I just found a large memory leak in that patch. This patch fixes it.
I'll add it to the BZ
Signed-Off-By: Christine Caulfield <ccaulfie@xxxxxxxxxx>
On 10/08/12 09:14, Fabio M. Di Nitto wrote:
Patch looks good to me.
ACK
Fabio
On 8/10/2012 10:00 AM, Christine Caulfield wrote:
confdb as it stands has a limit of 256 bytes in a returned value - anything larger than that will cause corosync to crash.
This patch adds checks toprevent this happening, fixes some API calls so that they return the full value and adds some new ones where it wasn't possible to safely return am arbitrarily long string.
It refers to Red Hat Bugzilla #845626
include/corosync/confdb.h | 18 +++
include/corosync/ipc_confdb.h | 54 +++++++++
lib/confdb.c | 232 +++++++++++++++++++++++++++++++++++++-----
lib/sa-confdb.c | 33 ++++-
lib/sa-confdb.h | 6 -
services/confdb.c | 215 ++++++++++++++++++++++++++++++++++----
tools/corosync-objctl.c | 27 +++-
7 files changed, 514 insertions(+), 71 deletions(-)
Signed-Off-By: Christine Caulfield <ccaulfie@xxxxxxxxxx>
_______________________________________________
discuss mailing list
discuss@xxxxxxxxxxxx
http://lists.corosync.org/mailman/listinfo/discuss
_______________________________________________
discuss mailing list
discuss@xxxxxxxxxxxx
http://lists.corosync.org/mailman/listinfo/discuss
diff --git a/lib/confdb.c b/lib/confdb.c
index 0dbecb9..ba4caf8 100644
--- a/lib/confdb.c
+++ b/lib/confdb.c
@@ -1115,6 +1115,7 @@ cs_error_t confdb_key_get_typed2 (
*value_len = response->value_length;
*type = response->type;
}
+ coroipcc_msg_send_reply_receive_in_buf_put(confdb_inst->handle);
error_exit:
(void)hdb_handle_put (&confdb_handle_t_db, handle);
@@ -1778,6 +1779,7 @@ cs_error_t confdb_key_iter_typed2 (
*value_len = response->value_length;
*type = response->type;
}
+ coroipcc_msg_send_reply_receive_in_buf_put(confdb_inst->handle);
sa_exit:
context->next_entry++;
_______________________________________________
discuss mailing list
discuss@xxxxxxxxxxxx
http://lists.corosync.org/mailman/listinfo/discuss