From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1751777Ab2IOW6d (ORCPT ); Sat, 15 Sep 2012 18:58:33 -0400 Received: from mail.active-venture.com ([67.228.131.205]:58803 "EHLO mail.active-venture.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1750907Ab2IOW6c (ORCPT ); Sat, 15 Sep 2012 18:58:32 -0400 X-Originating-IP: 108.223.40.66 Date: Sat, 15 Sep 2012 15:58:41 -0700 From: Guenter Roeck To: Parag Warudkar Cc: lm-sensors@lm-sensors.org, linux-kernel@vger.kernel.org, rydberg@euromail.se, khali@linux-fr.org Subject: Re: [PATCH] applesmc: Bump max wait and rearrange udelay Message-ID: <20120915225841.GA3816@roeck-us.net> References: MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: User-Agent: Mutt/1.5.21 (2010-09-15) Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On Sat, Sep 15, 2012 at 06:42:30PM -0400, Parag Warudkar wrote: > I have been getting a steady stream of wait_read timeouts on my 2010 MBP. > > After playing around with various values of APPLESMC_MAX_WAIT a value of > 0x10000 reduces the wait_read failures to zero under most normal workloads > - with and without AC power plugged in, at idle and and at make -j4 loads. > > While there I noticed we don't really need to udelay before first inb() - > so I moved it down to after first and subsequent failures. > > Been running this for couple days without any issues. > > Signed-off-by: Parag Warudkar > > diff --git a/drivers/hwmon/applesmc.c b/drivers/hwmon/applesmc.c > index 2827088..46cb458 100644 > --- a/drivers/hwmon/applesmc.c > +++ b/drivers/hwmon/applesmc.c > @@ -56,7 +56,7 @@ > /* wait up to 32 ms for a status change. */ > #define APPLESMC_MIN_WAIT 0x0010 > #define APPLESMC_RETRY_WAIT 0x0100 > -#define APPLESMC_MAX_WAIT 0x8000 > +#define APPLESMC_MAX_WAIT 0x10000 > > #define APPLESMC_READ_CMD 0x10 > #define APPLESMC_WRITE_CMD 0x11 > @@ -170,11 +170,11 @@ static int wait_read(void) > u8 status; > int us; > for (us = APPLESMC_MIN_WAIT; us < APPLESMC_MAX_WAIT; us <<= 1) { I wonder if it would make sense to keep APPLESMC_MAX_WAIT as it is now and replace the loop termination conditions with us <= APPLESMC_MAX_WAIT That would accomplish the same, and APPLESMC_MAX_WAIT would reflect the real maximum wait time. Also, since the delay time can get quite large, would it make sense to replace udelay with usleep_range() ? Guenter > - udelay(us); > status = inb(APPLESMC_CMD_PORT); > /* read: wait for smc to settle */ > if (status & 0x01) > return 0; > + udelay(us); > } > > pr_warn("wait_read() fail: 0x%02x\n", status); > @@ -192,11 +192,12 @@ static int send_byte(u8 cmd, u16 port) > > outb(cmd, port); > for (us = APPLESMC_MIN_WAIT; us < APPLESMC_MAX_WAIT; us <<= 1) { > - udelay(us); > status = inb(APPLESMC_CMD_PORT); > /* write: wait for smc to settle */ > - if (status & 0x02) > + if (status & 0x02) { > + udelay(us); > continue; > + } > /* ready: cmd accepted, return */ > if (status & 0x04) > return 0; >