Hi,Im using the NET+50 with Linux-2.0.38. About every 4-20 GB of Network transfer the netarm suddenly stops transfering data. Network is dead. I have to restart network (ifconfig eth0 down; ifconfig eth0 … up). Seem like the netarm doesn’t generate any RX-Interrupts anymore. Sometimes when the ethernet interface is idle I get a lot of Not Ready Interrupts from the DMA-Controller. May be the strange timer Handler for restarting the FIFO is responsible for this, but may be this is unrelated to the hangs. Does anybody have the same problem ? Jochen
Check the 50 errata to see the stall workarounds…
I am using NET+50 with NET OS4.0 to implement SNMP agent and using SNMP browser to test it. About 2~7 days the network is dead, no response to ping and SNMP call. Network relive after reset NET+50, what’s wrong with it? James.
I have found the problem. Its not a missing RX-Interrupt. The TX-Interrupt is lost and the dev->tbusy flag in the netarmeth Linux driver will be 1 forever. I have created a workaround, which restarts sending a packet after a timeout. But I don’t know why the Interrupt is lost. Other drivers don’t loose Interrupt (CAN-Bus driver has no lost Interrupts on 100 Million transmits). Im using linux-2.0.38