]> git.baikalelectronics.ru Git - arm-tf.git/commit
feat(fvp): dts: drop 32-bit .dts files
authorAndre Przywara <andre.przywara@arm.com>
Fri, 19 Aug 2022 09:26:00 +0000 (10:26 +0100)
committerAndre Przywara <andre.przywara@arm.com>
Tue, 11 Oct 2022 15:09:35 +0000 (16:09 +0100)
commitb92033075aa27031091e184b54f4dc278ecb27bc
tree19a00fa44f96ab3987cc19d62ff455349e0dba69
parent08f3c2bcdd452b6ea9e0415f91679d5aa124a4a3
feat(fvp): dts: drop 32-bit .dts files

Conceptually the DT is a hardware description, as such it's independent
from the instruction set that a DT client uses. So having separate DTs
for aarch32 and aarch64 does not make sense and is not needed.

Probably due to historic reasons (a Linux bug fixed in 2016 with Linux
commit ba6dea4f7ced, in Linux v4.8) the CPU reg property was using a
different size between aarch64 and aarch32, even though the size of it
is solely governed by the parent's #address-cells property.

Consolidate this to be always 2, and always use two cells to describe
the CPU's MPIDR register.

This removes the last difference of the -aarch32 versions of the FVP
DT files, so just remove all of them. The respective versions without
that suffix can now be used with AArch32 DT clients as well.

Also remove the respective part in the documentation.

Signed-off-by: Andre Przywara <andre.przywara@arm.com>
Change-Id: I45d3a2cbba8e04595a741e1cf41900377952673e
docs/plat/arm/fvp/index.rst
fdts/fvp-base-gicv2-psci-aarch32.dts [deleted file]
fdts/fvp-base-gicv3-psci-aarch32-1t.dts [deleted file]
fdts/fvp-base-gicv3-psci-aarch32-common.dtsi [deleted file]
fdts/fvp-base-gicv3-psci-aarch32.dts [deleted file]
fdts/fvp-defs.dtsi
fdts/rtsm_ve-motherboard-aarch32.dtsi [deleted file]