Currently the code can race as there is a read/write thread handling the stdio
but
there is no guarantee that when the process exits, the thread has handled all
the
data. This results in output where "END:" isn't actually at the end of the logs
but somewhere in the middle of the output.
Synchronisation is hard. The easiest way I can see to fix this is to have a
mutex
for the output and then in the main thread, after the child exits, read any
remaining
data. This avoids concurrent writes corrupting the output and ensures END: is
actually at the end of the test data.
Signed-off-by: Richard Purdie <richard.pur...@linuxfoundation.org>
---
utils.c | 25 +++++++++++++++++++++----
1 file changed, 21 insertions(+), 4 deletions(-)
diff --git a/utils.c b/utils.c
index ec57fa4..65b1df3 100644
--- a/utils.c
+++ b/utils.c
@@ -63,6 +63,7 @@ static struct {
int timeouted;
pid_t pid;
int padding1;
+ pthread_mutex_t fd_lock;
} _child_reader;
static inline char *
@@ -317,12 +318,13 @@ read_child(void *arg)
do {
r = poll(pfds, 2, _child_reader.timeout*1000);
+ pthread_mutex_lock(&_child_reader.fd_lock);
if (r > 0) {
char buf[WAIT_CHILD_BUF_MAX_SIZE];
ssize_t n;
if (pfds[0].revents != 0) {
- n = read(_child_reader.fds[0], buf,
WAIT_CHILD_BUF_MAX_SIZE);
+ n = read(_child_reader.fds[0], buf,
WAIT_CHILD_BUF_MAX_SIZE);
if (n > 0)
fwrite(buf, (size_t)n, 1,
_child_reader.fps[0]);
}
@@ -338,11 +340,13 @@ read_child(void *arg)
// as much data from the system as possible and kill
the test
collect_system_state(_child_reader.fps[0]);
_child_reader.timeouted = 1;
+ pthread_mutex_unlock(&_child_reader.fd_lock);
kill(-_child_reader.pid, SIGKILL);
}
fflush(_child_reader.fps[0]);
fflush(_child_reader.fps[1]);
+ pthread_mutex_unlock(&_child_reader.fd_lock);
} while (1);
return NULL;
@@ -444,6 +448,8 @@ run_ptests(struct ptest_list *head, const struct
ptest_options opts,
int slave;
int pgid = -1;
pthread_t tid;
+ ssize_t n;
+ char buf[WAIT_CHILD_BUF_MAX_SIZE];
if (opts.xml_filename) {
xh = xml_create(ptest_list_length(head), opts.xml_filename);
@@ -453,10 +459,10 @@ run_ptests(struct ptest_list *head, const struct
ptest_options opts,
do
{
- if ((rc = pipe(pipefd_stdout)) == -1)
+ if ((rc = pipe2(pipefd_stdout, O_NONBLOCK)) == -1)
break;
- if ((rc = pipe(pipefd_stderr)) == -1) {
+ if ((rc = pipe2(pipefd_stderr, O_NONBLOCK)) == -1) {
close(pipefd_stdout[0]);
close(pipefd_stdout[1]);
break;
@@ -466,6 +472,11 @@ run_ptests(struct ptest_list *head, const struct
ptest_options opts,
fprintf(fp, "ERROR: Unable to detach from controlling
tty, %s\n", strerror(errno));
}
+ if (pthread_mutex_init(&_child_reader.fd_lock, NULL) != 0) {
+ printf("Failed to init mutex\n");
+ exit(EXIT_FAILURE);
+ }
+
_child_reader.fds[0] = pipefd_stdout[0];
_child_reader.fds[1] = pipefd_stderr[0];
_child_reader.fps[0] = fp;
@@ -535,8 +546,13 @@ run_ptests(struct ptest_list *head, const struct
ptest_options opts,
entime = time(NULL);
duration = entime - sttime;
- /* Now the child has exited, ensure buffers are
in sync before writing */
+ pthread_mutex_lock(&_child_reader.fd_lock);
+ while ((n = read(_child_reader.fds[0], buf,
WAIT_CHILD_BUF_MAX_SIZE)) > 0)
+ fwrite(buf, (size_t)n, 1,
_child_reader.fps[0]);
+ while ((n = read(_child_reader.fds[1], buf,
WAIT_CHILD_BUF_MAX_SIZE)) > 0)
+ fwrite(buf, (size_t)n, 1,
_child_reader.fps[1]);
fflush(NULL);
+ pthread_mutex_unlock(&_child_reader.fd_lock);
if (status) {
fprintf(fp, "\nERROR: Exit status is
%d\n", status);
@@ -558,6 +574,7 @@ run_ptests(struct ptest_list *head, const struct
ptest_options opts,
pthread_cancel(tid);
pthread_join(tid, NULL);
+ pthread_mutex_destroy(&_child_reader.fd_lock);
close(pipefd_stdout[0]); close(pipefd_stdout[1]);
close(pipefd_stderr[0]); close(pipefd_stderr[1]);
--
2.39.2
-=-=-=-=-=-=-=-=-=-=-=-
Links: You receive all messages sent to this group.
View/Reply Online (#60347): https://lists.yoctoproject.org/g/yocto/message/60347
Mute This Topic: https://lists.yoctoproject.org/mt/99625182/21656
Group Owner: yocto+ow...@lists.yoctoproject.org
Unsubscribe:
https://lists.yoctoproject.org/g/yocto/leave/6691583/21656/737036229/xyzzy
[arch...@mail-archive.com]
-=-=-=-=-=-=-=-=-=-=-=-
