Date: Thu, 22 Dec 2011 18:51:47 -0800 From: Garrett Cooper <yanegomi@gmail.com> To: Alexander Best <arundel@freebsd.org> Cc: "freebsd-current@freebsd.org" <freebsd-current@freebsd.org>, Dimitry Andric <dim@freebsd.org>, Benjamin Kaduk <kaduk@MIT.EDU> Subject: Re: [patch] Cleaning up amd64 kernel optimization options Message-ID: <25FBBF23-CDFA-429E-966D-A90409D8F2BD@gmail.com> In-Reply-To: <20111223005932.GA65042@freebsd.org> References: <4EF34E52.2040905@FreeBSD.org> <20111222192549.GA19008@freebsd.org> <alpine.GSO.1.10.1112221948430.882@multics.mit.edu> <20111223005932.GA65042@freebsd.org>
next in thread | previous in thread | raw e-mail | index | archive | help
On Dec 22, 2011, at 4:59 PM, Alexander Best <arundel@freebsd.org> wrote: > On Thu Dec 22 11, Benjamin Kaduk wrote: >> On Thu, 22 Dec 2011, Alexander Best wrote: >>=20 >>> On Thu Dec 22 11, Dimitry Andric wrote: >>>> Hi, >>>>=20 >>>> I would like to ask some feedback on the attached patch, which cleans u= p >>>> the kernel optimization options for amd64. This was touched upon >>>> earlier by Alexander Best in freebsd-toolchain, here: >>>=20 >>> i've been using such settings for a few months now and haven't noticed a= ny >>> problems. >>>=20 >>> however bruce evans raised a good point (in a private mail). when you=20= >>> compile a >>> kernel without debugging enabled, -O2 is the default. if you experience=20= >>> issues, >>> and enable debugging, -O0 now becomes the default. in case the problems=20= >>> with >>> the kernel were caused by the -O2 option and aren't present with the -O0= >>> option, the newly built kernel with debugging enabled will not help you=20= >>> fix the >>> problems. in that case you would need to set -O2 explicitly in CFLAGS. h= is >>> exact words were: >>>=20 >>> " >>> I don't like -O2 for anything. However, if it is only a default, it is >>> not a problem provided it can be canceled easily. And for debugging, yo= u >>> want the default to be the same as without debugging, so that (by defaul= t) >>> you debug the same code that caused the problem. >>> " >>>=20 >>> however i don't think this is fixable. using -O0 for debuggable and >>> non-debuggable kernels will cause too much of a slowdown. >>>=20 >>> having -O2 as the default flag for non-debuggable kernels and -O2 -g for= >>> debuggable kernels might cause situations, where debugging isn't possibl= e, >>> where with -O0 -g it would have been. >>>=20 >>> so i guess although bruces concerns are valid, they are impossible to=20= >>> solve. >>=20 >> Where does -O0 come in? I only see talk of -O (i.e. -O1) versus -O2. >=20 > sorry. of course i meant -O: >=20 > .if defined(DEBUG) > _MINUS_O=3D -O > CTFFLAGS+=3D -g > .else > [..] Back in the 7.x days, I ran into some code that wasn't easily to debug becau= se the compiler optimized things out with -O2 by inlining and otherwise shif= ting around code, so setting breakpoints in gdb became difficult. So from th= at point on I've gotten into the habit of doing -O explicitly in make.conf i= f DEBUG_FLAGS was specified. Just a thought.. -Garrett=
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?25FBBF23-CDFA-429E-966D-A90409D8F2BD>