Subject: Re: building 2.0-RELEASE for amiga fails
To: Ignatios Souvatzis <ignatios@cs.uni-bonn.de>
From: Hauke Fath <hf@spg.tu-darmstadt.de>
List: netbsd-users
Date: 12/21/2004 14:16:35
Am 21.12.2004 um 13:43 Uhr +0100 schrieb Ignatios Souvatzis:
>On Tue, Dec 21, 2004 at 01:19:06PM +0100, Hauke Fath wrote:
>>  Am 21.12.2004 um 12:40 Uhr +0100 schrieb Ignatios Souvatzis:
>>  >Thou shalt not meddle with the kernel optimization settings!
>>  >(with the exception of using options M68060 etc., which add the appropriate
>>  > optimizations to the kernel building.)
>>
>>  Then this is new with NetBSD 2 / gcc 3.x. I've been running 1.5
>>  kernel & userland with '-O3 -m68030' for a long time now.
>
>This will work, provided you're using the right assembler CPU type
>calls and inserting them in the right make variables. A lot of the
>people who try, don't, and at some point I got tired of explaining
>over and over again... in general I assume that whoever can't read
>the makefiles himself, shouldn't bother.

That may well be, but I wasn't talking about the problem in front of 
the keyboard.

Setting (any) global CFLAGS/COPTS breaks 2.0 builds because some 
subsystems want -O or -Os or whatever but do not enforce it. OpenSSL 
was an example a few months ago. Kernels for my sparcstation 10 or my 
Quadra 700 built with -O3 blow up.

gcc 2.95.x didn't do that.

	hauke

-- 
/~\  The ASCII Ribbon Campaign                    Hauke Fath
\ /    No HTML/RTF in email	        Institut für Nachrichtentechnik
  X     No Word docs in email	                  TU Darmstadt
/ \  Respect for open standards              Ruf +49-6151-16-3281