975232 - [Tarako] USB is unstable and sometimes is hard to connect

Status: VERIFIED FIXED

(Firefox OS Graveyard :: General, defect)

Description

[Danny Liang [:dliang]]

There are two cases to cause we cannot connect device by USB:

Case 1: We cannot recognize the device and there is nothing by lsusb.
In this case, we have pre-setting as below:
- Add rule in /etc/udev/rules.d/51-android.rules:
  SUBSYSTEM=="usb", ATTR{idVendor}=="1782", ATTR{idProduct}=="<idProduct>", MODE="0666"
- Add device's PID in ~/.android/adb_usb.ini
  $echo "0x1782" >> ~/.android/adb_usb.ini

Case 2: We can recognize the device, but adb still does not work.
We use two cables, lsusb showed different results.

Cable A:
Bus 001 Device 002: ID 8087:0024 Intel Corp. Integrated Rate Matching Hub
Bus 002 Device 002: ID 8087:0024 Intel Corp. Integrated Rate Matching Hub
Bus 003 Device 002: ID 0461:4de2 Primax Electronics, Ltd
Bus 001 Device 001: ID 1d6b:0002 Linux Foundation 2.0 root hub
Bus 002 Device 001: ID 1d6b:0002 Linux Foundation 2.0 root hub
Bus 003 Device 001: ID 1d6b:0002 Linux Foundation 2.0 root hub
Bus 004 Device 001: ID 1d6b:0003 Linux Foundation 3.0 root hub
Bus 001 Device 017: ID 1782:5d05 Spreadtrum Communications Inc.
Bus 001 Device 005: ID 0bda:0161 Realtek Semiconductor Corp. Mass Storage Device

Cable B:
Bus 001 Device 002: ID 8087:0024 Intel Corp. Integrated Rate Matching Hub
Bus 002 Device 002: ID 8087:0024 Intel Corp. Integrated Rate Matching Hub
Bus 003 Device 002: ID 0461:4de2 Primax Electronics, Ltd
Bus 003 Device 011: ID 1782:5d06 Spreadtrum Communications Inc.
Bus 001 Device 001: ID 1d6b:0002 Linux Foundation 2.0 root hub
Bus 002 Device 001: ID 1d6b:0002 Linux Foundation 2.0 root hub
Bus 003 Device 001: ID 1d6b:0002 Linux Foundation 2.0 root hub
Bus 004 Device 001: ID 1d6b:0003 Linux Foundation 3.0 root hub
Bus 001 Device 005: ID 0bda:0161 Realtek Semiconductor Corp. Mass Storage Device

Comment 1

[Stephen Donner [:stephend] Not actively reading bugmail]

FWIW, we have our active device set up and accessible (as long as you have access to Mozilla Mountain View's VPN).

The IP is visible from http://qa-selenium.mv.mozilla.com:8080/computer/b2g-6/log; check the first line "[02/19/14 16:45:16] [SSH] Opening SSH connection to [xxx.xxx.xxx.xxx]" - I'll give you the credentials offline, in case it helps us all troubleshoot.

Comment 2

[Stephen Donner [:stephend] Not actively reading bugmail]

Danny, Thomas: this doesn't bode well for continuous integrations (via Jenkins); can we/have we escalate(d) this to our vendor?

Comment 3

[Stephen Donner [:stephend] Not actively reading bugmail]

(In reply to Danny Liang [:dliang] from comment #0)
> There are two cases to cause we cannot connect device by USB:

> Case 2: We can recognize the device, but adb still does not work.
> We use two cables, lsusb showed different results.
 
Any notable differences in the cables?  (There shouldn't be?)  What's your setup look like?  We have no such problem with the Inaris or Hamachis we've set up this way, for what it's worth.

Comment 4

[thomas tsai]

We have informed the issue to vendor last week and an internal bug is created in vendor. I have similar issue, too. I wonder the usb signal quality, timing and usb descriptors are not well configured in kernel.

Comment 5

[James Zhang (Spreadtrum)]

(In reply to thomas tsai from comment #4)
> We have informed the issue to vendor last week and an internal bug is
> created in vendor. I have similar issue, too. I wonder the usb signal
> quality, timing and usb descriptors are not well configured in kernel.

Thomas, we need reproduce path. For example, script.

Comment 6

[cmtalbert]

(In reply to James Zhang from comment #5)
> (In reply to thomas tsai from comment #4)
> > We have informed the issue to vendor last week and an internal bug is
> > created in vendor. I have similar issue, too. I wonder the usb signal
> > quality, timing and usb descriptors are not well configured in kernel.
> 
> Thomas, we need reproduce path. For example, script.

Hi James, thanks for being involved here. I don't believe there is a script to reproduce the issue. We simply plug the phone into usb and sometimes we can see it via the adb devices command and sometimes we cannot. To make matters worse, simply performing the lsusb command will sometimes show the device and sometimes not, and sometimes the lsusb reports different device IDs for the same device. :Dliang, is that a correct summary?

Comment 7

[Danny Liang [:dliang]]

(In reply to Clint Talbert ( :ctalbert ) from comment #6)
> (In reply to James Zhang from comment #5)
> > (In reply to thomas tsai from comment #4)
> > > We have informed the issue to vendor last week and an internal bug is
> > > created in vendor. I have similar issue, too. I wonder the usb signal
> > > quality, timing and usb descriptors are not well configured in kernel.
> > 
> > Thomas, we need reproduce path. For example, script.
> 
> Hi James, thanks for being involved here. I don't believe there is a script
> to reproduce the issue. We simply plug the phone into usb and sometimes we
> can see it via the adb devices command and sometimes we cannot. To make
> matters worse, simply performing the lsusb command will sometimes show the
> device and sometimes not, and sometimes the lsusb reports different device
> IDs for the same device. :Dliang, is that a correct summary?
Yes, it's correct
As I know, this issue has fail rate and might be depended on device/cable, so I don't think we need a script to reproduce this issue.
Pre-discussed with James, he might need more information below due to he cannot reproduce on his side.
1. How is the charging icon when issue happened
2. Please provide slog with timestamp to analysis.

How to get slog:
1. The rom must be eng/userdebug build, slog was default enable.
2. Insert SD card.
3. slog is located /sdcard/slog/

Comment 8

[Stephen Donner [:stephend] Not actively reading bugmail]

(In reply to Danny Liang [:dliang] from comment #7)
> How to get slog:
> 1. The rom must be eng/userdebug build, slog was default enable.
> 2. Insert SD card.
> 3. slog is located /sdcard/slog/

Raymond's out tomorrow + Thursday, so need-info'ing myself to try to get this info back into this bug.

Comment 9

[Stephen Donner [:stephend] Not actively reading bugmail]

I've asked Raymond to take a look at comment 7, this coming Monday (I'll be working remotely, and didn't have luck today).

Comment 10

[raymond [:retornam] (inactive)]

(In reply to Danny Liang [:dliang] from comment #7)
> (In reply to Clint Talbert ( :ctalbert ) from comment #6)
> > (In reply to James Zhang from comment #5)
> > > (In reply to thomas tsai from comment #4)
> > > > We have informed the issue to vendor last week and an internal bug is
> > > > created in vendor. I have similar issue, too. I wonder the usb signal
> > > > quality, timing and usb descriptors are not well configured in kernel.
> > > 
> > > Thomas, we need reproduce path. For example, script.
> > 
> > Hi James, thanks for being involved here. I don't believe there is a script
> > to reproduce the issue. We simply plug the phone into usb and sometimes we
> > can see it via the adb devices command and sometimes we cannot. To make
> > matters worse, simply performing the lsusb command will sometimes show the
> > device and sometimes not, and sometimes the lsusb reports different device
> > IDs for the same device. :Dliang, is that a correct summary?
> Yes, it's correct
> As I know, this issue has fail rate and might be depended on device/cable,
> so I don't think we need a script to reproduce this issue.
> Pre-discussed with James, he might need more information below due to he
> cannot reproduce on his side.
> 1. How is the charging icon when issue happened
> 2. Please provide slog with timestamp to analysis.
> 
> How to get slog:
> 1. The rom must be eng/userdebug build, slog was default enable.
> 2. Insert SD card.
> 3. slog is located /sdcard/slog/

There wasn't a folder named slog found on the SD card. Please needinfo me if you need more information. Thanks

Comment 11

[Danny Liang [:dliang]]

(In reply to raymond [:retornam] (needinfo? me) from comment #10)

> > How to get slog:
> > 1. The rom must be eng/userdebug build, slog was default enable.
> > 2. Insert SD card.
> > 3. slog is located /sdcard/slog/
> 
> There wasn't a folder named slog found on the SD card. Please needinfo me if
> you need more information. Thanks

James, is there any possibility to cause no slog in sdcard?

Comment 12

[James Zhang (Spreadtrum)]

(In reply to Danny Liang [:dliang] from comment #11)
> (In reply to raymond [:retornam] (needinfo? me) from comment #10)
> 
> > > How to get slog:
> > > 1. The rom must be eng/userdebug build, slog was default enable.
> > > 2. Insert SD card.
> > > 3. slog is located /sdcard/slog/
> > 
> > There wasn't a folder named slog found on the SD card. Please needinfo me if
> > you need more information. Thanks
> 
> James, is there any possibility to cause no slog in sdcard?

Remove sdcard or user build.

Comment 13

[raymond [:retornam] (inactive)]

(In reply to James Zhang from comment #12)
> (In reply to Danny Liang [:dliang] from comment #11)
> > (In reply to raymond [:retornam] (needinfo? me) from comment #10)
> > 
> > > > How to get slog:
> > > > 1. The rom must be eng/userdebug build, slog was default enable.
> > > > 2. Insert SD card.
> > > > 3. slog is located /sdcard/slog/
> > > 
> > > There wasn't a folder named slog found on the SD card. Please needinfo me if
> > > you need more information. Thanks
> > 
> > James, is there any possibility to cause no slog in sdcard?
> 
> Remove sdcard or user build.

I removed the sdcard  and inserted it into an sdcard reader so I could check. I had to do that because the host machine could not detect the phone.

Comment 14

[James Zhang (Spreadtrum)]

Danny and Thomas, we have a usb patch today, you can repo sync and get this patch in device/sprd.

commit 27a4b15ae8c526f538409504acb86643c0cd4daa
Author: lianxiang.zhou <[e-mail]>
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

Comment 15

[cmtalbert]

James, thank you!

Danny and Thomas, please need-info myself and Raymond when you get this patch into a build for us to try out.

Thanks!

Comment 16

[raymond [:retornam] (inactive)]

Hi All,
Has the patch been merged? If so please let me know the steps forward. Thanks

Comment 17

[thomas tsai]

yes, it's in. Please repo sync again and rebuild the whole image.

Comment 18

[raymond [:retornam] (inactive)]

(In reply to thomas tsai from comment #17)
> yes, it's in. Please repo sync again and rebuild the whole image.

Please let me know where the images are located and if I need some special permissions to access them.

Comment 19

[raymond [:retornam] (inactive)]

Hi I downloaded the flash script and flashed the device I have on hand. The device is now detected by both the lsusb and adb commands after several flashes. We can mark this bug as resolved.

Comment 20

[James Zhang (Spreadtrum)]

OK, I think Lianxiang has fixed this issue.

Comment 21

[raymond [:retornam] (inactive)]

marking as verified fixed. Thanks everyone