posix: Fix pthread_barrier_wait() return value #9976
Conversation
The return value from this function is supposed to return "PTHREAD_BARRIER_SERIAL_WAIT" from exactly one of the calling threads, otherwise zero or an error code. Fixes zephyrproject-rtos#9953 Signed-off-by: Andy Ross <andrew.j.ross@intel.com>
andyross
requested review from
nniranjhana and
ramakrishnapallala
as code owners
| @@ -11,7 +11,7 @@ | |||
|
|
|||
| int pthread_barrier_wait(pthread_barrier_t *b) | |||
| { | |||
| int key = irq_lock(); | |||
| int key = irq_lock(), ret, serial = 0; | |||
There was a problem hiding this comment.
Not unless I'm missing something. ret needs to match the return type of the function, and serial is just a flag that is 0 or 1 depending on whether this is the last thread to arrive.
There was a problem hiding this comment.
I am talking about return of irq_lock
| } | ||
| } | ||
|
|
||
| #define PTHREAD_BARRIER_SERIAL_THREAD 0 |
There was a problem hiding this comment.
Okay, according to doc
The constant PTHREAD_BARRIER_SERIAL_THREAD is defined in <pthread.h> and its value shall be distinct from any other value returned by pthread_barrier_wait().
and
the pthread_barrier_wait() function shall return PTHREAD_BARRIER_SERIAL_THREAD for a single (arbitrary) thread synchronized at the barrier and zero for each of the other threads.
So shouldn't the value be something other than 0?
There was a problem hiding this comment.
Uh... yeah. This is what I get for pushing a feature without a test, I guess. Will fix. I also see I misspelled the name of the symbol in the commit message (though it looks correct here).
|
@andyross can you please update this |
|
This is patched in #14003. Closing as a duplicate fix. |
ghost
closed this
The return value from this function is supposed to return
"PTHREAD_BARRIER_SERIAL_WAIT" from exactly one of the calling threads,
otherwise zero or an error code.
Fixes #9953
Signed-off-by: Andy Ross andrew.j.ross@intel.com