Raspberry Pi 3 status update

Gerd Hoffmann kraxel at redhat.com
Thu Jun 16 00:55:07 PDT 2016


  Hi,

> As far as our status, the Pi 3 is approaching being ready for 4.8.
> Today I've merged the Kconfig changes, so in -next tomorrow you'll be
> able to build an arm64 Pi 3 image by enabling all the BCM2835 bits.  The
> next steps I see that we need for Pi3:

Rebasing my rpi3 branch on your for-next I have these remaining bits:

d94a99d pinctrl: walk into bcm subdir unconditionally

Linus picked up that one.

c0a50e3 arm64: Add BCM2835 support to the defconfig.

Guess I should just send this one.

950c904 ARM: bcm2835: Add devicetree for the Raspberry Pi 3, for arm
(32)

Same here.

0f62ee0 ARM: bcm2385: Don't force pl011 onto pins 14/15.

See below.

0128845 i2c: bcm2835: Set up the rising/falling edge delays.
86a6a6a i2c: bcm2835: Set up the clock stretching timeout at boot.
fe554b4 clk: bcm2835: Mark the VPU clock as critical
0db14ab clk: bcm2835: Mark GPIO clocks enabled at boot as critical
ad775cd clk: bcm2835: Mark the CM SDRAM clock's parent as critical
4fe62f0 clk: bcm2835: Skip PLLC clocks when deciding on a new clock
parent

These are picked up from the mailing list.  Are they one the way
upstream via i2c/clk branches?

> - Build a new arm64 defconfig patch taking into account Arnd's feedback
>   (https://patchwork.kernel.org/patch/9151913/).  Feel free to take
>   authorship on the patch if you do.

I'll send my patch.

> - Update the sdhost driver for Arnd and Rob's feedback
>   (https://lkml.org/lkml/2016/6/1/862)
> 
> - Write a new GPIO driver using the firmware's GET_GPIO_STATE and
>   SET_GPIO_STATE calls.  This won't be able to control direction (not
>   that I think there's anything on the expander that we'd want to set
>   direction for), but it'll give us access to the HDMI HPD pin necessary
>   for graphics.

While talking about gpio: what is the status/plan for all those gpio
device tree changes?  Defining & using the standard groups?  The sdhost
patches have device tree changes depending on them ...

> - Consider "ARM: bcm2385: Don't force pl011 onto pins 14/15" from my
>   bcm2837-64-next branch

Looks reasonable to me ;)

cheers,
  Gerd



More information about the linux-rpi-kernel mailing list