KVM: selftests: Fix target thread to be migrated in rseq_test
PATCH v4 1/1 KVM: selftests: Fix target thread to be migrated in rseq_test
In rseq_test, there are two threads, which are vCPU thread and migration worker separately. Unfortunately, the test has the wrong PID passed to sched_setaffinity() in the migration worker. It forces migration on the migration worker because zeroed PID represents the calling thread, which is the migration worker itself. It means the vCPU thread is never enforced to migration and it can migrate at any time, which eventually leads to failure as the following logs show.
host# uname -r
5.19.0-rc6-gavin+
host# # cat /proc/cpuinfo | grep processor | tail -n 1
processor : 223
host# pwd /home/gavin/sandbox/linux.main/tools/testing/selftests/kvm
host# for i in `seq 1 100`; do \
echo "--------> $i"; ./rseq_test; done
--------> 1
--------> 2
--------> 3
--------> 4
--------> 5
--------> 6
==== Test Assertion Failure ====
rseq_test.c:265: rseq_cpu == cpu
pid=3925 tid=3925 errno=4 - Interrupted system call
1 0x0000000000401963: main at rseq_test.c:265 (discriminator 2)
2 0x0000ffffb044affb: ?? ??:0
3 0x0000ffffb044b0c7: ?? ??:0
4 0x0000000000401a6f: _start at ??:?
rseq CPU = 4, sched CPU = 27Fix the issue by passing correct parameter, TID of the vCPU thread, to sched_setaffinity() in the migration worker.
---
v4: Pick the code change as Sean suggested.
--- tools/testing/selftests/kvm/rseq_test.c | 8 +++++--- 1 file changed, 5 insertions(+), 3 deletions(-)
tools/testing/selftests/kvm/rseq_test.c +5 -3
@@ -82,8 +82,9 @@ static int next_cpu(int cpu)return cpu; } -static void *migration_worker (void *ign) +static void *migration_worker (void *__rseq_tid) { +pid_t rseq_tid = (pid_t )(unsigned long )__rseq_tid;cpu_set_t allowed_mask;int r, i, cpu;
@@ -106,7 +107,7 @@ static void *migration_worker(void *ign)* stable, i.e. while changing affinity is in-progress. */ smp_wmb (); - r =sched_setaffinity (0 ,sizeof (allowed_mask), &allowed_mask); + r =sched_setaffinity (rseq_tid,sizeof (allowed_mask), &allowed_mask);TEST_ASSERT (!r,"sched_setaffinity failed, errno = %d (%s)" , errno,strerror (errno));smp_wmb ();
@@ -231,7 +232,8 @@ int main(int argc, char *argv[]) vm =vm_create_default (VCPU_ID ,0 , guest_code);ucall_init (vm,NULL ); -pthread_create (&migration_thread,NULL , migration_worker,0 ); +pthread_create (&migration_thread,NULL , migration_worker, + (void *)(unsigned long )gettid ());for (i =0 ; !done; i++) {vcpu_run (vm,VCPU_ID );
Signature · 6 lines
2.23.0 _______________________________________________ kvmarm mailing list [email protected] https://lists.cs.columbia.edu/mailman/listinfo/kvmarm
Quoted 80 lines · Gavin Shan
In
rseq_test, there are two threads, which are vCPU thread and migration worker separately. Unfortunately, the test has the wrong PID passed tosched_setaffinity()in the migration worker. It forces migration on the migration worker because zeroed PID represents the calling thread, which is the migration worker itself. It means the vCPU thread is never enforced to migration and it can migrate at any time, which eventually leads to failure as the following logs show.host# uname -r
5.19.0-rc6-gavin+
host# # cat/proc/cpuinfo| grep processor | tail -n 1
processor : 223
host# pwd/home/gavin/sandbox/linux.main/tools/testing/selftests/kvmhost# for i in `seq 1 100`; do \ echo "--------> $i"; ./rseq_test; done --------> 1 --------> 2 --------> 3 --------> 4 --------> 5 --------> 6 ==== Test Assertion Failure ==== rseq_test.c:265: rseq_cpu == cpu pid=3925 tid=3925 errno=4 - Interrupted system call 1 0x0000000000401963: main at rseq_test.c:265 (discriminator 2) 2 0x0000ffffb044affb: ?? ??:0 3 0x0000ffffb044b0c7: ?? ??:0 4 0x0000000000401a6f: _start at ??:? rseq CPU = 4, sched CPU = 27Fix the issue by passing correct parameter, TID of the vCPU thread, to
sched_setaffinity()in the migration worker.Fixes 61e52f1630f5 ("KVM: selftests: Add a test for KVM_RUN+rseq to detect task migration bugs")Suggested-by Sean Christopherson <[email protected]>Signed-off-by Gavin Shan <[email protected]>Reviewed-by Oliver Upton <[email protected]>---
v4: Pick the code change as Sean suggested.--- tools/testing/selftests/kvm/rseq_test.c | 8 +++++--- 1 file changed, 5 insertions(+), 3 deletions(-)tools/testing/selftests/kvm/rseq_test.c +5
-3@@ -82,8 +82,9 @@ static int next_cpu(int cpu)return cpu; } -static void *migration_worker (void *ign) +static void *migration_worker (void *__rseq_tid) { +pid_t rseq_tid = (pid_t )(unsigned long )__rseq_tid;cpu_set_t allowed_mask;int r, i, cpu;@@ -106,7 +107,7 @@ static void *migration_worker(void *ign)* stable, i.e. while changing affinity is in-progress. */ smp_wmb (); - r =sched_setaffinity (0 ,sizeof (allowed_mask), &allowed_mask); + r =sched_setaffinity (rseq_tid,sizeof (allowed_mask), &allowed_mask);TEST_ASSERT (!r,"sched_setaffinity failed, errno = %d (%s)" , errno,strerror (errno));smp_wmb ();@@ -231,7 +232,8 @@ int main(int argc, char *argv[]) vm =vm_create_default (VCPU_ID ,0 , guest_code);ucall_init (vm,NULL ); -pthread_create (&migration_thread,NULL , migration_worker,0 ); +pthread_create (&migration_thread,NULL , migration_worker, + (void *)(unsigned long )gettid ());for (i =0 ; !done; i++) {vcpu_run (vm,VCPU_ID );Signature · 1 line
2.23.0
Thanks for figuring this out Gavin and Sean!
Quoted 69 lines · Gavin Shan
In
rseq_test, there are two threads, which are vCPU thread and migration worker separately. Unfortunately, the test has the wrong PID passed tosched_setaffinity()in the migration worker. It forces migration on the migration worker because zeroed PID represents the calling thread, which is the migration worker itself. It means the vCPU thread is never enforced to migration and it can migrate at any time, which eventually leads to failure as the following logs show.host# uname -r 5.19.0-rc6-gavin+ host# # cat /proc/cpuinfo | grep processor | tail -n 1 processor : 223 host# pwd /home/gavin/sandbox/linux.main/tools/testing/selftests/kvm host# for i in `seq 1 100`; do \ echo "--------> $i"; ./rseq_test; done --------> 1 --------> 2 --------> 3 --------> 4 --------> 5 --------> 6 ==== Test Assertion Failure ==== rseq_test.c:265: rseq_cpu == cpu pid=3925 tid=3925 errno=4 - Interrupted system call 1 0x0000000000401963: main at rseq_test.c:265 (discriminator 2) 2 0x0000ffffb044affb: ?? ??:0 3 0x0000ffffb044b0c7: ?? ??:0 4 0x0000000000401a6f: _start at ??:? rseq CPU = 4, sched CPU = 27Fix the issue by passing correct parameter, TID of the vCPU thread, to
sched_setaffinity()in the migration worker.Fixes 61e52f1630f5 ("KVM: selftests: Add a test for KVM_RUN+rseq to detect task migration bugs")Suggested-by Sean Christopherson <[email protected]>Signed-off-by Gavin Shan <[email protected]>Reviewed-by Oliver Upton <[email protected]>---
v4: Pick the code change as Sean suggested.--- tools/testing/selftests/kvm/rseq_test.c | 8 +++++--- 1 file changed, 5 insertions(+), 3 deletions(-)tools/testing/selftests/kvm/rseq_test.c +3
-2@@ -82,8 +82,9 @@ static int next_cpu(int cpu)return cpu; } -static void *migration_worker (void *ign) +static void *migration_worker (void *__rseq_tid) { +pid_t rseq_tid = (pid_t )(unsigned long )__rseq_tid;cpu_set_t allowed_mask;int r, i, cpu;@@ -106,7 +107,7 @@ static void *migration_worker(void *ign)* stable, i.e. while changing affinity is in-progress. */ smp_wmb (); - r =sched_setaffinity (0 ,sizeof (allowed_mask), &allowed_mask); + r =sched_setaffinity (rseq_tid,sizeof (allowed_mask), &allowed_mask);TEST_ASSERT (!r,"sched_setaffinity failed, errno = %d (%s)" , errno,strerror (errno));smp_wmb ();
tools/testing/selftests/kvm/rseq_test.c int main(int argc, char *argv[]) · line 239
Quoted 9 lines
tools/testing/selftests/kvm/rseq_test.c +2
-1@@ -231,7 +232,8 @@ int main(int argc, char *argv[]) vm =vm_create_default (VCPU_ID ,0 , guest_code);ucall_init (vm,NULL ); -pthread_create (&migration_thread,NULL , migration_worker,0 ); +pthread_create (&migration_thread,NULL , migration_worker, + (void *)(unsigned long )gettid ());for (i =0 ; !done; i++) {vcpu_run (vm,VCPU_ID );
Queued, thanks.
Paolo
_______________________________________________
kvmarm mailing list
[email protected]
https://lists.cs.columbia.edu/mailman/listinfo/kvmarm