Add pthread_get_name_np(3).

[kib]

Summary

The function retrieves the thread name previously set by pthread_set_name_np(3). The name is cached in the process memory.

Requested by: Willem Jan Withagen <[e-mail]>
Man page update: Yuri Pankov <[e-mail]>

[bjk]

Just a couple of grammar nits on the manpage.

share/man/man3/pthread_set_name_np.3

52 ↗

(On Diff #46621)

"function"

72 ↗

(On Diff #46621)

No "functions" here.

[kib]

Grammar fixes.

[kib]

Mention authors.

[• ian]

lib/libthr/thread/thr_info.c

91 ↗

(On Diff #46624)

What about when thread->name is NULL? Doesn't this need handling, like...

else if (len > 0)
  buf[0] = 0;

[pluknet]

share/man/man3/pthread_set_name_np.3

72 ↗

(On Diff #46624)

Please note that this duplicates what's already (perhaps, vaguely) documented in pthread(3)'s STANDARDS section:

The functions with the pthread_ prefix and _np suffix are non-portable
extensions to POSIX threads.

[kib]

share/man/man3/pthread_set_name_np.3

72 ↗

(On Diff #46624)

Yes, but what do you want me to change there ? I do not see a harm in repeating this information.

[kib]

Handle nameless threads.

[pluknet]

share/man/man3/pthread_set_name_np.3

72 ↗

(On Diff #46624)

Just a FYI, I see no problem here.

[arichardson]

lib/libthr/thread/thr_info.c

57 ↗

(On Diff #46629)

Could this be thread->name = strdup(name) instead? It will also set the value to null on allocation failure.

[kib]

Use strdup(3).

I was so concerned with another use of malloc(3) in libthr, so when I decided that it is fine, failed to realize that I reimplemented strdup(3).