Pi: Put display to sleep?

In Configure/Screen, there is a “DPMS” screen saver which currently does nothing more than turning the screen black. Presumably it’s supposed to put the display to sleep, but I also assume that there have historically been complications with the Pi’s odd graphics architecture.

CVS was updated a few days ago with an implementation of VCGenCmd. It’s now possible to enter *VCGenCmd display_power 0 to put the screen to sleep, and *VCGenCmd display_power 1 to wake it up again. This works on my Pi 1, at least.

The graphics driver is supposed to enable/disable the display in response to GraphicsV 4. I haven’t looked at the source, but presumably the DPMS screen saver uses this vector. How practical is it to update BCMVideo to effectively issue a VCGenCmd display_power command in response to GraphicsV 4? If everything works the way that I think/hope it does, then that should be enough to get proper display sleep working. And that means I’ll probably stop turning the machine off every night :)

Thoughts? Do any experts know how much work this would be?

And that means I’ll probably stop turning the machine off every night :)

I just turn the monitor off! But it’s a good point!

There’s an undocumented hdmi_blanking config.txt option which will control whether DPMS is used when the screen is blanked. Setting it to 1 should do the trick.

Apparently they can’t enable that option by default because it causes problems with software which doesn’t use the standard framebuffer: https://github.com/raspberrypi/linux/issues/487

I’ll have a think about whether there’s an easy way of making the OS use display_power – possibly this issue needs fixing first (GraphicsV_SetBlank needs to support being called from an interrupt handler, but *vcgencmd is a bit heavyweight for that, so delaying the blank/unblank until a callback or RTSupport routine would be required – for which we’d also need to make sure no interrupt handlers can get stuck waiting for VSync)

There’s an undocumented hdmi_blanking config.txt option which will control whether DPMS is used when the screen is blanked. Setting it to 1 should do the trick.

That doesn’t seem to make any difference – the screen just blanks the same as before – that is, it’s still working.

It worked when I tried it – perhaps your firmware is too old? (over 1 year old?) I’m not sure when the option was first introduced.

There’s an undocumented hdmi_blanking config.txt option which will control whether DPMS is used when the screen is blanked. Setting it to 1 should do the trick.

Thanks. I’ll give that a go tomorrow.

GraphicsV_SetBlank needs to support being called from an interrupt handler, but *vcgencmd is a bit heavyweight for that

Nothing is ever simple! :) Don’t worry about it; if there’s already a functional solution then there’s no reason to “hack around” in VCGenCmd.

That doesn’t seem to make any difference – the screen just blanks the same as before – that is, it’s still working.

Speaking from memory of use of screen blanking on older machines and older displays then perhaps what you describe is a reflection of the capabilities of the display?
It’s certainly the case that I’ve had different screens blank or shutdown with the same machine settings. Since the machine hadn’t changed the difference had to be the display unit.

Some units seem to power down the display when blanked while others simply blank (blackout) the screen.

There’s an undocumented hdmi_blanking config.txt option which will control whether DPMS is used when the screen is blanked. Setting it to 1 should do the trick.

Works here on an RPi3 with the following firmware :-

bootcode.bin  31May16
fixup.dat     11Oct16
start.elf     13Oct16

Thanks for the tip, a welcome feature.

Works here on an RPi3 with the following firmware :-

Where may I find the latest firmware?

Where may I find the latest firmware?

Belay that!

https://github.com/raspberrypi/firmware/tree/master/boot

hdmi_blanking=1 confirmed working on my system. Thanks for the tip.

As far as I’m concerned, this “wish” can now be cancelled :)

hdmi_blanking=1 confirmed working on my system. Thanks for the tip.
As far as I’m concerned, this “wish” can now be cancelled :)

OK, having upgraded to firmware 9th Dec 16, I find that adding the Config line

hdmi_blanking=1

makes no difference here.

I already had a screenblanker set in Config>Screen of DPMS blank delay 10 min.

This, when operational, blanked the screen on my PL2409HD (Iyama) monitor¹ to give a slightly illuminated dark grey screen – the monitor was already displaying what, I assume, the RPi was sending as “black”.

I had hoped that the Config line would make this even blacker by either sending nothing at all, or actually switching the monitor to a sleep mode – but, apparently, not so.

So – what am I missing here? Is my monitor incapable of receiving a sleep mode command, are the other respondents mistaking a previously-untried DPMS setting for a miracle – or what?

I am, to be completely transparent, using the AnyMode module to achieve my preferred screen mode. Could this make a difference?

¹I think of it as quite modern – but then I remember the Coronation!

It’s definitely going into full sleep for me, with a Philips of some description (I don’t have the exact model number handy). I’m using an HDMI cable, connected via an adapter to the DVI port. I do not currently have AnyMode installed.

It’s definitely going into full sleep for me, with a Philips of some description (I don’t have the exact model number handy).

Well I guess I’m just a bit unlucky then – nice to have it clarified! Thanks!

hdmi_blanking=1 … makes no difference here.

This works now! The reason was that I had somehow switched the active boot partition to another location on the boot disc, and I was updating files in a ghost DOS image within Boot! This also caused other problems as well!

My apologies!