Date: Sat, 8 Feb 1997 14:20:03 -0800 (PST) From: Bruce Evans <bde@zeta.org.au> To: freebsd-bugs Subject: Re: kern/2695: sio1 (16540 serial port) is not recognized by GENERIC boot kernel Message-ID: <199702082220.OAA03953@freefall.freebsd.org>
next in thread | raw e-mail | index | archive | help
The following reply was made to PR kern/2695; it has been noted by GNATS. From: Bruce Evans <bde@zeta.org.au> To: freebsd-gnats-submit@freebsd.org, garman@phs.k12.ar.us Cc: Subject: Re: kern/2695: sio1 (16540 serial port) is not recognized by GENERIC boot kernel Date: Sun, 9 Feb 1997 09:06:46 +1100 >I turned on verbose probe using flags 0x80 and it tells me: > >sio1: probe test 5 failed >sio1: probe test 6 failed >sio1 not found at 0x2f8 > >My working 2.1.6.1-RELEASE kernel reports: > >sio1 at 0x2f8-0x2ff irq 3 on isa >sio1: type 16450 >>How-To-Repeat: >boot a 2.2-gamma kernel with my modem installed on com2 :-) >>Fix: >not enough delay to allow the modem to respond to probes, perhaps? 2.2 delays 1000 usec longer than 2.1.6 between test4 and test 5. I have seen failures for test 3 and test 4 on a UMC8669F (16550 incompatible). I guess this is caused by fifo mode and possibly some other things not being cleared after a crash, and bugs in UMC's control registers. I recently found that the UMC's fifo control register doesn't work either - if it is written to while input is arriving, then the fifo trigger level is not changed until input stops arriving. There are also problems with writing to the divisor latch registers while input is arriving. So initialization of the divisor latch and the fifo in the probe can't be trusted to work (the fifo isn't initialized now). Bruce
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199702082220.OAA03953>