gettimeoffset questions.

C. Scott Ananian (cananian@lcs.mit.edu)
Mon, 25 May 1998 12:59:16 -0400 (EDT)

[This question is x86-implementation specific]
It seems the design of gettimeoffset is somewhat suspect on SMP systems.
Can we really guarantee that TSC values are identical in all CPUs of a
multiple-CPU system?

The current gettimeoffset code computes the time by comparing the Time
Stamp Counter on the local CPU with the timestamp counter value saved by
the timer_interrupt, which is always run on the boot CPU, no? There
doesn't seem to be any guarantee on which processor gettimeofday is
executed on, though. Other than bogus intra-jiffy timings, I don't see
any way that this problem would be visible, though.

Also -- what purpose were the 'delay' outb instructions for in the outb_p
macros for x86? I'm not entirely clear on the circumstances in which I
can safely eliminate them.
--Scott
@ @
=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-oOO-(_)-OOo-=-=-=-=-=
C. Scott Ananian: cananian@lcs.mit.edu / Declare the Truth boldly and
Laboratory for Computer Science/Crypto / without hindrance.
Massachusetts Institute of Technology /META-PARRESIAS AKOLUTOS:Acts 28:31
-.-. .-.. .. ..-. ..-. --- .-. -.. ... -.-. --- - - .- -. .- -. .. .- -.
PGP key available via finger and from http://www.pdos.lcs.mit.edu/~cananian

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu