Bricked Goflex Home. Need help

Hello all,

I have this Goflex Home which seems to be bricked. On power up, the device shows no blinking green lights, only a white light. The ethernet port shows only a red light, and the usb port still has power. When I put in a hard drive, I can feel it spinning. The CPU is still hot when I touch it. I tried to connect it with serial using this guide but nothing showed up in Putty. My serial adapter is PL2303 and I'm ordering a CP2102 which will arrive in a week.
I believe the goflex home was running stock FW but I'm not so sure since the device have been in storage for a long time. I uploaded some photos of the unit here

Is there anything I can do to unbrick it? Any help will be greatly appreciated.
Thanks

gigas,

The only way to unbrick this is with a working serial console. Once that's wotking, this box can be unbricked easily.

If somehow u-boot in NAND was damaged, then you need another Linux box to connect serial console and run UART booting:

https://forum.doozan.com/read.php?3,51739,51919#msg-51919

===

Also see this discussion about serial console:

http://forum.doozan.com/read.php?8,13263

Check the connection from your PL2303 to the GFH, try swaping TX and RX wires to see if it anything show up in serial console.

-bodhi
===========================
Forum Wiki
bodhi's corner (buy bodhi a beer)

I read through those discussions and tried everything but nothing seems to work.
I connected serial console to my laptop running Debian live with persistence and used kwboot but the device couldn't be found. It just stuck at "Sending boot message. Please reboot the target...\", even when I swap TX and RX wires. There is nothing showing up when I use the command "screen /dev/ttyUSB0 115200" either. I believe that there must be a hardware problem preventing my unit to boot or my PL2303 is broken. The only solution I have now is to wait for my CP2102 to arrive.

Luckily, I have another Goflex home unit still in working order, I want to run Debian with it. However, since I don't have a working serial connection, I want to boot it with stock U-boot using this guide:

Is it possible to use fw_setenv and fw_printenv included in stock FW instead of setenv and printenv? I assume that because I still be running stock uboot, it is possible to recover the device with the reset button. Is this correct?

gigas,

> "screen /dev/ttyUSB0 115200" either. I believe
> that there must be a hardware problem preventing
> my unit to boot or my PL2303 is broken. The only
> solution I have now is to wait for my CP2102 to
> arrive.

Yes, probably the PL2303 is broken.

> Luckily, I have another Goflex home unit still in
> working order, I want to run Debian with it.
> However, since I don't have a working serial
> connection, I want to boot it with stock U-boot
> using this guide:
>
> Is it possible to use fw_setenv and fw_printenv
> included in stock FW instead of setenv and
> printenv?

Not with stock FW. It does not have fw_setenv/printenv.

> I assume that because I still be running
> stock uboot, it is possible to recover the device
> with the reset button. Is this correct?

Not sure about that! it's been too long ago I can't recall how the reset button triggers the recovery.

You could install new u-boot and Debian rootfs on USB from inside stock OS. And only if you are OK with connecting serial console to recover if thing goes wrong.

-bodhi
===========================
Forum Wiki
bodhi's corner (buy bodhi a beer)

I think that you're incorrect. My GoFlex does have fw_setenv and fw_printenv on it (proof attached). I installed it with the help of this guide
http://goflexhome.blogspot.com/2019/01/firmware-reflash-without-seagateshare.html
I will install the new u-boot when my CP2102 arrive. But for now, I will leave the unit with stock uboot to avoid having 2 bricked GoFlex.
Show I follow the guide above and use fw_setenv and fw_printenv instead of setenv and printenv?



Edited 1 time(s). Last edit at 08/27/2021 01:02PM by gigas.

gigas,

> I think that you're incorrect. My GoFlex does have
> fw_setenv and fw_printenv on it (proof attached).
> I installed it with the help of this guide
> http://goflexhome.blogspot.com/2019/01/firmware-reflash-without-seagateshare.html

That needs an extra step to make fw_setenv/printenv availble. I could have remembered it incorrectly, but I recalled out of the box, none of the Kirkwood boxes has u-boot-tools installed.

And during new u-boot installation, we would have fw_setenv/printenv installed (by just copying the binary from a tarball).

> I will install the new u-boot when my CP2102
> arrive. But for now, I will leave the unit with
> stock uboot to avoid having 2 bricked GoFlex.
> Show I follow the guide above and use fw_setenv
> and fw_printenv instead of setenv and printenv?

No. You should make sure there is a rescue mechanism (i.e. serial console) before changing envs.

Or you could "living dangerously" and go ahead with the change, but with the expectation that it is risky.

-bodhi
===========================
Forum Wiki
bodhi's corner (buy bodhi a beer)

Oh, I didn't know that changing envs could lead to a brick. I always thought the most dangerous part is flash_erase and nandwrite. Thanks for clearing it up.
I have another question. According to the uboot install guide, current env suppose to have mtdparts in it, but mine doesn't have it.
Here is the output of fw_printenv

baudrate=115200
loads_echo=0
ipaddr=169.254.254.253
serverip=169.254.254.254
rootpath=/mnt/ARM_FS/
netmask=255.255.0.0
run_diag=yes
CASset=min
MALLOC_len=1
ethprime=egiga0
bootargs_end=:::DB88FXX81:eth0:none
image_name=uImage
standalone=fsload 0x2000000 $(image_name);setenv bootargs $(console) root=/dev/mtdblock0 rw ip=$(ipaddr):$(serverip)$(bootargs_end) $(mvPhoneConfig); bootm 0x2000000;
ethmtu=1500
usb0Mode=host
nandEcc=1bit
netretry=no
rcvrip=169.254.100.100
loadaddr=0x02000000
autoload=no
ethact=egiga0
rtc_mode=active
vm_ver_smt=2.0.381
ethaddr=XX:XX:XX:XX:XX:XX
explrpk=XXXX-XXXX-XXXX-XXXX
explrsn=NA1X0G0A
asm1_tag=12:09:10:01:04:46:09
console=console=ttyS0,115200
rtc_tz=PST
ntp_default=pool.ntp.org
smt_tag=12:09:10:01:04:47:49
stdin=serial
stdout=serial
stderr=serial
mainlineLinux=no
enaMonExt=no
enaCpuStream=no
enaWrAllo=no
pexMode=RC
disL2Cache=no
setL2CacheWT=yes
disL2Prefetch=yes
enaICPref=yes
enaDCPref=yes
sata_dma_mode=yes
netbsd_en=no
vxworks_en=no
bootdelay=3
disaMvPnp=no
enaAutoRecovery=no
pcieTune=no
filesize=33AA2E
bootargs=console=ttyS0,115200 root=/dev/ram0 rw ramdisk=0x01100000,8M install_type=nand
bootargs_root=ubi.mtd=2,2048 root=ubi0:rootfs rootfstype=ubifs init=/linuxrc
console2=console=ttyS0,115200
bootcmd=nand read.e 0x800000 0x100000 0x600000; setenv bootargs $(console2) $(bootargs_root); bootm 0x800000

Also, My mtd partitions seem to be different from the guide. Here is the output of cat /proc/mtd

dev:    size   erasesize  name
mtd0: 00100000 00020000 "u-boot"
mtd1: 00600000 00020000 "uImage"
mtd2: 0f900000 00020000 "root"

Is this normal?

gigas

> Oh, I didn't know that changing envs could lead to
> a brick. I always thought the most dangerous part
> is flash_erase and nandwrite. Thanks for clearing
> it up.

It is safe to change envs in serial console, because it is non-destructive. setenvs are temporary change, so when you recycle power or reboot, they will be gone. In Debian, changing envs with fw_setenv are permanent, so a typo (or power outage) would brick the box until you connect serial console and fix it.


> I have another question. According to the uboot
> install guide, current env suppose to have
> mtdparts in it, but mine doesn't have it.


> mtd0: 00100000 00020000 "u-boot"
> mtd1: 00600000 00020000 "uImage"
> mtd2: 0f900000 00020000 "root"
> [/code]

it is OK. The mtd0 (where u-boot and envs are) is all that counts. As long as it is 1MB then it is good to go.

mtd0: 00100000 00020000 "u-boot"

-bodhi
===========================
Forum Wiki
bodhi's corner (buy bodhi a beer)

Since my current env doesn't have mtdparts, What value should I use in step 8 e.2 of the uboot flashing instruction?

Quote

e.2 Box specific envs:

Then for all boxes, restore these 2 envs using the saved envs text in step c (replace xxx with the real saved values)

fw_setenv mtdparts 'xxxxxxxxx'
fw_setenv ethaddr 'xx:xx:xx:xx:xx:xx'

gigas,

> Since my current env doesn't have mtdparts, What
> value should I use in step 8 e.2 of the uboot
> flashing instruction?

fw_setenv mtdparts 'mtdparts=orion_nand:1M(u-boot),4M(uImage),32M(rootfs),-(data)'

-bodhi
===========================
Forum Wiki
bodhi's corner (buy bodhi a beer)

Because most of the fw_setenv commands from part 8 e are long and easy to be mistyped, Should I just make a shell script of all fw_setenv commands from 8 e and execute it? Are the results of each fw_setenv command important?

There are only 3 important envs

fw_setenv mtdparts 'xxxxxxxxx'
fw_setenv ethaddr 'xx:xx:xx:xx:xx:xx'
fw_setenv dtb_file '/boot/dts/kirkwood-xxx.dtb'

The rest can be done after installation. And copy/paste, don't retype.

-bodhi
===========================
Forum Wiki
bodhi's corner (buy bodhi a beer)

So, my CP2102 arrived today and I was able to boot into Debian using kwboot. However, there is no /dev/mtd0 so I can't flash the lastest uBoot. What should I do next?
Output of cat /proc/mtd

dev:    size   erasesize  name

Output of fw_printenv ethaddr

Cannot open /dev/mtd0: No such file or directory

I forgot to mention. My GoFlex showed this error, too.

U-Boot 2017.07-tld-1 (Sep 05 2017 - 00:21:31 -0700)
Seagate GoFlex Home

SoC:   Kirkwood 88F6281_A1
DRAM:  128 MiB
WARNING: Caches not enabled
NAND:  0 MiB
*** Warning - readenv() failed, using default environment

In:    serial
Out:   serial
Err:   serial
Net:   egiga0
88E1116 Initialized on egiga0
Hit any key to stop autoboot:  0

The NAND is dead right?



Edited 1 time(s). Last edit at 08/30/2021 09:05AM by gigas.

gigas,

> U-Boot 2017.07-tld-1 (Sep 05 2017 - 00:21:31
> -0700)
> Seagate GoFlex Home
>
> SoC: Kirkwood 88F6281_A1
> DRAM: 128 MiB
> WARNING: Caches not enabled
> NAND: 0 MiB
> *** Warning - readenv() failed, using default
> environment

This error is normal. When you kwboot the new u-boot, the envs are at a different location.

At this point, you can alreay boot the rootfs Debian-5.2.9-kirkwood-tld-1-rootfs-bodhi.tar.bz2 on USB.

The instruction is general to all Kirkwood boxes. Without kwboot capability you can install insde stock OS with some preparation such as define the mtd inserial console so stock OS can see them.

So by running kwboot, you can prepare the rootfs and boot directly into Debian first, and then flash u-boot and envs images in Debian. The mtd will be visible in Debian since the u-boot is new.

Please post the serial console log for verification. And

cat /proc/mtd
cat /etc/fw_env.config

-bodhi
===========================
Forum Wiki
bodhi's corner (buy bodhi a beer)

I already prepared the rootfs and booted into Debian fine. However, as I mentioned before, there is no /dev/mtd0 so I can't flash the new u-boot.
The serial console log is attached.
cat /proc/mtd

dev:    size   erasesize  name

cat /etc/fw_env.config

 # MTD device name	Device offset	Env. size	Flash sector size	Number of sectors 
/dev/mtd0 0xc0000 0x20000 0x20000

gigas,

Ah! I missed this:

U-Boot 2017.07-tld-1 (Sep 05 2017 - 00:21:31 -0700)
Seagate GoFlex Home

SoC:   Kirkwood 88F6281_A1
DRAM:  128 MiB
WARNING: Caches not enabled
NAND:  0 MiB

That's bad news. Looks like it's dead. But let me think about this a little bit and come back to this topic.

In the mean time, each time you reboot, you need to run kwboot. It is not bad, considering you have no NAND at this point.

-bodhi
===========================
Forum Wiki
bodhi's corner (buy bodhi a beer)



Edited 1 time(s). Last edit at 08/30/2021 04:49PM by bodhi.

gigas,

Indeed,

[   10.486787][    T1] nand: second ID read did not match 98,d8 against fd,fd
[   10.493768][    T1] nand: No NAND device found

Did you get this box used from somebody else, or you are the original owner?

-bodhi
===========================
Forum Wiki
bodhi's corner (buy bodhi a beer)

I bought this one used from someone else a long time ago. I used it until Seagate stopped supporting it, and kept it in storage after that. Recently, I bought another one on a deal, so I decided to dig this one out of storage. That's how I got 2 GoFlex

gigas,

> I bought this one used from someone else a long
> time ago. I used it until Seagate stopped
> supporting it,

This means NAND was working.

> and kept it in storage after that.

So NAND is broken sometime between the time you used it wih stock FW, and

> so I decided to dig this one out of storage.

That's quite rare, that NAND is bad just sitting on the shelf. I'm quite curious.

Have you done anything more (right after you digged it out) until the 1st time you tried to run kwboot?

-bodhi
===========================
Forum Wiki
bodhi's corner (buy bodhi a beer)

> Have you done anything more (right after you
> digged it out) until the 1st time you tried to run
> kwboot?

I did try to get serial connection with the metal shielding still on with the PL2303. Maybe I short circuited the NAND?

gigas,

> I did try to get serial connection with the metal
> shielding still on with the PL2303. Maybe I short
> circuited the NAND?

I did not remove the metal shielding either, only remove the case when I connected serial console. My GFHome boxes have pre-installed serial header so nothing was touched other than the pins. But I guess also it depends how careful you were.

-bodhi
===========================
Forum Wiki
bodhi's corner (buy bodhi a beer)

After some days messing with my dead NAND GoFlex, there are few things I want to ask:

- Are there any binaries of kwboot for other systems (such as android or mips router)? If there is none, are there any guides to help me build kwboot for those systems?( I was super lost when trying to compile it because of so many options I don't know the purpose of). I'm able to use kwboot wirelessly with the help of the android app "TCPUART transparent Bridge" and socat so driver is not an issue.

- I believe the reset button on the GoFlex has no function in debian, and safe to remove. Am I correct? I intend to drill out the reset button hole and route cable for easier serial access. Will this cause problems?

- Since my GoFlex has a dead NAND, is it possible that replacing the NAND chip can make it boot normally again?

Thank you so much.



Edited 1 time(s). Last edit at 09/05/2021 10:17AM by gigas.

gigas,

> - Are there any binaries of kwboot for other
> systems (such as android or mips router)?

No, kwboot is specific to Marvell SoCs. The BootROM in these boards use xmodem protocol to handshake with the sender to receive the u-boot image over UART.

For other ARM SoCs, usually you would find JTAG.

These Kirkwood boards also has a JTAG header. Most Kirwood board have the pins exposed, and you can solder a standard JTAG header to it.


> - I believe the reset button on the GoFlex has no
> function in debian, and safe to remove. Am I
> correct? I intend to drill out the reset button
> hole and route cable for easier serial access.
> Will this cause problems?

The reset button can be used in Debian also, as with all types of buttons. However, this GFH reset button is not yet programmed in the kernel or u-boot at the moment. So I guess, yes, you could expand the hole to be larger to route the serial wires without worrying.


> - Since my GoFlex has a dead NAND, is it possible
> that replacing the NAND chip can make it boot
> normally again?

It is. Some people are skilled and they replaced the NAND chip using one with more memory. But this is a difficult task for hardware beginners.

-bodhi
===========================
Forum Wiki
bodhi's corner (buy bodhi a beer)

I think you misinterpret my first question.
I am wondering if there is a way to boot my GoFlex without the need of a PC (since kwboot is required)? I do have an Android phone and router running OpenWRT or DDWRT and want to compile kwboot (or u-boot-tools package) on them.

I just need a working binary for those devices because I have a way to run kwboot without a CP2102 physically attached to the devices. It is as follow:
Uart <=> TCP server <=> socat client <=> kwboot

This way I don't have to care about CP2102 driver for each system I want to run kwboot on

gigas,

> I think you misinterpret my first question.
> I am wondering if there is a way to boot my GoFlex
> without the need of a PC (since kwboot is
> required)? I do have an Android phone and router
> running OpenWRT or DDWRT and want to compile
> kwboot (or u-boot-tools package) on them.

I see.

>
> I just need a working binary for those devices
> because I have a way to run kwboot without a
> CP2102 physically attached to the devices. It is
> as follow:
> Uart <=> TCP server <=> socat client <=> kwboot
>
> This way I don't have to care about CP2102 driver
> for each system I want to run kwboot on

How does this "Uart <=> TCP server" connection work? is the CP2102 attached to the "TCP server"?

You need at least one computer with the CP2102 or some module converter attached to it. This computer is physically near the GoFlex Home to the wires can be connected. And then ssh into that computer to run kwboot.

OpenWrt should have u-boot-tools in the repo. I believe Android might too.

Did I understand what you want to do correctly? perhaps I am not clear about "Uart <=> TCP server".

-bodhi
===========================
Forum Wiki
bodhi's corner (buy bodhi a beer)



Edited 1 time(s). Last edit at 09/06/2021 04:23AM by bodhi.

Yes, you understood it correctly. I want to ssh into Openwrt or Android to use u-boot-tools. Could you link me to it?

The "Uart <=> TCP server" connection works with the help of the android app "TCPUART transparent Bridge". Here is a link to it.
https://play.google.com/store/apps/details?id=com.hardcodedjoy.tcpuart&hl=en_US&gl=US
You will understand everything after reading the app's description.

gigas,

> Yes, you understood it correctly. I want to ssh
> into Openwrt or Android to use u-boot-tools. Could
> you link me to it?

It is a standard Linux install. You just install it in OpenWrt.

> The "Uart <=> TCP server" connection works with
> the help of the android app "TCPUART transparent
> Bridge".

I'm afraid you are making it harder than necessary. It is simple. Install u-boot-tools in your OpenWrt router, and then ssh into the router to run kwboot.

If you want to run ssh in Android, then install terminal app.

-bodhi
===========================
Forum Wiki
bodhi's corner (buy bodhi a beer)



Edited 1 time(s). Last edit at 09/06/2021 02:48AM by bodhi.

I don't think it is that simple.
- First, there is no u-boot-tools package in the official OpenWRT Package database. Here is the link for that.
https://openwrt.org/packages/table/start
There is uboot-envtools but it only includes fw_setenv and fw_printenv.
- Second, I do have termux (terminal app) running on my phone but termux doesn't have u-boot-tools on its package list either.
- Third, the reason for "Uart <=> TCP server" connection is mainly the lack of CP2102 driver on Android (there is no /dev/ttyUSB0 in termux). I believe there is a CP2102 driver for OpenWRT, but I haven't looked into it yet because there is no kwboot binary to test.

Which router do you have? can you find it on Wikidevi?

https://wikidevi.wi-cat.ru/Main_Page

-bodhi
===========================
Forum Wiki
bodhi's corner (buy bodhi a beer)