kern/121720: [wpi] wpi doesnt work if kernel has optionsSCHED_ULE

[Yousif Hassan]

The following reply was made to PR kern/121720; it has been noted by GNATS.

From: "Yousif Hassan" <yousif at alumni.jmu.edu>
To: "Andrew Lankford" <lankfordandrew at charter.net>,
	<bug-followup at FreeBSD.org>,
	<Helko.Glathe at freenet.de>
Cc:  
Subject: Re: kern/121720: [wpi] wpi doesnt work if kernel has optionsSCHED_ULE
Date: Tue, 18 Mar 2008 11:34:09 -0400

 ----- Original Message ----- 
 From: "Andrew Lankford" <lankfordandrew at charter.net>
 To: <freebsd-net at FreeBSD.org>
 Sent: Monday, March 17, 2008 3:20 AM
 Subject: Re: kern/121720: [wpi] wpi doesnt work if kernel has 
 optionsSCHED_ULE
 
 
 > The following reply was made to PR kern/121720; it has been noted by 
 > GNATS.
 >
 > From: Andrew Lankford <lankfordandrew at charter.net>
 > To: bug-followup at FreeBSD.org, Helko.Glathe at freenet.de
 > Cc:
 > Subject: Re: kern/121720: [wpi] wpi doesnt work if kernel has options 
 > SCHED_ULE
 > Date: Mon, 17 Mar 2008 02:49:43 -0400
 >
 > Same problem here with my 7-STABLE machine.  Settling for SCHED_4BSD did
 > the trick.  If this is a problem with the driver, then perhaps a MFC of
 > some recent (as in several days old) "stability fixes" to the wpi source
 > will fix the problem:
 > http://www.freebsd.org/cgi/cvsweb.cgi/src/sys/dev/wpi/if_wpi.c
 >
 > wpi in STABLE has panicked my laptop at least once.
 >
 > If you're impatient, I imagine that the latest cvs patches may work with
 > 7-STABLE.
 >
 > Andrew Lankford
 
 I'm adding this as a point of reference in case it helps a developer:
 
 My HP dv9700t runs SCHED_ULE and wpi works great.   This leads me to suspect 
 that this is probably not just as simple as if_wpi and the scheduler; I 
 believe this has to do with a combination of the wpi driver and the BIOS 
 (and/or ACPI).  There have been other problem reports related to different 
 loading / attach / buggy behavior with wpi - different computers seem to 
 produce different results.  This sounds related (albeit SCHED_4BSD appears 
 to solve this particular issue).
 
 FWIW.
 
 --Y