Messages in this thread | | | Date | Mon, 15 Mar 1999 11:05:07 -0500 (EST) | From | Rafael Reilova <> | Subject | Re: [patch] recover lost ticks |
| |
Hi,
On Mon, 15 Mar 1999, Ingo Molnar wrote:
> > On Mon, 15 Mar 1999, Andrea Arcangeli wrote: > > > If the tsc goes as hell gettimeofday will die too, and the TSC is > > explicity declared from Intel as a really safely trustable thing, so if > > the TSC will die I expect many other part of the CPU to die too. > > (i only pointed out that TSC bugs/inconsistencies (which happened in the > past) now might have yet another side-effect. [there was a robustness > claim])
Let me jump in here, real quickly, and just on the matter of broken TSC's...
There are two (known) cases of these:
1) On Cyrix/IBM/Centaur the TSC will stop if Suspend-on-Halt is enabled and a HLT is executed. This is usually a very good thing since power comsuption goes down dramatically, but the time code has to compensate (and does since somewhere in 2.1.12x).
2) A machine that does a suspend-to-disk or similar, where the CPU is actually shutdown. TSC will reset. This can happen on any x86 that does APM.
This is why, IMHO, is not a good idea for other parts of the kernel to use the TSC directly as a time measure, but instead go through gettimeofday(). The time code should take care of the above.
Cheers,
-Rafael
- 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/
| |