[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [PATCH v2] Prevent vhost-user-blk-test hang
From: |
address@hidden |
Subject: |
Re: [PATCH v2] Prevent vhost-user-blk-test hang |
Date: |
Tue, 7 Sep 2021 16:54:41 -0500 |
User-agent: |
NeoMutt/20210205-739-420e15 |
On Mon, Sep 06, 2021 at 01:25:20PM +0000, Raphael Norwitz wrote:
> In the vhost-user-blk-test, as of now there is nothing stoping
> vhost-user-blk in QEMU writing to the socket right after forking off the
> storage daemon before it has a chance to come up properly, leaving the
> test hanging forever. This intermittently hanging test has caused QEMU
> automation failures reported multiple times on the mailing list [1].
>
> This change makes the storage-daemon notify the vhost-user-blk-test
> that it is fully initialized and ready to handle client connections by
> creating a pidfile on initialiation. This ensures that the storage-daemon
> backend won't miss vhost-user messages and thereby resolves the hang.
>
> [1]
> https://lore.kernel.org/qemu-devel/CAFEAcA8kYpz9LiPNxnWJAPSjc=nv532bEdyfynaBeMeohqBp3A@mail.gmail.com/
>
> Signed-off-by: Raphael Norwitz <raphael.norwitz@nutanix.com>
> ---
> tests/qtest/vhost-user-blk-test.c | 33 ++++++++++++++++++++++++++++++-
> 1 file changed, 32 insertions(+), 1 deletion(-)
>
> diff --git a/tests/qtest/vhost-user-blk-test.c
> b/tests/qtest/vhost-user-blk-test.c
> index 6f108a1b62..78140e6f28 100644
> --- a/tests/qtest/vhost-user-blk-test.c
> +++ b/tests/qtest/vhost-user-blk-test.c
> @@ -24,6 +24,9 @@
> #define TEST_IMAGE_SIZE (64 * 1024 * 1024)
> #define QVIRTIO_BLK_TIMEOUT_US (30 * 1000 * 1000)
> #define PCI_SLOT_HP 0x06
> +#define PIDFILE_RETRIES 5
> +
> +const char *pidfile_format = "/tmp/daemon-%d";
Why is this not static? In fact...
>
> typedef struct {
> pid_t pid;
> @@ -885,7 +888,8 @@ static void start_vhost_user_blk(GString *cmd_line, int
> vus_instances,
> int num_queues)
> {
> const char *vhost_user_blk_bin = qtest_qemu_storage_daemon_binary();
> - int i;
> + int i, err, retries;
> + char *daemon_pidfile_path;
> gchar *img_path;
> GString *storage_daemon_command = g_string_new(NULL);
> QemuStorageDaemonState *qsd;
> @@ -898,6 +902,12 @@ static void start_vhost_user_blk(GString *cmd_line, int
> vus_instances,
> " -object memory-backend-memfd,id=mem,size=256M,share=on "
> " -M memory-backend=mem -m 256M ");
>
> + err = asprintf(&daemon_pidfile_path, pidfile_format, getpid());
...action at a distance makes it harder for gcc to warn about bad
formats. I'd just inline "/tmp/daemon-%d" here in the asprintf call,
and drop pidfile_format altogether.
Why are we using bare asprintf instead of glib's g_strdup_printf?
> + if (err == -1) {
> + fprintf(stderr, "Failed to format storage-daemon pidfile name %m");
%m in printf is a glibc-ism; not portable to non-Linux. Do we care?
> + abort();
Rather than directly abort, since this is a glib test runner, is there
a glib function we should be using? For example, using
g_assert_cmpint((err, !=, -1)?
> + }
> +
> for (i = 0; i < vus_instances; i++) {
> int fd;
> char *sock_path = create_listen_socket(&fd);
> @@ -914,6 +924,9 @@ static void start_vhost_user_blk(GString *cmd_line, int
> vus_instances,
> i + 1, sock_path);
> }
>
> + g_string_append_printf(storage_daemon_command, "--pidfile %s",
Missing a space on the tail end if there are more arguments to append
to the command line.
> + daemon_pidfile_path);
> +
> g_test_message("starting vhost-user backend: %s",
> storage_daemon_command->str);
> pid_t pid = fork();
> @@ -930,7 +943,25 @@ static void start_vhost_user_blk(GString *cmd_line, int
> vus_instances,
> execlp("/bin/sh", "sh", "-c", storage_daemon_command->str, NULL);
> exit(1);
> }
> +
> + /*
> + * Ensure the storage-daemon has come up properly before allowing the
> + * test to proceed.
> + */
> + retries = 0;
> + while (access(daemon_pidfile_path, F_OK) != 0) {
> + if (retries > PIDFILE_RETRIES) {
> + fprintf(stderr, "The storage-daemon failed to come up after %d "
> + "seconds - killing the test", PIDFILE_RETRIES);
> + abort();
Again, would some form of g_assert*() be better than bare abort?
> + }
> +
> + retries++;
> + usleep(1000);
We're inconsistent on whether qtest files use bare usleep or g_usleep.
> + }
> +
> g_string_free(storage_daemon_command, true);
> + free(daemon_pidfile_path);
Do you want to unlink() the file (if it exists)?
>
> qsd = g_new(QemuStorageDaemonState, 1);
> qsd->pid = pid;
> --
> 2.20.1
>
--
Eric Blake, Principal Software Engineer
Red Hat, Inc. +1-919-301-3266
Virtualization: qemu.org | libvirt.org