(resending without the html)
On Wed, Apr 16, 2014 at 9:36 AM, John Stultz <john.stultz@xxxxxxxxxx> wrote:
>
> On 04/16/2014 07:39 AM, Joe Perches wrote:
> > On Wed, 2014-04-16 at 23:27 +0900, Seunghun Lee wrote:
> >> This patch fix checkpatch.pl warnings and errors.
> > []
> >> diff --git a/drivers/staging/android/uapi/binder.h b/drivers/staging/android/uapi/binder.h
> > []
> >> @@ -169,7 +169,7 @@ struct binder_ptr_cookie {
> >> struct binder_handle_cookie {
> >> __u32 handle;
> >> binder_uintptr_t cookie;
> >> -} __attribute__((packed));
> >> +} __packed;
> > If this .h file is meant to be a user-space #include,
> > then it should not use the kernel specific __packed
> > but keep the __attribute__((packed))
>
> Agreed.
>
> > It does use __u32 though and that's generally
> > kernel specific.
>
> Hmm. Theres a ton of __u32 usage in include/uapi/* as well as typedefs
> for it too.
> include/uapi/asm-generic/int-l64.h:typedef unsigned int __u32;
> include/uapi/asm-generic/int-ll64.h:typedef unsigned int __u32;
>
> > John? Does any of these binder uapi files need a
> > bit more sorting out?
> I suspect this is ok, but Cc'ing Colin to give him a heads up, as it
> would probably cause trouble w/ their libc headers first.
__u32 is explicitly for use in userspace headers so it doesn't
conflict with userspace definitions of u32, uint32_t, etc. It gets
set through #include <linux/types.h> when a copy of include/uapi is in
the include path.
>From Documentation/CodingStyle:
(e) Types safe for use in userspace.
In certain structures which are visible to userspace, we cannot
require C99 types and cannot use the 'u32' form above. Thus, we
use __u32 and similar types in all structures which are shared
with userspace.
_______________________________________________
devel mailing list
devel@xxxxxxxxxxxxxxxxxxxxxx
http://driverdev.linuxdriverproject.org/mailman/listinfo/driverdev-devel
