send/receive packet

Hi,

If I understand,   cc3100 on TivaC is client,  which connected  successfully to the  server,  send packet, receive  packet , and failed with the seconds send.

Can you sniff  the all session ?  ( from the first SYN )

Does server is your implementation ? if yes, can you make sure server receive  the first packet correctly , and make sure server does not call close immediately?

Regards

Nir  

Hi Nir

Yes, I made one test server on my local laptop and I'm sure that I can get the first data correctly and the server isn't closed until my test done.

I find the cause of error return, that should be over the failed criteria of g_pCB->SocketTXFailure. I guess the algorithm should be if the SocketTXFailure over the criteria, the socket will be treated as the invalid and it can't work anymore.

In my observation, g_pCB->SocketTXFailure is 0 at the bootup and it will be increased in some operations. In my test case, g_pCB->SocketTXFailure is 0x0A before the first send packet, and g_pCB->SocketTXFailure becomes 0x0B after the first receive packet, so the second send packet hits the criteria and gets the error return.

May I know the design propose of g_pCB->SocketTXFailure and is it the necessary?

thanks
Gavin

Hi,

Does your application use more than single socket ? what is the socket id you recieved ?

If you remove the checking  g_pCB->SocketTXFailure at  _SlDrvDataWriteOp, does server recieve the second packet ?

Can you sniff packets ?  

 Regards

Nir

Hi Nir

No, I only use one socket in my test case, and the socket ID is 0.

If I remove the checking  g_pCB->SocketTXFailure at  _SlDrvDataWriteOp, and then the server can receive the second packet and the following packets.

So I'd like to know if the checking  g_pCB->SocketTXFailure is the necessary item in cc3100 implementation?

thanks

Gavin

Hi Gavin,

Does socket return 0 (ID)  ? (should be something else )

The  g_pCB->SocketTXFailure is important, you should use  g_pCB->SocketTXFailure mechanism

Pls increas stacks size ,  cc3100 driver stacks, application stack and try again , is it help ?

Nir 

Hi Nir

I'm sorry, the socket return is 0x10, not 0.

Do you know how to increase cc3100 driver stack?

I'm curious about one point, if I only increase the stack but not reset g_pCB->SocketTXFailure, g_pCB->SocketTXFailure still will hit the criteria in someday, right?

thanks
Gavin

Hi,

locate all places, you use FREE RTOS create thread (xTaskCreate), increase usStackDepth

Regards

Nir 

Hi Nir

I have increased the doubled usStackDepth in VStartSimpleLinkSpawnTask, but it can't help this issue.

when g_pCB->SocketTXFailure is over 0x0a, the error return will occur there.

I also verify the g_pCB->SocketTXFailure in the sample project of SDK 0.5, after several operations, g_pCB->SocketTXFailure will be 5 or 6. I guess the perfect case should be 0, right?

Could you please have more comment about g_pCB->SocketTXFailure? Why it is important in cc3100 implementation? 

thanks
Gavin 

Hi,

I will call you offline.

Meanwhile, can you let me know the version you use, call the following code after sl_start  

SlVersionFull ver;

unsigned long

statusWlan;

unsigned char

pConfigOpt, pConfigLen;

pConfigOpt = SL_DEVICE_GENERAL_VERSION;

sl_DevGet(SL_DEVICE_GENERAL_CONFIGURATION,&pConfigOpt,&pConfigLen,(

unsigned char

*)(&ver));

printf(

"CHIP %d\nMAC 31.%d.%d.%d.%d\nPHY %d.%d.%d.%d\nNWP %d.%d.%d.%d\nROM %d\nHOST %d.%d.%d.%d\n"

,

ver.ChipFwAndPhyVersion.ChipId,

ver.ChipFwAndPhyVersion.FwVersion[0],ver.ChipFwAndPhyVersion.FwVersion[1],

ver.ChipFwAndPhyVersion.FwVersion[2],ver.ChipFwAndPhyVersion.FwVersion[3],

ver.ChipFwAndPhyVersion.PhyVersion[0],ver.ChipFwAndPhyVersion.PhyVersion[1],

ver.ChipFwAndPhyVersion.PhyVersion[2],ver.ChipFwAndPhyVersion.PhyVersion[3],

ver.NwpVersion[0],ver.NwpVersion[1],ver.NwpVersion[2],ver.NwpVersion[3],

ver.RomVersion,

SL_MAJOR_VERSION_NUM,SL_MINOR_VERSION_NUM,SL_VERSION_NUM,SL_SUB_VERSION_NUM);

Thanks

Nir

Hi Nir

Please refer to the attached screenshot

HOST: 0.0.5.1

Hi,

I am interesting with NwpVersion , it`s not reasonable values, you may need to refresh it,  or can you check it with SimpleLinkStudio environment ?

Regards

Nir

Hi Nir

I don't know why I can't init cc3100 in SimpleLinkStudio environment, I don't debug it.

Where can I get the latest binary to reflash? with uniflash?

BTW, I use cc3100 bp 2.0B.

thanks
Gavin

Hi,

Pls send me, the version you can see on the top of the chip.

Regards

Nir

Hi Nir

XCC3100GZ

thanks

Gavin

Gavin,

The board version you have was intended for early access development.  We highly recommend you get an updated board to use with the latest SDK and tools.

It is possible to use Uniflash and upgrade the board you have and test out again, however, Uniflash currently does not support service pack upgrade.

I will make an announcement this Friday or beginning of next week for a new Uniflash version that supports service pack upgrade.

Shlomi

Hi Shlomi

About this issue, I'm still interesting in g_pCB->SocketTXFailure. 

Could you please provide more debug information about this counter?

Or everything will be fine in the latest board?

thanks

Gavin

Hi Gavin,

Please note I'm going to announce this week a new Uniflash version that is going to support ServicePack programming (along with some other improvements). I suggest then that you upgrade your board and it should work fine.

Regarding g_pCB->SocketTXFailure, it is not a counter but a bit mask per socket. It belongs to the flow control mechanism on the TX side. If you have any error on the socket (for example sl_socket fails or you get a disconnect from other side), you should have the corresponding bit set to 1.

Regards,

Shlomi