summaryrefslogtreecommitdiff
path: root/source4/torture/auth/pac.c
diff options
context:
space:
mode:
Diffstat (limited to 'source4/torture/auth/pac.c')
-rw-r--r--source4/torture/auth/pac.c10
1 files changed, 6 insertions, 4 deletions
diff --git a/source4/torture/auth/pac.c b/source4/torture/auth/pac.c
index 21c3119e31..59bf008970 100644
--- a/source4/torture/auth/pac.c
+++ b/source4/torture/auth/pac.c
@@ -215,6 +215,9 @@ static BOOL torture_pac_saved_check(void)
return False;
}
+ /* The krbtgt key in use when the above PAC was generated.
+ * This is an arcfour-hmac-md5 key, extracted with our 'net
+ * samdump' tool. */
krbtgt_bytes = smbpasswd_gethexpwd(mem_ctx, "B286757148AF7FD252C53603A150B7E7");
if (!krbtgt_bytes) {
DEBUG(0, ("Could not interpret krbtgt key"));
@@ -296,10 +299,9 @@ static BOOL torture_pac_saved_check(void)
dump_data(10,validate_blob.data,validate_blob.length);
- /* all we can check is the length of the buffers,
- * to check that the alignment and padding is ok,
- * we can't compare the bytes, because we use a different algorithm
- * to create the pointer values
+ /* compare both the length and the data bytes after a
+ * pull/push cycle. This ensures we use the exact same
+ * pointer, padding etc algorithms as win2k3.
*/
if (tmp_blob.length != validate_blob.length) {
DEBUG(0, ("PAC push failed: orignial buffer length[%u] != created buffer length[%u]\n",