Skip site navigation (1)Skip section navigation (2)
Date:      Tue, 19 Aug 1997 10:57:47 -0400
From:      Brian McGovern <bmcgover@cisco.com>
To:        hackers@freebsd.org
Subject:   Committers: New Cyclades Z driver ready
Message-ID:  <199708191457.KAA00543@bmcgover-pc.cisco.com>

next in thread | raw e-mail | index | archive | help
I've reached the point where my Cyclades Cyclom-Z driver is stable, and is
actually doing quite well (we use it for modem testing here at Cisco in
several groups). For those who don't know, the Zs come in one of three 
forms:

8Zo (V1)	- 8 ports per card to 460K per port

8Zo (V2)	- 8 ports per card to 920K per port

8Ze 		- up to 64 ports per card to 920K per port


So far, I've run 32 ports (4x8Zo V1) in a Pentium Pro 200 using about 30% of
the CPU at 115200 baud (this includes running PPP with FTP sessions saturating
all of the links).

As you can see, the boards are quite nice.

Anyhow, I promised Cyclades that once the driver was stable, I'd offer it up
the FreeBSD core for inclusion in to the standard operating system. I've 
uploaded the driver to ftp.freebsd.org in /pub/FreeBSD/incoming. The archive
name is cz-0.N.tgz. It includes everything needed to drop it in to the
kernel source, has a RELNOTES file, a makedev program (my shell scripting 
skills haven't been used for a couple of years, so I wrote it in C. Someone
may want to translate it back), the firmware, headers, etc. The only thing
that wasn't included prefab was the change to files.i386 in the conf
directory (simply because it was straight forward enough).

If someone could take a look at it, and either commit it, or reject it with
comments on "whats wrong" (don't be too harsh, this is my first "real" Unix
driver, and you don't want to scare me off from doing more :) ), and how 
you'd like it fixed, I'd appreciate it.

I also expect to continue to support the driver, and make some changes (for
instance, I'll be trying to add DMA capabilities, so the board can move the
data directly in to kernel buffers, so the kernel can avoid PCI bus
arbitration on an interrupt/poll), so comments, suggestions, and diffs are
very much welcome.

	-Brian



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199708191457.KAA00543>