Re: It's time to kill statistical profiling

From: Steve Kargl <sgk_at_troutmask.apl.washington.edu>
Date: Mon, 21 Jun 2021 17:57:01 -0700
On Mon, Jun 21, 2021 at 11:25:29AM -0400, Ed Maste wrote:
> On Sun, 20 Jun 2021 at 17:08, Steve Kargl
> <sgk_at_troutmask.apl.washington.edu> wrote:
> >
> > Are there plans to replace the _p.a with something
> > that allows profiling code?  If not, are you upstreaming
> > patches to GCC to disable -pg?  AFAIK, 'gcc10 -pg ... -lm'
> > causes gcc to look for at least libc_p.a and libm_p.a.
> 
> I believe this is how GCC operates indeed, but I'm not aware of the
> implementation or specific details.

For part of the answer, one is looking for gcc/gcc/config/freebsd-spec.h.
This is where -pg maps libc.a to libc_p.a.  I can't easily track down
libm.a to libm_p.a until later this week.

A bigger problem for FreeBSD and GCC is that there appears to be
no active FreeBSD GCC maintainer(s).  I submitted a GCC patch
for the C language that fixes a few hundred C language testsuite
failures, and a C++ patch that fixes complex arithmetic.  Both 
are languishing in GCC bugzilla. 

-- 
Steve
Received on Tue Jun 22 2021 - 00:57:01 UTC

Original text of this message