BeagleY-AI yocto meta layers

I would like to known if there is a doc, guide, meta-layers … to use yoctoi based image on the beagley-ai ?

Regards

Today that will be fun, you might be able to modify TI’s latest 9.x sdk and update the git repo to our kernel and u-boot changes… (Ti’s 9.x uses same v6.x kernel and u-boot major version)

Regards,

How different is the BB factory kernel than the j722 ti stuff, I was looking into using that with the .dtb of the beagley.

Just swap in our u-boot and device tree and it should boot

That is what my thoughts were on that, thank you.
My plans are to install a yocto image with my dev-meta-layer onto NVMe so the remote builds should be smoking fast.

You’ll need our pcie clock enable patch, TI is looking for another way to do it

Okay, thank you for the heads up on that.

Will NVMe run out of the box with the factory image and adapter board?

Yeap, and we have setup a user selectable (over serial) boot option, to do that.

Network boot too?

1:	microSD (production test)
2:	transfer microSD rootfs to NVMe (advanced)
3:	microSD (debug)
4:	microSD (default)
1 Like

No onboard storage, so today you need a fat 32 partition on a microSD card… but u-boot dhcp worked last I tried

1 Like

I was testing this one today… Amazon.com other than getting warm (as another fan somewhere) the poe/NVMe combo works nice

voodoo@beagley-ai4g-01:~$ uname -r ; lspci
6.1.80-ti-arm64-r55
00:00.0 PCI bridge: Texas Instruments Device b010
01:00.0 Non-Volatile memory controller: Sandisk Corp WD Black SN770 NVMe SSD (rev 01)
voodoo@beagley-ai4g-01:~$ lsblk
NAME        MAJ:MIN RM   SIZE RO TYPE MOUNTPOINTS
mmcblk1     179:0    0  29.7G  0 disk 
├─mmcblk1p1 179:1    0   256M  0 part /boot/firmware
├─mmcblk1p2 179:2    0     4G  0 part 
└─mmcblk1p3 179:3    0  25.5G  0 part 
nvme0n1     259:0    0 465.8G  0 disk 
├─nvme0n1p1 259:1    0     4G  0 part [SWAP]
└─nvme0n1p2 259:2    0 461.8G  0 part /
1 Like

Hi Robert,

I use sdk 9.2 on a sk-am62a-lp with TI yocto edgeai sdk. The idea is to adapt this on the BBY-AI.
If i understand, i just need to use your kernel and your uboot.

The kernel source is here : Robert Nelson / ti-linux-kernel-dev · GitLab

And for u-boot ?

Regards

u-boot here: Files · v2023.04-ti-09.02.00.009-BeagleY-AI-Production · BeagleBoard.org / u-boot · GitLab

Thank you !

About the kernel, can i just keep the ti linux kernel edgeAi and apply your patch ?
Then, Which machine should I declare for yocto build ?
image

J722S-evm