Issues with USB connectivity etc. and root / A2SD

[Guest jdouce]

i used recovery

[Guest afiorillo]

Another thing I was wondering is: which class the used MicroSD was? Since many core OS pieces are moved to the memory card by A2SD, a Class4+ card is mandatory: the stock provided one is a Class2 card, not suitable to be used with A2SD, so are many cards found on the market.

[Guest jdouce]

a2sd doesnt move core os it just moves the apps HOWEVER is u move the dalvik cache of a NON-ODEXED rom ull have issues since the odex files go to the SD

im using a 16gb Class 2.... but its san disk, speed test is a read speed of 3.8 (i tested it) .... not enuf to be branded class 4 but not far off and im using an odex rom so the .odex files are in /system/app

This is a REALLLLYY anoying issue since i use my phone tethered for internet.

Edited May 16, 2010 by jdouce

[Guest afiorillo]

a2sd doesnt move core os it just moves the apps HOWEVER is u move the dalvik cache of a NON-ODEXED rom ull have issues since the odex files go to the SD [...]

Dalvik-cache actually IS a core OS component, since the Android Java VM ("Dalvik") relies on it for program execution. It's a critical component, so MAYBE a faulty or slow memory card can mess up the entire thing while some initial setup.

That can explain the randominess of this issue.

[Guest jdouce]

ok maybe ... but i wonder if any using class 4+ is having this issue and if so then why wouldn't removing the a2sd fix it

Edited May 16, 2010 by jdouce

[Guest jdouce]

ok the USB-MS function in recovery

Press to enable

Press again to disable

Could this be disabling the USB-MS functionality completly?

cat /sys/devices/platform/msm_hsusb/usb_function_switch

this file IS there on my device which explains why ADB is working but WHY arnt the other usb functions working from within android

usb_mass_storage:enable

adb:enable

diag:disable

serial:disable

ether:disable

modem:disable

nmea:disable

also my

/sys/devices/platform/msm_hsusb/power/wakeup is blank duno if its supposed to be

device manager only shows ADB as a device on windows

Edited May 16, 2010 by jdouce

[Guest geek78]

I have never used the a2sd or the USB mount from the recovery and my desire has been bricked....

[Guest kiokoman]

ok the USB-MS function in recovery

Press to enable

Press again to disable

Could this be disabling the USB-MS functionality completly?

imho it's the culprit, probably people unaware of that enabled it and then restart the phone without disabling it , after that you are not able to switch that function again. :?

[Guest Specialist79]

Just my 2 cents:

AFAIK the phone comes with a configuration data area most likely in NVRAM (probably connected via I²C). This configuration contains parameters such as the device serial #, the regional branding, wifi rf calibration data and the kernel command line. Having had a look at the kernel sources, it contains much more. Some of these parameters are interpreted by the kernel, some are just ignored (cf. dmesg output). The same applies to the kernel command line parameters.

IMHO, some (yet unknown) event has altered this configuration data on the affected devices. IMHO, just re-flashing will not help as this operation will almost certainly not overwrite device-specific factory settings.

I do have a couple of ideas what could cause this change, though:

The first one depends on the type of exploit used to root the device: Does that involve any buffer overflows? If so, the exploit could have unexpected side effects. @Paul?

Another idea is that the device could detect that it is stuck in a reboot loop and fall back into a kind of safe mode. Did anyone face the issue without having had a reboot loop before the issue manifested itself?

As far as a fix is concerned: I'd probably disassemble the hboot code to better understand what's going on, but unfortunately my trial version of IDA Pro does not seem to come with ARM support (contrary to what the web site says). There also seems to be a factory mode built into the device, which can apparently be activated by providing the androidboot.mode=factory2 on the kernel command line. This seems to activate a kind of serial console according to the source. Use at your own risk...

Edited May 16, 2010 by Specialist79

[Guest kiokoman]

Just my 2 cents:

AFAIK the phone comes with a configuration data area most likely in NVRAM (probably connected via I²C). This configuration contains parameters such as the device serial #, the regional branding, wifi rf calibration data and the kernel command line. Having had a look at the kernel sources, it contains much more. Some of these parameters are interpreted by the kernel, some are just ignored (cf. dmesg output). The same applies to the kernel command line parameters.

IMHO, some (yet unknown) event has altered this configuration data on the affected devices. IMHO, just re-flashing will not help as this operation will almost certainly not overwrite device-specific factory settings.

I do have a couple of ideas what could cause this change, though:

The first one depends on the type of exploit used to root the device: Does that involve any buffer overflows? If so, the exploit could have unexpected side effects. @Paul?

Another idea is that the device could detect that it is stuck in a reboot loop and fall back into a kind of safe mode. Did anyone face the issue without having had a reboot loop before the issue manifested itself?

As far as a fix is concerned: I'd probably disassemble the hboot code to better understand what's going on, but unfortunately my trial version of IDA Pro does not seem to come with ARM support (contrary to what the web site says). There also seems to be a factory mode built into the device, which can apparently be activated by providing the androidboot.mode=factory2 on the kernel command line. This seems to activate a kind of serial console according to the source. Use at your own risk...

when i was creating a custom rom for myself i had a lot of loop restart and i don't have any issue, have flashed mine over 20 times so we can exclude a fall back into safe mode

[Guest mardox]

why is it that usb bricked devices suffer such a battery drain? not that i am a programmer but what would the connection between the two be? it seems my standby time is about the same as before but while using the device it drains a huge chunk more battery then before.

[Guest yabolek]

Another thing I was wondering is: which class the used MicroSD was? Since many core OS pieces are moved to the memory card by A2SD, a Class4+ card is mandatory: the stock provided one is a Class2 card, not suitable to be used with A2SD, so are many cards found on the market.

I have a class 2 16GB microsd card and all is fine (my usb is not bricked), I did not see any performance decrease when dalvik cache was moved to SD.

[Guest NOFX]

It seems so many people are joining this debate now - which is good :)

But are we actually nearing a solution? It seems to me that all that has been met are dead-ends, which completely sucks!

As my Market disappears every time I restart my phone, I really hate my Desire at this point. Do you think we will ever find a solution to this? I'm getting a bit sceptical. Is Paul still working on a solution, or are people starting to give up on os?

[Guest staba]

The worst thing is the fact that it's not completly clear what is causing this :). If we knew that people could prevent it or solve the issue completly. Really needs to be solved or it will hurt development for the desire big time in the long run.

[Guest NOFX]

The worst thing is the fact that it's not completly clear what is causing this ;). If we knew that people could prevent it or solve the issue completly. Really needs to be solved or it will hurt development for the desire big time in the long run.

That's exactly what I'm thinking. I'm hoping I will be able to unroot my device - will try later tonight. If that is doable I will send my device in for repair, and if I get a new one I'm not sure I have the guts to re-root it. Which means I lose 50% of the fun of my device.

This really sucks. Which is why I hope Paul and the rest of the developers will find a fix for this - sooooooooon :)

[Guest staba]

Let's hope so. I mean, my device is fine but as I said earlyer in the post, I poop my pants everytime I flash something because it could brick. Even though I don't have issues now, I wouldnt mind pledging in some kind of bounty thread for 20 euros for the one to solve it. Just so I can rest easy when I flash new fun things to my device.

[Guest NOFX]

Let's hope so. I mean, my device is fine but as I said earlyer in the post, I poop my pants everytime I flash something because it could brick. Even though I don't have issues now, I wouldnt mind pledging in some kind of bounty thread for 20 euros for the one to solve it. Just so I can rest easy when I flash new fun things to my device.

Same here - if 20 euros can save us, that would be excellent - but is this just a 'money-issue'?

[Guest staba]

Same here - if 20 euros can save us, that would be excellent - but is this just a 'money-issue'?

No ofcourse it's not a money issue. I'm not trying to buy Paul or anyone. I just wanted to say that it's very important to me and alot of people to get it out of the way even for people who do not have these issues (yet).

[Guest chrmrt]

Is it possible to unroot one of these USB bricked devices? If i can do that i am going to have a shot at getting a new one from the store this week...

The ROM i need installed is the generic android 2.1 build 1.15.405.4 without root

Thank you.

[Guest NOFX]

Is it possible to unroot one of these USB bricked devices? If i can do that i am going to have a shot at getting a new one from the store this week...

The ROM i need installed is the generic android 2.1 build 1.15.405.4 without root

Thank you.

Yep - just follow the 'Unroot guide' here on MoDaCo. You have to do it via the Fastboot menu - but apart from that the guide should be selv-explanatory.

Good luck!

[Guest afiorillo]

I have never used the a2sd or the USB mount from the recovery and my desire has been bricked....

This is a major tombstone thrown upon the A2SD theory, thank you for sharing. At least, no more time-wasting researches in the wrong direction.

[Guest jdouce]

this as a really anoying problem ...

and i cant remember WHEN it appeared i flashed countless times.

seems noone can pinpoint WHEN exactly it occured however i can say it DID NOT occur during my initial root as i used the USB to transfer further roms and flashed sucessfully

i belive this occured when i flashed an a2sd update zip but i flashed a r2 rom at the same time without reboots in between

[Guest Specialist79]

Just spotted a couple of interesting strings in the hboot file:

usbtestmode: <value>

value=0 -> disable

value=1 -> enable

readconfig

Can somebody with a bricked device test whether that is a valid fastboot command?

Edited May 16, 2010 by Specialist79

[Guest Specialist79]

this as a really anoying problem ...

and i cant remember WHEN it appeared i flashed countless times.

seems noone can pinpoint WHEN exactly it occured however i can say it DID NOT occur during my initial root as i used the USB to transfer further roms and flashed sucessfully

i belive this occured when i flashed an a2sd update zip but i flashed a r2 rom at the same time without reboots in between

Could it be that flashing twice in the same fastboot session triggers the issue?

[Guest jdouce]

2 flashed in same fastboot rules out i flashed twice during my initial root ... instantly flashed r1 and was fine

btw the guys over at acer liquid had a similar issue .... caused by a2sd

every1 on here

go to android market

get USB Switcher

it says get pro or somthing ignore it

connect to pc

use it (its a widget u gota add it to home)

see if usb ms works ... it worked for me

now i got usb mass storage back

if this works for every1

its a HTC software issue

actually ... is every here using a non sense version?

saying that im only partiually usb bricked .. still have acess in fastboot

Edited May 16, 2010 by jdouce