On 09/10/2013 09:48 AM, Sascha Hauer wrote:
On Mon, Sep 09, 2013 at 06:28:48PM +0200, Sebastian Hesselbarth wrote:
I have seen this driver and think it can be improved by exploiting
its config1/2 registers where synthesis settings are stored. Please
also look at [1] pp.1229; you can read out number of ports, port
width, interrupt, and debounce features.
Ok, the port width feature is easy enough to implement. Will add it in
the next version.
I also added a check for multiple ports, but just ignore the additional
ports for now since I don't know the register offsets and also have no
hardware to actually test it (The SoCFPGA only has one port)
Ok, fair enough. IIRC the Armada 1500 uses multiple instances instead
of multiple ports, too.
Unfortunately, I am not ready to run barebox on Armada 1500 and I
don't have a socfpga to test. So if you are not eager to modify it now,
I can add config reg based read-out later.
[1] http://www.altera.com/literature/hb/arria-v/hps.pdf
[...]
+static __maybe_unused struct of_device_id dwgpio_match[] = {
+ {
+ .compatible = "snps,dw-gpio",
+ }, {
Please add "snps,dw-apb-gpio", that's what the linux driver
will look for if make to pick it up for Armada 1500.
I'll better just change it to "snps,dw-apb-gpio". There already is a
dtsi file in the kernel which uses this binding (although it looks like
it hasn't been approved by anyone)
The driver got Acks, but must have slipped through. I have seen your
v2 but it still uses "snps,dw-gpio".
Sebastian
_______________________________________________
barebox mailing list
barebox@xxxxxxxxxxxxxxxxxxx
http://lists.infradead.org/mailman/listinfo/barebox
