hardware cached reads

David Woodhouse dwmw2 at infradead.org
Sat Aug 23 02:10:37 EDT 2003


On Sat, 2003-08-23 at 03:04, Jared Hulbert wrote:
> I noticed the new map_info.cached in
> /include/linux/mtd/map.h is only used in
> /drivers/mtd/maps/ocelot.c.  I'm assuming this is just
> incomplete.  It would seem the goal is to provide a
> dual mapping of the flash to uncached and cached
> regions.  
> 
> Am I reading this correctly?

Yes. Note that the Ocelot map driver used to handle flushing the cache
-- but did it far too often. The chip drivers need to learn when to do
it.

> Any tips, warnings, parallel efforts, flames, jokes?

We've discussed it briefly before, haven't we? Used cached mapping only
for data reads, flush it from the chip driver (probably with a new
map_info method for flushing the cache) upon write or erase.

Do remember that i386en get very unhappy if you ever map the same region
with both a cached and uncached mapping.

We may also want to investigate putting burst-capable chips (and
chipsets) into burst mode for ourselves.

-- 
dwmw2





More information about the linux-mtd mailing list