https://bugzilla.kernel.org/show_bug.cgi?id=208959 --- Comment #1 from Johan Hovold (johan@xxxxxxxxxx) --- On Wed, Aug 19, 2020 at 02:52:47AM +0000, bugzilla-daemon@xxxxxxxxxxxxxxxxxxx wrote: > https://bugzilla.kernel.org/show_bug.cgi?id=208959 > > Bug ID: 208959 > Summary: USB-serial (mct_u232): bit7=1 when tty is set to cs7? > Product: Drivers > Version: 2.5 > Kernel Version: 5.7.0 > Hardware: All > OS: Linux > Tree: Mainline > Status: NEW > Severity: low > Priority: P1 > Component: USB > Assignee: drivers_usb@xxxxxxxxxxxxxxxxxxxxxx > Reporter: mslade@xxxxxxxxxxxxxxxx > Regression: No > > I'm not 100% certain this is a bug, please let me know if this is expected > behavior. > > I have hooked an old serial mouse up to a USB serial device, USB ID 0711:0230 > (part of an circa 2000 Targus port replicator) > > when I configure it with the appropriate 1200 baud and 7 bit char width, the > bytes that come through /dev/ttyUSB0 all have bit7 set to 1. > > This behavior is in contrast to an old serial port on a box running a 2.4.19 > kernel. > > Is there a standard for the value that non-data bits should be set to? This sounds like a configuration issue. If the mouse is indeed using 7n1 and the converter is expecting 8n1 you could end up with bit 7 always set on input. Can you verify that the converter is configured correctly, for example, by reading out the termios settings, enabling debugging and/or connecting it to another serial port configured for 7n1? Johan -- You are receiving this mail because: You are watching the assignee of the bug.
