This is just the same as PAGE_ALIGN(), but rounds the address down, not up. Suggested-by: Dmitry Vyukov <dvyukov@xxxxxxxxxx> Signed-off-by: David Gow <davidgow@xxxxxxxxxx> Acked-by: Andrew Morton <akpm@xxxxxxxxxxxxxxxxxxxx> --- Please take this patch as part of the UML tree, along with patch #2, thanks! Changes since v2: https://lore.kernel.org/lkml/20220527185600.1236769-1-davidgow@xxxxxxxxxx/ - Add Andrew's Acked-by tag. v2 was the first version of this patch (it having been introduced as part of v2 of the UML/KASAN series). There are almost certainly lots of places where this macro should be used: just look for ALIGN_DOWN(..., PAGE_SIZE). I haven't gone through to try to replace them all. --- include/linux/mm.h | 3 +++ 1 file changed, 3 insertions(+) diff --git a/include/linux/mm.h b/include/linux/mm.h index 9f44254af8ce..9abe5975ad11 100644 --- a/include/linux/mm.h +++ b/include/linux/mm.h @@ -221,6 +221,9 @@ int overcommit_policy_handler(struct ctl_table *, int, void *, size_t *, /* to align the pointer to the (next) page boundary */ #define PAGE_ALIGN(addr) ALIGN(addr, PAGE_SIZE) +/* to align the pointer to the (prev) page boundary */ +#define PAGE_ALIGN_DOWN(addr) ALIGN_DOWN(addr, PAGE_SIZE) + /* test whether an address (unsigned long or pointer) is aligned to PAGE_SIZE */ #define PAGE_ALIGNED(addr) IS_ALIGNED((unsigned long)(addr), PAGE_SIZE) -- 2.37.0.rc0.161.g10f37bed90-goog
