[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: External monitor with Compaq Presario 1600



Hrm, I do have [M] nvram support, but I'm not sure how to do what you
suggested. Thanks for the tip, I'll see what I can learn about this so I
might try it.

Regards,

Arjen

On Mon, 7 Jul 2003, Heather Stern wrote:

> On Mon, Jul 07, 2003 at 11:20:02PM +0200, Arjen Verweij wrote:
> > Welp, I recompiled a kernel without fb enabled, but when push comes to
> > shove, I can NOT swap physical displays after choosing an option (i.e.
> > one of the kernel image flavors) from the lilo menu. I am at a loss here.
> >
> > I'm inclined to think that if it is possible in the lilo menu, this should
> > be possible without any special software, but then not being able to do if
> > while the machine is booting/after it has booted leaves me flabbergasted.
> >
> > The reason why this functionality seems suddenly lost is completely
> > unclear to me.
> >
> > Does anyone have suggestions on who to bother with this intriguing
> > problem?
> >
> > Thanks.
> > Arjen
>
> Well, perhaps it is controlled by some code in the BIOS?  Linux gets
> into protected mode ASAP and tries to not look back at the BIOS ever
> again - which means ordinary services like certain keys, running the
> floopy, etc all have to be dealt with by the kernel.  That's the only
> theory I have that covers the feature dying during the kernel load.
>
> For a reasonable analogy, laptop users whose floppy is really a USB
> gadget, where the BIOS normally fakes things out so it looks like a
> floppy... either need a kernel that really does have USB and USB-storage
> built in, or can't use that floppy until the module is loadable.  No
> TomsRtBt for them unless they customize it.
>
> Anyways if this is the case you might be able to do it *just as* the
> kernel gets started, but not after very far in and all, and not able to
> change the state after that.
>
> You could, however, try compiling the /dev/nvram feature in and see if
> you can see a diff in the results of reading it when you did switch it,
> or didn't switch it, but did nothing else special to the bott process.
> If so, maybe there's hope, since you can tickle /dev/nvram to hit the
> signal (CAREFULLY! don't hit any other values - dd bs=1 and skip), and
> see if it just... does the right thing.
>
> * Heather Stern * star@ many places...
>                 * Starshine Technical Services -*- 800 938 4078
>



Reply to: