bing@galbp.LBP.HARRIS.COM (Bing Bang) (01/26/88)
well, i finally found that one bug that kept me from posting my tty/rs232 driver. too late though it seems. i'll report the bug anyway. seems the reason we lose characters from the rs232 ports is that the interrupts are off in too many places for too long a time. i fixed this by leaving the ints on and masking off the unwanted ones at the controller. this caused the bug to show up. in the routine cp_mess, there should be a "cld" right after the "cli" instruction, in fact, i went through all movs[bw] instructions in the assembler source files and made sure the direction flag is set correctly. this fix made my rs232 fast enough for me to use a simple cu to talk to my local bbs's at 2400 baud without dropping a single character. i have a feeling the same could be done for jim's driver... bing -- Bing H. Bang +--------------------------------------------------------+ Harris/Lanier |OS/2 on PS/2: Half an operating system on half a machine| Atlanta GA +--------------------------------------------------------+
jaime@killer.UUCP (James Dasilva) (02/01/88)
In article <3468@galbp.LBP.HARRIS.COM> bing@galbp.UUCP (Bing Bang) writes: >well, i finally found that one bug that kept me from posting my tty/rs232 >driver. too late though it seems. i'll report the bug anyway. ^^^^^^^^^^^^^^^^^^^^^^^^ It's never too late. If you have a working tty driver that you've been testing for a long time, Post it! The net can only benefit from having a choice of drivers. It may be that good ideas from both your drivers could be used in a "definitive" tty driver. As a well-known desperate politian says: "Let the voters decide" --------------------------------------------------------------------------- James A. da Silva Sysop: The Mars Hotel BBS 301-470-3569 ihnp4!killer!jaime featuring Comp.Os.Minix traffic online