Re: CVS commit: src

[Joerg Sonnenberger <joerg%bec.de@localhost>]

On Tue, Oct 26, 2021 at 04:00:41AM +0900, Ryo Shimizu wrote:
> 
> >On Mon, Oct 25, 2021 at 07:54:45AM +0000, Ryo Shimizu wrote:
> >> Module Name:	src
> >> Committed By:	ryo
> >> Date:		Mon Oct 25 07:54:44 UTC 2021
> >> 
> >> Modified Files:
> >> 	src/share/mk: bsd.README bsd.lib.mk
> >> 	src/sys/conf: Makefile.kern.inc
> >> 	src/sys/lib/libkern: Makefile.libkern
> >> 
> >> Log Message:
> >> In some arch, _mcount() would be called recursively when built with COPTS=-O0.
> >> 
> >> Normally, functions called from mcount.c are expected to be expanded inline,
> >> so _mcount() will never be called recursively. But when build with COPTS=-O0,
> >> `static inline' functions aren't inlined, and _mcount() will be called
> >> recursively.
> >
> >So why not fix that by actually using always_inline (i.e.
> >__always_inline)?
> >
> >Joerg
> 
> Yes, that is correct. That method is also valid and should be done separately.
> 
> However, it is more direct to not add -pg to mcount.c.
> Also, I think this commit is valid because it is useful to be able to choose
> not to do per-file profiling.

Personally, I would prefer to just kill -pg support completely, but
that's a separate discussion. I don't think per-file profiling is that
useful and just begging for very confusing results. So I would strongly
prefer if this was done properly.

Joerg