x86/kasan: Print original address on #GP

Make #GP exceptions caused by out-of-bounds KASAN shadow accesses easier
to understand by computing the address of the original access and
printing that. More details are in the comments in the patch.

This turns an error like this:

  kasan: CONFIG_KASAN_INLINE enabled
  kasan: GPF could be caused by NULL-ptr deref or user memory access
  general protection fault, probably for non-canonical address
      0xe017577ddf75b7dd: 0000 [#1] PREEMPT SMP KASAN PTI

into this:

  general protection fault, probably for non-canonical address
      0xe017577ddf75b7dd: 0000 [#1] PREEMPT SMP KASAN PTI
  KASAN: maybe wild-memory-access in range
      [0x00badbeefbadbee8-0x00badbeefbadbeef]

The hook is placed in architecture-independent code, but is currently
only wired up to the X86 exception handler because I'm not sufficiently
familiar with the address space layout and exception handling mechanisms
on other architectures.

Signed-off-by: Jann Horn <[email protected]>
Signed-off-by: Borislav Petkov <[email protected]>
Reviewed-by: Dmitry Vyukov <[email protected]>
Cc: Alexander Potapenko <[email protected]>
Cc: Andrew Morton <[email protected]>
Cc: Andrey Konovalov <[email protected]>
Cc: Andrey Ryabinin <[email protected]>
Cc: Andy Lutomirski <[email protected]>
Cc: Dave Hansen <[email protected]>
Cc: "H. Peter Anvin" <[email protected]>
Cc: Ingo Molnar <[email protected]>
Cc: [email protected]
Cc: linux-mm <[email protected]>
Cc: Peter Zijlstra <[email protected]>
Cc: Sean Christopherson <[email protected]>
Cc: Thomas Gleixner <[email protected]>
Cc: x86-ml <[email protected]>
Link: https://lkml.kernel.org/r/[email protected]

Author: thejh

arch/x86/kernel/dumpstack.c

@@ -427,6 +427,8 @@ void die_addr(const char *str, struct pt_regs *regs, long err, long gp_addr)
 	int sig = SIGSEGV;
 
 	__die_header(str, regs, err);
+	if (gp_addr)
+		kasan_non_canonical_hook(gp_addr);
 	if (__die_body(str, regs, err))
 		sig = 0;
 	oops_end(flags, regs, sig);

arch/x86/mm/kasan_init_64.c

@@ -288,23 +288,6 @@ static void __init kasan_shallow_populate_pgds(void *start, void *end)
 	} while (pgd++, addr = next, addr != (unsigned long)end);
 }
 
-#ifdef CONFIG_KASAN_INLINE
-static int kasan_die_handler(struct notifier_block *self,
-			     unsigned long val,
-			     void *data)
-{
-	if (val == DIE_GPF) {
-		pr_emerg("CONFIG_KASAN_INLINE enabled\n");
-		pr_emerg("GPF could be caused by NULL-ptr deref or user memory access\n");
-	}
-	return NOTIFY_OK;
-}
-
-static struct notifier_block kasan_die_notifier = {
-	.notifier_call = kasan_die_handler,
-};
-#endif
-
 void __init kasan_early_init(void)
 {
 	int i;
@@ -341,10 +324,6 @@ void __init kasan_init(void)
 	int i;
 	void *shadow_cpu_entry_begin, *shadow_cpu_entry_end;
 
-#ifdef CONFIG_KASAN_INLINE
-	register_die_notifier(&kasan_die_notifier);
-#endif
-
 	memcpy(early_top_pgt, init_top_pgt, sizeof(early_top_pgt));
 
 	/*

include/linux/kasan.h

@@ -228,4 +228,10 @@ static inline void kasan_release_vmalloc(unsigned long start,
 					 unsigned long free_region_end) {}
 #endif
 
+#ifdef CONFIG_KASAN_INLINE
+void kasan_non_canonical_hook(unsigned long addr);
+#else /* CONFIG_KASAN_INLINE */
+static inline void kasan_non_canonical_hook(unsigned long addr) { }
+#endif /* CONFIG_KASAN_INLINE */
+
 #endif /* LINUX_KASAN_H */

mm/kasan/report.c

@@ -512,3 +512,43 @@ void __kasan_report(unsigned long addr, size_t size, bool is_write, unsigned lon
 
 	end_report(&flags);
 }
+
+#ifdef CONFIG_KASAN_INLINE
+/*
+ * With CONFIG_KASAN_INLINE, accesses to bogus pointers (outside the high
+ * canonical half of the address space) cause out-of-bounds shadow memory reads
+ * before the actual access. For addresses in the low canonical half of the
+ * address space, as well as most non-canonical addresses, that out-of-bounds
+ * shadow memory access lands in the non-canonical part of the address space.
+ * Help the user figure out what the original bogus pointer was.
+ */
+void kasan_non_canonical_hook(unsigned long addr)
+{
+	unsigned long orig_addr;
+	const char *bug_type;
+
+	if (addr < KASAN_SHADOW_OFFSET)
+		return;
+
+	orig_addr = (addr - KASAN_SHADOW_OFFSET) << KASAN_SHADOW_SCALE_SHIFT;
+	/*
+	 * For faults near the shadow address for NULL, we can be fairly certain
+	 * that this is a KASAN shadow memory access.
+	 * For faults that correspond to shadow for low canonical addresses, we
+	 * can still be pretty sure - that shadow region is a fairly narrow
+	 * chunk of the non-canonical address space.
+	 * But faults that look like shadow for non-canonical addresses are a
+	 * really large chunk of the address space. In that case, we still
+	 * print the decoded address, but make it clear that this is not
+	 * necessarily what's actually going on.
+	 */
+	if (orig_addr < PAGE_SIZE)
+		bug_type = "null-ptr-deref";
+	else if (orig_addr < TASK_SIZE)
+		bug_type = "probably user-memory-access";
+	else
+		bug_type = "maybe wild-memory-access";
+	pr_alert("KASAN: %s in range [0x%016lx-0x%016lx]\n", bug_type,
+		 orig_addr, orig_addr + KASAN_SHADOW_MASK);
+}
+#endif