Re: [RFC PATCH] libata: PIO via bounce buffer

[Alan Cox <alan@lxorguk.ukuu.org.uk>]

> Inside ata_bounce_pio_xfer(), get the buffer for your CPU.  If NULL or 
> too small, allocate new buffer, otherwise re-use existing buffer.
> 
> This method makes the common case _not_ allocate anything, which should 
> be obviously more efficient.

Obviously ?

kmalloc will usually return us a cache hot buffer very very fast. A
private allocator will return us an almost certainly cache cold buffer
slightly faster.

I'm not sure its that simple.

Alan