Closed
Bug 973046
Opened 11 years ago
Closed 11 years ago
[perf][Tarako] Unable to flash Tarako - stuck in fastboot trying to flash/send 'system' (95085 KB)
Categories
(Firefox OS Graveyard :: General, defect)
Tracking
(Not tracked)
VERIFIED
FIXED
People
(Reporter: stephend, Unassigned)
References
()
Details
(Whiteboard: [perf-reviewed][POVB])
Attachments
(1 file)
401.64 KB,
application/octet-stream
|
Details |
Raymond and I set up a new Tarako device (properly flashed with the vendor's base build), and attempted to resurrect the b2g.tarako.mozilla-aurora.v1.3.perf job; in doing so, we ran into the same problem, apparently, that caused the 1st device we set up to suddenly fail: when the 1.3 build-flash job runs, it tries to push the "system" app/files, like so:
13:01:02 sending 'system' (95085 KB)...
Both devices are stuck, here, in fastboot mode -- is the image too big for the device, perhaps? There's a note at the bottom of https://intranet.mozilla.org/B2G_Team/Tarako: "Fastboot can flash system.img < 100MB only; Otherwise the device will be bricked. In case of device is bricked, it can be recovered by using flashing tool to flash a base builds."
Al, are you seeing this too? :-(
Reporter | ||
Comment 1•11 years ago
|
||
http://qa-selenium.mv.mozilla.com:8080/view/B2G%20Tarako/job/b2g.tarako.mozilla-aurora.v1.3.perf/112/console is the most recent flash attempt, which fails.
Summary: [Tarako] Unable to flash Tarako - stuck in fastboot trying to flash/send 'system' (95085 KB) → [perf][Tarako] Unable to flash Tarako - stuck in fastboot trying to flash/send 'system' (95085 KB)
Reporter | ||
Updated•11 years ago
|
Flags: needinfo?(atsai)
Comment 2•11 years ago
|
||
For what I knew, Partner set the partition as 100MB, which should around 110000000 Bytes. The size of system image seems available for flashing. set the NI to Thomas Tsai and Danny to see if they have any suggestion.
Also, you will need to update kernel image for building. You will need to run the command once "./repo kernel update"
Flags: needinfo?(ttsai)
Flags: needinfo?(dliang)
Flags: needinfo?(atsai)
Comment 3•11 years ago
|
||
There is a bug in u-boot that caused flash system.ing fail. sprd has fixed this bug and please update u-boot then flash system.img again.
How to update u-boot:
#adb reboot bootloader
#fastboot flash 2ndbl u-boot.bin
#fastboot reboot
With the latest u-boot, the limitation of flash size is around 105M. By my test, I can flash the system.img with size is 101M.
Stephen, could you help to update u-boot then flash system.img again?
Attachment #8378090 -
Flags: feedback?(stephen.donner)
Flags: needinfo?(ttsai)
Flags: needinfo?(dliang)
Comment 4•11 years ago
|
||
I just tried the suggestion in comment 3 but I believe the device is in a bad state that will first need to be resolved. The fastboot commands are all returning with "< waiting for device >". Can we recover using the flashing tool as mentioned in https://intranet.mozilla.org/B2G_Team/Tarako#Q.26A
Updated•11 years ago
|
Whiteboard: [perf-reviewed]
Comment 5•11 years ago
|
||
(In reply to Dave Hunt (:davehunt) from comment #4)
> I just tried the suggestion in comment 3 but I believe the device is in a
> bad state that will first need to be resolved. The fastboot commands are all
> returning with "< waiting for device >". Can we recover using the flashing
> tool as mentioned in https://intranet.mozilla.org/B2G_Team/Tarako#Q.26A
I think your issue might be another one. May I have some questions to identify your issue?
1. Could you recognize the device by lsusb?
If it's recognized, it will show something like "Bus 001 Device 073: ID 1782:5d05 Spreadtrum Communications Inc."
2. Do you follow https://intranet.mozilla.org/B2G_Team/Tarako#Q.26A to setup your environment?
3. How is the status of adb connection?
4. Does the screen show "fastboot mode" when you try fastboot command?
Comment 6•11 years ago
|
||
I'll leave :stephend to answer these or delegate as I don't have physical access to the device.
Flags: needinfo?(stephen.donner)
Reporter | ||
Comment 7•11 years ago
|
||
Here are, I believe (Raymond can correct me) the udev rules we're using:
webqa@b2g-6:~$ cat /etc/udev/rules.d/51-android.rules
SUBSYSTEM=="usb", ATTR{idVendor}=="18d1", MODE="0666", GROUP="plugdev"
SUBSYSTEM=="usb", ATTR{idVendor}=="19d2", MODE="0666", GROUP="plugdev"
SUBSYSTEM=="usb", ATTR{idVendor}=="05c6", MODE="0666", GROUP="plugdev"
SUBSYSTEM=="usb", ATTR{idVendor}=="1782", ATTR{idProduct}=="5d06", MODE="0666", GROUP="plugdev"
(In reply to Danny Liang [:dliang] from comment #5)
> I think your issue might be another one. May I have some questions to
> identify your issue?
> 1. Could you recognize the device by lsusb?
> If it's recognized, it will show something like "Bus 001 Device 073: ID
> 1782:5d05 Spreadtrum Communications Inc."
It's not currently recognized, no.
> 2. Do you follow https://intranet.mozilla.org/B2G_Team/Tarako#Q.26A to setup
> your environment?
Yes, I believe Raymond did.
> 3. How is the status of adb connection?
adb devices doesn't return anything
> 4. Does the screen show "fastboot mode" when you try fastboot command?
Nope :-(
webqa@b2g-6:~$ fastboot mode
< waiting for device >
Flags: needinfo?(stephen.donner)
Comment 8•11 years ago
|
||
(In reply to Stephen Donner [:stephend] from comment #7)
> Here are, I believe (Raymond can correct me) the udev rules we're using:
>
> webqa@b2g-6:~$ cat /etc/udev/rules.d/51-android.rules
> SUBSYSTEM=="usb", ATTR{idVendor}=="18d1", MODE="0666", GROUP="plugdev"
> SUBSYSTEM=="usb", ATTR{idVendor}=="19d2", MODE="0666", GROUP="plugdev"
> SUBSYSTEM=="usb", ATTR{idVendor}=="05c6", MODE="0666", GROUP="plugdev"
> SUBSYSTEM=="usb", ATTR{idVendor}=="1782", ATTR{idProduct}=="5d06",
> MODE="0666", GROUP="plugdev"
>
Suggest to add two more lines in /etc/udev/rules.d/51-android.rules
SUBSYSTEM=="usb", ATTR{idVendor}=="1782", ATTR{idProduct}=="5d04", MODE="0666", GROUP="plugdev"
SUBSYSTEM=="usb", ATTR{idVendor}=="1782", ATTR{idProduct}=="5d05", MODE="0666", GROUP="plugdev"
And try to plugin/plugout the cable few times. In my experiment, sometimes device will be disconnected, and sometimes is good.
Reporter | ||
Comment 9•11 years ago
|
||
(In reply to Danny Liang [:dliang] from comment #8)
> (In reply to Stephen Donner [:stephend] from comment #7)
> > Here are, I believe (Raymond can correct me) the udev rules we're using:
> >
> > webqa@b2g-6:~$ cat /etc/udev/rules.d/51-android.rules
> > SUBSYSTEM=="usb", ATTR{idVendor}=="18d1", MODE="0666", GROUP="plugdev"
> > SUBSYSTEM=="usb", ATTR{idVendor}=="19d2", MODE="0666", GROUP="plugdev"
> > SUBSYSTEM=="usb", ATTR{idVendor}=="05c6", MODE="0666", GROUP="plugdev"
> > SUBSYSTEM=="usb", ATTR{idVendor}=="1782", ATTR{idProduct}=="5d06",
> > MODE="0666", GROUP="plugdev"
> >
> Suggest to add two more lines in /etc/udev/rules.d/51-android.rules
> SUBSYSTEM=="usb", ATTR{idVendor}=="1782", ATTR{idProduct}=="5d04",
> MODE="0666", GROUP="plugdev"
> SUBSYSTEM=="usb", ATTR{idVendor}=="1782", ATTR{idProduct}=="5d05",
> MODE="0666", GROUP="plugdev"
>
> And try to plugin/plugout the cable few times. In my experiment, sometimes
> device will be disconnected, and sometimes is good.
1) Done
2) Done
Still, no dice; the device isn't appearing when issuing lsusb, even (same host machine, same USB cable, etc.).
We even did webqa@b2g-6:~$ sudo udevadm control --reload-rules for good measure :-(
We understand this is pretty important to the Perf team, so perhaps we can make a local, consolidated effort? cc:ing Tony and Fabrice for potential outreach help.
Comment 10•11 years ago
|
||
Not sure if this helps, but I have a "~/.android/adb_usb.ini" with string "0x1782" (idVendor of Spreadtrum) on my Ubuntu.
Comment 11•11 years ago
|
||
Sometimes usb device is not found by lsusb, so it can't be solved by any setting.
Sometimes usb device is found by lsusb, but adb devices can not find any devices.
So not the only one condition happens in this case.
Comment 12•11 years ago
|
||
(In reply to Ting-Yu Chou [:ting] from comment #10)
> Not sure if this helps, but I have a "~/.android/adb_usb.ini" with string
> "0x1782" (idVendor of Spreadtrum) on my Ubuntu.
Please use this method or use spreadtrum download tool in Win7.
Comment 13•11 years ago
|
||
Thomas, I think you should distribute the spreadtrum research download tool to your guys.
Flags: needinfo?(ttsai)
Comment 14•11 years ago
|
||
Please refer to internal wiki : https://intranet.mozilla.org/B2G_Team/Tarako
about how to install pac tool and usb driver
Flags: needinfo?(ttsai)
It turns out that you are in adb mode when this occurs. You can do a:
adb reboot bootloader
To get into fastboot and then flash a build that has a system.img that is < 100 MB.
For the automation, we should probably have a check to see what the file size of the system.img is before we flash.
For the b2g, perhaps we should have a check to see how big the system.img is before we try to flash for tarako?
You don't necessarily have to use the tool to reset it to the base image.
Comment 16•11 years ago
|
||
Thomas, could you find someone to verify this bug again given Stephen is on PTO?
Flags: needinfo?(ttsai)
Comment 17•11 years ago
|
||
Hi Stephen, can you update the new build again and test the usb issue?
By https://bugzilla.mozilla.org/show_bug.cgi?id=975232
SPRD has improved the usb compatibilty. The USB compatible issue highly depends on the PC host, so if your adb/fastboot still doesn't work by a new build, please need info me again.
Flags: needinfo?(ttsai)
Comment 18•11 years ago
|
||
Is this bug fixed on my side?
commit 27a4b15ae8c526f538409504acb86643c0cd4daa
Author: lianxiang.zhou <lianxiang.zhou@spreadtrum.com>
Date: Mon Mar 3 15:25:24 2014 +0800
Bug #275374 usb tethering can not work
[bug number ] 275374
[root cause ]
[changes ]
[side effects]
[reviewers ]
Change-Id: Id490f6450d6f755e94b97b1e5adfa5c1e722af34
Flags: needinfo?(ttsai)
Reporter | ||
Comment 20•11 years ago
|
||
Comment on attachment 8378090 [details]
u-boot.bin
Once we have a new build from Release Engineering, with the correct size (so it'll fully flash), we'll confirm.
Attachment #8378090 -
Flags: feedback?(stephen.donner)
Comment 21•11 years ago
|
||
Release engineering is producing flashable builds now; can we try to get this running again?
Comment 22•11 years ago
|
||
(In reply to Jonathan Griffin (:jgriffin) from comment #21)
> Release engineering is producing flashable builds now; can we try to get
> this running again?
What about the system image size? It should be less than 105MB.
Comment 24•11 years ago
|
||
system.img is now 97MB (user build) with woff fonts and english, Bengali locales.
Here are the fonts included in the 97MB:
CharisSILCompact-BI.woff DroidSerif-Italic.woff Lohit-Kannada.woff
CharisSILCompact-B.woff DroidSerif-Regular.woff Lohit-Malayalam.woff
CharisSILCompact-I.woff FiraMonoOT-Bold.otf Lohit-Oriya.woff
CharisSILCompact-R.woff FiraMonoOT-Regular.otf Lohit-Punjabi.woff
DroidNaskh-Regular.woff FiraSansOT-BoldItalic.otf Lohit-Tamil.woff
DroidSansArmenian.woff FiraSansOT-Bold.otf Lohit-Telugu.woff
DroidSansEthiopic-Regular.woff FiraSansOT-LightItalic.otf MTLc3m.woff
DroidSansFallback.woff FiraSansOT-Light.otf MTLmr3m.woff
DroidSansGeorgian.woff FiraSansOT-MediumItalic.otf Padauk-bold.woff
DroidSansHebrew-Bold.woff FiraSansOT-Medium.otf Padauk.woff
DroidSansHebrew-Regular.woff FiraSansOT-RegularItalic.otf Roboto-BoldItalic.ttf
DroidSansMono.woff FiraSansOT-Regular.otf Roboto-Bold.ttf
DroidSansThai.woff Lohit-Bengali.woff Roboto-Italic.ttf
DroidSerif-BoldItalic.woff Lohit-Devanagari.woff Roboto-Regular.ttf
DroidSerif-Bold.woff Lohit-Gujarati.woff
Flags: needinfo?(ttsai)
Comment 25•11 years ago
|
||
Hi Stephen, After update u-boot.bin, I think it is able to flash tarako in fastboot with system.img less than 105MB. Can we resolve this bug as fixed with [POVB] in white board?
Flags: needinfo?(kli) → needinfo?(stephen.donner)
Comment 26•11 years ago
|
||
We will provide research download tool on Linux, Sam and our owner is testing it now.
Reporter | ||
Comment 27•11 years ago
|
||
Raymond, can you try flashing the Tarako with the base build (1st) and then 1.3T nightly build (2nd) again, today? Thanks!
Flags: needinfo?(stephen.donner) → needinfo?(mozbugs.retornam)
Comment 28•11 years ago
|
||
(In reply to Stephen Donner [:stephend] from comment #27)
> Raymond, can you try flashing the Tarako with the base build (1st) and then
> 1.3T nightly build (2nd) again, today? Thanks!
I've done both and connected the device to its node in Jenkins.
Flags: needinfo?(mozbugs.retornam)
Reporter | ||
Comment 29•11 years ago
|
||
(In reply to raymond [:retornam] (needinfo? me) from comment #28)
> (In reply to Stephen Donner [:stephend] from comment #27)
> > Raymond, can you try flashing the Tarako with the base build (1st) and then
> > 1.3T nightly build (2nd) again, today? Thanks!
>
> I've done both and connected the device to its node in Jenkins.
Thanks, Raymond! James, Kal-Zhen, can one of you take this bug and resolve it as FIXED, now? Thanks!
Flags: needinfo?(kli)
Updated•11 years ago
|
Status: NEW → RESOLVED
Closed: 11 years ago
Flags: needinfo?(kli)
Resolution: --- → FIXED
Whiteboard: [perf-reviewed] → [perf-reviewed][POVB]
Reporter | ||
Comment 30•11 years ago
|
||
Verified; thanks -- this is fine in our CI/Jenkins instance, now.
Status: RESOLVED → VERIFIED
You need to log in
before you can comment on or make changes to this bug.
Description
•