Messages in this thread | | | Date | Tue, 27 Jul 1999 07:15:08 +0200 (CEST) | From | Robert de Vries <> | Subject | Re: POSIX.1b timers comments |
| |
Robert de Vries rhdv@rhdv.cistron.nl
On 26 Jul 1999, Ulrich Drepper wrote:
> I appreciate seeing finally kernel support for this but have a few > problems with the actual implementation: > > - CLONE_ITIMERS > > The meaning of the flag should be reverted. Since sharing itimers > among all threads in a process is the POSIX behaviour this bit must > be set. But it is not in existing implementation for obvious reasons. > > This is a general problem with extensions to the clone interface.
I agree, for every POSIX extension to come a CLONE flag is required, so apart from the CLONE_SIGNALS flag, there is coming an CLONE_SEM, CLONE_SHM and CLONE_MQ.
> > > - MAX_ITIMERS > > This should be TIMER_MAX. POSIX does not define MAX_ITIMERS.
I agree, the internal kernel definition should match the POSIX definition.
> > > - There is no DELAYTIMER_MAX definition. > > This might be missing since the implementation is not yet finished > but in case this is not the reason please add it.
OK, it should be equal to INT_MAX.
BTW it looks like the kernel is lacking a sysconf system call. Or would everybody prefer it to be in the C library? One of the reasons for being in the kernel would be that certain kernel compile time constants are only known in the kernel and not in the C-library. For instance the number of TIMERS could be changed in the kernel, while the C-library is unaware.
Robert
-- Robert de Vries rhdv@rhdv.cistron.nl
- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu Please read the FAQ at http://www.tux.org/lkml/
| |