https://bugs.winehq.org/show_bug.cgi?id=47832
--- Comment #12 from qsniyg qsniyg@mail.com --- (In reply to Joe Davison from comment #10)
This was fixed in 4.21 but it appears there's been a regression since. The last version of Wine that Mod Organizer's USVFS is functional in, as far as I can tell, is 5.0.
In an initial attempt to figure out the cause I looked at kernelbase/file.c where most of kernel32/file.c is now located and found some discrepancies. I made a frankensteined file.c containing Myah's patch but it had no effect.
At any rate, I think this bug should be reopened as there definitely seems to be a regression.
The patch is still there, the important bit for this issue is that the `info->data_len < info->data_size - max_entry_size` check has been removed, which is still the case.
If USVFS fails, then something else is the cause. There could be a number of different causes. An issue that was exhibited in one of the later staging releases was that wine would call internal versions of the syscalls, rather than the exported thunks, which meant that it would bypass USVFS's hooked calls. This has since been fixed, but it's not unlikely that this might be exhibited somewhere else by accident.