Re: It's time to kill statistical profiling

From: Poul-Henning Kamp <phk_at_phk.freebsd.dk>
Date: Fri, 18 Jun 2021 13:43:13 +0000
--------
Warner Losh writes:

> Deleting the code from the kernel is the easy bit, I guess. I generally support
> this notion, but don't have the time to hunt down all the stuff in base and
> ports that uses these and make appropriate changes (up to and including git
> rm).

It's no rush, we've had this obsolete facility for 20-25 years already, we
can sweep the corners as we get to them.

I think the main decision is that we want to do it, and if we want to keep
profiled libararies for gcov or if we want to ditch profiled libraries.

The main reason I prefer to change to gcov libraries is that it
keeps the .mk infrastructure for having systematic variant librararies
in place, over the years people have abused that for a lot of
interesting experiments and hacks.

I also suspect that s/-p/-fprofile-arcs -ftest-coverage/ is a LOT
easier than removing that infrastructure :-)

-- 
Poul-Henning Kamp       | UNIX since Zilog Zeus 3.20
phk_at_FreeBSD.ORG         | TCP/IP since RFC 956
FreeBSD committer       | BSD since 4.3-tahoe    
Never attribute to malice what can adequately be explained by incompetence.
Received on Fri Jun 18 2021 - 13:43:13 UTC

Original text of this message