USB hangs up when video conference call lasting for several seconds in performance mode

cancel
Showing results for 
Show  only  | Search instead for 
Did you mean: 

USB hangs up when video conference call lasting for several seconds in performance mode

1,208 Views
lqw39
Contributor III

Hi, all,

We are using mx6q MCIMX6Q6AVT10AC platform with SMSC9514 usb ethernet controller.

1. if system working in performance mode, USB host always hangs up after video conference call started, so SMSC ethernet controller stopped, the video call only lasts for several seconds.

   kernel log:

ADDRCONF(NETDEV_CHANGE): eth0: link becomes ready
smsc95xx 2-1.1:1.0: eth0: link up, 100Mbps, full-duplex, lpa 0xCDE1

root@android:/ # ------------[ cut here ]------------
WARNING: at net/sched/sch_generic.c:257 dev_watchdog+0x278/0x29c()
NETDEV WATCHDOG: eth0 (smsc95xx): transmit queue 0 timed out
Modules linked in:
[<c004a634>] (unwind_backtrace+0x0/0xfc) from [<c0075ffc>] (warn_slowpath_common+0x4c/0x64)
[<c0075ffc>] (warn_slowpath_common+0x4c/0x64) from [<c00760a8>] (warn_slowpath_fmt+0x30/0x40)
[<c00760a8>] (warn_slowpath_fmt+0x30/0x40) from [<c0454a7c>] (dev_watchdog+0x278/0x29c)
[<c0454a7c>] (dev_watchdog+0x278/0x29c) from [<c0081e08>] (run_timer_softirq+0x124/0x244)
[<c0081e08>] (run_timer_softirq+0x124/0x244) from [<c007bf18>] (__do_softirq+0xa8/0x170)
[<c007bf18>] (__do_softirq+0xa8/0x170) from [<c007c22c>] (irq_exit+0x90/0x98)
[<c007c22c>] (irq_exit+0x90/0x98) from [<c003d378>] (do_local_timer+0x70/0x90)
[<c003d378>] (do_local_timer+0x70/0x90) from [<c004300c>] (__irq_svc+0x4c/0xe8)
Exception stack(0xc077df68 to 0xc077dfb0)
df60:                   c07e76e0 80000093 00000001 00000000 c077c000 c07e0444
df80: c05887cc c0796314 1000406a 412fc09a 00000000 00000000 00000000 c077dfb0
dfa0: c0052174 c004415c 40000013 ffffffff
[<c004300c>] (__irq_svc+0x4c/0xe8) from [<c004415c>] (default_idle+0x24/0x28)
[<c004415c>] (default_idle+0x24/0x28) from [<c0044534>] (cpu_idle+0xbc/0xfc)
[<c0044534>] (cpu_idle+0xbc/0xfc) from [<c000898c>] (start_kernel+0x2a0/0x2f0)
[<c000898c>] (start_kernel+0x2a0/0x2f0) from [<10008040>] (0x10008040)
---[ end trace b3d5190fb3501ae2 ]---
imx-ipuv3 imx-ipuv3.0: IPU Warning - IPU_INT_STAT_10 = 0x00000001

2. if system working in balance mode, video conference call can last for very long time (half an hour), but the video quality becomes pool, and has high latency.

however, we can play online video and upload/download huge data without any problem.

Anybody has ideas?

Labels (2)
12 Replies

945 Views
YixingKong
Senior Contributor IV

Qingwei, what is your update after our enginers' response? We would expect your response in 3 days, otherwise your discussion will be closed.

Thanks,

Yixing

0 Kudos

945 Views
sinanakman
Senior Contributor III

Hi Yixing, I have a general question : what is a "closing" of a discussion ? Isn't this a community discussion board ? Can you please point me at an official Freescale document that defines how the discussions are "closed" and what all that rules are. This way we will all know and can refer to others.

Regards

Sinan Akman

0 Kudos

945 Views
YixingKong
Senior Contributor IV

Sinan

When a discussion issue has been answered by anyone and the orignator of the dicussion issue click Correct Answer/Helpful Answer, we will say it is closed. The Freescale engineers who answered the discussion issue may not come here again. However, anyone who would like to make comment on the closed discussion issue could do it continuously. After a discussion issue is closed, it stays in the community for anyone to share ot refer to.

Unfortunately, the precise definition is not available now. We may add it some time late. Hopefully my explanation will make you clear about your question.

Regards,

Yixing

0 Kudos

945 Views
sinanakman
Senior Contributor III

Hi Yixing, thanks for the explanations and all that makes sense.

I was referring to part of what you said earlier : "expect your response in 3 days, otherwise your discussion will be closed."

Understandable after couple days your engineers can stop tracking the question however I'd expect no one other than the originator himself/herself should mark an answer "Correct" or "Helpful" to make it "closed".

That's the point I was trying to make it clear.

Thanks

Sinan Akman

0 Kudos

945 Views
YixingKong
Senior Contributor IV

Sinan, thank you for your comment/suggestion. My comment  "expect your response in 3 days, otherwise your discussion will be closed." would be for those orignators who would not come back here again once his/her discussion has been created. If you or someone else indicate that it is necessary for the discussion to stay, we would let the discssion open for enough time. As you can see that the discssion orginator has not been seen, we may close the discussion after a period of time.

Thanks,

Yixing

0 Kudos

945 Views
sinanakman
Senior Contributor III

Closing the unattended discussion after some time is fine but as mentioned earlier marking an answer "Correct" or "Helpful" just to close it is what makes no sense. That's all I am bringing up. Please always let the originator decide if the answer is correct or helpful. If you like to close it, do it without marking correct or helpful arbitrarily. Otherwise the "correct" or "helpful" marking will not be accurate.

Regards

Sinan Akman

0 Kudos

945 Views
YixingKong
Senior Contributor IV

Sinan, you are right. We close a discussion with no Correct/Helpful answer selected while only if the originator would not come back.

Thanks you again for your suggestion.

Yixing

0 Kudos

945 Views
sinanakman
Senior Contributor III

Hi Yixing, thanks for confirming this. So to summarize :

- if you respond to a question but if the originator does not follow up for some time, you "close" the discussion internally meaning that your engineers will stop tracking it

- only the originator marks an answer as correct or helpful

- if the originator does not follow up, the question stays unanswered.

Hope this covers it all. Thanks for clearing up.

Regards

Sinan Akman

0 Kudos

945 Views
YixingKong
Senior Contributor IV

Sinan

Nice summary.

Thanks,

Yixing

945 Views
LeonardoSandova
Specialist I

I wonder if this is a temperature related problem.

0 Kudos

945 Views
qiang_li-mpu_se
NXP Employee
NXP Employee

Did you try this USB ethernet adapter on Freescale refernce board with last BSP?

0 Kudos

945 Views
qiang_li-mpu_se
NXP Employee
NXP Employee

Another thing can be tried: in kernel config, enable the followed feature:

CONFIG_USB_EHCI_TT_NEWSCHED=y

0 Kudos