With the recent ffs() and ffs64() optimization a logical and operation was
removed, which allows the compiler to tell that the return value range of
both functions. This may lead to compile warnings as reported by the kernel
test robot.
Instead of only adding the not needed mask again, also add an __assume()
statement to tell newer compilers that they can assume a specific value
range. This allows newer compilers to optimize the not-needed logical and
operation away.
Also change the return type of flogr() to unsigned long and add the const
attribute to the function.
With this the reported warning is away, and in addition the kernel image
size is reduced by ~4kb.
Reported-by: kernel test robot <lkp@intel.com>
Closes: https://lore.kernel.org/oe-kbuild-all/202508211859.UoYsJbLN-lkp@intel.com/
Fixes: de88e74889a3 ("s390/bitops: Slightly optimize ffs() and fls64()")
Suggested-by: Juergen Christ <jchrist@linux.ibm.com>
Signed-off-by: Heiko Carstens <hca@linux.ibm.com>
---
arch/s390/include/asm/bitops.h | 11 +++++++----
1 file changed, 7 insertions(+), 4 deletions(-)
diff --git a/arch/s390/include/asm/bitops.h b/arch/s390/include/asm/bitops.h
index 9dfb687ba620..46b90b545986 100644
--- a/arch/s390/include/asm/bitops.h
+++ b/arch/s390/include/asm/bitops.h
@@ -130,11 +130,12 @@ static inline bool test_bit_inv(unsigned long nr,
* where the most significant bit has bit number 0.
* If no bit is set this function returns 64.
*/
-static __always_inline unsigned char __flogr(unsigned long word)
+static __always_inline __attribute_const__ unsigned long __flogr(unsigned long word)
{
- if (__builtin_constant_p(word)) {
- unsigned long bit = 0;
+ unsigned long bit;
+ if (__builtin_constant_p(word)) {
+ bit = 0;
if (!word)
return 64;
if (!(word & 0xffffffff00000000UL)) {
@@ -169,7 +170,9 @@ static __always_inline unsigned char __flogr(unsigned long word)
asm volatile(
" flogr %[rp],%[rp]\n"
: [rp] "+d" (rp.pair) : : "cc");
- return rp.even;
+ bit = rp.even;
+ __assume(bit <= 64);
+ return bit & 127;
}
}
--
2.48.1