Fix OOB read in stdlib thousand grouping parsing [BZ #29727]

__correctly_grouped_prefixmb only worked with thousands_len == 1,
otherwise it read past the end of cp or thousands.

This affects scanf formats like %'d, %'f and the internal but
exposed __strto{l,ul,f,d,..}_internal with grouping flag set
and an LC_NUMERIC locale where thousands_len > 1.

Avoid OOB access by considering thousands_len when initializing cp.
This fixes bug 29727.

Found by the morello port with strict bounds checking where

FAIL: stdlib/tst-strtod4
FAIL: stdlib/tst-strtod5i

crashed using a locale with thousands_len==3.

diff --git a/stdlib/grouping.c b/stdlib/grouping.c
index be7922f5fdcd1e12c97d916969dd8e5e77dba2f1..06cbe7b9c7a6e3b96af1b27c0aa41bbaf3747f43 100644 (file)
--- a/stdlib/grouping.c
+++ b/stdlib/grouping.c
@@ -52,21 +52,19 @@ __correctly_grouped_prefixmb (const STRING_TYPE *begin, const STRING_TYPE *end,
 #endif
                              const char *grouping)
 {
-#ifndef USE_WIDE_CHAR
-  size_t thousands_len;
-  int cnt;
-#endif
-
   if (grouping == NULL)
     return end;
 
-#ifndef USE_WIDE_CHAR
-  thousands_len = strlen (thousands);
+#ifdef USE_WIDE_CHAR
+  size_t thousands_len = 1;
+#else
+  size_t thousands_len = strlen (thousands);
+  int cnt;
 #endif
 
-  while (end > begin)
+  while (end - begin >= thousands_len)
     {
-      const STRING_TYPE *cp = end - 1;
+      const STRING_TYPE *cp = end - thousands_len;
       const char *gp = grouping;
 
       /* Check first group.  */