cube@cubidou.net (Quentin Garnier) writes:

>No.  Calls to {c,b}devsw_lookup are done at each call (for a good reason
>anyway;  a LKM can disappear).

Sounds like missing LKM reference counts.

>That means that in order for case 2 to still work, we'll require that
>one step of the COMPAT_NETBSD32 setup is to populate /emul/netbsd32/dev
>with i386's MAKEDEV.  Through the compat rewrite of path, i386 binaries
>will get the correct path and access the correct device.

Is there a reason to keep the minor numbering for /i386 other than
100% compatibility? To me the amd64 way sounds more logical.

Path rewriting will make i386 binaries work, but then you create
device aliases with possible different ownership/permissions.
How do you keep them in sync without creating security issues?