Skip site navigation (1)Skip section navigation (2)
Date:               Tue, 12 Sep 1995 16:10:29 +0000
From:      "Peter May" <peter@osix.com.au>
To:        J Wunsch <j@uriah.heep.sax.de>, hackers@freebsd.org
Subject:         Re: higher density diskettes
Message-ID:  <199509121557.PAA17050@thumper.osix.com.au>

next in thread | raw e-mail | index | archive | help
> As Peter May wrote:
> 
> > That should speed the performance of this already flaky driver up no 
> > end! :-)
> 
> Badly designed: yes.  But flakey?  How? Why?

The "Task" I have assigned myself over the last 18 months or so has 
been the MCA support. The major stumbling block here was getting the 
driver to work with an architecture that supported all of the 
features that the '765 was supposed to - like the diskette changed 
line and so forth.

Then add to that the layer of issues brought about by level sensitive 
interrupt handlers. IRQ's that are not reset by the driver (in the 
current architecture, which uses fdstate like fdintr :-( ) interrupt 
requests are not services at appropriate times, and fdstate fails to 
transition in many places or supply proper commands in all 
circumstances (like diskette changed).

Now, I had this working reasonably on 1.1.5.1R, then along came 2.0R. 
A week later, I had "repaired" all of the issues (none of which I 
corrected in any form of elegance, BTW). The 2.0.5 - more major 
changes to the FD driver. 

Now I use SUP .... and follow -current, I am sure I am going to win 
the race to get the driver that I "need" working!

Mind you, a lot of these issues only become apparent when you use an 
MCA machine ... but it leaves me wondering if some of the 
inconsistencies in the ISA stuff become obvious on MCA.

> We've got one of the fastest floppy disk drivers. :-) It's even about
> 10 % faster than Linux (for sequential reads), i've heard that old
> Xenix' are as fast, and it's way faster than all commercial Unices
> i've seen.  (For sequential reads, and i consider them the most
> typical _unix_ usage [tar floppy], we get 30 KB/s, that's 2/3 of the
> raw data rate, and basically proves that we don't use revolutions.)

On ISA, performance is great.
 
> Of course, all this assumes that you're issuing larger requests than
> single-sector, otherwise the o/s overhead may come into your way and
> make you lose revolutions.
> -- 
> cheers, J"org
> 
> joerg_wunsch@uriah.heep.sax.de -- http://www.sax.de/~joerg/
> Never trust an operating system you don't have sources for. ;-)

Best Regards,
---------------------------------------------------------------->>>>>

Peter May                           OSIX Pty Ltd
Director                            Level 1, 261-263 Pacific Highway
Technical Services                  North Sydney. NSW. Australia. 2060.

Home: +61-18-782858                 Internet: peter@osix.com.au
Work: +61-2-9922-3999               Fax:  +61-2-9922-3385

          >>>> PGP Public key available upon request <<<<

---------------------------------------------------------------->>>>>



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199509121557.PAA17050>