Date: Sun, 20 Jun 2021 12:17:33 -0400 From: Ed Maste <emaste@freebsd.org> To: Poul-Henning Kamp <phk@phk.freebsd.dk> Cc: "freebsd-arch@freebsd.org" <arch@freebsd.org> Subject: Re: It's time to kill statistical profiling Message-ID: <CAPyFy2B=ybCgy_gq=EF==1K=AKE2eTzoUnx7VFLRPFXTdYERwQ@mail.gmail.com> In-Reply-To: <202106180736.15I7aYmk068064@critter.freebsd.dk> References: <202106180736.15I7aYmk068064@critter.freebsd.dk>
next in thread | previous in thread | raw e-mail | index | archive | help
On Fri, 18 Jun 2021 at 03:36, Poul-Henning Kamp <phk@phk.freebsd.dk> wrote: > > Instead of documenting stathz, I suggest we retire statistical > profiling and convert the profiled libraries to code-coverage > profiling (-fprofile-arcs and -ftest-coverage) I brought up retiring the PROFILE option to build _p libraries last year: https://lists.freebsd.org/pipermail/freebsd-current/2020-January/075105.html There were a couple of objections in that thread, but if it doesn't produce accurate data there's little value in keeping it. > Instead of documenting stathz, I suggest we retire statistical > profiling and convert the profiled libraries to code-coverage > profiling (-fprofile-arcs and -ftest-coverage) Having built-in support for code coverage would be great, but IMO not done by just swapping the current libraries / build infrastructure / options over.
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?CAPyFy2B=ybCgy_gq=EF==1K=AKE2eTzoUnx7VFLRPFXTdYERwQ>