On 09/04/2018 01:10 PM, Andrey Ryabinin wrote:

On 09/04/2018 09:59 AM, Kyeongdon Kim wrote:

quoted

quoted

quoted

+#undef strncmp
+int strncmp(const char *cs, const char *ct, size_t len)
+{
+ check_memory_region((unsigned long)cs, len, false, _RET_IP_);
+ check_memory_region((unsigned long)ct, len, false, _RET_IP_);

This will cause false positives. Both 'cs', and 'ct' could be less than len bytes.

There is no need in these interceptors, just use the C implementations from lib/string.c
like you did in your first patch.
The only thing that was wrong in the first patch is that assembly implementations
were compiled out instead of being declared week.

Well, at first I thought so..
I would remove diff code in /mm/kasan/kasan.c then use C implementations in lib/string.c
w/ assem implementations as weak :

diff --git a/lib/string.c b/lib/string.c
index 2c0900a..a18b18f 100644
--- a/lib/string.c
+++ b/lib/string.c

@@ -312,7 +312,7 @@ size_t strlcat(char *dest, const char *src, size_t count)

?EXPORT_SYMBOL(strlcat);
?#endif

-#ifndef __HAVE_ARCH_STRCMP
+#if (defined(CONFIG_ARM64) && defined(CONFIG_KASAN)) || !defined(__HAVE_ARCH_STRCMP)

No. What part of "like you did in your first patch" is unclear to you?

Just to be absolutely clear, I meant #ifdef out __HAVE_ARCH_* defines like it has been done in this patch
http://lkml.kernel.org/r/[ref]