Other Parts Discussed in Thread: AM3352, TPS65217, TRF7970A, DAC7512, PMP, CC2520, TCA6416, TCA8418, LM8323, LM8333, ADS7846, TSC2004, TSC2005, TSC2007, TPS65218, PCF8574, DRV2665, DRV2667, TPIC2810, LP8727, BQ24190, BQ24257, BQ24735, BQ25890, LM63, LM70, LM73, LM77, LM80, LM83, LM87, LM90, LM92, LM93, LM95234, LM95241, LM95245, LM25066, TPS40422, ADC128D818, ADS1015, ADS7828, ADS7871, AMC6821, INA209, INA3221, TMP102, TMP103, TMP108, TMP401, TMP421, LP3943, TPS65010, TPS65086, TPS65090, TPS65910, TPS65912, TWL6040, LM3533, LP3971, LP3972, TPS51632, TPS62360, TPS65023, TVP5150, TFP410, DA8XX, LM3630A, LM3639, PCM1681, PCM3168A, TAS2552, TAS5086, TLV320AIC23, TS3A227E, TPA6130A2, LM3530, LM3642, LP3944, LP5521, LP5523, LP5562, LP8501, DAC124S085, TCA6507, CDCE706, CDCE925, ADC128S052, ADC161S626, ADS7950, ADS8688, TLC4541, AFE4403, AFE4404, OPT3001, LMP91000,
I have been chasing what appears to be packet corruption when using the CAN bus at reasonably high utilization.
We have a similar product already in production for a few years utilizing am AM3352 CPU that does not exhibit this issue when running the same software and kernel, on hardware of the same design.
The issues arise when sending/receiving frames from the same ID's with different data at intervals of 50us to 1ms pacing.
I have verified the actual bus traffic with external tools and a logic analyzer, on RX we are getting packets with corrupt data, but the DLC and byte count is still correct. On the TX side, the actual bus traffic is wrong and not what we load to the driver via SocketCAN.
I feel like I'm casing my tail on this issue, so I'm putting it out there for potential ideas from other users.