libsemanage direct_api can return errors < 0.

libsemanage direct_api can return errors < 0.

[Daniel J Walsh]

[-- Attachment #1: Type: text/plain, Size: 0 bytes --]



[-- Attachment #2: libsemanage-directapi.patch --]
[-- Type: text/plain, Size: 1177 bytes --]

--- nsalibsemanage/src/direct_api.c	2009-01-13 08:45:35.000000000 -0500
+++ libsemanage-2.0.31/src/direct_api.c	2009-06-02 14:54:20.000000000 -0400
@@ -991,7 +991,7 @@
 
 	int retval = -1;
 	char *data = NULL;
-	size_t data_len = 0;
+	ssize_t data_len = 0;
 	int compressed = 0;
 	int in_fd = -1;
 
@@ -999,7 +999,7 @@
 		return -1;
 	}
 
-	if ((data_len = map_file(in_fd, &data, &compressed)) == 0) {
+	if ((data_len = map_file(in_fd, &data, &compressed)) <= 0) {
 		goto cleanup;
 	}
 		
@@ -1117,7 +1117,7 @@
 {
 	int retval = -1;
 	char *data = NULL;
-	size_t data_len = 0;
+	ssize_t data_len = 0;
 	int compressed = 0;
 	int in_fd = -1;
 
@@ -1125,7 +1125,7 @@
 		return -1;
 	}
 
-	if ((data_len = map_file(in_fd, &data, &compressed)) == 0) {
+	if ((data_len = map_file(in_fd, &data, &compressed)) <= 0) {
 		goto cleanup;
 	}
 
@@ -1187,7 +1187,7 @@
 {
 	int retval = -1;
 	char *data = NULL;
-	size_t data_len = 0;
+	ssize_t data_len = 0;
 	int compressed = 0;
 	int in_fd;
 
@@ -1195,7 +1195,7 @@
 		return -1;
 	}
 
-	if ((data_len = map_file(in_fd, &data, &compressed)) == 0) {
+	if ((data_len = map_file(in_fd, &data, &compressed)) <= 0) {
 		goto cleanup;
 	}
 		

Re: libsemanage direct_api can return errors < 0.

[Joshua Brindle]

Daniel J Walsh wrote:


Merged.

--
This message was distributed to subscribers of the selinux mailing list.
If you no longer wish to subscribe, send mail to majordomo@tycho.nsa.gov with
the words "unsubscribe selinux" without quotes as the message.

Possible OpenSSH exploit

[Tracy Reed]

[-- Attachment #1: Type: text/plain, Size: 567 bytes --]

So far just rumors but this looks potentially scary:

http://seclists.org/fulldisclosure/2009/Jul/0028.html

http://isc.sans.org/diary.html?storyid=6742

One of my biggest reasons for using SE Linux for so long has been the
hopeful mitigation of such exploits. If this is real and SE Linux
mitigates it this could be serious validation for the whole SE Linux
philosophy. If somehow we still get owned while running the latest SE
Linux confined ssh I will be very disappointed.

Hopefully it's just a false alarm.

-- 
Tracy Reed
http://tracyreed.org

[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]