As discussed on LKML http://marc.info/?i=54611D86.4040306%40de.ibm.com ACCESS_ONCE might fail with specific compiler for non-scalar accesses. Here is a set of patches to tackle that problem. The first patch introduce READ_ONCE and ASSIGN_ONCE. If the data structure is larger than the machine word size memcpy is used and a warning is emitted. The next patches fix up all in-tree users of ACCESS_ONCE on non-scalar types. The last patch forces ACCESS_ONCE to work only on scalar types. I have cross-compiled the resulting kernel with defconfig and gcc 4.9 for microblaze, m68k, alpha, s390,x86_64, i686, sparc, sparc64, mips, ia64, arm and arm64. Runtime tested on s390x and x86_64. I have also verified that ASSIGN_ONCE works as expected with some test changes as there are no user in this patch series. Linus, ok for the next merge window? Christian Borntraeger (9): kernel: Provide READ_ONCE and ASSIGN_ONCE mm: replace ACCESS_ONCE with READ_ONCE or barriers x86/spinlock: Replace ACCESS_ONCE with READ_ONCE x86/gup: Replace ACCESS_ONCE with READ_ONCE mips/gup: Replace ACCESS_ONCE with READ_ONCE arm64/spinlock: Replace ACCESS_ONCE READ_ONCE arm/spinlock: Replace ACCESS_ONCE with READ_ONCE s390/kvm: REPLACE ACCESS_ONCE with READ_ONCE kernel: tighten rules for ACCESS ONCE arch/arm/include/asm/spinlock.h | 4 +-- arch/arm64/include/asm/spinlock.h | 4 +-- arch/mips/mm/gup.c | 2 +- arch/s390/kvm/gaccess.c | 14 ++++---- arch/x86/include/asm/spinlock.h | 8 ++--- arch/x86/mm/gup.c | 2 +- include/linux/compiler.h | 74 ++++++++++++++++++++++++++++++++++++++- mm/gup.c | 2 +- mm/memory.c | 2 +- mm/rmap.c | 3 +- 10 files changed, 95 insertions(+), 20 deletions(-) -- 1.9.3 -- To unsubscribe from this list: send the line "unsubscribe linux-arch" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
