Installing FVDW on a Seagate Personal Cloud

[fvdw]

thank you, BTW, little fvdw is bigger then me :-D

about setenv :whistle sorry my mistake yes of course "setenv" should be used to set an environment variable
I mixed it up with the syntaxc of bourne shell, stupid.

So we can poke around with the environment variable.

Thanks for the full printenv output, more clear now

The usbstart command clears gpio pin that apparently enables boot via usb but it is not clear what is required on the usb stick to let it boot

Lets focus for the moment on "bootcmd"
Commonly used to boot the device, is says "run nexus_boot". A long and complicated command, which need the specifice seagate setup with nexus. Not of use for us. We have to go another route.

If we set bootcmd to for instant to "run fvdw-sl" and put a environment variabel "fvdw-sl" like this the boot would as on current lacie box.

Code: Select all

setenv bootcmd 'disk 0x800000 0:6 ; setenv bootargs console=ttyS0,115200 root=/dev/sda7 rw ; bootm 0x2000000;'

However it could be that bootcmd is reset by u-boot to the default at each boot. Which is the case on lacie nas.
On the Lacie's it could prevented by setting a special environment variabel "LaBackdoor" to yes

Code: Select all

setenv LaBackdoor yes
saveenv

To re-enable "normal" boot

Code: Select all

sentenv LaBackdoor
saveenv

This something you could try. It should not harm u-boot.

PS an alternative is not changing bootcmd but booting the nas from the uboot prompt, the disadvantage of that is that you will need to repeat it at every boot

In the meantime I will look if compiling a kernel is easy for this device, we could use the one in rescue partition in case compile fails

[TonyP]

However it could be that bootcmd is reset by u-boot to the default at each boot. Which is the case on lacie nas.

Bootcmd is saved ok and not overwritten automatically after reboot, I had a play and did

Code: Select all

setenv wait wait 20
setenv customboot run nexus_boot
setenv bootcmd run wait customboot
saveenv


It did not boot as expected, but it might have needed an ide reset to be issued, or uboot doesn't like to wait. In any case, it survives a reboot without LaBackdoor needing to be turned on.

Edit - it should be sleep in unix, not wait!

Code: Select all

setenv xxsleep sleep 20
setenv xxresetide reset ide
setenv xxcustomboot run nexus_boot
setenv bootcmd run xxsleep xxresetide xxcustomboot
savenv


causes the box to pause for 20 seconds at boot time as expected. I have added xx to keep track of my changes.

[fvdw]

Oke :thumbup nice. So now we only need a kernel
The compiler is running to produce a standalone kernel for this device. It is present in the kernel tree. Hopefully my guess configuration will bring it to live

[fvdw]

Request: Can you post in a zip file
the files from sda2
armada-370-n090102.dtb
armada-370-n090103.dtb
armada-370-n090201.dtb
armada-370-n090203.dtb
armada-370-n090401.dtb

and
target.dtb

Put them in a zip file first

This to put one of those in the kernel, in the standaard linux tree there are several variants of the armada370 each one needs a specific device tree setup that needs to be added to the kernel, I can compile the dtb file of the standards present in the kernel but no idea if they conatin the right setup. Using of those of lacie should work

furthermore the output of dmesg when running the lacie kernel is useful to see what kind of main board is in there

[TonyP]

Sure, this device uses armada-370-n090103.dtb

Please find attached. Thank you.

[fvdw]

ok thks I will add that and post a kernel in a few minutes

PS think I already found the dts files, is your device a personal cloud with 1 disk ?

[TonyP]

Yes, a 3TB single disk.

I'm sorry though, I need to go to bed now!

It's 9.26pm here in Melbourne, Australia. I will be back after I drop the kiddies at school tomorrow at about 9.30am....so 12 hours from now.

Thank you for your help fvdw.

[fvdw]

no problem

attached a first trial standalone kernel with the dtb file you have send

This evening I will send one with own compiled dts file

You could laod and start it using the fvdw-sl-console but it could be that it loads it at a wrong location in memory overwriting the ram were uboot is loaded and then it will fail. (ps uboot in flash will not be affected)
This because I noticed in printenv addres the address of loading the rescue kernel is 2000000

So maybe better for first trial to load it manually via uboot console of the fvdw-sl console.
Put the standalone kernel in the tftp folder of the fvdw-sl console.
Be sure the tftp server (from fvdw-sl console) is also running
Then at command prompt give these commands
(do not replace ip addr and serverip and so on by actual ip's, type the command exactly as given)

Code: Select all

setenv ipaddr ${ipaddr}; setenv serverip ${serverip}
setenv bootargs console=ttyS0,115200 netconsole=6666@${ipaddr}/,6666@${serverip}/ root=/dev/sda7

(you can paste it in the window, right click on title bar of the window and choose edit)

Then upload the kernel

Code: Select all

tftpboot 0x2000000 uImage-4.2.3-sgpc-standalone

this will give output that it uploads the kernel and mention when it is done

next boot the kernel

Code: Select all

bootm

if that doesn't work try

Code: Select all

bootm 2000000

and see what happens. If it loads and produce output we need that output to see what can be done to solve the problem
When it loads without errors you should be able to log in with telnet client, using the ip adress mentioned in fvdw-sl console or the one assigned by you dhpc server.
(the static addresses programmed in the initramfs in the standalone kernel are in range of 192.68.1 and 192.168.0)

uImage-4.2.3-sgpc-standalone.zip

uImage-4.2.3-sgpc2-standalone.zip

[TonyP]

It says 'starting kernel' (using the dts compiled kernel) and I can see a telnet server but can't connect to it. Having a play now to make sure it's not Windows or firewall problem. Telnet just results in blank window, no login prompt.

Edit - it appears broken? Using Putty to telnet hangs with no response at all, and fails after about a minute with 'software caused connection abort' error, which it also does if if I try and connect to a non-existent IP on my home network.

Windows telnet client appears to connect (but only results in the cursor moving when I type, but no text is visible. It falls back to the Windows command line prompt after 1 minute.

This behaviour is different to when I try and connect to a non-existent IP. It says 'Could not open connection to the host, on port 23: Connect failed' when I try this.

JuiceSSH android app says 'connected' and accepts commands but with no response. It eventually closes the session.

I'm using Fing on my phone to scan for telnet, and it detects a change in the NAS IP to 192.168.1.8 +1 when I boot the new kernel, and says 'device has multiple IP addresses'. It shows the IP addresses as 192.168.1.8 and 192.168.1.251

It appears the kernel has 2 IP addresses for some reason. I tried connecting to 192.168.1.251 as well, unsuccessfully.

Edit - the above steps were completed using the 'bootm' command.

Using 'bootm 2000000' results in a telnet server being spawned, and 2 IP addresses - 192.168.1.251 and 192.168.1.252, cannot connect on either.

The n090103.dtb kernel doesn't spawn a telnet server at all, and gets no IP address.

[fvdw]

oke when the kernel boots do you get output in the uboot window ? showing progress of loading of the kernel
Something like the output you get when giving command "dmesg" when running normal kernel

Anyhow the kernel seem to be up and running in case the box gets an ip address :mrgreen: , and yes it gets 2 ip addresses one assigned via dhcp (the 192.168.1.8) and one assigned statically (192.168.1.251). The last one is added to have access in case that there is no dhcp server in the LAN.

Strange that the telnet connection fails, did you disable your firewall for the telnet client, as dhcp ip setup and telnet use a busybox applet I do not see a reason yet why the telnet server does not run. Can you check if port 21 is open on the NAS ? For instance with using nmap on your windows PC).But I guess it is there as you connected with two applications.