Messages in this thread | | | From | "Ulrich Windl" <> | Date | Thu, 12 Aug 1999 08:23:34 +0200 | Subject | 2.3 code freeze: time mega patch? |
| |
Dear hackers,
I have made a patch since 2.2pre7 that significantly updates the kernel time keeping. Specifically it presents virtual nanosecond resolution and extreme fine tuning of clock drifts.
The adjtimex function had been changed, but only very few, highly specialized applications would require recompilation.
The last release of the patch was "PPSkit-0.7.0" for Linux-2.2.9, and that version runs extremly fine on i386: One machine running for weeks keeps time within 5ms; that's quite fine for an old i386SX/16MHz with a WD8013 LAN card and an nnon-enhanced IDE harddrive using the old XT disk driver, don't you agree?
At home my Pentium-100 runs with an accurace of less than 1us.
Unfortunately I can't validate other architectures. Specifically the code to interpolate between ticks has to be adapted (if it was too much assembler or too hardware-specific). I did as much of the work as I was aware, but in the past nobody seemed to care. The other part that needs attention is reading and updating the battery-buffered clock of the hardware.
Yesterday I successfully validated the patch for Linux-2.2.11.
Now the question is: Can this patch be a candidate for 2.4???
If so, do I need to get the most recent 2.3.x kernel and merge my patch into it, or will someone else try it.
Anyway, people are invited to send updates for PPSkit-0.7.0 (if required), and others please wait for the next release that should be available real soon (either as 0.7.1 or 0.8.0, depending on some information I don't have yet).
As I'm not subscribed to this list, please send feedback CC'd to me.
Thank you.
Ulrich
- 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/
| |