Just a quick note about the LAN LED behavior while I have this in mind. As I previously mentioned, the LAN led doesn't exhibit the expected behavior once the eth0 link is up. I inspected many things, including the Marvell PHY driver "/drivers/net/phy/marvell.c". It seems that its default config is to use the LED[0] as the 1000M link LED, and it is indeed the case here as the LEby Robin - Debian
U-Boot seems to work fine. As the envs are not finished yet it failed to boot on HDD and booted the stock kernel from SPI flash but otherwise it boots without major issues. robin@raketa:~$ sudo kwboot -t -B 115200 /dev/ttyUSB0 -b uboot.2024.04-tld-1.ds216.kwb -p Mot de passe de robinĀ : kwboot version 2023.01 Patching image boot signature to UART Sending boot message. Please reboot thby Robin - Debian
Hi Tony, Yes, the goal was to extend nodes already declared in armada-38x.dtsi instead of re-declaring them to avoid warning messages during the dtb compilation. As for the smaller changes: - There is no EEPROM in the ds216, and it may refer to non-existing hardware in the ds116 too. If that's the case, the whole i2c node could be removed for the ds116. - I bumped the i2c frequency toby Robin - Debian
Hi Tony, I've been busy for the last few days, hopefully I should have more time to try and improve things now. Wow, it seems like you made a lot of progress on the u-boot side, I'll give it a shot. Digging deeper in the Synology GPL tarballs, I get more confused about this code in the DS216 DTS : /* use only one pin for UART1, as mpp20 is used by sata0 */ uart1_pins: uart-by Robin - Debian
Those GPIOs turn their respective SATA bays properly under the stock u-boot. However, even though both drives are powered, only drives in the Disk 2 bay will be detected by u-boot and thus be bootable. I think that the stock u-boot SATA config is just bad : - Setting GPIO 26 and 27 HIGH powers on both drives under u-boot. - Drives in Disk 1 bay are not detected no matter what. - Drives in Disby Robin - Debian
Hi Tony, I'm currently preparing a complete guide that I will post here in a few days. There are few other minor quirks I didn't mention previously that I will sum up in it, such as the fact that the stock u-boot only detects drives in the DISK 2 bay, even with an appropriate config. I hope you didn't missed my latest DTS edit (for the Seiko RTC) as the message landed at the vby Robin - Debian
It just works I followed your procedure with the stock kwboot of my debian 12 PC and it worked fine. I had to keep the power button pressed until kwboot stated loading. Any other way to power on the NAS ended up in a normal boot, ignoring kwboot. robin@raketa:~$ sudo kwboot -t -B 115200 /dev/ttyUSB0 -b mtd0.ds216 -p Mot de passe de robinĀ : kwboot version 2023.01 Patching image bootby Robin - Debian
UPDATE Even though the MVEBU RTC works fine, the system is intended to work with the dedicated SEIKO S35390A I2C RTC chip. Only the SEIKO RTC can wake the system. I edited the DTS to enable the use of the SEIKO RTC and disable de MVEBU RTC by default.by Robin - Debian
It works perfectly. root@debian:~# cat /proc/mtd dev: size erasesize name mtd0: 00100000 00001000 "u-boot" mtd1: 006e0000 00001000 "kernel" mtd2: 00010000 00001000 "u-boot-env" mtd3: 00010000 00001000 "data" root@debian:~# cat /etc/fw_env.config # MTD device name Device offset Env. size Flash sector size Number of sectors # DS216 /devby Robin - Debian
Hi everyone, Just a quick message to say that I dug a bit into the DS214 DTS and found several issues. The most important one probably being the RAM set to 1GB instead of 512MB. I'll fix all I can and do some testing to validate my changes, and post a revised DTS later on this tread. I also started a thread for mainline u-boot support on this device.by Robin - Debian
Hi everyone, As some may know, I recently tinkered on the Synology DS216 and managed to get Debian up and running on it. I also own a Synology DS214 but I didn't do much on it until now as i was unable to get the UART interface to work, until I got it to... UART TTL Someone already got Debian to run on the DS214 here and properly described the UART interface pinout. It happens to beby Robin - uBoot
Hi Tony, I reverted most of the changes I've done when I switched to my latest DTS as those were no more required. In the end, all the changes I made persistent are the following ones: setenv load_image_addr 0x02000000 setenv load_initrd_addr 0x3000000 setenv load_image 'echo loading uImage from raw HDD ...; scsi device 0; scsi read $load_image_addr 0x5000 0x2F00' setenvby Robin - Debian
As I did a few edits to my uboot environment, I performed a resetenv before getting the following log: BootROM - 1.73 Booting from SPI flash General initialization - Version: 1.0.0 AVS selection from EFUSE disabled (Skip reading EFUSE values) Overriding default AVS value to: 0x23 Detected Device ID 6820 High speed PHY - Version: 2.0 Init Synology board board SerDes lanes topologyby Robin - Debian
SO FAR, EVERYTHING WORKS I made a lot of minor edits to my DTS file, the latest one is attached to this post. I listened to your advice and tested the fan speeds carefully, LSB and MSB were swapped so I fixed it in no time. Everything seems to work fine when it comes to hwmon, and even though the LED config is not the way we would like it to be, it's probably the most convenient we canby Robin - Debian
LEDs and GPIOs The green disk activity LEDs are hardware triggered, I previously said otherwise as I was confused by the GPIO logic to enable it. The only thing is that those are not enabled by default, but setting the pin high in u-boot env seems like a simple persistent solution even under debian. By default the stock u-boot enables 3 GPIOs: - The fan speed LSB to spin the fan at the lowby Robin - Debian
Hi Tony, Thanks for your comments and tips. I spent most of my day refining my DTS and discovered a few things on the way... EEPROM For my first DTS, the one I shared previously, I foolishly copy-pasted some bindings from other DTS files I based my work on in the first place, including: i2c@11000 { pinctrl-names = "default"; pinctrl-0 = <&aby Robin - Debian
GOOD NEWS After spending all day refining my DTS, almost everything works properly. USB 2.0 and 3.0 are both fixed, SATA ports are powered, SPI NOR flash works at a much higher frequency thanks to FAST READ. There is now only one ethernet interface, and many other little things have been improved significantly. Here are my current DTS and bootlog, if you find something i may have missed i'by Robin - Debian
HWMON I think most of the issue is related to GPIO conflicts and/or bad definitions. I'll get back into this when debian will run a kernel with my edited DTB. LEDs AND GPIOs Okay so after a bit of trial and error, and time spent digging into the Synology GPL tarballs (especially synobios.txz), I finally determined the GPIO config for the DS216. I'll edit the DTS accordingly andby Robin - Debian
RTC It was indeed a non-issue. After setting the date and time, a full unplug and a fresh boot, hwclock returns the correct date and time. HWmon root@debian:~# ls /sys/class/hwmon/hwmon hwmon0/ hwmon1/ root@debian:~# ls /sys/class/hwmon/ . .. hwmon0 hwmon1 root@debian:~# ls /sys/class/hwmon/hwmon0 . device of_node subsystem temp1_input uevent .. name powerby Robin - Debian
Okay, thank you all for the hints. If it should be quite easy for you to compile u-boot for the DS216, that is some really good news ! I started to look at the DTS of the DS218j and DS116 and found something i'm not sure about. In "armada-385-synology-ds218j.dts" one can read: ethernet@70000 { compatable = "marvell,neta"; pinctrlby Robin - Debian
Hi Bodhi, Thanks for your quick reply. The original RTC battery was dead, I replaced it before the first boot and I didn't set the time since. I'll check tomorrow but it's probably nothing to be concerned about. I felt like the SPI flash issue was indeed minor, if LEDs and Buttons requires us to edit a separate DTS for this box I'll look forward to refine this on the wayby Robin - Debian
Hi everyone, I recently found a Synology DS216 and discovered that the software in it was really outdated, so i started to search if there was a way to put a fresh debian on it. Apparently no efforts has been put on this device so far, so i started to hack it by myself. Familiar with kirkwood device hacking, i found many interesting details about the MVEBU platforms on this forum. I got it to bby Robin - Debian