On Tue, Jul 26, 2011 at 11:41 AM, Al Viro <viro@xxxxxxxxxxxxxxxxxx> wrote:
>
> Hmm... We still have 19 spare bits in ->i_flags and at least 22 in ->i_state.
> Do we need to bother with new field at all? ->i_state is unsigned long...
Do we *need* it? No. But it's often in the next cache-line. It would
be beautiful to put the bits right next to i_mode, and basically make
a lot of the common operations just touch those four bytes for doing
the S_ISDIR tests and the "can I use generic_permission()" tests etc.
And the whole point of that is to make for a smaller cacheline
footprint (and admittedly also the few cycles to avoid the dereference
- and that you'd get from using i_state).
Linus
--
To unsubscribe from this list: send the line "unsubscribe linux-fsdevel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
