Current-Users archive

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index][Old Index]

namei heads up

I just committed another round of namei hacking, this time a 41-part
patchbomb affecting almost exclusively the internals. This probably
would break everything if it was going to break anything, but with
code of this type(*) one never knows. So if problems crop up, let me
know. (The usual ways.)

(*) Today's example: it turns out that if you set LOCKPARENT but not
LOCKLEAF, and you do a lookup that returns the same vnode as the
parent (ndp->ni_dvp) and the leaf (ndp->ni_vp)... then it does return
you the parent (without LOCKPARENT ndp->ni_dvp is NULL) but it doesn't
lock it! And the code in vfs_syscalls.c "knows" this so it can't be
changed easily or safely...

David A. Holland

Home | Main Index | Thread Index | Old Index