>
> Looking at the header files, it looks like doing cpu_to_le32 on that variable
> and doing le32_to_cpu will actually reverse the order twice, on big endian
> systems, thus technically all way would not swap the order at all.
>
> I'm not really sure how to manage the sparse warnings here.
Not sure whether the maintainers would like it, but we can do something like
this:
case 'l’:
#ifdef __LITTLE_ENDIAN
val = orig;
#else
orig = swab32(orig);
val = orig;
#endif
break;
case 'b’:
#ifdef __LITTLE_ENDIAN
orig = swab32(orig);
val = orig;
#else
val = orig;
#endif
break;
