Hi Johan, Our last exchange regarding my proposed ftdi_sio driver patch adding the needed quirk for DUART28C took place 3 weeks ago on October 5. Your last argument against my patch was this one: https://marc.info/?l=linux-usb&m=160189545216969&w=2 And here is my rebuttal to that argument made the same day: https://marc.info/?l=linux-usb&m=160192817717108&w=2 In your October 5 argument you wrote: > Let me give this some more thought. It has been 3 weeks - do you have any more thoughts that address my not yet answered rebuttal arguments in defense of my patch? My key arguments are: * The "standard" behaviour of Linux and other Unix-derived OSes of unconditionally asserting DTR & RTS on tty port open (or upon leaving B0 state) without giving userspace any ability to say "no, please don't do it" is a philosophical design bug, one that goes all the way back to original 1970s UNIX - but long-standing tradition does not make right. * Particularly in the present age of USB-serial adapters with LVCMOS rather than RS-232 electrical signals, this Unix/POSIX/Linux serial port handling philosophical design bug is hampering hardware engineers' ability to produce otherwise clean and elegant circuit designs. * Because Windows does NOT exhibit the same philosophical design bug in this regard as Unix/POSIX/Linux, there may be hw devices that were made for use with Windows, that depend on glitch-free DTR and/or RTS, and which will fail to work correctly with current Linux. When such cases occur, the party at fault is Linux and not the hardware design - the hardware engineers were merely exercising their natural right to make simple and elegant circuit designs that work well with OSes that are free of philosophical design bugs, it is not our fault as hw engineers that Linux inherited a philosophical design bug from Ancient UNIX by way of POSIX. * A minimally invasive surgical solution in the form of driver quirks that suppress the traditional DTR & RTS behaviour for those specific hw devices for which it is unacceptable is more practical than trying to fix the root-cause Unix philosophical design bug 45 y too late. I would appreciate a response to my arguments. Sincerely, Mychaela, custom hardware designer, she/her/hers
