Subject: Re: Musings on ld.elf_so and symbol lookup caching
To: <>
From: David Laight <>
List: tech-userlevel
Date: 10/04/2002 19:50:45
> BTW, I count the debugging output changes as part of the `useless'
> part.  You can already get all of this information from ktrace(1), and
> it adds *less* overhead and so is more accurate.

You can't - because there are no system calls involved in symbol
fixup, neither does ktrace timestamp its tracepoints.
The timestamp was VERY useful in sorting out where the
bottlenecks were.

Being able to enable different levels of tracing at run time
is also benefitial since it allows different timings to
be measured from the same built.

The timestamp code is only complicated because the i386
(aka PC) timer is a very slow device.


David Laight: