This document explains how to bring-up u-boot & Linux via JTAG This procedure has been tested on: i.MX6 Solo X Sabre SD i.MX6UL EVK Prerequistes: Get the latest BSP for your board. This procedure was tested with L4.1.15. Build the 'core-image-minimal' image to bring-up your board (Detailed steps here) Optional- Build a meta-toolchain for your device 1.- Set board to boot from Serial dowloader mode or set it to boot from the SD card and remove the sd card We basically want the board to stall in boot ROM to attach to the target. 2.- Connect JTAG probe and turn on the board The device should stall trying to establish a connection to download an image, this will allow us to attach to the target. 3.- Load Device Configuration Data In 'normal' boot sequence the boot ROM takes care of reading the DCD and configuring the device accordingly, but in this case we are skipping this sequence and we need to configure the device manually. The script used by Lauterbach to parse and configure the device is called dcd_interpreter.cmm and can be found here. Search for the package for your specific device. The DCD configuration for your board should be on your u-boot directory: yocto_build_dir/tmp/work/<your board>imx6ulevk/u-boot-imx/<u-boot_version>2016.03-r0/git under board/freescale/<name of your board>mx6ul_14x14_evk/imximage.cfg This file (imximage.cfg) contains all the data to bring up DRAM among other early configuration options. 4.- Load U-boot If an SREC file of U-boot is not present build it (meta-toolchain installed required) the SREC file contains all the information required by the probe to load it and makes this process easier. To build the SREC simply type: make <your board defconfig>mx6ul_14x14_evk_defconfig  (all supported boards are found under u-boot_dir/configs) make If you cannot build an SREC or do not want to, you can use the u-boot.imx (located under yocto_build_dir/tmp/deploy/images/<your board name>/) or u-boot.bin files but you will need to figure out the start address and load address for these files, this can be done by examining the IVT on u-boot.imx (here is a useful document explaining the structure of the IVT). Let U-boot run and you should see its output on the console I will try to boot from several sources but it will fail and show you the prompt. 5.- Create RAMDisk After building the core-image-minimal you will have all the required files under yocto_build_dir/tmp/deploy/images/<your board name>/ You will need: zImage.bin - zImage--<Linux Version>--<your board>.bin Device tree blob - zImage--<Linux Version>--<your board>.dtb Root file system - core-image-minimal-<your board>.rootfs.ext4 We need to create a RAMDisk out of the root file system we now have, these are the steps to do so: Compress current Root file system using gzip: gzip core-image-minimal-<your board>.rootfs.ext4 If you want to keep the original file use: gzip -c core-image-minimal-<your board>.rootfs.ext4 > core-image-minimal-<your board>.rootfs.ext4.gz Create RAMDisk using mkimage: mkimage -A arm -O linux -T ramdisk -C gzip -n core-image-minimal -d core-image-minimal-<your board>.rootfs.ext4.gz core-image-minimal-RAMDISK.rootfs.ext4.gz.u-boot Output: Image Name: core-image-minimal Created: Tue May 23 11:28:55 2017 Image Type: ARM Linux RAMDisk Image (gzip compressed) Data Size: 3017939 Bytes = 2947.21 kB = 2.88 MB Load Address: 00000000 Entry Point: 00000000 Here are some details on mkimage usage Usage: mkimage -l image -l ==> list image header information mkimage [-x] -A arch -O os -T type -C comp -a addr -e ep -n name -d data_file[:data_file...] image -A ==> set architecture to 'arch' -O ==> set operating system to 'os' -T ==> set image type to 'type' -C ==> set compression type 'comp' -a ==> set load address to 'addr' (hex) -e ==> set entry point to 'ep' (hex) -n ==> set image name to 'name' -d ==> use image data from 'datafile' -x ==> set XIP (execute in place) mkimage [-D dtc_options] [-f fit-image.its|-F] fit-image -D => set options for device tree compiler -f => input filename for FIT source Signing / verified boot not supported (CONFIG_FIT_SIGNATURE undefined) mkimage -V ==> print version information and exit 6.- Modify U-boot's environment variables Now we need to modify U-boot's bootargs as follows: setenv bootargs console=${console},${baudrate} root=/dev/ram rw We need to find out the addresses where u-boot will expect the zImage, the device tree and the initial RAMDisk, we can do it as follows: => printenv fdt_addr fdt_addr=0x83000000 => printenv initrd_addr initrd_addr=0x83800000 => printenv loadaddr loadaddr=0x80800000 Where: fdt_addr -> Device tree blob load address initrd_addr -> RAMDisk load address loadaddr -> zImage load address 7.- Load zImage, DTB and RAMDisk Now we know where to load our zImage, device tree blob and RAMDisk, on Lauterbach this can be achieved by running the following commands: Stop the target and execute: data.load.binary zImage.bin 0x80800000 data.load.binary Your_device.dtb 0x83000000 data.load.binary core-image-minimal-RAMDISK.rootfs.ext4.gz.u-boot 0x83800000 Let the device run again and deattach from the device in lauterbach this is achieved by: go SYStem.mode.NoDebug start the boot process on u-boot as follows: bootz ${loadaddr} ${initrd_addr} ${fdt_addr} You should now see the Linux kernel boot process on your terminal: After the kernel boots you should see its prompt on your terminal: Since we are running out of RAM there is no way for us to save u-boot's environment variables, but you can modify the source and compile u-boot with the new bootargs, by doing so you can create a Load script that loads all the binaries hits go and the boot process will continue automatically. One way to achieve this is to modify the configuration file under U-boot_dir/include/configs/<your board>.h find the mfgtool_args and modify accordingly. The images attached to this thread have been modified as mentioned.

In FSL i.MX53 reference design, it is configured as: static struct mxc_audio_platform_data sgtl5000_data = { .ssi_num = 1, .src_port = 2, .ext_port = 5, .hp_irq = gpio_to_irq(HEADPHONE_DEC_B), .hp_status = headphone_det_status, .init = mxc_sgtl5000_init, .ext_ram_rx = 1, }; by default. If change the configuration to be : static struct mxc_audio_platform_data sgtl5000_data = { .ssi_num = 0, .src_port = 1, .ext_port = 5, .hp_irq = gpio_to_irq(HEADPHONE_DEC_B), .hp_status = headphone_det_status, .init = mxc_sgtl5000_init, .ext_ram_rx = 1, }; There will prompt "imx_ssi_irq mxc_ssi SISR 8003a3 SIER 180100 fifo_errs=XXXX"  constantly, and audio is greatly distorted. The root cause of this issue is that SSI1/3 use SDMA, and also use IPMUX, but there is not the clock dependency between SDMA and IPMUX, so sometimes IPMUX clock is closed automatically. The attached patch may fix this issue. NOTE: If use SSI2 .ssi_num = 1,             .src_port = 2, If use SSI1 .ssi_num = 0,             .src_port = 1,

Procrank can be used to check if a process has memory leakage. Procrank will list four types of memory usage. For details refer to: http://elinux.org/Android_Memory_Usage Vss = virtual set size Rss = resident set size Pss = proportional set size Uss = unique set size Uss can be used to check if a process has memory leakage. If the Uss increases when some operations start and stop, this means there could be memory leakage. Procrank can get from: <myandroid>/out/target/product/<product_name>/system/xbin/procrank and also needs to push to the library you target: <myandroid>/out/target/product/< product_name >/system/lib/libpagemap.so

Q: What is the Min LPDDR2 clock frequency allowed by the i.MX6? The Jedec Spec for LPDDR2 allows for a min tck period of 100ns. Are there any required relashionship between the DDR clock frequency and other clocks in the i.MX6? A: The JEDEC maximum period for the MX6 is 100nS as per the LPDDR2 specification.  There is a minimum period during boot, before everything is configured and fully up to speed of 18nS. Are you saying the imx6 memory controller can operatate down to the min frequecies specified in the LPDDR2 JEDEC spec? Given that there is no limit specified in the data sheet, it should operate that slowly, provided the clocking can be set for it to operate so slowly. I would imagine that the core will need to be running slowly as well, since it does not make sense to slow the memory bus without slowing the core down as well.

JTAG Hardware and Software There are many opened and proprietary JTAG solutions. Here are some of them: Proprietary IAR Systems In-Circuit Debugging Probes Macraigor usb2Demon Segger - Jlink Free and Open Source Software GDB OpenOCD Open Hardware Turtelizer

A new release of the manufacturing tool is was recently made available, "imx-3.10.53_1.1.0_ga-mfg-tools". It can be found in the software download sections for the iMX6 family. However, it can be used to program an iMX28 in a Win7 64-bit host by adding a few files. The steps to do so are listed below and can be checked against the script in ucl2.xml.   Download the attached "28.vbs" file and place it into where the manufacturing tool was installed, typically in  <install_dir>\mfgtools\   Replace <install_dir>\mfgtools\Profiles\Linux\OS Firmware\ucl2.xml with the attached ucl2.xml.    Copy the attached files "updater_ivt.sb" and "fdisk-u.input" into <install_dir>\mfgtools\Profiles\Linux\OS Firmware\firmware Copy your iMX28 image file into <install_dir>\mfgtools\Profiles\Linux\OS Firmware\files.  The file should be renamed to "linux.sb" to conform with the ucl2.xml script. Copy your "rootfs.tar.bz2" file into <install_dir>\mfgtools\Profiles\Linux\OS Firmware\files To launch the manufacturing tool, double click on "28.vbs". Issue: After MfgTool has finished and the progress bars have turned green, clock on the Stop button or the program will start another cycle.

Kernel provides mtdoops to dump kmsg to MTD device, but MMC card is not a MTD device. We can let user-space program to execute the write operation to dump kmsg into block storage. The sample code is below. kernel space -- #include <linux/kernel.h> #include <linux/module.h> #include <linux/console.h> #include <linux/vmalloc.h> #include <linux/seq_file.h> #include <linux/workqueue.h> #include <linux/sched.h> #include <linux/wait.h> #include <linux/delay.h> #include <linux/interrupt.h> #include <linux/kmsg_dump.h> #include <linux/proc_fs.h> static struct kmsg_dumper dump; static struct proc_dir_entry *my_proc; static int is_panic = 0; static int my_proc_show(struct seq_file *m, void *v) {     seq_printf(m, "%d", is_panic);     return 0; } static int my_proc_open(struct inode *inode, struct file *file) {     return single_open(file, my_proc_show, NULL); } static const struct file_operations my_proc_ops = {     .open        = my_proc_open,     .read        = seq_read,     .llseek        = seq_lseek,     .release    = single_release, }; static void oops_do_dump(struct kmsg_dumper *dumper,         enum kmsg_dump_reason reason, const char *s1, unsigned long l1,         const char *s2, unsigned long l2) {     int i;     printk("### [%s:%d] reason = %d\n", __func__, __LINE__, reason);     is_panic = 1;     for (i = 0;i < 10; i++)         msleep(1000);     printk("### [%s:%d] should be done\n", __func__, __LINE__); } static int __init my_oops_init(void) {     int err;     dump.dump = oops_do_dump;     err = kmsg_dump_register(&dump);     if (err) {         printk(KERN_ERR "oops: registering kmsg dumper failed, error %d\n", err);         return -EINVAL;     }     my_proc = proc_create("dump_tester", 0, NULL, &my_proc_ops);     return 0; } static void __exit my_oops_exit(void) {     printk("### [%s:%d]\n", __func__, __LINE__);     if (my_proc)         remove_proc_entry( "dump_tester", NULL);     kmsg_dump_unregister(&dump); } module_init(my_oops_init); module_exit(my_oops_exit); MODULE_LICENSE("GPL"); User space -- #include <stdio.h> #include <stdlib.h> #include <string.h> #include <sys/types.h> #include <sys/stat.h> #include <unistd.h> #include <fcntl.h> #include <poll.h> #define BUF_LEN 40960 void main(int argc, char **argv) {     char tmp = 'X';     char buf[BUF_LEN];     int fd_src, fd_trg;     int fd = open("/proc/dump_tester", O_RDONLY, 0);     while(1) {         lseek(fd, 0, SEEK_SET);         read(fd, &tmp, 1);         //printf("### [%s:%d] ==> '%c'\n", __FUNCTION__, __LINE__, tmp);         if (tmp == '1') {             fd_src = open("/proc/kmsg", O_RDONLY, 0);             fd_trg = open("/dev/block/mmcblk0p6",  O_RDWR, 0);             memset(buf, 0, BUF_LEN);             write(fd_trg, buf, BUF_LEN);             lseek(fd_trg, 0, SEEK_SET);             read(fd_src, buf, BUF_LEN);             write(fd_trg, buf, BUF_LEN);             close(fd_src);             close(fd_trg);             sleep(1);             printf("### dump panic log into %s\n", "/dev/block/mmcblk0p6");             break;         }         sleep(1);     }     close(fd); }

We have validated Toshiba Smart NAND in our i.MX6SX platform, and boot successfully. The results are as below: 1. chip part number: THGBR2G5D1JTA00,  page_size: 16k+64  pages_per_block: 256 2. test platform: i.MX6SX Some information to take care of: 1. The pin assignment of smart nand is different from common raw nand, that is, Nand pin1 must connect to Vcc, pin2 connects to Vss, pin23 connects to VssQ, pin24 connects to VccQ, pin38 connects to VccQ 2. The ECC layout of FCB page itself must be set according to the i.MX6SX RM, otherwise FCB can't be read correctly. 3. EccBlock0EccType and EccBlockNEccType in FCB must be set as 0, and raw data can be put in DBBT and firmware without any ECC check codes.

This document is describe the control method of USB Power based on I.MX6ULL-EVK.   1. Hardware Design USB OTG can be working as device mode, host mode, or switching between both device mode and host mode. (a) USB OTG Device mode When USB OTG is working as device mode, the USB OTG ID pin should be pulled up by a resistor. And the 5V power comes from the VBUS pin of MicroUSB which is the 5V from the external USB HOST. From the above schematic :- When the USB OTG1 ID is pulled to high, the G/S of the NMOSFET Q2SK3018 is turned on. The ENA pin of U1101 is low, and then no 5V output from OUTA. If USB OTG1 is connected to the external USB HOST, a 5V power source will enter the board and supply to the USB OTG1 VBUS pin. (b) Switch between Device and Host mode This is also called USB OTG’s Dual Role. When we plug in MicroUSB with USB TYPE-AB to USB TYPE-A-F cable, USB OTG1 will switch from USB device to Host mode. At this time, the USB OTG1 ID pin will be pulled down to low, the G / S of the NMOSFET will be turned off, the ENA pin of U1101 will be pulled low, and OUTA will output 5V voltage to the VBUS pin of MicroUSB. (c) USB OTG Host mode When USB OTG is working as host mode,  USB OTG1 ID pin connect an external resistor to pull it down or use internal resistor to pull it down. If using external resistor, 2.2K/3.3K ohm resistor is recommended. The USB_OTG1_PWR should output High to enable ENA of U1101, then OUTA will output 5V to USB OTG1 VBUS. On i.MX6ULL-EVK, USB OTG2 port is designed to be Host mode. It can be a design reference. When the USB_OTG2_PWR pulled to high, the U1101 supplies 5V to the outside. 2. Software modification The GPIO1_IO04 can be used to control USB OTG1 power.

Add the above line to conf/local.conf file: PREFERRED_PROVIDER_virtual/kernel = "linux-fslc" Check this page to see what the mainline Linux kernel supports for a particular Freescale board.

Introduction This is a brief guide showing how to integrate the driver for the WF111 module to the i.MX6 BSP Release. In this case the WF111 driver is available on a repository and it’s in accordance with the Yocto Project, which allows to easily customize a linux distribution for your board. Requirements WF111 Documentation – Silicon Labs have made a great job of documenting the steps to add the WF111 driver to a Linux distribution and have created Application Note 996 (link below), which we will use as reference. http://www.silabs.com/documents/login/application-notes/AN996.pdf WF111 Driver - We will also be using the Yocto layer included on the following repository: https://github.com/engicam-stable/meta-engicam i.MX6 3.14.52 BSP Release – In out scenario the WF111 layer that will be imported includes a driver that it’s compatible with Linux Kernel 2.6.24 up to 4.1., which it’s important to keep in mind.   Installing the 3.14.52 BSP Release First, setup the 3.14.52 BSP as described on the i.MX Yocto Project User’s Guide.   Adding the WF111 Driver Layer Clone the WF111 Driver Layer to your sources folder inside the BSP Release directory. Since the 3.14.52 BSP Release is based on Fido we will clone the Fido branch of the driver repository. $ cd <BSP_RELEASE_DIR>/sources $ git clone https://github.com/engicam-stable/meta-engicam -b fido‍‍  Once the layer is cloned you would need to add the new later editing the bblayers.conf file located the following path: <BSP_RELEASE_DIR>/<BUILD_DIR>/conf/bblayers.conf By adding the following line to add the new layer.   BBLAYERS += " ${BSPDIR}/sources/meta-engicam "‍   This should make the wf111-driver available through bitbake since bitbake will now look into this layer for all available recipes. You can then add the driver to your image by adding the following line to the <BUILD_DIR>/conf/local.conf   IMAGE_INSTALL_append += "wf111-driver"‍ Or you may create a new image recipe that includes the wf111-driver package. However, there are certain kernel options that must be enabled for the driver to work.   Creating an append to configure the kernel options Before we can bake an image with the WF111 driver we would need to edit the kernel options as mentioned on Silabs AN996. The following kernel options must be enabled:   CONFIG_WIRELESS_EXT CONFIG_MODULES CONFIG_FW_LOADER We would need to add the CONFIG_WIRELESS_EXT as the other two options are enabled on the BSP by default.   This involves adding an addendum to the kernel recipe to change its configuration. You may either add this append to any layer. The best way to handle it would be using a new layer for all your customization. You can find how to create a new layer on the following document: https://community.nxp.com/docs/DOC-331917 We’ll use a new layer called meta-newlayer for this example. It’s important that this layer has a high priority so the changes from the bbappend are not overridden. The following alternative was suggested by Chris Hossack on the following thread: https://community.nxp.com/thread/376369 First, run the menuconfig tool on the bitbake environment: bitbake linux-imx -c menuconfig Enable the necessary options: Networking Support > Wireless > cfg80211 wireless extensions compatibility   Save the configuration and exit. Then run the following bitbake command, which will create a config fragment file that contains the changed made to the default kernel options. bitbake linux-imx -c diffconfig We’ll make an append file that adds the required options.  Content of the config fragment:   CONFIG_WIRELESS_EXT=y CONFIG_WEXT_CORE=y CONFIG_WEXT_PROC=y CONFIG_WEXT_SPY=y CONFIG_WEXT_PRIV=y CONFIG_CFG80211_WEXT=y CONFIG_LIB80211=y CONFIG_LIB80211_CRYPT_WEP=y CONFIG_LIB80211_CRYPT_CCMP=y CONFIG_LIB80211_CRYPT_TKIP=y # CONFIG_LIB80211_DEBUG is not set CONFIG_HOSTAP=y # CONFIG_HOSTAP_FIRMWARE is not set‍‍‍‍‍‍‍‍‍‍‍‍‍    Since we are appending the kernel layer we need to add the addendum on the same path as that of the original kernel recipe but within our layer and create the append file there. Also add the WF111.cfg file to the linux-imx directory:   We would need to copy (and you may rename it as well) to the folder where are will be creating the append recipe for the kernel. Copy:  <BSP_RELEASE>/<BUILD_DIR>/tmp/work/<MACHINE>-poky-Linux-gnueabi/linux-imx/<KERNEL_VERSION>/fragment.cfg To: <BSP_RELEASE>/sources/meta-newlayer/recipes-kernel/linux/linux-imx/WF111.cfg You can do so suing the following command: cp <BSP_RELEASE>/<BUILD_DIR>/tmp/work/<MACHINE>-poky-Linux-gnueabi/linux-imx/<KERNEL_VERSION>/fragment.cfg <BSP_RELEASE>/sources/meta-newlayer/recipes-kernel/linux/linux-imx/WF111.cfg‍ (Please note that the file was renamed for ease, but you may use any name for the config fragment)   We need to create the bbappend file on the following path (as it must be the same relative path as the original recipe it is appending) <BSP_RELEASE>/sources/meta-newlayer/recipes-kernel/linux/linux-imx_3.14.52.bbappend   The linux-imx_3.14.52.bbappend file would contain the following:   SRC_URI += "file://WF111.cfg"  do_configure_append() {          #this is run from         #./tmp/work/<MACHINE>-poky-linux-gnueabi/linux-imx/3.14.52-r0/git          cat ../*.cfg >> ${B}/.config  }‍‍‍‍‍‍    After creating this recipe you should be able to bake any image from the BSP and see the driver there. I tested with the core-minimal-image and found that the files were indeed added to /lib/firmware. $ bitbake core-image-minimal ‍‍‍

On imx8qm there are two DPUs(display process unit) and one ISI(image subsystem interface), ISI has 5 inputs and two of them are from DPU0 and DPU1.   This document demonstrates on how to loopback DPU1 outputs to ISI. Note that only mipi dsi0 of dpu0 and lvds1 of dpu1 can be loopbacked to isi.   Platform:                            imx8qm b0 mek OS:                                    yocto 4.14.78 ga hardware connection:        imx8qm lvds1 ====> it6263 cable =====> hdmi display.   1st: isi has 8 pipelines which can be assigned to any of the 5 inputs, this doc takes the 5th pipeline to sink the dpu1 input. So you will need to configure the isi_4( start from 0) source in the dts and write a simple v4l2 subdev for capture testing, the default isi_4 device will be /dev/video4.   2st: configure both framegen0 of dpu1 and lvds1's link to pixellink 3.   3st: write a v4l2 userspace program to capture from /dev/video4 device, take this vulkan-capture as an example. Note that Vulkan-capture is rendered by vulkan api, you can also take opengl es for rendering.   See the atttachments for details.   ======================================== 2019/11/12 update patches. ======================================== 2019/12/19 add patch. Support connect real display to DC1-LVDS1   Note: for ISI loopback,  it needs output of 2x GPIO (4x for HDMI-TX or combo PHY) to pixel_link_receiver_address: For iMX8QM: o LVDS: pixel_link_receiver_address[1:0] = do_gpio_dr[7:6]  o MIPI-DSI: pixel_link_receiver_address[1:0] = do_gpio_dr[7:6] o HDMI-TX: odd_pixel_link_receiver_address[1:0] = do_gpio_dr[7:6],even_pixel_link_receiver_address[1:0] = do_gpio_dr[5:4]   For iMX8QXP: o Combo MIPI-DSI / LVDS: pixel_link0_receiver_address[1:0] = do_gpio_dr[7:6], pixel_link1_receiver_address[1:0] = do_gpio_dr[5:4] 

    Xenomai is real-time framework, which can run seamlessly side-by-side Linux as a co-kernel system, or natively over mainline Linux kernels (with or without PREEMPT-RT patch). The dual kernel nicknamed Cobalt, is a significant rework of the Xenomai 2.x system. Cobalt implements the RTDM specification for interfacing with real-time device drivers. The native linux version, an enhanced implementation of the experimental Xenomai/SOLO work, is called Mercury. In this environment, only a standalone implementation of the RTDM specification in a kernel module is required, for interfacing the RTDM-compliant device drivers with the native kernel. You can get more detailed information from Home · Wiki · xenomai / xenomai · GitLab       I have ported xenomai 3.1 to i.MX Yocto 4.19.35-1.1.0, and currently support ARMv7 and tested on imx6ulevk/imx6ull14x14evk/imx6qpsabresd/imx6dlsabresd/imx6sxsabresdimx6slevk boards. I also did stress test by tool stress-ng on some boards.      You need to git clone https://gitee.com/zxd2021-imx/xenomai-arm.git, and git checkout Linux-4.19.35-1.1.0. (which inlcudes all patches and bb file) and add the following variable in conf/local.conf before build xenomai by command bitake xenomai.  XENOMAI_KERNEL_MODE = "cobalt"  PREFERRED_VERSION_linux-imx = "4.19-${XENOMAI_KERNEL_MODE}" IMAGE_INSTALL_append += " xenomai" DISTRO_FEATURES_remove = "optee" or XENOMAI_KERNEL_MODE = "mercury" PREFERRED_VERSION_linux-imx = "4.19-${XENOMAI_KERNEL_MODE}" IMAGE_INSTALL_append += " xenomai" DISTRO_FEATURES_remove = "optee" If XENOMAI_KERNEL_MODE = "cobalt", you can build dual kernel version. And If XENOMAI_KERNEL_MODE = "mercury", it is single kernel with PREEMPT-RT patch. The following is test result by the command (/usr/xenomai/demo/cyclictest -p 50 -t 5 -m -n -i 1000 😞 //Mecury on 6ULL with stress-ng --cpu 4 --io 2 --vm 1 --vm-bytes 128M --metrics-brief policy: fifo: loadavg: 6.08 2.17 0.81 8/101 534 T: 0 (  530) P:99 I:1000 C:  74474 Min:     23 Act:  235 Avg:   77 Max:    8278 T: 1 (  531) P:99 I:1500 C:  49482 Min:     24 Act:   32 Avg:   56 Max:    8277 T: 2 (  532) P:99 I:2000 C:  36805 Min:     24 Act:   38 Avg:   79 Max:    8170 T: 3 (  533) P:99 I:2500 C:  29333 Min:     25 Act:   41 Avg:   54 Max:    7069 T: 4 (  534) P:99 I:3000 C:  24344 Min:     24 Act:   51 Avg:   60 Max:    7193   //Cobalt on 6ULL with stress-ng --cpu 4 --io 2 --vm 1 --vm-bytes 128M --metrics-brief policy: fifo: loadavg: 7.02 6.50 4.01 8/100 660 T: 0 (  652) P:50 I:1000 C: 560348 Min:      1 Act:   10 Avg:   15 Max:      71 T: 1 (  653) P:50 I:1500 C: 373556 Min:      1 Act:    9 Avg:   17 Max:      78 T: 2 (  654) P:50 I:2000 C: 280157 Min:      2 Act:   14 Avg:   20 Max:      64 T: 3 (  655) P:50 I:2500 C: 224120 Min:      1 Act:   12 Avg:   15 Max:      57 T: 4 (  656) P:50 I:3000 C: 186765 Min:      1 Act:   31 Avg:   19 Max:      53   //Cobalt on 6qp with stress-ng --cpu 4 --io 2 --vm 1 --vm-bytes 512M --metrics-brief policy: fifo: loadavg: 8.11 7.44 4.45 8/156 1057 T: 0 (  917) P:50 I:1000 C: 686106 Min:      0 Act:    3 Avg:    5 Max:      53 T: 1 (  918) P:50 I:1500 C: 457395 Min:      0 Act:    3 Avg:    5 Max:      49 T: 2 (  919) P:50 I:2000 C: 342866 Min:      0 Act:    2 Avg:    4 Max:      43 T: 3 (  920) P:50 I:2500 C: 274425 Min:      0 Act:    3 Avg:    5 Max:      58 T: 4 (  921) P:50 I:3000 C: 228682 Min:      0 Act:    2 Avg:    6 Max:      46   //Cobalt on 6dl with stress-ng --cpu 2 --io 2 --vm 1 --vm-bytes 256M --metrics-brief policy: fifo: loadavg: 3.35 4.15 2.47 1/122 850 T: 0 (  729) P:50 I:1000 C: 608088 Min:      0 Act:    1 Avg:    3 Max:      34 T: 1 (  730) P:50 I:1500 C: 405389 Min:      0 Act:    0 Avg:    4 Max:      38 T: 2 (  731) P:50 I:2000 C: 304039 Min:      0 Act:    1 Avg:    4 Max:      45 T: 3 (  732) P:50 I:2500 C: 243225 Min:      0 Act:    0 Avg:    4 Max:      49 T: 4 (  733) P:50 I:3000 C: 202683 Min:      0 Act:    0 Avg:    5 Max:      38   //Cobalt on 6SX stress-ng --cpu 4 --io 2 --vm 1 --vm-bytes 512M  --metrics-brief policy: fifo: loadavg: 7.51 7.19 6.66 8/123 670 T: 0 (  598) P:50 I:1000 C:2314339 Min:      0 Act:    3 Avg:    8 Max:      60 T: 1 (  599) P:50 I:1500 C:1542873 Min:      0 Act:   15 Avg:    8 Max:      72 T: 2 (  600) P:50 I:2000 C:1157152 Min:      0 Act:    4 Avg:    9 Max:      55 T: 3 (  601) P:50 I:2500 C: 925721 Min:      0 Act:    5 Avg:    9 Max:      57 T: 4 (  602) P:50 I:3000 C: 771434 Min:      0 Act:    6 Avg:    6 Max:      41   //Cobalt on 6Solo lite stress-ng --cpu 4 --io 2 --vm 1 --vm-bytes 512M  --metrics-brief policy: fifo: loadavg: 7.01 7.04 6.93 8/104 598 T: 0 (  571) P:50 I:1000 C:3639967 Min:      0 Act:    9 Avg:    7 Max:      60 T: 1 (  572) P:50 I:1500 C:2426642 Min:      0 Act:    9 Avg:   11 Max:      66 T: 2 (  573) P:50 I:2000 C:1819980 Min:      0 Act:   11 Avg:   10 Max:      57 T: 3 (  574) P:50 I:2500 C:1455983 Min:      0 Act:   12 Avg:   10 Max:      56 T: 4 (  575) P:50 I:3000 C:1213316 Min:      0 Act:    7 Avg:    9 Max:      43   //Cobalt on 7d with stress-ng --cpu 2 --io 2 --vm 1 --vm-bytes 256M --metrics-brief policy: fifo: loadavg: 5.03 5.11 5.15 6/107 683 T: 0 (  626) P:50 I:1000 C:6842938 Min:      0 Act:    1 Avg:    2 Max:      63 T: 1 (  627) P:50 I:1500 C:4561953 Min:      0 Act:    4 Avg:    2 Max:      66 T: 2 (  628) P:50 I:2000 C:3421461 Min:      0 Act:    0 Avg:    2 Max:      69 T: 3 (  629) P:50 I:2500 C:2737166 Min:      0 Act:    3 Avg:    2 Max:      71 T: 4 (  630) P:50 I:3000 C:2280969 Min:      0 Act:    2 Avg:    1 Max:      33   //////////////////////////////////////// Update for Yocto L5.10.52 2.1.0  /////////////////////////////////////////////////////////// New release for Yocto release L5.10.52 2.1.0. You need to git clone https://gitee.com/zxd2021-imx/xenomai-arm and git checkout xenomai-5.10.52-2.1.0. Updating: 1, Upgrade Xenomai to v3.2 2, Enable Dovetail instead of ipipe. Copy xenomai-arm to <Yocto folder>/sources/meta-imx/meta-bsp/recipes-kernel, and add the following variable in conf/local.conf before build Image with xenomai enable by command bitake imx-image-multimedia. XENOMAI_KERNEL_MODE = "cobalt" IMAGE_INSTALL_append += " xenomai" or XENOMAI_KERNEL_MODE = "mercury" IMAGE_INSTALL_append += " xenomai" Notice: If XENOMAI_KERNEL_MODE = "cobalt", you can build dual kernel version. And If XENOMAI_KERNEL_MODE = "mercury", it is single kernel with PREEMPT-RT patch. //////////////////////////////////////// Update for Yocto L5.15.71 2.2.0  /////////////////////////////////////////////////////////// New release for Yocto release L5.15.71 2.2.0. You need to git clone https://gitee.com/zxd2021-imx/xenomai-arm and git checkout xenomai-5.15.71-2.2.0. Updating: 1, Upgrade Xenomai to v3.2.2 Copy xenomai-arm to <Yocto folder>/sources/meta-imx/meta-bsp/recipes-kernel, and add the following variable in conf/local.conf before build Image with xenomai enable by command bitake imx-image-multimedia. XENOMAI_KERNEL_MODE = "cobalt" IMAGE_INSTALL:append += " xenomai" or XENOMAI_KERNEL_MODE = "mercury" IMAGE_INSTALL:append += " xenomai" Notice: If XENOMAI_KERNEL_MODE = "cobalt", you can build dual kernel version. And If XENOMAI_KERNEL_MODE = "mercury", it is single kernel with PREEMPT-RT patch.   ///////// Later update for Later Yocto release, please refer to the following community post //////////// 移植实时Linux方案Xenomai到i.MX ARM64平台 (Enable real-time Linux Xenomai on i.MX ARM64 Platform)   

[中文翻译版] 见附件   原文链接： https://community.nxp.com/docs/DOC-343102 

Copy Redboot binary to /tftpboot. In this case: redboot.bin Load binary file to i.MX ram memory: RedBoot> load -v -r -b 0x100000 /tftpboot/redboot.bin Run the loaded image RedBoot> run 0x100000 Enable NOR, NAND or MMC flash media for Redboot. In this case, NAND is beeing used. RedBoot> factive nand Update Redboot in the flash with currently running image RedBoot> romupdate Copy redboot binary to /tftpboot. In this case: redboot.bin Load binary file to i.MX ram memory: RedBoot> load -v -r -b 0x100000 /tftpboot/redboot.bin Run the loaded image RedBoot> run 0x100000 Enable NOR, NAND, or MMC flash media for Redboot. In this case, NAND is being used. RedBoot> factive nand Update Redboot in the flash with currently running image RedBoot> romupdate

Features Additional Information Features The i.MX31 PDK, with Smart Speed™ technology, is a completely integrated hardware and software solution that simplifies product development so you can focus on your critical differentiation needed for market success. Reduce development time, and design products that have power to spare, even when running multiple applications simultaneously. Receive stellar image and graphic performance in a system design that dramatically reduces power consumption. The i.MX31 PDK provides: Modular hardware enabling multiple connectivity technologies Optimized development software for Linux®, Windows® CE 5.0 and Windows Embedded CE 6.0 operating systems Out-of-box experience, complete with demonstration software and performance data Maximum performance and power savings Complete "Design. Debug. Demo." capability as simple as 1,2,3 i.MX31 Applications Processor Module i.MX31 Applications Processor - ARM11™ 128 MB DDR SDRAM 256 MB NAND FLASH Power Management (PMIC MC13783) + Power Circuitry Audio HS USB PHY Touch Controller Connector Debug Module (Software Development) Debug Ethernet Port Debug Serial Port JTAG Reset, Interrupt, Boot Switches Debug LEDs CodeTest Interface Power Source Current/Power Monitoring Personality Module (Demo-ready) Acceleromater MMA7450L (Freescale) User I/O Connectivity (FM, 802.11, Bluetooth, USB OTG, USB HS) Button 2.7"TFT Display 2MP Camera Module SDcard, ATA HDD External Connectors (dock, headphones, TV out, GPS) Microphone Speaker Additional Information i.MX31 PDK Contents If you are new to i.MX31PDK development we suggest checking out:Not authorized to view the specified document 1673 To flash BootLoader: i.MX31 PDK Board Flashing Miscellaneous Tutorials Blink i.MX 31PDK LEDs Using U-Boot i.MX31 Testing RNGA I.MX31 Testing TvOut I.MX31 Using CLKO

Here is a summary including kdump + crash porting in i.mx, the tool is very useful in crash issue. Overview What is Kdump + Crash Preconditions kernel kexec Step actions Crash tool Analysis of use cases   Besides that, I summary the dumper tools including kdump and pstore, the respective patches shown as below: Kdump: Customer can apply below patch in config and cmdline, which has been confirmed on linux os. As memory is very precious to android, so kdump is not worth adopting. Config:        First kernel:       CONFIG_KEXEC=y       CONFIG_SYSFS=y       CONFIG_DEBUG_INFO=Y        Capture kernel:       CONFIG_CRASH_DUMP=y       CONFIG_PROC_VMCORE=y Cmdline:        crashkernel=512M Pstore: Customer can apply below patch in config and dts showing as below, which has been confirmed on linux and android os. Config: CONFIG_PSTORE_PMSG=y Dts: +               ramoops@0x91f00000 { +                       compatible = "ramoops"; +                       reg = <0 0x91f00000 0 0x00100000>; +                       record-size     = <0x00020000>; +                       console-size    = <0x00020000>; +                       ftrace-size     = <0x00020000>; +                       pmsg-size       = <0x00020000>; +               }; +                 decoder_boot: decoder-boot@84000000 {                         reg = <0 0x84000000 0 0x2000000>;                         no-map;   Reproduce steps Reboot the system by enter below command:       $ reboot Check the related log by enter below command:       $ ls /sys/fs/pstore/          console-ramoops-0 pmsg-ramoops-0

This is ov5645 driver and tested with i.MX6 L3.0.35 BSP .  It is modified based on ov5640.c. P.S. The power down function for OV5645 is different from the OV5640. So modify the function in your_board.c like this: static void mx6q_mipi_powerdown(int powerdown) {     if (!powerdown)         gpio_set_value(MIPI_PWDN, 1);     else         gpio_set_value(MIPI_PWDN, 0);     msleep(5); }

For Fedora Users: Open a terminal as root Edit tftp file -> #gedit /etc/xinetd.d/tftp Add these lines: service tftp     {   socket_type = dgram   protocol = udp   wait = yes   user = root   server = /usr/sbin/in.tftpd   server_args = /tftpboot   disable = no   per_source = 100 2   flags = IPv4 } Restart the service: # /etc/init.d/xinetd restart OR # service xinetd restart

Q: What is latency figures for the VPU to decode Device: i.MX6Q OS: Linux Resolution:                         1920x1080(HD) Frame rate:                        30 FPS Function:                             Overlay messages. Input/Output:                   8 bit / YUV 4:2:0 / NAL stream Profile level:                      4.1. Constrained Baseline. I and P frames support. A: It depend on the syntax in H.264, includes num_reorder_frames,max_dec_frame_buffering,num_ref_frames,MaxDpbSize,etc. for start latency: it cover vpu driver loading, allocate buffers, init, decoding the first frame less than 100ms on iMX6/Linux. This document was generated from the following discussion: VPU Latency i.MX6