Hi,

Olimex has an eval board called SAM7EX256 and they promote freertos. I have the board and have downloaded the freertos version 5.1.2. It turns out that Olimex have a code example that is based on freertos version 4.0.2.As the Olimex changes done to version 4.0.2 appeared to be minimal. I attempted to include the changes in version 5.1.2.

The web server page comes up but the network statistics page shows many dropped packages.

I was curious to see if you have any recommendation for what version I should use?

My compile environment is set up with Eclipse and Yagarto. The evaluation is almost working so it may be a minor detail or it could be related to that I am not using the best version for the compiler setup. Have you heard about a similar problem before and if so what did it relate to?

My changes to 5.1.2 are minimal, namely PHY_ID, powerdown bit inversion and PHY address. In addition I adjusted the clock value to 48.054857Mhz.

I see that some Makefiles are using arm_source only and others thumb_arm mixed. Could this relate to the problem with dropped packages?

Thanks!
JMA

ARM or Thumb mode should not make any difference.

Its always best to use the latest version, but I don't think this will make any difference to the dropped packets you are seeing.

Can you step through the code to see why the packets are being dropped, you can search for the code that increments the drop statistic, place a break point there, then work backwards to find the reason? Is there a heavy traffic load on the network.?

I did a quick test by placing a function call to "abc(int a){tm=a;}" at each place "++uip_stat.ip.drop" is done in uip.c. ( tm is a "volatile int tm";)

Then placed a breakpoint on abc....the app never stops but the "uip_stat.ip.drop" does increment.

I also noted that the stat page in the web browser has a number in the right column that is one line above the first text line in the left column...and one extra number below the last left text line. (24 text lines and 26 numbers)

No load on the network. Could be a stack problem....any thoughts?