On Thu, Apr 27, 2023 at 1:11 PM Peter Xu <peterx@xxxxxxxxxx> wrote:
>
> There's one PER_VCPU_DEBUG in per-vcpu uffd threads but it's never hit.
>
> Trigger that when quit in normal ways (kick pollfd[1]), meanwhile fix the
> number of nanosec calculation.
>
> Signed-off-by: Peter Xu <peterx@xxxxxxxxxx>
Reviewed-by: James Houghton <jthoughton@xxxxxxxxxx>
> ---
> tools/testing/selftests/kvm/lib/userfaultfd_util.c | 4 ++--
> 1 file changed, 2 insertions(+), 2 deletions(-)
>
> diff --git a/tools/testing/selftests/kvm/lib/userfaultfd_util.c b/tools/testing/selftests/kvm/lib/userfaultfd_util.c
> index 92cef20902f1..271f63891581 100644
> --- a/tools/testing/selftests/kvm/lib/userfaultfd_util.c
> +++ b/tools/testing/selftests/kvm/lib/userfaultfd_util.c
> @@ -70,7 +70,7 @@ static void *uffd_handler_thread_fn(void *arg)
> r = read(pollfd[1].fd, &tmp_chr, 1);
> TEST_ASSERT(r == 1,
> "Error reading pipefd in UFFD thread\n");
> - return NULL;
> + break;
> }
>
> if (!(pollfd[0].revents & POLLIN))
> @@ -103,7 +103,7 @@ static void *uffd_handler_thread_fn(void *arg)
> ts_diff = timespec_elapsed(start);
> PER_VCPU_DEBUG("userfaulted %ld pages over %ld.%.9lds. (%f/sec)\n",
> pages, ts_diff.tv_sec, ts_diff.tv_nsec,
> - pages / ((double)ts_diff.tv_sec + (double)ts_diff.tv_nsec / 100000000.0));
> + pages / ((double)ts_diff.tv_sec + (double)ts_diff.tv_nsec / NSEC_PER_SEC));
I almost confused this fix for [1]. Thanks for catching this!
[1]: https://lore.kernel.org/kvm/20230223001805.2971237-1-amoorthy@xxxxxxxxxx/
>
> return NULL;
> }
> --
> 2.39.1
>
