> In my over night test the workaround triggered 31 times. The test has found no problems. This is frustrating: how did you manage to get that much triggers ;-) I ran 4 targets in parallel and got 3 fifo errors in total after 20 h and no test failures. I will now switch my setup from doing the selft-receiption test to a two-node-same-host transmission test which is where I first detected the error. BTW: "... (applies) against current net/main" - is this kh-lingo for using this: git clone https://git.kernel.org/pub/scm/linux/kernel/git/netdev/net.git/ ? Oh and I "backported" the patched driver source into my 6.0.12 raspbian kernel. mfg Stefan