Re: Bug in compiling

[Keith Owens]

On Wed, 3 Apr 2002 21:34:59 -0800 (PST), 
Robin Johnson <[e-mail]> wrote:
>On Wed, 3 Apr 2002, Keith Owens wrote:
>> On Wed, 3 Apr 2002 00:57:06 -0800 (PST),
>> Robin Johnson <[e-mail]> wrote:
>> >gcc -D__KERNEL__ -I/usr/src/linux-2.4.19-pre4-ac3/include -Wall
>> >-Wstrict-prototypes -Wno-trigraphs -O2 -fomit-frame-pointer
>> >-fno-strict-aliasing -fno-common -pipe -mpreferred-stack-boundary=2
>> >-march=i686   -DKBUILD_BASENAME=exec_domain  -DEXPORT_SYMTAB -c
>> >exec_domain.c
>> >exec_domain.c:234: parse error before `register_exec_domain'
>> >exec_domain.c:235: parse error before `unregister_exec_domain'
>> >exec_domain.c:236: parse error before `__set_personality'
>> >exec_domain.c:287: parse error before `abi_defhandler_coff'
>> >...
>>
>> All EXPORT_SYMBOL.  You would get that behaviour if gcc did not
>> recognise EXPORT_SYMBOL as a macro.  Probably random data corruption.
>Of the pair of machines, I upgraded one of them to GCC 3.0.4, while
>leaving the other at 2.95.3.

On the failing system,

cd /usr/src/linux-2.4.19-pre4-ac3/kernel
gcc -D__KERNEL__ -I/usr/src/linux-2.4.19-pre4-ac3/include -Wall \
        -Wstrict-prototypes -Wno-trigraphs -O2 -fomit-frame-pointer \
        -fno-strict-aliasing -fno-common -pipe \
        -mpreferred-stack-boundary=2 -march=i686 \
        -DKBUILD_BASENAME=exec_domain -DEXPORT_SYMTAB \
	-E -c exec_domain.c -o exec_domain.i

and send exec_domain.i to [e-mail] (not the list).