hw/block/xen-block.c | 2 +- qom/qom-qmp-cmds.c | 2 +- 2 files changed, 2 insertions(+), 2 deletions(-)
In qmp_object_add(), user_creatable_add_type() may set errp with some error message and
return NULL. In this case, qmp_object_add() still alloc memory to *ret_data which return
to the caller and causes a memory leak.
This patch do this alloc() action only if obj is not NULL to fix it. And initialize ret_data
in xen-block to avoid a possible uninitialized error.
The Leak stack:
Direct leak of 4120 byte(s) in 1 object(s) allocated from:
#0 0x7f6106ce5970 in __interceptor_calloc (/lib64/libasan.so.5+0xef970)
#1 0x7f6105e6a49d in g_malloc0 (/lib64/libglib-2.0.so.0+0x5249d)
#2 0x55d2c58c17fd in qdict_new /mnt/sdb/qemu-new/qemu_test/qemu/qobject/qdict.c:29
#3 0x55d2c53a0051 in qmp_object_add /mnt/sdb/qemu-new/qemu_test/qemu/qom/qom-qmp-cmds.c:291
#4 0x55d2c57b47da in do_qmp_dispatch /mnt/sdb/qemu-new/qemu_test/qemu/qapi/qmp-dispatch.c:132
#5 0x55d2c57b47da in qmp_dispatch /mnt/sdb/qemu-new/qemu_test/qemu/qapi/qmp-dispatch.c:175
#6 0x55d2c52f1430 in monitor_qmp_dispatch /mnt/sdb/qemu-new/qemu_test/qemu/monitor/qmp.c:145
#7 0x55d2c52f3087 in monitor_qmp_bh_dispatcher /mnt/sdb/qemu-new/qemu_test/qemu/monitor/qmp.c:234
#8 0x55d2c58e6153 in aio_bh_call /mnt/sdb/qemu-new/qemu_test/qemu/util/async.c:136
Fixes: 5f07c4d60d091320186e7b0edaf9ed2cc16b2d1e
Reported-by: Euler Robot <euler.robot@huawei.com>
Signed-off-by: Pan Nengyuan <pannengyuan@huawei.com>
---
Cc: Kevin Wolf <kwolf@redhat.com>
---
hw/block/xen-block.c | 2 +-
qom/qom-qmp-cmds.c | 2 +-
2 files changed, 2 insertions(+), 2 deletions(-)
diff --git a/hw/block/xen-block.c b/hw/block/xen-block.c
index 3885464513..041866b846 100644
--- a/hw/block/xen-block.c
+++ b/hw/block/xen-block.c
@@ -860,7 +860,7 @@ static XenBlockIOThread *xen_block_iothread_create(const char *id,
XenBlockIOThread *iothread = g_new(XenBlockIOThread, 1);
Error *local_err = NULL;
QDict *opts;
- QObject *ret_data;
+ QObject *ret_data = NULL;
iothread->id = g_strdup(id);
diff --git a/qom/qom-qmp-cmds.c b/qom/qom-qmp-cmds.c
index 435193b036..6bd137ccbf 100644
--- a/qom/qom-qmp-cmds.c
+++ b/qom/qom-qmp-cmds.c
@@ -287,8 +287,8 @@ void qmp_object_add(QDict *qdict, QObject **ret_data, Error **errp)
visit_free(v);
if (obj) {
object_unref(obj);
+ *ret_data = QOBJECT(qdict_new());
}
- *ret_data = QOBJECT(qdict_new());
}
void qmp_object_del(const char *id, Error **errp)
--
2.18.2
Pan Nengyuan <pannengyuan@huawei.com> writes: > In qmp_object_add(), user_creatable_add_type() may set errp with some error message and > return NULL. In this case, qmp_object_add() still alloc memory to *ret_data which return > to the caller and causes a memory leak. > > This patch do this alloc() action only if obj is not NULL to fix it. And initialize ret_data > in xen-block to avoid a possible uninitialized error. > > The Leak stack: > Direct leak of 4120 byte(s) in 1 object(s) allocated from: > #0 0x7f6106ce5970 in __interceptor_calloc (/lib64/libasan.so.5+0xef970) > #1 0x7f6105e6a49d in g_malloc0 (/lib64/libglib-2.0.so.0+0x5249d) > #2 0x55d2c58c17fd in qdict_new /mnt/sdb/qemu-new/qemu_test/qemu/qobject/qdict.c:29 > #3 0x55d2c53a0051 in qmp_object_add /mnt/sdb/qemu-new/qemu_test/qemu/qom/qom-qmp-cmds.c:291 > #4 0x55d2c57b47da in do_qmp_dispatch /mnt/sdb/qemu-new/qemu_test/qemu/qapi/qmp-dispatch.c:132 > #5 0x55d2c57b47da in qmp_dispatch /mnt/sdb/qemu-new/qemu_test/qemu/qapi/qmp-dispatch.c:175 > #6 0x55d2c52f1430 in monitor_qmp_dispatch /mnt/sdb/qemu-new/qemu_test/qemu/monitor/qmp.c:145 > #7 0x55d2c52f3087 in monitor_qmp_bh_dispatcher /mnt/sdb/qemu-new/qemu_test/qemu/monitor/qmp.c:234 > #8 0x55d2c58e6153 in aio_bh_call /mnt/sdb/qemu-new/qemu_test/qemu/util/async.c:136 > > Fixes: 5f07c4d60d091320186e7b0edaf9ed2cc16b2d1e > Reported-by: Euler Robot <euler.robot@huawei.com> > Signed-off-by: Pan Nengyuan <pannengyuan@huawei.com> Reviewed-by: Markus Armbruster <armbru@redhat.com>
Paolo, looks like this has fallen through the cracks. If you'd prefer me to take it, let me know. Markus Armbruster <armbru@redhat.com> writes: > Pan Nengyuan <pannengyuan@huawei.com> writes: > >> In qmp_object_add(), user_creatable_add_type() may set errp with some error message and >> return NULL. In this case, qmp_object_add() still alloc memory to *ret_data which return >> to the caller and causes a memory leak. >> >> This patch do this alloc() action only if obj is not NULL to fix it. And initialize ret_data >> in xen-block to avoid a possible uninitialized error. >> >> The Leak stack: >> Direct leak of 4120 byte(s) in 1 object(s) allocated from: >> #0 0x7f6106ce5970 in __interceptor_calloc (/lib64/libasan.so.5+0xef970) >> #1 0x7f6105e6a49d in g_malloc0 (/lib64/libglib-2.0.so.0+0x5249d) >> #2 0x55d2c58c17fd in qdict_new /mnt/sdb/qemu-new/qemu_test/qemu/qobject/qdict.c:29 >> #3 0x55d2c53a0051 in qmp_object_add /mnt/sdb/qemu-new/qemu_test/qemu/qom/qom-qmp-cmds.c:291 >> #4 0x55d2c57b47da in do_qmp_dispatch /mnt/sdb/qemu-new/qemu_test/qemu/qapi/qmp-dispatch.c:132 >> #5 0x55d2c57b47da in qmp_dispatch /mnt/sdb/qemu-new/qemu_test/qemu/qapi/qmp-dispatch.c:175 >> #6 0x55d2c52f1430 in monitor_qmp_dispatch /mnt/sdb/qemu-new/qemu_test/qemu/monitor/qmp.c:145 >> #7 0x55d2c52f3087 in monitor_qmp_bh_dispatcher /mnt/sdb/qemu-new/qemu_test/qemu/monitor/qmp.c:234 >> #8 0x55d2c58e6153 in aio_bh_call /mnt/sdb/qemu-new/qemu_test/qemu/util/async.c:136 >> >> Fixes: 5f07c4d60d091320186e7b0edaf9ed2cc16b2d1e >> Reported-by: Euler Robot <euler.robot@huawei.com> >> Signed-off-by: Pan Nengyuan <pannengyuan@huawei.com> > > Reviewed-by: Markus Armbruster <armbru@redhat.com>
On 06/04/20 08:34, Markus Armbruster wrote:
> Paolo, looks like this has fallen through the cracks. If you'd prefer
> me to take it, let me know.
>
> Markus Armbruster <armbru@redhat.com> writes:
Actually it was in my latest pull request, but between this version,
Marc-André and mine you might have missed it:
commit 7f5d9b206d1e86425faa5b84b551068bf044b823
Author: Paolo Bonzini <pbonzini@redhat.com>
Date: Thu Mar 26 10:41:21 2020 +0100
object-add: don't create return value if failed
No need to return an empty value from object-add (it would also leak
if the command failed). While at it, remove the "if" around object_unref
since object_unref handles NULL arguments just fine.
Reported-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Message-Id: <20200325184723.2029630-4-marcandre.lureau@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Thanks,
Paolo
© 2016 - 2024 Red Hat, Inc.