Skip site navigation (1)Skip section navigation (2)
Date:      Mon, 15 Mar 2004 15:39:30 -0500
From:      John Baldwin <jhb@FreeBSD.org>
To:        Ken Smith <kensmith@cse.Buffalo.EDU>
Cc:        Wilko Bulte <wkb@freebie.xs4all.nl>
Subject:   Re: cvs commit: src/sys/kern sys_process.c syscalls.master
Message-ID:  <200403151539.30298.jhb@FreeBSD.org>
In-Reply-To: <20040315192139.GB28070@electra.cse.Buffalo.EDU>
References:  <200403151848.i2FImSFe047106@repoman.freebsd.org> <20040315190853.GA33607@freebie.xs4all.nl> <20040315192139.GB28070@electra.cse.Buffalo.EDU>

next in thread | previous in thread | raw e-mail | index | archive | help
On Monday 15 March 2004 02:21 pm, Ken Smith wrote:
> On Mon, Mar 15, 2004 at 08:08:53PM +0100, Wilko Bulte wrote:
> > On Mon, Mar 15, 2004 at 10:48:28AM -0800, John Baldwin wrote:
> > > jhb         2004/03/15 10:48:28 PST
> > >
> > >   FreeBSD src repository
> > >
> > >   Modified files:
> > >     sys/kern             sys_process.c syscalls.master
> > >   Log:
> > >   Drop the proc lock around calls to the MD functions
> > > ptrace_single_step(), ptrace_set_pc(), and cpu_ptrace() so that those
> > > functions are free to acquire Giant, sleep, etc.  We already do a
> > > PHOLD/PRELE around them so that it is safe to sleep inside of these
> > > routines if necessary.  This allows ptrace() to be marked MP safe again
> > > as it no longer triggers lock order reversals on Alpha.
> >
> > Someone with a SMP alpha box might also want to give this a spin I
> > suppose.
>
> beast (alpha ref machine in the cluster) has dual processors, if something
> special needs to be done to test this we could arrange something.

You would need to dink with gdb (setting breakpoints, detach/attach, etc.) on 
a kernel with WITNESS turned on and see if it breaks or whines.

-- 
John Baldwin <jhb@FreeBSD.org>  <><  http://www.FreeBSD.org/~jhb/
"Power Users Use the Power to Serve"  =  http://www.FreeBSD.org



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?200403151539.30298.jhb>