Question about the frame buffer

So, on my laptop, if the display goes to sleep (blanks), if I ssh in, and tell it to reboot, or shut down, the display wakes up and shows me all the various services stopping and so forth… What controls that? Is it something in kernel? Is it something in a package? I ask because I have another machine that when I tell it to reboot, or shut down, the display doesn’t wake up, so I have no idea what the status of it is, without looking at it. Anyone out there know?

    • James Le Cuirot
    • October 13th, 2010

    Hmm. Not sure. Is one of them changing VT? Maybe that could trigger a wakeup. Also, is one of them using KMS? I suspect that would make a difference.

    • Duncan
    • October 13th, 2010

    Another vote for KMS being the difference, here.

    The one machine I run sshd on I start it manually, to ssh updates over to it from my build machine mainly, and I don’t run it otherwise so I seldom initiate reboot via ssh and thus don’t know how it reacts to it, visually, but I do run KMS on both machines now, and the KMS mode screen blanking behavior at the console is markedly different and more X-like than it was with VGACON/UMS.

    With VGACON (and with the legacy non-KMS framebuffers I’ve run) there’s an obscure terminal setting, set by default, that tells the screen to blank after so many minutes idle, but that’s all it does — the video card apparently remains active, simply sending all-black to the display, as the displays never go into sleep mode despite being blanked.

    With KMS OTOH, as with X, the displays actually physically go into sleep mode, powering down the backlights, indicating that the graphics card has dropped sync as well.

    Based on that, I’d guess that one or the other wakes the screen back up not based on physical input at the console, but on new output. I’m not sure which way it’d go, tho, without actually checking.

    • jsbronder
    • October 14th, 2010

    If you want to control this see setterm(1).

  1. No trackbacks yet.