| Message ID | 20240124142857.4146716-4-libaokun1@huawei.com (mailing list archive) |
|---|---|
| State | Accepted |
| Headers | show |
| Series | fs: make the i_size_read/write helpers be smp_load_acquire/store_release() | expand |
diff --git a/include/asm-generic/barrier.h b/include/asm-generic/barrier.h index 961f4d88f9ef..0c0695763bea 100644 --- a/include/asm-generic/barrier.h +++ b/include/asm-generic/barrier.h @@ -193,7 +193,6 @@ do { \ #ifndef smp_store_release #define smp_store_release(p, v) \ do { \ - compiletime_assert_atomic_type(*p); \ barrier(); \ WRITE_ONCE(*p, v); \ } while (0) @@ -203,7 +202,6 @@ do { \ #define smp_load_acquire(p) \ ({ \ __unqual_scalar_typeof(*p) ___p1 = READ_ONCE(*p); \ - compiletime_assert_atomic_type(*p); \ barrier(); \ (typeof(*p))___p1; \ })
If CONFIG_SMP is not enabled, the smp_load_acquire/smp_store_release is implemented as READ_ONCE/READ_ONCE and barrier() and type checking. READ_ONCE/READ_ONCE already checks the pointer type, and then checks it more stringently outside, but the non-SMP case simply isn't relevant, so remove the extra compiletime_assert_atomic_type() to avoid compilation errors. Reported-by: kernel test robot <lkp@intel.com> Closes: https://lore.kernel.org/oe-kbuild-all/202401230837.TXro0PHi-lkp@intel.com/ Suggested-by: Linus Torvalds <torvalds@linux-foundation.org> Signed-off-by: Baokun Li <libaokun1@huawei.com> --- include/asm-generic/barrier.h | 2 -- 1 file changed, 2 deletions(-)