From: Arjun Shankar <arjun@redhat.com>
Date: Wed, 4 Nov 2020 11:19:38 +0000 (+0100)
Subject: iconv: Accept redundant shift sequences in IBM1364 [BZ #26224]
X-Git-Url: http://git.ipfire.org/cgi-bin/gitweb.cgi?a=commitdiff_plain;h=050022910be1d1f5c11cd5168f1685ad4f9580d2;p=thirdparty%2Fglibc.git

iconv: Accept redundant shift sequences in IBM1364 [BZ #26224]

The IBM1364, IBM1371, IBM1388, IBM1390 and IBM1399 character sets
share converter logic (iconvdata/ibm1364.c) which would reject
redundant shift sequences when processing input in these character
sets.  This led to a hang in the iconv program (CVE-2020-27618).

This commit adjusts the converter to ignore redundant shift sequences
and adds test cases for iconv_prog hangs that would be triggered upon
their rejection.  This brings the implementation in line with other
converters that also ignore redundant shift sequences (e.g. IBM930
etc., fixed in commit 692de4b3960d).

Reviewed-by: Carlos O'Donell <carlos@redhat.com>
(cherry picked from commit 9a99c682144bdbd40792ebf822fe9264e0376fb5)
---

diff --git a/NEWS b/NEWS
index 8b186756833..4da2a0cbda6 100644
--- a/NEWS
+++ b/NEWS
@@ -7,6 +7,8 @@ using `glibc' in the "product" field.
 
 The following bugs are resolved with this release:
 
+  [26224] iconv hangs when converting some invalid inputs from several IBM
+    character sets (CVE-2020-27618)
   [26534] libm.so 2.32 SIGILL in pow() due to FMA4 instruction on non-FMA4
     system
   [26555] string: strerrorname_np does not return the documented value
@@ -199,6 +201,10 @@ Security related changes:
   Dytrych of the Cisco Security Assessment and Penetration Team (See
   TALOS-2020-1019).
 
+  CVE-2020-27618: An infinite loop has been fixed in the iconv program when
+  invoked with input containing redundant shift sequences in the IBM1364,
+  IBM1371, IBM1388, IBM1390, or IBM1399 character sets.
+
 The following bugs are resolved with this release:
 
   [9809] localedata: ckb_IQ: new Kurdish Sorani locale
diff --git a/iconv/tst-iconv_prog.sh b/iconv/tst-iconv_prog.sh
index 8298136b7f4..d8db7b335c1 100644
--- a/iconv/tst-iconv_prog.sh
+++ b/iconv/tst-iconv_prog.sh
@@ -102,12 +102,16 @@ hangarray=(
 "\x00\x80;-c;IBM1161;UTF-8//TRANSLIT//IGNORE"
 "\x00\xdb;-c;IBM1162;UTF-8//TRANSLIT//IGNORE"
 "\x00\x70;-c;IBM12712;UTF-8//TRANSLIT//IGNORE"
-# These are known hangs that are yet to be fixed:
-# "\x00\x0f;-c;IBM1364;UTF-8"
-# "\x00\x0f;-c;IBM1371;UTF-8"
-# "\x00\x0f;-c;IBM1388;UTF-8"
-# "\x00\x0f;-c;IBM1390;UTF-8"
-# "\x00\x0f;-c;IBM1399;UTF-8"
+"\x00\x0f;-c;IBM1364;UTF-8"
+"\x0e\x0e;-c;IBM1364;UTF-8"
+"\x00\x0f;-c;IBM1371;UTF-8"
+"\x0e\x0e;-c;IBM1371;UTF-8"
+"\x00\x0f;-c;IBM1388;UTF-8"
+"\x0e\x0e;-c;IBM1388;UTF-8"
+"\x00\x0f;-c;IBM1390;UTF-8"
+"\x0e\x0e;-c;IBM1390;UTF-8"
+"\x00\x0f;-c;IBM1399;UTF-8"
+"\x0e\x0e;-c;IBM1399;UTF-8"
 "\x00\x53;-c;IBM16804;UTF-8//TRANSLIT//IGNORE"
 "\x00\x41;-c;IBM274;UTF-8//TRANSLIT//IGNORE"
 "\x00\x41;-c;IBM275;UTF-8//TRANSLIT//IGNORE"
diff --git a/iconvdata/ibm1364.c b/iconvdata/ibm1364.c
index 49e7267ab45..521f0825b7f 100644
--- a/iconvdata/ibm1364.c
+++ b/iconvdata/ibm1364.c
@@ -158,24 +158,14 @@ enum
 									      \
     if (__builtin_expect (ch, 0) == SO)					      \
       {									      \
-	/* Shift OUT, change to DBCS converter.  */			      \
-	if (curcs == db)						      \
-	  {								      \
-	    result = __GCONV_ILLEGAL_INPUT;				      \
-	    break;							      \
-	  }								      \
+	/* Shift OUT, change to DBCS converter (redundant escape okay).  */   \
 	curcs = db;							      \
 	++inptr;							      \
 	continue;							      \
       }									      \
     if (__builtin_expect (ch, 0) == SI)					      \
       {									      \
-	/* Shift IN, change to SBCS converter.  */			      \
-	if (curcs == sb)						      \
-	  {								      \
-	    result = __GCONV_ILLEGAL_INPUT;				      \
-	    break;							      \
-	  }								      \
+	/* Shift IN, change to SBCS converter (redundant escape okay).  */    \
 	curcs = sb;							      \
 	++inptr;							      \
 	continue;							      \