I discovered another latent bug in the RX->TX forwarding path while working on RSC. This one already has sufficient spinxsk test coverage to hit, but we simply had a low rate of crash dumps being successfully uploaded to AzWatson. Eventually I was able to repro locally and fix.
I discovered another latent bug in the RX->TX forwarding path while working on RSC. This one already has sufficient spinxsk test coverage to hit, but we simply had a low rate of crash dumps being successfully uploaded to AzWatson. Eventually I was able to repro locally and fix.