fix interrupts on the devices hanging off the SPMI bus. The interrupts
one should head into upstream as well at some point, for the touchpad one
I'm not so sure.
ok tobhe@
Includes a revert of the changes in linux 5.15 that switched the cpsw(4)
bindings to a different format. This allows us to update the dtb port
without breaking cpsw(4) on the Beaglebone Black. If we ever adjust
cpsw(4) to work with the new bindings, this could be dropped.
ok tobhe@
Updating this port has a high chance of breaking drivers as breaking
changes are frequently made to device trees in linux. For example the
description of BeagleBone Black hardware changed in linux 5.15 to be
compatible with a new cpsw driver in linux. This would break cpsw(4) if
the port were updated to 5.15.
No need for the entire linux source when we only need the
device-tree sources; this cuts WRKSRC from ~1G to ~62M.
While here, simply build target code a bit.
OK jsg
adds EMAC Ethernet to Allwinner A64/H5/H3
changes Allwinner A10/A20 clocks to CCU bindings
fixes address of i2c2 register for Allwinner H5/H3
adds dtbs for Wandboard rev D1
Switch from using ports gcc to clang. Base gcc can't be used as the
preprocessor doesn't handle '#' not being in the first column.
Problem in initial patch spotted by and ok sthen@
This adds "allwinner,sun8i-h3-ccu" to existing allwinner h3 dtbs that
H3 support in -current requires.
And notably adds new dtbs for:
imx6q-utilite-pro
sun7i-a20-bananapi-m1-plus
sun8i-h3-bananapi-m2-plus
bcm2837-rpi-3-b
Tested on cubox, bbb and panda.
falls back to /aliases/serial0.
Add patches to set stdout-path for dts files that don't set stdout-path
and are known to use a serial device other than serial0 as a console.
Confirmed to restore console output on pandaboard es by abieber@
ok sthen@ naddy@
ARM/MIPS/PowerPC boards. DTBs are binary representations of Flattened
Device Trees (FDTs) that describe hardware properties such as memory
ranges, interrupts and processors.
DTB files are used by U-Boot and other firmware interfaces to provide
this information to the kernel.
Feedback from jca@ and sthen@. ok sthen@
