Re: [PATCH] ppbus/ppc locking

From: John Baldwin <jhb_at_freebsd.org>
Date: Wed, 21 Jan 2009 15:36:08 -0500
On Thursday 04 December 2008 9:42:53 am Bruce M. Simpson wrote:
> John Baldwin wrote:
> > Please test!  This is the last non-MPSAFE network driver at this point.  
This 
> > patch adds locking for the ppbus(4)/ppc(4) devices and the various ppbus 
> > child devices (lpt, vpo, lpbb, ppi, pps).  The basic model is that a 
single 
> > mutex in the ppc(4) driver protects the ppc0 hardware and is shared with 
the 
> > various child drivers.  Two drivers now have detach methods that did not 
have 
> > them before (plip and ppi).  I've done some simple testing on my laptop 
(able 
> > to load the drivers and do some simple things w/o panic'ing or tripping 
> > assertions), but I am not really able to test the peripheral drivers 
fully.
> >
> > http://www.FreeBSD.org/~jhb/patches/ppc_locking.patch
> >
> >   
> 
> Runway lpt Giant is an occasionally show stopping issue for me because 
> my printer is attached via the plt port.  I may get time to look at this 
> later on...
> 
> I tried applying these patches against 7-STABLE.
> 
> ppc_cleanup.patch applied OK to 7.
> 
> ppc_intr.patch applied OK to 7 with interrupt.h change manually merged,
> and some fixups to ppc.c for earlier intr_event kpi.
> 
> ppc_locking.patch does not apply cleanly, and it's too much for me to 
> deal with right now.
> 
> I found I had to hack up an existing 7 tree in /usr/src to get things to 
> compile because of the wide scope of the changes (touching kern, sys 
> etc), I couldn't just use an svn checkout to work from.

Try http://www.FreeBSD.org/~jhb/patches/ppc_mpsafe_7.patch  It is a complete 
backport generated against a fresh stable/7 tree.  It does not need the 
interrupt changes since my locking patches actually undo them.

-- 
John Baldwin
Received on Wed Jan 21 2009 - 22:11:12 UTC

This archive was generated by hypermail 2.4.0 : Wed May 19 2021 - 11:39:40 UTC