bootstd regression from distro_bootcmd for removable EFI boot and weird return codes for fs

[Da Xue]

Hi,

I switch to bootstd and I am experiencing an issue with bootstd not
detected EFI bootable file in /EFI/BOOT/BOOTAA64.EFI in certain image,
buildroot specifically. The same setup works fine with distro_bootcmd.
I have attached the logs. I have tried the following:

1) Directory and filename case upper and lower
2) FAT32 and FAT16 filesystem
3) Changing the size of the filesystem

I noticed some particular weirdness with u-boot's return code for
functions dealing with fs_exists that might be related.

1) FAT/FAT32 driver uses fs_ls_generic. When you do `ls mmc 1
/filename`, it returns 1 if the file exists and returns 1 when the
file doesn't exist.
2) This issue follows fs_exists. Is this a bug or intended behavior?

Best,
Da
-------------- next part --------------
A non-text attachment was scrubbed...
Name: bootstd.log
Type: application/octet-stream
Size: 21638 bytes
Desc: not available
URL: <https://lists.denx.de/pipermail/u-boot/attachments/20230724/2d76451a/attachment.obj>