arm64: Add support for Cavium Thunder ARMv8 platform
v2 cover arm64: Add support for Cavium Thunder ARMv8 platform
From: Radha Mohan Chintakuntla <[email protected]>
This patch adds preliminary support for Cavium's Thunder SoC which is based on ARMv8 core.
The Thunder platform is full GICv3 compliant and uses 48-bit Physical Addresses.
This patch depends on the following submitted patches
1) GICv3 driver from Marc Zyngier
[PATCH 01/18] arm64: initial support for GICv3
http://www.spinics.net/lists/arm-kernel/msg305027.html
2) GICv3 devicetree bindings
[PATCH 02/18] arm64: GICv3 device tree binding documentation
http://www.spinics.net/lists/arm-kernel/msg305023.html
3) Boot time configuration of Intermediate Physical Address size [RFC PATCH 1/1] arm64: Add boot time configuration of Intermediate Physical Address size
http://www.spinics.net/lists/arm-kernel/msg312370.html
Changes from v1:
- Removed Kconfig option for Thunder.
Radha Mohan Chintakuntla (1):
arm64: Add initial dts for Cavium Thunder SoC
arch/arm64/boot/dts/Makefile | 1 + arch/arm64/boot/dts/thunder.dts | 158 +++++++++++++++++++++++++++++++++++++++ 2 files changed, 159 insertions(+), 0 deletions(-) create mode 100644 arch/arm64/boot/dts/thunder.dts
PATCH v2 1/1 arm64: Add initial dts for Cavium Thunder SoC
From: Radha Mohan Chintakuntla <[email protected]>
--- arch/arm64/boot/dts/Makefile | 1 + arch/arm64/boot/dts/thunder.dts | 158 +++++++++++++++++++++++++++++++++++++++ 2 files changed, 159 insertions(+), 0 deletions(-)
arch/arm64/boot/dts/Makefile +1 -0
@@ -1,5 +1,6 @@ dtb-$(CONFIG_ARCH_VEXPRESS) += rtsm_ve-aemv8a.dtb foundation-v8.dtb dtb-$(CONFIG_ARCH_XGENE) += apm-mustang.dtb +dtb-$(CONFIG_ARCH_THUNDER) += thunder.dtb targets += dtbs targets += $(dtb-y)
arch/arm64/boot/dts/thunder.dts new file +158 -0
158 lines not on this page · raw message
Signature · 1 line
1.7.1
PATCH v2 1/1 arm64: Add initial dts for Cavium Thunder SoC
Quoted 13 lines · On Mar 3, 2014, at 1:19 AM, mohun106 at gmail.com wrote:
From: Radha Mohan Chintakuntla <[email protected]>
Signed-off-by Radha Mohan Chintakuntla <[email protected]>---
arch/arm64/boot/dts/Makefile| 1 +arch/arm64/boot/dts/thunder.dts| 158 +++++++++++++++++++++++++++++++++++++++2 files changed, 159 insertions(+), 0 deletions(-)arch/arm64/boot/dts/Makefile +0
-0@@ -1,5 +1,6 @@
Quoted 3 lines
dtb-$(CONFIG_ARCH_VEXPRESS) += rtsm_ve-aemv8a.dtb foundation-v8.dtb dtb-$(CONFIG_ARCH_XGENE) += apm-mustang.dtb +dtb-$(CONFIG_ARCH_THUNDER) += thunder.dtb
Where is the Kconfig associated with CONFIG_ARCH_THUNDER? Also, I know on arch/arm we are moving to CONFIG_SOC_ instead of CONFIG_ARCH_, I assume the same is true for arm64.
- k
Signature · 2 lines
Employee of Qualcomm Innovation Center, Inc. Qualcomm Innovation Center, Inc. is a member of Code Aurora Forum, hosted by The Linux Foundation
PATCH v2 1/1 arm64: Add initial dts for Cavium Thunder SoC
Quoted 11 lines · Kumar Gala
Quoted 8 lines · On Mar 3, 2014, at 1:19 AM, mohun106 at gmail.com wrote:
From: Radha Mohan Chintakuntla <[email protected]>
Signed-off-by Radha Mohan Chintakuntla <[email protected]>---
arch/arm64/boot/dts/Makefile| 1 +arch/arm64/boot/dts/thunder.dts| 158 +++++++++++++++++++++++++++++++++++++++2 files changed, 159 insertions(+), 0 deletions(-)
Quoted 10 lines
Quoted 8 lines
arch/arm64/boot/dts/Makefile +0
-0@@ -1,5 +1,6 @@dtb-$(CONFIG_ARCH_VEXPRESS) += rtsm_ve-aemv8a.dtb foundation-v8.dtb dtb-$(CONFIG_ARCH_XGENE) += apm-mustang.dtb +dtb-$(CONFIG_ARCH_THUNDER) += thunder.dtb
Where is the
Kconfigassociated withCONFIG_ARCH_THUNDER? Also, I know onarch/armwe are moving toCONFIG_SOC_instead ofCONFIG_ARCH_, I assume the same is true for arm64.
Yes, I deliberately removed it as I thought we are trying to get rid of CONFIG_ARCH_ and have a single image boot on all platforms.
Quoted 7 lines
- k
--
Employee of Qualcomm Innovation Center, Inc.
Qualcomm Innovation Center, Inc. is a member of Code Aurora Forum, hosted by The Linux Foundation
PATCH v2 1/1 arm64: Add initial dts for Cavium Thunder SoC
In reply to mohun106 at gmail.comQuoted 25 lines · On Mon, Mar 3, 2014 at 1:19 AM, <[email protected]> wrote:
From: Radha Mohan Chintakuntla <[email protected]>
Signed-off-by Radha Mohan Chintakuntla <[email protected]>--- arch/arm64/boot/dts/Makefile | 1 + arch/arm64/boot/dts/thunder.dts | 158 +++++++++++++++++++++++++++++++++++++++ 2 files changed, 159 insertions(+), 0 deletions(-)arch/arm64/boot/dts/Makefile +1
-0@@ -1,5 +1,6 @@ dtb-$(CONFIG_ARCH_VEXPRESS) += rtsm_ve-aemv8a.dtb foundation-v8.dtb dtb-$(CONFIG_ARCH_XGENE) += apm-mustang.dtb +dtb-$(CONFIG_ARCH_THUNDER) += thunder.dtb targets += dtbs targets += $(dtb-y)arch/arm64/boot/dts/thunder.dts new file +0
-0
arch/arm64/boot/dts/thunder.dts line 9
Quoted 9 lines
arch/arm64/boot/dts/thunder.dts new file +9
-0@@ -0,0 +1,158 @@ +/* + * Cavium Thunder 8xxx DTS file + * + * Copyright (C) 2013, Cavium Inc. + * + * This program is free software; you can redistribute it and/or + * modify it under the terms of the GNU General Public License as + * published by the Free Software Foundation; either version 2 of + * the License, or (at your option) any later version.
DTS files should really be BSD licensed for use on other OSs.
arch/arm64/boot/dts/thunder.dts line 13
Quoted 4 lines
arch/arm64/boot/dts/thunder.dts new file +4
-0+ */ +/dts-v1/; +/ { + model="Thunder";
whitespace around the =.
arch/arm64/boot/dts/thunder.dts line 14
Quoted 1 line
arch/arm64/boot/dts/thunder.dts new file +1
-0+ compatible = "cavium, thunder";
Please document compatible strings. There should not be a space in the string. This should be more specific. The comments refer to 8xxx so I assume there is a part number.
Quoted 5 lines
arch/arm64/boot/dts/thunder.dts new file +5
-0+ interrupt-parent = <&gic0>; + #address-cells = <2>; + #size-cells = <2>; + + aliases {
arch/arm64/boot/dts/thunder.dts line 29
Quoted 10 lines
arch/arm64/boot/dts/thunder.dts new file +10
-0+ serial0 = &uaa0; + serial1 = &uaa1; + }; + + cpus { + #address-cells = <2>; + #size-cells = <0>; + cpu at 0 { + device_type = "cpu"; + compatible = "arm,armv8";
Custom core right? Add a compatible string for the core name.
arch/arm64/boot/dts/thunder.dts line 39
Quoted 10 lines
arch/arm64/boot/dts/thunder.dts new file +10
-0+ reg = <0x0 0x0>; + enable-method = "spin-table"; + cpu-release-addr = <0x0 0xfff8>; + }; + cpu at 1 { + device_type = "cpu"; + compatible = "arm,armv8"; + reg = <0x0 0x1>; + enable-method = "spin-table"; + cpu-release-addr = <0x0 0xfff8>;
Same address for all cores? I don't see how that will work for hotplug.
Quoted 36 lines
arch/arm64/boot/dts/thunder.dts new file +36
-0+ }; + cpu at 2 { + device_type = "cpu"; + compatible = "arm,armv8"; + reg = <0x0 0x2>; + enable-method = "spin-table"; + cpu-release-addr = <0x0 0xfff8>; + }; + cpu at 3 { + device_type = "cpu"; + compatible = "arm,armv8"; + reg = <0x0 0x3>; + enable-method = "spin-table"; + cpu-release-addr = <0x0 0xfff8>; + }; + cpu at 4 { + device_type = "cpu"; + compatible = "arm,armv8"; + reg = <0x0 0x4>; + enable-method = "spin-table"; + cpu-release-addr = <0x0 0xfff8>; + }; + cpu at 5 { + device_type = "cpu"; + compatible = "arm,armv8"; + reg = <0x0 0x5>; + enable-method = "spin-table"; + cpu-release-addr = <0x0 0xfff8>; + }; + cpu at 6 { + device_type = "cpu"; + compatible = "arm,armv8"; + reg = <0x0 0x6>; + enable-method = "spin-table"; + cpu-release-addr = <0x0 0xfff8>; + };
arch/arm64/boot/dts/thunder.dts line 85
Quoted 10 lines
arch/arm64/boot/dts/thunder.dts new file +10
-0+ cpu at 7 { + device_type = "cpu"; + compatible = "arm,armv8"; + reg = <0x0 0x7>; + enable-method = "spin-table"; + cpu-release-addr = <0x0 0xfff8>; + }; + + }; + memory at 0x0 {
Drop the '0x'
arch/arm64/boot/dts/thunder.dts line 90
Quoted 5 lines
arch/arm64/boot/dts/thunder.dts new file +5
-0+ device_type = "memory"; + reg = <0x0 0x0 0x0 0x80000000>; + }; + + gic0: interrupt-controller at 0x801000000000 {
Ditto
arch/arm64/boot/dts/thunder.dts line 96
Quoted 6 lines
arch/arm64/boot/dts/thunder.dts new file +6
-0+ compatible = "arm,gic-v3"; + #interrupt-cells = <3>; + #address-cells = <0>; + interrupt-controller; + reg = <0x8010 0x0 0x0 0x10000>, // GICD + <0x8010 0x80000000 0x0 0x100000>; // GICR
Umm, so no virtualization support?
This should be under a bus node.
Quoted 10 lines
arch/arm64/boot/dts/thunder.dts new file +10
-0+ interrupts = <1 9 0xf04>; + }; + + timer { + compatible = "arm,armv8-timer"; + interrupts = <1 13 0xff01>, + <1 14 0xff01>, + <1 11 0xff01>, + <1 10 0xff01>; + };
arch/arm64/boot/dts/thunder.dts line 116
Quoted 10 lines
arch/arm64/boot/dts/thunder.dts new file +10
-0+ + on-chip-devices { + compatible = "simple-bus"; + #address-cells = <2>; + #size-cells = <2>; + ranges; + + ahci0: host-bus-adapter at 810000000000 { + compatible = "snps,spear-ahci"; + reg-io-width = <4>;
This property is not used for this binding.
Quoted 4 lines
arch/arm64/boot/dts/thunder.dts new file +4
-0+ reg = <0x8100 0x0 0x0 0x1100>; + interrupts = <1 32 4>; + }; +
arch/arm64/boot/dts/thunder.dts line 130
Quoted 10 lines
arch/arm64/boot/dts/thunder.dts new file +10
-0+ nic0: ethernet at 843000000000 { + compatible = "smsc,lan9115"; + reg-io-width = <4>; + reg = <0x8430 0x0 0x0 0x1000>; + interrupts = <1 31 4>; + }; + }; + + amba { + compatible = "arm,amba-bus";
should be "simple-bus"
arch/arm64/boot/dts/thunder.dts line 135
Quoted 5 lines
arch/arm64/boot/dts/thunder.dts new file +5
-0+ #address-cells = <2>; + #size-cells = <2>; + ranges; + + refclk24mhz: refclk24mhz {
move to top level or a clocks node.
arch/arm64/boot/dts/thunder.dts line 142
Quoted 7 lines
arch/arm64/boot/dts/thunder.dts new file +7
-0+ compatible = "fixed-clock"; + #clock-cells = <0>; + clock-frequency = <24000000>; + clock-output-names = "refclk24mhz"; + }; + + uaa0: uart at 87e024000000 {
use ranges so the bus address is 0x87e0_00000000.
arch/arm64/boot/dts/thunder.dts line 147
Quoted 5 lines
arch/arm64/boot/dts/thunder.dts new file +5
-0+ compatible = "arm,pl011", "arm,primecell"; + reg = <0x87e0 0x24000000 0x0 0x1000>; + interrupts = <1 21 4>; + clocks = <&refclk24mhz>; + clock-names = "apb_pclk";
Your APB bus frequency is really only 24MHz? There should 2 clocks defined here.
Quoted 19 lines
arch/arm64/boot/dts/thunder.dts new file +11
-1+ }; + + uaa1: uart at 87e025000000 { + compatible = "arm,pl011", "arm,primecell"; + reg = <0x87e0 0x25000000 0x0 0x1000>; + interrupts = <1 22 4>; + clocks = <&refclk24mhz>; + clock-names = "apb_pclk"; + }; + }; +}; --1.7.1
_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel at lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
PATCH v2 1/1 arm64: Add initial dts for Cavium Thunder SoC
Quoted 29 lines · Rob Herring
Quoted 28 lines · On Mon, Mar 3, 2014 at 1:19 AM, <[email protected]> wrote:
From: Radha Mohan Chintakuntla <[email protected]>
Signed-off-by Radha Mohan Chintakuntla <[email protected]>--- arch/arm64/boot/dts/Makefile | 1 + arch/arm64/boot/dts/thunder.dts | 158 +++++++++++++++++++++++++++++++++++++++ 2 files changed, 159 insertions(+), 0 deletions(-)arch/arm64/boot/dts/Makefile +1
-0@@ -1,5 +1,6 @@ dtb-$(CONFIG_ARCH_VEXPRESS) += rtsm_ve-aemv8a.dtb foundation-v8.dtb dtb-$(CONFIG_ARCH_XGENE) += apm-mustang.dtb +dtb-$(CONFIG_ARCH_THUNDER) += thunder.dtb targets += dtbs targets += $(dtb-y)arch/arm64/boot/dts/thunder.dts new file +3
-0@@ -0,0 +1,158 @@ +/* + * Cavium Thunder 8xxx DTS file + *
Quoted 9 lines
arch/arm64/boot/dts/thunder.dts line 9
Quoted 6 lines
arch/arm64/boot/dts/thunder.dts new file +6
-0+ * Copyright (C) 2013, Cavium Inc. + * + * This program is free software; you can redistribute it and/or + * modify it under the terms of the GNU General Public License as + * published by the Free Software Foundation; either version 2 of + * the License, or (at your option) any later version.DTS files should really be BSD licensed for use on other OSs.
Is this a new requirement? I haven't seen any DTS (atleast in ARM/ARM64) that uses BSD license.
Quoted 7 lines
arch/arm64/boot/dts/thunder.dts line 13
Quoted 4 lines
arch/arm64/boot/dts/thunder.dts new file +4
-0+ */ +/dts-v1/; +/ { + model="Thunder";whitespace around the =.
Quoted 5 lines
arch/arm64/boot/dts/thunder.dts line 14
Quoted 1 line
arch/arm64/boot/dts/thunder.dts new file +1
-0+ compatible = "cavium, thunder";Please document compatible strings. There should not be a space in the string. This should be more specific. The comments refer to 8xxx so I assume there is a part number.
OK will do.
Quoted 10 lines
Quoted 10 lines
arch/arm64/boot/dts/thunder.dts new file +9
-0+ interrupt-parent = <&gic0>; + #address-cells = <2>; + #size-cells = <2>; + + aliases { + serial0 = &uaa0; + serial1 = &uaa1; + }; +
Quoted 8 lines
arch/arm64/boot/dts/thunder.dts line 29
Quoted 6 lines
arch/arm64/boot/dts/thunder.dts new file +6
-0+ cpus { + #address-cells = <2>; + #size-cells = <0>; + cpu at 0 { + device_type = "cpu"; + compatible = "arm,armv8";Custom core right? Add a compatible string for the core name.
OK, I guess that should be also documented, right? But I didn't find other ARM64 platforms doing it.
Quoted 12 lines
arch/arm64/boot/dts/thunder.dts line 39
Quoted 11 lines
arch/arm64/boot/dts/thunder.dts new file +10
-0+ reg = <0x0 0x0>; + enable-method = "spin-table"; + cpu-release-addr = <0x0 0xfff8>; + }; + cpu at 1 { + device_type = "cpu"; + compatible = "arm,armv8"; + reg = <0x0 0x1>; + enable-method = "spin-table"; + cpu-release-addr = <0x0 0xfff8>;
Quoted 1 line
Same address for all cores? I don't see how that will work for hotplug.
Hotplug will work if we use PSCI boot-method. For spin-table there is no hotplug support. And we don't enable it too.
Quoted 57 lines
Quoted 37 lines
arch/arm64/boot/dts/thunder.dts new file +36
-0+ }; + cpu at 2 { + device_type = "cpu"; + compatible = "arm,armv8"; + reg = <0x0 0x2>; + enable-method = "spin-table"; + cpu-release-addr = <0x0 0xfff8>; + }; + cpu at 3 { + device_type = "cpu"; + compatible = "arm,armv8"; + reg = <0x0 0x3>; + enable-method = "spin-table"; + cpu-release-addr = <0x0 0xfff8>; + }; + cpu at 4 { + device_type = "cpu"; + compatible = "arm,armv8"; + reg = <0x0 0x4>; + enable-method = "spin-table"; + cpu-release-addr = <0x0 0xfff8>; + }; + cpu at 5 { + device_type = "cpu"; + compatible = "arm,armv8"; + reg = <0x0 0x5>; + enable-method = "spin-table"; + cpu-release-addr = <0x0 0xfff8>; + }; + cpu at 6 { + device_type = "cpu"; + compatible = "arm,armv8"; + reg = <0x0 0x6>; + enable-method = "spin-table"; + cpu-release-addr = <0x0 0xfff8>; + };arch/arm64/boot/dts/thunder.dts line 85
Quoted 10 lines
arch/arm64/boot/dts/thunder.dts new file +10
-0+ cpu at 7 { + device_type = "cpu"; + compatible = "arm,armv8"; + reg = <0x0 0x7>; + enable-method = "spin-table"; + cpu-release-addr = <0x0 0xfff8>; + }; + + }; + memory at 0x0 {Drop the '0x'
arch/arm64/boot/dts/thunder.dts line 90
Quoted 5 lines
arch/arm64/boot/dts/thunder.dts new file +5
-0+ device_type = "memory"; + reg = <0x0 0x0 0x0 0x80000000>; + }; + + gic0: interrupt-controller at 0x801000000000 {Ditto
Quoted 8 lines
arch/arm64/boot/dts/thunder.dts line 96
Quoted 6 lines
arch/arm64/boot/dts/thunder.dts new file +6
-0+ compatible = "arm,gic-v3"; + #interrupt-cells = <3>; + #address-cells = <0>; + interrupt-controller; + reg = <0x8010 0x0 0x0 0x10000>, // GICD + <0x8010 0x80000000 0x0 0x100000>; // GICRUmm, so no virtualization support?
Why do you think there is no virtualization support? Our platform is not backward compatible with GICv2
Quoted 2 lines
This should be under a bus node.
Why GIC should be under a bus node?
Quoted 24 lines
Quoted 11 lines
arch/arm64/boot/dts/thunder.dts new file +10
-0+ interrupts = <1 9 0xf04>; + }; + + timer { + compatible = "arm,armv8-timer"; + interrupts = <1 13 0xff01>, + <1 14 0xff01>, + <1 11 0xff01>, + <1 10 0xff01>; + };arch/arm64/boot/dts/thunder.dts line 116
Quoted 10 lines
arch/arm64/boot/dts/thunder.dts new file +10
-0+ + on-chip-devices { + compatible = "simple-bus"; + #address-cells = <2>; + #size-cells = <2>; + ranges; + + ahci0: host-bus-adapter at 810000000000 { + compatible = "snps,spear-ahci"; + reg-io-width = <4>;This property is not used for this binding.
Quoted 16 lines
Quoted 4 lines
arch/arm64/boot/dts/thunder.dts new file +4
-0+ reg = <0x8100 0x0 0x0 0x1100>; + interrupts = <1 32 4>; + }; +arch/arm64/boot/dts/thunder.dts line 130
Quoted 10 lines
arch/arm64/boot/dts/thunder.dts new file +10
-0+ nic0: ethernet at 843000000000 { + compatible = "smsc,lan9115"; + reg-io-width = <4>; + reg = <0x8430 0x0 0x0 0x1000>; + interrupts = <1 31 4>; + }; + }; + + amba { + compatible = "arm,amba-bus";should be "simple-bus"
PL011 will not get probed if not "amba-bus"
Quoted 8 lines
arch/arm64/boot/dts/thunder.dts line 135
Quoted 6 lines
arch/arm64/boot/dts/thunder.dts new file +5
-0+ #address-cells = <2>; + #size-cells = <2>; + ranges; + + refclk24mhz: refclk24mhz {move to top level or a clocks node.
OK.
Quoted 11 lines
arch/arm64/boot/dts/thunder.dts line 142
Quoted 8 lines
arch/arm64/boot/dts/thunder.dts new file +7
-0+ compatible = "fixed-clock"; + #clock-cells = <0>; + clock-frequency = <24000000>; + clock-output-names = "refclk24mhz"; + }; + + uaa0: uart at 87e024000000 {use ranges so the bus address is 0x87e0_00000000.
OK, I need to look at this.
Quoted 8 lines
arch/arm64/boot/dts/thunder.dts line 147
Quoted 5 lines
arch/arm64/boot/dts/thunder.dts new file +5
-0+ compatible = "arm,pl011", "arm,primecell"; + reg = <0x87e0 0x24000000 0x0 0x1000>; + interrupts = <1 21 4>; + clocks = <&refclk24mhz>; + clock-names = "apb_pclk";Your APB bus frequency is really only 24MHz? There should 2 clocks defined here.
This clock is a dummy one as this DTS is based on a simulator. Is 2 clocks really a necessity? Then don't the drivers needs to do something if we don't define 2 clocks?
Quoted 19 lines
Quoted 19 lines
arch/arm64/boot/dts/thunder.dts new file +11
-1+ }; + + uaa1: uart at 87e025000000 { + compatible = "arm,pl011", "arm,primecell"; + reg = <0x87e0 0x25000000 0x0 0x1000>; + interrupts = <1 22 4>; + clocks = <&refclk24mhz>; + clock-names = "apb_pclk"; + }; + }; +}; --1.7.1
_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel at lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
PATCH v2 1/1 arm64: Add initial dts for Cavium Thunder SoC
Quoted 40 lines · Radha Mohan
Quoted 29 lines · Rob Herring
Quoted 28 lines · On Mon, Mar 3, 2014 at 1:19 AM, <[email protected]> wrote:
From: Radha Mohan Chintakuntla <[email protected]>
Signed-off-by Radha Mohan Chintakuntla <[email protected]>---
arch/arm64/boot/dts/Makefile| 1 +arch/arm64/boot/dts/thunder.dts| 158 +++++++++++++++++++++++++++++++++++++++2 files changed, 159 insertions(+), 0 deletions(-)arch/arm64/boot/dts/Makefile +0
-0@@ -1,5 +1,6 @@dtb-$(CONFIG_ARCH_VEXPRESS) += rtsm_ve-aemv8a.dtb foundation-v8.dtb dtb-$(CONFIG_ARCH_XGENE) += apm-mustang.dtb +dtb-$(CONFIG_ARCH_THUNDER) += thunder.dtb
targets += dtbs
targets += $(dtb-y)arch/arm64/boot/dts/thunder.dts new file +3
-0@@ -0,0 +1,158 @@ +/* + * Cavium Thunder 8xxx DTS file + *Quoted 9 lines
arch/arm64/boot/dts/thunder.dts line 9
Quoted 6 lines
arch/arm64/boot/dts/thunder.dts new file +6
-0+ * Copyright (C) 2013, Cavium Inc. + * + * This program is free software; you can redistribute it and/or + * modify it under the terms of the GNU General Public License as + * published by the Free Software Foundation; either version 2 of + * the License, or (at your option) any later version.DTS files should really be BSD licensed for use on other OSs.
Quoted 2 lines
Is this a new requirement? I haven't seen any DTS (atleast in ARM/ARM64) that uses BSD license.
Its not a new requirement, just a reasonable suggestion if you think that someone other OS might be booted on your hardware that would use DT. Some of the PowerPC DTs are licensed this way to allow BSD based OS and some 3rd party OSes to use the device tree.
Quoted 35 lines
Quoted 8 lines
arch/arm64/boot/dts/thunder.dts line 13
Quoted 5 lines
arch/arm64/boot/dts/thunder.dts new file +4
-0+ */ +/dts-v1/; +/ { + model="Thunder";whitespace around the =.
Quoted 5 lines
arch/arm64/boot/dts/thunder.dts line 14
Quoted 1 line
arch/arm64/boot/dts/thunder.dts new file +1
-0+ compatible = "cavium, thunder";Please document compatible strings. There should not be a space in the string. This should be more specific. The comments refer to 8xxx so I assume there is a part number.
OK will do.
Quoted 10 lines
Quoted 10 lines
arch/arm64/boot/dts/thunder.dts new file +9
-0+ interrupt-parent = <&gic0>; + #address-cells = <2>; + #size-cells = <2>; + + aliases { + serial0 = &uaa0; + serial1 = &uaa1; + }; +Quoted 8 lines
arch/arm64/boot/dts/thunder.dts line 29
Quoted 6 lines
arch/arm64/boot/dts/thunder.dts new file +6
-0+ cpus { + #address-cells = <2>; + #size-cells = <0>; + cpu at 0 { + device_type = "cpu"; + compatible = "arm,armv8";Custom core right? Add a compatible string for the core name.
Quoted 2 lines
OK, I guess that should be also documented, right? But I didn't find other ARM64 platforms doing it.
The apm-storm.dtsi does it. The only in kernel .dts for arm64 are for the foundation model.
Quoted 84 lines
Quoted 13 lines
arch/arm64/boot/dts/thunder.dts line 39
Quoted 12 lines
arch/arm64/boot/dts/thunder.dts new file +10
-0+ reg = <0x0 0x0>; + enable-method = "spin-table"; + cpu-release-addr = <0x0 0xfff8>; + }; + cpu at 1 { + device_type = "cpu"; + compatible = "arm,armv8"; + reg = <0x0 0x1>; + enable-method = "spin-table"; + cpu-release-addr = <0x0 0xfff8>;Quoted 1 line
Same address for all cores? I don't see how that will work for hotplug.
Hotplug will work if we use PSCI boot-method. For spin-table there is no hotplug support. And we don't enable it too.
Quoted 57 lines
Quoted 37 lines
arch/arm64/boot/dts/thunder.dts new file +36
-0+ }; + cpu at 2 { + device_type = "cpu"; + compatible = "arm,armv8"; + reg = <0x0 0x2>; + enable-method = "spin-table"; + cpu-release-addr = <0x0 0xfff8>; + }; + cpu at 3 { + device_type = "cpu"; + compatible = "arm,armv8"; + reg = <0x0 0x3>; + enable-method = "spin-table"; + cpu-release-addr = <0x0 0xfff8>; + }; + cpu at 4 { + device_type = "cpu"; + compatible = "arm,armv8"; + reg = <0x0 0x4>; + enable-method = "spin-table"; + cpu-release-addr = <0x0 0xfff8>; + }; + cpu at 5 { + device_type = "cpu"; + compatible = "arm,armv8"; + reg = <0x0 0x5>; + enable-method = "spin-table"; + cpu-release-addr = <0x0 0xfff8>; + }; + cpu at 6 { + device_type = "cpu"; + compatible = "arm,armv8"; + reg = <0x0 0x6>; + enable-method = "spin-table"; + cpu-release-addr = <0x0 0xfff8>; + };arch/arm64/boot/dts/thunder.dts line 85
Quoted 10 lines
arch/arm64/boot/dts/thunder.dts new file +10
-0+ cpu at 7 { + device_type = "cpu"; + compatible = "arm,armv8"; + reg = <0x0 0x7>; + enable-method = "spin-table"; + cpu-release-addr = <0x0 0xfff8>; + }; + + }; + memory at 0x0 {Drop the '0x'
arch/arm64/boot/dts/thunder.dts line 90
Quoted 5 lines
arch/arm64/boot/dts/thunder.dts new file +5
-0+ device_type = "memory"; + reg = <0x0 0x0 0x0 0x80000000>; + }; + + gic0: interrupt-controller at 0x801000000000 {Ditto
Quoted 8 lines
arch/arm64/boot/dts/thunder.dts line 96
Quoted 6 lines
arch/arm64/boot/dts/thunder.dts new file +6
-0+ compatible = "arm,gic-v3"; + #interrupt-cells = <3>; + #address-cells = <0>; + interrupt-controller; + reg = <0x8010 0x0 0x0 0x10000>, // GICD + <0x8010 0x80000000 0x0 0x100000>; // GICRUmm, so no virtualization support?
Quoted 7 lines
Why do you think there is no virtualization support? Our platform is not backward compatible with GICv2
Quoted 2 lines
This should be under a bus node.
Why GIC should be under a bus node?
Typically we?ve collected on chip peripheral registers in a ?soc? container that is compatible with ?simple-bus"
Quoted 95 lines
Quoted 25 lines
Quoted 12 lines
arch/arm64/boot/dts/thunder.dts new file +10
-0+ interrupts = <1 9 0xf04>; + }; + + timer { + compatible = "arm,armv8-timer"; + interrupts = <1 13 0xff01>, + <1 14 0xff01>, + <1 11 0xff01>, + <1 10 0xff01>; + };arch/arm64/boot/dts/thunder.dts line 116
Quoted 10 lines
arch/arm64/boot/dts/thunder.dts new file +10
-0+ + on-chip-devices { + compatible = "simple-bus"; + #address-cells = <2>; + #size-cells = <2>; + ranges; + + ahci0: host-bus-adapter at 810000000000 { + compatible = "snps,spear-ahci"; + reg-io-width = <4>;This property is not used for this binding.
Quoted 16 lines
Quoted 4 lines
arch/arm64/boot/dts/thunder.dts new file +4
-0+ reg = <0x8100 0x0 0x0 0x1100>; + interrupts = <1 32 4>; + }; +arch/arm64/boot/dts/thunder.dts line 130
Quoted 10 lines
arch/arm64/boot/dts/thunder.dts new file +10
-0+ nic0: ethernet at 843000000000 { + compatible = "smsc,lan9115"; + reg-io-width = <4>; + reg = <0x8430 0x0 0x0 0x1000>; + interrupts = <1 31 4>; + }; + }; + + amba { + compatible = "arm,amba-bus";should be "simple-bus"
PL011 will not get probed if not "amba-bus"
Quoted 8 lines
arch/arm64/boot/dts/thunder.dts line 135
Quoted 6 lines
arch/arm64/boot/dts/thunder.dts new file +5
-0+ #address-cells = <2>; + #size-cells = <2>; + ranges; + + refclk24mhz: refclk24mhz {move to top level or a clocks node.
OK.
Quoted 11 lines
arch/arm64/boot/dts/thunder.dts line 142
Quoted 8 lines
arch/arm64/boot/dts/thunder.dts new file +7
-0+ compatible = "fixed-clock"; + #clock-cells = <0>; + clock-frequency = <24000000>; + clock-output-names = "refclk24mhz"; + }; + + uaa0: uart at 87e024000000 {use ranges so the bus address is 0x87e0_00000000.
OK, I need to look at this.
Quoted 8 lines
arch/arm64/boot/dts/thunder.dts line 147
Quoted 5 lines
arch/arm64/boot/dts/thunder.dts new file +5
-0+ compatible = "arm,pl011", "arm,primecell"; + reg = <0x87e0 0x24000000 0x0 0x1000>; + interrupts = <1 21 4>; + clocks = <&refclk24mhz>; + clock-names = "apb_pclk";Your APB bus frequency is really only 24MHz? There should 2 clocks defined here.
This clock is a dummy one as this DTS is based on a simulator. Is 2 clocks really a necessity? Then don't the drivers needs to do something if we don't define 2 clocks?
Quoted 13 lines
Quoted 13 lines
arch/arm64/boot/dts/thunder.dts new file +11
-0+ }; + + uaa1: uart at 87e025000000 { + compatible = "arm,pl011", "arm,primecell"; + reg = <0x87e0 0x25000000 0x0 0x1000>; + interrupts = <1 22 4>; + clocks = <&refclk24mhz>; + clock-names = "apb_pclk"; + }; + }; +};?
1.7.1
Signature · 2 lines
Employee of Qualcomm Innovation Center, Inc. Qualcomm Innovation Center, Inc. is a member of Code Aurora Forum, hosted by The Linux Foundation
PATCH v2 1/1 arm64: Add initial dts for Cavium Thunder SoC
Quoted 43 lines · Kumar Gala
Quoted 40 lines · Radha Mohan
Quoted 38 lines · Rob Herring
Quoted 25 lines · On Mon, Mar 3, 2014 at 1:19 AM, <[email protected]> wrote:
From: Radha Mohan Chintakuntla <[email protected]>
Signed-off-by Radha Mohan Chintakuntla <[email protected]>---
arch/arm64/boot/dts/Makefile| 1 +arch/arm64/boot/dts/thunder.dts| 158 +++++++++++++++++++++++++++++++++++++++2 files changed, 159 insertions(+), 0 deletions(-)arch/arm64/boot/dts/Makefile +0
-0@@ -1,5 +1,6 @@dtb-$(CONFIG_ARCH_VEXPRESS) += rtsm_ve-aemv8a.dtb foundation-v8.dtb dtb-$(CONFIG_ARCH_XGENE) += apm-mustang.dtb +dtb-$(CONFIG_ARCH_THUNDER) += thunder.dtb
targets += dtbs
targets += $(dtb-y)arch/arm64/boot/dts/thunder.dts new file +0
-0arch/arm64/boot/dts/thunder.dts line 9
Quoted 9 lines
arch/arm64/boot/dts/thunder.dts new file +9
-0@@ -0,0 +1,158 @@ +/* + * Cavium Thunder 8xxx DTS file + * + * Copyright (C) 2013, Cavium Inc. + * + * This program is free software; you can redistribute it and/or + * modify it under the terms of the GNU General Public License as + * published by the Free Software Foundation; either version 2 of + * the License, or (at your option) any later version.DTS files should really be BSD licensed for use on other OSs.
Quoted 5 lines
Quoted 2 lines
Is this a new requirement? I haven't seen any DTS (atleast in ARM/ARM64) that uses BSD license.
Its not a new requirement, just a reasonable suggestion if you think that someone other OS might be booted on your hardware that would use DT. Some of the PowerPC DTs are licensed this way to allow BSD based OS and some 3rd party OSes to use the device tree.
OK got it.
Quoted 35 lines
Quoted 35 lines
Quoted 8 lines
arch/arm64/boot/dts/thunder.dts line 13
Quoted 5 lines
arch/arm64/boot/dts/thunder.dts new file +4
-0+ */ +/dts-v1/; +/ { + model="Thunder";whitespace around the =.
Quoted 5 lines
arch/arm64/boot/dts/thunder.dts line 14
Quoted 1 line
arch/arm64/boot/dts/thunder.dts new file +1
-0+ compatible = "cavium, thunder";Please document compatible strings. There should not be a space in the string. This should be more specific. The comments refer to 8xxx so I assume there is a part number.
OK will do.
Quoted 18 lines
Quoted 6 lines
arch/arm64/boot/dts/thunder.dts new file +5
-0+ interrupt-parent = <&gic0>; + #address-cells = <2>; + #size-cells = <2>; + + aliases {arch/arm64/boot/dts/thunder.dts line 29
Quoted 10 lines
arch/arm64/boot/dts/thunder.dts new file +10
-0+ serial0 = &uaa0; + serial1 = &uaa1; + }; + + cpus { + #address-cells = <2>; + #size-cells = <0>; + cpu at 0 { + device_type = "cpu"; + compatible = "arm,armv8";Custom core right? Add a compatible string for the core name.
Quoted 4 lines
Quoted 2 lines
OK, I guess that should be also documented, right? But I didn't find other ARM64 platforms doing it.
The
apm-storm.dtsidoes it. The only in kernel .dts for arm64 are for the foundation model.
Doesn't this need documentation then. I didn't find "potenza" anywhere in the Documentation.
Quoted 89 lines
Quoted 89 lines
Quoted 14 lines
arch/arm64/boot/dts/thunder.dts line 39
Quoted 13 lines
arch/arm64/boot/dts/thunder.dts new file +10
-0+ reg = <0x0 0x0>; + enable-method = "spin-table"; + cpu-release-addr = <0x0 0xfff8>; + }; + cpu at 1 { + device_type = "cpu"; + compatible = "arm,armv8"; + reg = <0x0 0x1>; + enable-method = "spin-table"; + cpu-release-addr = <0x0 0xfff8>;Quoted 1 line
Same address for all cores? I don't see how that will work for hotplug.
Hotplug will work if we use PSCI boot-method. For spin-table there is no hotplug support. And we don't enable it too.
Quoted 57 lines
Quoted 37 lines
arch/arm64/boot/dts/thunder.dts new file +36
-0+ }; + cpu at 2 { + device_type = "cpu"; + compatible = "arm,armv8"; + reg = <0x0 0x2>; + enable-method = "spin-table"; + cpu-release-addr = <0x0 0xfff8>; + }; + cpu at 3 { + device_type = "cpu"; + compatible = "arm,armv8"; + reg = <0x0 0x3>; + enable-method = "spin-table"; + cpu-release-addr = <0x0 0xfff8>; + }; + cpu at 4 { + device_type = "cpu"; + compatible = "arm,armv8"; + reg = <0x0 0x4>; + enable-method = "spin-table"; + cpu-release-addr = <0x0 0xfff8>; + }; + cpu at 5 { + device_type = "cpu"; + compatible = "arm,armv8"; + reg = <0x0 0x5>; + enable-method = "spin-table"; + cpu-release-addr = <0x0 0xfff8>; + }; + cpu at 6 { + device_type = "cpu"; + compatible = "arm,armv8"; + reg = <0x0 0x6>; + enable-method = "spin-table"; + cpu-release-addr = <0x0 0xfff8>; + };arch/arm64/boot/dts/thunder.dts line 85
Quoted 10 lines
arch/arm64/boot/dts/thunder.dts new file +10
-0+ cpu at 7 { + device_type = "cpu"; + compatible = "arm,armv8"; + reg = <0x0 0x7>; + enable-method = "spin-table"; + cpu-release-addr = <0x0 0xfff8>; + }; + + }; + memory at 0x0 {Drop the '0x'
arch/arm64/boot/dts/thunder.dts line 90
Quoted 5 lines
arch/arm64/boot/dts/thunder.dts new file +5
-0+ device_type = "memory"; + reg = <0x0 0x0 0x0 0x80000000>; + }; + + gic0: interrupt-controller at 0x801000000000 {Ditto
Quoted 8 lines
arch/arm64/boot/dts/thunder.dts line 96
Quoted 6 lines
arch/arm64/boot/dts/thunder.dts new file +6
-0+ compatible = "arm,gic-v3"; + #interrupt-cells = <3>; + #address-cells = <0>; + interrupt-controller; + reg = <0x8010 0x0 0x0 0x10000>, // GICD + <0x8010 0x80000000 0x0 0x100000>; // GICRUmm, so no virtualization support?
Why do you think there is no virtualization support? Our platform is not backward compatible with GICv2
Quoted 1 line
Quoted 6 lines
Quoted 3 lines
Quoted 1 line
This should be under a bus node.
Why GIC should be under a bus node?
Typically we've collected on chip peripheral registers in a "soc" container that is compatible with "simple-bus"
Right. But I thought Rob was referring to GIC being under a bus node. Am I missing something here?
Quoted 100 lines
Quoted 82 lines
Quoted 25 lines
Quoted 12 lines
arch/arm64/boot/dts/thunder.dts new file +10
-0+ interrupts = <1 9 0xf04>; + }; + + timer { + compatible = "arm,armv8-timer"; + interrupts = <1 13 0xff01>, + <1 14 0xff01>, + <1 11 0xff01>, + <1 10 0xff01>; + };arch/arm64/boot/dts/thunder.dts line 116
Quoted 10 lines
arch/arm64/boot/dts/thunder.dts new file +10
-0+ + on-chip-devices { + compatible = "simple-bus"; + #address-cells = <2>; + #size-cells = <2>; + ranges; + + ahci0: host-bus-adapter at 810000000000 { + compatible = "snps,spear-ahci"; + reg-io-width = <4>;This property is not used for this binding.
Quoted 16 lines
Quoted 4 lines
arch/arm64/boot/dts/thunder.dts new file +4
-0+ reg = <0x8100 0x0 0x0 0x1100>; + interrupts = <1 32 4>; + }; +arch/arm64/boot/dts/thunder.dts line 130
Quoted 10 lines
arch/arm64/boot/dts/thunder.dts new file +10
-0+ nic0: ethernet at 843000000000 { + compatible = "smsc,lan9115"; + reg-io-width = <4>; + reg = <0x8430 0x0 0x0 0x1000>; + interrupts = <1 31 4>; + }; + }; + + amba { + compatible = "arm,amba-bus";should be "simple-bus"
PL011 will not get probed if not "amba-bus"
Quoted 8 lines
arch/arm64/boot/dts/thunder.dts line 135
Quoted 6 lines
arch/arm64/boot/dts/thunder.dts new file +5
-0+ #address-cells = <2>; + #size-cells = <2>; + ranges; + + refclk24mhz: refclk24mhz {move to top level or a clocks node.
OK.
Quoted 11 lines
arch/arm64/boot/dts/thunder.dts line 142
Quoted 8 lines
arch/arm64/boot/dts/thunder.dts new file +7
-0+ compatible = "fixed-clock"; + #clock-cells = <0>; + clock-frequency = <24000000>; + clock-output-names = "refclk24mhz"; + }; + + uaa0: uart at 87e024000000 {use ranges so the bus address is 0x87e0_00000000.
OK, I need to look at this.
Quoted 8 lines
arch/arm64/boot/dts/thunder.dts line 147
Quoted 5 lines
arch/arm64/boot/dts/thunder.dts new file +5
-0+ compatible = "arm,pl011", "arm,primecell"; + reg = <0x87e0 0x24000000 0x0 0x1000>; + interrupts = <1 21 4>; + clocks = <&refclk24mhz>; + clock-names = "apb_pclk";Your APB bus frequency is really only 24MHz? There should 2 clocks defined here.
This clock is a dummy one as this DTS is based on a simulator. Is 2 clocks really a necessity? Then don't the drivers needs to do something if we don't define 2 clocks?
Quoted 13 lines
Quoted 5 lines
Quoted 5 lines
arch/arm64/boot/dts/thunder.dts new file +5
-0+ }; + + uaa1: uart at 87e025000000 { + compatible = "arm,pl011", "arm,primecell"; + reg = <0x87e0 0x25000000 0x0 0x1000>;Quoted 8 lines
Quoted 8 lines
arch/arm64/boot/dts/thunder.dts new file +6
-1+ interrupts = <1 22 4>; + clocks = <&refclk24mhz>; + clock-names = "apb_pclk"; + }; + }; +}; --1.7.1
--
Employee of Qualcomm Innovation Center, Inc.
Qualcomm Innovation Center, Inc. is a member of Code Aurora Forum, hosted by The Linux Foundation
PATCH v2 1/1 arm64: Add initial dts for Cavium Thunder SoC
In reply to Radha MohanQuoted 40 lines · Radha Mohan
Quoted 29 lines · Rob Herring
Quoted 28 lines · On Mon, Mar 3, 2014 at 1:19 AM, <[email protected]> wrote:
From: Radha Mohan Chintakuntla <[email protected]>
Signed-off-by Radha Mohan Chintakuntla <[email protected]>--- arch/arm64/boot/dts/Makefile | 1 + arch/arm64/boot/dts/thunder.dts | 158 +++++++++++++++++++++++++++++++++++++++ 2 files changed, 159 insertions(+), 0 deletions(-)arch/arm64/boot/dts/Makefile +1
-0@@ -1,5 +1,6 @@ dtb-$(CONFIG_ARCH_VEXPRESS) += rtsm_ve-aemv8a.dtb foundation-v8.dtb dtb-$(CONFIG_ARCH_XGENE) += apm-mustang.dtb +dtb-$(CONFIG_ARCH_THUNDER) += thunder.dtb targets += dtbs targets += $(dtb-y)arch/arm64/boot/dts/thunder.dts new file +3
-0@@ -0,0 +1,158 @@ +/* + * Cavium Thunder 8xxx DTS file + *Quoted 9 lines
arch/arm64/boot/dts/thunder.dts line 9
Quoted 6 lines
arch/arm64/boot/dts/thunder.dts new file +6
-0+ * Copyright (C) 2013, Cavium Inc. + * + * This program is free software; you can redistribute it and/or + * modify it under the terms of the GNU General Public License as + * published by the Free Software Foundation; either version 2 of + * the License, or (at your option) any later version.DTS files should really be BSD licensed for use on other OSs.
Quoted 2 lines
Is this a new requirement? I haven't seen any DTS (atleast in ARM/ARM64) that uses BSD license.
I'm not saying you have to. Considering the DTB should come from the firmware and you could want to boot other OS's, BSD license makes more sense. Also at some point, we plan to move dts files out of the kernel to better support other projects like u-boot or FreeBSD. It may become more of an issue then and dts files could need to be re-licensed. I'd rather avoid the whole issue of trying to re-license things.
Quoted 34 lines
Quoted 7 lines
arch/arm64/boot/dts/thunder.dts line 13
Quoted 4 lines
arch/arm64/boot/dts/thunder.dts new file +4
-0+ */ +/dts-v1/; +/ { + model="Thunder";whitespace around the =.
Quoted 5 lines
arch/arm64/boot/dts/thunder.dts line 14
Quoted 1 line
arch/arm64/boot/dts/thunder.dts new file +1
-0+ compatible = "cavium, thunder";Please document compatible strings. There should not be a space in the string. This should be more specific. The comments refer to 8xxx so I assume there is a part number.
OK will do.
Quoted 10 lines
Quoted 10 lines
arch/arm64/boot/dts/thunder.dts new file +9
-0+ interrupt-parent = <&gic0>; + #address-cells = <2>; + #size-cells = <2>; + + aliases { + serial0 = &uaa0; + serial1 = &uaa1; + }; +Quoted 8 lines
arch/arm64/boot/dts/thunder.dts line 29
Quoted 6 lines
arch/arm64/boot/dts/thunder.dts new file +6
-0+ cpus { + #address-cells = <2>; + #size-cells = <0>; + cpu at 0 { + device_type = "cpu"; + compatible = "arm,armv8";Custom core right? Add a compatible string for the core name.
Quoted 2 lines
OK, I guess that should be also documented, right? But I didn't find other ARM64 platforms doing it.
arm,potenza for APM. Everything else is models, so they don't really count.
Quoted 82 lines
Quoted 13 lines
arch/arm64/boot/dts/thunder.dts line 39
Quoted 12 lines
arch/arm64/boot/dts/thunder.dts new file +10
-0+ reg = <0x0 0x0>; + enable-method = "spin-table"; + cpu-release-addr = <0x0 0xfff8>; + }; + cpu at 1 { + device_type = "cpu"; + compatible = "arm,armv8"; + reg = <0x0 0x1>; + enable-method = "spin-table"; + cpu-release-addr = <0x0 0xfff8>;Quoted 1 line
Same address for all cores? I don't see how that will work for hotplug.
Hotplug will work if we use PSCI boot-method. For spin-table there is no hotplug support. And we don't enable it too.
Quoted 64 lines
Quoted 37 lines
arch/arm64/boot/dts/thunder.dts new file +36
-0+ }; + cpu at 2 { + device_type = "cpu"; + compatible = "arm,armv8"; + reg = <0x0 0x2>; + enable-method = "spin-table"; + cpu-release-addr = <0x0 0xfff8>; + }; + cpu at 3 { + device_type = "cpu"; + compatible = "arm,armv8"; + reg = <0x0 0x3>; + enable-method = "spin-table"; + cpu-release-addr = <0x0 0xfff8>; + }; + cpu at 4 { + device_type = "cpu"; + compatible = "arm,armv8"; + reg = <0x0 0x4>; + enable-method = "spin-table"; + cpu-release-addr = <0x0 0xfff8>; + }; + cpu at 5 { + device_type = "cpu"; + compatible = "arm,armv8"; + reg = <0x0 0x5>; + enable-method = "spin-table"; + cpu-release-addr = <0x0 0xfff8>; + }; + cpu at 6 { + device_type = "cpu"; + compatible = "arm,armv8"; + reg = <0x0 0x6>; + enable-method = "spin-table"; + cpu-release-addr = <0x0 0xfff8>; + };arch/arm64/boot/dts/thunder.dts line 85
Quoted 10 lines
arch/arm64/boot/dts/thunder.dts new file +10
-0+ cpu at 7 { + device_type = "cpu"; + compatible = "arm,armv8"; + reg = <0x0 0x7>; + enable-method = "spin-table"; + cpu-release-addr = <0x0 0xfff8>; + }; + + }; + memory at 0x0 {Drop the '0x'
arch/arm64/boot/dts/thunder.dts line 90
Quoted 5 lines
arch/arm64/boot/dts/thunder.dts new file +5
-0+ device_type = "memory"; + reg = <0x0 0x0 0x0 0x80000000>; + }; + + gic0: interrupt-controller at 0x801000000000 {Ditto
Quoted 6 lines
arch/arm64/boot/dts/thunder.dts new file +6
-0+ compatible = "arm,gic-v3"; + #interrupt-cells = <3>; + #address-cells = <0>; + interrupt-controller; + reg = <0x8010 0x0 0x0 0x10000>, // GICD + <0x8010 0x80000000 0x0 0x100000>; // GICR
Quoted 4 lines
Quoted 1 line
Umm, so no virtualization support?
Why do you think there is no virtualization support? Our platform is not backward compatible with GICv2
Okay.
Quoted 5 lines
Quoted 3 lines
This should be under a bus node.
Why GIC should be under a bus node?
Because it is on some bus.
Quoted 42 lines
Quoted 25 lines
Quoted 12 lines
arch/arm64/boot/dts/thunder.dts new file +10
-0+ interrupts = <1 9 0xf04>; + }; + + timer { + compatible = "arm,armv8-timer"; + interrupts = <1 13 0xff01>, + <1 14 0xff01>, + <1 11 0xff01>, + <1 10 0xff01>; + };arch/arm64/boot/dts/thunder.dts line 116
Quoted 10 lines
arch/arm64/boot/dts/thunder.dts new file +10
-0+ + on-chip-devices { + compatible = "simple-bus"; + #address-cells = <2>; + #size-cells = <2>; + ranges; + + ahci0: host-bus-adapter at 810000000000 { + compatible = "snps,spear-ahci"; + reg-io-width = <4>;This property is not used for this binding.
Quoted 16 lines
Quoted 4 lines
arch/arm64/boot/dts/thunder.dts new file +4
-0+ reg = <0x8100 0x0 0x0 0x1100>; + interrupts = <1 32 4>; + }; +arch/arm64/boot/dts/thunder.dts line 130
Quoted 10 lines
arch/arm64/boot/dts/thunder.dts new file +10
-0+ nic0: ethernet at 843000000000 { + compatible = "smsc,lan9115"; + reg-io-width = <4>; + reg = <0x8430 0x0 0x0 0x1000>; + interrupts = <1 31 4>; + }; + }; + + amba { + compatible = "arm,amba-bus";should be "simple-bus"
Quoted 1 line
PL011 will not get probed if not "amba-bus"
Yes, it will. It is "arm,primecell" on the devices that is used for probing. There is really no such thing as an AMBA bus from a software view. There are no registers to program the bus (typically) or discovery mechanism. The ID registers within devices give you device type, but nothing describes where devices are.
Quoted 32 lines
Quoted 9 lines
arch/arm64/boot/dts/thunder.dts line 135
Quoted 7 lines
arch/arm64/boot/dts/thunder.dts new file +5
-0+ #address-cells = <2>; + #size-cells = <2>; + ranges; + + refclk24mhz: refclk24mhz {move to top level or a clocks node.
OK.
Quoted 11 lines
arch/arm64/boot/dts/thunder.dts line 142
Quoted 8 lines
arch/arm64/boot/dts/thunder.dts new file +7
-0+ compatible = "fixed-clock"; + #clock-cells = <0>; + clock-frequency = <24000000>; + clock-output-names = "refclk24mhz"; + }; + + uaa0: uart at 87e024000000 {use ranges so the bus address is 0x87e0_00000000.
OK, I need to look at this.
Quoted 6 lines
Quoted 5 lines
arch/arm64/boot/dts/thunder.dts new file +5
-0+ compatible = "arm,pl011", "arm,primecell"; + reg = <0x87e0 0x24000000 0x0 0x1000>; + interrupts = <1 21 4>; + clocks = <&refclk24mhz>; + clock-names = "apb_pclk";
Quoted 6 lines
Quoted 2 lines
Your APB bus frequency is really only 24MHz? There should 2 clocks defined here.
This clock is a dummy one as this DTS is based on a simulator. Is 2 clocks really a necessity? Then don't the drivers needs to do something if we don't define 2 clocks?
Accurately describing the h/w is a necessity.
Rob