On Wed, Feb 21, 2024 at 06:37:13PM -0800, Charlie Jenkins wrote: > The parisc implementation of csum_partial previously folded the result > into 16 bits instead of returning all 32 bits and letting consumers like > ip_compute_csum do the folding. Since ip_compute_csum no longer depends > on this requirement, remove the folding so that the parisc > implementation operates the same as other architectures. > > Signed-off-by: Charlie Jenkins <charlie@xxxxxxxxxxxx> Tested-by: Guenter Roeck <linux@xxxxxxxxxxxx> > --- > arch/parisc/lib/checksum.c | 5 +---- > 1 file changed, 1 insertion(+), 4 deletions(-) > > diff --git a/arch/parisc/lib/checksum.c b/arch/parisc/lib/checksum.c > index 05f5ca4b2f96..eaa660491e24 100644 > --- a/arch/parisc/lib/checksum.c > +++ b/arch/parisc/lib/checksum.c > @@ -95,14 +95,11 @@ unsigned int do_csum(const unsigned char *buff, int len) > /* > * computes a partial checksum, e.g. for TCP/UDP fragments > */ > -/* > - * why bother folding? > - */ > __wsum csum_partial(const void *buff, int len, __wsum sum) > { > unsigned int result = do_csum(buff, len); > addc(result, sum); > - return (__force __wsum)from32to16(result); > + return (__force __wsum)result; > } > > EXPORT_SYMBOL(csum_partial); > > -- > 2.34.1 >