The use case from one customer we are supporting requires to run i.MX93 M33 code with a combination of DRAM and on chip TCM.  However, the examples in the official MCUXpresso SDK release support to run M33 code in TCM only. So we did some customization to achieve that. User can refer to the the attached slides and patches for the details.

Hello everyone, in this post we will control one (or more) WS2812 LEDs from a Linux UserSpace application, communicating with Cortex M33 through RPMSG.  Required materials: FRDM-IMX93 WS2812 LED/Strip Attached wsled.c (userspace application) Attached ws2812_m33_project (Source code of firmware for M33)   Environment: Linux Kernel 6.6 arm-gnu-toolchain-13.3.rel1-x86_64-arm-none-eabi SDK 24.12.00   This is the block diagram of the application:    Diagrama sin título.drawio.png     CORTEX M33 SIDE.  From WS2812  LED datasheet, we can see the timing must be precise, that is the reason we will use the Cortex M33 (Real time applications):  Manuel_Salas_3-1749596540788.png   As we can see, the timing just accept tolerance of 150ns, that is the reason is preferred to use the Cortex M33 instead of control with Linux that is in charge of other tasks.    The function that we will implement to send Zero and One is:  void WS2812B_send_bit(uint8_t bit) { if (bit) { // Send "1" RGPIO_PinWrite(WS2812B_LED_RGPIO, WS2812B_LED_RGPIO_PIN, 1); for (int i = 0; i < HIGH_ONE_CYCLES; i++) __asm__("NOP"); RGPIO_PinWrite(WS2812B_LED_RGPIO, WS2812B_LED_RGPIO_PIN, 0); for (int i = 0; i < LOW_ONE_CYCLES; i++) __asm__("NOP"); } else { // Send "0" RGPIO_PinWrite(WS2812B_LED_RGPIO, WS2812B_LED_RGPIO_PIN, 1); for (int i = 0; i < HIGH_ZERO_CYCLES; i++) __asm__("NOP"); RGPIO_PinWrite(WS2812B_LED_RGPIO, WS2812B_LED_RGPIO_PIN, 0); for (int i = 0; i < LOW_ZERO_CYCLES; i++) __asm__("NOP"); } }     Testing and measuring the toggle timing with the RGPIO driver of Cortex M33, we could find the best number of NOP instructions to adjust with the required timing according to datasheet, and the results with core running at 200MHz are:    Parameter  Number of NOP times  HIGH_ONE_CYCLES  22  (T1H: 700 ns)  LOW_ONE_CYCLES  12  (T1L: 600 ns)  LOW_ZERO_CYCLES  20  (T0L: 800 ns)  HIGH_ZERO_CYCLES  6    (T0H: 350 ns)      Zero: Manuel_Salas_4-1749596660278.png Manuel_Salas_5-1749596678175.png   One: Manuel_Salas_6-1749596709146.png Manuel_Salas_7-1749596723057.png   Taking in mind this information, we can start to develop our application. From Cortex M33, we created a little protocol, that will wait for instructions to do.    For example, the expected string to fill the RPMSG buffer is:    Index  Purpose  Example  app_buf[0]  Command ('c' or 't')  'c' = store color in buffer  't' = show colors on LEDs  app_buf[1]  Red component (0–255)  0xFF  app_buf[2]  Green component (0–255)  0x00  app_buf[3]  Blue component (0–255)  0x80  app_buf[4]  LED index (target LED position)  e.g. 0, 1, 2    Commands:    'c'  "Color store"  This command stores the RGB color value into a local buffer, targeting a specific LED (app_buf[4]).  It doesn't update the LED strip immediately, it just prepares the data.  Example of RPMSG received buffer from Cortex A (Linux):   app_buf = { 'c', 0xFF, 0x00, 0x00, 2 };  // Store red color for LED 2     't'  "Transfer"  This command sends the full color buffer to the WS2812 strip, updating the LED colors.  The RGB values in the buffer are ignored, it simply triggers an update based on previously stored data.  Example of RPMSG received buffer from Cortex A (Linux):  app_buf = { 't', 0, 0, 0, 0 };  // Transfer the buffered colors to the LED strip     You can modify the code to add your custom commands like change number of LEDs on the strip, or change the GPIO, etc. Currently, the number of WS2812 LEDs supported is defined in software by NUM_RGB_LEDS. As default is 8 LEDs.    I will attach the full code for Cortex M33 firmware, it was tested with arm-gnu-toolchain-13.3.rel1-x86_64-arm-none-eabi and SDK 2_15, SDK 2_16 and SDK 24.    In this example, the EXP_GPIO_IO24 of FRDM-IMX93 was used to control the LED Strip. Connection.png   CORTEX A55 SIDE.  For Cortex A55 (Linux side), we developed an userspace application, that will create the structure for the buffer to send to the Cortex M33 and send through RPMSG using the imx_rpmsg_tty included on the NXP BSP. You can find the userspace application attached.    As we know, the imx_rpmsg_tty creates the tty channel called ttyRPMSG31 under /dev. So, this userspace application communicates directly with the device /dev/ttyRPMSG31 to send the buffer with the required structure for Cortex M33.    EXAMPLE OF USAGE.    STEP 1. Compile firmware for Cortex M33 and store in FRDM-IMX93 under /lib/firmware to use with remoteproc:   STEP 2.  Compile or cross-compile the userspace application (attached.)  To compile on the platform:  root@imx93evk:~# gcc wsled.c -o wsled root@imx93evk:~# ls wsled wsled.c Also, we can copy the wsled under /bin to practicity. root@imx93evk:~# wsled Error: -c <red> <green> <blue> is required. WS2812 Usage: wsled -c <red> <green> <blue> [-n <led_position>] Load the color of exact LED to the buffer wsled -t Shows the colors to the LEDs   STEP 3. To run the entire example: boot the board and load the firmware for cortex M33 with remoteproc: root@imx93evk:~# cd /sys/class/remoteproc/remoteproc0 root@imx93evk:/sys/class/remoteproc/remoteproc0# echo ws2812_m33.elf > firmware root@imx93evk:/sys/class/remoteproc/remoteproc0# echo start > state  Then, we will receive this from Cortex M33 terminal: RPMSG String Communication Channel FreeRTOS RTOS M33 clock is at: 200000000 Hz Nameservice sent, ready for incoming messages from Cortex-A55! WSLED   Load the imx_rpmsg_tty module in Linux to initialize RPMSG: root@imx93evk:~# modprobe imx_rpmsg_tty   Then, we will have the hello world message from Cortex A (Linux side), but with our buffer format in Cortex M33 console: CMD: h [1]:0x65 [2]:0x6c [3]:0x6c [4]: 111 LEN = 12   Finally, we are ready change colors of the LEDs: Load the buffer (LED 1 = RED, LED 2 = Green, LED 3 = Blue). root@imx93evk:~# wsled -c 0x55 0x00 0x00 -n 1 root@imx93evk:~# wsled -c 0x00 0x55 0x00 -n 2 root@imx93evk:~# wsled -c 0x00 0x00 0x55 -n 3   Show colors on LEDs. root@imx93evk:~# wsled -t   Manuel_Salas_8-1749597921154.png  

This article describe i.M93 RGMII to PHY connection, delay adding tips. i.MX93 don't support delay in both FEC and QOS port in i.MX93 side. It also provide solution on i.MX93 how to connect MAC to MAC in HW & SW. Thanks! 

this article write down the steps when help customer bring dual camera under single mipi csi with NXP switch NX3DV642, just for a reference, connect OV5640 + os08a20 with mipi csi0 via ISP, as default, mipi csi0 just has one port with one camera, this document improve how to enable os08a20 with second port under mipi csi0, refer to the document and patch as attachment, for how to enable ov5640 with ISP, pls refer to my another document

This is about the case studay from two USB related issues: #1. Plugin detection issue caused by an errata in i.MX8/8X (host mode).       Very limited unit may encounter this problem. When issue happens, there will be totally no action on USB host port when a debug plugged in. #2: High Speed disconnection detection issue (host mode):       This may happen on some special USB design which have complex circuit and connector design on the USB path and long USB cable. The USB enumeration might be interrupted by an un-expected disconnection event. System log shows USB recognition started but failed at several different stages during handshake. For details, please refer to the doc attached (#1/#2). A reference patch for each has also been made by David.

This document shows how to build genivi step by step, but I haven’t tested the images yet, before building the images, pls refer to the host setup and host packages according to the yocto project user’s guide, I don’t mention here again, this is for imx8mq as example, you can choose the different board name to build   Before building the genivi package, customer also can refer to the kernel and image name from: https://github.com/GENIVI/meta-ivi/tree/master   4.9.88 IMAGE   1. Create a bin folder in the home directory $ mkdir ~/bin (this step may not be needed if the bin folder already exists) $ curl https://storage.googleapis.com/git-repo-downloads/repo > ~/bin/repo $ chmod a+x ~/bin/repo    2. Add the following line to the .bashrc file to ensure that the ~/bin folder is in your PATH variable. export PATH=~/bin:$PATH    3. Yocto Project Setup $ mkdir imx-yocto-bsp $ cd imx-yocto-bsp $ repo init -u https://source.codeaurora.org/external/imx/imx-manifest -b imx-linux-rocko -m imx-4.9.88-2.0.0_genivi.xml $ repo sync   4.update Weston 3.0.0 to Weston 4.0.0 $ git clone https://git.yoctoproject.org/git/meta-freescale -b warrior   then replace the wayland directory in "imx-yocto-bsp/sources/meta-fsl-bsp-release/imx/meta-bsp/recipes-graphics/wayland" with the "meta-freescale/recipes-graphics/wayland" in cloned directory.   5.image build DISTRO=nxp-imx-genivi-wayland MACHINE=imx8mqevk source ./nxp-setup-genivi.sh -b genivi-wayland   $bitbake  pulsar-image    6.Error fix if you don’t update Weston, you should get the error message like joanxie_1-1615271019981.png The error shows required Weston >=4.0.0, but current bsp includes Weston version is 3.0.0, so you need to update the Weston to the 4.0.0 step by step $ git clone https://git.yoctoproject.org/git/meta-freescale -b warrior $ rm -rf ../sources/meta-fsl-bsp-release/imx/meta-bsp/recipes-graphics/wayland $ cp -r meta-freescale/recipes-graphics/wayland ../sources/meta-fsl-bsp-release/imx/meta-bsp/recipes-graphics/   $ bitbake -c cleansstate wayland-ivi-extension $ bitbake  wayland-ivi-extension $ bitbake  pulsar-image   4.14.95 IMAGE   1. Create a bin folder in the home directory $ mkdir ~/bin (this step may not be needed if the bin folder already exists) $ curl https://storage.googleapis.com/git-repo-downloads/repo > ~/bin/repo $ chmod a+x ~/bin/repo   2. Add the following line to the .bashrc file to ensure that the ~/bin folder is in your PATH variable. export PATH=~/bin:$PATH   3.Yocto Project Setup $ mkdir imx-yocto-bsp $ cd imx-yocto-bsp $ repo init -u https://source.codeaurora.org/external/imx/imx-manifest -b imx-linux-warrior -m imx-4.19.35-1.1.0_genivi.xml $ repo sync   4. change Weston 6.0.1 to Weston 5.0.0 $ git clone https://git.yoctoproject.org/git/meta-freescale -b zeus   then replace the wayland directory in "imx-yocto-bsp/sources/meta-fsl-bsp-release/imx/meta-bsp/recipes-graphics/wayland" with the "meta-freescale/recipes-graphics/wayland" in cloned directory.   5.image build $ DISTRO=fsl-imx-wayland MACHINE=imx8mqevk source fsl-setup-release.sh -b build-wayland   $ bitbake  meta-ivi-image    6.Error fix if you don’t change Weston, you should get the error message like joanxie_2-1615271127756.png so try to change the Weston to the 5.0.0 step by step $ git clone https://git.yoctoproject.org/git/meta-freescale -b zeus $ rm -rf ../sources/meta-fsl-bsp-release/imx/meta-bsp/recipes-graphics/wayland $ cp -r meta-freescale/recipes-graphics/wayland ../sources/meta-fsl-bsp-release/imx/meta-bsp/recipes-graphics   $ bitbake -c cleansstate weston $ bitbake  weston $ bitbake  meta-ivi-image     5.4.24 IMAGE   1. Create a bin folder in the home directory $ mkdir ~/bin (this step may not be needed if the bin folder already exists) $ curl https://storage.googleapis.com/git-repo-downloads/repo > ~/bin/repo $ chmod a+x ~/bin/repo   2. Add the following line to the .bashrc file to ensure that the ~/bin folder is in your PATH variable. export PATH=~/bin:$PATH   3.Yocto Project Setup $ mkdir imx-yocto-bsp $ cd imx-yocto-bsp $ repo init -u https://source.codeaurora.org/external/imx/imx-manifest -b imx-linux-zeus -m imx-5.4.24-2.1.0_genivi.xml $ repo sync   4. image build $ DISTRO=fsl-imx-wayland MACHINE=imx8mqevk source imx-setup-release.sh -b build-wayland   $ bitbake  meta-ivi-image    

Follow the PyeIQ Introduction Guide on eIQ Community: PyeIQ v1.0 for BSP 5.4.3_2.0.0 Release  Follow an introduction video for PyeIQ:

      Minicom is a terminal emulation that can access a remote serial console enabling the configuration of RedBoot or the flash file system of the board.   Configuring       Reach the configuration by typing CTRL-A Z       Press key Z after releasing CTRL and A!       Configure Minicom to work with i.MX by following the procedures below.     Set the Serial Port       At the screen configuration, type O, choosing cOnfigure Minicom       In menu, choose Serial Port Setup       Below, the configuration option:   +-----------------------------------------------------------------------+   | A - Serial Device  : /dev/ttyS0                             |   | B - Lockfile Location  : /var/lock                          |   | C - Callin Program  :                                          |   | D - Callout Program  :                                        |   | E - Bps/Par/Bits  : 115200 8N1                           |   | F - Hardware Flow Control : No                           |   | G - Software Flow Control : No                           |   |                                                                        |   | Change which setting?                                      |   +-----------------------------------------------------------------------+   Type the letter of option to enable the modification.   Remember to choose the right Serial Device

Overview Resources Download Ubuntu 12.04.01 Download i.MX28EVK BSP and Documents Ubuntu Host Setup Host Package Update Ubuntu Configuration PDF Sudo Priviledges Default Shell CCACHE Directory Layout Extract SDK and Documents Install BSP Sources Ubuntu Software Packages for LTIB Patching LTIB Create SD Card Using Ubuntu Host Media Booting Selection Cable Connections   Overview Freescale's i.M28EVK development kit provides a platform for running software and evaluating features of the i.MX28 processor. This document provides the details for running the Linux Board Support Package (BSP) on the Ubuntu 12.04 64-bit Precise Pangolin Host on an Intel/AMD architecture computer. The 32-bit host is not covered in this document and does have different configuration steps than described here.   An Ubuntu Linux host is used to cross-compile the BSP creating ARM images. The BSP provides a build system named Linux Target Image Builder, (LTIB),  the GNU tool suite for compiling and debugging, U-Boot boot loader, Linux kernel, and a root file system. Resources i.M28EVK- i.MX28 Evaluation Kit Web Page MCIMX28EVKJ Product Summary Page- i.MX28 Download Collateral L2.6.36_MX28_SDK_10.12_Source- BSP Source Download Linux documentation - i.MX28EVK Documentation Ubuntu 12.04.1 LTS (Precise Pangolin)- Ubuntu 12.04 Release Download Ubuntu 12.04.01 A dedicated computer running Ubuntu or a Virtual Machine, (VMware or VirtualBox), can be used for running the Host Ubuntu software. The Ubuntu image is available for downloaded from the Ubuntu site: Ubuntu 12.04.1 LTS (Precise Pangolin).   This Ubuntu host ISO was used with the md5 checksum: ubuntu-12.04.1-desktop-amd64.iso  06472ddf11382c8da1f32e9487435c3d   One way to acquire the ISO is to use zsync to download: zsync http://releases.ubuntu.com/12.04/ubuntu-12.04.1-desktop-amd64.iso.zsync  Once downloaded, installing the ISO is user preference - either a dedicated Linux PC or in a Virtual Machine.   Download i.MX28EVK BSP and Documents The BSP download is from this site L2.6.36_MX28_SDK_10.12_Source and the documents from Linux documentation that requires a free registration to specify login credentials,   436e0b8e1c7976c657d530a45f9dbd0c L2.6.35_10.12.01_SDK_source_bundle.tar.gz de0274320a17c1e989d1ef5c088973e2 L2.6.35_10.12.01_SDK_docs.tar.gz   Ubuntu Host Setup Ubuntu login credentials of User: user Password: user are used for this documents. Host Package Update Once logged in to the Ubuntu host, the existing packages are brought up to date to the latest version before installing the BSP. The Ubuntu package manager used is apt-get. $ sudo apt-get update $ sudo apt-get upgrade  01. Check all installed packages for new revisions 02. all newer packages found are installed.   Addtional packages are required for the ltib build system. Ubuntu Configuration PDF evince is the default pdf reader, another option is zathura. $ sudo apt-get install zathura Sudo Priviledges LTIB requires super user priviledges for some operations. To enable a visudo entry is added to the sudo'ers file. For more information run 'man visudo'.   $ sudo visudo  The first word, user, is the login account 'user' This can be changed to whatever login you used, or if you have groups configured you can provide a group that developers are in - refer to the man page for sudo for details. Add this line:   user ALL =NOPASSWD: /usr/bin/rpm/ /opt/freescale/ltib/usr/bin/rpm   Default Shell Ubuntu uses the default shell 'dash'. This however causes failures on bash scripting which is part of the ltib system. Change the default shell from 'dash' to 'bash'   $ sudo update-alternatives --install /bin/sh sh /bin/bash 1  CCACHE ccache provides a fast C/C++ compiler cache which is supported in the ltib system. To configure once the ccache package has been installed: $ sudo apt-get install ccache $ ccache -M 50M $ ccache -c  02. Set the cache limit to 50 Meg 03. Clear the cache folder   Directory Layout The following directory structure is used: /home/user/freescale/imx28/ |-- archive |-- L2.6.35_10.12.01_ER_source |-- L2.6.35_10.12.01_SDK_docs |-- L2.6.35_10.12.01_SDK_scripts |-- ltib |-- ubuntu-imx28-ltib-patch   The archive directory is where the BSP and documents are stored; command to create the directory: $ mkdir -p ~/freescale/imx28/archive   Extract SDK and Documents The following instructions were used to extract the contents of the Software Development Kit:   $ cd ~/freescale/imx28/archive $ tar -zxf L2.6.35_10.12.01_SDK_source_bundle.tar.gz -C ..    01. Change into the directory containing the tar ball that is compressed. 02. Extract the contents into the directory above (-C ..) the current directory -z unzip -x extract -f L2.6.35_10.12.01_SDK_source_bundle.tar.gz   $ tar -zxf L2.6.35_10.12.01_SDK_docs.tar.gz  01. Extract the contents into the directory above (-C ..) the current directory     -z unzip     -x extract     -f L2.6.35_10.12.01_SDK_docs.tar.gz this file The contents of both tar files are now in the directory /home/user/freescale/imx28. Install BSP Sources After extracting the content from the L2.6.35_10.12.01_SDK_source_bundle.tar.gz the file L2.6.35_10.12.01_SDK.source.tar.gz contains all the sources and the build system. Extract the contents and install. This will create the ltib directory which is the build system. $ tar -zxf L2.6.35_10.12.01_SDK_source.tar.gz $ cd L2.6.35_10.12.01_ER_source $ ./install  Read the license information and accept by entering YES. An installation directory is then asked for, providing:  .. which is the parent directory. The installation script copies the packages and will inform you that 'Installation complete, your ltib installation has been placed in ../ltib, to complete the installation: cd .../ltib ./ltib  HOWEVER before doing this, there are packages and patches that need to be applied to run ltib on Ubuntu 12.04.01. Ubuntu Software Packages for LTIB The following packages are required. The script pkg-setup.sh attached below has these packages which can be downloaded and executed to install. $ sh pkg-setup.sh  sudo apt-get -y install gettext libgtk2.0-dev rpm bison m4 libfreetype6-dev sudo apt-get -y install libdbus-glib-1-dev liborbit2-dev intltool sudo apt-get -y install ccache zlib1g zlib1g-dev gcc g++ libtool sudo apt-get -y install uuid-dev liblzo2-dev tcl wget libncurses5-dev sudo apt-get -y install libncursesw5-dev lib32z1 libglib2.0-dev xsltproc sudo apt-get -y install ia32-libs libc6-dev-i386 The file pkg2-setup.sh contains optional packages for development. To install, download and execute: $ sh pkg2-setup.sh Please refer to the document ltib_build_host_setup.pdf for more information on host setup. Patching LTIB The location of files from the glibc-devel and zlib Ubuntu 12.04 packages has changed from 9.0.4 Ubuntu which the original ltib was released against. To update ltib operation the following patches are implemented from the directory ~/freescale/imx28/ltib 1. The file ltib is changed at line 2387 adding the '-v' option to the rpm call OLD:     system('rpm --force-debian 2>/dev/null') == 0? NEW:     system('rpm -v --force-debian 2>/dev/null') == 0? 2. The file bin/Ltibutils.pm is updated to support glibc-devel and zlib.   glibc-devel update: Line 563 add check for /usr/lib32/libm.so 'glibc-devel' => sub {-f 'usr/lib/libm.so' || -f '/usr/lib64/libz.so' || -f '/usr/lib32/libm.so'},   zlibc update: Line 584 add /lib/x86_64-linux-gnu/libz.so* zlib => sub{my @f = (glob('/usr/lib/libz.so*'),               glob('/lib/x86_64-linux-gnu/libz.so*'),               glob('/lib/libz.so*'),   The above patches are also in the attachment 0001-patches-for-12.04-ubuntu.patch.   LTIB packages also need adjustments to correctly build on Ubuntu. The tar file below, ubuntu-imx28-ltib-patch.tgz contains all the updates. Download and extract the contents at the same directory level as your ltib source directory. $ tar -zxf ubuntu-imx28-ltib-patch.tgz ├── ltib ├── ubuntu-imx28-ltib-patch └── ubuntu-imx28-ltib-patch.tgz Change directories to ubuntu-imx28-ltib-patch and then run the install-patches.sh script. $ cd ubuntu-imx28-ltib-patch $ ./install-patches.sh   The following packages are updated: lkc mtd-utils mux_server sparse Create SD Card Using Ubuntu Host The tar file L2.6.35_10.12.01_SDK_scripts.tar.gz contains scripts for writing the images from the ltib build to a SD card. Extract the content, copy the scripts to the ltib directory, and update the mk_mx28_sd script to work with the updated fdisk command.   $ tar -zxf L2.6.35_10.12.01_SDK_scripts.tar.gz $ cd L2.6.35_10.12.01_SDK_scripts $ cp mk_hdr.sh ~/freescale/imx28/ltib $ cp mk_mx28_sd ~/freescale/imx28/ltib $ cd ~/freescale/imx28/ltib  Edit mk_mx28_sd script and add the 'u' at line 177 then the o command after. This changes cylinders to sectors.   OLD: echo "o n   NEW: echo "u o n   Once updated to create the SD card which is at /dev/sdb: $ ./mk_mx28_sd /dev/sdb  NOTE: if mounted automatically, you need to unmount for the script to work $ sudo umount /dev/sdb*      Media Booting Selection The i.MX28EVK has a boot option to execute from the SD Card in Slot 0 which is located on the bottom of the EVK. On the top of the EVK there are switches that are read during the start up process to determine what boot media to use. The SD Card in slot 0 is used for this example which requires the settings: B3/DIP1 B2/DIP2 B1/DIP3 B0/DIP4 1 0 0 1 Refer to the user guide, i.MX28_Linux_BSP_UG.pdf section 3.2.1. Boot Modes for all options. The user guide is found in the Linux documentation bundle documentation.  Refer to the next section for a picture showing the boot switch location and the SD Card Slot 0 location. Cable Connections A computer serial port is connected to the i.MX28EVK serial port. The communication setting is 115200 baud, 8 data bits, No parity, and 1 stop bit. There is NO flow control set for this port. This is typically shown as 115200, 8N1. The power supply is connected  

I've recently done some digging into fstrim feature on sabresd-6q. Here list it below detailed: Background:      When Nexus 7 first came out, many users felt it was speedy and high-performance. But after months of installing applications and using the tablet, things began slowing down.This was a friction point that many hoped would be fixed in the new Nexus 7 (2013) model, which it was. There’s even more to the story though, it turns out Google has fixed that storage I/O aging problem on all Nexus devices with the Android 4.3 update which comes with a little-hyped feature that could be huge for performance in the future – TRIM support via fstrim. Why slowing down？      Remember that deleting a file in software isn't actually communicated to solid state storage (whether SSD or eMMC). The space is freed up from the user's perspective, but the eMMC controller in this case still treats the pages in NAND as having valid data. Let's say you copy a 3GB movie to your internal storage, watch the movie and later delete it. You'd have 3GB free to re-use, but until you re-write those blocks the eMMC controller would treat all 3GB as valid data. There's a data structure used by the eMMC controller that tracks mapping logical locations to physical locations in NAND. I won't go into great detail here but the more complex that mapping becomes, and the more locations that have to be tracked, the slower internal NAND management works. Why SSD has such issue? Root cause is "Flash memory is divided into blocks, which is further divided in pages. The minimum write unit is a page, but the smallest erase unit is a block" which will cause fragments as time goes. Why fstrim?      Fstrim is one mechanism of linux filesystem which derives from TRIM. TRIM is already introduced since linux2.6.33. It is the name of a command that the operating system can send to tell the SSD which blocks are free in the filesystem.The SSD uses this information to internally defragment the blocks and keep free pages available to be written quickly and efficiently. How framework triggers fstrim in android4.3?      Android4.3 introduces a new service IdleMaintenanceService to manage when and how trigger fstrim in right time. Once system satisfies the following conditions, it will send out one "ACTION_IDLE_MAINTENANCE_START" intent which the MountService listens for. Last time  "ACTION_IDLE_MAINTENANCE_START" intent has been sent before 24 hours. The device is either off-charger with 80% battery or on-charger with 30% battery The device can go idle state. User don't use it for 71 minutes. Test fstrim on Imx.6      In order to wait little time to trigger fstrim, I tried to change the following variables not to wait 71 minutes and 24 hours. After changing, it will trigger one time fstrim two minutes after system goes into idle state every 30 minutes. --- a/services/java/com/android/server/IdleMaintenanceService.java +++ b/services/java/com/android/server/IdleMaintenanceService.java -    private static final long MIN_IDLE_MAINTENANCE_INTERVAL_MILLIS = 24 * 60 * 60 * 1000; // 1 day +    private static final long MIN_IDLE_MAINTENANCE_INTERVAL_MILLIS = 30 * 60 * 1000; // 30 minute -    private static final long MIN_USER_INACTIVITY_IDLE_MAINTENANCE_START = 71 * 60 * 1000; // 71 min +    private static final long MIN_USER_INACTIVITY_IDLE_MAINTENANCE_START = 2 * 60 * 1000; // 2 min -    private static final long MAX_IDLE_MAINTENANCE_DURATION = 71 * 60 * 1000; // 71 min +    private static final long MAX_IDLE_MAINTENANCE_DURATION = 2 * 60 * 1000; // 2 min Firstly I  run Quadrant apk to see my I/O performance at beginning, Its score is 3544. During this 30 minutes, you can try to mess your system. Try your best to install most apks. Run monkey test in background. Then I  run Quadrant again to check my I/O performance, its score is only a little decline 3538. After that I run "logcat | busybox grep -i fstrim" in console to wait for the action of fstrim, here is my logcat: 130|root@sabresd_6dq:/ # logcat -d | busybox grep -i fstrim I/fstrim  ( 2344): Starting fstrim work... I/fstrim  ( 2344): Invoking FITRIM ioctl on /data I/fstrim  ( 2344): Trimmed 6367305728 bytes on /data I/fstrim  ( 2344): Invoking FITRIM ioctl on /cache I/fstrim  ( 2344): Trimmed 526355456 bytes on /cache I/fstrim  ( 2344): Finished fstrim work. OK, I checked the I/O information again, its score is restored to 3546: Conclusion      Good, fstrim takes effect. So on imx.6 using android4.3, fstrim will also work automatically. We will also enjoy good performance in I/O memory.Great!

This article will describe one suggestion for one issue that UART continuously generate RX interrupts and receive 0xFF even when Rx line is continuously high in some cases on imx6 series. Below I will explain with imx6DL. Some settings are just to make it easier to reproduce. BSP version: L5.4.70-2.3.0 Board HW: MCIMX6DL-SDB When issue happen Config imx6DL UART3 as the serial port to 1200 baud, 8-N-1 format. Keep the RX Line high. Make the RX line low and keep it for a short time (360 usec-370 usec).   At this condition, you will find that the UART will continuously generate RX interrupts and show receiving 0xFF even you make the RX line return to be high. Why issue happen The low time is not in the correct range and out of our spec. In the imx6DL AEC document, there is one chapter named UART Receiver like blow hongting_dong_0-1630284809892.png   If using 1200 band, that means one valid bit time is 833 usec. And there is a definition that “tolerate 1/(16 x Fbaud_rate) tolerance in each bit”. That’s means in the case of 1200 baud. A range of valid bit is 781 to 885 usec. But is reproducing, the Low level time is 360 usec. This time out of range will make UART state machine to be confused. How to fix Actually, the best way is following our spec. If there is such an unknown situation in the customer’s environment, then the following method could be regarded as a suggestion to fix the issue meet by the customer. The interrupt handler will check USR1[AWAKE]. 2    If AWAKE is asserted, clear it and proceed as usual (assume we have valid data), else, check if USR1[AGTIM] is asserted. 3    if AGTIM is asserted, clear it and proceed as usual, else perform software reset (assume we have invalid data). Checking AGTIM is for one race condition when the RX fifo has some characters (less than RXTL) but no more data is coming in. When following this procedure, the UART will perform a software reset when a block interrupt occurs. Notes: From customer report, error could be cleared if a valid start-bit is detected on the RX line. This needs to be verified by the customer themselves. Test code has been included in attachment.   Besst Regards

Question: 1)      Is WDOG2 somehow accessible from customers code in normal mode? Is it only used within the trust zone to protect against DOS attacks from normal world?) 2)      Is there any mechanism preventing changing of the watchdog registers like being able to write them only during a period after Reset? If yes, which registers are protected? Answer: WDOG1 and WDOG2 are identical the only difference is how the signals are connected. There is nothing presenting someone from using both WDOGs for non-secure purposes. There is no time window for WDOG access but there are some write-once only bits in the register, which can be written only once after reset, after that all following writes will be ignored. They are clearly described in the RM. The bits are: WDZST, WDBG, WDW, WDE, WDT, WIE, WICT, and PDE.

Structures to be modified The main table that needs to be modified is in C:\WINCE600\PLATFORM\COMMON\SRC\SOC\COMMON_FSL_V2_PDK1_9\NAND\INC\NANDTYPES.h. Add the Nand specific information to the following structure. typedef struct _NandChipInfo { FlashInfo fi; //@<<info> FlashInfo structure BYTE NANDCode[NANDID_LENGTH];//@<<info> NAND full ID BYTE NumBlockCycles; //@<<info> flash erase address cycle BYTE ChipAddrCycleNum; //@<<info> flash access address cycle BYTE DataWidth; //@<<info> 8/16 bits data width BYTE BBMarkNum; //@<<info> MAX_MARK_NUM = 4 BYTE BBMarkPage[MAX_MARK_NUM];//@<<info> MAX_MARK_NUM = 4 BYTE StatusBusyBit; //@<<info> interleave mode support BYTE StatusErrorBit; //@<<info> interleave mode support WORD SpareDataLength; //@<<info> spare area size BYTE CmdReadStatus; //@<<command> read status BYTE CmdRead1; //@<<command> read first 256 bytes data BYTE CmdRead2; //@<<command> read last 256 bytes data BYTE CmdReadId; //@<<command> read device ID BYTE CmdReset; //@<<command> reset nand flash BYTE CmdWrite1; //@<<command> sequence data input BYTE CmdWrite2; //@<<command> page program BYTE CmdErase1; //@<<command> block erase BYTE CmdErase2; //@<<command> block erase NANDTiming timings; //@<<info> NAND timing parameters }NandChipInfo, *pNandChipInfo; All information from NANDCode to CmdErase2, can be obtained from the Nand Datasheet. The structure FlashInfo, is filled in with data obtained from the Nand Analysis sheet. Please check this link to see how to create this spreadsheet from Nand Datasheets. typedef struct _FlashInfo { FLASH_TYPE flashType; DWORD dwNumBlocks; DWORD dwBytesPerBlock; WORD wSectorsPerBlock; WORD wDataBytesPerSector; }FlashInfo, *PFlashInfo; In the similar way the Nand timings calculated from the spreadsheet are add into this structure. typedef struct _NANDTiming { BYTE DataSetup; BYTE DataHold; BYTE AddressSetup; BYTE DataSample; }NANDTiming, *PNANDTiming; Remember to do a clean sysgen and build.

Background Some customers connect their imx running Android to the cloud, so reboot-bootloader and the use of fastboot oem unlock is not acceptable. This way the system is disconnected and cannot be operated remotely using adb. This article therefore proposes to execute oem unlock before the bootloader is started when the system is first run, which can be unlocked directly.   Patch diff --git a/drivers/fastboot/fb_fsl/fb_fsl_boot.c b/drivers/fastboot/fb_fsl/fb_fsl_boot.c index 86b919775a..d60cd248ee 100644 --- a/drivers/fastboot/fb_fsl/fb_fsl_boot.c +++ b/drivers/fastboot/fb_fsl/fb_fsl_boot.c @@ -529,6 +529,42 @@ bool __weak is_power_key_pressed(void) { return false; } +static void wipe_all_userdata(void) +{ + char response[FASTBOOT_RESPONSE_LEN]; + + /* Erase all user data */ + printf("Start userdata wipe process....\n"); + /* Erase /data partition */ + fastboot_wipe_data_partition(); + +#if defined (CONFIG_ANDROID_SUPPORT) || defined (CONFIG_ANDROID_AUTO_SUPPORT) + /* Erase the misc partition. */ + process_erase_mmc(FASTBOOT_PARTITION_MISC, response); +#endif + +#ifndef CONFIG_ANDROID_AB_SUPPORT + /* Erase the cache partition for legacy imx6/7 */ + process_erase_mmc(FASTBOOT_PARTITION_CACHE, response); +#endif + +#if defined(AVB_RPMB) && !defined(CONFIG_IMX_TRUSTY_OS) + printf("Start stored_rollback_index wipe process....\n"); + rbkidx_erase(); + printf("Wipe stored_rollback_index completed.\n"); +#endif + process_erase_mmc(FASTBOOT_PARTITION_METADATA, response); + printf("Wipe userdata completed.\n"); +} +void do_unlock(void) +{ + int status; + wipe_all_userdata(); + status = fastboot_set_lock_stat(FASTBOOT_UNLOCK); + if (status < 0) + return FASTBOOT_LOCK_ERROR; + printf("Unlock device\n"); +} int do_boota(struct cmd_tbl *cmdtp, int flag, int argc, char * const argv[]) { ulong addr = 0; @@ -563,6 +599,9 @@ int do_boota(struct cmd_tbl *cmdtp, int flag, int argc, char * const argv[]) { fastboot_set_lock_stat(FASTBOOT_LOCK); lock_status = FASTBOOT_LOCK; } + if (lock_status == FASTBOOT_LOCK){ + do_unlock(); + } bool allow_fail = (lock_status == FASTBOOT_UNLOCK ? true : false); avb_metric = get_timer(0);      

Added two functions in DDR_Stress_Tester V1.0.2, find the attachment for the details. 1. Check the actual boot mode settings in H/W target board 2. Add test option for bit flip and bit spread walking test

i.MX evaluation board can be a simple solution to program i.MX boards in a factory for instance. i.MX evaluation board are not for industrial usage, but you can find plenty of cheap i.MX insdustrial boards on the web. Here I am using an i.MX8QXP rev B0 MEK board and I will program an i.MX6Q SABRE SD board. The first step is to generate your image. Follow the documentation steps to generate the "validation" image. You will have to customize a little bit the local.conf file (in conf/local.conf) to have git, cmake, gcc and other missing package. edit local.conf and add the following lines at the end of the file: IMAGE_INSTALL_append = " git cmake htop packagegroup-core-buildessential xz p7zip rsync"‍‍‍‍‍ I have added rsync package in local, it can replace cp (copy) but with the --progress option you can see the copy progression. P7zip replace unzip for our images archives avaialable on nxp.com as unzip as issues with big files. then rebake your image: bitbake -k fsl-image-validation-imx‍‍‍‍‍ When it is done, go in tmp/deploy/image/<your image generated> and use uuu to program your board (I use a sd card; thus I can increase the partition esily): sudo ./uuu -b sd_all imx-boot-imx8qxpmek-sd.bin-flash fsl-image-validation-imx-imx8qxpmek.sdcard.bz2/*‍‍‍‍‍ As the rootfs can be too small, use gparted under Linux for instance to increase the size of the partition. Put the SD card and start your board. Here here the dirty part... You may know archlinux|ARM websitesite (Arch Linux ARM ), you have a lots of precompiled packages. Thus on the board you can download it, and copy the file in /usr folder (you can use it to have the latest openSSL for  instance!). Plug an ethernet cable on the board and check if it is up: ifconfig -a ifconfig eth0 up‍‍‍‍‍‍‍‍‍‍ Now you should have access to the internet. On uuu webpage you can find all the packages you need (here I am using a 4.14.98_2.0.0 Linux): mkdir missinglibs cd missinglibs wget http://mirror.archlinuxarm.org/aarch64/core/bzip2-1.0.8-2-aarch64.pkg.tar.xz wget http://mirror.archlinuxarm.org/aarch64/core/nettle-3.5.1-1-aarch64.pkg.tar.xz wget http://mirror.archlinuxarm.org/aarch64/core/libusb-1.0.22-1-aarch64.pkg.tar.xz wget http://mirror.archlinuxarm.org/aarch64/extra/libzip-1.5.2-2-aarch64.pkg.tar.xz wget http://mirror.archlinuxarm.org/aarch64/core/zlib-1:1.2.11-3-aarch64.pkg.tar.xz wget http://mirror.archlinuxarm.org/aarch64/extra/p7zip-16.02-5-aarch64.pkg.tar.xz cd ..‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍ Wait all the archives are downloaded (otherwise you'll decompress before the archive is downloaded) as wget is running in background! Now untar the archives and copy it in the rootfs (dirty): tar -xJf libzip-1.5.2-2-aarch64.pkg.tar.xz tar -xJf libusb-1.0.22-1-aarch64.pkg.tar.xz tar -xJf nettle-3.5.1-1-aarch64.pkg.tar.xz tar -xJf bzip2-1.0.8-2-aarch64.pkg.tar.xz cp zlib-1:1.2.11-3-aarch64.pkg.tar.xz zlib tar -xJf zlib tar -xJf p7zip-16.02-5-aarch64.pkg.tar.xz cd usr sudo cp -R . /usr cd ../../ ‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍ Download and compile uuu: git clone git://github.com/NXPmicro/mfgtools.git cd mfgtools/ cmake . make‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍ Download an image on nxp.com for instance. I have downloaded on the i.MX6 4.14.98_2.0.0 image and put it on a usb key. then unzip it in the uuu folder: 7z e L4.14.98_2.0.0_ga_images_MX6QPDLSOLOX.zip‍‍‍‍ As mentionned before unzip cannot hadle big files... so use 7z as me plug the i.MX6Q SABRE SD to the i.MX8X and program your i.MX6 board: ./uuu uuu.auto-imx6qsabresd‍ uuu (Universal Update Utility) for nxp imx chips -- libuuu_1.3.74-0-g64eeca1 Success 1 Failure 0 ‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍‍