CODEWITHSUNDEEP
c++linux
Desperado17
Desperado17

Reputation: 1015

clock_nanosleep - same pointer for request and remain?

Is it safe to submit the same object to clock_nanosleep in the request and remain parameters?

do {
   ret = clock_nanosleep(CLOCK_MONOTONIC, 0,  &t, &t);
} while (ret == EINTR);

I'm currently researching a phenomenon where the system sometimes never leaves that loop (although it usually does).

Regards

Upvotes: 0

Views: 381

Answers (1)

Useless
Useless

Reputation: 67762

Is it safe to submit the same object to clock_nanosleep in the request and remain parameters?

In a POSIX-conforming environment at least, it's fine.

From the POSIX clock_nanosleep doc,

The rqtp and rmtp arguments can point to the same object.

For LINUX specifically, I can't see anything in my local manpage about this, at least not explicitly prohibiting it. The non-NULL remain timespec is only written to if the call is already going to return, with an incomplete relative sleep: it doesn't need to read the request timespec again during or after this.

The manpage does additionally say (of the remain timespec)

This value can then be used to call clock_nanosleep() again and complete a (relative) sleep.

in which case it's certainly desirable for clock_nanosleep to write it back into the same variable.

Upvotes: 1

Related Questions