[comp.unix.i386] ISC asy/floppy drivers still broken?

john@jwt.UUCP (John Temples) (08/26/90)

In article <3687@rsiatl.UUCP> jgd@rsiatl.UUCP (John G. DeArmond) writes:
>Hmm, seems to me that if ISC would just [...]
>concentrate their efforts (and our money) on
>things like fixing the ASY drivers and [...]
>make the floppy driver catch more than one 
>sector per revolution and so on and so forth, we'd all be lightyears ahead.

Does ISC *still* not have a working asy driver or usable floppy
driver?  I was using ISC 1.0.6, and found the asy drivers completely
broken (especially for dial in/dial out on the same line), and the
floppy drivers absurdly slow.  I have seen people continue to
complain on the net about how they can't get bidirectional modem
access working, even on later versions of ISC.

I recently bough ESIX, and the asy drivers work perfectly.  They are
100% solid, work with bidirectional modem access "right out of the
box" without using the uugetty kludge, and support the 16550.  The
floppy drivers are not quite as fast as DOS, but are dramatically
improved over the ISC drivers I have experience with.

I paid $618 for an unlimited user ESIX development system.  The most
recent price I saw for a 1-2 user ISC 2.2 system was $1,895.
Granted, the ISC package has NFS, VP/ix, and a couple other things,
rather than the unlimited user license.  But how can they justify
such prices if even the most fundamental drivers are broken?
And not to mention the difference in support costs: $600/year for
ISC versus $0/year for ESIX.

I'd much rather have fewer features that work well than lots of
features that don't work at all.
-- 
John W. Temples -- john@jwt.UUCP (uunet!jwt!john)

johnl@esegue.segue.boston.ma.us (John R. Levine) (08/26/90)

In article <2050@jwt.UUCP> you write:
>Does ISC *still* not have a working asy driver or usable floppy driver?

ISC finally cleaned up the asy driver in the X5 update to 386/ix 2.0.x.
The throughput is OK, particularly if you have a 16550 FIFO serial chip.
It has separate /dev/ttyN and /dev/acuN special files so that dialing in
and out on the same line is locked in the kernel without having to fool
around with uugetty.  I have an internal Telebit modem that I upgraded to
a 16550 and get respectable uucp throughput above 1100 cps.

The 2.2 floppy driver isn't superb but it is at least reasonable.  Disk
throughput under vp/ix is great, it apparently reads and writes full tracks
on the raw disk.  The driver also makes a respectable try at determining the
density of a disk when you open it.

-- 
Regards,
John Levine, johnl@esegue.segue.boston.ma.us, {spdcc|ima|world}!esegue!johnl

larry@nstar.uucp (Larry Snyder) (08/27/90)

john@jwt.UUCP (John Temples) writes:

>I recently bough ESIX, and the asy drivers work perfectly.  They are
>100% solid, work with bidirectional modem access "right out of the
>box" without using the uugetty kludge, and support the 16550.  The

Do the ESIX ASY drivers fully support hardware only flow control?

-- 
      Larry Snyder, Northern Star Communications, Notre Dame, IN USA 
            uucp: iuvax!ndmath!nstar!larry  -or-  larry@nstar
     Public Access Unix Site (219) 289-0282 (5 lines/PEP/HST/Hayes-V)