0), read_regn_el0(pmevcntr, 1),
+			read_sysreg(pmovsclr_el0));
+}
+
 #endif
 
 /*
@@ -402,6 +638,18 @@ int main(int argc, char *argv[])
 		report_prefix_push(argv[1]);
 		test_event_introspection();
 		report_prefix_pop();
+	} else if (strcmp(argv[1], "pmu-event-counter-config") == 0) {
+		report_prefix_push(argv[1]);
+		test_event_counter_config();
+		report_prefix_pop();
+	} else if (strcmp(argv[1], "pmu-basic-event-count") == 0) {
+		report_prefix_push(argv[1]);
+		test_basic_event_count();
+		report_prefix_pop();
+	} else if (strcmp(argv[1], "pmu-mem-access") == 0) {
+		report_prefix_push(argv[1]);
+		test_mem_access();
+		report_prefix_pop();
 	} else {
 		report_abort("Unknown sub-test '%s'", argv[1]);
 	}
diff --git a/arm/unittests.cfg b/arm/unittests.cfg
index f993548..32ab8c6 100644
--- a/arm/unittests.cfg
+++ b/arm/unittests.cfg
@@ -72,6 +72,24 @@ groups = pmu
 arch = arm64
 extra_params = -append 'pmu-event-introspection'
 
+[pmu-event-counter-config]
+file = pmu.flat
+groups = pmu
+arch = arm64
+extra_params = -append 'pmu-event-counter-config'
+
+[pmu-basic-event-count]
+file = pmu.flat
+groups = pmu
+arch = arm64
+extra_params = -append 'pmu-basic-event-count'
+
+[pmu-mem-access]
+file = pmu.flat
+groups = pmu
+arch = arm64
+extra_params = -append 'pmu-mem-access'
+
 # Test PMU support (TCG) with -icount IPC=1
 #[pmu-tcg-icount-1]
 #file = pmu.flat
diff --git a/lib/arm64/asm/sysreg.h b/lib/arm64/asm/sysreg.h
index a45eebd..378bf7e 100644
--- a/lib/arm64/asm/sysreg.h
+++ b/lib/arm64/asm/sysreg.h
@@ -49,6 +49,12 @@
 	asm volatile("msr_s " xstr(r) ", %x0" : : "rZ" (__val));\
 } while (0)
 
+#define write_regn_el0(__reg, __n, __val) \
+	write_sysreg((__val), __reg ## __n ## _el0)
+
+#define read_regn_el0(__reg, __n) \
+	read_sysreg(__reg ## __n ## _el0)
+
 asm(
 "	.irp	num,0,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30\n"
 "	.equ	.L__reg_num_x\\num, \\num\n"
-- 
2.20.1

[kvm-unit-tests PATCH v3 07/12] arm: pmu: Basic event counter Tests

[Eric Auger]

Adds the following tests:
- event-counter-config: test event counter configuration
- basic-event-count:
  - programs counters #0 and #1 to count 2 required events
  (resp. CPU_CYCLES and INST_RETIRED). Counter #0 is preset
  to a value close enough to the 32b
  overflow limit so that we check the overflow bit is set
  after the execution of the asm loop.
- mem-access: counts MEM_ACCESS event on counters #0 and #1
  with and without 32-bit overflow.

Signed-off-by: Eric Auger <eric.auger@redhat.com>
Reviewed-by: Andre Przywara <andre.przywara@arm.com>

---

v2 -> v3:
- used BIT() for PMEVTYPER_EXCLUDE_EL1/0
- print the event id in hexa (Peter)
- added prefix pop
- s/write_regn/write_regn_el0
- s/read_regn/read_regn_el0
- remove print_pmevtyper
- moved write/read_regn_el0 into lib/arm64/asm/sysreg.h
- added pmu prefix to the test names
- Kept Andre's R-b as functional code has not changed

v1 -> v2:
- fix PMCNTENSET_EL0 and PMCNTENCLR_EL0 op0
- print PMEVTYPER SH
- properly clobber used regs and add "cc"
- simplify mem_access_loop
---
 arm/pmu.c              | 248 +++++++++++++++++++++++++++++++++++++++++
 arm/unittests.cfg      |  18 +++
 lib/arm64/asm/sysreg.h |   6 +
 3 files changed, 272 insertions(+)

diff --git a/arm/pmu.c b/arm/pmu.c
index 8c49e50..45dccf7 100644
--- a/arm/pmu.c
+++ b/arm/pmu.c
@@ -19,9 +19,14 @@
 #include "asm/sysreg.h"
 #include "asm/processor.h"
 #include <bitops.h>
+#include <asm/gic.h>
 
 #define PMU_PMCR_E         (1 << 0)
+#define PMU_PMCR_P         (1 << 1)
 #define PMU_PMCR_C         (1 << 2)
+#define PMU_PMCR_D         (1 << 3)
+#define PMU_PMCR_X         (1 << 4)
+#define PMU_PMCR_DP        (1 << 5)
 #define PMU_PMCR_LC        (1 << 6)
 #define PMU_PMCR_N_SHIFT   11
 #define PMU_PMCR_N_MASK    0x1f
@@ -38,6 +43,7 @@
 #define SW_INCR			0x0
 #define INST_RETIRED		0x8
 #define CPU_CYCLES		0x11
+#define MEM_ACCESS		0x13
 #define INST_PREC		0x1B
 #define STALL_FRONTEND		0x23
 #define STALL_BACKEND		0x24
@@ -47,6 +53,10 @@
 #define EXT_COMMON_EVENTS_LOW	0x4000
 #define EXT_COMMON_EVENTS_HIGH	0x403F
 
+#define ALL_SET			0xFFFFFFFF
+#define ALL_CLEAR		0x0
+#define PRE_OVERFLOW		0xFFFFFFF0
+
 struct pmu {
 	unsigned int version;
 	unsigned int nb_implemented_counters;
@@ -127,6 +137,9 @@ static inline void precise_instrs_loop(int loop, uint32_t pmcr)
 
 /* event counter tests only implemented for aarch64 */
 static void test_event_introspection(void) {}
+static void test_event_counter_config(void) {}
+static void test_basic_event_count(void) {}
+static void test_mem_access(void) {}
 
 #elif defined(__aarch64__)
 #define ID_AA64DFR0_PERFMON_SHIFT 8
@@ -175,6 +188,11 @@ static inline void precise_instrs_loop(int loop, uint32_t pmcr)
 }
 
 #define PMCEID1_EL0 sys_reg(3, 3, 9, 12, 7)
+#define PMCNTENSET_EL0 sys_reg(3, 3, 9, 12, 1)
+#define PMCNTENCLR_EL0 sys_reg(3, 3, 9, 12, 2)
+
+#define PMEVTYPER_EXCLUDE_EL1 BIT(31)
+#define PMEVTYPER_EXCLUDE_EL0 BIT(30)
 
 static bool is_event_supported(uint32_t n, bool warn)
 {
@@ -228,6 +246,224 @@ static void test_event_introspection(void)
 	report(required_events, "Check required events are implemented");
 }
 
+/*
+ * Extra instructions inserted by the compiler would be difficult to compensate
+ * for, so hand assemble everything between, and including, the PMCR accesses
+ * to start and stop counting. isb instructions are inserted to make sure
+ * pmccntr read after this function returns the exact instructions executed
+ * in the controlled block. Loads @loop times the data at @address into x9.
+ */
+static void mem_access_loop(void *addr, int loop, uint32_t pmcr)
+{
+asm volatile(
+	"       msr     pmcr_el0, %[pmcr]\n"
+	"       isb\n"
+	"       mov     x10, %[loop]\n"
+	"1:     sub     x10, x10, #1\n"
+	"       ldr	x9, [%[addr]]\n"
+	"       cmp     x10, #0x0\n"
+	"       b.gt    1b\n"
+	"       msr     pmcr_el0, xzr\n"
+	"       isb\n"
+	:
+	: [addr] "r" (addr), [pmcr] "r" (pmcr), [loop] "r" (loop)
+	: "x9", "x10", "cc");
+}
+
+static void pmu_reset(void)
+{
+	/* reset all counters, counting disabled at PMCR level*/
+	set_pmcr(pmu.pmcr_ro | PMU_PMCR_LC | PMU_PMCR_C | PMU_PMCR_P);
+	/* Disable all counters */
+	write_sysreg_s(ALL_SET, PMCNTENCLR_EL0);
+	/* clear overflow reg */
+	write_sysreg(ALL_SET, pmovsclr_el0);
+	/* disable overflow interrupts on all counters */
+	write_sysreg(ALL_SET, pmintenclr_el1);
+	isb();
+}
+
+static void test_event_counter_config(void)
+{
+	int i;
+
+	if (!pmu.nb_implemented_counters) {
+		report_skip("No event counter, skip ...");
+		return;
+	}
+
+	pmu_reset();
+
+	/*
+	 * Test setting through PMESELR/PMXEVTYPER and PMEVTYPERn read,
+	 * select counter 0
+	 */
+	write_sysreg(1, PMSELR_EL0);
+	/* program this counter to count unsupported event */
+	write_sysreg(0xEA, PMXEVTYPER_EL0);
+	write_sysreg(0xdeadbeef, PMXEVCNTR_EL0);
+	report((read_regn_el0(pmevtyper, 1) & 0xFFF) == 0xEA,
+		"PMESELR/PMXEVTYPER/PMEVTYPERn");
+	report((read_regn_el0(pmevcntr, 1) == 0xdeadbeef),
+		"PMESELR/PMXEVCNTR/PMEVCNTRn");
+
+	/* try to configure an unsupported event within the range [0x0, 0x3F] */
+	for (i = 0; i <= 0x3F; i++) {
+		if (!is_event_supported(i, false))
+			break;
+	}
+	if (i > 0x3F) {
+		report_skip("pmevtyper: all events within [0x0, 0x3F] are supported");
+		return;
+	}
+
+	/* select counter 0 */
+	write_sysreg(0, PMSELR_EL0);
+	/* program this counter to count unsupported event */
+	write_sysreg(i, PMXEVCNTR_EL0);
+	/* read the counter value */
+	read_sysreg(PMXEVCNTR_EL0);
+	report(read_sysreg(PMXEVCNTR_EL0) == i,
+		"read of a counter programmed with unsupported event");
+
+}
+
+static bool satisfy_prerequisites(uint32_t *events, unsigned int nb_events)
+{
+	int i;
+
+	if (pmu.nb_implemented_counters < nb_events) {
+		report_skip("Skip test as number of counters is too small (%d)",
+			    pmu.nb_implemented_counters);
+		return false;
+	}
+
+	for (i = 0; i < nb_events; i++) {
+		if (!is_event_supported(events[i], false)) {
+			report_skip("Skip test as event 0x%x is not supported",
+				    events[i]);
+			return false;
+		}
+	}
+	return true;
+}
+
+static void test_basic_event_count(void)
+{
+	uint32_t implemented_counter_mask, non_implemented_counter_mask;
+	uint32_t counter_mask;
+	uint32_t events[] = {CPU_CYCLES, INST_RETIRED};
+
+	if (!satisfy_prerequisites(events, ARRAY_SIZE(events)))
+		return;
+
+	implemented_counter_mask = BIT(pmu.nb_implemented_counters) - 1;
+	non_implemented_counter_mask = ~(BIT(31) | implemented_counter_mask);
+	counter_mask = implemented_counter_mask | non_implemented_counter_mask;
+
+	write_regn_el0(pmevtyper, 0, CPU_CYCLES | PMEVTYPER_EXCLUDE_EL0);
+	write_regn_el0(pmevtyper, 1, INST_RETIRED | PMEVTYPER_EXCLUDE_EL0);
+
+	/* disable all counters */
+	write_sysreg_s(ALL_SET, PMCNTENCLR_EL0);
+	report(!read_sysreg_s(PMCNTENCLR_EL0) && !read_sysreg_s(PMCNTENSET_EL0),
+		"pmcntenclr: disable all counters");
+
+	/*
+	 * clear cycle and all event counters and allow counter enablement
+	 * through PMCNTENSET. LC is RES1.
+	 */
+	set_pmcr(pmu.pmcr_ro | PMU_PMCR_LC | PMU_PMCR_C | PMU_PMCR_P);
+	isb();
+	report(get_pmcr() == (pmu.pmcr_ro | PMU_PMCR_LC), "pmcr: reset counters");
+
+	/* Preset counter #0 to pre overflow value to trigger an overflow */
+	write_regn_el0(pmevcntr, 0, PRE_OVERFLOW);
+	report(read_regn_el0(pmevcntr, 0) == PRE_OVERFLOW,
+		"counter #0 preset to pre-overflow value");
+	report(!read_regn_el0(pmevcntr, 1), "counter #1 is 0");
+
+	/*
+	 * Enable all implemented counters and also attempt to enable
+	 * not supported counters. Counting still is disabled by !PMCR.E
+	 */
+	write_sysreg_s(counter_mask, PMCNTENSET_EL0);
+
+	/* check only those implemented are enabled */
+	report((read_sysreg_s(PMCNTENSET_EL0) == read_sysreg_s(PMCNTENCLR_EL0)) &&
+		(read_sysreg_s(PMCNTENSET_EL0) == implemented_counter_mask),
+		"pmcntenset: enabled implemented_counters");
+
+	/* Disable all counters but counters #0 and #1 */
+	write_sysreg_s(~0x3, PMCNTENCLR_EL0);
+	report((read_sysreg_s(PMCNTENSET_EL0) == read_sysreg_s(PMCNTENCLR_EL0)) &&
+		(read_sysreg_s(PMCNTENSET_EL0) == 0x3),
+		"pmcntenset: just enabled #0 and #1");
+
+	/* clear overflow register */
+	write_sysreg(ALL_SET, pmovsclr_el0);
+	report(!read_sysreg(pmovsclr_el0), "check overflow reg is 0");
+
+	/* disable overflow interrupts on all counters*/
+	write_sysreg(ALL_SET, pmintenclr_el1);
+	report(!read_sysreg(pmintenclr_el1),
+		"pmintenclr_el1=0, all interrupts disabled");
+
+	/* enable overflow interrupts on all event counters */
+	write_sysreg(implemented_counter_mask | non_implemented_counter_mask,
+		     pmintenset_el1);
+	report(read_sysreg(pmintenset_el1) == implemented_counter_mask,
+		"overflow interrupts enabled on all implemented counters");
+
+	/* Set PMCR.E, execute asm code and unset PMCR.E */
+	precise_instrs_loop(20, pmu.pmcr_ro | PMU_PMCR_E);
+
+	report_info("counter #0 is 0x%lx (CPU_CYCLES)",
+		    read_regn_el0(pmevcntr, 0));
+	report_info("counter #1 is 0x%lx (INST_RETIRED)",
+		    read_regn_el0(pmevcntr, 1));
+
+	report_info("overflow reg = 0x%lx", read_sysreg(pmovsclr_el0));
+	report(read_sysreg(pmovsclr_el0) & 0x1,
+		"check overflow happened on #0 only");
+}
+
+static void test_mem_access(void)
+{
+	void *addr = malloc(PAGE_SIZE);
+	uint32_t events[] = {MEM_ACCESS, MEM_ACCESS};
+
+	if (!satisfy_prerequisites(events, ARRAY_SIZE(events)))
+		return;
+
+	pmu_reset();
+
+	write_regn_el0(pmevtyper, 0, MEM_ACCESS | PMEVTYPER_EXCLUDE_EL0);
+	write_regn_el0(pmevtyper, 1, MEM_ACCESS | PMEVTYPER_EXCLUDE_EL0);
+	write_sysreg_s(0x3, PMCNTENSET_EL0);
+	isb();
+	mem_access_loop(addr, 20, pmu.pmcr_ro | PMU_PMCR_E);
+	report_info("counter #0 is %ld (MEM_ACCESS)", read_regn_el0(pmevcntr, 0));
+	report_info("counter #1 is %ld (MEM_ACCESS)", read_regn_el0(pmevcntr, 1));
+	/* We may measure more than 20 mem access depending on the core */
+	report((read_regn_el0(pmevcntr, 0) == read_regn_el0(pmevcntr, 1)) &&
+	       (read_regn_el0(pmevcntr, 0) >= 20) && !read_sysreg(pmovsclr_el0),
+	       "Ran 20 mem accesses");
+
+	pmu_reset();
+
+	write_regn_el0(pmevcntr, 0, PRE_OVERFLOW);
+	write_regn_el0(pmevcntr, 1, PRE_OVERFLOW);
+	write_sysreg_s(0x3, PMCNTENSET_EL0);
+	isb();
+	mem_access_loop(addr, 20, pmu.pmcr_ro | PMU_PMCR_E);
+	report(read_sysreg(pmovsclr_el0) == 0x3,
+	       "Ran 20 mem accesses with expected overflows on both counters");
+	report_info("cnt#0 = %ld cnt#1=%ld overflow=0x%lx",
+			read_regn_el0(pmevcntr, 0), read_regn_el0(pmevcntr, 1),
+			read_sysreg(pmovsclr_el0));
+}
+
 #endif
 
 /*
@@ -402,6 +638,18 @@ int main(int argc, char *argv[])
 		report_prefix_push(argv[1]);
 		test_event_introspection();
 		report_prefix_pop();
+	} else if (strcmp(argv[1], "pmu-event-counter-config") == 0) {
+		report_prefix_push(argv[1]);
+		test_event_counter_config();
+		report_prefix_pop();
+	} else if (strcmp(argv[1], "pmu-basic-event-count") == 0) {
+		report_prefix_push(argv[1]);
+		test_basic_event_count();
+		report_prefix_pop();
+	} else if (strcmp(argv[1], "pmu-mem-access") == 0) {
+		report_prefix_push(argv[1]);
+		test_mem_access();
+		report_prefix_pop();
 	} else {
 		report_abort("Unknown sub-test '%s'", argv[1]);
 	}
diff --git a/arm/unittests.cfg b/arm/unittests.cfg
index f993548..32ab8c6 100644
--- a/arm/unittests.cfg
+++ b/arm/unittests.cfg
@@ -72,6 +72,24 @@ groups = pmu
 arch = arm64
 extra_params = -append 'pmu-event-introspection'
 
+[pmu-event-counter-config]
+file = pmu.flat
+groups = pmu
+arch = arm64
+extra_params = -append 'pmu-event-counter-config'
+
+[pmu-basic-event-count]
+file = pmu.flat
+groups = pmu
+arch = arm64
+extra_params = -append 'pmu-basic-event-count'
+
+[pmu-mem-access]
+file = pmu.flat
+groups = pmu
+arch = arm64
+extra_params = -append 'pmu-mem-access'
+
 # Test PMU support (TCG) with -icount IPC=1
 #[pmu-tcg-icount-1]
 #file = pmu.flat
diff --git a/lib/arm64/asm/sysreg.h b/lib/arm64/asm/sysreg.h
index a45eebd..378bf7e 100644
--- a/lib/arm64/asm/sysreg.h
+++ b/lib/arm64/asm/sysreg.h
@@ -49,6 +49,12 @@
 	asm volatile("msr_s " xstr(r) ", %x0" : : "rZ" (__val));\
 } while (0)
 
+#define write_regn_el0(__reg, __n, __val) \
+	write_sysreg((__val), __reg ## __n ## _el0)
+
+#define read_regn_el0(__reg, __n) \
+	read_sysreg(__reg ## __n ## _el0)
+
 asm(
 "	.irp	num,0,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30\n"
 "	.equ	.L__reg_num_x\\num, \\num\n"
-- 
2.20.1

_______________________________________________
kvmarm mailing list
kvmarm@lists.cs.columbia.edu
https://lists.cs.columbia.edu/mailman/listinfo/kvmarm

[kvm-unit-tests PATCH v3 08/12] arm: pmu: Test SW_INCR event count

[Eric Auger]

Add tests dedicated to SW_INCR event counting.

Signed-off-by: Eric Auger <eric.auger@redhat.com>

---

v3: new
- Formerly in chained counter tests but as QEMU does not
  support chained counters, the whole test was failing. Peter
  split the test.
---
 arm/pmu.c         | 47 +++++++++++++++++++++++++++++++++++++++++++++++
 arm/unittests.cfg |  6 ++++++
 2 files changed, 53 insertions(+)

diff --git a/arm/pmu.c b/arm/pmu.c
index 45dccf7..c954c71 100644
--- a/arm/pmu.c
+++ b/arm/pmu.c
@@ -140,6 +140,7 @@ static void test_event_introspection(void) {}
 static void test_event_counter_config(void) {}
 static void test_basic_event_count(void) {}
 static void test_mem_access(void) {}
+static void test_sw_incr(void) {}
 
 #elif defined(__aarch64__)
 #define ID_AA64DFR0_PERFMON_SHIFT 8
@@ -464,6 +465,48 @@ static void test_mem_access(void)
 			read_sysreg(pmovsclr_el0));
 }
 
+static void test_sw_incr(void)
+{
+	uint32_t events[] = {SW_INCR, SW_INCR};
+	int i;
+
+	if (!satisfy_prerequisites(events, ARRAY_SIZE(events)))
+		return;
+
+	pmu_reset();
+
+	write_regn_el0(pmevtyper, 0, SW_INCR | PMEVTYPER_EXCLUDE_EL0);
+	write_regn_el0(pmevtyper, 1, SW_INCR | PMEVTYPER_EXCLUDE_EL0);
+	/* enable counters #0 and #1 */
+	write_sysreg_s(0x3, PMCNTENSET_EL0);
+
+	write_regn_el0(pmevcntr, 0, PRE_OVERFLOW);
+
+	for (i = 0; i < 100; i++)
+		write_sysreg(0x1, pmswinc_el0);
+
+	report_info("SW_INCR counter #0 has value %ld", read_regn_el0(pmevcntr, 0));
+	report(read_regn_el0(pmevcntr, 0) == PRE_OVERFLOW,
+		"PWSYNC does not increment if PMCR.E is unset");
+
+	pmu_reset();
+
+	write_regn_el0(pmevcntr, 0, PRE_OVERFLOW);
+	write_sysreg_s(0x3, PMCNTENSET_EL0);
+	set_pmcr(pmu.pmcr_ro | PMU_PMCR_E);
+
+	for (i = 0; i < 100; i++)
+		write_sysreg(0x3, pmswinc_el0);
+
+	report(read_regn_el0(pmevcntr, 0)  == 84, "counter #1 after + 100 SW_INCR");
+	report(read_regn_el0(pmevcntr, 1)  == 100,
+		"counter #0 after + 100 SW_INCR");
+	report_info("counter values after 100 SW_INCR #0=%ld #1=%ld",
+		    read_regn_el0(pmevcntr, 0), read_regn_el0(pmevcntr, 1));
+	report(read_sysreg(pmovsclr_el0) == 0x1,
+		"overflow reg after 100 SW_INCR");
+}
+
 #endif
 
 /*
@@ -650,6 +693,10 @@ int main(int argc, char *argv[])
 		report_prefix_push(argv[1]);
 		test_mem_access();
 		report_prefix_pop();
+	} else if (strcmp(argv[1], "pmu-sw-incr") == 0) {
+		report_prefix_push(argv[1]);
+		test_sw_incr();
+		report_prefix_pop();
 	} else {
 		report_abort("Unknown sub-test '%s'", argv[1]);
 	}
diff --git a/arm/unittests.cfg b/arm/unittests.cfg
index 32ab8c6..175afe6 100644
--- a/arm/unittests.cfg
+++ b/arm/unittests.cfg
@@ -90,6 +90,12 @@ groups = pmu
 arch = arm64
 extra_params = -append 'pmu-mem-access'
 
+[pmu-sw-incr]
+file = pmu.flat
+groups = pmu
+arch = arm64
+extra_params = -append 'pmu-sw-incr'
+
 # Test PMU support (TCG) with -icount IPC=1
 #[pmu-tcg-icount-1]
 #file = pmu.flat
-- 
2.20.1

[kvm-unit-tests PATCH v3 08/12] arm: pmu: Test SW_INCR event count

[Eric Auger]

Add tests dedicated to SW_INCR event counting.

Signed-off-by: Eric Auger <eric.auger@redhat.com>

---

v3: new
- Formerly in chained counter tests but as QEMU does not
  support chained counters, the whole test was failing. Peter
  split the test.
---
 arm/pmu.c         | 47 +++++++++++++++++++++++++++++++++++++++++++++++
 arm/unittests.cfg |  6 ++++++
 2 files changed, 53 insertions(+)

diff --git a/arm/pmu.c b/arm/pmu.c
index 45dccf7..c954c71 100644
--- a/arm/pmu.c
+++ b/arm/pmu.c
@@ -140,6 +140,7 @@ static void test_event_introspection(void) {}
 static void test_event_counter_config(void) {}
 static void test_basic_event_count(void) {}
 static void test_mem_access(void) {}
+static void test_sw_incr(void) {}
 
 #elif defined(__aarch64__)
 #define ID_AA64DFR0_PERFMON_SHIFT 8
@@ -464,6 +465,48 @@ static void test_mem_access(void)
 			read_sysreg(pmovsclr_el0));
 }
 
+static void test_sw_incr(void)
+{
+	uint32_t events[] = {SW_INCR, SW_INCR};
+	int i;
+
+	if (!satisfy_prerequisites(events, ARRAY_SIZE(events)))
+		return;
+
+	pmu_reset();
+
+	write_regn_el0(pmevtyper, 0, SW_INCR | PMEVTYPER_EXCLUDE_EL0);
+	write_regn_el0(pmevtyper, 1, SW_INCR | PMEVTYPER_EXCLUDE_EL0);
+	/* enable counters #0 and #1 */
+	write_sysreg_s(0x3, PMCNTENSET_EL0);
+
+	write_regn_el0(pmevcntr, 0, PRE_OVERFLOW);
+
+	for (i = 0; i < 100; i++)
+		write_sysreg(0x1, pmswinc_el0);
+
+	report_info("SW_INCR counter #0 has value %ld", read_regn_el0(pmevcntr, 0));
+	report(read_regn_el0(pmevcntr, 0) == PRE_OVERFLOW,
+		"PWSYNC does not increment if PMCR.E is unset");
+
+	pmu_reset();
+
+	write_regn_el0(pmevcntr, 0, PRE_OVERFLOW);
+	write_sysreg_s(0x3, PMCNTENSET_EL0);
+	set_pmcr(pmu.pmcr_ro | PMU_PMCR_E);
+
+	for (i = 0; i < 100; i++)
+		write_sysreg(0x3, pmswinc_el0);
+
+	report(read_regn_el0(pmevcntr, 0)  == 84, "counter #1 after + 100 SW_INCR");
+	report(read_regn_el0(pmevcntr, 1)  == 100,
+		"counter #0 after + 100 SW_INCR");
+	report_info("counter values after 100 SW_INCR #0=%ld #1=%ld",
+		    read_regn_el0(pmevcntr, 0), read_regn_el0(pmevcntr, 1));
+	report(read_sysreg(pmovsclr_el0) == 0x1,
+		"overflow reg after 100 SW_INCR");
+}
+
 #endif
 
 /*
@@ -650,6 +693,10 @@ int main(int argc, char *argv[])
 		report_prefix_push(argv[1]);
 		test_mem_access();
 		report_prefix_pop();
+	} else if (strcmp(argv[1], "pmu-sw-incr") == 0) {
+		report_prefix_push(argv[1]);
+		test_sw_incr();
+		report_prefix_pop();
 	} else {
 		report_abort("Unknown sub-test '%s'", argv[1]);
 	}
diff --git a/arm/unittests.cfg b/arm/unittests.cfg
index 32ab8c6..175afe6 100644
--- a/arm/unittests.cfg
+++ b/arm/unittests.cfg
@@ -90,6 +90,12 @@ groups = pmu
 arch = arm64
 extra_params = -append 'pmu-mem-access'
 
+[pmu-sw-incr]
+file = pmu.flat
+groups = pmu
+arch = arm64
+extra_params = -append 'pmu-sw-incr'
+
 # Test PMU support (TCG) with -icount IPC=1
 #[pmu-tcg-icount-1]
 #file = pmu.flat
-- 
2.20.1

[kvm-unit-tests PATCH v3 08/12] arm: pmu: Test SW_INCR event count

[Eric Auger]

Add tests dedicated to SW_INCR event counting.

Signed-off-by: Eric Auger <eric.auger@redhat.com>

---

v3: new
- Formerly in chained counter tests but as QEMU does not
  support chained counters, the whole test was failing. Peter
  split the test.
---
 arm/pmu.c         | 47 +++++++++++++++++++++++++++++++++++++++++++++++
 arm/unittests.cfg |  6 ++++++
 2 files changed, 53 insertions(+)

diff --git a/arm/pmu.c b/arm/pmu.c
index 45dccf7..c954c71 100644
--- a/arm/pmu.c
+++ b/arm/pmu.c
@@ -140,6 +140,7 @@ static void test_event_introspection(void) {}
 static void test_event_counter_config(void) {}
 static void test_basic_event_count(void) {}
 static void test_mem_access(void) {}
+static void test_sw_incr(void) {}
 
 #elif defined(__aarch64__)
 #define ID_AA64DFR0_PERFMON_SHIFT 8
@@ -464,6 +465,48 @@ static void test_mem_access(void)
 			read_sysreg(pmovsclr_el0));
 }
 
+static void test_sw_incr(void)
+{
+	uint32_t events[] = {SW_INCR, SW_INCR};
+	int i;
+
+	if (!satisfy_prerequisites(events, ARRAY_SIZE(events)))
+		return;
+
+	pmu_reset();
+
+	write_regn_el0(pmevtyper, 0, SW_INCR | PMEVTYPER_EXCLUDE_EL0);
+	write_regn_el0(pmevtyper, 1, SW_INCR | PMEVTYPER_EXCLUDE_EL0);
+	/* enable counters #0 and #1 */
+	write_sysreg_s(0x3, PMCNTENSET_EL0);
+
+	write_regn_el0(pmevcntr, 0, PRE_OVERFLOW);
+
+	for (i = 0; i < 100; i++)
+		write_sysreg(0x1, pmswinc_el0);
+
+	report_info("SW_INCR counter #0 has value %ld", read_regn_el0(pmevcntr, 0));
+	report(read_regn_el0(pmevcntr, 0) == PRE_OVERFLOW,
+		"PWSYNC does not increment if PMCR.E is unset");
+
+	pmu_reset();
+
+	write_regn_el0(pmevcntr, 0, PRE_OVERFLOW);
+	write_sysreg_s(0x3, PMCNTENSET_EL0);
+	set_pmcr(pmu.pmcr_ro | PMU_PMCR_E);
+
+	for (i = 0; i < 100; i++)
+		write_sysreg(0x3, pmswinc_el0);
+
+	report(read_regn_el0(pmevcntr, 0)  == 84, "counter #1 after + 100 SW_INCR");
+	report(read_regn_el0(pmevcntr, 1)  == 100,
+		"counter #0 after + 100 SW_INCR");
+	report_info("counter values after 100 SW_INCR #0=%ld #1=%ld",
+		    read_regn_el0(pmevcntr, 0), read_regn_el0(pmevcntr, 1));
+	report(read_sysreg(pmovsclr_el0) == 0x1,
+		"overflow reg after 100 SW_INCR");
+}
+
 #endif
 
 /*
@@ -650,6 +693,10 @@ int main(int argc, char *argv[])
 		report_prefix_push(argv[1]);
 		test_mem_access();
 		report_prefix_pop();
+	} else if (strcmp(argv[1], "pmu-sw-incr") == 0) {
+		report_prefix_push(argv[1]);
+		test_sw_incr();
+		report_prefix_pop();
 	} else {
 		report_abort("Unknown sub-test '%s'", argv[1]);
 	}
diff --git a/arm/unittests.cfg b/arm/unittests.cfg
index 32ab8c6..175afe6 100644
--- a/arm/unittests.cfg
+++ b/arm/unittests.cfg
@@ -90,6 +90,12 @@ groups = pmu
 arch = arm64
 extra_params = -append 'pmu-mem-access'
 
+[pmu-sw-incr]
+file = pmu.flat
+groups = pmu
+arch = arm64
+extra_params = -append 'pmu-sw-incr'
+
 # Test PMU support (TCG) with -icount IPC=1
 #[pmu-tcg-icount-1]
 #file = pmu.flat
-- 
2.20.1

_______________________________________________
kvmarm mailing list
kvmarm@lists.cs.columbia.edu
https://lists.cs.columbia.edu/mailman/listinfo/kvmarm

[kvm-unit-tests PATCH v3 09/12] arm: pmu: Test chained counters

[Eric Auger]

Add 2 tests exercising chained counters. The first one uses
CPU_CYCLES and the second one uses SW_INCR.

Signed-off-by: Eric Auger <eric.auger@redhat.com>

---

v2 -> v3:
- added prefix pop
- added pmu prefix to the test names
- defines, event array ...
---
 arm/pmu.c         | 97 +++++++++++++++++++++++++++++++++++++++++++++++
 arm/unittests.cfg | 12 ++++++
 2 files changed, 109 insertions(+)

diff --git a/arm/pmu.c b/arm/pmu.c
index c954c71..ccca31b 100644
--- a/arm/pmu.c
+++ b/arm/pmu.c
@@ -47,6 +47,7 @@
 #define INST_PREC		0x1B
 #define STALL_FRONTEND		0x23
 #define STALL_BACKEND		0x24
+#define CHAIN			0x1E
 
 #define COMMON_EVENTS_LOW	0x0
 #define COMMON_EVENTS_HIGH	0x3F
@@ -141,6 +142,8 @@ static void test_event_counter_config(void) {}
 static void test_basic_event_count(void) {}
 static void test_mem_access(void) {}
 static void test_sw_incr(void) {}
+static void test_chained_counters(void) {}
+static void test_chained_sw_incr(void) {}
 
 #elif defined(__aarch64__)
 #define ID_AA64DFR0_PERFMON_SHIFT 8
@@ -507,6 +510,92 @@ static void test_sw_incr(void)
 		"overflow reg after 100 SW_INCR");
 }
 
+static void test_chained_counters(void)
+{
+	uint32_t events[] = {CPU_CYCLES, CHAIN};
+
+	if (!satisfy_prerequisites(events, ARRAY_SIZE(events)))
+		return;
+
+	pmu_reset();
+
+	write_regn_el0(pmevtyper, 0, CPU_CYCLES | PMEVTYPER_EXCLUDE_EL0);
+	write_regn_el0(pmevtyper, 1, CHAIN | PMEVTYPER_EXCLUDE_EL0);
+	/* enable counters #0 and #1 */
+	write_sysreg_s(0x3, PMCNTENSET_EL0);
+	write_regn_el0(pmevcntr, 0, PRE_OVERFLOW);
+
+	precise_instrs_loop(22, pmu.pmcr_ro | PMU_PMCR_E);
+
+	report(read_regn_el0(pmevcntr, 1) == 1, "CHAIN counter #1 incremented");
+	report(!read_sysreg(pmovsclr_el0), "check no overflow is recorded");
+
+	/* test 64b overflow */
+
+	pmu_reset();
+	write_sysreg_s(0x3, PMCNTENSET_EL0);
+
+	write_regn_el0(pmevcntr, 0, PRE_OVERFLOW);
+	write_regn_el0(pmevcntr, 1, 0x1);
+	precise_instrs_loop(22, pmu.pmcr_ro | PMU_PMCR_E);
+	report_info("overflow reg = 0x%lx", read_sysreg(pmovsclr_el0));
+	report(read_regn_el0(pmevcntr, 1) == 2, "CHAIN counter #1 incremented");
+	report(!read_sysreg(pmovsclr_el0), "check no overflow is recorded");
+
+	write_regn_el0(pmevcntr, 0, PRE_OVERFLOW);
+	write_regn_el0(pmevcntr, 1, ALL_SET);
+
+	precise_instrs_loop(22, pmu.pmcr_ro | PMU_PMCR_E);
+	report_info("overflow reg = 0x%lx", read_sysreg(pmovsclr_el0));
+	report(!read_regn_el0(pmevcntr, 1), "CHAIN counter #1 wrapped");
+	report(read_sysreg(pmovsclr_el0) == 0x2,
+		"check no overflow is recorded");
+}
+
+static void test_chained_sw_incr(void)
+{
+	uint32_t events[] = {SW_INCR, CHAIN};
+	int i;
+
+	if (!satisfy_prerequisites(events, ARRAY_SIZE(events)))
+		return;
+
+	pmu_reset();
+
+	write_regn_el0(pmevtyper, 0, SW_INCR | PMEVTYPER_EXCLUDE_EL0);
+	write_regn_el0(pmevtyper, 1, CHAIN | PMEVTYPER_EXCLUDE_EL0);
+	/* enable counters #0 and #1 */
+	write_sysreg_s(0x3, PMCNTENSET_EL0);
+
+	write_regn_el0(pmevcntr, 0, PRE_OVERFLOW);
+	set_pmcr(pmu.pmcr_ro | PMU_PMCR_E);
+	for (i = 0; i < 100; i++)
+		write_sysreg(0x1, pmswinc_el0);
+
+	report(!read_sysreg(pmovsclr_el0) && (read_regn_el0(pmevcntr, 1) == 1),
+		"overflow reg after 100 SW_INCR/CHAIN");
+	report_info("overflow=0x%lx, #0=%ld #1=%ld", read_sysreg(pmovsclr_el0),
+		    read_regn_el0(pmevcntr, 0), read_regn_el0(pmevcntr, 1));
+
+	/* 64b SW_INCR and overflow on CHAIN counter*/
+	pmu_reset();
+
+	write_regn_el0(pmevtyper, 1, events[1] | PMEVTYPER_EXCLUDE_EL0);
+	write_regn_el0(pmevcntr, 0, PRE_OVERFLOW);
+	write_regn_el0(pmevcntr, 1, ALL_SET);
+	write_sysreg_s(0x3, PMCNTENSET_EL0);
+	set_pmcr(pmu.pmcr_ro | PMU_PMCR_E);
+	for (i = 0; i < 100; i++)
+		write_sysreg(0x1, pmswinc_el0);
+
+	report((read_sysreg(pmovsclr_el0) == 0x2) &&
+		(read_regn_el0(pmevcntr, 1) == 0) &&
+		(read_regn_el0(pmevcntr, 0) == 84),
+		"overflow reg after 100 SW_INCR/CHAIN");
+	report_info("overflow=0x%lx, #0=%ld #1=%ld", read_sysreg(pmovsclr_el0),
+		    read_regn_el0(pmevcntr, 0), read_regn_el0(pmevcntr, 1));
+}
+
 #endif
 
 /*
@@ -697,6 +786,14 @@ int main(int argc, char *argv[])
 		report_prefix_push(argv[1]);
 		test_sw_incr();
 		report_prefix_pop();
+	} else if (strcmp(argv[1], "pmu-chained-counters") == 0) {
+		report_prefix_push(argv[1]);
+		test_chained_counters();
+		report_prefix_pop();
+	} else if (strcmp(argv[1], "pmu-chained-sw-incr") == 0) {
+		report_prefix_push(argv[1]);
+		test_chained_sw_incr();
+		report_prefix_pop();
 	} else {
 		report_abort("Unknown sub-test '%s'", argv[1]);
 	}
diff --git a/arm/unittests.cfg b/arm/unittests.cfg
index 175afe6..d31dcbf 100644
--- a/arm/unittests.cfg
+++ b/arm/unittests.cfg
@@ -96,6 +96,18 @@ groups = pmu
 arch = arm64
 extra_params = -append 'pmu-sw-incr'
 
+[pmu-chained-counters]
+file = pmu.flat
+groups = pmu
+arch = arm64
+extra_params = -append 'pmu-chained-counters'
+
+[pmu-chained-sw-incr]
+file = pmu.flat
+groups = pmu
+arch = arm64
+extra_params = -append 'pmu-chained-sw-incr'
+
 # Test PMU support (TCG) with -icount IPC=1
 #[pmu-tcg-icount-1]
 #file = pmu.flat
-- 
2.20.1

[kvm-unit-tests PATCH v3 09/12] arm: pmu: Test chained counters

[Eric Auger]

Add 2 tests exercising chained counters. The first one uses
CPU_CYCLES and the second one uses SW_INCR.

Signed-off-by: Eric Auger <eric.auger@redhat.com>

---

v2 -> v3:
- added prefix pop
- added pmu prefix to the test names
- defines, event array ...
---
 arm/pmu.c         | 97 +++++++++++++++++++++++++++++++++++++++++++++++
 arm/unittests.cfg | 12 ++++++
 2 files changed, 109 insertions(+)

diff --git a/arm/pmu.c b/arm/pmu.c
index c954c71..ccca31b 100644
--- a/arm/pmu.c
+++ b/arm/pmu.c
@@ -47,6 +47,7 @@
 #define INST_PREC		0x1B
 #define STALL_FRONTEND		0x23
 #define STALL_BACKEND		0x24
+#define CHAIN			0x1E
 
 #define COMMON_EVENTS_LOW	0x0
 #define COMMON_EVENTS_HIGH	0x3F
@@ -141,6 +142,8 @@ static void test_event_counter_config(void) {}
 static void test_basic_event_count(void) {}
 static void test_mem_access(void) {}
 static void test_sw_incr(void) {}
+static void test_chained_counters(void) {}
+static void test_chained_sw_incr(void) {}
 
 #elif defined(__aarch64__)
 #define ID_AA64DFR0_PERFMON_SHIFT 8
@@ -507,6 +510,92 @@ static void test_sw_incr(void)
 		"overflow reg after 100 SW_INCR");
 }
 
+static void test_chained_counters(void)
+{
+	uint32_t events[] = {CPU_CYCLES, CHAIN};
+
+	if (!satisfy_prerequisites(events, ARRAY_SIZE(events)))
+		return;
+
+	pmu_reset();
+
+	write_regn_el0(pmevtyper, 0, CPU_CYCLES | PMEVTYPER_EXCLUDE_EL0);
+	write_regn_el0(pmevtyper, 1, CHAIN | PMEVTYPER_EXCLUDE_EL0);
+	/* enable counters #0 and #1 */
+	write_sysreg_s(0x3, PMCNTENSET_EL0);
+	write_regn_el0(pmevcntr, 0, PRE_OVERFLOW);
+
+	precise_instrs_loop(22, pmu.pmcr_ro | PMU_PMCR_E);
+
+	report(read_regn_el0(pmevcntr, 1) == 1, "CHAIN counter #1 incremented");
+	report(!read_sysreg(pmovsclr_el0), "check no overflow is recorded");
+
+	/* test 64b overflow */
+
+	pmu_reset();
+	write_sysreg_s(0x3, PMCNTENSET_EL0);
+
+	write_regn_el0(pmevcntr, 0, PRE_OVERFLOW);
+	write_regn_el0(pmevcntr, 1, 0x1);
+	precise_instrs_loop(22, pmu.pmcr_ro | PMU_PMCR_E);
+	report_info("overflow reg = 0x%lx", read_sysreg(pmovsclr_el0));
+	report(read_regn_el0(pmevcntr, 1) == 2, "CHAIN counter #1 incremented");
+	report(!read_sysreg(pmovsclr_el0), "check no overflow is recorded");
+
+	write_regn_el0(pmevcntr, 0, PRE_OVERFLOW);
+	write_regn_el0(pmevcntr, 1, ALL_SET);
+
+	precise_instrs_loop(22, pmu.pmcr_ro | PMU_PMCR_E);
+	report_info("overflow reg = 0x%lx", read_sysreg(pmovsclr_el0));
+	report(!read_regn_el0(pmevcntr, 1), "CHAIN counter #1 wrapped");
+	report(read_sysreg(pmovsclr_el0) == 0x2,
+		"check no overflow is recorded");
+}
+
+static void test_chained_sw_incr(void)
+{
+	uint32_t events[] = {SW_INCR, CHAIN};
+	int i;
+
+	if (!satisfy_prerequisites(events, ARRAY_SIZE(events)))
+		return;
+
+	pmu_reset();
+
+	write_regn_el0(pmevtyper, 0, SW_INCR | PMEVTYPER_EXCLUDE_EL0);
+	write_regn_el0(pmevtyper, 1, CHAIN | PMEVTYPER_EXCLUDE_EL0);
+	/* enable counters #0 and #1 */
+	write_sysreg_s(0x3, PMCNTENSET_EL0);
+
+	write_regn_el0(pmevcntr, 0, PRE_OVERFLOW);
+	set_pmcr(pmu.pmcr_ro | PMU_PMCR_E);
+	for (i = 0; i < 100; i++)
+		write_sysreg(0x1, pmswinc_el0);
+
+	report(!read_sysreg(pmovsclr_el0) && (read_regn_el0(pmevcntr, 1) == 1),
+		"overflow reg after 100 SW_INCR/CHAIN");
+	report_info("overflow=0x%lx, #0=%ld #1=%ld", read_sysreg(pmovsclr_el0),
+		    read_regn_el0(pmevcntr, 0), read_regn_el0(pmevcntr, 1));
+
+	/* 64b SW_INCR and overflow on CHAIN counter*/
+	pmu_reset();
+
+	write_regn_el0(pmevtyper, 1, events[1] | PMEVTYPER_EXCLUDE_EL0);
+	write_regn_el0(pmevcntr, 0, PRE_OVERFLOW);
+	write_regn_el0(pmevcntr, 1, ALL_SET);
+	write_sysreg_s(0x3, PMCNTENSET_EL0);
+	set_pmcr(pmu.pmcr_ro | PMU_PMCR_E);
+	for (i = 0; i < 100; i++)
+		write_sysreg(0x1, pmswinc_el0);
+
+	report((read_sysreg(pmovsclr_el0) == 0x2) &&
+		(read_regn_el0(pmevcntr, 1) == 0) &&
+		(read_regn_el0(pmevcntr, 0) == 84),
+		"overflow reg after 100 SW_INCR/CHAIN");
+	report_info("overflow=0x%lx, #0=%ld #1=%ld", read_sysreg(pmovsclr_el0),
+		    read_regn_el0(pmevcntr, 0), read_regn_el0(pmevcntr, 1));
+}
+
 #endif
 
 /*
@@ -697,6 +786,14 @@ int main(int argc, char *argv[])
 		report_prefix_push(argv[1]);
 		test_sw_incr();
 		report_prefix_pop();
+	} else if (strcmp(argv[1], "pmu-chained-counters") == 0) {
+		report_prefix_push(argv[1]);
+		test_chained_counters();
+		report_prefix_pop();
+	} else if (strcmp(argv[1], "pmu-chained-sw-incr") == 0) {
+		report_prefix_push(argv[1]);
+		test_chained_sw_incr();
+		report_prefix_pop();
 	} else {
 		report_abort("Unknown sub-test '%s'", argv[1]);
 	}
diff --git a/arm/unittests.cfg b/arm/unittests.cfg
index 175afe6..d31dcbf 100644
--- a/arm/unittests.cfg
+++ b/arm/unittests.cfg
@@ -96,6 +96,18 @@ groups = pmu
 arch = arm64
 extra_params = -append 'pmu-sw-incr'
 
+[pmu-chained-counters]
+file = pmu.flat
+groups = pmu
+arch = arm64
+extra_params = -append 'pmu-chained-counters'
+
+[pmu-chained-sw-incr]
+file = pmu.flat
+groups = pmu
+arch = arm64
+extra_params = -append 'pmu-chained-sw-incr'
+
 # Test PMU support (TCG) with -icount IPC=1
 #[pmu-tcg-icount-1]
 #file = pmu.flat
-- 
2.20.1

[kvm-unit-tests PATCH v3 09/12] arm: pmu: Test chained counters

[Eric Auger]

Add 2 tests exercising chained counters. The first one uses
CPU_CYCLES and the second one uses SW_INCR.

Signed-off-by: Eric Auger <eric.auger@redhat.com>

---

v2 -> v3:
- added prefix pop
- added pmu prefix to the test names
- defines, event array ...
---
 arm/pmu.c         | 97 +++++++++++++++++++++++++++++++++++++++++++++++
 arm/unittests.cfg | 12 ++++++
 2 files changed, 109 insertions(+)

diff --git a/arm/pmu.c b/arm/pmu.c
index c954c71..ccca31b 100644
--- a/arm/pmu.c
+++ b/arm/pmu.c
@@ -47,6 +47,7 @@
 #define INST_PREC		0x1B
 #define STALL_FRONTEND		0x23
 #define STALL_BACKEND		0x24
+#define CHAIN			0x1E
 
 #define COMMON_EVENTS_LOW	0x0
 #define COMMON_EVENTS_HIGH	0x3F
@@ -141,6 +142,8 @@ static void test_event_counter_config(void) {}
 static void test_basic_event_count(void) {}
 static void test_mem_access(void) {}
 static void test_sw_incr(void) {}
+static void test_chained_counters(void) {}
+static void test_chained_sw_incr(void) {}
 
 #elif defined(__aarch64__)
 #define ID_AA64DFR0_PERFMON_SHIFT 8
@@ -507,6 +510,92 @@ static void test_sw_incr(void)
 		"overflow reg after 100 SW_INCR");
 }
 
+static void test_chained_counters(void)
+{
+	uint32_t events[] = {CPU_CYCLES, CHAIN};
+
+	if (!satisfy_prerequisites(events, ARRAY_SIZE(events)))
+		return;
+
+	pmu_reset();
+
+	write_regn_el0(pmevtyper, 0, CPU_CYCLES | PMEVTYPER_EXCLUDE_EL0);
+	write_regn_el0(pmevtyper, 1, CHAIN | PMEVTYPER_EXCLUDE_EL0);
+	/* enable counters #0 and #1 */
+	write_sysreg_s(0x3, PMCNTENSET_EL0);
+	write_regn_el0(pmevcntr, 0, PRE_OVERFLOW);
+
+	precise_instrs_loop(22, pmu.pmcr_ro | PMU_PMCR_E);
+
+	report(read_regn_el0(pmevcntr, 1) == 1, "CHAIN counter #1 incremented");
+	report(!read_sysreg(pmovsclr_el0), "check no overflow is recorded");
+
+	/* test 64b overflow */
+
+	pmu_reset();
+	write_sysreg_s(0x3, PMCNTENSET_EL0);
+
+	write_regn_el0(pmevcntr, 0, PRE_OVERFLOW);
+	write_regn_el0(pmevcntr, 1, 0x1);
+	precise_instrs_loop(22, pmu.pmcr_ro | PMU_PMCR_E);
+	report_info("overflow reg = 0x%lx", read_sysreg(pmovsclr_el0));
+	report(read_regn_el0(pmevcntr, 1) == 2, "CHAIN counter #1 incremented");
+	report(!read_sysreg(pmovsclr_el0), "check no overflow is recorded");
+
+	write_regn_el0(pmevcntr, 0, PRE_OVERFLOW);
+	write_regn_el0(pmevcntr, 1, ALL_SET);
+
+	precise_instrs_loop(22, pmu.pmcr_ro | PMU_PMCR_E);
+	report_info("overflow reg = 0x%lx", read_sysreg(pmovsclr_el0));
+	report(!read_regn_el0(pmevcntr, 1), "CHAIN counter #1 wrapped");
+	report(read_sysreg(pmovsclr_el0) == 0x2,
+		"check no overflow is recorded");
+}
+
+static void test_chained_sw_incr(void)
+{
+	uint32_t events[] = {SW_INCR, CHAIN};
+	int i;
+
+	if (!satisfy_prerequisites(events, ARRAY_SIZE(events)))
+		return;
+
+	pmu_reset();
+
+	write_regn_el0(pmevtyper, 0, SW_INCR | PMEVTYPER_EXCLUDE_EL0);
+	write_regn_el0(pmevtyper, 1, CHAIN | PMEVTYPER_EXCLUDE_EL0);
+	/* enable counters #0 and #1 */
+	write_sysreg_s(0x3, PMCNTENSET_EL0);
+
+	write_regn_el0(pmevcntr, 0, PRE_OVERFLOW);
+	set_pmcr(pmu.pmcr_ro | PMU_PMCR_E);
+	for (i = 0; i < 100; i++)
+		write_sysreg(0x1, pmswinc_el0);
+
+	report(!read_sysreg(pmovsclr_el0) && (read_regn_el0(pmevcntr, 1) == 1),
+		"overflow reg after 100 SW_INCR/CHAIN");
+	report_info("overflow=0x%lx, #0=%ld #1=%ld", read_sysreg(pmovsclr_el0),
+		    read_regn_el0(pmevcntr, 0), read_regn_el0(pmevcntr, 1));
+
+	/* 64b SW_INCR and overflow on CHAIN counter*/
+	pmu_reset();
+
+	write_regn_el0(pmevtyper, 1, events[1] | PMEVTYPER_EXCLUDE_EL0);
+	write_regn_el0(pmevcntr, 0, PRE_OVERFLOW);
+	write_regn_el0(pmevcntr, 1, ALL_SET);
+	write_sysreg_s(0x3, PMCNTENSET_EL0);
+	set_pmcr(pmu.pmcr_ro | PMU_PMCR_E);
+	for (i = 0; i < 100; i++)
+		write_sysreg(0x1, pmswinc_el0);
+
+	report((read_sysreg(pmovsclr_el0) == 0x2) &&
+		(read_regn_el0(pmevcntr, 1) == 0) &&
+		(read_regn_el0(pmevcntr, 0) == 84),
+		"overflow reg after 100 SW_INCR/CHAIN");
+	report_info("overflow=0x%lx, #0=%ld #1=%ld", read_sysreg(pmovsclr_el0),
+		    read_regn_el0(pmevcntr, 0), read_regn_el0(pmevcntr, 1));
+}
+
 #endif
 
 /*
@@ -697,6 +786,14 @@ int main(int argc, char *argv[])
 		report_prefix_push(argv[1]);
 		test_sw_incr();
 		report_prefix_pop();
+	} else if (strcmp(argv[1], "pmu-chained-counters") == 0) {
+		report_prefix_push(argv[1]);
+		test_chained_counters();
+		report_prefix_pop();
+	} else if (strcmp(argv[1], "pmu-chained-sw-incr") == 0) {
+		report_prefix_push(argv[1]);
+		test_chained_sw_incr();
+		report_prefix_pop();
 	} else {
 		report_abort("Unknown sub-test '%s'", argv[1]);
 	}
diff --git a/arm/unittests.cfg b/arm/unittests.cfg
index 175afe6..d31dcbf 100644
--- a/arm/unittests.cfg
+++ b/arm/unittests.cfg
@@ -96,6 +96,18 @@ groups = pmu
 arch = arm64
 extra_params = -append 'pmu-sw-incr'
 
+[pmu-chained-counters]
+file = pmu.flat
+groups = pmu
+arch = arm64
+extra_params = -append 'pmu-chained-counters'
+
+[pmu-chained-sw-incr]
+file = pmu.flat
+groups = pmu
+arch = arm64
+extra_params = -append 'pmu-chained-sw-incr'
+
 # Test PMU support (TCG) with -icount IPC=1
 #[pmu-tcg-icount-1]
 #file = pmu.flat
-- 
2.20.1

_______________________________________________
kvmarm mailing list
kvmarm@lists.cs.columbia.edu
https://lists.cs.columbia.edu/mailman/listinfo/kvmarm

[kvm-unit-tests PATCH v3 10/12] arm: pmu: test 32-bit <-> 64-bit transitions

[Eric Auger]

Test configurations where we transit from 32b to 64b
counters and conversely. Also tests configuration where
chain counters are configured but only one counter is
enabled.

Signed-off-by: Eric Auger <eric.auger@redhat.com>

---

v2 -> v3:
- added prefix pop
---
 arm/pmu.c         | 138 ++++++++++++++++++++++++++++++++++++++++++++++
 arm/unittests.cfg |   6 ++
 2 files changed, 144 insertions(+)

diff --git a/arm/pmu.c b/arm/pmu.c
index ccca31b..79c2a0f 100644
--- a/arm/pmu.c
+++ b/arm/pmu.c
@@ -57,6 +57,7 @@
 #define ALL_SET			0xFFFFFFFF
 #define ALL_CLEAR		0x0
 #define PRE_OVERFLOW		0xFFFFFFF0
+#define PRE_OVERFLOW2		0xFFFFFFDC
 
 struct pmu {
 	unsigned int version;
@@ -144,6 +145,7 @@ static void test_mem_access(void) {}
 static void test_sw_incr(void) {}
 static void test_chained_counters(void) {}
 static void test_chained_sw_incr(void) {}
+static void test_chain_promotion(void) {}
 
 #elif defined(__aarch64__)
 #define ID_AA64DFR0_PERFMON_SHIFT 8
@@ -596,6 +598,138 @@ static void test_chained_sw_incr(void)
 		    read_regn_el0(pmevcntr, 0), read_regn_el0(pmevcntr, 1));
 }
 
+static void test_chain_promotion(void)
+{
+	uint32_t events[] = {MEM_ACCESS, CHAIN};
+	void *addr = malloc(PAGE_SIZE);
+
+	if (!satisfy_prerequisites(events, ARRAY_SIZE(events)))
+		return;
+
+	/* Only enable CHAIN counter */
+	pmu_reset();
+	write_regn_el0(pmevtyper, 0, MEM_ACCESS | PMEVTYPER_EXCLUDE_EL0);
+	write_regn_el0(pmevtyper, 1, CHAIN | PMEVTYPER_EXCLUDE_EL0);
+	write_sysreg_s(0x2, PMCNTENSET_EL0);
+	isb();
+
+	mem_access_loop(addr, 20, pmu.pmcr_ro | PMU_PMCR_E);
+	report(!read_regn_el0(pmevcntr, 0),
+		"chain counter not counting if even counter is disabled");
+
+	/* Only enable even counter */
+	pmu_reset();
+	write_regn_el0(pmevcntr, 0, PRE_OVERFLOW);
+	write_sysreg_s(0x1, PMCNTENSET_EL0);
+	isb();
+
+	mem_access_loop(addr, 20, pmu.pmcr_ro | PMU_PMCR_E);
+	report(!read_regn_el0(pmevcntr, 1) && (read_sysreg(pmovsclr_el0) == 0x1),
+		"odd counter did not increment on overflow if disabled");
+	report_info("MEM_ACCESS counter #0 has value %ld",
+		    read_regn_el0(pmevcntr, 0));
+	report_info("CHAIN counter #1 has value %ld",
+		    read_regn_el0(pmevcntr, 1));
+	report_info("overflow counter %ld", read_sysreg(pmovsclr_el0));
+
+	/* start at 0xFFFFFFDC, +20 with CHAIN enabled, +20 with CHAIN disabled */
+	pmu_reset();
+	write_sysreg_s(0x3, PMCNTENSET_EL0);
+	write_regn_el0(pmevcntr, 0, PRE_OVERFLOW2);
+	isb();
+
+	mem_access_loop(addr, 20, pmu.pmcr_ro | PMU_PMCR_E);
+	report_info("MEM_ACCESS counter #0 has value 0x%lx",
+		    read_regn_el0(pmevcntr, 0));
+
+	/* disable the CHAIN event */
+	write_sysreg_s(0x2, PMCNTENCLR_EL0);
+	mem_access_loop(addr, 20, pmu.pmcr_ro | PMU_PMCR_E);
+	report_info("MEM_ACCESS counter #0 has value 0x%lx",
+		    read_regn_el0(pmevcntr, 0));
+	report(read_sysreg(pmovsclr_el0) == 0x1,
+		"should have triggered an overflow on #0");
+	report(!read_regn_el0(pmevcntr, 1),
+		"CHAIN counter #1 shouldn't have incremented");
+
+	/* start at 0xFFFFFFDC, +20 with CHAIN disabled, +20 with CHAIN enabled */
+
+	pmu_reset();
+	write_sysreg_s(0x1, PMCNTENSET_EL0);
+	write_regn_el0(pmevcntr, 0, PRE_OVERFLOW2);
+	isb();
+	report_info("counter #0 = 0x%lx, counter #1 = 0x%lx overflow=0x%lx",
+		    read_regn_el0(pmevcntr, 0), read_regn_el0(pmevcntr, 1),
+		    read_sysreg(pmovsclr_el0));
+
+	mem_access_loop(addr, 20, pmu.pmcr_ro | PMU_PMCR_E);
+	report_info("MEM_ACCESS counter #0 has value 0x%lx",
+		    read_regn_el0(pmevcntr, 0));
+
+	/* enable the CHAIN event */
+	write_sysreg_s(0x3, PMCNTENSET_EL0);
+	isb();
+	mem_access_loop(addr, 20, pmu.pmcr_ro | PMU_PMCR_E);
+	report_info("MEM_ACCESS counter #0 has value 0x%lx",
+		    read_regn_el0(pmevcntr, 0));
+
+	report((read_regn_el0(pmevcntr, 1) == 1) && !read_sysreg(pmovsclr_el0),
+		"CHAIN counter #1 should have incremented and no overflow expected");
+
+	report_info("CHAIN counter #1 = 0x%lx, overflow=0x%lx",
+		read_regn_el0(pmevcntr, 1), read_sysreg(pmovsclr_el0));
+
+	/* start as MEM_ACCESS/CPU_CYCLES and move to CHAIN/MEM_ACCESS */
+	pmu_reset();
+	write_regn_el0(pmevtyper, 0, MEM_ACCESS | PMEVTYPER_EXCLUDE_EL0);
+	write_regn_el0(pmevtyper, 1, CPU_CYCLES | PMEVTYPER_EXCLUDE_EL0);
+	write_sysreg_s(0x3, PMCNTENSET_EL0);
+	write_regn_el0(pmevcntr, 0, PRE_OVERFLOW2);
+	isb();
+
+	mem_access_loop(addr, 20, pmu.pmcr_ro | PMU_PMCR_E);
+	report_info("MEM_ACCESS counter #0 has value 0x%lx",
+		    read_regn_el0(pmevcntr, 0));
+
+	/* 0 becomes CHAINED */
+	write_sysreg_s(0x0, PMCNTENSET_EL0);
+	write_regn_el0(pmevtyper, 1, CHAIN | PMEVTYPER_EXCLUDE_EL0);
+	write_sysreg_s(0x3, PMCNTENSET_EL0);
+	write_regn_el0(pmevcntr, 1, 0x0);
+
+	mem_access_loop(addr, 20, pmu.pmcr_ro | PMU_PMCR_E);
+	report_info("MEM_ACCESS counter #0 has value 0x%lx",
+		    read_regn_el0(pmevcntr, 0));
+
+	report((read_regn_el0(pmevcntr, 1) == 1) && !read_sysreg(pmovsclr_el0),
+		"CHAIN counter #1 should have incremented and no overflow expected");
+
+	report_info("CHAIN counter #1 = 0x%lx, overflow=0x%lx",
+		read_regn_el0(pmevcntr, 1), read_sysreg(pmovsclr_el0));
+
+	/* start as CHAIN/MEM_ACCESS and move to MEM_ACCESS/CPU_CYCLES */
+	pmu_reset();
+	write_regn_el0(pmevtyper, 0, MEM_ACCESS | PMEVTYPER_EXCLUDE_EL0);
+	write_regn_el0(pmevtyper, 1, CHAIN | PMEVTYPER_EXCLUDE_EL0);
+	write_regn_el0(pmevcntr, 0, PRE_OVERFLOW2);
+	write_sysreg_s(0x3, PMCNTENSET_EL0);
+
+	mem_access_loop(addr, 20, pmu.pmcr_ro | PMU_PMCR_E);
+	report_info("counter #0=0x%lx, counter #1=0x%lx",
+			read_regn_el0(pmevcntr, 0), read_regn_el0(pmevcntr, 1));
+
+	write_sysreg_s(0x0, PMCNTENSET_EL0);
+	write_regn_el0(pmevtyper, 1, CPU_CYCLES | PMEVTYPER_EXCLUDE_EL0);
+	write_sysreg_s(0x3, PMCNTENSET_EL0);
+
+	mem_access_loop(addr, 20, pmu.pmcr_ro | PMU_PMCR_E);
+	report(read_sysreg(pmovsclr_el0) == 1,
+		"overflow is expected on counter 0");
+	report_info("counter #0=0x%lx, counter #1=0x%lx overflow=0x%lx",
+			read_regn_el0(pmevcntr, 0), read_regn_el0(pmevcntr, 1),
+			read_sysreg(pmovsclr_el0));
+}
+
 #endif
 
 /*
@@ -794,6 +928,10 @@ int main(int argc, char *argv[])
 		report_prefix_push(argv[1]);
 		test_chained_sw_incr();
 		report_prefix_pop();
+	} else if (strcmp(argv[1], "pmu-chain-promotion") == 0) {
+		report_prefix_push(argv[1]);
+		test_chain_promotion();
+		report_prefix_pop();
 	} else {
 		report_abort("Unknown sub-test '%s'", argv[1]);
 	}
diff --git a/arm/unittests.cfg b/arm/unittests.cfg
index d31dcbf..1b0c8c8 100644
--- a/arm/unittests.cfg
+++ b/arm/unittests.cfg
@@ -108,6 +108,12 @@ groups = pmu
 arch = arm64
 extra_params = -append 'pmu-chained-sw-incr'
 
+[pmu-chain-promotion]
+file = pmu.flat
+groups = pmu
+arch = arm64
+extra_params = -append 'pmu-chain-promotion'
+
 # Test PMU support (TCG) with -icount IPC=1
 #[pmu-tcg-icount-1]
 #file = pmu.flat
-- 
2.20.1

[kvm-unit-tests PATCH v3 10/12] arm: pmu: test 32-bit <-> 64-bit transitions

[Eric Auger]

Test configurations where we transit from 32b to 64b
counters and conversely. Also tests configuration where
chain counters are configured but only one counter is
enabled.

Signed-off-by: Eric Auger <eric.auger@redhat.com>

---

v2 -> v3:
- added prefix pop
---
 arm/pmu.c         | 138 ++++++++++++++++++++++++++++++++++++++++++++++
 arm/unittests.cfg |   6 ++
 2 files changed, 144 insertions(+)

diff --git a/arm/pmu.c b/arm/pmu.c
index ccca31b..79c2a0f 100644
--- a/arm/pmu.c
+++ b/arm/pmu.c
@@ -57,6 +57,7 @@
 #define ALL_SET			0xFFFFFFFF
 #define ALL_CLEAR		0x0
 #define PRE_OVERFLOW		0xFFFFFFF0
+#define PRE_OVERFLOW2		0xFFFFFFDC
 
 struct pmu {
 	unsigned int version;
@@ -144,6 +145,7 @@ static void test_mem_access(void) {}
 static void test_sw_incr(void) {}
 static void test_chained_counters(void) {}
 static void test_chained_sw_incr(void) {}
+static void test_chain_promotion(void) {}
 
 #elif defined(__aarch64__)
 #define ID_AA64DFR0_PERFMON_SHIFT 8
@@ -596,6 +598,138 @@ static void test_chained_sw_incr(void)
 		    read_regn_el0(pmevcntr, 0), read_regn_el0(pmevcntr, 1));
 }
 
+static void test_chain_promotion(void)
+{
+	uint32_t events[] = {MEM_ACCESS, CHAIN};
+	void *addr = malloc(PAGE_SIZE);
+
+	if (!satisfy_prerequisites(events, ARRAY_SIZE(events)))
+		return;
+
+	/* Only enable CHAIN counter */
+	pmu_reset();
+	write_regn_el0(pmevtyper, 0, MEM_ACCESS | PMEVTYPER_EXCLUDE_EL0);
+	write_regn_el0(pmevtyper, 1, CHAIN | PMEVTYPER_EXCLUDE_EL0);
+	write_sysreg_s(0x2, PMCNTENSET_EL0);
+	isb();
+
+	mem_access_loop(addr, 20, pmu.pmcr_ro | PMU_PMCR_E);
+	report(!read_regn_el0(pmevcntr, 0),
+		"chain counter not counting if even counter is disabled");
+
+	/* Only enable even counter */
+	pmu_reset();
+	write_regn_el0(pmevcntr, 0, PRE_OVERFLOW);
+	write_sysreg_s(0x1, PMCNTENSET_EL0);
+	isb();
+
+	mem_access_loop(addr, 20, pmu.pmcr_ro | PMU_PMCR_E);
+	report(!read_regn_el0(pmevcntr, 1) && (read_sysreg(pmovsclr_el0) == 0x1),
+		"odd counter did not increment on overflow if disabled");
+	report_info("MEM_ACCESS counter #0 has value %ld",
+		    read_regn_el0(pmevcntr, 0));
+	report_info("CHAIN counter #1 has value %ld",
+		    read_regn_el0(pmevcntr, 1));
+	report_info("overflow counter %ld", read_sysreg(pmovsclr_el0));
+
+	/* start at 0xFFFFFFDC, +20 with CHAIN enabled, +20 with CHAIN disabled */
+	pmu_reset();
+	write_sysreg_s(0x3, PMCNTENSET_EL0);
+	write_regn_el0(pmevcntr, 0, PRE_OVERFLOW2);
+	isb();
+
+	mem_access_loop(addr, 20, pmu.pmcr_ro | PMU_PMCR_E);
+	report_info("MEM_ACCESS counter #0 has value 0x%lx",
+		    read_regn_el0(pmevcntr, 0));
+
+	/* disable the CHAIN event */
+	write_sysreg_s(0x2, PMCNTENCLR_EL0);
+	mem_access_loop(addr, 20, pmu.pmcr_ro | PMU_PMCR_E);
+	report_info("MEM_ACCESS counter #0 has value 0x%lx",
+		    read_regn_el0(pmevcntr, 0));
+	report(read_sysreg(pmovsclr_el0) == 0x1,
+		"should have triggered an overflow on #0");
+	report(!read_regn_el0(pmevcntr, 1),
+		"CHAIN counter #1 shouldn't have incremented");
+
+	/* start at 0xFFFFFFDC, +20 with CHAIN disabled, +20 with CHAIN enabled */
+
+	pmu_reset();
+	write_sysreg_s(0x1, PMCNTENSET_EL0);
+	write_regn_el0(pmevcntr, 0, PRE_OVERFLOW2);
+	isb();
+	report_info("counter #0 = 0x%lx, counter #1 = 0x%lx overflow=0x%lx",
+		    read_regn_el0(pmevcntr, 0), read_regn_el0(pmevcntr, 1),
+		    read_sysreg(pmovsclr_el0));
+
+	mem_access_loop(addr, 20, pmu.pmcr_ro | PMU_PMCR_E);
+	report_info("MEM_ACCESS counter #0 has value 0x%lx",
+		    read_regn_el0(pmevcntr, 0));
+
+	/* enable the CHAIN event */
+	write_sysreg_s(0x3, PMCNTENSET_EL0);
+	isb();
+	mem_access_loop(addr, 20, pmu.pmcr_ro | PMU_PMCR_E);
+	report_info("MEM_ACCESS counter #0 has value 0x%lx",
+		    read_regn_el0(pmevcntr, 0));
+
+	report((read_regn_el0(pmevcntr, 1) == 1) && !read_sysreg(pmovsclr_el0),
+		"CHAIN counter #1 should have incremented and no overflow expected");
+
+	report_info("CHAIN counter #1 = 0x%lx, overflow=0x%lx",
+		read_regn_el0(pmevcntr, 1), read_sysreg(pmovsclr_el0));
+
+	/* start as MEM_ACCESS/CPU_CYCLES and move to CHAIN/MEM_ACCESS */
+	pmu_reset();
+	write_regn_el0(pmevtyper, 0, MEM_ACCESS | PMEVTYPER_EXCLUDE_EL0);
+	write_regn_el0(pmevtyper, 1, CPU_CYCLES | PMEVTYPER_EXCLUDE_EL0);
+	write_sysreg_s(0x3, PMCNTENSET_EL0);
+	write_regn_el0(pmevcntr, 0, PRE_OVERFLOW2);
+	isb();
+
+	mem_access_loop(addr, 20, pmu.pmcr_ro | PMU_PMCR_E);
+	report_info("MEM_ACCESS counter #0 has value 0x%lx",
+		    read_regn_el0(pmevcntr, 0));
+
+	/* 0 becomes CHAINED */
+	write_sysreg_s(0x0, PMCNTENSET_EL0);
+	write_regn_el0(pmevtyper, 1, CHAIN | PMEVTYPER_EXCLUDE_EL0);
+	write_sysreg_s(0x3, PMCNTENSET_EL0);
+	write_regn_el0(pmevcntr, 1, 0x0);
+
+	mem_access_loop(addr, 20, pmu.pmcr_ro | PMU_PMCR_E);
+	report_info("MEM_ACCESS counter #0 has value 0x%lx",
+		    read_regn_el0(pmevcntr, 0));
+
+	report((read_regn_el0(pmevcntr, 1) == 1) && !read_sysreg(pmovsclr_el0),
+		"CHAIN counter #1 should have incremented and no overflow expected");
+
+	report_info("CHAIN counter #1 = 0x%lx, overflow=0x%lx",
+		read_regn_el0(pmevcntr, 1), read_sysreg(pmovsclr_el0));
+
+	/* start as CHAIN/MEM_ACCESS and move to MEM_ACCESS/CPU_CYCLES */
+	pmu_reset();
+	write_regn_el0(pmevtyper, 0, MEM_ACCESS | PMEVTYPER_EXCLUDE_EL0);
+	write_regn_el0(pmevtyper, 1, CHAIN | PMEVTYPER_EXCLUDE_EL0);
+	write_regn_el0(pmevcntr, 0, PRE_OVERFLOW2);
+	write_sysreg_s(0x3, PMCNTENSET_EL0);
+
+	mem_access_loop(addr, 20, pmu.pmcr_ro | PMU_PMCR_E);
+	report_info("counter #0=0x%lx, counter #1=0x%lx",
+			read_regn_el0(pmevcntr, 0), read_regn_el0(pmevcntr, 1));
+
+	write_sysreg_s(0x0, PMCNTENSET_EL0);
+	write_regn_el0(pmevtyper, 1, CPU_CYCLES | PMEVTYPER_EXCLUDE_EL0);
+	write_sysreg_s(0x3, PMCNTENSET_EL0);
+
+	mem_access_loop(addr, 20, pmu.pmcr_ro | PMU_PMCR_E);
+	report(read_sysreg(pmovsclr_el0) == 1,
+		"overflow is expected on counter 0");
+	report_info("counter #0=0x%lx, counter #1=0x%lx overflow=0x%lx",
+			read_regn_el0(pmevcntr, 0), read_regn_el0(pmevcntr, 1),
+			read_sysreg(pmovsclr_el0));
+}
+
 #endif
 
 /*
@@ -794,6 +928,10 @@ int main(int argc, char *argv[])
 		report_prefix_push(argv[1]);
 		test_chained_sw_incr();
 		report_prefix_pop();
+	} else if (strcmp(argv[1], "pmu-chain-promotion") == 0) {
+		report_prefix_push(argv[1]);
+		test_chain_promotion();
+		report_prefix_pop();
 	} else {
 		report_abort("Unknown sub-test '%s'", argv[1]);
 	}
diff --git a/arm/unittests.cfg b/arm/unittests.cfg
index d31dcbf..1b0c8c8 100644
--- a/arm/unittests.cfg
+++ b/arm/unittests.cfg
@@ -108,6 +108,12 @@ groups = pmu
 arch = arm64
 extra_params = -append 'pmu-chained-sw-incr'
 
+[pmu-chain-promotion]
+file = pmu.flat
+groups = pmu
+arch = arm64
+extra_params = -append 'pmu-chain-promotion'
+
 # Test PMU support (TCG) with -icount IPC=1
 #[pmu-tcg-icount-1]
 #file = pmu.flat
-- 
2.20.1

[kvm-unit-tests PATCH v3 10/12] arm: pmu: test 32-bit <-> 64-bit transitions

[Eric Auger]

Test configurations where we transit from 32b to 64b
counters and conversely. Also tests configuration where
chain counters are configured but only one counter is
enabled.

Signed-off-by: Eric Auger <eric.auger@redhat.com>

---

v2 -> v3:
- added prefix pop
---
 arm/pmu.c         | 138 ++++++++++++++++++++++++++++++++++++++++++++++
 arm/unittests.cfg |   6 ++
 2 files changed, 144 insertions(+)

diff --git a/arm/pmu.c b/arm/pmu.c
index ccca31b..79c2a0f 100644
--- a/arm/pmu.c
+++ b/arm/pmu.c
@@ -57,6 +57,7 @@
 #define ALL_SET			0xFFFFFFFF
 #define ALL_CLEAR		0x0
 #define PRE_OVERFLOW		0xFFFFFFF0
+#define PRE_OVERFLOW2		0xFFFFFFDC
 
 struct pmu {
 	unsigned int version;
@@ -144,6 +145,7 @@ static void test_mem_access(void) {}
 static void test_sw_incr(void) {}
 static void test_chained_counters(void) {}
 static void test_chained_sw_incr(void) {}
+static void test_chain_promotion(void) {}
 
 #elif defined(__aarch64__)
 #define ID_AA64DFR0_PERFMON_SHIFT 8
@@ -596,6 +598,138 @@ static void test_chained_sw_incr(void)
 		    read_regn_el0(pmevcntr, 0), read_regn_el0(pmevcntr, 1));
 }
 
+static void test_chain_promotion(void)
+{
+	uint32_t events[] = {MEM_ACCESS, CHAIN};
+	void *addr = malloc(PAGE_SIZE);
+
+	if (!satisfy_prerequisites(events, ARRAY_SIZE(events)))
+		return;
+
+	/* Only enable CHAIN counter */
+	pmu_reset();
+	write_regn_el0(pmevtyper, 0, MEM_ACCESS | PMEVTYPER_EXCLUDE_EL0);
+	write_regn_el0(pmevtyper, 1, CHAIN | PMEVTYPER_EXCLUDE_EL0);
+	write_sysreg_s(0x2, PMCNTENSET_EL0);
+	isb();
+
+	mem_access_loop(addr, 20, pmu.pmcr_ro | PMU_PMCR_E);
+	report(!read_regn_el0(pmevcntr, 0),
+		"chain counter not counting if even counter is disabled");
+
+	/* Only enable even counter */
+	pmu_reset();
+	write_regn_el0(pmevcntr, 0, PRE_OVERFLOW);
+	write_sysreg_s(0x1, PMCNTENSET_EL0);
+	isb();
+
+	mem_access_loop(addr, 20, pmu.pmcr_ro | PMU_PMCR_E);
+	report(!read_regn_el0(pmevcntr, 1) && (read_sysreg(pmovsclr_el0) == 0x1),
+		"odd counter did not increment on overflow if disabled");
+	report_info("MEM_ACCESS counter #0 has value %ld",
+		    read_regn_el0(pmevcntr, 0));
+	report_info("CHAIN counter #1 has value %ld",
+		    read_regn_el0(pmevcntr, 1));
+	report_info("overflow counter %ld", read_sysreg(pmovsclr_el0));
+
+	/* start at 0xFFFFFFDC, +20 with CHAIN enabled, +20 with CHAIN disabled */
+	pmu_reset();
+	write_sysreg_s(0x3, PMCNTENSET_EL0);
+	write_regn_el0(pmevcntr, 0, PRE_OVERFLOW2);
+	isb();
+
+	mem_access_loop(addr, 20, pmu.pmcr_ro | PMU_PMCR_E);
+	report_info("MEM_ACCESS counter #0 has value 0x%lx",
+		    read_regn_el0(pmevcntr, 0));
+
+	/* disable the CHAIN event */
+	write_sysreg_s(0x2, PMCNTENCLR_EL0);
+	mem_access_loop(addr, 20, pmu.pmcr_ro | PMU_PMCR_E);
+	report_info("MEM_ACCESS counter #0 has value 0x%lx",
+		    read_regn_el0(pmevcntr, 0));
+	report(read_sysreg(pmovsclr_el0) == 0x1,
+		"should have triggered an overflow on #0");
+	report(!read_regn_el0(pmevcntr, 1),
+		"CHAIN counter #1 shouldn't have incremented");
+
+	/* start at 0xFFFFFFDC, +20 with CHAIN disabled, +20 with CHAIN enabled */
+
+	pmu_reset();
+	write_sysreg_s(0x1, PMCNTENSET_EL0);
+	write_regn_el0(pmevcntr, 0, PRE_OVERFLOW2);
+	isb();
+	report_info("counter #0 = 0x%lx, counter #1 = 0x%lx overflow=0x%lx",
+		    read_regn_el0(pmevcntr, 0), read_regn_el0(pmevcntr, 1),
+		    read_sysreg(pmovsclr_el0));
+
+	mem_access_loop(addr, 20, pmu.pmcr_ro | PMU_PMCR_E);
+	report_info("MEM_ACCESS counter #0 has value 0x%lx",
+		    read_regn_el0(pmevcntr, 0));
+
+	/* enable the CHAIN event */
+	write_sysreg_s(0x3, PMCNTENSET_EL0);
+	isb();
+	mem_access_loop(addr, 20, pmu.pmcr_ro | PMU_PMCR_E);
+	report_info("MEM_ACCESS counter #0 has value 0x%lx",
+		    read_regn_el0(pmevcntr, 0));
+
+	report((read_regn_el0(pmevcntr, 1) == 1) && !read_sysreg(pmovsclr_el0),
+		"CHAIN counter #1 should have incremented and no overflow expected");
+
+	report_info("CHAIN counter #1 = 0x%lx, overflow=0x%lx",
+		read_regn_el0(pmevcntr, 1), read_sysreg(pmovsclr_el0));
+
+	/* start as MEM_ACCESS/CPU_CYCLES and move to CHAIN/MEM_ACCESS */
+	pmu_reset();
+	write_regn_el0(pmevtyper, 0, MEM_ACCESS | PMEVTYPER_EXCLUDE_EL0);
+	write_regn_el0(pmevtyper, 1, CPU_CYCLES | PMEVTYPER_EXCLUDE_EL0);
+	write_sysreg_s(0x3, PMCNTENSET_EL0);
+	write_regn_el0(pmevcntr, 0, PRE_OVERFLOW2);
+	isb();
+
+	mem_access_loop(addr, 20, pmu.pmcr_ro | PMU_PMCR_E);
+	report_info("MEM_ACCESS counter #0 has value 0x%lx",
+		    read_regn_el0(pmevcntr, 0));
+
+	/* 0 becomes CHAINED */
+	write_sysreg_s(0x0, PMCNTENSET_EL0);
+	write_regn_el0(pmevtyper, 1, CHAIN | PMEVTYPER_EXCLUDE_EL0);
+	write_sysreg_s(0x3, PMCNTENSET_EL0);
+	write_regn_el0(pmevcntr, 1, 0x0);
+
+	mem_access_loop(addr, 20, pmu.pmcr_ro | PMU_PMCR_E);
+	report_info("MEM_ACCESS counter #0 has value 0x%lx",
+		    read_regn_el0(pmevcntr, 0));
+
+	report((read_regn_el0(pmevcntr, 1) == 1) && !read_sysreg(pmovsclr_el0),
+		"CHAIN counter #1 should have incremented and no overflow expected");
+
+	report_info("CHAIN counter #1 = 0x%lx, overflow=0x%lx",
+		read_regn_el0(pmevcntr, 1), read_sysreg(pmovsclr_el0));
+
+	/* start as CHAIN/MEM_ACCESS and move to MEM_ACCESS/CPU_CYCLES */
+	pmu_reset();
+	write_regn_el0(pmevtyper, 0, MEM_ACCESS | PMEVTYPER_EXCLUDE_EL0);
+	write_regn_el0(pmevtyper, 1, CHAIN | PMEVTYPER_EXCLUDE_EL0);
+	write_regn_el0(pmevcntr, 0, PRE_OVERFLOW2);
+	write_sysreg_s(0x3, PMCNTENSET_EL0);
+
+	mem_access_loop(addr, 20, pmu.pmcr_ro | PMU_PMCR_E);
+	report_info("counter #0=0x%lx, counter #1=0x%lx",
+			read_regn_el0(pmevcntr, 0), read_regn_el0(pmevcntr, 1));
+
+	write_sysreg_s(0x0, PMCNTENSET_EL0);
+	write_regn_el0(pmevtyper, 1, CPU_CYCLES | PMEVTYPER_EXCLUDE_EL0);
+	write_sysreg_s(0x3, PMCNTENSET_EL0);
+
+	mem_access_loop(addr, 20, pmu.pmcr_ro | PMU_PMCR_E);
+	report(read_sysreg(pmovsclr_el0) == 1,
+		"overflow is expected on counter 0");
+	report_info("counter #0=0x%lx, counter #1=0x%lx overflow=0x%lx",
+			read_regn_el0(pmevcntr, 0), read_regn_el0(pmevcntr, 1),
+			read_sysreg(pmovsclr_el0));
+}
+
 #endif
 
 /*
@@ -794,6 +928,10 @@ int main(int argc, char *argv[])
 		report_prefix_push(argv[1]);
 		test_chained_sw_incr();
 		report_prefix_pop();
+	} else if (strcmp(argv[1], "pmu-chain-promotion") == 0) {
+		report_prefix_push(argv[1]);
+		test_chain_promotion();
+		report_prefix_pop();
 	} else {
 		report_abort("Unknown sub-test '%s'", argv[1]);
 	}
diff --git a/arm/unittests.cfg b/arm/unittests.cfg
index d31dcbf..1b0c8c8 100644
--- a/arm/unittests.cfg
+++ b/arm/unittests.cfg
@@ -108,6 +108,12 @@ groups = pmu
 arch = arm64
 extra_params = -append 'pmu-chained-sw-incr'
 
+[pmu-chain-promotion]
+file = pmu.flat
+groups = pmu
+arch = arm64
+extra_params = -append 'pmu-chain-promotion'
+
 # Test PMU support (TCG) with -icount IPC=1
 #[pmu-tcg-icount-1]
 #file = pmu.flat
-- 
2.20.1

_______________________________________________
kvmarm mailing list
kvmarm@lists.cs.columbia.edu
https://lists.cs.columbia.edu/mailman/listinfo/kvmarm

[kvm-unit-tests PATCH v3 11/12] arm: gic: Introduce gic_irq_set_clr_enable() helper

[Eric Auger]

Allows to set or clear the enable state of a PPI/SGI/SPI.

Signed-off-by: Eric Auger <eric.auger@redhat.com>

---
---
 lib/arm/asm/gic.h |  4 ++++
 lib/arm/gic.c     | 31 +++++++++++++++++++++++++++++++
 2 files changed, 35 insertions(+)

diff --git a/lib/arm/asm/gic.h b/lib/arm/asm/gic.h
index 922cbe9..57e81c6 100644
--- a/lib/arm/asm/gic.h
+++ b/lib/arm/asm/gic.h
@@ -82,5 +82,9 @@ extern void gic_ipi_send_single(int irq, int cpu);
 extern void gic_ipi_send_mask(int irq, const cpumask_t *dest);
 extern enum gic_irq_state gic_irq_state(int irq);
 
+void gic_irq_set_clr_enable(int irq, bool enable);
+#define gic_enable_irq(irq) gic_irq_set_clr_enable(irq, true);
+#define gic_disable_irq(irq) gic_irq_set_clr_enable(irq, false);
+
 #endif /* !__ASSEMBLY__ */
 #endif /* _ASMARM_GIC_H_ */
diff --git a/lib/arm/gic.c b/lib/arm/gic.c
index c3c5f6b..8a1a8c8 100644
--- a/lib/arm/gic.c
+++ b/lib/arm/gic.c
@@ -147,6 +147,36 @@ void gic_ipi_send_mask(int irq, const cpumask_t *dest)
 	gic_common_ops->ipi_send_mask(irq, dest);
 }
 
+void gic_irq_set_clr_enable(int irq, bool enable)
+{
+	u32 offset, split = 32, shift = (irq % 32);
+	u32 reg, mask = BIT(shift);
+	void *base;
+
+	assert(irq < 1020);
+
+	switch (gic_version()) {
+	case 2:
+		offset = enable ? GICD_ISENABLER : GICD_ICENABLER;
+		base = gicv2_dist_base();
+		break;
+	case 3:
+		if (irq < 32) {
+			offset = enable ? GICR_ISENABLER0 : GICR_ICENABLER0;
+			base = gicv3_sgi_base();
+		} else {
+			offset = enable ? GICD_ISENABLER : GICD_ICENABLER;
+			base = gicv3_dist_base();
+		}
+		break;
+	default:
+		assert(0);
+	}
+	base += offset + (irq / split) * 4;
+	reg = readl(base);
+	writel(reg | mask, base);
+}
+
 enum gic_irq_state gic_irq_state(int irq)
 {
 	enum gic_irq_state state;
@@ -191,3 +221,4 @@ enum gic_irq_state gic_irq_state(int irq)
 
 	return state;
 }
+
-- 
2.20.1

[kvm-unit-tests PATCH v3 11/12] arm: gic: Introduce gic_irq_set_clr_enable() helper

[Eric Auger]

Allows to set or clear the enable state of a PPI/SGI/SPI.

Signed-off-by: Eric Auger <eric.auger@redhat.com>

---
---
 lib/arm/asm/gic.h |  4 ++++
 lib/arm/gic.c     | 31 +++++++++++++++++++++++++++++++
 2 files changed, 35 insertions(+)

diff --git a/lib/arm/asm/gic.h b/lib/arm/asm/gic.h
index 922cbe9..57e81c6 100644
--- a/lib/arm/asm/gic.h
+++ b/lib/arm/asm/gic.h
@@ -82,5 +82,9 @@ extern void gic_ipi_send_single(int irq, int cpu);
 extern void gic_ipi_send_mask(int irq, const cpumask_t *dest);
 extern enum gic_irq_state gic_irq_state(int irq);
 
+void gic_irq_set_clr_enable(int irq, bool enable);
+#define gic_enable_irq(irq) gic_irq_set_clr_enable(irq, true);
+#define gic_disable_irq(irq) gic_irq_set_clr_enable(irq, false);
+
 #endif /* !__ASSEMBLY__ */
 #endif /* _ASMARM_GIC_H_ */
diff --git a/lib/arm/gic.c b/lib/arm/gic.c
index c3c5f6b..8a1a8c8 100644
--- a/lib/arm/gic.c
+++ b/lib/arm/gic.c
@@ -147,6 +147,36 @@ void gic_ipi_send_mask(int irq, const cpumask_t *dest)
 	gic_common_ops->ipi_send_mask(irq, dest);
 }
 
+void gic_irq_set_clr_enable(int irq, bool enable)
+{
+	u32 offset, split = 32, shift = (irq % 32);
+	u32 reg, mask = BIT(shift);
+	void *base;
+
+	assert(irq < 1020);
+
+	switch (gic_version()) {
+	case 2:
+		offset = enable ? GICD_ISENABLER : GICD_ICENABLER;
+		base = gicv2_dist_base();
+		break;
+	case 3:
+		if (irq < 32) {
+			offset = enable ? GICR_ISENABLER0 : GICR_ICENABLER0;
+			base = gicv3_sgi_base();
+		} else {
+			offset = enable ? GICD_ISENABLER : GICD_ICENABLER;
+			base = gicv3_dist_base();
+		}
+		break;
+	default:
+		assert(0);
+	}
+	base += offset + (irq / split) * 4;
+	reg = readl(base);
+	writel(reg | mask, base);
+}
+
 enum gic_irq_state gic_irq_state(int irq)
 {
 	enum gic_irq_state state;
@@ -191,3 +221,4 @@ enum gic_irq_state gic_irq_state(int irq)
 
 	return state;
 }
+
-- 
2.20.1

[kvm-unit-tests PATCH v3 11/12] arm: gic: Introduce gic_irq_set_clr_enable() helper

[Eric Auger]

Allows to set or clear the enable state of a PPI/SGI/SPI.

Signed-off-by: Eric Auger <eric.auger@redhat.com>

---
---
 lib/arm/asm/gic.h |  4 ++++
 lib/arm/gic.c     | 31 +++++++++++++++++++++++++++++++
 2 files changed, 35 insertions(+)

diff --git a/lib/arm/asm/gic.h b/lib/arm/asm/gic.h
index 922cbe9..57e81c6 100644
--- a/lib/arm/asm/gic.h
+++ b/lib/arm/asm/gic.h
@@ -82,5 +82,9 @@ extern void gic_ipi_send_single(int irq, int cpu);
 extern void gic_ipi_send_mask(int irq, const cpumask_t *dest);
 extern enum gic_irq_state gic_irq_state(int irq);
 
+void gic_irq_set_clr_enable(int irq, bool enable);
+#define gic_enable_irq(irq) gic_irq_set_clr_enable(irq, true);
+#define gic_disable_irq(irq) gic_irq_set_clr_enable(irq, false);
+
 #endif /* !__ASSEMBLY__ */
 #endif /* _ASMARM_GIC_H_ */
diff --git a/lib/arm/gic.c b/lib/arm/gic.c
index c3c5f6b..8a1a8c8 100644
--- a/lib/arm/gic.c
+++ b/lib/arm/gic.c
@@ -147,6 +147,36 @@ void gic_ipi_send_mask(int irq, const cpumask_t *dest)
 	gic_common_ops->ipi_send_mask(irq, dest);
 }
 
+void gic_irq_set_clr_enable(int irq, bool enable)
+{
+	u32 offset, split = 32, shift = (irq % 32);
+	u32 reg, mask = BIT(shift);
+	void *base;
+
+	assert(irq < 1020);
+
+	switch (gic_version()) {
+	case 2:
+		offset = enable ? GICD_ISENABLER : GICD_ICENABLER;
+		base = gicv2_dist_base();
+		break;
+	case 3:
+		if (irq < 32) {
+			offset = enable ? GICR_ISENABLER0 : GICR_ICENABLER0;
+			base = gicv3_sgi_base();
+		} else {
+			offset = enable ? GICD_ISENABLER : GICD_ICENABLER;
+			base = gicv3_dist_base();
+		}
+		break;
+	default:
+		assert(0);
+	}
+	base += offset + (irq / split) * 4;
+	reg = readl(base);
+	writel(reg | mask, base);
+}
+
 enum gic_irq_state gic_irq_state(int irq)
 {
 	enum gic_irq_state state;
@@ -191,3 +221,4 @@ enum gic_irq_state gic_irq_state(int irq)
 
 	return state;
 }
+
-- 
2.20.1

_______________________________________________
kvmarm mailing list
kvmarm@lists.cs.columbia.edu
https://lists.cs.columbia.edu/mailman/listinfo/kvmarm

[kvm-unit-tests PATCH v3 12/12] arm: pmu: Test overflow interrupts

[Eric Auger]

Test overflows for MEM_ACCESS and SW_INCR events. Also tests
overflows with 64-bit events.

Signed-off-by: Eric Auger <eric.auger@redhat.com>

---

v2 -> v3:
- added prefix pop
- added pmu_stats.unexpected
- added pmu- prefix
- remove traces in irq_handler()

v1 -> v2:
- inline setup_irq() code
---
 arm/pmu.c         | 139 ++++++++++++++++++++++++++++++++++++++++++++++
 arm/unittests.cfg |   6 ++
 2 files changed, 145 insertions(+)

diff --git a/arm/pmu.c b/arm/pmu.c
index 79c2a0f..628249c 100644
--- a/arm/pmu.c
+++ b/arm/pmu.c
@@ -59,12 +59,20 @@
 #define PRE_OVERFLOW		0xFFFFFFF0
 #define PRE_OVERFLOW2		0xFFFFFFDC
 
+#define PMU_PPI			23
+
 struct pmu {
 	unsigned int version;
 	unsigned int nb_implemented_counters;
 	uint32_t pmcr_ro;
 };
 
+struct pmu_stats {
+	unsigned long bitmap;
+	uint32_t interrupts[32];
+	bool unexpected;
+};
+
 static struct pmu pmu;
 
 #if defined(__arm__)
@@ -146,6 +154,7 @@ static void test_sw_incr(void) {}
 static void test_chained_counters(void) {}
 static void test_chained_sw_incr(void) {}
 static void test_chain_promotion(void) {}
+static void test_overflow_interrupt(void) {}
 
 #elif defined(__aarch64__)
 #define ID_AA64DFR0_PERFMON_SHIFT 8
@@ -276,6 +285,43 @@ asm volatile(
 	: "x9", "x10", "cc");
 }
 
+static struct pmu_stats pmu_stats;
+
+static void irq_handler(struct pt_regs *regs)
+{
+	uint32_t irqstat, irqnr;
+
+	irqstat = gic_read_iar();
+	irqnr = gic_iar_irqnr(irqstat);
+
+	if (irqnr == PMU_PPI) {
+		unsigned long overflows = read_sysreg(pmovsclr_el0);
+		int i;
+
+		for (i = 0; i < 32; i++) {
+			if (test_and_clear_bit(i, &overflows)) {
+				pmu_stats.interrupts[i]++;
+				pmu_stats.bitmap |= 1 << i;
+			}
+		}
+		write_sysreg(ALL_SET, pmovsclr_el0);
+	} else {
+		pmu_stats.unexpected = true;
+	}
+	gic_write_eoir(irqstat);
+}
+
+static void pmu_reset_stats(void)
+{
+	int i;
+
+	for (i = 0; i < 32; i++)
+		pmu_stats.interrupts[i] = 0;
+
+	pmu_stats.bitmap = 0;
+	pmu_stats.unexpected = false;
+}
+
 static void pmu_reset(void)
 {
 	/* reset all counters, counting disabled at PMCR level*/
@@ -286,6 +332,7 @@ static void pmu_reset(void)
 	write_sysreg(ALL_SET, pmovsclr_el0);
 	/* disable overflow interrupts on all counters */
 	write_sysreg(ALL_SET, pmintenclr_el1);
+	pmu_reset_stats();
 	isb();
 }
 
@@ -730,6 +777,94 @@ static void test_chain_promotion(void)
 			read_sysreg(pmovsclr_el0));
 }
 
+static bool expect_interrupts(uint32_t bitmap)
+{
+	int i;
+
+	if (pmu_stats.bitmap ^ bitmap || pmu_stats.unexpected)
+		return false;
+
+	for (i = 0; i < 32; i++) {
+		if (test_and_clear_bit(i, &pmu_stats.bitmap))
+			if (pmu_stats.interrupts[i] != 1)
+				return false;
+	}
+	return true;
+}
+
+static void test_overflow_interrupt(void)
+{
+	uint32_t events[] = {MEM_ACCESS, SW_INCR};
+	void *addr = malloc(PAGE_SIZE);
+	int i;
+
+	if (!satisfy_prerequisites(events, ARRAY_SIZE(events)))
+		return;
+
+	gic_enable_defaults();
+	install_irq_handler(EL1H_IRQ, irq_handler);
+	local_irq_enable();
+	gic_enable_irq(23);
+
+	pmu_reset();
+
+	write_regn_el0(pmevtyper, 0, MEM_ACCESS | PMEVTYPER_EXCLUDE_EL0);
+	write_regn_el0(pmevtyper, 1, SW_INCR | PMEVTYPER_EXCLUDE_EL0);
+	write_sysreg_s(0x3, PMCNTENSET_EL0);
+	write_regn_el0(pmevcntr, 0, PRE_OVERFLOW);
+	write_regn_el0(pmevcntr, 1, PRE_OVERFLOW);
+	isb();
+
+	/* interrupts are disabled */
+
+	mem_access_loop(addr, 200, pmu.pmcr_ro | PMU_PMCR_E);
+	report(expect_interrupts(0), "no overflow interrupt received");
+
+	set_pmcr(pmu.pmcr_ro | PMU_PMCR_E);
+	for (i = 0; i < 100; i++)
+		write_sysreg(0x2, pmswinc_el0);
+
+	set_pmcr(pmu.pmcr_ro);
+	report(expect_interrupts(0), "no overflow interrupt received");
+
+	/* enable interrupts */
+
+	pmu_reset_stats();
+
+	write_regn_el0(pmevcntr, 0, PRE_OVERFLOW);
+	write_regn_el0(pmevcntr, 1, PRE_OVERFLOW);
+	write_sysreg(ALL_SET, pmintenset_el1);
+	isb();
+
+	mem_access_loop(addr, 200, pmu.pmcr_ro | PMU_PMCR_E);
+	for (i = 0; i < 100; i++)
+		write_sysreg(0x3, pmswinc_el0);
+
+	mem_access_loop(addr, 200, pmu.pmcr_ro);
+	report_info("overflow=0x%lx", read_sysreg(pmovsclr_el0));
+	report(expect_interrupts(0x3),
+		"overflow interrupts expected on #0 and #1");
+
+	/* promote to 64-b */
+
+	pmu_reset_stats();
+
+	write_regn_el0(pmevtyper, 1, CHAIN | PMEVTYPER_EXCLUDE_EL0);
+	write_regn_el0(pmevcntr, 0, PRE_OVERFLOW);
+	isb();
+	mem_access_loop(addr, 200, pmu.pmcr_ro | PMU_PMCR_E);
+	report(expect_interrupts(0),
+		"no overflow interrupt expected on 32b boundary");
+
+	/* overflow on odd counter */
+	pmu_reset_stats();
+	write_regn_el0(pmevcntr, 0, PRE_OVERFLOW);
+	write_regn_el0(pmevcntr, 1, ALL_SET);
+	isb();
+	mem_access_loop(addr, 400, pmu.pmcr_ro | PMU_PMCR_E);
+	report(expect_interrupts(0x2),
+		"expect overflow interrupt on odd counter");
+}
 #endif
 
 /*
@@ -932,6 +1067,10 @@ int main(int argc, char *argv[])
 		report_prefix_push(argv[1]);
 		test_chain_promotion();
 		report_prefix_pop();
+	} else if (strcmp(argv[1], "pmu-overflow-interrupt") == 0) {
+		report_prefix_push(argv[1]);
+		test_overflow_interrupt();
+		report_prefix_pop();
 	} else {
 		report_abort("Unknown sub-test '%s'", argv[1]);
 	}
diff --git a/arm/unittests.cfg b/arm/unittests.cfg
index 1b0c8c8..455fd10 100644
--- a/arm/unittests.cfg
+++ b/arm/unittests.cfg
@@ -114,6 +114,12 @@ groups = pmu
 arch = arm64
 extra_params = -append 'pmu-chain-promotion'
 
+[pmu-overflow-interrupt]
+file = pmu.flat
+groups = pmu
+arch = arm64
+extra_params = -append 'pmu-overflow-interrupt'
+
 # Test PMU support (TCG) with -icount IPC=1
 #[pmu-tcg-icount-1]
 #file = pmu.flat
-- 
2.20.1

[kvm-unit-tests PATCH v3 12/12] arm: pmu: Test overflow interrupts

[Eric Auger]

Test overflows for MEM_ACCESS and SW_INCR events. Also tests
overflows with 64-bit events.

Signed-off-by: Eric Auger <eric.auger@redhat.com>

---

v2 -> v3:
- added prefix pop
- added pmu_stats.unexpected
- added pmu- prefix
- remove traces in irq_handler()

v1 -> v2:
- inline setup_irq() code
---
 arm/pmu.c         | 139 ++++++++++++++++++++++++++++++++++++++++++++++
 arm/unittests.cfg |   6 ++
 2 files changed, 145 insertions(+)

diff --git a/arm/pmu.c b/arm/pmu.c
index 79c2a0f..628249c 100644
--- a/arm/pmu.c
+++ b/arm/pmu.c
@@ -59,12 +59,20 @@
 #define PRE_OVERFLOW		0xFFFFFFF0
 #define PRE_OVERFLOW2		0xFFFFFFDC
 
+#define PMU_PPI			23
+
 struct pmu {
 	unsigned int version;
 	unsigned int nb_implemented_counters;
 	uint32_t pmcr_ro;
 };
 
+struct pmu_stats {
+	unsigned long bitmap;
+	uint32_t interrupts[32];
+	bool unexpected;
+};
+
 static struct pmu pmu;
 
 #if defined(__arm__)
@@ -146,6 +154,7 @@ static void test_sw_incr(void) {}
 static void test_chained_counters(void) {}
 static void test_chained_sw_incr(void) {}
 static void test_chain_promotion(void) {}
+static void test_overflow_interrupt(void) {}
 
 #elif defined(__aarch64__)
 #define ID_AA64DFR0_PERFMON_SHIFT 8
@@ -276,6 +285,43 @@ asm volatile(
 	: "x9", "x10", "cc");
 }
 
+static struct pmu_stats pmu_stats;
+
+static void irq_handler(struct pt_regs *regs)
+{
+	uint32_t irqstat, irqnr;
+
+	irqstat = gic_read_iar();
+	irqnr = gic_iar_irqnr(irqstat);
+
+	if (irqnr == PMU_PPI) {
+		unsigned long overflows = read_sysreg(pmovsclr_el0);
+		int i;
+
+		for (i = 0; i < 32; i++) {
+			if (test_and_clear_bit(i, &overflows)) {
+				pmu_stats.interrupts[i]++;
+				pmu_stats.bitmap |= 1 << i;
+			}
+		}
+		write_sysreg(ALL_SET, pmovsclr_el0);
+	} else {
+		pmu_stats.unexpected = true;
+	}
+	gic_write_eoir(irqstat);
+}
+
+static void pmu_reset_stats(void)
+{
+	int i;
+
+	for (i = 0; i < 32; i++)
+		pmu_stats.interrupts[i] = 0;
+
+	pmu_stats.bitmap = 0;
+	pmu_stats.unexpected = false;
+}
+
 static void pmu_reset(void)
 {
 	/* reset all counters, counting disabled at PMCR level*/
@@ -286,6 +332,7 @@ static void pmu_reset(void)
 	write_sysreg(ALL_SET, pmovsclr_el0);
 	/* disable overflow interrupts on all counters */
 	write_sysreg(ALL_SET, pmintenclr_el1);
+	pmu_reset_stats();
 	isb();
 }
 
@@ -730,6 +777,94 @@ static void test_chain_promotion(void)
 			read_sysreg(pmovsclr_el0));
 }
 
+static bool expect_interrupts(uint32_t bitmap)
+{
+	int i;
+
+	if (pmu_stats.bitmap ^ bitmap || pmu_stats.unexpected)
+		return false;
+
+	for (i = 0; i < 32; i++) {
+		if (test_and_clear_bit(i, &pmu_stats.bitmap))
+			if (pmu_stats.interrupts[i] != 1)
+				return false;
+	}
+	return true;
+}
+
+static void test_overflow_interrupt(void)
+{
+	uint32_t events[] = {MEM_ACCESS, SW_INCR};
+	void *addr = malloc(PAGE_SIZE);
+	int i;
+
+	if (!satisfy_prerequisites(events, ARRAY_SIZE(events)))
+		return;
+
+	gic_enable_defaults();
+	install_irq_handler(EL1H_IRQ, irq_handler);
+	local_irq_enable();
+	gic_enable_irq(23);
+
+	pmu_reset();
+
+	write_regn_el0(pmevtyper, 0, MEM_ACCESS | PMEVTYPER_EXCLUDE_EL0);
+	write_regn_el0(pmevtyper, 1, SW_INCR | PMEVTYPER_EXCLUDE_EL0);
+	write_sysreg_s(0x3, PMCNTENSET_EL0);
+	write_regn_el0(pmevcntr, 0, PRE_OVERFLOW);
+	write_regn_el0(pmevcntr, 1, PRE_OVERFLOW);
+	isb();
+
+	/* interrupts are disabled */
+
+	mem_access_loop(addr, 200, pmu.pmcr_ro | PMU_PMCR_E);
+	report(expect_interrupts(0), "no overflow interrupt received");
+
+	set_pmcr(pmu.pmcr_ro | PMU_PMCR_E);
+	for (i = 0; i < 100; i++)
+		write_sysreg(0x2, pmswinc_el0);
+
+	set_pmcr(pmu.pmcr_ro);
+	report(expect_interrupts(0), "no overflow interrupt received");
+
+	/* enable interrupts */
+
+	pmu_reset_stats();
+
+	write_regn_el0(pmevcntr, 0, PRE_OVERFLOW);
+	write_regn_el0(pmevcntr, 1, PRE_OVERFLOW);
+	write_sysreg(ALL_SET, pmintenset_el1);
+	isb();
+
+	mem_access_loop(addr, 200, pmu.pmcr_ro | PMU_PMCR_E);
+	for (i = 0; i < 100; i++)
+		write_sysreg(0x3, pmswinc_el0);
+
+	mem_access_loop(addr, 200, pmu.pmcr_ro);
+	report_info("overflow=0x%lx", read_sysreg(pmovsclr_el0));
+	report(expect_interrupts(0x3),
+		"overflow interrupts expected on #0 and #1");
+
+	/* promote to 64-b */
+
+	pmu_reset_stats();
+
+	write_regn_el0(pmevtyper, 1, CHAIN | PMEVTYPER_EXCLUDE_EL0);
+	write_regn_el0(pmevcntr, 0, PRE_OVERFLOW);
+	isb();
+	mem_access_loop(addr, 200, pmu.pmcr_ro | PMU_PMCR_E);
+	report(expect_interrupts(0),
+		"no overflow interrupt expected on 32b boundary");
+
+	/* overflow on odd counter */
+	pmu_reset_stats();
+	write_regn_el0(pmevcntr, 0, PRE_OVERFLOW);
+	write_regn_el0(pmevcntr, 1, ALL_SET);
+	isb();
+	mem_access_loop(addr, 400, pmu.pmcr_ro | PMU_PMCR_E);
+	report(expect_interrupts(0x2),
+		"expect overflow interrupt on odd counter");
+}
 #endif
 
 /*
@@ -932,6 +1067,10 @@ int main(int argc, char *argv[])
 		report_prefix_push(argv[1]);
 		test_chain_promotion();
 		report_prefix_pop();
+	} else if (strcmp(argv[1], "pmu-overflow-interrupt") == 0) {
+		report_prefix_push(argv[1]);
+		test_overflow_interrupt();
+		report_prefix_pop();
 	} else {
 		report_abort("Unknown sub-test '%s'", argv[1]);
 	}
diff --git a/arm/unittests.cfg b/arm/unittests.cfg
index 1b0c8c8..455fd10 100644
--- a/arm/unittests.cfg
+++ b/arm/unittests.cfg
@@ -114,6 +114,12 @@ groups = pmu
 arch = arm64
 extra_params = -append 'pmu-chain-promotion'
 
+[pmu-overflow-interrupt]
+file = pmu.flat
+groups = pmu
+arch = arm64
+extra_params = -append 'pmu-overflow-interrupt'
+
 # Test PMU support (TCG) with -icount IPC=1
 #[pmu-tcg-icount-1]
 #file = pmu.flat
-- 
2.20.1

[kvm-unit-tests PATCH v3 12/12] arm: pmu: Test overflow interrupts

[Eric Auger]

Test overflows for MEM_ACCESS and SW_INCR events. Also tests
overflows with 64-bit events.

Signed-off-by: Eric Auger <eric.auger@redhat.com>

---

v2 -> v3:
- added prefix pop
- added pmu_stats.unexpected
- added pmu- prefix
- remove traces in irq_handler()

v1 -> v2:
- inline setup_irq() code
---
 arm/pmu.c         | 139 ++++++++++++++++++++++++++++++++++++++++++++++
 arm/unittests.cfg |   6 ++
 2 files changed, 145 insertions(+)

diff --git a/arm/pmu.c b/arm/pmu.c
index 79c2a0f..628249c 100644
--- a/arm/pmu.c
+++ b/arm/pmu.c
@@ -59,12 +59,20 @@
 #define PRE_OVERFLOW		0xFFFFFFF0
 #define PRE_OVERFLOW2		0xFFFFFFDC
 
+#define PMU_PPI			23
+
 struct pmu {
 	unsigned int version;
 	unsigned int nb_implemented_counters;
 	uint32_t pmcr_ro;
 };
 
+struct pmu_stats {
+	unsigned long bitmap;
+	uint32_t interrupts[32];
+	bool unexpected;
+};
+
 static struct pmu pmu;
 
 #if defined(__arm__)
@@ -146,6 +154,7 @@ static void test_sw_incr(void) {}
 static void test_chained_counters(void) {}
 static void test_chained_sw_incr(void) {}
 static void test_chain_promotion(void) {}
+static void test_overflow_interrupt(void) {}
 
 #elif defined(__aarch64__)
 #define ID_AA64DFR0_PERFMON_SHIFT 8
@@ -276,6 +285,43 @@ asm volatile(
 	: "x9", "x10", "cc");
 }
 
+static struct pmu_stats pmu_stats;
+
+static void irq_handler(struct pt_regs *regs)
+{
+	uint32_t irqstat, irqnr;
+
+	irqstat = gic_read_iar();
+	irqnr = gic_iar_irqnr(irqstat);
+
+	if (irqnr == PMU_PPI) {
+		unsigned long overflows = read_sysreg(pmovsclr_el0);
+		int i;
+
+		for (i = 0; i < 32; i++) {
+			if (test_and_clear_bit(i, &overflows)) {
+				pmu_stats.interrupts[i]++;
+				pmu_stats.bitmap |= 1 << i;
+			}
+		}
+		write_sysreg(ALL_SET, pmovsclr_el0);
+	} else {
+		pmu_stats.unexpected = true;
+	}
+	gic_write_eoir(irqstat);
+}
+
+static void pmu_reset_stats(void)
+{
+	int i;
+
+	for (i = 0; i < 32; i++)
+		pmu_stats.interrupts[i] = 0;
+
+	pmu_stats.bitmap = 0;
+	pmu_stats.unexpected = false;
+}
+
 static void pmu_reset(void)
 {
 	/* reset all counters, counting disabled at PMCR level*/
@@ -286,6 +332,7 @@ static void pmu_reset(void)
 	write_sysreg(ALL_SET, pmovsclr_el0);
 	/* disable overflow interrupts on all counters */
 	write_sysreg(ALL_SET, pmintenclr_el1);
+	pmu_reset_stats();
 	isb();
 }
 
@@ -730,6 +777,94 @@ static void test_chain_promotion(void)
 			read_sysreg(pmovsclr_el0));
 }
 
+static bool expect_interrupts(uint32_t bitmap)
+{
+	int i;
+
+	if (pmu_stats.bitmap ^ bitmap || pmu_stats.unexpected)
+		return false;
+
+	for (i = 0; i < 32; i++) {
+		if (test_and_clear_bit(i, &pmu_stats.bitmap))
+			if (pmu_stats.interrupts[i] != 1)
+				return false;
+	}
+	return true;
+}
+
+static void test_overflow_interrupt(void)
+{
+	uint32_t events[] = {MEM_ACCESS, SW_INCR};
+	void *addr = malloc(PAGE_SIZE);
+	int i;
+
+	if (!satisfy_prerequisites(events, ARRAY_SIZE(events)))
+		return;
+
+	gic_enable_defaults();
+	install_irq_handler(EL1H_IRQ, irq_handler);
+	local_irq_enable();
+	gic_enable_irq(23);
+
+	pmu_reset();
+
+	write_regn_el0(pmevtyper, 0, MEM_ACCESS | PMEVTYPER_EXCLUDE_EL0);
+	write_regn_el0(pmevtyper, 1, SW_INCR | PMEVTYPER_EXCLUDE_EL0);
+	write_sysreg_s(0x3, PMCNTENSET_EL0);
+	write_regn_el0(pmevcntr, 0, PRE_OVERFLOW);
+	write_regn_el0(pmevcntr, 1, PRE_OVERFLOW);
+	isb();
+
+	/* interrupts are disabled */
+
+	mem_access_loop(addr, 200, pmu.pmcr_ro | PMU_PMCR_E);
+	report(expect_interrupts(0), "no overflow interrupt received");
+
+	set_pmcr(pmu.pmcr_ro | PMU_PMCR_E);
+	for (i = 0; i < 100; i++)
+		write_sysreg(0x2, pmswinc_el0);
+
+	set_pmcr(pmu.pmcr_ro);
+	report(expect_interrupts(0), "no overflow interrupt received");
+
+	/* enable interrupts */
+
+	pmu_reset_stats();
+
+	write_regn_el0(pmevcntr, 0, PRE_OVERFLOW);
+	write_regn_el0(pmevcntr, 1, PRE_OVERFLOW);
+	write_sysreg(ALL_SET, pmintenset_el1);
+	isb();
+
+	mem_access_loop(addr, 200, pmu.pmcr_ro | PMU_PMCR_E);
+	for (i = 0; i < 100; i++)
+		write_sysreg(0x3, pmswinc_el0);
+
+	mem_access_loop(addr, 200, pmu.pmcr_ro);
+	report_info("overflow=0x%lx", read_sysreg(pmovsclr_el0));
+	report(expect_interrupts(0x3),
+		"overflow interrupts expected on #0 and #1");
+
+	/* promote to 64-b */
+
+	pmu_reset_stats();
+
+	write_regn_el0(pmevtyper, 1, CHAIN | PMEVTYPER_EXCLUDE_EL0);
+	write_regn_el0(pmevcntr, 0, PRE_OVERFLOW);
+	isb();
+	mem_access_loop(addr, 200, pmu.pmcr_ro | PMU_PMCR_E);
+	report(expect_interrupts(0),
+		"no overflow interrupt expected on 32b boundary");
+
+	/* overflow on odd counter */
+	pmu_reset_stats();
+	write_regn_el0(pmevcntr, 0, PRE_OVERFLOW);
+	write_regn_el0(pmevcntr, 1, ALL_SET);
+	isb();
+	mem_access_loop(addr, 400, pmu.pmcr_ro | PMU_PMCR_E);
+	report(expect_interrupts(0x2),
+		"expect overflow interrupt on odd counter");
+}
 #endif
 
 /*
@@ -932,6 +1067,10 @@ int main(int argc, char *argv[])
 		report_prefix_push(argv[1]);
 		test_chain_promotion();
 		report_prefix_pop();
+	} else if (strcmp(argv[1], "pmu-overflow-interrupt") == 0) {
+		report_prefix_push(argv[1]);
+		test_overflow_interrupt();
+		report_prefix_pop();
 	} else {
 		report_abort("Unknown sub-test '%s'", argv[1]);
 	}
diff --git a/arm/unittests.cfg b/arm/unittests.cfg
index 1b0c8c8..455fd10 100644
--- a/arm/unittests.cfg
+++ b/arm/unittests.cfg
@@ -114,6 +114,12 @@ groups = pmu
 arch = arm64
 extra_params = -append 'pmu-chain-promotion'
 
+[pmu-overflow-interrupt]
+file = pmu.flat
+groups = pmu
+arch = arm64
+extra_params = -append 'pmu-overflow-interrupt'
+
 # Test PMU support (TCG) with -icount IPC=1
 #[pmu-tcg-icount-1]
 #file = pmu.flat
-- 
2.20.1

_______________________________________________
kvmarm mailing list
kvmarm@lists.cs.columbia.edu
https://lists.cs.columbia.edu/mailman/listinfo/kvmarm