On Fri, Nov 08, 2019 at 10:24:52PM +0000, Alistair Francis wrote:
On Thu, 2019-11-07 at 17:49 -0500, Tom Rini wrote:
On Tue, Nov 05, 2019 at 11:30:45AM -0800, Alistair Francis wrote:
To fix these failures when building with musl: include/env.h:166:1: error: unknown type name 'ulong'; did you mean 'long'? ensure that ulong is defined.
Signed-off-by: Alistair Francis alistair.francis@wdc.com
include/env.h | 2 ++ 1 file changed, 2 insertions(+)
diff --git a/include/env.h b/include/env.h index b72239f6a5..5ca49a3456 100644 --- a/include/env.h +++ b/include/env.h @@ -13,6 +13,8 @@ #include <stdbool.h> #include <linux/types.h>
+typedef unsigned long ulong;
struct environment_s;
/* Value for environment validity */
What exactly is the case you're hitting this under? <linux/types.h> is where we have that typedef. I'd almost rather spell out unsigned
That only applies if __KERNEL_STRICT_NAMES isn't defined (I'm guessing it somehow is for musl).
https://github.com/u-boot/u-boot/blob/master/include/linux/types.h#L8
OK, so lets first understand for sure how we're getting in to this position.
long in a few more places than add a typedef here, if we can't solve this via correcting some include order or some other underlying problem.
Is there somehwere else you would prefer it?
If, once we know why the existing typedef's are used, and it's a feature not a bug then we should probably clean function prototypes to be more portably correct rather than adding a typedef for this case.