*BSD News Article 1770


Return to BSD News archive

Path: sserve!manuel!munnari.oz.au!mips!mips!swrinde!gatech!bloom-beacon!eru.mt.luth.se!lunic!sunic2!mcsun!uknet!axion!rtf.bt.co.uk!duplain
From: duplain@rtf.bt.co.uk (Andy Duplain)
Newsgroups: comp.unix.bsd
Subject: Re: Cross compiling for 386bsd using gcc 2.2.2
Message-ID: <1992Jul1.135130.1127@rtf.bt.co.uk>
Date: 1 Jul 92 13:51:30 GMT
References: <1992Jun29.101909.25381@rtf.bt.co.uk> <1992Jul1.091811.10113@rtf.bt.co.uk> <1992Jul1.120819.23372@rtf.bt.co.uk>
Organization: BT Customer Systems, Brighton, UK
Lines: 21

In article <1992Jul1.120819.23372@rtf.bt.co.uk> duplain@rtf.bt.co.uk (Andy Duplain) writes:
>In article <1992Jul1.091811.10113@rtf.bt.co.uk> duplain@rtf.bt.co.uk (Andy Duplain) writes:
>  
>	As  an  adendum to my last message there seems to be a problem with
>	~/src/lib/csu.i386/gmon.c:    It    needs    the   "extern   mcount
>	asm("mcount")"  declaration  removed  in  order to allow references
>	to _mcount to be resolved... can anyone confirm this ?

	Forget  this  garbage  I wrote earlier! There is nothing wrong with
	the  crt0.o  files  (gcrt0.o  should  be  used  for both -p and -pg
	compilation).  It  seems  that  gcc  makes  calls to _mcount rather
	than  mcount,  and  this can be fixed by hacking config/i386.h. The
	gcc  that  is  distributed  with  Net/2  sets-up the address of the
	label  to  profile  in  %eax  then calls mcount, where as gcc 2.2.2
	sets-up  the  address  in  %edx  and  then  calls _mcount... so I'm
	hacking  gcc  2.2.2.  Did  you  guys  who  set-up gcc 2.2.2 on your
	386BSD systems change this also ?

-- 
Andy Duplain, BT Customer Systems, Brighton, UK.            duplain@rtf.bt.co.uk
#define	DISCLAIMER      My views and opinions are my own, and not my company's