Re: [PATCH AUTOSEL 5.8 20/20] riscv: Fix Kendryte K210 device tree

From: Palmer Dabbelt
Date: Sat Sep 26 2020 - 15:48:45 EST


On Tue, 22 Sep 2020 17:27:42 PDT (-0700), Damien Le Moal wrote:
On 2020/09/21 23:41, Sasha Levin wrote:
From: Damien Le Moal <damien.lemoal@xxxxxxx>

[ Upstream commit f025d9d9934b84cd03b7796072d10686029c408e ]

The Kendryte K210 SoC CLINT is compatible with Sifive clint v0
(sifive,clint0). Fix the Kendryte K210 device tree clint entry to be
inline with the sifive timer definition documented in
Documentation/devicetree/bindings/timer/sifive,clint.yaml.
The device tree clint entry is renamed similarly to u-boot device tree
definition to improve compatibility with u-boot defined device tree.
To ensure correct initialization, the interrup-cells attribute is added
and the interrupt-extended attribute definition fixed.

This fixes boot failures with Kendryte K210 SoC boards.

Note that the clock referenced is kept as K210_CLK_ACLK, which does not
necessarilly match the clint MTIME increment rate. This however does not
seem to cause any problem for now.

Signed-off-by: Damien Le Moal <damien.lemoal@xxxxxxx>
Signed-off-by: Palmer Dabbelt <palmerdabbelt@xxxxxxxxxx>
Signed-off-by: Sasha Levin <sashal@xxxxxxxxxx>
---
arch/riscv/boot/dts/kendryte/k210.dtsi | 6 ++++--
1 file changed, 4 insertions(+), 2 deletions(-)

diff --git a/arch/riscv/boot/dts/kendryte/k210.dtsi b/arch/riscv/boot/dts/kendryte/k210.dtsi
index c1df56ccb8d55..d2d0ff6456325 100644
--- a/arch/riscv/boot/dts/kendryte/k210.dtsi
+++ b/arch/riscv/boot/dts/kendryte/k210.dtsi
@@ -95,10 +95,12 @@ sysctl: sysctl@50440000 {
#clock-cells = <1>;
};
- clint0: interrupt-controller@2000000 {
+ clint0: clint@2000000 {
+ #interrupt-cells = <1>;
compatible = "riscv,clint0";
reg = <0x2000000 0xC000>;
- interrupts-extended = <&cpu0_intc 3>, <&cpu1_intc 3>;
+ interrupts-extended = <&cpu0_intc 3 &cpu0_intc 7
+ &cpu1_intc 3 &cpu1_intc 7>;
clocks = <&sysctl K210_CLK_ACLK>;
};


Sasha,

This is a fix for a problem in 5.9 tree. 5.8 kernel is fine without this patch.
And I think applying it to 5.8 might actually break things since the proper
clint driver was added to kernel 5.9 and does not exist in 5.8.

IIUC this won't actually break anything on 5.8, as the reason nobody noticed
that the old one was broken is because the old CLINT driver just didn't care
about what's in the device tree. These interrupt numbers are defined by the
ISA manual so we jut had them encoded into the arch/riscv first-level interrupt
controller driver.

That said, it definately doesn't fix anything so it seems safer to just not
backport it.