s390/uaccess: avoid __ashlti3() call

__cmpxchg_user_key() uses 128 bit types which, depending on compiler
and config options, may lead to an __ashlti3() library call.

Get rid of that by simply casting the 128 bit values to 32 bit values.

Reported-by: kernel test robot <lkp@intel.com>
Suggested-by: Janis Schoetterl-Glausch <scgl@linux.ibm.com>
Fixes: 51098f0eb22e ("s390/cmpxchg: make loop condition for 1,2 byte cases precise")
Link: https://lore.kernel.org/all/4b96b112d5415d08a81d30657feec2c8c3000f7c.camel@linux.ibm.com/
Signed-off-by: Heiko Carstens <hca@linux.ibm.com>
This commit is contained in:
Heiko Carstens 2023-01-04 16:55:53 +01:00
parent 739ad2e4e1
commit b33d59fb37

View File

@ -407,8 +407,8 @@ static __always_inline int __cmpxchg_user_key(unsigned long address, void *uval,
shift = (3 ^ (address & 3)) << 3;
address ^= address & 3;
_old = (old & 0xff) << shift;
_new = (new & 0xff) << shift;
_old = ((unsigned int)old & 0xff) << shift;
_new = ((unsigned int)new & 0xff) << shift;
mask = ~(0xff << shift);
asm volatile(
" spka 0(%[key])\n"
@ -455,8 +455,8 @@ static __always_inline int __cmpxchg_user_key(unsigned long address, void *uval,
shift = (2 ^ (address & 2)) << 3;
address ^= address & 2;
_old = (old & 0xffff) << shift;
_new = (new & 0xffff) << shift;
_old = ((unsigned int)old & 0xffff) << shift;
_new = ((unsigned int)new & 0xffff) << shift;
mask = ~(0xffff << shift);
asm volatile(
" spka 0(%[key])\n"