Skip site navigation (1)Skip section navigation (2)
Date:      Fri, 30 Jan 2004 11:10:02 -0700 (MST)
From:      "M. Warner Losh" <imp@bsdimp.com>
To:        jhb@FreeBSD.org
Cc:        current@FreeBSD.org
Subject:   Re: API to turn off the display
Message-ID:  <20040130.111002.26275922.imp@bsdimp.com>
In-Reply-To: <200401301306.57986.jhb@FreeBSD.org>
References:  <20040130092529.GA725@laptop.6bone.nl> <20040130.083614.75932761.imp@bsdimp.com> <200401301306.57986.jhb@FreeBSD.org>

next in thread | previous in thread | raw e-mail | index | archive | help
In message: <200401301306.57986.jhb@FreeBSD.org>
            John Baldwin <jhb@FreeBSD.org> writes:
: On Friday 30 January 2004 10:36 am, M. Warner Losh wrote:
: > In message: <20040130092529.GA725@laptop.6bone.nl>
: >
: >             Mark Santcroos <marks@ripe.net> writes:
: > : On Thu, Jan 29, 2004 at 11:32:54AM -0500, John Baldwin wrote:
: > : > I read the spec yesterday, and what is supposed to happen is this:  The
: > : > display (LCD, CRT, etc.) is supposed to be powered down using DPMS. 
: > : > The actual adapter is then supposed to be powered down using either PCI
: > : > or ACPI sleep states, and the adapter should not be powered down to a
: > : > lower sleep state (like D3) w/o using DPMS to power the display down to
: > : > at least that state first.  This means that the kernel might need to
: > : > grow a dummy vga driver of some sort and some simple dpms support for
: > : > suspend/resume.
: > :
: > : I second this.
: > :
: > : Nate, can you put it on your acpi-todo list?
: >
: > The only problem with having a vga device for suspend/resume is that
: > it will claim devices that drm wants to claim later and X won't work.
: >
: > I have about the simplest driver possible for vga in one of my trees,
: > and found this to be a problem on my dell i8k.
: 
: Yes, we need a way I think to have multiple drivers attach to a device, or 
: maybe have the drm drivers attach as a child of the vga device instead.  For 
: the intel motherboards with onboard graphics, we already have this problem 
: where drm and agp both want the same device.  Having a model of:
: 
: vga0
:   drm0
:   agp0
: 
: Might be the sanest way to do this cleanly.

Yes.  I like that model.  While it might be nice to have a reauction
of devices in the newbus level, I think that for this case there's
limited usefulness in doing that.  Having vga driver that others
attach to will also allow drm drivers to possibly attach to a vga
driver that's attached to something other than a pci bus.

Warner



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20040130.111002.26275922.imp>