Re: [libvirt] [PATCH] openvz_driver.c: avoid leak on OOM error path
Matthias Bolte wrote:
...
Why do you think openvzFreeDriver must not VIR_FREE the driver
object?
The driver object is allocated per connection in openvzOpen, so
openvzClose has to VIR_FREE it.
I was lazy, and assumed there wouldn't be so many leaks.
I should have looked.
Here's a better patch. Thanks for keeping me honest.
From 2b865b8f955d88bb687a429efd636dd2212cd13b Mon Sep 17 00:00:00 2001
From: Jim Meyering <meyering(a)redhat.com>
Date: Mon, 15 Feb 2010 18:33:38 +0100
Subject: [PATCH] openvz (openvzFreeDriver): avoid leaks
* src/openvz/openvz_conf.c (openvzFreeDriver): Also free driver buffer.
Based on a suggestion from Matthias Bolte.
---
src/openvz/openvz_conf.c | 1 +
1 files changed, 1 insertions(+), 0 deletions(-)
diff --git a/src/openvz/openvz_conf.c b/src/openvz/openvz_conf.c
index f4b8199..51dbde5 100644
--- a/src/openvz/openvz_conf.c
+++ b/src/openvz/openvz_conf.c
@@ -427,6 +427,7 @@ openvzFreeDriver(struct openvz_driver *driver)
virDomainObjListDeinit(&driver->domains);
virCapabilitiesFree(driver->caps);
+ VIR_FREE(driver);
}
--
1.7.0.181.g41533