On Aug 1, 2019, at 8:07 AM, Qian Cai [off-list ref] wrote:

quoted

On Aug 1, 2019, at 8:01 AM, Will Deacon [off-list ref] wrote:

On Wed, Jul 31, 2019 at 04:38:18PM -0400, Qian Cai wrote:

quoted

When CONFIG_KASAN_SW_TAGS=n, set_tag() is compiled away. GCC throws a
warning,

mm/kasan/common.c: In function '__kasan_kmalloc':
mm/kasan/common.c:464:5: warning: variable 'tag' set but not used
[-Wunused-but-set-variable]
u8 tag = 0xff;
   ^~~

Fix it by making __tag_set() a static inline function.

Signed-off-by: Qian Cai <redacted>
---
arch/arm64/include/asm/memory.h | 6 +++++-
1 file changed, 5 insertions(+), 1 deletion(-)

diff --git a/arch/arm64/include/asm/memory.h b/arch/arm64/include/asm/memory.h
index b7ba75809751..9645b1340afe 100644
--- a/arch/arm64/include/asm/memory.h
+++ b/arch/arm64/include/asm/memory.h

@@ -210,7 +210,11 @@ static inline unsigned long kaslr_offset(void)

#define __tag_reset(addr)	untagged_addr(addr)
#define __tag_get(addr)		(__u8)((u64)(addr) >> 56)
#else
-#define __tag_set(addr, tag)	(addr)
+static inline const void *__tag_set(const void *addr, u8 tag)
+{
+	return addr;
+}

Why doesn't this trigger a warning in page_to_virt(), which passes an
unsigned long for the address parameter?

#define page_to_virt(page) … __tag_set(__addr, page_kasan_tag(page)); …

static inline u8 page_kasan_tag(const struct page *page)
{	
	return 0xff;
}

GCC will see that “page” is used.

If you are talking about “addr”,

#define __tag_set(addr, tag)	(addr)

It is actually used.


_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel