Messages in this thread | | | Date | Sat, 11 May 2013 14:12:25 -0700 | Subject | Re: [RFC] next cycle fun: ->release() API change | From | Linus Torvalds <> |
| |
On Sat, May 11, 2013 at 2:06 PM, Al Viro <viro@zeniv.linux.org.uk> wrote: > > Less boilerplate? We used to pass inode to fput() as well, but > switched to passing file alone...
.. and that was painful.
The advantage has to be balanced against the pain it causes. I'm not seeing the advantage here as being worth it. If this kind of thing not only causes way more churn, _and_ it causes us to pick a new (worse) name just because it also forces a non-compatible ABI, I'm really doubtful.
I mean, if we had *other* reasons for the churn, and the name needed to change anyway, then maybe, but..
Linus
| |