Again, tried to submit i2c-pport.c but the usage of
udelay(#) has to change to schedule_timeout(#).
I think this is going to affect lots of drivers and needs
converting before we will get sensors included into the
kernel.
Albert
-------- Original Message --------
Subject: Re: [patch 3/3] 2.5.36 i2c new adapter i2c-pport driver
Date: Wed, 18 Sep 2002 12:33:46 -0400
From: Jeff Garzik <jgarzik at mandrakesoft.com>
Organization: MandrakeSoft
To: Albert Cranford <ac9410 at attbi.com>
CC: Linus Torvalds <torvalds at transmeta.com>,Kernel mailing list <linux-kernel at vger.kernel.org>
References: <Pine.LNX.4.44.0209180205270.358-200000 at home1>
Albert Cranford wrote:
> +static int bit_pport_init(void)
> +{
> + if (!request_region((base+2),1, "i2c (PPORT adapter)")) {
> + return -ENODEV;
> + } else {
> +
> + /* test for PPORT adap. */
> +
> +
> + PortData=inb(base+2);
> + PortData= (PortData SET_SDA) SET_SCL;
> + outb(PortData,base+2);
> +
> + if (!(inb(base+2) | 0x06)) { /* SDA and SCL will be high */
> + DEBINIT(printk("i2c-pport.o: SDA and SCL was low.\n"));
> + return -ENODEV;
> + } else {
> +
> + /*SCL high and SDA low*/
> + PortData = PortData SET_SCL CLR_SDA;
> + outb(PortData,base+2);
> + udelay(400);
use schedule_timeout() instead of udelay() since you're in process
context. otherwise, looks ok...
