P-SeriesのトピックRe: Packet loss on eTSEC controller (P1020RDB)

Packet loss on eTSEC controller (P1020RDB)

ayazshahid — Mon, 29 Apr 2013 14:56:29 GMT

Hi,

I've ran into a strange issue with eTSEC controller on P1020RDB. It seems like there's a fair amount of packet loss when I PING my target. PING with smaller payloads are okay but the problem becomes severe when I try to use a ping with 1024 bytes payload.

While troubleshooting the issue, It seems like the packets are being properly received by the controller but response packets are not transferred on the wire by eTSEC.

Here's some info about my environment.

Tx and Rx buffers are 8-byte aligned.
I've checked the RMON MIB registers and there's no collision / overflow / underflow condition.
The data buffers and descriptors are being updated properly.
PING with smaller payload is working properly.

I'll appreciate your comments.

Thanks!

Re: Packet loss on eTSEC controller (P1020RDB)

lunminliang — Fri, 03 May 2013 00:58:17 GMT

Please see below technical support comment:

Based on this information it is not possible to figure out the root cause of the problem. PING is a user-level utility that accesses the network through the services provided by the IP stack, which is not implemented in Freescale hardware.

If you believe that the problem is related to eTSEC hardware, please provide a hardware level problem description, namely:

1. Actions taken by the software with respect to eTSEС (register or shared memory structure

updates) that immediately precede the problem manifestation and/or suspected to

cause the problem.

2. Observed response of the hardware (if any) to the actions specified in p. 1 and why

this response is considered incorrect.

3. Any explicit error indications returned by eTSEC in BDs and IEVENT.