| Message ID | 20180929084311.15600-4-peterx@redhat.com (mailing list archive) |
|---|---|
| State | Accepted |
| Headers | show |
| Series | userfaultfd: selftests: cleanups and trivial fixes | expand |
On Sat, Sep 29, 2018 at 04:43:11PM +0800, Peter Xu wrote: > Now we recycle the uffd servicing threads earlier than the lock > threads. It might happen that when the lock thread is still blocked at > a pthread mutex lock while the servicing thread has already quitted for > the cpu so the lock thread will be blocked forever and hang the test > program. To fix the possible race, recycle the lock threads first. > > This never happens with current missing-only tests, but when I start to > run the write-protection tests (the feature is not yet posted upstream) > it happens every time of the run possibly because in that new test we'll > need to service two page faults for each lock operation. > > Signed-off-by: Peter Xu <peterx@redhat.com> Acked-by: Mike Rapoport <rppt@linux.vnt.ibm.com> > --- > tools/testing/selftests/vm/userfaultfd.c | 11 ++++++----- > 1 file changed, 6 insertions(+), 5 deletions(-) > > diff --git a/tools/testing/selftests/vm/userfaultfd.c b/tools/testing/selftests/vm/userfaultfd.c > index f79706f13ce7..a388675b15af 100644 > --- a/tools/testing/selftests/vm/userfaultfd.c > +++ b/tools/testing/selftests/vm/userfaultfd.c > @@ -623,6 +623,12 @@ static int stress(unsigned long *userfaults) > if (uffd_test_ops->release_pages(area_src)) > return 1; > > + > + finished = 1; > + for (cpu = 0; cpu < nr_cpus; cpu++) > + if (pthread_join(locking_threads[cpu], NULL)) > + return 1; > + > for (cpu = 0; cpu < nr_cpus; cpu++) { > char c; > if (bounces & BOUNCE_POLL) { > @@ -640,11 +646,6 @@ static int stress(unsigned long *userfaults) > } > } > > - finished = 1; > - for (cpu = 0; cpu < nr_cpus; cpu++) > - if (pthread_join(locking_threads[cpu], NULL)) > - return 1; > - > return 0; > } > > -- > 2.17.1 >
diff --git a/tools/testing/selftests/vm/userfaultfd.c b/tools/testing/selftests/vm/userfaultfd.c index f79706f13ce7..a388675b15af 100644 --- a/tools/testing/selftests/vm/userfaultfd.c +++ b/tools/testing/selftests/vm/userfaultfd.c @@ -623,6 +623,12 @@ static int stress(unsigned long *userfaults) if (uffd_test_ops->release_pages(area_src)) return 1; + + finished = 1; + for (cpu = 0; cpu < nr_cpus; cpu++) + if (pthread_join(locking_threads[cpu], NULL)) + return 1; + for (cpu = 0; cpu < nr_cpus; cpu++) { char c; if (bounces & BOUNCE_POLL) { @@ -640,11 +646,6 @@ static int stress(unsigned long *userfaults) } } - finished = 1; - for (cpu = 0; cpu < nr_cpus; cpu++) - if (pthread_join(locking_threads[cpu], NULL)) - return 1; - return 0; }
Now we recycle the uffd servicing threads earlier than the lock threads. It might happen that when the lock thread is still blocked at a pthread mutex lock while the servicing thread has already quitted for the cpu so the lock thread will be blocked forever and hang the test program. To fix the possible race, recycle the lock threads first. This never happens with current missing-only tests, but when I start to run the write-protection tests (the feature is not yet posted upstream) it happens every time of the run possibly because in that new test we'll need to service two page faults for each lock operation. Signed-off-by: Peter Xu <peterx@redhat.com> --- tools/testing/selftests/vm/userfaultfd.c | 11 ++++++----- 1 file changed, 6 insertions(+), 5 deletions(-)