-
Notifications
You must be signed in to change notification settings - Fork 237
OE4T Meeting Notes 2025‐01‐10
Dan Walkes edited this page Jan 10, 2025
·
1 revision
8
- UEFI variable write/Reclaim() issue https://github.com/NVIDIA/edk2-nvidia/issues/114
- Have seen general problems with boot on EMMC with Xavier which may be related
- Also wondering if https://github.com/OE4T/meta-tegra/pull/1734 could be related.
- Dan will follow up with Kurt’s patch on the issue.
- Will revisit next month to decide whether to put an optional non-default workaround in meta-tegra or continue to wait for NVIDIA to respond.
- Ilies to test on AGX Orin with test script.
- initrd-flash issue https://github.com/OE4T/meta-tegra/issues/1794
- Possible issue at https://forums.developer.nvidia.com/t/mb1-bl-crash-when-rebooting-to-rcm-from-b-slot/309503/17 when using reboot recovery.
- Workaround
busybox devmem 0x0C3903CC 32 0x00000000 && reboot forced-recovery
- May be good to start a wiki page with a list of common mistakes, helpful suggestions, could ultimately link this from initrd-flash failures.
- meta-ros issues enabling cuda and opencv
- Should be handled through MACHINE_ARCH
- Other packages using opencv are broken because they don’t have a reference to cuda.
- Needed to have a conditional include for cuda
inherit ${@bb.utils.contains("MACHINE_FEATURES", "cuda", "cuda", "", d)}
- May be possible to have cmake files generated by opencv automatically include cuda for calling layers rather than relying on changing recipe.