On Tue, 2018-11-06 at 17:34 -0800, Joe Perches wrote:
On Tue, 2018-11-06 at 15:48 -0800, Bart Van Assche wrote:
> On Mon, 2018-11-05 at 13:12 -0800, Alexander Duyck wrote:
> > One change I made in addition is I replaced the use of "bool X:1" to
define
> > the bitfield to a "u8 X:1" setup in order to resolve some checkpatch
> > warnings.
>
> Please use "bool X:1" instead of "u8 X:1". I think it was a bad
idea to make
> checkpatch complain about "bool X:1" since "bool X:1" should
only be avoided
> in structures for which alignment must be architecture-independent. For struct
> device it is fine if member alignment differs per architecture. Additionally,
> changing "bool X:1" into "u8 X:1" will reduce performance on
architectures that
> cannot do byte addressing.
I generally agree. But the checkpatch warning _could_
be useful in those cases where alignment should be
architecture-independent.
Any suggestion on how to improve the message?
It would be great if a heuristic could be developed that recognizes structs
for which the data layout must be architecture independent. If such a
heuristic could be developed it could be used to only display warn about
"bool X:n" for such structures.
Bart.