Alex Hornung <ahornung@gmail.com> added the comment:
Was not a race.
Problem is fixed in 879172271d5105425ca7df8c599f59b035bfcb8e. It was basica=
lly=20
related to the fact that before devfs, calling make_dev twice, returned the=
=20
same device twice. As make_dev is now asynchronous, there's no way of doing=
=20
this. Instead, a warning is printed if vfs.devfs.debug >=3D 1 about a doubl=
e=20
make_dev.
While I fixed most of the drivers relying on double make_dev, it seems that=
I=20
still missed some ;)
Cheers,
Alex Hornung
----------
status: chatting -> resolved
_____________________________________________________
DragonFly issue tracker <bugs@lists.dragonflybsd.org>
<http://bugs.dragonflybsd.org/issue1485>
_____________________________________________________