09:18 <paulbarker> marex: Are the mmc patches you sent to the list the same as https://source.denx.de/u-boot/custodians/u-boot-sh/-/commits/sdhi-scc-fixes?

09:18 <paulbarker> I have an RZ/G2L set up on my desk so I can give them a quick test

09:26 <marex> paulbarker: yeah

09:26 <marex> paulbarker: that would be much appreciated

11:55 <xypron> calebccff: Could you, please, test if https://patchwork.ozlabs.org/project/uboot/patch/20240220115423.61417-1-heinrich.schuchardt@canonical.com/ solves your problem with ext2.

12:03 <calebccff> xypron: that looks about right! thanks for submitting a patch for this. I'll try it out :D

12:45 <LeSpocky> hello

12:45 <LeSpocky> just came across a strange thing with redundand env in (raw) nand

12:48 <LeSpocky> but I think I have to look deeper into it again

14:59 <xypron> calebccff: What fs_get_size() should return for a dictionary is hard to say. I guess the remaining issue needs to be fixed in EFI. Which service is systemd calling to get the size?

15:10 <xypron> calebccff: In EDK II EFI_FILE_PROTOCOL.GetInfo() via FatGetDirEntInfo() returns the size of the directory on disk.

15:13 <calebccff> xypron: here's a trace of efi_file.c running systemd-boot with your patch https://p.calebs.dev/930378@raw

15:13 <calebccff> the issue is that efi_file_getinfo() fails with status 7

15:13 <calebccff> i followed this and i'm pretty sure it's because that codepath winds up calling info fs_get_size() which fails because ext4fs doesn't expect you to do that for a directory

15:27 <xypron> calebccff: Yes it does. We probably need to adjust the ext4_open code to return the size of directories. What I really find tricky in the ext4 code is that ext4fs_iterate_dir() makes allocations and we should take care to avoid memory leaks. The current code for ext4fs_exists() and ext4fs_size() is leaky.

15:28 <xypron> calebccff: We should also test soft links.