Firmware/hardware split

[Deleted User][Deleted User] Posts: 4
Forum Member
According to the manual, the standby button switches between operation and standby modes. However:
  1. recording can take place in standby mode, and
  2. on entering operation mode, the firmware is loaded.
This implies the recording functions are hardware controlled (unless some of the firmware remains resident in standby mode).

If so, I wonder how best to think about the machine as it is clearly not just a special purpose PC. Anyone have some insights?

It also would affect what kinds of things might change as the firmware is upgraded.

Just trying to understand this impressive machine better... :)

Glyn

Comments

  • [Deleted User][Deleted User] Posts: 4,545
    Forum Member
    ✭✭✭
    When the machine is in standby but the a scheduled recording needs to start, the PVR seems to start up in almost exactly the same way but maybe with certain things disabled - like video/sound output and whatever else is not needed. I mean the loader counts up etc as usual and taking the machine out of 'standby' while its recording takes the same amount of time as a channel change. If the machine is recording when you switch it to standby and then switch it on from standby if it is still recording the EPG is still fully populated. It seems the software is either partially or (more likely) fully loaded when the machine is recording but in standby - and is controlling what is activated/available when in this state.
  • [Deleted User][Deleted User] Posts: 4
    Forum Member
    Interesting - thanks. Maybe the machine is more like a PC than I thought.

    I wonder if anyone has ported Linux to it, written their own firmware, etc.? ;-)
  • [Deleted User][Deleted User] Posts: 4,545
    Forum Member
    ✭✭✭
    Like an xbox? :) I've seen firmware modifications done on other more mature Humax machines but it still feel like early days for the 9200t. Humax Zone and Humax Community are a good source for looking at whats been done with modified firmware for other Humax devices.
  • [Deleted User][Deleted User] Posts: 455
    Forum Member
    My theory is that in standby the machine has the main processor etc. totally powered off but with something handling a timed wake-up (alarm module on SoC? And extra 8/16 bit micro in the system?)

    It only needs one alarm - as it suspends it just needs to check for the 1st scheduled recording, or scheduled wake-up, and set that as a time to boot.

    After it boots, it checks why it woke up to see whether to act as if awake (wake-up due to button, remote or scheduled wake-up) or whether to make a recording and then standby again when it's finished.

    Ian
  • [Deleted User][Deleted User] Posts: 5,528
    Forum Member
    The way PVRs tend to work is that when power is first applied there's just a small amount of special program visible to the CPU known as the bootloader (as it pulls up the hardware by it's boot straps!). This goes to the main place where the firmware lives (Nor flash, Nand flash or the hard drive) and loads the main operating system (in effect it's very like the BIOS in a PC loading Windows!). Once that main operating system is loaded in RAM it will stay there until the power cord is pulled and you start the process over again. The On/standby switch does nothing much more than switching on/off the video output and possibly spinning down/up the hard drive. But other than that a box in "standby" is pretty much identical to a box that is "on". When you press the front/remote button to transition from "standby" to "on" it does nothing much more than spinning up the HDD and enabling the scart switch (and maybe a few other things brought out of a "low power" mode). This is why a standby->on startup is far far quicker than a mains off->on startup.

    So that comment "2. on entering operation mode, the firmware is loaded." sounds like a bit of a white lie to me (unless they are talking about a cold reset)

    BTW, as the 9200 is (I believe) based on the NEC Emma2 chipset I think there's a strong chance that the operating system is actually Nucleus Plus though I guess it might be VxWorks

    Cliff
  • [Deleted User][Deleted User] Posts: 4,545
    Forum Member
    ✭✭✭
    Observing two units one of them is definately pretty much on when in standby. The 9200t standby seems to be in a much more comatose state that my Fusion when in standby. In the Fusion the machine is basically still on but with video and sound cut. Switching it on from standby takes the same amount of time the Fusion takes for say a channel change when on.

    When the 9200t is switched on from the back power switch, or switched on from the standby button on the front, there is a digital read out on the front panel indicating the boot loader version and a percentage indicator going up 10, 20, 30 ..100 until its finished loading what it needs - takes 10 seconds, then the machine goes through a "starting" (this is the readout on the front of the unit at this point) phase for 5 seconds until its able to display picture and sound.

    When the unit is in 'standby' but needs to do a timer recording it goes through the same process. I.e. the boot loader process described above on switch on happens with the count up progress indicator while it’s loading everything necessary up before the recordings start. I think in this case the machine is pretty much on normally but without displaying picture and sound, and switching on from this state is the same as a channel change. I dont think the EPG polulates in the background though even though at least one of the tuners is being used for a recordings and EPG one of the tuners is active so maybe some functions will not perform when the machine is on for a recording but stil in standby.
  • [Deleted User][Deleted User] Posts: 5,528
    Forum Member
    Ah that's interesting, I'll take back what I said about the "white lie" then. I was just going on my experience of what my Fusion does and now actually debugging what a Sky PVR does (it most definitely remains "alive" in standby and the 45 second bootloader only runs at first power application)

    Cliff
  • [Deleted User][Deleted User] Posts: 195
    Forum Member
    CJL wrote:
    BTW, as the 9200 is (I believe) based on the NEC Emma2 chipset I think there's a strong chance that the operating system is actually Nucleus Plus though I guess it might be VxWorks

    Cliff


    I believe it's VxWorks, as Humax are advertised on the Wind River website (well... guess it could be embedded linux?)
  • [Deleted User][Deleted User] Posts: 195
    Forum Member
    When in true standby, the box goes completely comatose, as when it comes out of standby, or starts to record, the firmware is loaded by the bootloader. The recording timer is handled by a seperate on-board RTC, which wakes the box up. When recording the 'GUI' functions are disabled, and the box 'appears' to be in standby, however, switching it 'on' with ther remote skips the bootloader (already loaded) and enables GUI functions.
  • [Deleted User][Deleted User] Posts: 4,545
    Forum Member
    ✭✭✭
    CJL wrote:
    Ah that's interesting, I'll take back what I said about the "white lie" then. I was just going on my experience of what my Fusion does and now actually debugging what a Sky PVR does (it most definitely remains "alive" in standby and the 45 second bootloader only runs at first power application)

    Cliff

    Yes. It would be nice if the 9200t took 5 seconds rather than 15 seconds to start up mind.
  • [Deleted User][Deleted User] Posts: 34
    Forum Member
    I believe it's VxWorks, as Humax are advertised on the Wind River website (well... guess it could be embedded linux?)
    Boeing plan to use this OS in their new 787 :eek:
  • [Deleted User][Deleted User] Posts: 195
    Forum Member
    I can just picture it... flight RA187 locked up on BBC1 at 17:59!
  • [Deleted User][Deleted User] Posts: 5,528
    Forum Member
    horaldo wrote:
    Boeing plan to use this OS in their new 787 :eek:
    Well any jet in the last 10 years (at least) is almost certainly already running some form of RTOS. The only things that have changed in the RTOS world in the last 10+ years is that WindRiver have gone round buying up the competition right left and centre in a very Microsoft-like way to control the market (and even prepare their own Linux strategy on the basis that they can't buy them out and some developers WILL use it as an operating system - so they'll offer you WindRiver flavoured Linux if you like).

    However I'd have no worries about a jet running VxWorks - it is a very good RTOS and is so well established in 1,000's and 1,000's of commercial products that its stability must rate orders of magnitude above Microsoft software.

    Now if Boeing announced they were going to design jets around Microsoft Windows CE I *would* get worried (very worried!)

    Cliff
Sign In or Register to comment.