Are posix regcomp and regexec threadsafe? In specific, on GNU libc?

Question

Two separate questions here really: Can I use regexes in a multithreaded program without locking and, if so, can I use the same regex_t at the same time in multiple threads? I can't find an answer on Google or the manpages.

Answer 1

http://www.opengroup.org/onlinepubs/9699919799/functions/V2_chap02.html

2.9.1 Thread-Safety

All functions defined by this volume of POSIX.1-2008 shall be thread-safe, except that the following functions1 need not be thread-safe.

...

regexec and regcomp are not in that list, so they are required to be thread-safe.

See also: http://www.opengroup.org/onlinepubs/9699919799/functions/regcomp.html

Part of the rationale text reads:

The interface is defined so that the matched substrings rm_sp and rm_ep are in a separate regmatch_t structure instead of in regex_t. This allows a single compiled RE to be used simultaneously in several contexts; in main() and a signal handler, perhaps, or in multiple threads of lightweight processes.

Answer 2

Can I use regexes in a multithreaded program without locking

Different ones, yes.

can I use the same regex_t at the same time in multiple threads?

In general: If you plan on doing so, you will have to do the locking around the functions, since few data structures do the locking for you.

regexec: Since regexec however takes a const regex_t, executing regexec seems safe for concurrent execution without locking. (After all, this is POSIX.1-2001, where stupid stuff like static buffers as used in the early BSD APIs usually don't occur anymore.)