Re: CVS commit: src

[Matt Thomas <matt%3am-software.com@localhost>]

On Sep 1, 2010, at 12:48 AM, Adam Hoka wrote:

> On Tue, 31 Aug 2010 21:57:27 +0200
> Joerg Sonnenberger <joerg%britannica.bec.de@localhost> wrote:
> 
>> On Tue, Aug 31, 2010 at 08:43:33PM +0400, Valeriy E. Ushakov wrote:
>>> On Tue, Aug 31, 2010 at 16:24:40 +0000, Tom Spindler wrote:
>>> 
>>>>>   src/sys/sys: exec_elf.h
>>>>> 
>>>>> Log Message:
>>>>> Replace the current usage of Elf64_Half with Elf64_Word and rename
>>>>> NetBSD specific Elf64_Quarter to Elf64_Half. This restores compatibility
>>>>> with the common ELF specifications.
>>>> 
>>>> I believe this change is what's causing alpha to blow up:
>>>> 
>>>> /src/nbsrc/external/bsd/libdwarf/dist/dwarf_dump.c:884: warning: format 
>>>> '%u' expects type 'unsigned int', but argument 2 has type 'Elf64_Word'
>>>> /src/nbsrc/external/bsd/libdwarf/dist/dwarf_dump.c:886: warning: format 
>>>> '%u' expects type 'unsigned int', but argument 3 has type 'Elf64_Word'
>>>> 
>>>> There's (still) a ton of lint warnings in libdwarf, but they don't seem to
>>>> have changed for i386, at least.
>>> 
>>> 
>>> sys/arch/alpha/include/elf_machdep.h:
>>> 
>>> /*
>>> * Alpha ELF uses different (non-standard) definitions of Elf64_Sword
>>> * and Elf64_Word.
>>> */
>>> typedef     int64_t         Elf64_Sword;
>>> #define     ELF64_FSZ_SWORD 8
>>> typedef     uint64_t        Elf64_Word;
>>> #define     ELF64_FSZ_WORD  8
>> 
>> That violates the ELF API as specified by SCO again. Is there any reason
>> why Alpha has to be special and can't just use the standard types?
>> 
>> Joerg
> 
> also, why does alpha use non standard platform magic number?

Because alpha's use of ELF predates the assignment of it.