[rt2x00-users] [PATCHv2 6/6] rt2x00: Use an enum instead of u16 for the rate_mode TX descriptor field

[Ivo Van Doorn]

Hi,

>> On Thu, Feb 24, 2011 at 10:10 AM, Helmut Schaa
>> <helmut.schaa at googlemail.com> wrote:
>> > Am Donnerstag, 24. Februar 2011 schrieb Johannes Stezenbach:
>> >> On Wed, Feb 23, 2011 at 11:46:43PM +0100, Helmut Schaa wrote:
>> >> > Am Mittwoch, 23. Februar 2011 schrieb Helmut Schaa:
>> >> > > Am Mittwoch, 23. Februar 2011 schrieb Johannes Stezenbach:
>> >> > > > On Wed, Feb 23, 2011 at 01:46:23PM +0100, Helmut Schaa wrote:
>> >> > > > > This makes the code less error-prone and might allow the compiler to use
>> >> > > > > a u8 for the rate_mode.
>> >> > > > ...
>> >> > > > > -     u16 rate_mode;
>> >> > > > > +     enum rate_modulation rate_mode;
>> >> > > >
>> >> > > > I think on most platforms an enum is the same size as an int.
>> >> > >
>> >> > > Unless you specify __attribute__ ((packed)) in the enum definition.
>> >> >
>> >> > JFI calling gcc with "-fshort-enums" will also make use of smaller
>> >> > types for enums.
>> >>
>> >> This is not used on purpose because it creates binary compatibility
>> >> issues in userspace ABI (various drivers uses enum in ioctls).
>> >
>> > Ok, thanks for pointing this out.
>> >
>> > Ivo, if you're applying this could you please remove the misleading comment
>> > "might allow the compiler to use a u8 for the rate_mode" from the commit log?
>>
>> Sure no problem, the patch didn't apply cleanly anyway :P
>
> Urgs, sorry bout that. Did the other patches apply cleanly?

Oh this was patch 6... sorry, haven't tried that one yet.
Patch 1 applied cleanly, but patch 2 failed. I hadn't spend any
aditional time on it yet,
to apply them manually. So I will do that later today (then I will
also check if the other
patches apply cleanly or not).

Ivo