Assembler warning: Warning: .space repeat count is zero, ignored

[Andrew Pinski]

That is what I am doing right now:
#0  assemble_zeros (size=4294967294) at 
/Volumes/UFS_Partition/pinskia/src/fsf/gcc/src/gcc/varasm.c:1235
#1  0x000583a0 in output_constant (exp=0x109fc30, size=4294967294, 
align=16) at 
/Volumes/UFS_Partition/pinskia/src/fsf/gcc/src/gcc/varasm.c:3918

(gdb) p (long)size
$2 = -2

Looks like someone said size is unsigned when it really is signed (can 
be less than 0).

 From gcc/varasm.c:
1234        if (size > 0)
1235          ASM_OUTPUT_SKIP (asm_out_file, size);

Thanks,
Andrew Pinski


On Monday, May 12, 2003, at 14:46 US/Eastern, Jan Hubicka wrote:

>> On hppa64-hp-hpux11.00 with gas style assembler output, I see
>> the warnings are coming from lines like:
>>
>> 	.zero   18446744073709551614
>>
>> This is a rather large skip!
>>
>> hppa64-hp-hpux11.00 is using the define for ASM_OUTPUT_SKIP from 
>> elfos.h.
>
> That is what puzzles me.  The definition:
> #define ASM_OUTPUT_SKIP(FILE, SIZE) \
>    fprintf ((FILE), "%s"HOST_WIDE_INT_PRINT_UNSIGNED"\n",\
>                SKIP_ASM_OP, (SIZE))
> Really should match the uses.  Perhaps I messed up the tree so 
> something
> didn't get commited as for me it works.  Or perhaps something 
> underflows
> because of unsigned math...  Can you drop this into debugger and figure
> out from where the wrong value comes?
>
> Honza
>>
>> Dave
>> -- 
>> J. David Anglin                                  
>> dave.anglin@nrc-cnrc.gc.ca
>> National Research Council of Canada              (613) 990-0752 (FAX: 
>> 952-6602)
>
>