Calling clock_gettime very often with rr

Robert O'Callahan robert at ocallahan.org
Mon May 28 21:52:48 UTC 2018


On Tue, May 29, 2018 at 5:38 AM, Benjamin King <benjaminking at web.de> wrote:

> I was considering using rdtsc to avoid too frequent calls to clock_gettime.
> Something along the lines of 'assume that 1 microsecond did not pass unless
> time stamp counter changed by at least X'. But this is also nondeterminism
> that would need to be tracked by rr, no?
>

Yes, and handling RDTSC is even slower than clock_gettime because we must
context switch to the tracer. Unfortunately there is no way to selectively
trap on RDTSCs like there is for system calls --- no seccomp filters for
RDTSC. That could be implemented but would require new kernel APIs. With
the right kernel APIs though, I think we could make it fast under rr.

Rob
-- 
Su ot deraeppa sah dna Rehtaf eht htiw saw hcihw, efil lanrete eht uoy ot
mialcorp ew dna, ti ot yfitset dna ti nees evah ew; deraeppa efil eht. Efil
fo Drow eht gninrecnoc mialcorp ew siht - dehcuot evah sdnah ruo dna ta
dekool evah ew hcihw, seye ruo htiw nees evah ew hcihw, draeh evah ew
hcihw, gninnigeb eht morf saw hcihw taht.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.mozilla.org/pipermail/rr-dev/attachments/20180529/81362aa2/attachment.html>


More information about the rr-dev mailing list