EK-TM4C1294XL: TM4C1294

Hello Felix,

Those aren't file we provide. I checked various resources and we don't provide any .cgi files.

Those lines are for 'get' calls so it makes me think those files are being received from the webserver itself.

I will try and ask my more Ethernet savvy colleague for ideas but we aren't webserver experts, I am not sure if he has sufficient background to add more value.

Best Regards,

Ralph Jacobi

Felix Valladares said:

We need to modify webserver for our application and need to modify these cgi files.

Section 8.5 Writing CGI Functions in TM4C129x Ethernet Applications for Socket-Based TIRTOS NDK has some information on writing CGI functions, albeit using the TI-RTOS NDK rather than the FreeRTOS+lwIP used in SPMA072

Felix Valladares said:

Do you know where can I find these files?

The CGI files are not actually files in the embedded file system on the Tiva HTTPD server, but rather the server recognises the .cgi file extension and calls a function in C to handle the request.

Looking in the config.c file within smpa072.zip there is the following array which maps the .cgi filename to a function which is called:

//*****************************************************************************
//
//! This array is passed to the HTTPD server to inform it of special URIs
//! that are treated as common gateway interface (CGI) scripts.  Each URI name
//! is defined along with a pointer to the function which is to be called to
//! process it.
//
//*****************************************************************************
static const tCGI g_psConfigCGIURIs[] =
{
    { "/config.cgi", ConfigCGIHandler },            // CGI_INDEX_CONFIG
    { "/misc.cgi", ConfigMiscCGIHandler },          // CGI_INDEX_MISC
    { "/defaults.cgi", ConfigDefaultsCGIHandler },  // CGI_INDEX_DEFAULTS
    { "/ip.cgi", ConfigIPCGIHandler },              // CGI_INDEX_IP
};

Thanks for clarifying Chester! I will take a look.

Felix

Thanks Ralph.

Thanks Charles! You are correct. This fixed the issue of not accepting new static IP address. It works fine as long as I don't re-start. If I cycle the power on the TIVA, forcing it to reset, it goes back to DHCP mode trying to acquire a dynamic IP address. I need it to remain in static IP mode. I tried to disable LWIP_DHCP in lwipopts.h but didn't help.

Thanks again for your help.

Felix

Charles, is it possible to make static IP address the default? 

Felix

Hi Felix,

  You can configure for static IP by default. This is how you do it.

1. In the lwipopts.h file, you need to set the below two #define to 0.

#define LWIP_DHCP                       0
#define LWIP_AUTOIP                     0

2. In the lwip_task.c file, go to the lwIPTaskInit() function.

Replace:
lwIPInit(g_ui32SysClock, pui8MAC, 0, 0, 0, IPADDR_USE_DHCP);

By:
lwIPInit(g_ui32SysClock, pui8MAC, (192u << 24) | (168u << 16) | (254 << 8) | 136, 0xFFFF0000, 0xFFFF0000, IPADDR_USE_STATIC);

The above static IP address is an available IP address that I use for testing on my network. You need to enter your own static IP address. 

3. Rebuild the project and then run.

Hi Charles, this works. However code above uses a fixed static IP address. We need to be able to modify static IP address using webserver. For some reason, as I said, I am able to change static IP address to any value using your webserver, however, when I reset TIVA, it reverts back to DHCP mode. It tries to acquire IP address from DHCP server and comes back with an arbitrary dynamic address.

It doesn't stay in static IP mode with the address I just programmed. Is this normal? It seems by looking at the code that webserver should keep last programmed static IP address. I don't know if you observe same behavior on your end.

Felix

Hi,

Felix Valladares said:

when I reset TIVA, it reverts back to DHCP mode. It tries to acquire IP address from DHCP server and comes back with an arbitrary dynamic address.

I don't think that is possible. If you followed the instructions to set LWIP_DHCP and LWIP_AUTOIP to 0 then the DHCP module is completed removed from the stack after you rebuild the project. You can confirm that in wireshark and you should not see any DHCP transactions. As I mentioned you also need to add the below line so that lwIP is initialized with your static address. After each power cycle, the IP address will be whatever you provided. In my below example, I will only see 192.168.254.136. 

lwIPInit(g_ui32SysClock, pui8MAC, (192u << 24) | (168u << 16) | (254 << 8) | 136, 0xFFFF0000, 0xFFFF0000, IPADDR_USE_STATIC);

Hi Charles, there seems to be a misunderstanding. The example code you provide works as you describe. If I follow your code, there is no DHCP transactions and IP address is static. The problem is that I need to change the address using the webserver. You see, our end customers need to change IP address to whatever value they want. This IP address needs to be saved in non-volatile memory and retrieved on power up. I am having difficulty doing this in the original code. I am trying to modify it. I think, after power up,  I need to provide lwIPInit() with the static IP address I entered using webserver. 

Felix

Hi Felix,

  Glad you solve the problem yourself. Yes, you can first retrieve the new static address that is saved in the EEprom before you call lwIPInit(). I have not tried it myself yet. Would you mind to share the code that it may benefit others in the future? Thanks.