[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]

Re: [PATCH] couple of small bugs



lwsync looks exactly like a sync except for an extra bit being set.  Older
hardware including PPC32 hardware will execute it as a normal sync
instruction. In other words the instruction is backward compatible.

For reference:
http://sources.redhat.com/ml/binutils/2000-12/msg00311.html
http://www-1.ibm.com/servers/esdd/articles/power4_mem.html

We can do some explicit testing on 32 bit hardware, but I believe it
is the correct solution.

On Tue, 29 Jul 2003, Steve Munroe wrote:

> Date: Tue, 29 Jul 2003 14:15:32 -0500
> From: Steve Munroe <sjmunroe us ibm com>
> To: Joel Schopp <jschopp austin ibm com>
> Cc: phil-list redhat com
> Subject: Re: [PATCH] couple of small bugs
>
> Joel Schopp writes:
>
> > /* lwsync is exactly like sync for all pre Power4 architectures
> > * for Power4 and greater it is a lot like sync but different
> > * enough to have much written about the differences.
>
> Joel this file is common to PPC32/PPC64 and there is no guarantee that
> this is true for PPC32 hardware. So until we can verify that lwsync will
> behave correctly on existing powerMacs we need to be careful about which
> code we use it in.
>
> So we need to hold onto this patch until we can discuss.
>




[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]