Fix some regex issues with out-of-range characters and large char ranges.

tglsfdc · tglsfdc · commit 3bb3f42f3749 · 2016-02-08T10:25:40.000-05:00
Previously, our regex code defined CHR_MAX as 0xfffffffe, which is a bad choice because it is outside the range of type "celt" (int32). Characters approaching that limit could lead to infinite loops in logic such as "for (c = a; c <= b; c++)" where c is of type celt but the range bounds are chr. Such loops will work safely only if CHR_MAX+1 is representable in celt, since c must advance to beyond b before the loop will exit. Fortunately, there seems no reason not to restrict CHR_MAX to 0x7ffffffe. It's highly unlikely that Unicode will ever assign codes that high, and none of our other backend encodings need characters beyond that either. In addition to modifying the macro, we have to explicitly enforce character range restrictions on the values of \u, \U, and \x escape sequences, else the limit is trivially bypassed. Also, the code for expanding case-independent character ranges in bracket expressions had a potential integer overflow in its calculation of the number of characters it could generate, which could lead to allocating too small a character vector and then overwriting memory. An attacker with the ability to supply arbitrary regex patterns could easily cause transient DOS via server crashes, and the possibility for privilege escalation has not been ruled out. Quite aside from the integer-overflow problem, the range expansion code was unnecessarily inefficient in that it always produced a result consisting of individual characters, abandoning the knowledge that we had a range to start with. If the input range is large, this requires excessive memory. Change it so that the original range is reported as-is, and then we add on any case-equivalent characters that are outside that range. With this approach, we can bound the number of individual characters allowed without sacrificing much. This patch allows at most 100000 individual characters, which I believe to be more than the number of case pairs existing in Unicode, so that the restriction will never be hit in practice. It's still possible for range() to take awhile given a large character code range, so also add statement-cancel detection to its loop. The downstream function dovec() also lacked cancel detection, and could take a long time given a large output from range(). Per fuzz testing by Greg Stark. Back-patch to all supported branches. Security: CVE-2016-0773
diff --git a/src/backend/regex/regc_lex.c b/src/backend/regex/regc_lex.c
@@ -813,13 +813,13 @@ lexescape(struct vars * v)
 			break;
 		case CHR('u'):
 			c = lexdigits(v, 16, 4, 4);
-			if (ISERR())
+			if (ISERR() || c < CHR_MIN || c > CHR_MAX)
 				FAILW(REG_EESCAPE);
 			RETV(PLAIN, c);
 			break;
 		case CHR('U'):
 			c = lexdigits(v, 16, 8, 8);
-			if (ISERR())
+			if (ISERR() || c < CHR_MIN || c > CHR_MAX)
 				FAILW(REG_EESCAPE);
 			RETV(PLAIN, c);
 			break;
@@ -837,7 +837,7 @@ lexescape(struct vars * v)
 		case CHR('x'):
 			NOTE(REG_UUNPORT);
 			c = lexdigits(v, 16, 1, 255);		/* REs >255 long outside spec */
-			if (ISERR())
+			if (ISERR() || c < CHR_MIN || c > CHR_MAX)
 				FAILW(REG_EESCAPE);
 			RETV(PLAIN, c);
 			break;
@@ -899,6 +899,9 @@ lexescape(struct vars * v)
 
 /*
  * lexdigits - slurp up digits and return chr value
+ *
+ * This does not account for overflow; callers should range-check the result
+ * if maxlen is large enough to make that possible.
  */
 static chr						/* chr value; errors signalled via ERR */
 lexdigits(struct vars * v,
diff --git a/src/backend/regex/regc_locale.c b/src/backend/regex/regc_locale.c
@@ -408,8 +408,7 @@ range(struct vars * v,			/* context */
 	int			nchrs;
 	struct cvec *cv;
 	celt		c,
-				lc,
-				uc;
+				cc;
 
 	if (a != b && !before(a, b))
 	{
@@ -427,24 +426,51 @@ range(struct vars * v,			/* context */
 
 	/*
 	 * When case-independent, it's hard to decide when cvec ranges are usable,
-	 * so for now at least, we won't try.  We allocate enough space for two
-	 * case variants plus a little extra for the two title case variants.
+	 * so for now at least, we won't try.  We use a range for the originally
+	 * specified chrs and then add on any case-equivalents that are outside
+	 * that range as individual chrs.
+	 *
+	 * To ensure sane behavior if someone specifies a very large range, limit
+	 * the allocation size to 100000 chrs (arbitrary) and check for overrun
+	 * inside the loop below.
 	 */
+	nchrs = b - a + 1;
+	if (nchrs <= 0 || nchrs > 100000)
+		nchrs = 100000;
 
-	nchrs = (b - a + 1) * 2 + 4;
-
-	cv = getcvec(v, nchrs, 0);
+	cv = getcvec(v, nchrs, 1);
 	NOERRN();
+	addrange(cv, a, b);
 
 	for (c = a; c <= b; c++)
 	{
-		addchr(cv, c);
-		lc = pg_wc_tolower((chr) c);
-		if (c != lc)
-			addchr(cv, lc);
-		uc = pg_wc_toupper((chr) c);
-		if (c != uc)
-			addchr(cv, uc);
+		cc = pg_wc_tolower((chr) c);
+		if (cc != c &&
+			(before(cc, a) || before(b, cc)))
+		{
+			if (cv->nchrs >= cv->chrspace)
+			{
+				ERR(REG_ETOOBIG);
+				return NULL;
+			}
+			addchr(cv, cc);
+		}
+		cc = pg_wc_toupper((chr) c);
+		if (cc != c &&
+			(before(cc, a) || before(b, cc)))
+		{
+			if (cv->nchrs >= cv->chrspace)
+			{
+				ERR(REG_ETOOBIG);
+				return NULL;
+			}
+			addchr(cv, cc);
+		}
+		if (CANCEL_REQUESTED(v->re))
+		{
+			ERR(REG_CANCEL);
+			return NULL;
+		}
 	}
 
 	return cv;
diff --git a/src/backend/regex/regcomp.c b/src/backend/regex/regcomp.c
@@ -1593,6 +1593,7 @@ dovec(struct vars * v,
 	{
 		ch = *p;
 		newarc(v->nfa, PLAIN, subcolor(v->cm, ch), lp, rp);
+		NOERR();
 	}
 
 	/* and the ranges */
@@ -1602,6 +1603,7 @@ dovec(struct vars * v,
 		to = *(p + 1);
 		if (from <= to)
 			subrange(v, from, to, lp, rp);
+		NOERR();
 	}
 }
 
diff --git a/src/include/regex/regcustom.h b/src/include/regex/regcustom.h
@@ -65,7 +65,8 @@ typedef int celt;				/* type to hold chr, or NOCELT */
 #define DIGITVAL(c) ((c)-'0')	/* turn chr digit into its value */
 #define CHRBITS 32				/* bits in a chr; must not use sizeof */
 #define CHR_MIN 0x00000000		/* smallest and largest chr; the value */
-#define CHR_MAX 0xfffffffe		/* CHR_MAX-CHR_MIN+1 should fit in uchr */
+#define CHR_MAX 0x7ffffffe		/* CHR_MAX-CHR_MIN+1 must fit in an int, and
+								 * CHR_MAX+1 must fit in both chr and celt */
 
 /* functions operating on chr */
 #define iscalnum(x) pg_wc_isalnum(x)
diff --git a/src/test/regress/expected/regex.out b/src/test/regress/expected/regex.out
@@ -495,3 +495,5 @@ select 'xyz' ~ 'x(\w)(?=\1)';  -- no backrefs in LACONs
 ERROR:  invalid regular expression: invalid backreference number
 select 'xyz' ~ 'x(\w)(?=(\1))';
 ERROR:  invalid regular expression: invalid backreference number
+select 'a' ~ '\x7fffffff';  -- invalid chr code
+ERROR:  invalid regular expression: invalid escape \ sequence
diff --git a/src/test/regress/sql/regex.sql b/src/test/regress/sql/regex.sql
@@ -121,3 +121,4 @@ select 'a' ~ '()+\1';
 -- Error conditions
 select 'xyz' ~ 'x(\w)(?=\1)';  -- no backrefs in LACONs
 select 'xyz' ~ 'x(\w)(?=(\1))';
+select 'a' ~ '\x7fffffff';  -- invalid chr code

Original file line number	Diff line number	Diff line change
`@@ -1593,6 +1593,7 @@ dovec(struct vars * v,`
`1593`	`1593`	`{`
`1594`	`1594`	`ch = *p;`
`1595`	`1595`	`newarc(v->nfa, PLAIN, subcolor(v->cm, ch), lp, rp);`
	`1596`	`+ NOERR();`
`1596`	`1597`	`}`
`1597`	`1598`
`1598`	`1599`	`/* and the ranges */`
`@@ -1602,6 +1603,7 @@ dovec(struct vars * v,`
`1602`	`1603`	`to = *(p + 1);`
`1603`	`1604`	`if (from <= to)`
`1604`	`1605`	`subrange(v, from, to, lp, rp);`
	`1606`	`+ NOERR();`
`1605`	`1607`	`}`
`1606`	`1608`	`}`
`1607`	`1609`