Re: [ANNOUNCE] ktimers subsystem

From: Christopher Friesen
Date: Mon Sep 19 2005 - 23:58:06 EST


Thomas Gleixner wrote:
On Mon, 2005-09-19 at 16:39 -0600, Christopher Friesen wrote:

Thomas Gleixner wrote:

We should rather ask glibc people why gettimeofday() / clock_getttime()
is called inside the library code all over the place for non obvious
reasons.

--flight-recorder style logs

If you want to implement such stuff efficiently you rely on rdtscll() on
x86 or other monotonic easy accessible time souces and not on a
permanent call to gettimeofday.

Not portable across architectures, and doesn't work across all smp/numa environments. Also not easy to compare with other nodes on the network, whereas with ntp-synch'd nodes you can use gettimeofday() for quite accurate correlations.

Please beware me of red herrings. If application developers code with
respect to random OS worst case behaviour then they should not complain
that OS N is having an additional add instruction in one of the pathes.

Actually I'm not complaining about additional add instructions. I was just suggesting some reasons why apps might reasonably want to know the time frequently.

Chris

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/