ramips: add initial support for TP-LINK TL-WR840N v4 #974

svobodavac · 2017-03-21T11:39:43Z

This patch adds initial support for wireless router TP-LINK TL-WR840N v4.

Specification:

SoC: MediaTek MT7628NN (580MHz)
RAM: 64 MiB (Winbond W9751G6JB-25)
Flash: 8 MiB (GigaDevice GD25Q64CSIG)
LAN: 4x 100M
WAN: 1x 100M
Antennas: 2x non-detachable
LED: 5x green
Buttons: 1x reset

Instalation:
TP-Link firmware uses recent firmware v3 header, please see
https://github.com/xdarklight/mktplinkfw3.
Therefore direct upgrade from TP-Link web UI is not supported yet.

Upgrade from the TP-Link firmware is possible using common TFTP recovery mode.
Rename file lede-ramips-mt7628-tl-wr840n-v4-squashfs-factory.bin
to tp_recovery.bin and put it on a TFTP server with IP 192.168.0.66.
Hold reset buton when connecting power until TFTP transfer begins.

For upgrade from LEDE use generic sysupgrade method.

LEDE uses uImage firmware header instead of the TP-Link v3 firmware
header. In order to recover the original TP-Link firmware please
use TFTP recovery mode. The header has to be removed from
the original firmware:
dd if=TL-WR840Nv4_0.9.1_3.16_up_boot(161011).bin
bs=512 skip=1 > tp_recovery.bin

Currently not working:
The only LAN LED indicates activity only on LAN port 2. The LAN LED
seems to be hardwired to the internal switch, port 2.

Signed-off-by: Vaclav Svoboda svoboda@neng.cz

heinzek

The only LAN LED indicates activity only on LAN port 2. The LAN LED
seems to be hardwired to the internal switch, port 2.

Hi. Did You try fix it by adding "ephy" or "p0led_an", "p1led_an", "p2led_an", "p3led_an", "p4led_an", to pinctrl0?

heinzek · 2017-03-21T14:40:21Z

target/linux/ramips/dts/TL-WR840NV4.dts

+			gpios = <&gpio1 6 GPIO_ACTIVE_LOW>;
+			linux,code = <KEY_RESTART>;
+		};
+


Drop this unnecessary line.

heinzek · 2017-03-21T15:14:23Z

target/linux/ramips/dts/TL-WR840NV4.dts

+		};
+
+		ethernet {
+			label = "tl-wr840n-v4:green:ethernet";


I think better name for this led is lan.

I do agree. User gude contains description "ethernet", but fixed.

heinzek · 2017-03-21T15:14:44Z

target/linux/ramips/dts/TL-WR840NV4.dts

+		};
+
+		internet {
+			label = "tl-wr840n-v4:green:internet";


and here wan

heinzek · 2017-03-21T15:18:29Z

target/linux/ramips/dts/TL-WR840NV4.dts

+/dts-v1/;
+
+#include <dt-bindings/input/input.h>
+#include <dt-bindings/gpio/gpio.h>


wrong alphabetical first gpio.h than input.h

heinzek · 2017-03-21T15:46:20Z

target/linux/ramips/base-files/etc/board.d/01_leds

+	set_wifi_led "$board:green:wireless"
+	ucidef_set_led_netdev "lan" "lan" "$board:green:ethernet" eth0.1
+	ucidef_set_led_netdev "wan" "wan" "$board:green:internet" eth0.2
+	ucidef_set_led_default "wps" "wps" "$board:green:wps" "0"


Drop ucidef_set_led_default here. They shouldn't be required.

svobodavac · 2017-03-22T13:41:51Z

Hi. Did You try fix it by adding "ephy" or "p0led_an", "p1led_an", "p2led_an", "p3led_an", "p4led_an", to pinctrl0?

I have tried all of them - ephy and pxled_an.
It seems the p2led_an is the right one for MT7628AN, port2:

&pinctrl {
         state_default: pinctrl0 {
                gpio {
                        ralink,group = "i2c", "i2s", "uart1", "uart2", "p2led_an" ;
                        ralink,function = "gpio";
                };
        };
 };

I'm unable to switch pins 39-43 to gpio and manually control LAN LEDS.
Wifi LED (pin 44, wled_an) is probably set to gpio and I can turn the LED on/off.

Please, any idea?

Update 2017/03/24:
I'm sorry, my stupid fault.
I used incorrect image for upgrade.
I can confirm that it is possible to switch pin p2led_an to GPIO mode and cotrol the LED manually.

The same the pin p0led_an for the WAN LED.
Both LAN and WAN LEDs are working now.

mkresin · 2017-03-23T09:46:36Z

target/linux/ramips/image/mt7628.mk

+  IMAGE_SIZE := 7808k
+  DEVICE_TITLE := TP-Link TL-WR840N v4
+  DEVICE_PACKAGES :=
+  KERNEL := $(KERNEL_DTB) | prepend-to 448 | uImage lzma


please use the existing pad-extra instead of your custom prepend-to.

pad-extra 448 | $(KERNEL_DTB) | uImage lzma

I'm afraid it's not possible.
KERNEL_DTB = kernel-bin | patch-dtb | lzma
Calling pad-extra before making KERNEL_DTB causes, that all leading zeroes crated by pad-extra are subject of lzma a will be comprimed:(
I can't find better/simpler solution.

This patch adds initial support for wireless router TP-LINK TL-WR840N v4. Specification: - SoC: MediaTek MT7628NN (580MHz) - RAM: 64 MiB (Winbond W9751G6JB-25) - Flash: 8 MiB (GigaDevice GD25Q64CSIG) - LAN: 4x 100M - WAN: 1x 100M - Antennas: 2x non-detachable - LED: 5x green - Buttons: 1x reset Instalation: TP-Link firmware uses recent firmware v3 header, please see https://github.com/xdarklight/mktplinkfw3. Therefore direct upgrade from TP-Link web UI is not supported yet. Upgrade from the TP-Link firmware is possible using common TFTP recovery mode. Rename file lede-ramips-mt7628-tl-wr840n-v4-squashfs-factory.bin to tp_recovery.bin and put it on a TFTP server with IP 192.168.0.66. Hold reset buton when connecting power until TFTP transfer begins. For upgrade from LEDE use generic sysupgrade method. LEDE uses uImage firmware header instead of the TP-Link v3 firmware header. In order to recover the original TP-Link firmware please use TFTP recovery mode. The header has to be removed from the original firmware: dd if=TL-WR840Nv4_0.9.1_3.16_up_boot$161011$.bin \ bs=512 skip=1 > tp_recovery.bin Signed-off-by: Vaclav Svoboda <svoboda@neng.cz>

mkresin · 2017-03-27T07:54:58Z

Totally missed that you managed to fix the LED issue. Better add a new comment next time so that everyone gets noticed.

I'm still struggling to understand why this extra padding is required. What I might to understand so far is that the 128KByte extra padding of the factory image is used to ensure that the tftp recovery doesn't overwrite the bootloader.

The 448 bytes extra padding is required since you don't create a valid tp-link header, right?

Assuming that the bootloader works the same way as the lantiq tp-link uboots, the padding can be dropped by using a valid tp-link header as it is done in target/linux/lantiq/image/tp-link.mk. It still doesn't create a valid v3 tp-link header that is accepted by the tp-link webinterface, but it creates a tp-link header where the update u-boot flag is unchecked.

Any chance that you give it a try or are you done with it?

Edit: To get the firmware splitter working with the TP-Link header, you need to add

CONFIG_MTD_SPLIT_TPLINK_FW=y

to target/linux/ramips/mt7628/config-4.4/target/linux/ramips/mt7628/config-4.9.

svobodavac · 2017-03-27T09:54:30Z

@mkresin
Extra padding:

Uboot TFTP recovery mode takes all the firmware file, but it strips TP-Link header (size of 0x200 B - 512 B) and boot partition (size of 0x20000 B - 128 kB). Then it writes rest of the firmware file to the firmware partition (including second TP-Link partition).
So we need prepend fake TP-Link header and fake boot partition to factory LEDE image.
It seems the uBoot TFTP recovery mode doesn't rewrite the boot partition. (I can confirm this behavior just for current bootloader version;-))
When we create the entire LEDE firmware we use uImage header (64B) instead of TP-Link header (512 B) in order to better parse firmware and to find the beginning of the root_fs.
It also help us to provide sysupgrade using uImage a standard way.
We have to put extra padding (512-64B) after the uImage header to ensure the code entry point is on address where the bootloadrer expects it.
Yes, I can give it a try to use the TP-Link header. Thank you for the hint with TP-Link MTD splitting.

mkresin · 2017-04-27T12:12:01Z

@svobodavac any update on using the TP-Link header instead of faking it? Would you please rebase your PR to fix the conflicting files.

sanitariu · 2017-05-04T09:13:01Z

So far router seems to work. I found only one major problem. This is wifi txpower. Seems it is working on 25mw or sort of power which is too low. You loose the device in the next room.
Anyone knows a way to make the txpower working normal like 50-100mw ?

root@lede-router:~# iwinfo wlan0 info
wlan0 ESSID: "LEDE-ROUTER"
Access Point: 18:D6:C7:DF:12:46
Mode: Master Channel: 5 (2.432 GHz)
Tx-Power: unknown Link Quality: unknown/70
Signal: unknown Noise: unknown
Bit Rate: unknown
Encryption: mixed WPA/WPA2 PSK (CCMP)
Type: nl80211 HW Mode(s): 802.11bgn
Hardware: unknown [Generic MAC80211]
TX power offset: unknown
Frequency offset: unknown
Supports VAPs: yes PHY name: phy0

root@lede-router:~# iwinfo wlan0 txpowerlist
0 dBm ( 1 mW)
1 dBm ( 1 mW)
2 dBm ( 1 mW)
3 dBm ( 1 mW)
4 dBm ( 2 mW)
5 dBm ( 3 mW)
6 dBm ( 3 mW)
7 dBm ( 5 mW)
8 dBm ( 6 mW)
9 dBm ( 7 mW)
10 dBm ( 10 mW)
11 dBm ( 12 mW)
12 dBm ( 15 mW)
13 dBm ( 19 mW)
14 dBm ( 25 mW)
15 dBm ( 31 mW)
16 dBm ( 39 mW)
17 dBm ( 50 mW)
18 dBm ( 63 mW)
19 dBm ( 79 mW)
20 dBm ( 100 mW)

Power selection is missing.

mkresin · 2017-05-04T18:28:06Z

I will close this Pull Request due to inactivity. Please reopen the PR if you have the time to fix the outlined issues.

fabriciopirini · 2017-05-06T17:27:31Z

oh boy ... It's a shame I know nothing abt this so I could help.I was looking foward to it because i bought this router 3 weeks ago w/o knowing it has no custom firmwares available =/ ... Well, I hope some good guy can keep this request going ... I'm cheering fo it =)

dim45 · 2017-05-15T04:18:03Z

help error "git apply 974.patch"

raxpyraxp · 2017-05-15T10:11:53Z

target/linux/ramips/dts/TL-WR840NV4.dts

+
+&wmac {
+	status = "okay";
+	mtd-mac-address = <&factory 0xf100>;


I'm trying to investigate poor wireless performance, which was described by @sanitariu
The thing I noticed here is that the EEPROM address was not defined, so it uses default value from mt7628an.dtsi, which is wrong for this hardware:
mediatek,mtd-eeprom = <&factory 0x0000>;
Real EEPROM is located in radio partition, so we must declare this:

&spi0 { ... m25p80@0 { ... radio: partition@07f0000 { ... &wmac { ... mediatek,mtd-eeprom = <&radio 0x0000>; };

But performance is still bad though.

Did you make some patch for the correct eeprom ?

You are the author of this https://raxp.ru/index.php/2017/06/21/lede-for-wr841n-v13wr840n-v4-ru/ ?
Can share the developments?

9000000 · 2017-06-10T03:35:24Z

Can you give me the link to download the firmware

gleidsonhenriques · 2017-06-14T17:14:14Z

I'm also interested in this firmware.

sanitariu · 2017-06-15T07:25:58Z

No link. Just get the source and compile. Wifi performance is really bad so do not use in production.
I tried to patch as worm202 pointed but no luck. Also tried other stuff from .dtsi but no luck.

gleidsonhenriques · 2017-06-22T16:12:54Z

@sanitariu Overall performance or distance performance ?

f85ac21de1 · 2017-06-23T22:27:45Z

@sanitariu im interested in computing performance not Wifi, how can i compile it?

sanitariu · 2017-06-28T11:25:39Z

Its not so hard. Just clone the code. Do make menuconfig choose the arch you like and then compile with make. Output files will be in bin folder.
About 840 ver.4 wifi - problem is with TX power. It's too low i think default like 15-30mw or internal antenna working only (distance performance).

alangregory · 2017-07-03T15:22:42Z

@sanitariu Do you think wifi performance is related to TX power problem ?

alangregory · 2017-07-03T15:26:19Z

Support for 840Nv4 and 841Nv13 was added recently , please check 24043a0 . I'm testing some wireless performance problems. It would be of great help if ppl could test and give feedback.

sanitariu · 2017-07-04T09:51:58Z

Wifi is still bad. The command "iwinfo wlan0 txpowerlist" does not show actual txpower but i think it is set default 30 or 50mw. I compared to tp-link 740 version 4 which is 63mw and i see it stronger in the next room.

mkresin · 2017-07-04T11:53:13Z

Please use the LEDE forum for any further discussion or send a mail to the LEDE devel mailing list.

This PR was never merged, support for the TL-WR840N was added with 24043a0.

heinzek reviewed Mar 21, 2017

View reviewed changes

pepe2k added the ramips label Mar 22, 2017

svobodavac force-pushed the TP-LINK_TL-WR840NV4_support branch from d52dff0 to f9afa6e Compare March 22, 2017 13:06

mkresin requested changes Mar 23, 2017

View reviewed changes

svobodavac force-pushed the TP-LINK_TL-WR840NV4_support branch from f9afa6e to 8706b60 Compare March 24, 2017 11:28

sanitariu approved these changes Apr 27, 2017

View reviewed changes

mkresin self-assigned this Apr 27, 2017

mkresin closed this May 4, 2017

raxpyraxp reviewed May 15, 2017

View reviewed changes

lede-project locked and limited conversation to collaborators Jul 4, 2017

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

ramips: add initial support for TP-LINK TL-WR840N v4 #974

ramips: add initial support for TP-LINK TL-WR840N v4 #974

svobodavac commented Mar 21, 2017

heinzek left a comment •

edited

heinzek Mar 21, 2017

svobodavac Mar 22, 2017

heinzek Mar 21, 2017

svobodavac Mar 22, 2017

heinzek Mar 21, 2017

svobodavac Mar 22, 2017

heinzek Mar 21, 2017

svobodavac Mar 22, 2017

heinzek Mar 21, 2017

svobodavac Mar 22, 2017

svobodavac commented Mar 22, 2017 •

edited

mkresin Mar 23, 2017

svobodavac Mar 24, 2017

mkresin commented Mar 27, 2017 •

edited

svobodavac commented Mar 27, 2017

mkresin commented Apr 27, 2017

sanitariu commented May 4, 2017 •

edited

mkresin commented May 4, 2017

fabriciopirini commented May 6, 2017 •

edited

dim45 commented May 15, 2017

raxpyraxp May 15, 2017

sanitariu May 19, 2017

Lipton2kirov Jul 4, 2017

9000000 commented Jun 10, 2017

gleidsonhenriques commented Jun 14, 2017

sanitariu commented Jun 15, 2017

gleidsonhenriques commented Jun 22, 2017

f85ac21de1 commented Jun 23, 2017

sanitariu commented Jun 28, 2017

alangregory commented Jul 3, 2017

alangregory commented Jul 3, 2017

sanitariu commented Jul 4, 2017

mkresin commented Jul 4, 2017 •

edited

ramips: add initial support for TP-LINK TL-WR840N v4 #974

ramips: add initial support for TP-LINK TL-WR840N v4 #974

Conversation

svobodavac commented Mar 21, 2017

heinzek left a comment • edited

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

svobodavac commented Mar 22, 2017 • edited

Choose a reason for hiding this comment

Choose a reason for hiding this comment

mkresin commented Mar 27, 2017 • edited

svobodavac commented Mar 27, 2017

mkresin commented Apr 27, 2017

sanitariu commented May 4, 2017 • edited

mkresin commented May 4, 2017

fabriciopirini commented May 6, 2017 • edited

dim45 commented May 15, 2017

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

9000000 commented Jun 10, 2017

gleidsonhenriques commented Jun 14, 2017

sanitariu commented Jun 15, 2017

gleidsonhenriques commented Jun 22, 2017

f85ac21de1 commented Jun 23, 2017

sanitariu commented Jun 28, 2017

alangregory commented Jul 3, 2017

alangregory commented Jul 3, 2017

sanitariu commented Jul 4, 2017

mkresin commented Jul 4, 2017 • edited

heinzek left a comment •

edited

svobodavac commented Mar 22, 2017 •

edited

mkresin commented Mar 27, 2017 •

edited

sanitariu commented May 4, 2017 •

edited

fabriciopirini commented May 6, 2017 •

edited

mkresin commented Jul 4, 2017 •

edited