- To: Jonathan Wakely <jwakely.gcc@xxxxxxxxx>
- Subject: Re: Side effect of -finline-functions
- From: Alexander Monakov <amonakov@xxxxxxxxx>
- Date: Thu, 3 May 2012 16:58:04 +0400 (MSK)
- Cc: Mathieu Malaterre <mathieu.malaterre@xxxxxxxxx>, gcc-help@xxxxxxxxxxx
- Comment: DKIM? See http://www.dkim.org
- Comment: DomainKeys? See http://antispam.yahoo.com/domainkeys
- In-reply-to: <CAH6eHdSO869G9D=HBHjTad4LMDeLXqCJiMwpJV2mP5+dyNaySQ@mail.gmail.com>
- User-agent: Alpine 2.00 (LNX 1167 2008-08-23)
> The manual says "Not all optimizations are controlled directly by a
> flag. Only optimizations that have a flag are listed in this section."
> So there may be additional optimisations at -O3 beyond those listed.
Yes, a statement like "breaks with -O3, works with -O3 -fno-inline-functions"
would be stronger.
Besides, inlining itself may not be doing a transformation that breaks your
code, but instead open an opportunity for one of later optimization passes.
Does your code break with -O2 -finline-functions? You may want to start with
-O2 -finline-functions (or -O3) and add -fno- options until one of them fixes
your code to possibly identify the breaking transformation more accurately.
Alexander
[Linux C Programming]
[Linux Kernel]
[eCos]
[Fedora Development]
[Fedora Announce]
[Autoconf]
[The DWARVES Debugging Tools]
[Yosemite Campsites]
[Yosemite News]
[Linux GCC]