Philip Blundell (philb@gnu.org)
Sun, 13 Sep 1998 16:54:01 +0100
Hi,
I've collected together some of the patches that people have come up with
recently. You can find the aggregate of these with some changes of my own at
http://www.tazenda.demon.co.uk/phil/linux-2.1.121-parport-diff-980913.gz
This includes:
* Amiga driver from Joerg Dorchain
* Atari driver from Andreas Schwab
* IRQ-handling changes from Andrea Arcangeli
* lp uses a spinlock not cli/sti
* lp timings are always done with udelay() not a simple counter; the delay
defaults to 1us for m68k and SPARC and 0 for everything else.
I had to fiddle with parport_amiga a bit to get it to fit in with Andrea's irq
code; please check I haven't damaged it too much. I think the IRQ handling is
basically the way I want it to be now.
The way parport_procfs handles writes to the `irq' file has changed a bit as
well. Each driver now has an extra method that parport_procfs can call to ask
it to change the IRQ. No driver currently implements this but the possibility
is there. This means that drivers can check the value to make sure it's
sensible, and reconfigure the hardware appropriately on systems where it's
supported (ie some Super-IO chipsets). Hopefully this will make the facility
useful for more than correcting mistakes made by the autoprobe code. :-)
I'm not intending to try to get these changes merged before 2.2. Instead they
will probably go into 2.3 and I will make an optional patchset for 2.2 for the
benefit of people who want them (probably m68k). I've also written to David
to find out the status of the magic he was working with chipset-specific code
so that ECP and EPP can finally become useful.
By the way, m68k people, feel free to create your own PARPORT_MODE_... values
if the behaviour of your hardware doesn't precisely match any of the options
that we have for PCs. It was always the intention that these could vary
depending on the driver which is why they're named the way they are.
p.
-- To unsubscribe, send mail to: linux-parport-request@torque.net --
-- with the single word "unsubscribe" in the body of the message. --
This archive was generated by hypermail 2.0b3 on Wed 30 Dec 1998 - 10:18:17 EST