On Fri, 14 Jan 2011, Richard Cochran wrote:

On Thu, Jan 13, 2011 at 06:03:24PM +0100, Thomas Gleixner wrote:

quoted

On Thu, 13 Jan 2011, Richard Cochran wrote:

quoted

 int posix_cpu_clock_getres(const clockid_t which_clock, struct timespec *ts);
 int posix_cpu_clock_get(const clockid_t which_clock, struct timespec *ts);
-int posix_cpu_clock_set(const clockid_t which_clock, const struct timespec *ts);
+int posix_cpu_clock_set(const clockid_t which_clock, struct timespec *ts);

Shouldn't we change the clock_set function to have *ts const in all places ?

Yes, your are right.

quoted

quoted

@@ -293,6 +261,11 @@ static __init int init_posix_timers(void)
 		.clock_adj = do_posix_clock_noadjtime,
 		.timer_create = no_timer_create,
 		.nsleep = no_nsleep,
+		/* defaults: */
+		.nsleep_restart	= common_nsleep_restart,
+		.timer_del	= common_timer_del,
+		.timer_get	= common_timer_get,
+		.timer_set	= common_timer_set,

Hmm, we do not need to set functional entries for clocks which neither
implement timer_create nor nsleep.

I know, but I wanted to be really pendantic about what the previous
code was, and what the new code does.

Before: By leaving the k_clock function pointer NULL, the clock
        selects common_xyz. This common function may or may make
        sense for that clock.

After:  By leaving the k_clock function pointer NULL, the clock will
        return EINVAL for that syscall.

Maybe it would be better to leave the cleaning up of the common crud
as a follow up patch. What do you think?

Fair enough.