On 06.07.2023 01:49, Stefano Stabellini wrote: >> --- a/xen/include/xen/bitops.h >> +++ b/xen/include/xen/bitops.h >> @@ -227,11 +227,11 @@ static inline __u32 ror32(__u32 word, unsigned int >> shift) >> } >> >> /* base-2 logarithm */ >> -#define __L2(_x) (((_x) & 0x00000002) ? 1 : 0) >> -#define __L4(_x) (((_x) & 0x0000000c) ? ( 2 + __L2( (_x)>> 2)) : __L2( _x)) >> -#define __L8(_x) (((_x) & 0x000000f0) ? ( 4 + __L4( (_x)>> 4)) : __L4( _x)) >> -#define __L16(_x) (((_x) & 0x0000ff00) ? ( 8 + __L8( (_x)>> 8)) : __L8( _x)) >> -#define ilog2(_x) (((_x) & 0xffff0000) ? (16 + __L16((_x)>>16)) : __L16(_x)) >> +#define __L2(_x) (((_x) & 0x00000002U) ? 1 : 0) >> +#define __L4(_x) (((_x) & 0x0000000cU) ? ( 2 + __L2( (_x)>> 2)) : __L2( >> _x)) >> +#define __L8(_x) (((_x) & 0x000000f0U) ? ( 4 + __L4( (_x)>> 4)) : __L4( >> _x)) >> +#define __L16(_x) (((_x) & 0x0000ff00U) ? ( 8 + __L8( (_x)>> 8)) : __L8( >> _x)) >> +#define ilog2(_x) (((_x) & 0xffff0000U) ? (16 + __L16((_x)>>16)) : >> __L16(_x)) > > Jan, would you want to replace _x with x here as well?
Yes. > If so, I think it would be best if done in a separate patch (keep > mechanical changes mechanical) and keep this patch as is: It was done in one go in patch 9, so I don't see why doing it here as well would do any harm. Additionally missing blanks before >> should be inserted at this occasion, and perhaps excess ones dropped from the first of the changed lines (or further ones added to align toe colons). While with that the lines still stay just below the 80 chars limit, dropping the leading underscores would get us a little father away from that boundary. Jan
