Issues with USB connectivity etc. and root / A2SD

[Guest drvdijk]

If everybody stops flashing, no progress will be made. Then again, no more bricked devices. Ghe :-)

I'm glad there still are people brave enough to flash!

[Guest Normola]

I was just checking the android kenel repos, and spotted a recent change:

adb: Add Qualcomm USB id

Is this likely to make any difference? I'd compile adb from source and give it a go, but my desire isn't bricked (sorry).

android.git.kernel.org

Edited May 25, 2010 by Normola

[Guest ciairo]

I was just checking the android kenel repos, and spotted a recent change:

adb: Add Qualcomm USB id

Is this likely to make any difference? I'd compile adb from source and give it a go, but my desire isn't bricked (sorry).

android.git.kernel.org

Are you familiar with compiling adb? There are plenty of people here who would be happy to try! :rolleyes:

[Guest Normola]

Are you familiar with compiling adb? There are plenty of people here who would be happy to try! :rolleyes:

I've never compiled adb before. I'm comfortable compiling stuff in linux, but windows stuff would take a bit longer as I'd have to get a proper environment set up.

Edited May 25, 2010 by Normola

[Guest adi90]

I was just checking the android kenel repos, and spotted a recent change:

adb: Add Qualcomm USB id

Is this likely to make any difference? I'd compile adb from source and give it a go, but my desire isn't bricked (sorry).

android.git.kernel.org

If you can compile it for linux i will try :rolleyes:

[Guest xTc is loVe]

Have you tried to use the same version as your carrier's?

Yes i have to wait until a new Update arrives from my Carrier :rolleyes: I asked them but without number and contract data and so on..... :D Have to wait until someone like Paul fix it -.-

[Guest ciairo]

If you can compile it for linux i will try :rolleyes:

so will I. Thanks Normola!

[Guest xTc is loVe]

I was just checking the android kenel repos, and spotted a recent change:

adb: Add Qualcomm USB id

Is this likely to make any difference? I'd compile adb from source and give it a go, but my desire isn't bricked (sorry).

android.git.kernel.org

Im so interested if this work! Like this can give us ADB Access via This Qualcomm thing!?

[Guest Normola]

Complied. Just looking for a place to upload it. I have no idea if this will work or not.

[Guest Normola]

Anyone care to test?

Just try ./adb devices or ./adb shell to see if you can make a connection.

adb.tar.gz

Make sure that any old versions aren't running before trying this. Use 'sudo ./adb kill-server' or 'sudo pkill adb' first, then launch the new adb with 'sudo adb root'

Edited May 26, 2010 by Normola

[Guest Biliskner]

Anyone care to test?

Just try ./adb devices or ./adb shell to see if you can make a connection.

adb.tar.gz

Ah crap I would test, but I'm at work, all WinXP machines here...

[Guest Normola]

Ah crap I would test, but I'm at work, all WinXP machines here...

Don't get your hopes up anyway :rolleyes:

It probably wont be any different from the original version.

[Guest adi90]

not working ... :rolleyes:

[Guest Biliskner]

Don't get your hopes up anyway :D

It probably wont be any different from the original version.

I'm running TestRUU ROM so have been anxiously waiting, but that's fading... I only want to flash custom ROMs, if anyone can sign ROMs with HTC keys (lol), post here pls! :rolleyes:

[Guest Normola]

not working ... :rolleyes:

I'm working on a few other ideas at the moment...

Out of interest, if you unplug your desire, then plug it back in, what does dmesg in linux report?

*edit*

Also, the output from lsusb would be useful.

*edit2*

Looking at my own lsusb output (working phone) the device id shows as 0x0bb4:.... which is the vendor ID for HTC. It's possible that when it moves to the Qualcomm test mode, this vendor id changes to something else.

If so, it may just be a case of adding this new id to the list of vendor ids in adb (assuming it isn't the latest Qualcomm addition).

Edited May 26, 2010 by Normola

[Guest Normola]

This would be useful from anyone with a non-working phone. You can get the device id in windows like this.

[Guest myth1001]

Would love to help but I don't have a bricked phone to start with. :rolleyes:

Hope someone out there will be able to supply the details you need.

And hope there will be some progress! :D

[Guest Normola]

Would love to help but I don't have a bricked phone to start with. :rolleyes:

Hope someone out there will be able to supply the details you need.

And hope there will be some progress! :(

This is the first time I ever wanted a bricked phone. I could check it myself then :D

[Guest MIYU-G]

I'm working on a few other ideas at the moment...

Out of interest, if you unplug your desire, then plug it back in, what does dmesg in linux report?

*edit*

Also, the output from lsusb would be useful.

*edit2*

Looking at my own lsusb output (working phone) the device id shows as 0x0bb4:.... which is the vendor ID for HTC. It's possible that when it moves to the Qualcomm test mode, this vendor id changes to something else.

If so, it may just be a case of adding this new id to the list of vendor ids in adb (assuming it isn't the latest Qualcomm addition).

may be it's a good way to go on

[Guest Normola]

Actually, looking at this post, it probably won't help. The id is the recently added Qualcomm one.

There are some pretty serious problems with the Qualcomm device vs. the HTC device.

iSerial 0

bInterfaceClass 255 Vendor Specific Class

bInterfaceSubClass 255 Vendor Specific Subclass

bInterfaceProtocol 255 Vendor Specific Protocol

iInterface 0

I don't think that adb can be made to talk to it :rolleyes:

Interestingly though: Changing the connection mode from Charge Only to any of the others makes the subclass change. Does this cause any additional error messages in dmesg on bricked phones?

Edited May 26, 2010 by Normola

[Guest Normola]

A quick grep through the kernel source files has thrown up some possible issues.

Fastboot.c

int match_fastboot(usb_ifc_info *info)

{

	if(!(vendor_id && (info->dev_vendor == vendor_id)) &&

	   (info->dev_vendor != 0x18d1) &&  // Google


	   (info->dev_vendor != 0x0451) &&

	   (info->dev_vendor != 0x22b8) &&  // Motorola

	   (info->dev_vendor != 0x0502) &&

	   (info->dev_vendor != 0x413c) &&  // DELL

	   (info->dev_vendor != 0x0bb4))	// HTC

			return -1;

	if(info->ifc_class != 0xff) return -1;

	if(info->ifc_subclass != 0x42) return -1;

	if(info->ifc_protocol != 0x03) return -1;

	// require matching serial number if a serial number is specified

	// at the command line with the -s option.

	if (serial && strcmp(serial, info->serial_number) != 0) return -1;

	return 0;

}

This would return -1 with the Qualcomm device

Not sure if altering this would have any effect though.

[Guest Normola]

It's late, I need to sleep before work and I've spammed the forum enough tonight.

I don't suppose anyone has tried the following?:

usage: fastboot [ <option> ] <command>

commands:

update <filename> reflash device from update.zip

flashall flash boot + recovery + system

flash <partition> [ <filename> ] write a file to a flash partition

erase <partition> erase a flash partition

getvar <variable> display a bootloader variable

boot <kernel> [ <ramdisk> ] download and boot kernel

flash:raw boot <kernel> [ <ramdisk> ] create bootimage and flash it

devices list all connected devices

reboot reboot device normally

reboot-bootloader reboot device into bootloader

options:

-w erase userdata and cache

-s <serial number> specify device serial number

-p <product> specify product name

-c <cmdline> override kernel commandline

-i <vendor id> specify a custom USB vendor id

-b <base_addr> specify a custom kernel base address

Perhaps modify the step1.bat/sh file to use the so that all the calls to fastboot use the qualcomm vendor id instead?

For example:

fastboot reboot -i 0x05c6

seems to work in linux. I'd think it would be identical in windows.

I'm off to bed.

Edited May 26, 2010 by Normola

[Guest eViL D:]

It's late, I need to sleep before work and I've spammed the forum enough tonight.

I don't suppose anyone has tried the following?:

Perhaps modify the step1.bat/sh file to use the so that all the calls to fastboot use the qualcomm vendor id instead?

For example:

fastboot reboot -i 0x05c6

seems to work in linux. I'd think it would be identical in windows.

I'm off to bed.

fastboot works fine for me, it's the adb that doesn't talk.

[Guest Normola]

fastboot works fine for me, it's the adb that doesn't talk.

Could you give my newer version a shot?

[Guest Biliskner]

Could you give my newer version a shot?

is that a recompiled version? or the fastboot command?