On 4/26/24 01:31, Krzysztof Kozlowski wrote:
On 25/04/2024 23:36, Eddie James wrote:Add the P11 FSI device tree for use in upcoming BMC systems. Unlike P10, there is no system with only two processors, so only the quad processor FSI layout is necessary. Signed-off-by: Eddie James <eajames@xxxxxxxxxxxxx> --- .../arm/boot/dts/aspeed/ibm-power11-quad.dtsi | 1696 +++++++++++++++++ 1 file changed, 1696 insertions(+) create mode 100644 arch/arm/boot/dts/aspeed/ibm-power11-quad.dtsi diff --git a/arch/arm/boot/dts/aspeed/ibm-power11-quad.dtsi b/arch/arm/boot/dts/aspeed/ibm-power11-quad.dtsi new file mode 100644 index 000000000000..c3a0ecf12aa0 --- /dev/null +++ b/arch/arm/boot/dts/aspeed/ibm-power11-quad.dtsi @@ -0,0 +1,1696 @@ +// SPDX-License-Identifier: GPL-2.0-or-later +// Copyright 2024 IBM Corp. + +&fsim0 {This does not make sense. You do not include any file here, so what do you want to override? How can you even test this file?
This is an include file, to be included in the new device tree files in the next two patches. It will be tested as part of those. Andrew requested I split this up, and I have to add this one first, even though nothing is referencing it yet. The same model is used for the P10 FSI devices.
+ status = "okay"; + + #address-cells = <2>; + #size-cells = <0>; + + cfam-reset-gpios = <&gpio0 ASPEED_GPIO(Q, 0) GPIO_ACTIVE_HIGH>; ++ +&cfam3_i2c16 { + fsi-i2cr@20 { + compatible = "ibm,i2cr-fsi-master"; + reg = <0x20>; + #address-cells = <2>; + #size-cells = <0>; + + cfam@0,0 { + reg = <0 0>; + #address-cells = <1>; + #size-cells = <1>; + chip-id = <0>; + + scom416: scom@1000 { + compatible = "ibm,i2cr-scom"; + reg = <0x1000 0x400>; + }; + + sbefifo416: sbefifo@2400 { + compatible = "ibm,odyssey-sbefifo"; + reg = <0x2400 0x400>; + #address-cells = <1>; + #size-cells = <0>; + }; + }; + }; +}; + +&cfam3_i2c17 {This looks randomly ordered.
Not sure what you mean. Everything is sequentially ordered?
+&fsi_occ1 { + reg = <2>; +}; + +&fsi_occ2 { + reg = <3>; +}; + +&fsi_occ3 { + reg = <4>; +}; + +/ {Nope. Root node never goes to end of file. Look at all modern DTS.
Ack.
+ aliases { + i2c100 = &cfam0_i2c0; + i2c101 = &cfam0_i2c1; + i2c110 = &cfam0_i2c10; + i2c111 = &cfam0_i2c11; + i2c112 = &cfam0_i2c12; + i2c113 = &cfam0_i2c13; + i2c114 = &cfam0_i2c14; + i2c115 = &cfam0_i2c15; + i2c202 = &cfam1_i2c2; + i2c203 = &cfam1_i2c3; + i2c210 = &cfam1_i2c10;Best regards, Krzysztof
