PROCESSOR-SDK-AM62A: Default SDK 11. WIC image does not boot with 128 GB uSD card

Hi Evyatar,

Please refer the below query

Link:- https://e2e.ti.com/support/processors-group/processors---internal/f/processors---internal-forum/1539027/sk-am62a-lp-sk-am62a-lp-does-not-boot-with-pre-built-linux-images?tisearch=e2e-sitesearch&keymatch=Default%252520SDK%25252011.%252520WIC%252520image%252520does%252520not%252520boot%252520with%252520128%252520GB%252520uSD%252520card#

Hello Dilna,

Thank you for replying, but I do not have access to the internal forum.

Hi Evyatar,

OK. No problem i will share the link information here.

1. Could you partition the SD card properly and try again. hope you are using 'create-sdcard.sh' script to prepare the SD card and flash the prebuild image

2. Could you please try different brand SD card with different size. 

Yes, we normally use the script in the SDK to partition the cards.

The same SanDisk Ultra 128 GB microSD cards worked with the bootloader until SDK 9.2. There was a SW regression starting from SDK 10. There is a closed thread about it here.

Since then we have been using 32 GB cards, but I wanted to re-open the issue since in the future such small cards may be harder to acquire.

Is there a known brand that TI tested with 128+ GB cards?

Hello,

Could you please share the Amazon link of the SD card for me to procure the same and try it out locally on the EVM?

Thanks!

Sure:
https://www.amazon.com/SanDisk-128GB-Ultra-microSDXC-Memory/dp/B0BDYVC5TD?th=1

Hello,

I do see the issue with this card

[12:07:53.662] U-Boot SPL 2025.01-00547-g3bcfad6cee95-dirty (Aug 19 2025 - 12:06:20 +0530)
[12:07:53.669] SYSFW ABI: 4.0 (firmware rev 0x000b '11.1.5--v11.01.05 (Fancy Rat)')
[12:07:53.741] SPL initial stack usage: 13568 bytes
[12:07:53.759] Trying to boot from MMC2
[12:07:53.762] selecting mode MMC legacy (freq : 0 MHz)
[12:07:53.789] selecting mode MMC legacy (freq : 25 MHz)
[12:07:53.805] sd card: widths [4, 1] modes [MMC legacy, SD High Speed (50MHz)]
[12:07:53.807] host: widths [4, 1] modes [MMC legacy, SD High Speed (50MHz), UHS SDR12 (25MHz), UHS SDR25 (50MHz), UHS SDR50 (100MHz), UHS DDR50 (50MHz), UHS SDR104 (208MHz)]
[12:07:53.829] trying mode SD High Speed (50MHz) width 4 (at 50 MHz)
[12:07:53.838] selecting mode SD High Speed (50MHz) (freq : 50 MHz)
[12:07:53.854] card capabilities: widths [4, 1] modes [MMC legacy, SD High Speed (50MHz)]
[12:07:53.857] host capabilities: widths [4, 1] modes [MMC legacy, SD High Speed (50MHz), UHS SDR12 (25MHz), UHS SDR25 (50MHz), UHS SDR50 (100MHz), UHS DDR50 (50MHz), UHS SDR104 (208MHz],
[12:07:53.965] Authentication passed
[12:07:53.968] Authentication passed
[12:07:53.981] Authentication passed
[12:07:54.013] Authentication passed
[12:07:54.029] Authentication passed
[12:07:54.045] Starting ATF on ARM64 core...

[12:07:54.046] NOTICE:  BL31: v2.13.0(release):v2.13.0-259-ge0c4d3903b-dirty
[12:07:54.049] NOTICE:  BL31: Built : 07:01:36, Jul  1 2025

[12:07:54.302] U-Boot SPL 2025.01-00547-g3bcfad6cee95-dirty (Aug 19 2025 - 12:06:36 +0530)
[12:07:54.309] SYSFW ABI: 4.0 (firmware rev 0x000b '11.1.5--v11.01.05 (Fancy Rat)')
[12:07:54.318] DM ABI: 3.0 (firmware ver 0x000b 'MSDK.11.01.00.16-dirty--v11.01.02' patch_ver: 2)
[12:07:54.365] Trying to boot from MMC2
[12:07:54.366] mmc@fa00000: No vqmmc supply
[12:07:54.381] selecting mode MMC legacy (freq : 0 MHz)
[12:07:54.653] selecting mode MMC legacy (freq : 25 MHz)
[12:07:54.662] sd card: widths [4, 1] modes [MMC legacy, SD High Speed (50MHz), UHS SDR12 (25MHz), UHS SDR25 (50MHz), UHS SDR50 (100MHz), UHS DDR50 (50MHz), UHS SDR104 (208MHz)]
[12:07:54.683] host: widths [4, 1] modes [MMC legacy, SD High Speed (50MHz), UHS SDR12 (25MHz), UHS SDR25 (50MHz), UHS SDR50 (100MHz), UHS DDR50 (50MHz), UHS SDR104 (208MHz)]
[12:07:54.698] trying mode UHS SDR104 (208MHz) width 4 (at 208 MHz)
[12:07:54.717] selecting mode UHS SDR104 (208MHz) (freq : 208 MHz)
[12:07:54.797] am654_sdhci mmc@fa00000: No passing ITAPDLY, return 0
[12:07:54.804] card capabilities: widths [4, 1] modes [MMC legacy, SD High Speed (50MHz), UHS SDR12 (25MHz), UHS SDR25 (50MHz), UHS SDR50 (100MHz), UHS DDR50 (50MHz), UHS SDR104 (208MHz],
[12:07:54.817] host capabilities: widths [4, 1] modes [MMC legacy, SD High Speed (50MHz), UHS SDR12 (25MHz), UHS SDR25 (50MHz), UHS SDR50 (100MHz), UHS DDR50 (50MHz), UHS SDR104 (208MHz],
[12:07:54.837] mmc_bread: Failed to set blocklen
[12:07:54.849] mmc_bread: Failed to set blocklen
[12:07:54.850] Partition 1 invalid on device 1
[12:07:54.862] spl_register_fat_device: fat register err - -1
[12:07:54.863] spl_load_image_fat: error reading image u-boot.img, err - -1
[12:07:54.867] SPL: failed to boot from all boot devices
[12:07:54.872] ### ERROR ### Please RESET the board ###

If you apply the first workaround from the following guide to enumerate the SD card in HS mode instead of the UHS (default) then it boots successfully

https://software-dl.ti.com/processor-sdk-linux/esd/AM62AX/11_01_07_05/exports/docs/linux/Foundational_Components/U-Boot/UG-Memory-K3.html#steps-for-working-around-sd-card-issues-in-u-boot

Evyatar Tamir said:

but does not boot with a 128 GB card of the same make and model (which works with SDK 9.2).

This was working OOB just because the U-Boot in the SDK 9.2 enumerates the SD card in HS mode only (default) instead of the UHS. 

Thank you for verifying!

Do you know why the 32 GB model boots correctly? I thought it was of the same speed class.

Hello,

Evyatar Tamir said:

Do you know why the 32 GB model boots correctly? I thought it was of the same speed class.

I don't have this card but not sure how it is booting for you. We already had a known issue of AM62A not booting with this 32GB card. We have now added 128GB one as well to this known issue list. The team will investigate to find the root cause of the issue.

For the time being, if a certain SD card is not working, it is recommended to apply the previously shared workarounds.

 SK-AM62A-LP: SD card compatibility issue
my new Kingston 128GB card with these specs wasn't booting up
100MB/s Read, UHS-I Speed Class, U1, V10 (64GB-128GB)

Above patch solved the problem