/dev/mem not exist and fast-gpio not work



  • I upgraded firmware, fast-gpio gives segmentation fault.
    Also power-down not work because use fast-gpio.
    Please fix firmware and release patch.

    Thanks



  • @Giovanni-Gentile said in /dev/mem not exist and fast-gpio not work:

    I upgraded firmware, fast-gpio gives segmentation fault.

    This is a well known issue already reported in a number of threads

    Also power-down not work because use fast-gpio.

    No, that is unrelated. There is no actual powerdown capability, however a previous manifestation of a design bug used to sort of accomplish a halt to meaningful software execution, that from a user perspective looked a bit like an off state.



  • You would think powering down would be useful with a small portable device like an Omega.



  • @Costas-Costas It's not an Omega-specific thing, it's a more general thing stemming from the SoC being a consumer-level router-SoC -- I have several different kinds of routers and they all behave this same way, with all of them rebooting, instead of shutting down. These SoCs rarely have any power-management features, other than turning internal peripherals on or off, and thus you'd have to build a more complicated power-delivery system that provides external power-down features and you'd need to then provide a way for the SoC to also communicate with this power-delivery system -- rarely worth the effort and added complexity, all things considered.



  • @Chris-Stratton said in /dev/mem not exist and fast-gpio not work:

    No, that is unrelated. There is no actual powerdown capability, however a previous manifestation of a design bug used to sort of accomplish a halt to meaningful software execution, that from a user perspective looked a bit like an off state.

    I assume it hasn't been done, because the SoC wouldn't actually be off. It'd still be running, it'd just be stuck in a never-ending for-loop and it would be terribly misleading and an ugly solution. But then again, I ain't a kernel-dev and I am just making an educated guess on their motives.



  • @WereCatf sorry I thought the Omega was a portable computer from the details on the Kickstarter campaign.

    Most routers don't have SD cards attached, although some do have USB data ports.

    Onion should at least write a routine to make it as safe as possible to "pull the plug" with SD cards and USB's in situ.



  • @Costas-Costas said in /dev/mem not exist and fast-gpio not work:

    Most routers don't have SD cards attached, although some do have USB data ports.

    The fact that there is no SD-card slot doesn't mean the SoC wouldn't support SD-cards, though. I haven't even heard of a single router-SoC from the last decade or so that didn't support such. It's just a matter of the router-manufacturers deeming that there's little demand for including an SD-card slot, when most consumers are more likely to want to plug in USB-devices.



  • @WereCatf I think you are missing the point, again.

    If router manufacturers had added SD card's they would have a "power down" routine, ergo Onion to provide the same.



  • Power down isn't going to happen.

    sync();
    unmount_everything()
    printk("it's safe to turn off now"); 
    set_leds(ALL_OFF);
    while(1);
    

    Might, and would presumably be useful for those using delicate filesystems.

    This is after all what resetting the Omega 2+ in ignorance of the flash address mode bug used to effectively do.



  • @Chris-Stratton so Onion just need to tidy that up and compile as power-down, right?



Looks like your connection to Community was lost, please wait while we try to reconnect.