MDEV-35688 UBSAN: SUMMARY: UndefinedBehaviorSanitizer: nullptr-with-offset in my_casedn_utf8mb3

The functions MY_CHARSET_HANDLER::caseup() and MY_CHARSET_HANDLER::casedn()
in their virtual imlementations do "const char *end= src + srclen"
in the very beginning. Therefore src cannot be NULL to avoid
"UBSAN: SUMMARY: UndefinedBehaviorSanitizer: nullptr-with-offset".

Adding DBUG_ASSERT(src != NULL) into all virtual implementations,
to catch this problem in regular Debug builds (without UBSAN).

Fixing Master_info_index::get_master_info() to check connection_name->str.
If it is NULL then passing empty_clex_str into IdentBufferCasedn
instead of *connection_name.

mysql-test/suite/sys_vars/r/replicate_do_db_basic.result

@@ -43,3 +43,9 @@ SELECT @@GLOBAL.replicate_do_db;
 
 # Cleanup.
 SET @@GLOBAL.replicate_do_db = @save_replicate_do_db;
+#
+# MDEV-35688 UBSAN: SUMMARY: UndefinedBehaviorSanitizer: nullptr-with-offset in my_casedn_utf8mb3
+#
+show variables like 'replicate_do_db';
+Variable_name	Value
+replicate_do_db	

mysql-test/suite/sys_vars/t/replicate_do_db_basic.test

@@ -43,3 +43,9 @@ SELECT @@GLOBAL.replicate_do_db;
 
 --echo # Cleanup.
 SET @@GLOBAL.replicate_do_db = @save_replicate_do_db;
+
+--echo #
+--echo # MDEV-35688 UBSAN: SUMMARY: UndefinedBehaviorSanitizer: nullptr-with-offset in my_casedn_utf8mb3
+--echo #
+
+show variables like 'replicate_do_db';

sql/rpl_mi.cc

@@ -1374,7 +1374,9 @@ Master_info_index::get_master_info(const LEX_CSTRING *connection_name,
               connection_name->str));
 
   /* Make name lower case for comparison */
-  IdentBufferCasedn<MAX_CONNECTION_NAME> buff(*connection_name);
+  IdentBufferCasedn<MAX_CONNECTION_NAME> buff(connection_name->str ?
+                                              *connection_name :
+                                              empty_clex_str);
   mi= (Master_info*) my_hash_search(&master_info_hash,
                                     (const uchar*) buff.ptr(), buff.length());
   if (!mi && warning != Sql_condition::WARN_LEVEL_NOTE)

strings/ctype-mb.c

@@ -125,6 +125,7 @@ size_t
 my_casedn_mb(CHARSET_INFO * cs, const char *src, size_t srclen,
                     char *dst, size_t dstlen)
 {
+  DBUG_ASSERT(src != NULL); /* Avoid UBSAN nullptr-with-offset */
   DBUG_ASSERT(dstlen >= srclen * cs->cset->casedn_multiply(cs));
   DBUG_ASSERT(src != dst || cs->cset->casedn_multiply(cs) == 1);
   return my_casefold_mb(cs, src, srclen, dst, dstlen, cs->to_lower, 0);
@@ -135,6 +136,7 @@ size_t
 my_caseup_mb(CHARSET_INFO * cs, const char *src, size_t srclen,
              char *dst, size_t dstlen)
 {
+  DBUG_ASSERT(src != NULL); /* Avoid UBSAN nullptr-with-offset */
   DBUG_ASSERT(dstlen >= srclen * cs->cset->caseup_multiply(cs));
   DBUG_ASSERT(src != dst || cs->cset->caseup_multiply(cs) == 1);
   return my_casefold_mb(cs, src, srclen, dst, dstlen, cs->to_upper, 1);

strings/ctype-simple.c

@@ -253,6 +253,7 @@ size_t my_caseup_8bit(CHARSET_INFO * cs, const char *src, size_t srclen,
 {
   const char *end= src + srclen;
   register const uchar *map= cs->to_upper;
+  DBUG_ASSERT(src != NULL); /* Avoid UBSAN nullptr-with-offset */
   DBUG_ASSERT(srclen <= dstlen);
   for ( ; src != end ; src++)
     *dst++= (char) map[(uchar) *src];
@@ -265,6 +266,7 @@ size_t my_casedn_8bit(CHARSET_INFO * cs, const char *src, size_t srclen,
 {
   const char *end= src + srclen;
   register const uchar *map=cs->to_lower;
+  DBUG_ASSERT(src != NULL); /* Avoid UBSAN nullptr-with-offset */
   DBUG_ASSERT(srclen <= dstlen);
   for ( ; src != end ; src++)
     *dst++= (char) map[(uchar) *src];

strings/ctype-ucs2.c

@@ -1310,6 +1310,7 @@ my_caseup_utf16(CHARSET_INFO *cs, const char *src, size_t srclen,
   const char *srcend= src + srclen;
   char *dstend= dst + dstlen;
   MY_CASEFOLD_INFO *uni_plane= cs->casefold;
+  DBUG_ASSERT(src != NULL); /* Avoid UBSAN nullptr-with-offset */
   DBUG_ASSERT(srclen <= dstlen);
   
   while ((src < srcend) &&
@@ -1368,6 +1369,7 @@ my_casedn_utf16(CHARSET_INFO *cs, const char *src, size_t srclen,
   const char *srcend= src + srclen;
   char *dstend= dst + dstlen;
   MY_CASEFOLD_INFO *uni_plane= cs->casefold;
+  DBUG_ASSERT(src != NULL); /* Avoid UBSAN nullptr-with-offset */
   DBUG_ASSERT(srclen <= dstlen);
 
   while ((src < srcend) &&
@@ -2180,6 +2182,7 @@ my_caseup_utf32(CHARSET_INFO *cs, const char *src, size_t srclen,
   const char *srcend= src + srclen;
   char *dstend= dst + dstlen;
   MY_CASEFOLD_INFO *uni_plane= cs->casefold;
+  DBUG_ASSERT(src != NULL); /* Avoid UBSAN nullptr-with-offset */
   DBUG_ASSERT(srclen <= dstlen);
   
   while ((src < srcend) &&
@@ -2237,6 +2240,7 @@ my_casedn_utf32(CHARSET_INFO *cs, const char *src, size_t srclen,
   const char *srcend= src + srclen;
   char *dstend= dst + dstlen;
   MY_CASEFOLD_INFO *uni_plane= cs->casefold;
+  DBUG_ASSERT(src != NULL); /* Avoid UBSAN nullptr-with-offset */
   DBUG_ASSERT(srclen <= dstlen);
 
   while ((res= my_utf32_uni(cs, &wc, (uchar*) src, (uchar*) srcend)) > 0)
@@ -3048,6 +3052,7 @@ static size_t my_caseup_ucs2(CHARSET_INFO *cs, const char *src, size_t srclen,
   const char *srcend= src + srclen;
   char *dstend= dst + dstlen;
   MY_CASEFOLD_INFO *uni_plane= cs->casefold;
+  DBUG_ASSERT(src != NULL); /* Avoid UBSAN nullptr-with-offset */
   DBUG_ASSERT(srclen <= dstlen);
   
   while ((src < srcend) &&
@@ -3100,6 +3105,7 @@ static size_t my_casedn_ucs2(CHARSET_INFO *cs, const char *src, size_t srclen,
   char *dstend= dst + dstlen;
   MY_CASEFOLD_INFO *uni_plane= cs->casefold;
   DBUG_ASSERT(srclen <= dstlen);
+  DBUG_ASSERT(src != NULL); /* Avoid UBSAN nullptr-with-offset */
 
   while ((src < srcend) &&
          (res= my_ucs2_uni(cs, &wc, (uchar*) src, (uchar*) srcend)) > 0)

strings/ctype-ujis.c

@@ -67219,6 +67219,7 @@ size_t
 my_casedn_ujis(CHARSET_INFO * cs, const char *src, size_t srclen,
                char *dst, size_t dstlen)
 {
+  DBUG_ASSERT(src != NULL); /* Avoid UBSAN nullptr-with-offset */
   DBUG_ASSERT(dstlen >= srclen * cs->cset->casedn_multiply(cs));
   DBUG_ASSERT(src != dst || cs->cset->casedn_multiply(cs) == 1);
   return my_casefold_ujis(cs, src, srclen, dst, dstlen, cs->to_lower, 0);
@@ -67232,6 +67233,7 @@ size_t
 my_caseup_ujis(CHARSET_INFO * cs, const char *src, size_t srclen,
                char *dst, size_t dstlen)
 {
+  DBUG_ASSERT(src != NULL); /* Avoid UBSAN nullptr-with-offset */
   DBUG_ASSERT(dstlen >= srclen * cs->cset->caseup_multiply(cs));
   DBUG_ASSERT(src != dst || cs->cset->caseup_multiply(cs) == 1);
   return my_casefold_ujis(cs, src, srclen, dst, dstlen, cs->to_upper, 1);

strings/ctype-utf8.c

@@ -676,6 +676,7 @@ static size_t my_caseup_utf8mb3(CHARSET_INFO *cs,
   const char *srcend= src + srclen;
   char *dstend= dst + dstlen, *dst0= dst;
   MY_CASEFOLD_INFO *uni_plane= cs->casefold;
+  DBUG_ASSERT(src != NULL); /* Avoid UBSAN nullptr-with-offset */
   DBUG_ASSERT(src != dst || cs->cset->caseup_multiply(cs) == 1);
 
   while ((src < srcend) &&
@@ -754,6 +755,7 @@ static size_t my_casedn_utf8mb3(CHARSET_INFO *cs,
   const char *srcend= src + srclen;
   char *dstend= dst + dstlen, *dst0= dst;
   MY_CASEFOLD_INFO *uni_plane= cs->casefold;
+  DBUG_ASSERT(src != NULL); /* Avoid UBSAN nullptr-with-offset */
   DBUG_ASSERT(src != dst || cs->cset->casedn_multiply(cs) == 1);
 
   while ((src < srcend) &&
@@ -3124,6 +3126,7 @@ my_caseup_utf8mb4(CHARSET_INFO *cs, const char *src, size_t srclen,
   const char *srcend= src + srclen;
   char *dstend= dst + dstlen, *dst0= dst;
   MY_CASEFOLD_INFO *uni_plane= cs->casefold;
+  DBUG_ASSERT(src != NULL); /* Avoid UBSAN nullptr-with-offset */
   DBUG_ASSERT(src != dst || cs->cset->caseup_multiply(cs) == 1);
 
   while ((src < srcend) &&
@@ -3218,6 +3221,7 @@ my_casedn_utf8mb4(CHARSET_INFO *cs,
   const char *srcend= src + srclen;
   char *dstend= dst + dstlen, *dst0= dst;
   MY_CASEFOLD_INFO *uni_plane= cs->casefold;
+  DBUG_ASSERT(src != NULL); /* Avoid UBSAN nullptr-with-offset */
   DBUG_ASSERT(src != dst || cs->cset->casedn_multiply(cs) == 1);
 
   while ((src < srcend) &&