[PATCH 2/2 v4] [BZ #10283] localedef: align fixed maps to SHMLBA

[Mike Frysinger]

On Wednesday 19 June 2013 13:36:58 Roland McGrath wrote:
> > after all, the issue we're talking about here isn't really founded in the
> > Linux port, but from the hardware itself.  it arises due to aliasing
> > issues with hardware memory maps.  so i don't see how the hurd guys
> > could even avoid it if they ever get around to doing a port for
> > sparc/arm/ia64/hppa/etc...
> 
> That's not the point.  The point is that SHMLBA being the way the system
> communicates these requirements is a Linuxism and not a proper GNU API.

we're talking about an internal knob here, not an external API.  we're also 
taking about something that, in practice, makes no difference as i outlined in 
the rest of the previous e-mail.  i understand your pov about trying to keep 
things logically separated, but i don't think this case warrants being super 
conservative -- in practice, every target is going to end up defining it to 
SHMLBA.  the fact that GNU/hurd hasn't yet is a reflection of it not getting 
beyond the x86 cpu, not that they're going to pick a different route.  if it 
ever does grow past x86, it's going to hit the same issues every other OS has, 
and if they have any semblance of sanity, just do what everyone else has -- 
use SHMLBA.  they need to have that properly defined in order for POSIX shared 
memory (shmat) to work which means it'll also work for MAP_FIXED.
-mike
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 836 bytes
Desc: This is a digitally signed message part.
URL: <http://sourceware.org/pipermail/libc-alpha/attachments/20130619/916ad644/attachment.sig>