[OpenWrt-Devel] [PATCH v2] ipq40xx: Add support for IPQ4019 ap-dk07.1-c1

Christian Lamparter chunkeey at gmail.com
Fri Sep 7 11:06:28 EDT 2018


On Friday, September 7, 2018 2:10:30 PM CEST Ram Chandra Jangir wrote:
>  * SoC: QCA IPQ4019
>  * RAM: 512 MiB
>  * Flash: 32 MiB of SPI NOR and 128 MiB NAND
>  * Wireless: 2.4-GHz 802.11ng, and two 5-GHz 802.11ac
>  * IoT :
> 	CSR8811 radio [ supports BT classic and BLE ]
> 	SiLabs EFR32MG radio [ supports 802.15.4 ZigBee ]
> 
>    Tested on IPQ ap-dk07.1-c1 Board:
> 	1) NOR boot and NAND boot
> 	2) Tested USB and PCIe interfaces
Wait, usb is not enabled in the dts.

> 	3) WDOG test
> 	4) cpu frequency scaling
Hey, that's great. Glad to hear that it is working. 
So, do you know of the upstream status of the cpu frequency
scaling fix that OpenWrt carries:
<https://patchwork.codeaurora.org/patch/473589/>

Because it would be a real shame if that wasn't fixed upstream too,
right?

> 	5) ethernet, 2G and 5G WiFi
>
> 
> The ap-dk07.1-c1 board is already available in upstream
> linux v4.17
> 
> Signed-off-by: Ram Chandra Jangir <rjangir at codeaurora.org>
I do have a few comments down below.

> ---
> Changes since v1:
>  Fixed commit message for Wireless 2.4-GHz 802.11ng
>  Updated description for 904-Fix-rootfs-*-mount.patch
> 
>  .../etc/hotplug.d/firmware/11-ath10k-caldata       |  6 +-
>  target/linux/ipq40xx/image/Makefile                | 13 +++
>  ...ipq4019-Add-ipq4019-ap.dk07.1-common-data.patch | 99 ++++++++++++++++++++++
>  ...4019-Add-qcom-ipq4019-ap.dk07.1-c1-board-.patch | 99 ++++++++++++++++++++++
>  ...x-rootfs-conflict-with-OpenWrt-auto-mount.patch | 40 +++++++++
>  ...5-dts-qcom-ipq4019-ap.dk07.1-enable-nodes.patch | 79 +++++++++++++++++
>  6 files changed, 334 insertions(+), 2 deletions(-)
>  create mode 100644 target/linux/ipq40xx/patches-4.14/902-ARM-dts-ipq4019-Add-ipq4019-ap.dk07.1-common-data.patch
>  create mode 100644 target/linux/ipq40xx/patches-4.14/903-ARM-dts-ipq4019-Add-qcom-ipq4019-ap.dk07.1-c1-board-.patch
>  create mode 100644 target/linux/ipq40xx/patches-4.14/904-Fix-rootfs-conflict-with-OpenWrt-auto-mount.patch
>  create mode 100644 target/linux/ipq40xx/patches-4.14/905-dts-qcom-ipq4019-ap.dk07.1-enable-nodes.patch
> 
> diff --git a/target/linux/ipq40xx/base-files/etc/hotplug.d/firmware/11-ath10k-caldata b/target/linux/ipq40xx/base-files/etc/hotplug.d/firmware/11-ath10k-caldata
> index 1d4fd9c..83518c9 100644
> --- a/target/linux/ipq40xx/base-files/etc/hotplug.d/firmware/11-ath10k-caldata
> +++ b/target/linux/ipq40xx/base-files/etc/hotplug.d/firmware/11-ath10k-caldata
> @@ -143,7 +143,8 @@ case "$FIRMWARE" in
>  		;;
>  	compex,wpj428 |\
>  	openmesh,a42 |\
> -	openmesh,a62)
> +	openmesh,a62 |\
> +	qcom,ap-dk07.1-c1)
>  		ath10kcal_extract "0:ART" 4096 12064
>  		;;
>  	zyxel,nbg6617 |\
> @@ -179,7 +180,8 @@ case "$FIRMWARE" in
>  		;;
>  	compex,wpj428 |\
>  	openmesh,a42 |\
> -	openmesh,a62)
> +	openmesh,a62 |\
> +	qcom,ap-dk07.1-c1)
>  		ath10kcal_extract "0:ART" 20480 12064
>  		;;
>  	zyxel,nbg6617 |\
> diff --git a/target/linux/ipq40xx/image/Makefile b/target/linux/ipq40xx/image/Makefile
> index 5cd11ca..bee2145 100644
> --- a/target/linux/ipq40xx/image/Makefile
> +++ b/target/linux/ipq40xx/image/Makefile
> @@ -207,6 +207,19 @@ define Device/qcom_ap-dk04.1-c1
>  endef
>  TARGET_DEVICES += qcom_ap-dk04.1-c1
>  
> +define Device/qcom_ap-dk07.1-c1
> +	$(call Device/FitImage)
> +	$(call Device/UbiFit)
> +	BOARD_NAME := ap-dk07.1-c1
This is a new device. So, why do you need a legacy BOARD_NAME for it?

> +	DEVICE_DTS := qcom-ipq4019-ap.dk07.1-c1
> +	KERNEL_INSTALL := 1
> +	KERNEL_SIZE := 4096k
> +	BLOCKSIZE := 128k
> +	PAGESIZE := 2048
> +	DEVICE_TITLE := QCA AP-DK07.1-C1
> +endef
> +TARGET_DEVICES += qcom_ap-dk07.1-c1
What's wrong with calling the device "qcom_ipq4019-ap.dk07.1-c1"?

By the way:
Also, John is preparing to drop ath10k: 
<https://git.openwrt.org/?p=openwrt/staging/blogic.git;a=commit;h=f9ccc8f64670e3e15b160cffb19f1ef1a8519e97>

| author	John Crispin <john at phrozen.org>	
| Wed, 5 Sep 2018 14:51:44 +0200 (14:51 +0200)
| committer	John Crispin <john at phrozen.org>	
| Thu, 6 Sep 2018 19:00:21 +0200 (19:00 +0200)
| commit	f9ccc8f64670e3e15b160cffb19f1ef1a8519e97
|
|treewide: drop ath10k support
|
|people should use ath10k-ct instead
|
|Signed-off-by: John Crispin <john at phrozen.org>

So please make sure that the Wi-Fis are indeed working with ath10k-ct.

>  define Device/zyxel_nbg6617
>  	$(call Device/FitImageLzma)
>  	DEVICE_DTS := qcom-ipq4018-nbg6617
> diff --git a/target/linux/ipq40xx/patches-4.14/902-ARM-dts-ipq4019-Add-ipq4019-ap.dk07.1-common-data.patch b/target/linux/ipq40xx/patches-4.14/902-ARM-dts-ipq4019-Add-ipq4019-ap.dk07.1-common-data.patch
> new file mode 100644
> index 0000000..bcc109f
> --- /dev/null
> +++ b/target/linux/ipq40xx/patches-4.14/902-ARM-dts-ipq4019-Add-ipq4019-ap.dk07.1-common-data.patch
> @@ -0,0 +1,99 @@
> +From f97b2aaaf0734a3a6dd6effff6ee0aaa0a69cab5 Mon Sep 17 00:00:00 2001
> +From: Sricharan R <sricharan at codeaurora.org>
> +Date: Fri, 25 May 2018 11:41:17 +0530
> +Subject: [PATCH] ARM: dts: ipq4019: Add ipq4019-ap.dk07.1 common data
> +
> +Add the common data for all dk07 based boards.
> +
> +Reviewed-by: Abhishek Sahu <absahu at codeaurora.org>
> +Signed-off-by: Sricharan R <sricharan at codeaurora.org>
> +Signed-off-by: Andy Gross <andy.gross at linaro.org>
> +---
> + arch/arm/boot/dts/qcom-ipq4019-ap.dk07.1.dtsi | 75 +++++++++++++++++++++++++++
> + 1 file changed, 75 insertions(+)
> + create mode 100644 arch/arm/boot/dts/qcom-ipq4019-ap.dk07.1.dtsi
> +
> +diff --git a/arch/arm/boot/dts/qcom-ipq4019-ap.dk07.1.dtsi b/arch/arm/boot/dts/qcom-ipq4019-ap.dk07.1.dtsi
> +new file mode 100644
> +index 0000000..9f1a5a66
> +--- /dev/null
> ++++ b/arch/arm/boot/dts/qcom-ipq4019-ap.dk07.1.dtsi
I guess you'll need to push those changes upstream.

> +@@ -0,0 +1,75 @@
> ++// SPDX-License-Identifier: GPL-2.0
> ++// Copyright (c) 2018, The Linux Foundation. All rights reserved.
> ++
> ++#include "qcom-ipq4019.dtsi"
> ++#include <dt-bindings/input/input.h>
> ++#include <dt-bindings/gpio/gpio.h>
I noticed that you included input.h put none of these reference
boards actually define any gpio-keys.

So, what's the reason here? Are these DTS actually for real "boards" or
are they just for show? (Also related: why is there no system/power LED
defined?)

> ++/ {
> ++	model = "Qualcomm Technologies, Inc. IPQ4019/AP-DK07.1";
> ++
> ++	memory {
> ++		device_type = "memory";
> ++		reg = <0x80000000 0x20000000>; /* 512MB */
> ++	};
> ++
> ++	aliases {
> ++		serial0 = &blsp1_uart1;
> ++		serial1 = &blsp1_uart2;
(The alias node and chosen node was removed later in a separate
patch.)
> ++	};
> ++
> ++	chosen {
> ++		stdout-path = "serial0:115200n8";
> ++	};
> ++
> ++	soc {
> ++		pinctrl at 1000000 {
> ++			serial_0_pins: serial0-pinmux {
> ++				pins = "gpio16", "gpio17";
> ++				function = "blsp_uart0";
> ++				bias-disable;
> ++			};
> ++
> ++			i2c_0_pins: i2c-0-pinmux {
> ++				pins = "gpio20", "gpio21";
> ++				function = "blsp_i2c0";
> ++				bias-disable;
> ++			};
> ++
> ++			nand_pins: nand-pins {
> ++				pins = "gpio53", "gpio55", "gpio56",
> ++				       "gpio57", "gpio58", "gpio59",
> ++				       "gpio60", "gpio62", "gpio63",
> ++				       "gpio64", "gpio65", "gpio66",
> ++				       "gpio67", "gpio68", "gpio69";
> ++				function = "qpic";
> ++                        };
> ++		};
bias-pull-ups and bias-pull-downs are missing here.

Also what happend to gpio52? The "qcom-ipq4019-ap.dk04.1.dtsi" still
defines it and from what I can tell, there seems to be only one valid
way to wire up a nand chip with the SoC.

> ++
> ++		serial at 78af000 {
> ++			pinctrl-0 = <&serial_0_pins>;
> ++			pinctrl-names = "default";
> ++			status = "ok";
> ++		};
I guess it would have been nice to use the label here.

&tlmm {
..
};

&blsp_dma {

};

&blsp1_spi1 {

};

&blsp1_uart1 {
..
};



but no biggy.

> ++
> ++		dma at 7884000 {
> ++			status = "ok";
> ++		};
> ++
> ++		i2c at 78b7000 { /* BLSP1 QUP2 */
> ++			pinctrl-0 = <&i2c_0_pins>;
> ++			pinctrl-names = "default";
> ++			status = "ok";
> ++		};
> ++
> ++		dma at 7984000 {
> ++			status = "ok";
> ++		};
> ++
> ++		qpic-nand at 79b0000 {
> ++			pinctrl-0 = <&nand_pins>;
> ++			pinctrl-names = "default";
> ++			status = "ok";
> ++		};
> ++	};
> ++};
> +-- 
> +1.9.1
> +
> diff --git a/target/linux/ipq40xx/patches-4.14/903-ARM-dts-ipq4019-Add-qcom-ipq4019-ap.dk07.1-c1-board-.patch b/target/linux/ipq40xx/patches-4.14/903-ARM-dts-ipq4019-Add-qcom-ipq4019-ap.dk07.1-c1-board-.patch
> new file mode 100644
> index 0000000..655a372
> --- /dev/null
> +++ b/target/linux/ipq40xx/patches-4.14/903-ARM-dts-ipq4019-Add-qcom-ipq4019-ap.dk07.1-c1-board-.patch
> @@ -0,0 +1,99 @@
> +From 5ade893ec03c6612a38d53eaf4324f21e18eeef5 Mon Sep 17 00:00:00 2001
> +From: Sricharan R <sricharan at codeaurora.org>
> +Date: Fri, 25 May 2018 11:41:18 +0530
> +Subject: [PATCH] ARM: dts: ipq4019: Add qcom-ipq4019-ap.dk07.1-c1 board file
> +
> +Reviewed-by: Abhishek Sahu <absahu at codeaurora.org>
> +Signed-off-by: Sricharan R <sricharan at codeaurora.org>
> +Signed-off-by: Andy Gross <andy.gross at linaro.org>
> +---
> + arch/arm/boot/dts/Makefile                      |  1 +
> + arch/arm/boot/dts/qcom-ipq4019-ap.dk07.1-c1.dts | 64 +++++++++++++++++++++++++
> + 2 files changed, 65 insertions(+)
> + create mode 100644 arch/arm/boot/dts/qcom-ipq4019-ap.dk07.1-c1.dts
> +
> +diff --git a/arch/arm/boot/dts/Makefile b/arch/arm/boot/dts/Makefile
> +index 2c002b3..eede5df 100644
> +--- a/arch/arm/boot/dts/Makefile
> ++++ b/arch/arm/boot/dts/Makefile
> +@@ -708,6 +708,7 @@
> + 	qcom-ipq4019-ap.dk01.1-c1.dtb \
> + 	qcom-ipq4019-a62.dtb \
> + 	qcom-ipq4019-ap.dk04.1-c1.dtb \
> ++	qcom-ipq4019-ap.dk07.1-c1.dtb \
> + 	qcom-ipq4028-wpj428.dtb \
> + 	qcom-ipq4029-gl-b1300.dtb \
> + 	qcom-ipq4029-mr33.dtb \
> +diff --git a/arch/arm/boot/dts/qcom-ipq4019-ap.dk07.1-c1.dts b/arch/arm/boot/dts/qcom-ipq4019-ap.dk07.1-c1.dts
> +new file mode 100644
> +index 0000000..8c7ef65
> +--- /dev/null
> ++++ b/arch/arm/boot/dts/qcom-ipq4019-ap.dk07.1-c1.dts
Same as with the .dtsi: please push those changes upstream.

> +@@ -0,0 +1,64 @@
> ++// SPDX-License-Identifier: GPL-2.0
> ++// Copyright (c) 2018, The Linux Foundation. All rights reserved.
> ++
> ++#include "qcom-ipq4019-ap.dk07.1.dtsi"
> ++
> ++/ {
> ++	model = "Qualcomm Technologies, Inc. IPQ4019/AP-DK07.1-C1";
> ++	compatible = "qcom,ipq4019-ap-dk07.1-c1";
Make that:

compatible = "qcom,ipq4019-ap-dk07.1-c1", "qcom,ipq4019";
> ++
> ++	soc {
> ++		pci at 40000000 {
> ++			status = "ok";
> ++			perst-gpio = <&tlmm 38 0x1>;
> ++		};
The phandle's 3'rd parameter should be GPIO_ACTIVE_LOW and not 0x1. 

> ++
> ++		spi at 78b6000 {
> ++			status = "ok";
> ++		};
> ++
> ++		pinctrl at 1000000 {
> ++			serial_1_pins: serial1-pinmux {
> ++				pins = "gpio8", "gpio9",
> ++					"gpio10", "gpio11";
> ++				function = "blsp_uart1";
> ++				bias-disable;
> ++			};
> ++
> ++			spi_0_pins: spi-0-pinmux {
> ++				pinmux {
> ++					function = "blsp_spi0";
> ++					pins = "gpio13", "gpio14", "gpio15";
> ++					bias-disable;
> ++				};
> ++				pinmux_cs {
> ++					function = "gpio";
> ++					pins = "gpio12";
> ++					bias-disable;
> ++					output-high;
> ++				};
> ++			};
> ++		};
> ++
> ++		serial at 78b0000 {
> ++			pinctrl-0 = <&serial_1_pins>;
> ++			pinctrl-names = "default";
> ++			status = "ok";
> ++		};
> ++
> ++		spi at 78b5000 {
> ++			pinctrl-0 = <&spi_0_pins>;
> ++			pinctrl-names = "default";
> ++			status = "ok";
> ++			cs-gpios = <&tlmm 12 0>;
that 0 would be GPIO_ACTIVE_HIGH (altough, the spi subsystem does not
care about high or low there. It has its own: "spi-cs-high; " property
for this purpose.

> ++
> ++			m25p80 at 0 {
> ++				#address-cells = <1>;
> ++				#size-cells = <1>;
> ++				reg = <0>;
> ++				compatible = "n25q128a11";
I think "jedec,spi-nor" is missing here. It must be included for
any SPI NOR flash that supports the JEDE READ ID Opcode.

> ++				spi-max-frequency = <24000000>;
> ++			};
> ++		};
> ++	};
> ++};
> +-- 
> +1.9.1
> +
> diff --git a/target/linux/ipq40xx/patches-4.14/904-Fix-rootfs-conflict-with-OpenWrt-auto-mount.patch b/target/linux/ipq40xx/patches-4.14/904-Fix-rootfs-conflict-with-OpenWrt-auto-mount.patch
> new file mode 100644
> index 0000000..92b4b3e
> --- /dev/null
> +++ b/target/linux/ipq40xx/patches-4.14/904-Fix-rootfs-conflict-with-OpenWrt-auto-mount.patch
> @@ -0,0 +1,40 @@
> +From ca96c645d61b4462fe6ee3d815717335214f27c2 Mon Sep 17 00:00:00 2001
> +From: Ram Chandra Jangir <rjangir at codeaurora.org>
> +Date: Sat, 9 Jun 2018 20:31:50 +0530
> +Subject: [PATCH] Fix rootfs conflict with OpenWrt auto mounting
> +
> +IPQ4019 nand flash has ubi partition name as 'rootfs', when we
> +boot IPQ4019 ap-dk07.1-c1 board from nand flash(with ubi image),
> +OpenWrt tries to attach mtd device named "ubi" on boot and it
> +fails.
> +Hence we rename this partition name from 'rootfs' to 'ubi' to
> +get it auto mounted.
> +
> +Signed-off-by: Ram Chandra Jangir <rjangir at codeaurora.org>
Really?

Try setting "ubi.mtd" bootargs:
<http://www.linux-mtd.infradead.org/faq/ubi.html#L_attachmtd>
It even comes with a fitting example:
"ubi.mtd=rootfs"

> +---
> + drivers/mtd/mtdpart.c | 9 ++++++---
> + 1 files changed, 6 insertions(+), 3 deletions(-)
> +
> +diff --git a/drivers/mtd/mtdpart.c b/drivers/mtd/mtdpart.c
> +index c87b71c..4980f93 100644
> +--- a/drivers/mtd/mtdpart.c
> ++++ b/drivers/mtd/mtdpart.c
> +@@ -887,9 +887,12 @@ static void mtd_partition_split(struct mtd_info *master, struct mtd_part *part)
> + 		return;
> + 
> + 	if (!strcmp(part->mtd.name, "rootfs")) {
> +-		run_parsers_by_type(part, MTD_PARSER_TYPE_ROOTFS);
> +-
> +-		rootfs_found = 1;
> ++		if(mtd_type_is_nand(master)) {
> ++			part->mtd.name="ubi";
> ++		} else {
> ++			run_parsers_by_type(part, MTD_PARSER_TYPE_ROOTFS);
> ++			rootfs_found = 1;
> ++		}
> + 	}
> + 
> + 	if (!strcmp(part->mtd.name, SPLIT_FIRMWARE_NAME) &&
> +-- 
> +2.7.2
> +
> diff --git a/target/linux/ipq40xx/patches-4.14/905-dts-qcom-ipq4019-ap.dk07.1-enable-nodes.patch b/target/linux/ipq40xx/patches-4.14/905-dts-qcom-ipq4019-ap.dk07.1-enable-nodes.patch
> new file mode 100644
> index 0000000..9a32719
> --- /dev/null
> +++ b/target/linux/ipq40xx/patches-4.14/905-dts-qcom-ipq4019-ap.dk07.1-enable-nodes.patch
> @@ -0,0 +1,79 @@
> +--- a/arch/arm/boot/dts/qcom-ipq4019-ap.dk07.1-c1.dts
> ++++ b/arch/arm/boot/dts/qcom-ipq4019-ap.dk07.1-c1.dts
> +@@ -5,7 +5,7 @@
> + 
> + / {
> + 	model = "Qualcomm Technologies, Inc. IPQ4019/AP-DK07.1-C1";
> +-	compatible = "qcom,ipq4019-ap-dk07.1-c1";
> ++	compatible = "qcom,ap-dk07.1-c1", "qcom,ipq4019";
"qcom,ipq4019-ap-dk07.1-c1" should be staying.

> + 
> + 	soc {
> + 		pci at 40000000 {
> +--- a/arch/arm/boot/dts/qcom-ipq4019-ap.dk07.1.dtsi
> ++++ b/arch/arm/boot/dts/qcom-ipq4019-ap.dk07.1.dtsi
> +@@ -13,15 +13,6 @@
> + 		reg = <0x80000000 0x20000000>; /* 512MB */
> + 	};
> + 
> +-	aliases {
> +-		serial0 = &blsp1_uart1;
> +-		serial1 = &blsp1_uart2;
> +-	};
> +-
> +-	chosen {
> +-		stdout-path = "serial0:115200n8";
> +-	};
> +-
Sad to see those nodes go. Altough serial1 was not enabled.
So why was it there in the first place? c&p?

> + 	soc {
> + 		pinctrl at 1000000 {
> + 			serial_0_pins: serial0-pinmux {
> +@@ -67,9 +71,49 @@
> + 		};
> + 
> + 		qpic-nand at 79b0000 {
> ++			compatible = "qcom,ipq4019-nand", "qcom,msm-nand";
> + 			pinctrl-0 = <&nand_pins>;
> + 			pinctrl-names = "default";
> + 			status = "ok";
> ++
> ++			nand at 0 {
> ++				compatible = "qcom,nandcs";
> ++				reg = <0>;
> ++				#address-cells = <1>;
> ++				#size-cells = <1>;
> ++
> ++				nand-ecc-strength = <4>;
> ++				nand-ecc-step-size = <512>;
> ++				nand-bus-width = <8>;
Is u-boot filling in the partitions here?

> ++			};
> ++		};
> ++
> ++		mdio at 90000 {
> ++			status = "okay";
> ++		};
> ++
> ++		ess-switch at c000000 {
> ++				status = "okay";
> ++				switch_mac_mode = <0x0>; /* mac mode for RGMII RMII */
The comment (RGMII/RMII mode) is either wrong or the
switch_mac_mode value is misleading.

> ++				switch_initvlas = <
> ++					0x0007c 0x54 /* PORT0_STATUS */
> ++				>;
This already is set? Also, the googlized ess-switch ar40xx.c driver we
are stuck with does not parse this property.

> ++			};
> ++
> ++		ess-psgmii at 98000 {
> ++			status = "okay";
> ++		};
> ++
> ++		edma at c080000 {
> ++			status = "okay";
> ++		};
> ++
> ++		wifi at a000000 {
> ++			status = "okay";
> ++		};
> ++
> ++		wifi at a800000 {
> ++			status = "okay";
> + 		};
> + 	};
> + };
> 

OT: A user reported a kernel panic with the IPQ4019 ap-dk01.1-c1 on
the forum: <https://forum.openwrt.org/t/ipq4018-dk01-1-kernel-panic/20060>
I wonder if this affects the  ap-dk04.1-c1 and  ap-dk07.1-c1 as well.
Can you please comment on that.




_______________________________________________
openwrt-devel mailing list
openwrt-devel at lists.openwrt.org
https://lists.openwrt.org/mailman/listinfo/openwrt-devel



More information about the openwrt-devel mailing list