🚀 go-pugleaf

RetroBBS NetNews Server

Inspired by RockSolid Light RIP Retro Guy

Article View: gnu.utils.bug
Article #174

binutils-920602 ld for coff-m68k produces a strange result

#174
From: (Hermann Hueni)
Date: Wed, 22 Jul 1992 10:22
59 lines
2550 bytes
I use gcc2.2.2/gas-1.92.3 for cross-compilation sparc->m68k..
I use ld from the binutils-920602 together with the newest bfd library
from gdb4.6 and confiured with --target=sun3.

Generating  srec output with ld works perfectly.
I can also generate a.out-sunos-big format files and look at them with size
and other tools form the binutils successfully.

	bash$ /usr/local/sun3/bin/size neuro6.aout 
	text       data      bss         dec         hex     filename
	180192  16384   1838900 2035476 1f0f14  neuro6.aout

Since our target system can only load coff format files, I tried to generate
coff-m68k. When I look at the generated file with size, I get:

	bash$ /usr/local/sun3/bin/size neuro6.coff 
	text    data    bss     dec       hex     filename
	0        11552   0        11552   2d20    neuro6.coff

Something must go wrong. I suppose, that the bfd part that implements coff-m68k
is buggy. When I investigate the file with objdump -x, I get the following;

	neuro6.coff:     file format coff-m68k
	neuro6.coff
	architecture: m68k:68020, flags 0x00000032:
	EXEC_P, HAS_SYMS, HAS_LOCALS
	start address 0x04040000
	SECTION 0 [.data]       : size 00002d20 vma 04020000 align 2**3
	 ALLOC, LOAD
	SECTION 1 []    : size 00000002 vma 00000404 align 2**3
	 ALLOC, LOAD
	bfd assertion fail ./coffcode.h:3059
	bfd assertion fail ./coffcode.h:3059
	bfd assertion fail ./coffcode.h:3059

The last message is repeated about 130 times. Thereafter, the symbol table gets dumped.
Here is the link script that I am using:

/* OUTPUT_FORMAT( "a.out-sunos-big" ) */
OUTPUT_FORMAT( "coff-m68k" )
OUTPUT_ARCH( m68k )
__dtof=0x ;
MEMORY
{
        ram: ORIGIN=0x4020000, LENGTH=0x3d0000
}
SECTIONS {
        .data 0x4020000 : { *(.data)}
        .text 0x4040000 : {*(.text) *(code)}
        .bss  0x4080000 : {*(.bss) [COMMON] }  
}


-----------------------------------------------------
Any suggestions what might be wrong ?

thanks,
hermann
hueni@optolab.unibe.ch
Technical Details
Message-ID: <1992Jul22.062207.752@aragorn.unibe.ch>
Path: arc1-ca.pugleaf.net!sparky!uunet!cis.ohio-state.edu!optolab.unibe.ch!hueni

Show Full Headers
Complete Article Headers
Newsgroups: gnu.utils.bug
Path: arc1-ca.pugleaf.net!sparky!uunet!cis.ohio-state.edu!optolab.unibe.ch!hueni
From: hueni@optolab.unibe.ch ( (Hermann Hueni))
Subject: binutils-920602 ld for coff-m68k produces a strange result
Message-ID: <1992Jul22.062207.752@aragorn.unibe.ch>
Keywords: bfd, ld, coff, m68k
Sender: gnulists@ai.mit.edu
Reply-To: hueni@optolab.unibe.ch
Organization: Physiological Institute, University of Berne, Switzerland
Distribution: gnu
Date: Wed, 22 Jul 1992 10:22:07 GMT
Approved: bug-gnu-utils@prep.ai.mit.edu
Lines: 60