This is the mail archive of the
ecos-patches@sources.redhat.com
mailing list for the eCos project.
RE: MPC8260 cache patch
- From: "Patrick Doyle" <wpd at delcomsys dot com>
- To: "Andrew Lunn" <andrew dot lunn at ascom dot ch>
- Cc: "eCos patches" <ecos-patches at sources dot redhat dot com>
- Date: Tue, 25 Mar 2003 09:56:06 -0500
- Subject: RE: MPC8260 cache patch
I assume that there is a 'HAL_DCACHE_SYNC_REGION()' type of macro to flush
just the portion of memory that needs to be synchronized. In the other
direction, I would expect the snooping hardware in the cache to handle
received packets.
--wpd
> -----Original Message-----
> From: Andrew Lunn [mailto:andrew dot lunn at ascom dot ch]
> Sent: Tuesday, March 25, 2003 9:41 AM
> To: Patrick Doyle
> Cc: eCos patches
> Subject: Re: MPC8260 cache patch
>
>
> > As I was thinking about this last night, I started to wonder,
> why does it
> > matter if the sync time is doubled? The effeciency expert in
> my cringes to
> > hear me write this (huh?) but, seriously, how often does the
> cache need to
> > be sync'd in eCos.
>
> Some network devices could be affected. Onces that are memory mapped.
> You need to sync the memory before telling the device to actually do
> the send, otherwise it sends the random junk in DRAM and not the real
> packet which could still be in cache.
>
> I don't know if this is true for the 8260 though.
>
> Andrew