This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: PATCH: Turn on -fomit-frame-pointer by default for 32bit Linux/x86
- From: Mike Stump <mikestump at comcast dot net>
- To: Jack Howarth <howarth at bromo dot med dot uc dot edu>
- Cc: Uros Bizjak <ubizjak at gmail dot com>, "H.J. Lu" <hjl dot tools at gmail dot com>, GCC Patches <gcc-patches at gcc dot gnu dot org>, Richard Guenther <rguenther at suse dot de>, Jakub Jelinek <jakub at redhat dot com>, Mark Mitchell <mark at codesourcery dot com>, Richard Henderson <rth at redhat dot com>, jh at suse dot cz, Andi Kleen <andi at firstfloor dot org>, Andrew Haley <aph at redhat dot com>, David Daney <ddaney at caviumnetworks dot com>, Xinliang David Li <davidxl at google dot com>, Chris Lattner <clattner at apple dot com>, Gerald Pfeifer <gerald at pfeifer dot com>
- Date: Thu, 12 Aug 2010 15:59:42 -0700
- Subject: Re: PATCH: Turn on -fomit-frame-pointer by default for 32bit Linux/x86
- References: <AANLkTinbviPWO-hDfgL=31x=st_LA2frCPPYXtscEAhX@mail.gmail.com> <AANLkTim1L=eJGVO92S8_g+VP1ZUF4-O2O8KBUMB1UcwB@mail.gmail.com> <AANLkTi=7pqkdQjtDbqTSh-TAEa1L3TjBXgzXX2=znS4M@mail.gmail.com> <AANLkTi=FxbGb=sqr-+L0_tHwunFQS+0AqabarkEG10n7@mail.gmail.com> <AANLkTikUX0WLkiBFF+1eE8X44PsKyVNL6GNShPapoGo-@mail.gmail.com> <20100812172246.GB25488@bromo.med.uc.edu> <AANLkTind4=ouDoCMRPuRs2yFXduX=s2vBV65q7ZWaD1A@mail.gmail.com> <20100812173911.GA25792@bromo.med.uc.edu>
On Aug 12, 2010, at 10:39 AM, Jack Howarth wrote:
>> We don't ENABLE it for 32bit MACHO now.
> The only comment I can find in Apple's documentation on this is at...
I'm fine with the configure option forcing the default for darwin, if people want to do that. The option is fine for people that know what they're doing. In fact, I'd argue that it should affect darwin equally.
On darwin, we have CrashReporter that walks the stack but has yet to be enhanced to read dwarf. When combined with system libraries that create new threads for you, and the fact when something crashes, it snaps _all_ stacks from all threads, we wind up wanting async thread stack walking as well. The reports can't be very complete, if people use this option, but, normal developers can never see those reports, so the fact they are less useful, is less interesting to most people. For developers, on the machine they built software for, they will have to live with their decision, we shouldn't interfere with it. Also, Shark (performance monitoring tool suitable for developers) isn't going to be able to walk the stack, but again, that is the developers choice. The general disclaimer could be enhanced to say that CrashReporter and Shark will provide less useful data if this configure option is used, if one wanted. Though, that would be done in a wiki article as well.