Date: Thu, 14 Jun 2007 21:22:53 +1000 (EST) From: Bruce Evans <brde@optusnet.com.au> To: Kip Macy <kip.macy@gmail.com> Cc: cvs-all@FreeBSD.org, cvs-src@FreeBSD.org, src-committers@FreeBSD.org, Bruce Evans <bde@FreeBSD.org>, Bruce Evans <brde@optusnet.com.au> Subject: Re: cvs commit: src/sys/libkern mcount.c Message-ID: <20070614211730.F33664@besplex.bde.org> In-Reply-To: <b1fa29170706131153k44b1e4a8h3101bf175f040012@mail.gmail.com> References: <200706130617.l5D6HncF038605@repoman.freebsd.org> <b1fa29170706122337y558ca741k273e4babb78c936d@mail.gmail.com> <20070613184656.N25269@delplex.bde.org> <b1fa29170706131014t6a9a3d20t228efbfc17b97483@mail.gmail.com> <20070614043846.T27413@besplex.bde.org> <b1fa29170706131153k44b1e4a8h3101bf175f040012@mail.gmail.com>
next in thread | previous in thread | raw e-mail | index | archive | help
On Wed, 13 Jun 2007, Kip Macy wrote: >> It's unlikely because 0.01% of users configure kernel profiling and >> you didn't report configuring it. >> Profiling doesn't hang for me, but I use uncommitted fixes for locking, >> and only use it on amd64 and i386, and haven't tried it under SMP >> lately. Profiling never worked right for SMP and hangs are one >> possibility for it not working. -pp is only supported for amd64 and >> i386. I only tried it lately to see if gcc-4.2 broke it. gcc-4.2 >> certainly broke -pp, but AFAIK not in a way that causes additional >> hangs or breaks -p. > > "-p" works "-pp" does not. This is a dual Woodcrest (2x2) running > x86_64. I guess I can try running with SMP disabled as another data > point. Try my fixes sent to freebsd-amd64. gcc-4.2 will give unusuable (but faster) results with -pp since the main known new problem with gcc-4.2 is it not emitting any profiler-epilogue calls. Bruce
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20070614211730.F33664>