Re: sparc64 vs. gcc 4.8

[Erik Fair <fair%netbsd.org@localhost>]

On Oct 3, 2014, at 09:15 , Eduardo Horvath <eeh%NetBSD.org@localhost> wrote:

> On Fri, 3 Oct 2014, Erik Fair wrote:
> 
>> The object files that differ from the same source but different compiler versions would, one hopes, constrain the bug search space.
> 
> That's probably not a good technique.  Most of the time object files have 
> timestamps in them so they never compare properly.

This suggests a need for "elfcmp" or "objcmp" or equivalent, to compare like ELF sections (text, data, bss, symbols) in object files against each other, while excluding sections that would always be different (e.g. notes containing time stamps, compiler versions).

However, if sparc64's memory barrier implementation has issues, that's certainly fundamental enough to require immediate attention. Eduardo, can you be more specific about what you thought was wrong with the existing implementation?

	Erik <fair%netbsd.org@localhost>