Search Results

Search results 1-20 of 1,000. There are more results available, please enhance your search parameters.

  • User Avatar

    This is strange. Does the freeze always happen at the same time when the rootfs is mounted? Or can it also be earlier or later? We assume that this may be caused by a floating line that triggers some interrupts. Can you e-mail us your schematics? Or at least the part of the schematics where the PicoCOMA5 is involved? Then we can check if there is anything unusual that may trigger the problem. Your F&S Support Team

  • User Avatar

    What is the last message before the freeze and the first message after the freeze? Maybe this can give a hint about the device that is causing the delay. Your F&S Support Team

  • User Avatar

    No, I do not think that your freeze is related to the i.MX6SX issue. On i.MX6 this was triggered when the bus frequency was lowered to 24 MHz, which is only possible with some RAM types and if no device driver requests a high bus frequency anymore. This was introduced by NXP in kernel 4.1, so I do not think that the Vybrid CPU on PicoCOMA5, that is still on an older kernel, has any similar power saving functions. Usually you try to isolate such a problem by disabling nearly all devices but the s…

  • User Avatar

    Usually you just look for the hardware specs of the device that you want to use. And when you have found a matching device, then you check if there is support for it in the kernel. Either by looking at the menuconfig help entries, by looking into the kernel documentation or by checking the website of the manufactorer or Google. Or even easier, by asking the hardware vendor if there is Linux support. Not all drivers are part of the Linux tree. For example touch drivers and WLAN drivers are often …

  • User Avatar

    Quote from leitmotiv: “Ok from this replies i assume you don't understand what i am saying or giving just political reply” Why do you think that? I tried to answer very elaborately, trying to address your direct complaints. Let me summarize Your complaint: Why does it not work across different versions? My answer: Because to improve NAND stability, the NAND layout needed to be changed. This change unfortunately could not be done without breaking compatibility. Your complaint: But why not leave i…

  • User Avatar

    i.MX6SX Linux Release V2.1 After over a year of preparation, it is finally available. We have uploaded a new Linux version for all boards and modules based on the i.MX6-SoloX CPU to our server, i.e. the fsimx6sx architecture: efusA9X and PicoCOMA9X. This release is running on all platforms of this architecture at the same time. So you can download the same binaries for NBoot, U-Boot, Linux kernel and Buildroot root filesystem to any of the i.MX6-SoloX boards and it will run. This is a maintenanc…

  • User Avatar

    Quote from leitmotiv: “1) i can't use mtd-utils and fw_printenv/setenv anymore because now uboot flash structure is changed, was base on previous uboot version you given” No, we have not changed Uboot flash structure, we have changed flash structure for *all* parts of the software. Is this so difficult to understand? The original NXP/Freescale code was very incomplete and unreliable. As long as there are no bit errors in the flash, everything seems to be OK. But we have NAND flash, and NAND flas…

  • User Avatar

    Quote from leitmotiv: “1) We use mtd-utils together with the tools called fw_printenv (and fw_setenv), to set uboot parameters from userspace. This is needed because on same terminal we use different displays (lvds 7", 12", 15") so it can be configured at any moment, like other parameters” OK. So where is the problem? You can use these tools. Quote: “2) We cannot give to our technicians another board, because the technician is in another country and sometime is very expensive and takes very long…

  • User Avatar

    The i.MX6-SoloX CPU (and therefore the efusA9X) does not have a hardware video processing unit (VPU). So all video processing must be done in software. But H.264 (MPEG-4 AVC) decoding in software does not make sense. The computing required to do this is so intense that the result will not have an acceptable frame rate. This is why we did not add any MPEG-4 codecs to our standard configuration in the first place. You can try MPEG-2, there the computation is less complex. But I would guess that ev…

  • User Avatar

    Quote from leitmotiv: “Problem 1) Some of our technicians only have Windows 10, and some versions of nboot on some cards does not permit to load uboot with serial...” That's simply not true. Serial download of U-Boot is possible on *every* NBoot version that we ever released. You just need to press 'd' (lower case). In previous versions it was a difference if you press 'd' (lower case) or 'D' (upper case). Lower case was for U-Boot, upper case for NBoot. On newer NBoot versions, they both are th…

  • User Avatar

    i.MX6 Linux Release V3.1 After over a year of preparation, it is finally available. We have uploaded a new Linux version for all boards and modules based on the i.MX6 CPU to our server, i.e. the fsimx6 architecture: armStoneA9, armStoneA9r2, PicoMODA9, efusA9, NetDCUA9, QBlissA9 and QBlissA9r2. This release is running on all platforms of this architecture at the same time. So you can download the same binaries for NBoot, U-Boot, Linux kernel and Buildroot root filesystem to any of the i.MX6 boar…

  • User Avatar

    Have you tried to use a download manager program? These programs can usually resume an interrupted download. Or under Linux (or CygWin) simply use wget to download the file. wget will also retry the file several times if the connection breaks and only fetches the remaining part. And even if you have to start wget again, you can give option -c (or --continue) to continue with the same file without having to retransmit everything again. Your F&S Support Team

  • User Avatar

    Let me explain again. In the past we had several releases for i.MX6 (usually called fsimx6): - V0.1 and V0.2 for QBlissA9 only (the V0.x series). This was 2012. - V1.0, V1.1, V1.2 (the V1.x series). This was 2013. - V2.0, V2.1 (the V2.x series). This was 2015. - V3.0. This was 2016. In the next few days there will be a new V3.1. (The 3.x series). Between V1.1 and V2.0, we totally re-implemented the NAND flash access that made everything much more stable and reliable than before. This unfortunate…

  • User Avatar

    The HTML page that you get is most probably a "file not found" page. So when Buildroot tries to download the package, it is not found on the server anymore. This may happen from time to time when package maintainers move to a different provider or change their directory structure on the server. Unfortunately when the server sends such a page, Buildroot can not detect the error, because the download itself was successful and only the content of the download was different. Therefore it does not st…

  • User Avatar

    These timings look absolutely correct. So the pixelclock should be right, about 9MHz. What I see from the data sheet, this display latches data on the rising edge of the pixel clock, which is rather unusual (and different to other ETM043 display variants). So you should invert the pixel clock polarity. You can do this by adding FB_SYNC_ON_GREEN to the .sync entry above. This is "misused" as pixel clock polarity in the Vybrid display driver. This should fix the jagged graphics, for example in ver…

  • User Avatar

    Please see also the other thread. You must not mix versiosn of NBoot, U-Boot and Linux. In fsimx6-V2.0 we implemented a completely new NAND access, which is not compatible to the previous version. So if you use Linux kernel and root filesystem based on fsimx6-V1.x release, you must use old NBoot (<=VN17) and old U-Boot. New NBoot and U-Boot are only possible if you use software based on fsimx6-V2.0 or newer. If you use the correct versions, then also the fw_printenv will work again. If you use d…

  • User Avatar

    VERY IMPORTANT! If you use software based on one of the releases of the V1.x series, then do *not* update NBoot (because there is no simple way to downgrade again). For fsimx6-V1.x, you need an NBoot from before VN20, the best is VN17. Only if you use software from release fsimx6-V2.0 or newer, you must use a newer NBoot, at the moment VN29 or newer is recommended. This was all explained in the release notes to fsimx6-V2.0. Nonetheless keep in mind that the NAND code of fsimx6-V1.x is much less …

  • User Avatar

    The SKIT for PicoCOM4 and PicoCOMA5 ist exactly the same. In fact it is already from PicoCOM2, with only a small modification on the VBUS voltage for the USB device port. Why the PicoCOMA5 is running slower on your base board, I can not explain. Are you sure that this is not the result of your display driver modification? How did you change the settings for the display? Because if the clock speed is too high, then the DMA transfers to the display could block the system bus so that the CPU core c…

  • User Avatar

    GPIOs general questions

    fs-support_HK - - efusA9 NXP i.MX6

    Post

    1.) Theoretically yes. If a pad is not used by another function, it can be accessed as GPIO. But actually it depends on the function it was set before, either in the bootloader or how the pad is initialized by the CPU at reset time. Most pads *are* initialized as GPIO, then it will work directly. But some pads are set to a different function, and then you also have to add a pad setting to the iomuxc section of the device tree, to set the pad to GPIO function. If I remember correctly, the LCD col…

  • User Avatar

    The question is: who is generating this error? Is this really the UART driver? Or is it simply the stty command line tool that can not handle these large baud rates? Probably you have to try a small UART test C program where you set the baud rate with the system function and then transmit some bytes. Then you know whether the system call fails or not. From the driver side I see no restrictions. If there is a divider that can be used to divide the base clock, then it should work. However the base…