Welcome to the Infineon Developer Community

I have a RPi 3b which is running a software package that hasn't had updates in many years and does not run on newer kernels. The system keeps randomly encountering a bug where the below message gets thrown and wlan0 is unresponsive until the system is rebooted (as outlined in the github thread found here). The firmware version the system currently reports is: 7.45.98.38 (r674442CY) FWID 01-e58d219f

brcmfmac: brcmf_sdio_hostmail: Unknown mailbox data content: 0x40012

That said, I am trying to compile a new driver to run on the old 4.14.79-v7+ kernel. I have followed the info on the backports github page and have not been successful as of yet (I am a complete novice here).

To summarize, I got the 4.14.79 kernel source, set variables, and compiled the kernel image just fine. When trying to build the backports package for the modules, I run into issues. The first time I tried building the kernel image, I did not specify the below variables and ran into a problem since I did not specify "CONFIG_ASYMMETRIC_KEY_TYPE" and associated variables in the config. The second time around I did use the variables and everything went smooth. However, when trying to insert the compat.ko module, I get a message about "Unknown symbol in module" and dmesg says "Unknown symbol hash_algo_name (err 0)". Now, as far as I can tell that module does not exist on my system (cat /proc/kallsyms | grep hash_algo_name and hash_info both return nothing) and I don't know how to get around it.

Ultimately, I am trying to update the Wi-Fi driver and firmware to fix the bug at the top of the post and nothing more. I want to keep the modifications to this old system to a minimum. If that means using an older driver source that doesn't need the "hash_algo_name" symbol or hash_info module, then that is fine by me. If there is a way to compile the latest driver so that it doesn't have any need for the hash_info module, that is also fine by me. Whatever it takes to get a newer driver on the system that doesn't have the issue I keep running into randomly. 

Any insight is appreciated!

Below are the commands and variables I have been using:

apt install git bc bison flex libssl-dev make
git clone --depth=1 --branch raspberrypi-kernel_1.20181112-1 https://github.com/raspberrypi/linux
cd linux
KERNEL=kernel7
make bcm2709_defconfig
# add the below lines to the .config file
	CONFIG_CFG80211=m
	CONFIG_BCMDHD=n
	CONFIG_ASYMMETRIC_KEY_TYPE=y
	CONFIG_ASYMMETRIC_PUBLIC_KEY_SUBTYPE=y
	CONFIG_X509_CERTIFICATE_PARSER=y
	CONFIG_PKCS7_MESSAGE_PARSER=y
make oldconfig
make -j4 zImage modules dtbs
make modules_install

cd ~/
git clone --depth=1 https://github.com/Infineon/ifx-backports
cd ifx-backports/v5.15.58-backports/
bash
MY_KERNEL=~/linux
make KLIB=$MY_KERNEL KLIB_BUILD=$MY_KERNEL defconfig-brcmfmac
make KLIB=$MY_KERNEL KLIB_BUILD=$MY_KERNEL modules

cp ~/ifx-backports/v5.15.58-backports/compat/compat.ko /lib/modules/4.14.79-v7+/compat.ko
insmod /lib/modules/4.14.79-v7+/compat.ko

Dear IFX,

EVAL_AUDAMP24 board has IRS20957 + 400V GaN (IGT40R070D1) and i wonder on 100V GaN connection.

Do we need any circuit change on Gate on/off once we use 100V or 600V GaN instead  of 400V GaN?

BR/Saint

Hello Infineon,

My project requires using the SMIF (Octal SPI) to transfer from external flash memory into XMC7200 internal RAM through P-DMA.

Based on the 1-D definition of P-DMA in the XMC7200, the maximum transfer size is dependent on the "X_COUNT",

Where "X_COUNT" have a maximum of 256 iterations,

Which equals a maximum of 256*1 Bytes = 256 B of data transferred every P-DMA transfer,  assuming 8-bit data element each transfer.

1. Is there a maximum sequential P-DMA transfer constraint that we should be aware through this approach (SMIF (Octal SPI) to Internal RAM through P-DMA)?

2. If I have 16kB (64 sets of 256 B) of total data to be transferred through P-DMA, is there any easy way to automate and chain this 64 sets of transfer together without the microcontroller being involved after P-DMA transfer have started?

3. Or must the microcontroller be involved after each 256B of data transferred through every P-DMA transfer?

4. If so, is the overhead time spent between each P-DMA reprogramming and triggering significant or in the microsecond order of magnitude?

HI

I would like to build a trigger control for a youtube maker project , using the TLE493D-W2B6

I have an TLE493DW2B6 shield2go board, and I connected it to the Arduino 

I installed the library TLE493D

I copied the DrillTrigger_SpeedAsNumber example

https://github.com/Infineon/TLE493D-3DMagnetic-Sensor/tree/master/examples/DrillTrigger_SpeedAsNumber

connected the sensor to the ARduino UNO R3   I2C pins (checked with the oscilloscope I can see SCL adn SDA moving correctly)

Connected the VDD pin of the board to 3V3 and the GND to GND

but all I can read are always the same number (see screenhsot below)

even if I put a magnet on the top of the sensor:

am I missing something?

thanks

Fabio

Hi,

I am using CY8C6245LQI-S3D42 and I used the DFU examples to boot the system and created the application to run on AM4 core and everything is working well. 

I wait for 5 sec in DFU to check if there is any updates or just move with the Application code.

I find that sometimes the device stops booting once in a blue moon. Now If I flash again just the boot loader, everything is fine , I need not have to write the Application, it moves from DFU in 5 sec to application.

I believe, there is some how the flash/DFU is getting corrupted.  I want to protect this scenario. Can some one share how do I protect my DFU so that it never erases . Any example's will be appreciable, I am using USB CDC transport in DFU.

Regards,

Madhav

I'm exploring the CSX Touchpad Tuning sample application using the PSoC4 MAX Pioneer kit. In the user manual for the CAPSENSE Tuner app, it says "The middleware supports simultaneous detection up to three touches for a CSX touchpad
touch, so the touch table displays all the detected touches." However, I can only get two touches to show in the CSX Touch Table on the Touchpad View in the Tuner app. I can see the third (and when appropriate, fourth) finger in the main Widget View graphic, but notice only two of the fingers have a colored "dot" in the middle representing the centroid of the touch. I suspect I need to modify the number of centroids that are being tracked and/or reported, but am not sure what parameter I need to change in the capsense_defines header file to do this.  

Also, what is the nature of the limitation of the middleware only supporting three simultaneous touches? Is there a theoretical limit of touches CAPSENSE can support? 

Hello,

I'm working on a project for a company where I need to implement Ethernet with UDP using Lwip. I noticed that there is no example on this particular subject. I only found examples of TCP echo. Nevertheless, I am attempting to write my own code for implementing a UDP server (code shown below). Then, I implement the function(udpServer_Init) in the main_cpu0 program. When I try to run the program, I can ping the microcontroller, but when I attempt to connect via Putty terminal, it says "connection refused." I am wondering if I am missing some settings in lwipopts.h or if I am overlooking something else. I would greatly appreciate it if anybody could provide me with some examples of UDP on the Aurix microcontroller or provide any help regarding this problem. I am also including important program files in the attachment.