Originally posted by: modl@whuts.UUCP (SMITH)
Article-I.D.: whuts.114
Posted: Wed Mar 13 16:57:10 1985
Date-Received: Thu, 14-Mar-85 06:17:29 EST
References: <5173@utzoo.UUCP> <161@sesame.UUCP>
Organization: AT&T Bell Laboratories
Lines: 82
> I have VENIX running on my AT, and in fact wrote the Hard Disk driver
> which is used by VENIX. Some responses to your questions follow:
>
>> I have recently acquired Venix for the PC/AT and have run into a few problems,
>> and was wondering if others out there (there MUST be others :-) ) had run
>> into them as well....
>> 1) date/time on bootup is sort-of correct..month, day and time are correct,
>> year is always 1970
> Not a problem on my machine. Check that you have used IBM's 'setup' properly.
>
>> 2) awk(1) does not handle the BEGIN pattern correctly..it hangs the process
> Sorry- I avoid using awk on any machine. It's too slow even when it works.
>
>> 3) curses(3) routine getstr does not return when RETURN pressed..have to
>> hit ^J (that took a while to stumble into)
> You probably want to use CBREAK rather than RAW mode.
>
>> 4) awk(1) does not handle this type of command line properly:
>> awk -f prog filename
>> it just hangs there
>>
>> Some observations:
>> 1) though I could not find it anywhere in the manuals, I have discovered
>> that kill -2 1 causes /etc/ttys to be re-read..sure beats re-booting as
>> the manual's way to re-configure your ports
> This is the historic method used by V7 derivatives.
>
>> 2) arbitrary-length reads directed at a tty in raw mode
>> block until the number of characters
>> requested are input. THIS IS MENTIONED in ONE (1) PLACE as far as I can
>> see, and not well marked.
>> (in thinking about this, it might be why getstr screws up)
>> This means, among other things, that stdio can not be used in raw
>> mode (I have some public software which does this), as the old
>> _filbuf does a read of BUFSIZ. cute
> This is very true, although it has hidden benefits for things like
> uucp that want to read 64 bytes RAW, and need all of them before anything
> can be done.
>
>> 3) performance is pretty good
>>
>> Has anyone put the floating point processor in? how does venix live with it?
>> How about the AST Advantage card??? Anyone with it?
>> And of course, the 64 dollar question, what about second source hard disks?
> I haven't been able to get the FPP, but it does work on machines that have it.
> I have an Advantage card in my machine, and am about to load it up with those
> $8.00 256K RAM chips that are now available. As far as second source hard
> disks and VENIX go, here's the poop: If the disk matches one of IBM's disk
> types as listed in the TR, fine. If it doesn't, you'll need a later version
> of my disk driver, that can handle 'odd ball' disks. In particular, the
> Maxtor-140 is one of these.
>
>>
>> I would appreciate hearing from those hearty souls who are running Venix
>> on a PC/AT....
>>
>> as they say, thanks in advance
>> --
>> Jerry J. Deroo
>> {allegra,ihnp4,linus,decvax}!utzoo!jerry
>
> One further bit of advice: Pop for the $1.59 and get a 15Mhz crystal for
> the AT. The 15Mhz seems to work well, and even though it runs the 80286
> over spec, it won't be running the 80287 over spec. The 286 will be running
> at 7.5Mhz and the 287 will be running at 5Mhz. I think this is
> a reasonable compromise. With a 15Mhz crystal, the 80286 out benchmarks
> a 10Mhz 68010 when running the drystone benchmark (balanced statement
> mix, variable types, but no floating point or disk).
>
> --
> Rick Richardson, PC Research, Inc.
>
> {genrad|ihnp4|ima}!wjh12!talcott!sesame!{rick|pcrat!rick}
> {cbosgd|harvard}!talcott!sesame!{rick|pcrat!rick}
> rick%sesame@harvard.ARPA
I'm currently running an AT with a 80287 and 512K, but no hard disk. Could you
give me a source for a 15Mhz crystal? Sure would appreciate it.
Max Robin AT&T BellLabs
ZZ
*** REPLACE THIS LINE WITH YOUR MESSAGE ***