On Mon, 7 Mar 2016 17:18:17 +0100
Boris Brezillon [off-list ref] wrote:
Hi,
This patchset aims at fixing a few minor bugs, and improving performances
of NAND accesses going through the sunxi NAND controller.
Note that patch 5 exports functions provided by the core which are needed
in patch 6 to still support raw OOB accesses. Other patches are just
fixes or improvements only touching the sunxi driver itself.
Here are the main improvements:
- queue CLE/ALE requests instead of forcing the controller to issue each
cmd and address cycle separately
- benefit from ECC correction on protected OOB bytes
- use polling instead of interrupt-based waiting (avoid scheduling
overhead for short wait period)
I'm applying this series since it only touches the sunxi NAND driver,
and I don't expect to have any external reviews :).
--
Boris Brezillon, Free Electrons
Embedded Linux and Kernel engineering
http://free-electrons.com