Hi,
I'm in the process of adding a device tree overlay for the PiFace
Digital Raspberry Pi daughter board [0]. It's an mcp23s17 SPI GPIO port
expander. In doing so, I noticed some errors with the mcp23s08 driver.
This series adds my fixups to (which builds successfully):
v5.8-13252-gcf4a66ae5e4a
But I have only tested them on (rpi-5.4.y):
v5.4.51-1078-g92834e4bb4ce
They're quite trivial and backwards compatible, although I might be
wrong about "interrupt-controller". Can someone please confirm?
Many thanks,
Thomas
[0] https://github.com/raspberrypi/linux/pull/3794
Thomas Preston (3):
pinctrl: mcp23s08: Fixup mcp23x17 regmap_config
pinctrl: mcp23s08: Remove interrupt-controller
devicetree: mcp23s08: Remove interrupt-controller
.../bindings/pinctrl/pinctrl-mcp23s08.txt | 8 -----
drivers/pinctrl/pinctrl-mcp23s08.c | 35 +++++++++----------
drivers/pinctrl/pinctrl-mcp23s08.h | 2 +-
3 files changed, 18 insertions(+), 27 deletions(-)
--
2.26.2
On 14/08/2020 11:03, Thomas Preston wrote:
> I'm in the process of adding a device tree overlay for the PiFace
> Digital Raspberry Pi daughter board [0]. It's an mcp23s17 SPI GPIO port
> expander. In doing so, I noticed some errors with the mcp23s08 driver.
[snip]
> They're quite trivial and backwards compatible, although I might be
> wrong about "interrupt-controller". Can someone please confirm?
[snip]
> [0] https://github.com/raspberrypi/linux/pull/3794
Actually I think I'm wrong about the interrupt-controller changes in
patches 0002 and 0003.
I think Patch 0001 fixups are fine still.
Sorry for the noise!
On Fri, Aug 14, 2020 at 02:56:54PM +0100, Thomas Preston wrote:
> On 14/08/2020 11:03, Thomas Preston wrote:
> > I'm in the process of adding a device tree overlay for the PiFace
> > Digital Raspberry Pi daughter board [0]. It's an mcp23s17 SPI GPIO port
> > expander. In doing so, I noticed some errors with the mcp23s08 driver.
> [snip]
> > They're quite trivial and backwards compatible, although I might be
> > wrong about "interrupt-controller". Can someone please confirm?
> [snip]
> > [0] https://github.com/raspberrypi/linux/pull/3794
>
> Actually I think I'm wrong about the interrupt-controller changes in patches
> 0002 and 0003.
You are. Looking at the datasheet, the GPIOs have interrupt capability.
GPIO controllers are typically both an interrupt client and provider.
Rob
On 17/08/2020 20:29, Rob Herring wrote:
> On Fri, Aug 14, 2020 at 02:56:54PM +0100, Thomas Preston wrote:
>> Actually I think I'm wrong about the interrupt-controller changes in patches
>> 0002 and 0003.
>
> You are. Looking at the datasheet, the GPIOs have interrupt capability.
> GPIO controllers are typically both an interrupt client and provider.
Thanks for the clarification.
I still think the patch 0001 is required. The precious and volatile
ranges are broken.