Skip to contentSkip to conversation

x86: pmu: Test WRMSR on a running counter

PATCH Applied Eric Hankland 3 participants 3 messages Feb 4, 2020 – Feb 5, 2020 KVM

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 x86: pmu: Test WRMSR on a running counter

Ensure that the value of the counter was successfully set to 0 after writing it while the counter was running.

---
 x86/pmu.c | 16 ++++++++++++++++
 1 file changed, 16 insertions(+)
x86/pmu.c +16 -0
index cb8c9e3..8a77993 100644
@@ -419,6 +419,21 @@ static void check_rdpmc(void)
 	report_prefix_pop();
 }
 
+static void check_running_counter_wrmsr(void)
+{
+	pmu_counter_t evt = {
+		.ctr = MSR_IA32_PERFCTR0,
+		.config = EVNTSEL_OS | EVNTSEL_USR | gp_events[1].unit_sel,
+		.count = 0,
+	};
+
+	start_event(&evt);
+	loop();
+	wrmsr(MSR_IA32_PERFCTR0, 0);
+	stop_event(&evt);
+	report("running counter wrmsr", evt.count < gp_events[1].min);
+}
+
 int main(int ac, char **av)
 {
 	struct cpuid id = cpuid(10);
@@ -453,6 +468,7 @@ int main(int ac, char **av)
 	check_counters_many();
 	check_counter_overflow();
 	check_gp_counter_cmask();
+	check_running_counter_wrmsr();
 
 	return report_summary();
 }
#1 lore raw
Quoted 36 lines · Eric Hankland

Ensure that the value of the counter was successfully set to 0 after writing it while the counter was running.

---
 x86/pmu.c | 16 ++++++++++++++++
 1 file changed, 16 insertions(+)
x86/pmu.c +15 -0
index cb8c9e3..8a77993 100644
@@ -419,6 +419,21 @@ static void check_rdpmc(void)
 	report_prefix_pop();
 }
 
+static void check_running_counter_wrmsr(void)
+{
+	pmu_counter_t evt = {
+		.ctr = MSR_IA32_PERFCTR0,
+		.config = EVNTSEL_OS | EVNTSEL_USR | gp_events[1].unit_sel,
+		.count = 0,
+	};
+
+	start_event(&evt);
+	loop();
+	wrmsr(MSR_IA32_PERFCTR0, 0);
+	stop_event(&evt);
+	report("running counter wrmsr", evt.count < gp_events[1].min);
+}
+
 int main(int ac, char **av)
 {
 	struct cpuid id = cpuid(10);

x86/pmu.c int main(int ac, char **av) · line 474

Quoted 7 lines
x86/pmu.c +1 -0
@@ -453,6 +468,7 @@ int main(int ac, char **av)
 	check_counters_many();
 	check_counter_overflow();
 	check_gp_counter_cmask();
+	check_running_counter_wrmsr();
 
 	return report_summary();
 }

You shall not pass [-Werror]:

gcc -mno-red-zone -mno-sse -mno-sse2 -m64 -O1 -g -MMD -MF x86/.pmu.d -fno-strict-aliasing -Wall -Wwrite-strings -Wempty-body -Wuninitialized -Wignored-qualifiers -Werror -fno-omit-frame-pointer -Wno-frame-address -fno-pic -no-pie -Wclobbered -Wunused-but-set-parameter -Wmissing-parameter-type -Wold-style-declaration -Woverride-init -Wmissing-prototypes -Wstrict-prototypes -std=gnu99 -ffreestanding -I /home/vitty/workspace/Upstream/kvm-unit-tests/lib -I /home/vitty/workspace/Upstream/kvm-unit-tests/lib/x86 -I lib -c -o x86/pmu.o x86/pmu.c
x86/pmu.c: In function ‘check_running_counter_wrmsr’:
x86/pmu.c:435:44: error: passing argument 2 of ‘report’ makes pointer from integer without a cast [-Werror=int-conversion]

  435 |  report("running counter wrmsr", evt.count < gp_events[1].min);
      |                                  ~~~~~~~~~~^~~~~~~~~~~~~~~~~~
      |                                            |
      |                                            int
In file included from /home/vitty/workspace/Upstream/kvm-unit-tests/lib/x86/processor.h:4,
                 from x86/pmu.c:3:
/home/vitty/workspace/Upstream/kvm-unit-tests/lib/libcflat.h:102:43: note: expected ‘const char *’ but argument is of type ‘int’
  102 | extern void report(bool pass, const char *msg_fmt, ...)
      |                               ~~~~~~~~~~~~^~~~~~~

cc1: all warnings being treated as errors

Signature · 1 line
Vitaly
maintainer Applied
#2 lore raw
In reply to Eric Hankland
Quoted 13 lines · Eric Hankland

Ensure that the value of the counter was successfully set to 0 after writing it while the counter was running.

---
 x86/pmu.c | 16 ++++++++++++++++
 1 file changed, 16 insertions(+)
x86/pmu.c +0 -0
index cb8c9e3..8a77993 100644

x86/pmu.c static void check_rdpmc(void) · line 419

Quoted 1 line
x86/pmu.c +0 -0
@@ -419,6 +419,21 @@ static void check_rdpmc(void)
 	report_prefix_pop();

Queued, thanks.

Paolo