This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
Re: A patch for linux 2.1.127
- To: davem at dm dot cobaltmicro dot com (David S. Miller)
- Subject: Re: A patch for linux 2.1.127
- From: kwrohrer at ce dot mediaone dot net
- Date: Mon, 9 Nov 1998 15:39:14 -0600 (EST)
- Cc: torvalds at transmeta dot com, rth at cygnus dot com, egcs at cygnus dot com, linux-kernel at vger dot rutgers dot edu
- Reply-to: kwrohrer at enteract dot com
And lo, David S. Miller saith unto me:
>
> Date: Sun, 8 Nov 1998 23:21:39 -0800 (PST)
> From: Linus Torvalds <torvalds@transmeta.com>
>
> How about something simple like:
>
> if (SMALL_REGISTER_SET)
> never_inline_functions_unless_the_user_asked_for_it();
>
> which means that even with -O6 you would not inline functions
> unless they were marked inline.
>
> I think I'd rather tell people "don't compile the kernel with -O6"
> than turn off -finline-functions for -O6 by default on any machine.
>
Why not just recommend "-O6 -fno-inline-functions" like we did with the
strength reduction bug? (Insert the right switch to disable the
aggressive inlining of -O6 but not disable explicitly requested
inlining here...and in the compiler code, if it's not there already.)
Maybe even add a "-OK" or "-OL" optimization level for code which has
already been hand-optimized in these ways? Of course, we'd also want
to just use -O6 for code that has not been so heavily optimized vis a
vis inlining on Intel.
Keith
--
"The avalanche has already started; |Linux: http://www.linuxhq.com |"Zooty,
it is too late for the pebbles to |KDE: http://www.kde.org | zoot
vote." Kosh, "Believers", Babylon 5 |Keith: kwrohrer@enteract.com | zoot!"
www.midwinter.com/lurk/lurker.html |http://www.enteract.com/~kwrohrer | --Rebo