I have an issue on a custom K64 board - when doing flood ping to it, it stopped responding after few seconds.
The tools I use: KDS 2.0.0, KSDK 1.1.0 on windows.
I tried to debug and get some info from the ENET_get_stats() it did not show me any error. Further testing shows me that
once it stopped, the total received packages keep increasing but total sent did not increase.
My questions are: does anyone notice the same issue on any K64 board or it is just me? How do I debug in this case? I suspect it may be related to TCPIP_task() not schedule the Ping reply.
updates: I disabled my extra tasks(mDns, NTP, Telnet, I2C, etc.), just run the minimum threads below - it seemed no issue at all. So I will add back my other task and see how it reacts.
SIZE | USED | TASK ID | NAME |
00001488(000005D0) | 00000740(000002E4) 00010001 | WATCHDOG service | |
00000320(00000140) | 00000124(0000007C) 00010002 | _mqx_idle_task | |
00001328(00000530) | 00000956(000003BC) 00010003 | Main | |
00001584(00000630) | 00001244(000004DC) 00010004 | TCP/IP |
The reason for the failure is because I have a IRQ (DMA) at priority level 3. The timer IRQ priority is 4 for K64 in MQX. Even though I changed to use 3 for the timer it still fails. But change to 4 for both then it works.
Now here comes the question: Why IRQ priority 3 does not work on K64? (it worked on K60).