[PATCH v2 00/10] ARM: mvebu: DT support, SPI, GPIO driver, and Dove DT

[Sascha Hauer]

On Thu, Jul 04, 2013 at 01:40:24PM +0200, Sebastian Hesselbarth wrote:
> On 07/04/13 09:37, Sascha Hauer wrote:
> >On Tue, Jul 02, 2013 at 08:30:39PM +0200, Sebastian Hesselbarth wrote:
> >>This patch set is a first attempt to bring full DT support to Marvell
> >>MVEBU SoCs. It also introduces a driver for the SPI controller found
> >>on these SoCs with special treatment of additional functions for Dove
> >>and Armada 370/XP SoCs. Also a GPIO driver is added that I have written
> >>during OF patches.
> >>
> >>First, current soc_init is moved from postcore initcall to core initcall,
> >>to allow clocks to be accessed early. Clock lookup for timers is converted
> >>to physbase instead of name-based lookup. Then DT support is added for both
> >>clocksource drivers found on Orion and Armada 370/XP SoCs, respectively.
> >>
> >>Also, a DT-only SPI driver is added to allow access to SPI devices found
> >>on boards comprising Marvell SoCs. Then a DT-only GPIO driver compatible
> >>with MVEBU SoCs and device tree parsing for gpio-leds is added.
> >>
> >>The last two patches extend SolidRun CuBox defconfig with a bunch of
> >>options for future drivers and debug commands and import DTS files from
> >>Linux with an additional node for the timer.
> >>
> >>Sebastian Hesselbarth (10):
> >>   ARM: mvebu: move soc_init to core_initcall
> >>   clocksource: orion: lookup clock by physbase
> >>   clocksource: orion: add DT support
> >>   clocksource: mvebu: lookup clock by physbase
> >>   clocksource: mvebu: add DT support
> >>   spi: add Marvell MVEBU SoC SPI driver
> >>   GPIO: add Marvell Orion/MVEBU SoC GPIO driver
> >>   LED: add support for device tree parsing of gpio-leds
> >>   ARM: mvebu: add more options to SolidRun CuBox defconfig
> >>   ARM: mvebu: import DT files for Dove SoC and SolidRun CuBox
> >
> >Applied this series except for the SPI driver.
> 
> Sascha,
> 
> currently mvebu on next is broken because of missing
> "OF: address and device related sync and cleanup" patch set.
> 
> I should have stated clearly that this patch set depends on
> the former set.
> 
> Can you re-test the OF address/device patches and check if
> they still break SPI chip select gpio on efikasb?

I applied the series and based the Marvell patches ontop of it. I
haven't tested on EfiksSB but on another board with a SPI controller, so
I assume this issue is fixed now.

Please note that your of address cleanup series adds about 1.8k to the
binary size. This doesn't sound too much, but such stuff really sums up
over time. I agree that we can't do without proper address translation
in the future, so I applied it. Just be aware in the future that we
still need a good compromise between binary size and sophisticated code.

Sascha


-- 
Pengutronix e.K.                           |                             |
Industrial Linux Solutions                 | http://www.pengutronix.de/  |
Peiner Str. 6-8, 31137 Hildesheim, Germany | Phone: +49-5121-206917-0    |
Amtsgericht Hildesheim, HRA 2686           | Fax:   +49-5121-206917-5555 |