varigit / linux-imx Goto Github PK

Load and run the hello_world binary under u-boot.
The system sends a hello world. message and responds to data from the uart ( note its on pins 3 and 5 on the symphonny module )
Boot into linux.
The Uart stops responding.
The remoteproc state responds with "attached".

Reboot the system without the m33 running and the remoteproc services run as expected, you can load /reload the firmware and start / stop the m33 as expected.

I got this error building yocto core-image-minimal for VAR-SOM-6UL:

ERROR: linux-variscite-5.15.71+gitAUTOINC+5ebe23fd15-r0 do_fetch: Fetcher failure: Fetch command export PSEUDO_DISABLED=1; export PATH="/workdir/build_wayland/tmp/sysroots-uninative/x86_64-linux/usr/bin:/workdir/sources/poky/scripts:/workdir/build_wayland/tmp/work/imx6ul_var_dart-fslc-linux-gnueabi/linux-variscite/5.15.71+gitAUTOINC+5ebe23fd15-r0/recipe-sysroot-native/usr/bin/arm-fslc-linux-gnueabi:/workdir/build_wayland/tmp/work/imx6ul_var_dart-fslc-linux-gnueabi/linux-variscite/5.15.71+gitAUTOINC+5ebe23fd15-r0/recipe-sysroot/usr/bin/crossscripts:/workdir/build_wayland/tmp/work/imx6ul_var_dart-fslc-linux-gnueabi/linux-variscite/5.15.71+gitAUTOINC+5ebe23fd15-r0/recipe-sysroot-native/usr/sbin:/workdir/build_wayland/tmp/work/imx6ul_var_dart-fslc-linux-gnueabi/linux-variscite/5.15.71+gitAUTOINC+5ebe23fd15-r0/recipe-sysroot-native/usr/bin:/workdir/build_wayland/tmp/work/imx6ul_var_dart-fslc-linux-gnueabi/linux-variscite/5.15.71+gitAUTOINC+5ebe23fd15-r0/recipe-sysroot-native/sbin:/workdir/build_wayland/tmp/work/imx6ul_var_dart-fslc-linux-gnueabi/linux-variscite/5.15.71+gitAUTOINC+5ebe23fd15-r0/recipe-sysroot-native/bin:/workdir/sources/poky/bitbake/bin:/workdir/build_wayland/tmp/hosttools"; export HOME="/home/vari"; LANG=C git -c core.fsyncobjectfiles=0 -c gc.autoDetach=false -c core.pager=cat clone --bare --mirror https://github.com/varigit/linux-imx /workdir/downloads//git2/github.com.varigit.linux-imx --progress failed with exit code 128, no output
ERROR: linux-variscite-5.15.71+gitAUTOINC+5ebe23fd15-r0 do_fetch: Bitbake Fetcher Error: FetchError('Unable to fetch URL from any source.', 'git://github.com/varigit/linux-imx;protocol=https;branch=lf-5.15.y_var01')
ERROR: Logfile of failure stored in: /workdir/build_wayland/tmp/work/imx6ul_var_dart-fslc-linux-gnueabi/linux-variscite/5.15.71+gitAUTOINC+5ebe23fd15-r0/temp/log.do_fetch.147
ERROR: Task (/workdir/sources/meta-variscite-bsp/recipes-kernel/linux/linux-variscite_5.15.bb:do_fetch) failed with exit code '1'

Trying to clone the linux-imx repository it fails if DEPTH is set greater than 1.

It seems not to be a network problem.

I'm tiring compile kernel for VAR-SOM-MX8 : NXP i.MX8 from commit 02be52b and got error related to irqchip. My devconfig is attached

|kernel-source/drivers/irqchip/irq-imx-irqsteer.c: In function 'imx_irqsteer_chans_enable':
| kernel-source/drivers/irqchip/irq-imx-irqsteer.c:194:34: error: 'irqsteer_data' undeclared (first use in this function)
|   194 |         ret = clk_prepare_enable(irqsteer_data->ipg_clk);
|       |                                  ^~~~~~~~~~~~~
| kernel-source/drivers/irqchip/irq-imx-irqsteer.c:194:34: note: each undeclared identifier is reported only once for each function it appears in
| In file included from kernel-source/include/linux/device.h:25,
|                  from kernel-source/include/linux/of_platform.h:9,
|                  from kernel-source/drivers/irqchip/irq-imx-irqsteer.c:15:
| kernel-source/drivers/irqchip/irq-imx-irqsteer.c: At top level:
|kernel-source/drivers/irqchip/irq-imx-irqsteer.c:386:28: error: 'imx_irqsteer_runtime_suspend' undeclared here (not in a function); did you mean 'pm_generic_runt
ime_suspend'?
|   386 |         SET_RUNTIME_PM_OPS(imx_irqsteer_runtime_suspend,
|       |                            ^~~~~~~~~~~~~~~~~~~~~~~~~~~~
| /home/zdeneks/eaton-platform/build-kirkstone/tmp/work-shared/eaton-xh303/kernel-source/include/linux/pm.h:341:28: note: in definition of macro 'SET_RUNTIME_PM_OPS'
|   341 |         .runtime_suspend = suspend_fn, \
|       |                            ^~~~~~~~~~
|   CC      fs/nfs/proc.o
| kernel-source/drivers/irqchip/irq-imx-irqsteer.c:387:28: error: 'imx_irqsteer_runtime_resume' undeclared here (not in a function)
|   387 |                            imx_irqsteer_runtime_resume, NULL)
|       |                            ^~~~~~~~~~~~~~~~~~~~~~~~~~~
| kernel-source/include/linux/pm.h:342:27: note: in definition of macro 'SET_RUNTIME_PM_OPS'
|   342 |         .runtime_resume = resume_fn, \
|       |                           ^~~~~~~~~
| make[2]: *** [kernel-source/scripts/Makefile.build:289: drivers/irqchip/irq-imx-irqsteer.o] Error 1

defconfig.txt

Hi All,

Could you guys see the botologo appears while booting the variscite devices?

I'm using var_dvk_mx6,

I have some drivers I'd like to pull in that are in v5.18 but the latest kernel here is v5.15. Any plans to update to something later in the 5 series?

Hay,

i patched the SN65DSI83 Driver in a newer Kernel Version (4.14.98_2.0.0_ga) from NXP for the 8MM.
The driver looks working without Issues but i get no picture. All Registers looking fine.

I am using the Linux kernel provided by variscite 5.4-2.1.x-imx_var01.
On which NXP Linux Kernel version this release is based on?

Setup:
SoC : imx7d
SoM : imx7-var-som
kernel version:5.4.127
keyutils version :1.6.1

Please find the attached kernel configuration(kernel_configuration.txt).

I am following AN12714 given by NXP[latest version] for encryption storage using CAAM secure keys.

I configure kernel as mentioned in application note and caam-keygen device registered correctly.

**root@imx7-var-som:~# dmesg | grep caam
[ 0.504972] caam 30900000.caam: Entropy delay = 3200
[ 0.564417] caam 30900000.caam: Instantiated RNG4 SH0
[ 0.623846] caam 30900000.caam: Instantiated RNG4 SH1
[ 0.623870] caam 30900000.caam: device ID = 0x0a16030000000000 (Era 8)
kernel_config.txt
kernel_config.txt

[ 0.623879] caam 30900000.caam: job rings = 3, qi = 0
[ 0.635929] caam algorithms registered in /proc/crypto
[ 0.637547] caam 30900000.caam: caam pkc algorithms registered in /proc/crypto
[ 0.637567] caam 30900000.caam: registering rng-caam
[ 0.683685] Device caam-keygen registered**

I have successfully generated keys using caam-keygen.

I successfully added a key based on the generated black key in the key retention service using "keyctl"

Following is the output:

root@imx7-var-som:/# cat /data/caam/randomkey | keyctl padd logon logkey: @s
765699361

but when I check using keyctl then same id doesn't appear

Following is output:

root@imx7-var-som:/# keyctl list @s
1 key in keyring:
24465593: --alswrv 0 65534 keyring: _uid.0

I also followed the old version of AN12714 but getting following error:
add_key: No such device

Please guide me to resolve the issue.

Regards,
dpatil

Hej
Building the "lf-5.10.y_var04" branch at current head (32517d8) for the DART-6UL SOM fails for two reasons. The "lf-5.10.y_var04" branch is the suggested version for DART-6UL by the wiki (https://variwiki.com/index.php?title=Yocto_Build_Linux&release=mx6ul-yocto-hardknott-5.10.72_2.2.1-v1.0).
Looks like both of the reason were present and solved on branch "5.4-2.1.x-imx_var01".
Note: using custom gcc 11.x toolchain and custom defconfig derived from the provided "imx_v7_var_defconfig".

Reasons for failure:

• irq-imx-irqsteer
  Solved by 391a7c6
• busfreq-imx
  Solved by the following commits:
  • 4f49200
  • e85ce31

Applying these 3 commits enables the build to finish and yields a bootable kernel.

Hello Variscite developers,

[The Problem Description]

According the the iMX8QM reference manual the main PLL for MIPI DSI subsystem is 864 MHz and is not supposed to be internally configurable (fixed clock, 24 MHz x 36). It is explicitly stated in the i.MX8QX RM on the page 447:
8.11.3.2 Subsystem PLLs ==>> Table 8-17. Subsystem PLLs.

The mipi_pll_div2_clk is derived from the main PLL divided by 2, which is actually 432 MHz as expected:

a72_clk 0 0 0 1296000000 0 0 50000
a53_clk 0 0 0 1200000000 0 0 50000
mipi_pll_div2_clk 2 2 0 432000000 0 0 50000
mipi0_dsi_rx_esc_clk 0 0 0 72000000 0 0 50000
mipi0_dsi_tx_esc_clk 1 1 0 18000000 0 0 50000
mipi0_dsi_phy_clk 1 1 0 27000000 0 0 50000
mipi1_dsi_phy_clk 0 0 0 27000000 0 0 50000

My problem is that mipi_pll_div2_clk (432 MHz clock) is not able to drive the non standard display on the MIPI DSI 0 channel.

[Proposed Solution Description]

So, I got an idea how to fix this problem using a .dtsi include file. Using arch/arm64/boot/dts/freescale/imx8qm-ss-mipi.dtsi , I decided to change mipi_pll_div2_clk from 432 MHz to 864 MHz.

And here is the actual patch, what is changed:

diff --git a/arch/arm64/boot/dts/freescale/imx8qm-ss-mipi.dtsi b/arch/arm64/boot/dts/freescale/imx8qm-ss-mipi.dtsi
index ba166abbab3c..537ffbc2422a 100644
--- a/arch/arm64/boot/dts/freescale/imx8qm-ss-mipi.dtsi
+++ b/arch/arm64/boot/dts/freescale/imx8qm-ss-mipi.dtsi
@@ -15,7 +15,8 @@
mipi_pll_div2_clk: clock-mipi-div2-pll {
compatible = "fixed-clock";
#clock-cells = <0>;
[del] clock-frequency = <432000000>;
[add] // clock-frequency = <432000000>;
[add] clock-frequency = <864000000>;
clock-output-names = "mipi_pll_div2_clk";
};

[The Actual Results]

I recompiled the kernel, and loaded it on the Android device I am using. The formal results are shown here:

a72_clk 0 0 0 1296000000 0 0 50000
a53_clk 0 0 0 1200000000 0 0 50000
mipi_pll_div2_clk 0 0 0 864000000 0 0 50000
mipi0_dsi_rx_esc_clk 0 0 0 72000000 0 0 50000
mipi0_dsi_tx_esc_clk 0 0 0 18000000 0 0 50000
mipi0_dsi_phy_clk 0 0 0 24000000 0 0 50000
mipi1_dsi_phy_clk 0 0 0 27000000 0 0 50000

But when I tested the display, I discovered that the actual mipi_pll_div2_clk did NOT change, and it's real value stayed 432 MHz.

As my best guess, the clk_summary command should reflect the real clocks, not a formal changes, introduced from the .dtsi and .dts files?!

This phenomenon comes from the fact that the SCU unit provides the i.MX8QX clocks, and my kludge/assumption is that SCU does NOT communicate clocks back to the i.MX8QX's main unit, should it?

Is there any method to change the mipi_pll_div2_clk on the i.MX8QX silicons, actually NOT reprogramming the SCU unit (even if this is still possible)?

Zoran

Hello, I'm trying to build a driver for external Wi-Fi from this repo https://github.com/gglluukk/rtl8188eus. It successfully build on Ubuntu 20.04 and Raspberry Pi OS lite (32-bit). So, now I'm trying to do it with 5.4.142-g489e299ef3e9 from varigit. make process throw some error in kernel-headers which I downloaded with the command sudo apt-get install linux-headers-$(uname -r).
Could somebody suggest or give a guide on how to correctly build something related to kernel headers for mx8mm?

The VAR-SOM-MX7 base device tree (imx7d-var-som.dtsi) contains entries that are specific to the VAR-STK-MX7 / VAR-DVK-MX7 evaluation kits. This makes it harder than necessary to adept the device tree since it's pretty much required to patch the base device tree.

Notable devkit specific entries:

• gpio-keys
• the ov5640_mipi
• the ft5x06_ts
• The pcie_phy assuming an external PCI clock

Some of these devices are documented as being on the evaluation kit on the variscite wiki but for other devices this is not very clear.

Having a generic dtsi device tree for the VAR_SOM-MX7 and a seperate dts device tree for the evaluation kits would make adapting the device tree for custom carriers a lot easier.