AM62P-Q1: AM62P secure boot

Hello,

Donghui Peng said:

After setting keyver 1(HS-FS => HS_SE), I can no longer boot from eMMC or SD card.

If you have not flashed the HSSE images then it is expected that you could not boot from the eMMC or SD card.

The BOOTMODE9 pin of my board is pulled low during power-up, preventing it from booting normally from the SD card. I need to reset MCU_PORZ (after powerup) to enable normal SD card booting.

I have already do this:

    echo 0>/sys/block/mmcblk0boot0/force_ro

    dd if=/dev/zero of=/dev/mmcblk0boot0 bs=512 count=8192

    sync

Donghui Peng said:

dd if=/dev/zero of=/dev/mmcblk0boot0 bs=512 count=8192

I am not sure what is the intention of writing zeros to the eMMC boot0 partition.

Are you trying to boot from the SD card or the eMMC?

I use SBL  falcon mode，priamary boot from emmc boot0; backup bootmode: SD card.

erase firmware in emmc boot0, and It will boot from   backup bootmode.

Donghui Peng said:

erase firmware in emmc boot0, and It will boot from   backup bootmode.

So, it means you are trying to boot from the SD card? If yes, have you tried booting from the SD card primary bootmode instead of the backup bootmode? Additionally, are you using the correct key for signing the images?

1) I have burned sbl_emm_linux_stage1.release.hs.image and sbl_emm_linux_stage2.release.appimage.hs to the eMMC, but I mistakenly used the EMC_SBL_ENABLED=yes option for compilation.So it do not work.

./gen_keywr_cert.sh -t tifek/SR_10/ti_fek_public.pem -s keys/v15/smpk.pem --keycnt 1 

2) I only have 2 AM62P boards available and want to try booting from an SD card with sbl_linux_sd to verify secure boot.

However, sbl_linux_sd works properly with SDK 10, but there is no log (no secure boot) when I use mcu_plus_sdk_am62px_11_01_01_08.

Donghui Peng said:

However, sbl_linux_sd works properly with SDK 10, but there is no log (no secure boot) when I use mcu_plus_sdk_am62px_11_01_01_08.

Could you please share the working and the non-working `sbl_linux_sd` images?

• [BOOTLOADER_PROFILE] Boot Media       : SD Card 
  [BOOTLOADER_PROFILE] Boot Image Size  : 185 KB 
  [BOOTLOADER_PROFILE] Cores present    : 
  wkup-r5f0-0
  a530-0
  [BOOTLOADER PROFILE] System_init                      :      24835us 
  [BOOTLOADER PROFILE] Drivers_open                     :      18468us 
  [BOOTLOADER PROFILE] Board_driversOpen                :          0us 
  [BOOTLOADER PROFILE] Sciclient Get Version            :         27us 
  [BOOTLOADER PROFILE] App_loadImages(CSL_CORE_ID_WKUP_R5FSS0_0) :     290074us 
  [BOOTLOADER PROFILE] App_loadLinuxImages              :    1836770us 
  [BOOTLOADER_PROFILE] SBL Total Time Taken             :    2170177us 
  
  Image loading done, switching to application ...
  Starting linux and RTOS/Baremetal applications
  Some tests have failed!!
  Hello World!

Files cannot be uploaded to the forum;

above are the logs :

Donghui Peng said:

However, sbl_linux_sd works properly with SDK 10, but there is no log (no secure boot) when I use mcu_plus_sdk_am62px_11_01_01_08.

Here you reported that you do not see any logs. This suggests that the SBL_SD_LNIUX is itself isn't booting.

Donghui Peng said:

above are the logs :

But these logs show that the SBL_SD_LINUX is at least booting.

Could you please confirm what is the exact issue? Is it that the SBL_SD_LINUX is always booting successfully but it says "Some tests have failed" with 11.1 SDK version but runs successfully all the way with SDK v10?

UART0:
<sbl_sd_linux> 1 App_OpenloadableImage /sd0/mcu_r5_image 
<sbl_sd_linux> 1 App_OpenloadableImage /sd0/wkup_r5_image 
<sbl_sd_linux> 2 Bootloader_open bootParams 
<sbl_sd_linux> 3 App_loadImages bootImageInfo return 0 
<sbl_sd_linux> 4 App_OpenloadableImage /sd0/linux_a53_image 
<sbl_sd_linux> 5 Bootloader_open bootParamsLinux 
<sbl_sd_linux> 6 App_loadImages bootImageInfoLinux return 0 
Image loading done, switching to application ...
Starting linux and RTOS/Baremetal applications
<sbl_sNOTIdCE:  BL_31:l v2.10i.0(renlease)u:v2x.10.>0-367-g00f1e c6b787-d irtyA
uOTpICE: p BL_31: Buirlt : 1u6:0n9:05L, Feib  9 20n24
 xCpu bootHandleLinux return 0 
<sbl_sd_linux> 8 App_runCpus bootHandle return -1 
Some tests have failed!!
[    0.000000] Booting Linux on physical CPU 0x0000000000 [0x410fd034]
[    0.000000] Linux version 6.6.32+ (peng@peng-VMware) (aarch64-oe-linux-gcc (GCC) 13.3.0, GNU ld (GNU Binutils) 2.42.0.20240716) #21 SMP PREEMPT Mon Dec  8 11:21:09 CST 2025
[    0.000000] earlycon: ns16550a0 at MMIO32 0x0000000002800000 (options '')
[    0.000000] printk: bootconsole [ns16550a0] enabled
....
....
^_^  hello 2025
 
(O(∩_∩)O) login:


WAKEUP_UART0:
<wkup_r5> main_thread
 Hello World!
 
 

SDK V10:

sbl_sd_linux:      OK

wkup_r5_image: OK

mcu_r5_iamge:  NG (no log)

linux_a53_image: OK

SDK V11:

sbl_sd_linux: NG (no log)

I have attempted copying the source code sbl_sd_linux from SDK10 to SDK11 for compilation  yet no log has been generated so far.

Thanks for the clarification.

Donghui Peng said:

sbl_sd_linux: NG (no log)

Let us the look at this issue first. Have you made sure you are using the correct key for signing in the SDK11? And the encryption is disable? If you could share the `sbl_sd_linux` image from the SDK10 (working) and SDK11 (non-working) then I could verify if they are signed by the same keys.

1）I am using other board in HS-FS mode (not HS-SE mode).

2) encryption is disable4010.tiboot3.zip

Hello,

The SBL_SD_LINUX not working is a known issue as also listed in the release notes

https://software-dl.ti.com/mcu-plus-sdk/esd/AM62PX/11_01_01_08/exports/docs/api_guide_am62px/RELEASE_NOTES_11_01_01_PAGE.html#autotoc_md148

Please apply the following patch to resolve the issue

github.com/.../d1fc5806e10

Thanks,   apply the patch above,   SBL_SD_LINUX in SDK 11.01 is ok now .

DEVICE_TYPE=HS

sbl_sd_linux.release.hs.tiimage  is OK  now when I use the board in HS mode.(use keywriter)

So, I guess the SBL_SD_LINUX is all booting fine.

Donghui Peng said:

mcu_r5_iamge:  NG (no log)

Is this issue still open?

This occurs in SDK V10 SBL_SD_LINUX, but it doesn’t matter,

we are not using mcu_r5_image , I will test it later.