Closed Bug 1026963 Opened 10 years ago Closed 10 years ago

[Flatfish] Need a bare boot image supporting fastboot mode (for solving the brick problem)

Categories

(Firefox OS Graveyard :: General, defect)

Product:
Firefox OS Graveyard
Component:
General
Type:
defect
Priority:
Not set
Severity:
normal

Tracking

(Not tracked)

Status:
RESOLVED FIXED

People

(Reporter: iean.lin, Unassigned, NeedInfo)

References

Details

(Whiteboard: [Flatfish][TCP]) 

User Agent: Mozilla/5.0 (Windows NT 5.1; rv:30.0) Gecko/20100101 Firefox/30.0 (Beta/Release)
Build ID: 20140605174243
Hi Sam,

Please help verify the image, thanks...

Download link:
https://www.dropbox.com/sh/fyffh1039820804/AABgNby_K8kjuU3uZFyzedH0a?n=78209401
Flags: needinfo?(ckjboy2003)
Blocks: flatfish
This works to me.
Flags: needinfo?(ckjboy2003)
Status: UNCONFIRMED → NEW
Ever confirmed: true
I am one of the Flatfish users who had a bricked tablet.

I used this and can report my tablet has been recovered. I thank you and I think future flatfish developers thank you!

I wanted to note a few things:
- The instructions in the readme (which I used) rely on PhoenixSuit which is a Windows-only application.  Does this solution depend on having access to Windows?  I see there is a bash script among the files.  I am not familiar enough with *nix to understand it.  Does it enable Linux and Mac users to also recover their devices?
- After booting into fastboot mode I experienced a couple issues:
  - "fastboot devices" worked (showing ????? instead of a name, but I have seen this previously) but with further commands fastboot was crashing and saying it was out of memory.  I rebooted my system and only ran the flashing commands and succeeded.
  - After flashing was successful I used the "Fastboot reboot" command. It said "rebooting..." but the device did not reboot. I used ^C and tried again without success. Eventually I used the power button to power down and up again and it booted into B2G/Firefox OS successfully.  As a test I then rebooted into fastboot mode and successfully used the fastboot command to reboot.

I suspect there may be some incompatibility or issue with the fastboot mode in this image - at least with the Windows fastboot.exe.
Anyway, I wanted to note this in case other Windows users ran into the same issues.
Hi, Caspy, it's good to know that it helps.

1. PhoenixSuit is for Windows only. The other tool, LiveSuit running on Linux, seems to be not available in the public domain.
2. The flash-flatfish.sh was tested on Linux only. We'll be grateful if any one can help test it on Mac. The 'bash' shell is required. I guess it won't work on Windows. (To use it, please download the script, and change the permission by 'chmod +x flash-flatfish.sh'. Then, run './flash-flatfish.sh <path-to-images>')
3. How did you do to do the flashing, on which host operating system? Did you use the script or you did it by using the fastboot commands? Could you provide the command procedure, better with log, so that we can help check the problem. 

I guess you were using Windows, since you mentioned 'fastboot.exe'. We've never tried running fastboot on Windows yet. Eva, could you help try fastboot on Windows? 

Thanks.
Flags: needinfo?(eva.chen.fx)
Hi William,
The fastboot work on Windows8.1

Eva
Flags: needinfo?(eva.chen.fx)
Blocks: 1028019
"fastboot.exe" also works on my XP system
A couple of thoughts, but they add complications

1) perhaps the required Windows exes will run on Linux under Wine?
2) Bash is available on Windows via GitHub Windows, which use MinGW. Another option is Cygwin. However other dependencies may not work out.
Hi, Steve, thanks for the input. 
We used to use VirtualBox to install the Windows system, in which the PhoenixSuit is installed and run. Wine can be a good idea. However, we've never tried that yet.
Bug 1028019 has been created for the Flatfish image flash script discussion.
Just wanted to report that I asked someone to install PhoenixSuit with Wine and they reported that it failed to install properly.

Also, just to add clarity for some reading this bug, the Bash script included with the download is not crucial to the recovery of the device, it is simply for flashing good image files (FxOS) once the device is in a proper fastboot mode.
This can be done with a script or manually from any platform (Mac, Linux, Windows).

I did try using fastboot in Windows again after flashing the recovery image and was successful with no issues. I can't say why there was the problem before, but I did originally try 'fastboot erase cache' the first time and not the second. I think it may have been what gave the problem originally.  I is a part of one of our protocols.
Thanks for the update.

Yes, the flash script is just provided to easy the flash process and automatically deal with several problems that the user may encounter.
I am assisting a user who has a bricked tablet.  He is on a Mac and running a Windows XP VM.
The tablet is detected and PhoenixSuit prompts him about Format mode, he says yes and sees an empty progress bar, the flash quickly fails and says "Wait for PnpFesIn timeout!"
Any idea what this is?  Are there known issue with running PhoenixSuit on Windows XP?
Iean, any idea? Can Kevin help on this?
Flags: needinfo?(iean.lin)
The problem might have to do with the VM you're using. If you're using VMWare, try VirtualBox instead. It usually works on VirtualBox.
Flags: needinfo?(iean.lin)
The user just reported to me that he successfully flashed his device.

He went to another computer with native Windows 7 and received the exact same error as before.
But he was able to get to this window in PhoenixSuit: http://i.imgur.com/3L86j1i.png
Which suggests pressing the power key repeatedly (10x) after connecting to the PC and this did the trick.

I do find it interesting/puzzling that we both apparently have the tablet hardware and both flashed the same broken build, but I did not experience the same issue in flashing.
Hi, Iean, is the problem that Caspy has met is due to the burning tool? Have you ever encountered similar problem (failing to burn the bare boot image from time to time)?
So far as I know, there are issues with some USB3.0 controller, to disable USB3.0 setting on BIOS can fix this problem.
Yes, I've just confirmed this issue and the solution. Thanks to Iean.
Thanks. That helps contributors clarify the problem that could be caused by the USB 3.0 host controller when the 'bare boot image' needs to be burned into a brick device by the 'PhoenixSuit tool'.
This bug seems like it could be marked RESOLVED FIXED.  Agree?
Agreed. Please.
Status: NEW → RESOLVED
Closed: 10 years ago
Resolution: --- → FIXED
Blocks: 1031983
Update the download link of the bare boot image (recover-image-for-fastboot).
https://www.dropbox.com/sh/lrwkifcira0608t/AAC_J0Du-GrpFkKx5oIdFg39a
Whiteboard: [Flatfish][TCP]
A new version of the bare boot image is now available in https://www.dropbox.com/sh/b2py1btcwstqldl/AABblbq_csa1IHQwdvLdfptTa/. The file is located under the recovery-image-for-fastboot/ directory, namely sun6i_boot-fastboot_v2.0.img. In this release, the following issue has been solved.

1. Support hardware key to enter the Fastboot mode (Bug 1017934)
2. Show a 'Fastboot' logo once the device enter the fastboot mode.
3. Unify the USB VID for both adb and fastboot. (Bug 1033619)
4. Solve the snowflake problem while flashing the system.img file.

Please check the README file for further information.
See Also: → 983033
William,

I tried follow the instructions at the README file from https://www.dropbox.com/sh/b2py1btcwstqldl/AABC4ZX-Fg9xqDwkJInGl-yba/recovery-image-for-fastboot?dl=0 but after

    4. Hold the volume(-) button then plug-in the microUSB cable to connect to the host PC.

nothing happens. The output of dmesg is

    [11895.770209] usb 3-2: new full-speed USB device number 17 using xhci_hcd
    [11895.899393] usb 3-2: New USB device found, idVendor=1f3a, idProduct=efe8
    [11895.899398] usb 3-2: New USB device strings: Mfr=0, Product=0, SerialNumber=0

but LiveSuit doesn't see the device. Any suggestion?
Flags: needinfo?(william.wyliang)
You need to log in before you can comment on or make changes to this bug.