On Thu, Jun 03, 2021 at 05:56:32PM -0500, Alex Villacís Lasso wrote: > El 3/6/21 a las 16:48, David Frey escribió: > > Hi Johan, > > > > I have added linux-usb@xxxxxxxxxxxxxxx to the CC as requested. > > > It appears that in the 5.12.8 kernel, there is a failure every time > > cp210x_tiocmset_port is called resulting from this: > > ret = cp210x_write_u16_reg(port, CP210X_SET_MHS, control); > > Sorry to butt in, but if you are comfortable compiling kernel code, try > manually reverting only commit f61309d9c96a308465bec9d2e5206da265b075a0 > (USB: serial: cp210x: set IXOFF thresholds) and loading the resulting > .ko module. I do that to keep working in my ESP32 projects while this > issue gets a proper fix. More details at bug report at > https://bugzilla.redhat.com/show_bug.cgi?id=1965527 Yeah, that would likely work here too. > Just a guess: maybe ulXonLimit and ulXoffLimit should only be non-zero > when actually enabling software flow control, and must be set to zero > otherwise? The faulty commit just assigns them unconditionally to 128. I had the same thought and that's why I asked you to try setting ixon and ixoff using stty. Unfortunately, that didn't seem to have any effect, judging from the last logs you sent me. Johan
