Re: ECP port driver

["Robert Creager" <Robert@CarolineCreager.com>]

Chris and all,

Here's the code and makefile.  I'm running late, so a full description 
will follow in a couple of hours.  Suffice to say, a /dev file will exists 
at some point, created through a script, and open/close/read will 
work on that dev file.  I'm thinking of creating a tassiv directory 
under the /dev directory, then maybe we can through some stamp 
driver nodes in there (device drivers are fun!).  insmod won't work 
just yet, as the read function isn't defined in the file.  I was just 
working on that.

Later,
Rob

> 
> Rob,
> 
> Sounds like you should keep working on it.  I am
> converting my Buffer Card driver from 2.2 to 2.4
> that's enough for now.  And yes "insb" with big
> blocks is a win.  I do it in the Buffer Driver
> the bottleneck there is now the ISA bus
> 
> I would like to see your driver.  I'd like to get
> a start on software that would use it.
> 
> Robert Creager wrote:
> > 
> > Chris,
> > 
> > Well, I was thinking more along the lines of "how far
> > along are you?".  If you haven't started yet, I'll
> > continue.  If you're nearly done, I won't continue.
> > That sort of thing.  As soon as I have something
> > working, I'll be happy to show everyone.
> > 
> > I will always support the code I write.
> > 
> > I am using the parport driver.  I am to the point of
> > implementing the read method on the device.  The 2.4.5
> > parport driver doesn't implement the read using the
> > FIFO/FIFO information.  I'm grabbing the 2.4.9 kernel
> > to see if it might be there, otherwise, I'll spin my
> > own from the 'broken' paraport block read.
> > 
> > While playing around with another, simpler parallel
> > port driver, I found that using the insb macro with 16
> > bytes transfered at a time gets my computer to around
> > 30 second transfer time for 16Mbytes.  Reading 1 byte
> > at a time using inb read method takes around 90
> > seconds for 16Mbytes.
> > 
> > The 2.4.5 parport driver only implements DMA on the
> > write side, not the read side, so that may be more of
> > an adventure to try to implement.
> > 
> > I am planning on implementing a specific device which
> > uses the FIFO test mode, which will stream data
> > without any hardware hooked up, which will provide a
> > device which behaves the same as the real thing.
> > 
> > You want to see my current code now, let me know and
> > I'll send it to you.
> > 
> > Later,
> > Rob
> > 
> >
> 
> -- 
> 
> --
>    Chris Albertson             
>    chrisalbertson90278@yahoo.com
>    Redondo Beach, California
>    home: 310-376-1029
>    cell: 310-990-7550
> 


Later,
Rob
The following section of this message contains a file attachment
prepared for transmission using the Internet MIME message format.
If you are using Pegasus Mail, or any another MIME-compliant system,
you should be able to save it or view it from within your mailer.
If you cannot, please ask your system administrator for assistance.

   ---- File information -----------
     File:  tassiv_parport.c
     Date:  4 Sep 2001, 13:34
     Size:  4922 bytes.
     Type:  Program-source

tassiv_parport.c

KERNELDIR = /usr/local/src/linux-2.4.5
INCLUDEDIR = $(KERNELDIR)/include
VERSIONFILE = $(INCLUDEDIR)/linux/version.h
VERSION = $(shell awk -F\" '/REL/ {print $$2}' $(VERSIONFILE))
INSTALLDIR = /lib/modules/$(VERSION)/misc

include $(KERNELDIR)/.config

CFLAGS = -D__KERNEL__ -DMODULE -I$(INCLUDEDIR) -O -Wall

ifdef CONFIG_SMP
	CFLAGS += -D__SMP__ -DSMP
endif

OBJS = tassiv_parport.o

all: $(OBJS)

install:
	install -d $(INSTALLDIR)
	install -c $(OBJS) $(INSTALLDIR)

clean:
	rm -f $(OBJS) *~ core