Messages in this thread |  | | | Date | Wed, 28 Jul 1999 13:01:22 +0200 (CEST) | | From | "Robert H. de Vries" <> | | Subject | Re: PATCH: POSIX 1003.1b timer minor fixes |
| |
Ulrich writes:
> Robert de Vries <rhdv@rhdv.cistron.nl> writes: > >> I disagree, because the system call overhead takes more time than the >> execution of a switch statement. > > You can add some clocks on user level but the handling of interrupts > for timers et cannot efficiently be done at user level.
You are absolutely right, but nobody says that you have to implement both clock and timer functionality for each CLOCK_<type>. So some CLOCKs have only clock functionality, some only timer, and the rest has both. On SGI IRIX 6.5 you have for instance three clocks: CLOCK_REALTIME (all clock & timer functions supported) CLOCK_SGI_CYCLE (clock_gettime() and clock_getres() only) CLOCK_SGI_FAST (timer_create() & clock_getres() only)
For each of these clocks there can be a mix of how and where each function is implemented. The decision of where the implementation of each function can be done best is made by the kernel side developer (=Robert) and the C library developer (=Ulrich) in close cooperation.
I am a performance freak, as I make real-time systems. So I am prepared to trade some simplicity of implementation for performance.
Robert
-- Robert H. de Vries PO/SIM Fokker Space B.V. e-mail: R.de.Vries@fokkerspace.nl tel: (+31)71-5245464 fax: (+31)71-5245498
- 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/
|  |