gcc 3.2.3 config/os/bits/i386/atomicity.h broken

[Andreas Jaeger]

Loren James Rittle <rittle@latour.rsch.comm.mot.com> writes:

>> Sorry, one question: Which file exactly is meant?
>
> Sorry, if I was unclear.  The patch in question was in the thread (you
> found it more-or-less, I recently CC'd on a "final" version).  It was
> not the old 3.2.X-version of config/os/bits/i386/atomicity.h but
> rather this new one that correctly uses instructions actually
> available on i386.
>
> [...]
>> I would prefer to use the i486 implementation and perhaps use this
>> only for i386.  The problem I see here is more performance,
>
> OK, agreed, but the issue is i386 code compiled on an i386 and then
> moved to a higher processor.  OK, someone that explicitly configures
> for i386 on an x86-64.  I am still concerned because of this text:
>
>>> xchgl will work correctly on x86-64.  The only problem (and that's why
>>> it was disabled temporarily in the patch you mentioned) is that
>>> arguments to it might have a different size.
>
> Do I have to do anything special to code against that statement?

No, nothing - if you use int's as arguments.  The problem arises when
using longs since those are different,

Andreas
-- 
 Andreas Jaeger
  SuSE Labs aj@suse.de
   private aj@arthur.inka.de
    http://www.suse.de/~aj