On 06/07/2016 11:56 AM, Lukas Slebodnik wrote:
On (06/06/16 18:55), Lukas Slebodnik wrote:
On (15/03/16 12:31), Lukas Slebodnik wrote:
On (15/03/16 11:26), Pavel Březina wrote:
On 03/07/2016 01:33 PM, Lukas Slebodnik wrote:
On (07/03/16 12:12), Pavel Březina wrote:
On 03/07/2016 10:14 AM, Lukas Slebodnik wrote: > ehlo, > > simple aptch is attached.
When there, can you also talloc_free(attrs) on error? Thanks.
See updated patch
Some time has passed now so I take it as you won't implement Sumit's suggestion.
I will but I have tasks with higher priority :-)
Updated patch is attached.
LS
From e616ea9e8e58d0ed70b56edc338184d783597004 Mon Sep 17 00:00:00 2001 From: Lukas Slebodnik lslebodn@redhat.com Date: Mon, 6 Jun 2016 18:15:44 +0200 Subject: [PATCH] TOOLS: Prevent dereference of null pointer
VAR_CHECK is called with (var, EOK, ...) EOK would be returned in case of "var != EOK" and output argument _attrs would not be initialized. Therefore there could be dereference of null pointer after calling function usermod_build_attrs.
src/tools/sss_sync_ops.c | 62 +++++++++++++++++++++--------------------------- 1 file changed, 27 insertions(+), 35 deletions(-)
diff --git a/src/tools/sss_sync_ops.c b/src/tools/sss_sync_ops.c index 5468929b691c6539cdf55f59be3560412e398f21..e47aef37d2b89b28b7ff18555473136bdf7596cf 100644 --- a/src/tools/sss_sync_ops.c +++ b/src/tools/sss_sync_ops.c
- if (lock == DO_UNLOCK) {
- if (ret == EOK && lock == DO_UNLOCK) {
attr_name = SYSDB_DISABLED; /* PAM code checks for 'false' value in SYSDB_DISABLED attribute */ ret = sysdb_attrs_add_string(attrs,
SYSDB_DISABLED,
attr_name, "false");
VAR_CHECK(ret, EOK, SYSDB_DISABLED,"Could not add attribute to changeset\n");
- }
- if (ret != EOK) {
DEBUG(SSSDBG_CRIT_FAILURE,"Could not add attribute [%s] to changeset.\n", attr_name);I forgot to return error here.
} *_attrs = attrs;Upodated patch is attached.
LS
Ack.