USB EP behave irradically after time.

I have a system with an iMx25/EHCI host connected to an HL8548. The Linux ttyACM0 has a pppd connection and an application issuing intermittent ‘AT+CSQ?’ and ‘AT+CREG?’ queries on ttyACM3. Everything works fine for some time (10-120min) and then the application indicates that it is getting no response. However, the pppd connection is still completely functional and I can run network traffic.

I stop the application and attached kermit to the port. Often there is no data. It seems that some times it recovers. The last case, I had the device still responding with ‘OK’. Ie, I issues an ‘ATI’ command from kermit and received only ‘OK’. I tried many command like ‘AT+GMR’, ‘ATZ’, etc and they all just returned ok (or ‘error’). I have instrumented with ‘usbmon.ko’ and I updated libpcap/tcpdump and grabbed ‘/dev/usbmon1’ captures. I looked at the captures in wireshark and it seems that the modem is actually just returning ‘OK’. Ie, it doesn’t seem to report any data for commands on this port.

Is the code on the HL8548 side having a separate task/thread per logical connection? Is it possible that the logical connection gets messed up? If I completely re-initialize the device, things recover. However, I would prefer to only reset the ‘ttyACM3’ by some means and not the whole device. The application is setup for a Telit device and may send some AT commands that are non-standard. Could the USB host controller I am using cause this issue somehow?Dont see this in traces Ie, Consistently kill a particular EP? We are using this USB controller with different USB devices without issues.

After reboot,

Any suggestions or recommendations for debugging and fixing this issue?

Here is another Tcpdump/Wireshark USB capture as the issue occurs,

Frame ‘137488’ is the first ‘bad AT’ command with response ‘137492/137494’.
Frame ‘137424’ has a good response at ‘137439/137446’. This is a summary from a 15MB capture.

A good ‘AT+CSQ’ reply (instead of ‘137494’) looks like,

Where the summary doesn’t show the ‘\x0a\x0d’ line breaks in the data. I disconnected the data ttyACM0 and the AT commands
behave the same way on that interface. ‘ATZ’ and ‘AT&F’ don’t seem to reset anything. The ‘AT&V’ actually prints something in this mode. It seems that the command processor on the HL8548 breaks after some time?

Hi,
Could you please check whether the issue is present in the latest FW version?

Thanks,
Rex

I am trying to update the firmware, but I have some issues. Mysteriously, my hardware engineer sent me a new firmware file yesterday. This has some Windows USB loader. I am trying to sniff the traffic on the Windows machine to get a working sample of the firmware update protocol. The exe (in RAR) is RHL85xx.5.5.11.0.201412191148.x6250_1.exe. I do not normally use Windows. I have the development board working with Linux and the ttyACM are installed automatically. The Windows developers tell me I need some ‘INI’ files (and multi-ttyACM/COM port) drivers for Windows. Did I miss a Sierra Wireless package some where?

When we run the update ‘exe’ with the development platform, then RHL*.exe times out and gives some ‘assert/debug’ to say it waited 1minute for an update. I think this is because we need to install windows drivers/files to update the chip soldered on the HL6/HL8 developer platform. We can see multiple CDC ACM nodes in the Windows ‘device manager’.

I have the USB Firmware guide. I will need to write the software to do the update on our iMx25 target and/or get a pin out of the developer platform socket so that we can hook our radio daughter board to the developer boards to update the firmware. I see there are ‘SL developer socket’ kit manuals; but I do not see an HL developer socket kit. It is referred to as AirPrime HL8548x Socket Board User Guide, Reference Number: 4114852. I can not find that document; but maybe I don’t need that (see below).

Is this the latest firmware ‘RHL85xx.5.5.11.0.201412191148.x6250_1’?
Can you point me to resources to get the HL6/HL8 developer board/system (or HL8548 USB) working with Windows?
Is there any info on updating the firmware (besides AirPrime_HL75xx_and_HL854xx_Firmware_Download_Protocol_Rev1_0.pdf)?
If we create a JIG to power/reset our radio daughter board and provide USB connection to a PC, is this enough to use the Windows firmware updater? If so, ignore next question.
Do you have a pin-out for the socket connectors CN200/CN201 for this developer board?

Apparently my Ubuntu 14.04 is more user friendly than Windows for USB serial ports.

I guess I am conditioned to think that the answers to these are no,

Do you have firmware update for a Linux Host system?
Do you give source code for this Linux host firmware update?

After some time, I understand different Firmware update processes. Without a Windows PC, the AirVantage system seems to be the only way to update firmware prior to revision 5.5.15.xx of the firmware.

It think the USB issue has disappeared, but now I have different issues with v5.5.14.xx firmware. I guess that the HL8548 software is still beta. No person told me this at my company; hw just delivers a board. Anyways, it seems that June-Dec 2015 might be a date to start expecting some stable software for these modules?