Posts by fs-support_ZU

Hello,

maybe I missunderstood. "We would like to create a process of a shell-like application which takes it command via STDIN and prints out the results via STDOUT" this is your "console".

Think this is the way "Pocket CMD v 8.00 " calls "cild processes". See below this works. Or what is your parent proccess?

Hello,

i wonder if it is not sufficient to create the process and just wait for it. STDIN, STDOUT should be still your "console"?

Pseudo code:

CreateProcess(_T("ndcucfg"), ... , nonewconsole, ... , processinfo, ...):

...

WaitForSingeObject(processinfo.process, oo);

Hello,

sorry for delay.

We are working on it and can offer you the kernel within the next two days.

Hello,

you can download eboot v2.6 now and we will do the WEC7 release for armStoneA9R4 within this week.

Hello,

you received a PCN from you distributor that for change from armStoneA9R2 to armStoneA9R4 new kernel an N-/EBoot are required!

You have to use EBoot V2.6 for armStoneA9R4. Previous EBoot do not know the "R4" board!

Sorry, there is an issue in kernel V3.5 concerning armStoneA9R4 and WEC7. We provide you corrected V3.6 asp.

>> We found new r4 are not exactly the same board as we was told (only PHY IC was changed) and seems there are many differences, even PCB route is totally new

<< Only routes concerning PHY should have changed. For details you may contact our HW department.

Hello,

>> in my case always say HKLM\System\RTC\BatteryGood=1, even disconnecting battery .. is this right ??

<< This value is only read while boot, so you have to reboot after remove Vbat.

>> ...power, "minutes" field are lost and start "00" again.

<< I am a bit helpless about the situation. I test it again and it works also there are not changes in HW refer to RTC. And i don't think RTC on two boards are demaged.

Does the issues persit if you connect only Ethernet and serial debug line? Maybe we have some side effects. You can get/set date and time via telnet -> command "date" and "time".

Can you send one board direct to my desk for investigation?

Hello,

sorry eboot menu for watchdog is invalid. HW watchdog is always on when kernel is loaded from Flash.

If kernel hangs or threads below priority 100 block for about 4s the HW watchdog fires.

The API refers the WEC SW watchdog.

Do you have threads running or adjust driver priorities below 100?

Are there any additional serial debug information in case of watchdog reboot?

Hello,

I did two checks with kernel V3.5 and armStoneA9 HW rev. 1.20. One with default time zone one with time zone UTC+1. In both cases I cannot reproduce the issue.

I did following steps in both cases:

1. power on set date and time

2. power off for about 1min

3. power on, check date and time -> OK

4. power off for about 10min

5. power on, check date and time -> OK

When I remove the battery the debug line shows the follwing information:

...

WARNING: SNVS was reset.

RTC: VBAT is low or not available!

[OAL] MACB: Disabled

[OAL] RestartReason: IPP

...

Ths information is also available under HKLM\System\RTC\BatteryGood=dword:<zero=NOK or one=OK>

Hello,

1) all setting you do via control panel (or ndcucfg) are stored in RAM. For store it into Flash you have to do a "reg save" (which wraps RegFlushKey()). That is valid for all our board and it was always like this all the time.

2) Ok, I will check. This is an unknown issue!

>> This new kernel version is showed as: V3.05 Built Jan 25 2023 in FSBoadConfig utility instead of V3.50

<< Thanks, this is a formating bug we try to remove it in the next version. Both are wrong it should be V3.5.

Hello,

a,b) the external battery buffers only the "sytem time" in the RTC. The time zone/DST options (local time) have to be saved in Registry. This is correct!

c) I am not sure if I understood. The hour does change after each reboot?

Can you please test with kernel V3.5. I just put it into the download.

Hello fstendike,

thanks for this information. But it is strange. On my side my suggestion works for several hours without any issues!

About your desire above, don't think you need the ndisconfig source.

Try DLL import similar according the functions below:

Samples for DLL import of "CreateFile", "CloseFile" and "DeviceIoControl" (note, may be overloaded in different ways) will be available here: PWM sample. Also assume all const. will be found in ndis.h.

Hello fstendike,

does the problem still persit on your side?

Or you found any workaround? This would be interesting for me too.

I test the kernel V3.4 with Ethernet driver V1.3 this does not change the behavior. But we can exclude Ethernet driver cause the issue.

If I use the control panel's applet "Network" to assign DHCP Ip <-> static IP the issue does not occur! I test it ten times.

But I must admit when I use following (pseudo code) it does not work reliable when I do not wait for some time at * (in my case I wait 6s). After additional 4s -15s I retreive a valid and useable "CurrentIpAddress" by GetAdatersInfo() :

do{

RegistrySet/ClearDHCP

*

IOCTL_NDIS_UNBIND_ADAPTER

IOCTL_NDIS_BIND_ADAPTER

Poll GetAdaptersInfo();

}while(true)

So I assume to use the Registry to enable DHCP is not the correct way because of a timing issue. As far as I know you use PB to create your image so maybe you can figure out how the control panel enabels DHCP;).

Hope this helps.

Hello,

ok, thanks for the update.

>> ... but for software ...

<< It should not related to the BSP. Assume if you use same catalog and project settings as for A9X it should behave in the same way on A9XR2.

>> This is what I did for efusA9XR2 to get the same behaviour as on efusA9X (since we did not use explicit RegFlushKey() on efusA9X either and registry persisted).

<< I can't believe this. The only difference between A9X and A9XR2 is the Ethernet phy!

And bit 1 is set. Flags it a evaluated bit wise and 3=1|2.

PS: Sorry you said RegistryFlags I thought about Flags. Nevertheless there are no differencies between A9A and A9XR2 and I would use RegFlushKey.

Hello,

thanks for this information.

Flash:

I don't know anything a "auto reg flush feature" but I would not use it at all.

Keep two things in mind, write cycles to Flash are limited and using FAT, power off while writing may demage the FAT. So I would prefer explicite call of "reg save" (which implicite calls OS function RegFlushKey()).

SD:

which SD Card you are talking about? Can you give me the pins or index in Registry.

PS: do delete your [HKEY_LOCAL_MACHINE\init\BootVars] in OSDesign.reg and use default. This is safe.

Else msdn is your friend.

Dear Vladimir,

thanks for this information. I will check it asp.

Differencies between Flash- and SD drivers on efusA9XR2 and efusA9X are marginal and our standard image works on efusA9XR2. So I assume it may be a configuration issue while the BSP creation.

Sorry, no i did not test it.

Hope I can spend some time next week. At the time we are bound on other projects.

Hello Bernard,

>> Our procedure was made for PicoCOM4 (2011) and we receive PicoCOM4.2 without WARNING message that indicates a board version change!!

<< You or your distributor got a PCN about the required changes between PicoCOM4 and PicoCOM4.2! And when you get PicoCOM4.2 you had explicit ordered them.

>> Is it a way to recover the possibility to upload this time the last eboot_pc4_112.nb0 file

<< If you can still step into NBoot (startup + DCUTerm + s) you can try to download EBoot via serial (DCUTerm->File->transmit binary file->...) or via USB (NetDCUUSBLoader). Use command "d" for this purpose.

Hope this helps.

Hello,

great if everything about configuration is clear now.

You can use "SetFilePointer" to jump from port to port *or* you can use DeviceIoControl(..., IOCTL_DIO_GET/SET/CLR_PIN, ...) to address one single IO by one single call.

You will find sample code here: DeviceDriverDoc.

Hello,

can you please tell me the driver version which is working/not working and if the problem is related to efusA9/efusA)R2 or both.

Thanks.