[Asterisk-code-review] VECTOR: Passing parameters with side effects to macros is da... (asterisk[13])
Richard Mudgett
asteriskteam at digium.com
Thu Jun 21 17:10:25 CDT 2018
Richard Mudgett has uploaded this change for review. ( https://gerrit.asterisk.org/9273
Change subject: VECTOR: Passing parameters with side effects to macros is dangerous.
......................................................................
VECTOR: Passing parameters with side effects to macros is dangerous.
* Fix several instances where we were bumping a ref in the parameter and
then unrefing the object if it failed. The way the AST_VECTOR_APPEND()
and AST_VECTOR_REPLACE() macros are implemented means if it fails the new
value was never evaluated.
Change-Id: I2847872a455b11ea7e5b7ce697c0a455a1d0ac9a
---
M res/res_pjsip/pjsip_options.c
M res/res_pjsip_history.c
M res/stasis/messaging.c
3 files changed, 8 insertions(+), 5 deletions(-)
git pull ssh://gerrit.asterisk.org:29418/asterisk refs/changes/73/9273/1
diff --git a/res/res_pjsip/pjsip_options.c b/res/res_pjsip/pjsip_options.c
index 2229664..891de89 100644
--- a/res/res_pjsip/pjsip_options.c
+++ b/res/res_pjsip/pjsip_options.c
@@ -1530,10 +1530,11 @@
ast_debug(3, "Adding endpoint compositor '%s' to AOR '%s'\n",
task_data->endpoint_state_compositor->name, task_data->aor_options->name);
+ ao2_ref(task_data->endpoint_state_compositor, +1);
if (AST_VECTOR_APPEND(&task_data->aor_options->compositors,
- ao2_bump(task_data->endpoint_state_compositor))) {
+ task_data->endpoint_state_compositor)) {
/* Failed to add so no need to update the endpoint status. Nothing changed. */
- ao2_cleanup(task_data->endpoint_state_compositor);
+ ao2_ref(task_data->endpoint_state_compositor, -1);
return 0;
}
diff --git a/res/res_pjsip_history.c b/res/res_pjsip_history.c
index 8f44d39..e6e9ac2 100644
--- a/res/res_pjsip_history.c
+++ b/res/res_pjsip_history.c
@@ -1135,7 +1135,8 @@
} else if (!res) {
continue;
} else {
- if (AST_VECTOR_APPEND(output, ao2_bump(entry))) {
+ ao2_bump(entry);
+ if (AST_VECTOR_APPEND(output, entry)) {
ao2_cleanup(entry);
}
}
diff --git a/res/stasis/messaging.c b/res/stasis/messaging.c
index 979b687..4a86d58 100644
--- a/res/stasis/messaging.c
+++ b/res/stasis/messaging.c
@@ -459,8 +459,9 @@
ao2_link(endpoint_subscriptions, sub);
} else {
ast_rwlock_wrlock(&tech_subscriptions_lock);
- if (AST_VECTOR_APPEND(&tech_subscriptions, ao2_bump(sub))) {
- /* Release the ao2_bump that was for the vector and allocation references. */
+ ao2_ref(sub, +1);
+ if (AST_VECTOR_APPEND(&tech_subscriptions, sub)) {
+ /* Release the refs that were for the vector and the allocation. */
ao2_ref(sub, -2);
sub = NULL;
}
--
To view, visit https://gerrit.asterisk.org/9273
To unsubscribe, visit https://gerrit.asterisk.org/settings
Gerrit-Project: asterisk
Gerrit-Branch: 13
Gerrit-MessageType: newchange
Gerrit-Change-Id: I2847872a455b11ea7e5b7ce697c0a455a1d0ac9a
Gerrit-Change-Number: 9273
Gerrit-PatchSet: 1
Gerrit-Owner: Richard Mudgett <rmudgett at digium.com>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.digium.com/pipermail/asterisk-code-review/attachments/20180621/81c1daf5/attachment-0001.html>
More information about the asterisk-code-review
mailing list