Skip site navigation (1)Skip section navigation (2)
Date:      Mon, 27 Jun 2005 16:47:32 -0400
From:      John Baldwin <jhb@FreeBSD.org>
To:        freebsd-current@freebsd.org
Cc:        cracauer@cons.org, Eric Anholt <eta@lclark.edu>, "M. Warner Losh" <imp@bsdimp.com>
Subject:   Re: 6.0-current panic: loading radeon module
Message-ID:  <200506271647.34208.jhb@FreeBSD.org>
In-Reply-To: <1119813460.3817.17.camel@leguin>
References:  <1119650444.1173.105.camel@leguin> <20050626.125729.127901479.imp@bsdimp.com> <1119813460.3817.17.camel@leguin>

next in thread | previous in thread | raw e-mail | index | archive | help
On Sunday 26 June 2005 03:17 pm, Eric Anholt wrote:
> On Sun, 2005-06-26 at 12:57 -0600, M. Warner Losh wrote:
> > In message: <1119809673.3817.3.camel@leguin>
> >
> >             Eric Anholt <eta@lclark.edu> writes:
> > : On Fri, 2005-06-24 at 19:59 -0400, Martin Cracauer wrote:
> > : > > I'd forgotten about this thread.  I've got a proposed fix for this
> > : > > at: http://www.freebsd.org/cgi/query-pr.cgi?pr=kern/80718
> > : >
> > : > I get the same panic and backtrace with that patch (on a 32 MB
> > : > moobile Radeon 7500).
> > : >
> > : > I will annotate the sourcecode now to verify (or not) the NULLs in
> > : > the arguments for the bus argument.
> > :
> > : Thanks for showing that that patch didn't work.  I should have
> > : remembered why it wouldn't.  The next step would be to try a
> > : workaround: Comment out the two calls to drm_initmap() in radeon_cp.c.
> > :
> > : We really need a nice call to grab the start/offset of a pci resource
> > : given the rid.  We can avoid it now, but we might not be able to given
> > : where the DRM is headed.
> >
> > rman_get_start() doesn't do the trick?  Or do you need to know what it
> > is before the bus_alloc_resource?  If so, you can call
> > bus_get_resource().  What is it that you need to know?
>
> bus_alloc_resource tries to map the 128+ MB framebuffer and pmap_mapdev
> falls over.  I'll take a look at bus_get_resource now.

Try doing it without RF_ACTIVE.  The problem there is that we really just want 
to allocate the resource so the backing physical addresses are reserved, but 
we don't want to actually map that into KVA.

-- 
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?200506271647.34208.jhb>