Skip site navigation (1)Skip section navigation (2)
Date:      Sun, 30 Oct 2005 11:45:14 -0500
From:      Sergey Babkin <babkin@verizon.net>
To:        Kris Kennaway <kris@obsecurity.org>
Cc:        gbergling@0xfce3.net, babkin@users.sourceforge.net, phk@phk.freebsd.dk, freebsd-arch@freebsd.org
Subject:   Re: wscons for FreeBSD?
Message-ID:  <4364F89A.29B607B4@verizon.net>
References:  <12768156.1130520253107.JavaMail.root@vms073.mailsrvcs.net> <200510281435.41700.jhb@freebsd.org> <4362B611.B4A12AF8@verizon.net> <20051028.215049.102576958.imp@bsdimp.com> <20051030104814.GA32477@xor.obsecurity.org>

next in thread | previous in thread | raw e-mail | index | archive | help
Kris Kennaway wrote:
> 
> On Fri, Oct 28, 2005 at 09:50:49PM -0600, M. Warner Losh wrote:
> > In message: <4362B611.B4A12AF8@verizon.net>
> >             Sergey Babkin <babkin@verizon.net> writes:
> > : John Baldwin wrote:
> > : >
> > : > On Friday 28 October 2005 01:24 pm, Sergey Babkin wrote:
> > : > > >From: John Baldwin <jhb@freebsd.org>
> > : > > >
> > :
> > : > > * When entering panic/debugger mode the console
> > : > >   should reset its video mode to the one where
> > : > >   the panic information is visible.
> > : >
> > : > I think this might be kind of hard since you really don't know what X has done
> > : > to the hardware unless you make X talk to the console driver to do
> > : > everything.
> > :
> > : Can you re-initialize the device from scratch? My knowledge
> > : about the video hardware dates to SVGA times and even that
> > : is not too deep.
> > :
> > : Maybe provide a basic SVGA re-initialization routine on x86 and have
> > : an interface for the third-party in-kernel drivers that
> > : would contain the re-initialization routine.
> >
> > When this has come up in the past, those in the know say that it takes
> > card specific registers and knowledge to reset these cards' video
> > modes.
> 
> Even if it doesn't work on all cards, if there's some method that will
> at least work for a subset of configurations it may be worth
> investigating: by that point the machine has panicked anyway, so
> what's the worst that can happen?

I think that if there are no endless busy waits, nothing should make
the situation any worse.

I guess for VESA-compatible cards the VESA BIOS couls be used too,
but that's much more complicated to do and much more prone to
things going wrong during resets.

But I guess in any case providing an interface for the in-kernel reset
drivers would solve the problem as long as there are those drivers.
Their implementation can probably be extracted from XF86, though 
of course it would need extra work.

-SB



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?4364F89A.29B607B4>