On Wed, Sep 12, 2018 at 6:33 PM, Dmitry Vyukov <dvyukov@xxxxxxxxxx> wrote: > On Wed, Aug 29, 2018 at 1:35 PM, Andrey Konovalov <andreyknvl@xxxxxxxxxx> wrote: >> +#ifdef CONFIG_KASAN_HW >> +#define KASAN_TAG_SHIFTED(tag) ((unsigned long)(tag) << 56) >> +#define KASAN_SET_TAG(addr, tag) (((addr) & ~KASAN_TAG_SHIFTED(0xff)) | \ >> + KASAN_TAG_SHIFTED(tag)) >> +#define KASAN_RESET_TAG(addr) KASAN_SET_TAG(addr, 0xff) >> +#endif >> + > > > Wouldn't it be better to > #define KASAN_RESET_TAG(addr) addr > when CONFIG_KASAN_HW is not enabled, and then not duplicate the macros > below? That's what we do in kasan.h for all hooks. > I see that a subsequent patch duplicates yet another macro in this > file. While we could use: > > #define __kimg_to_phys(addr) (KASAN_RESET_TAG(addr) - kimage_voffset) > > with and without kasan. Duplicating them increases risk that somebody > will change only the non-kasan version but forget kasan version. Will do in v7.
