Posts by @fterlife

    I noticed you already pulled in kernel 5.3 in the development, also for the Pi boards, and that you are now even working on kernel 5.4 integration. That's really impressive and quite bleeding edge I might say. :) Can you shed a light on current performance just so we can get a hint of the future? Thanks a lot for all your hard work.

    there won't be any RPi4 testbuilds for a while.

    That's understandable. The .002 build works quite well. Nevertheless curious as I am I made another image from the Github source. It boots and it's even running on Linux kernel 5.1.8 so that's quite amazing. However instead of scaling down the image my 720P tv shows only the upper left portion of the HDMI output when the Pi4 is set to 1080p. Other builds I made in the past few weeks had the exact same issue. The official test builds work fine though. Is there a reason why all builds straight from Github show this behavior? Is there something I can do with these builds that fixes the HDMI output so that my TV receives the entire output?

    Here's the original post:

    Raspberry Pi 4 temperature - Page 4 - Raspberry Pi Forums

    And here's an important follow-up:

    Raspberry Pi 4 - new VLI firmware issues - Raspberry Pi Forums

    Considering there are still some issues with the newer firmware I've flashed back to the original one, just to be sure.

    Based on the amount of issues I've seen with the Pi4 (heat, USB power delivery, USB power issue with GPIO) I am 1000% waiting for Pi 4.1 edition in 3 to 6 months.

    Also, sadfly, I wouldn't use one without a fan, period.

    And miss out on all the debugging action? :D I'm seriously enjoying my Pi 4 despite its initial issues. Minor issues that is. It's just so much more powerful than the Pi 3. Kodi has never been so responsive on my TV screen. It feels as if there's a full blown PC hiding behind the TV somewhere. My Pi 4 is not in a case yet and that's fine. That way there is no heat issue whatsoever. And no fan mind you, just a small heat sink on the SoC. USB power delivery is fine with an original RPi power source or any plain USB charger with high enough current output. I wouldn't want to use an expensive phone or laptop charger with it anyway. I'm patiently waiting for a good case to become available where I live. And for the time being I'd still pack my Pi 3 if I wanted to use Kodi on the road. Sure there are issues but simply don't expect to make full use of the hardware yet (such as running HDR 4K60 video) and just treat it as an overclocked Pi 3 for now. The rest of the awesomeness is on its way.

    And by the way, if you have lots of H.265 material such as myself then the Pi 4 actually has less issues with heat than a Pi 3. The thermometer icon regularly popped up with my Pi 3. I have yet to see it with my Pi 4.

    It is a known problem. Every generation of Pi has been a little bit hotter than the previous. Pi4 runs very hot - not just the CPU but several other components on the board are generating more heat than previous generation. The result is that many people consider active cooling necessary.

    Flirc case will help to cool the CPU a little bit but it is a bad design because it's a mostly-sealed metal box and only sinks heat from the CPU. With many other components generating more heat, the air inside will heat up something terrible and it will slowly cook your RPi. The material also makes it much harder to cut holes for venting or add a fan.

    I wonder how the Pi 4 will behave with enclosures such as these that are basically one large heat sink for all chips.Would make it easy to carry around also. I refuse to rely on active cooling. A Pi has to be completely silent to begin with, and I travel quite a lot so the less parts and wires that can break the better.

    Short update on my own nightly build that I made yesterday with all the latest code. It's quite stable and both H.264 and H.265 video play smoothly. There is a problem though with the HDMI output. My TV has a 720p display but when the HDMI output is 1080p instead of downscaling the image the TV crops the output and only displays the upper left part of the output. I'm not sure if feedback on a nightly build is appreciated but I wanted to share it anyway. ;)

    I don't expect output issues on a 1080p tv so for people who want to try this out I've made a torrent magnet of it. All this is is a build from the current source code in Github, for the RPi4 obviously.

    Magnet link:

    1. magnet:?xt=urn:btih:5788de8b21f45ffe32937138c19aef4940f2faa0&dn=LibreELEC-RPi4.arm-9.80-devel-20190719234137-235fdbc.img.gz&


    MD5 - d9fb3e6acc97780e2b5b0f5049f9888f

    SHA-256 - a021d8585df8e76a3bf21b9e78539946143e7e7169c8c3e31fec5e522a148397

    Figured it out! I ran make image again with full build log capture. From that I was able to determine that 'makeinfo' was missing on my system. After a quick search on the internet I found out I needed to install the 'texinfo' package. After that the build process succeeded and it's working hard again to finish all the rest. Thanks for the quick and necessary replies. :)

    "Doesn't seem to work" isn't too specific. I use this to build RPi4 images:

    1. ARCH=arm PROJECT=RPi DEVICE=RPi4 make image

    so long,


    I know. I couldn't be more specific because the attempt was several days ago and I just thought that perhaps I should use a different command. I'm on Kubuntu 19.04 by the way.

    Just minutes ago I started yet another attempt just so I could get the specific error during building. However I chose to download the source code first using ARCH=arm PROJECT=RPi DEVICE=RPi4 tools/download-tool and among the ongoing downloads I noticed a few errors such as this one:

    I don't know whether this is a local problem or a Github issue but it seems that not all source code is available. Perhaps that explains my failed attempt several days ago.

    Once all source code is fetched I will attempt another run.



    All sources seemed to have downloaded fine despite some errors. E.g. the previously mentioned flac-1.3.2.tar.xz package did end up in the sources directory after all. Unfortunately the same error occurred as last time:

    Both of the mentioned log files don't seem to give any extra info and simply repeat the error I quoted here. Hopefully this will shed some light on the issue.