rbthomas@athos.rutgers.edu (Rick Thomas) (07/29/89)
I an attempting to bring up Bruce Evans' protected mode Minix.
I have run into two snags.
1) When I try to compile the kernel with the Bruce's TTY driver I get
an undefined tag '_rs232' and it refuses to asld.
2) My printer (an Epson FX86e) seems to be dropping bunches of
characters.
Neither of these problems occurs when I am working with vanilla 1.3,
and problem 1 does not occur with the stock tty driver as modified by
the "non-'B-'" patch files.
All the CRC's came out right, so I am not working with munged files.
What is missing?
Rick
========================================================================
Rick Thomas, Manager
Rutgers University, College of Engineering
Piscataway, NJ 08855-0909
Phone: (201) 932-4301
Internet: rbthomas@jove.rutgers.edu
UUCP: {any backbone site}!rutgers!jove.rutgers.edu!rbthomas
========================================================================
--
Rick Thomas
uucp: {ames, att, harvard}!rutgers!jove.rutgers.edu!rbthomas
internet: rbthomas@JOVE.RUTGERS.EDU
bitnet: rbthomas@zodiac.bitnet
Phone: (201) 932-4301rbthomas@athos.rutgers.edu (Rick Thomas) (08/01/89)
I found the problem. I had forgotten to tell "axmakefile" that I had
the new tty driver. (When all else fails -- read the instructions!)
BTW -- the new tty driver line in axmakefile refers to "rs2.o", I had to
change it to "rs2.s" to get it to work. I guess Bruce has a "real"
(separate) assembler and loader.
Rick
--
Rick Thomas
uucp: {ames, att, harvard}!rutgers!jove.rutgers.edu!rbthomas
internet: rbthomas@JOVE.RUTGERS.EDU
bitnet: rbthomas@zodiac.bitnet
Phone: (201) 932-4301rbthomas@athos.rutgers.edu (Rick Thomas) (08/01/89)
So, after I got a protected mode kernel with the Bruce's TTY driver
made, I went to test it out on my genuine IBM PC/AT (Old slow model --
6MHz clock) with my Multitech V.32 9600 bps modem. After all, Bruce
claimed that it worked on a 4.77MHz XT at 19.2Kbps. An AT should be at
least as fast as that! (On the other hand, he did mention that access
to the Extended memory RAM disk might keep interrupts locked for more
than one character time on this configuration. So I was curious.)
The results are mixed. It's seems to be OK for reading netnews and
using vi for editing on the remote system, but bulk data transfer to
disks (RAM or Hard) has problems with dropped characters (less with
output to the RAM disk, more with output to the Hard disk.)
Running kermit in protocol mode to download from a SUN to Minix with
the receiving file on Hard disk using 1000 character packets gets
frequent garbled packets and retransmissions. The garbles seem to to
be associated with the moments when "update" wakes up and causes writes
to the disk. Putting the output file on the RAM disk did not have any
problems with garbles.
On the other hand, turning on kermit's "session logging" to just grab
everything that comes over the serial port and log it to RAM disk
resulted in dropping large chunks of data. It appears that something
is going on that is not fast enough to keep up with 9600bps without
pauses. (I didn't try logging to Hard disk.) (??) All that message
passing may be slowing things down (??).
The next thing to try is enabling hardware (CTS/RTS) flow control.
Rick
--
Rick Thomas
uucp: {ames, att, harvard}!rutgers!jove.rutgers.edu!rbthomas
internet: rbthomas@JOVE.RUTGERS.EDU
bitnet: rbthomas@zodiac.bitnet
Phone: (201) 932-4301rbthomas@athos.rutgers.edu (Rick Thomas) (08/01/89)
> BTW -- the new tty driver line in axmakefile refers to "rs2.o", I had to > change it to "rs2.s" to get it to work. I guess Bruce has a "real" > (separate) assembler and loader. For compatibility with UN*X makefiles, it sure would be nice to have "cc" (optionally) produce a packed assembler file with '.o' suffix and "libpack" produce a '.o' from a '.s' and asld accept '.o' as a suffix representing packed assembler. (alias "libpack" to "as" and "asld" to "ld" and you have *real* [PDP-11] compatibility!!!) It shouldn't be necessary to break any existing Minix style makefiles, just have asld accept both '.s' and '.o' as legal input file suffixes. Andy? How about it? Can you give us a binary patch to asld that will do that? I can handle the fixes to "cc" and "libpack", but I don't seem to have the sources for 'asld' lying around. Anybody else have the sources to asld? Enjoy! Rick PS -- does POSIX say anything about this? -- Rick Thomas uucp: {ames, att, harvard}!rutgers!jove.rutgers.edu!rbthomas internet: rbthomas@JOVE.RUTGERS.EDU bitnet: rbthomas@zodiac.bitnet Phone: (201) 932-4301