On Tue, Dec 10, 2019 at 04:38:54PM +1100, Michael Ellerman wrote: > Good question, I'll have a look. > > There seems to be confusion about what the type of the bit number is, > which is leading to sign extension in some cases and not others. Shiny. > It looks like the type should be unsigned long? I'm thinking unsigned makes most sense, I mean, negative bit offsets should 'work' but that's almost always guaranteed to be an out-of-bound operation. As to 'long' vs 'int', I'm not sure, 4G bits is a long bitmap. But I suppose since the bitmap itself is 'unsigned long', we might as well use 'unsigned long' for the bitnr too. > Documentation/core-api/atomic_ops.rst: void __clear_bit_unlock(unsigned long nr, unsigned long *addr); > arch/mips/include/asm/bitops.h:static inline void __clear_bit_unlock(unsigned long nr, volatile unsigned long *addr) > arch/powerpc/include/asm/bitops.h:static inline void arch___clear_bit_unlock(int nr, volatile unsigned long *addr) > arch/riscv/include/asm/bitops.h:static inline void __clear_bit_unlock(unsigned long nr, volatile unsigned long *addr) > arch/s390/include/asm/bitops.h:static inline void arch___clear_bit_unlock(unsigned long nr, > include/asm-generic/bitops/instrumented-lock.h:static inline void __clear_bit_unlock(long nr, volatile unsigned long *addr) > include/asm-generic/bitops/lock.h:static inline void __clear_bit_unlock(unsigned int nr, > > So I guess step one is to convert our versions to use unsigned long, so > we're at least not tripping over that difference when comparing the > assembly. Yeah, I'll look at fixing the generic code, bitops/atomic.h and bitops/non-atomic.h don't even agree on the type of bitnr.