[Qemu-devel] [PATCH v2] qemu: avoid memory leak while remove disk

[Qemu-devel] [PATCH v2] qemu: avoid memory leak while remove disk

[w00426999]

Memset vhost_dev to zero in the vhost_dev_cleanup function.
This causes dev.vqs to be NULL, so that
vqs does not free up space when calling the g_free function.
This will result in a memory leak. But you can't release vqs
directly in the vhost_dev_cleanup function, because vhost_net
will also call this function, and vhost_net's vqs is assigned by array.
In order to solve this problem, we first save the pointer of vqs,
and release the space of vqs after vhost_dev_cleanup is called.

Signed-off-by: Jian Wang <wangjian161@huawei.com>
---
 hw/block/vhost-user-blk.c | 7 +++++--
 hw/scsi/vhost-scsi.c      | 3 ++-
 hw/scsi/vhost-user-scsi.c | 3 ++-
 3 files changed, 9 insertions(+), 4 deletions(-)

diff --git a/hw/block/vhost-user-blk.c b/hw/block/vhost-user-blk.c
index 1451940..c3af28f 100644
--- a/hw/block/vhost-user-blk.c
+++ b/hw/block/vhost-user-blk.c
@@ -250,6 +250,7 @@ static void vhost_user_blk_device_realize(DeviceState *dev, Error **errp)
     VirtIODevice *vdev = VIRTIO_DEVICE(dev);
     VHostUserBlk *s = VHOST_USER_BLK(vdev);
     VhostUserState *user;
+    struct vhost_virtqueue *vqs = NULL;
     int i, ret;
 
     if (!s->chardev.chr) {
@@ -288,6 +289,7 @@ static void vhost_user_blk_device_realize(DeviceState *dev, Error **errp)
     s->dev.vqs = g_new(struct vhost_virtqueue, s->dev.nvqs);
     s->dev.vq_index = 0;
     s->dev.backend_features = 0;
+    vqs = s->dev.vqs;
 
     vhost_dev_set_config_notifier(&s->dev, &blk_ops);
 
@@ -314,7 +316,7 @@ static void vhost_user_blk_device_realize(DeviceState *dev, Error **errp)
 vhost_err:
     vhost_dev_cleanup(&s->dev);
 virtio_err:
-    g_free(s->dev.vqs);
+    g_free(vqs);
     virtio_cleanup(vdev);
 
     vhost_user_cleanup(user);
@@ -326,10 +328,11 @@ static void vhost_user_blk_device_unrealize(DeviceState *dev, Error **errp)
 {
     VirtIODevice *vdev = VIRTIO_DEVICE(dev);
     VHostUserBlk *s = VHOST_USER_BLK(dev);
+    struct vhost_virtqueue *vqs = s->dev.vqs;
 
     vhost_user_blk_set_status(vdev, 0);
     vhost_dev_cleanup(&s->dev);
-    g_free(s->dev.vqs);
+    g_free(vqs);
     virtio_cleanup(vdev);
 
     if (s->vhost_user) {
diff --git a/hw/scsi/vhost-scsi.c b/hw/scsi/vhost-scsi.c
index 7f21b4f..61e2e57 100644
--- a/hw/scsi/vhost-scsi.c
+++ b/hw/scsi/vhost-scsi.c
@@ -215,6 +215,7 @@ static void vhost_scsi_unrealize(DeviceState *dev, Error **errp)
 {
     VirtIODevice *vdev = VIRTIO_DEVICE(dev);
     VHostSCSICommon *vsc = VHOST_SCSI_COMMON(dev);
+    struct vhost_virtqueue *vqs = vsc->dev.vqs;
 
     migrate_del_blocker(vsc->migration_blocker);
     error_free(vsc->migration_blocker);
@@ -223,7 +224,7 @@ static void vhost_scsi_unrealize(DeviceState *dev, Error **errp)
     vhost_scsi_set_status(vdev, 0);
 
     vhost_dev_cleanup(&vsc->dev);
-    g_free(vsc->dev.vqs);
+    g_free(vqs);
 
     virtio_scsi_common_unrealize(dev, errp);
 }
diff --git a/hw/scsi/vhost-user-scsi.c b/hw/scsi/vhost-user-scsi.c
index 2e1ba4a..6728878 100644
--- a/hw/scsi/vhost-user-scsi.c
+++ b/hw/scsi/vhost-user-scsi.c
@@ -121,12 +121,13 @@ static void vhost_user_scsi_unrealize(DeviceState *dev, Error **errp)
     VirtIODevice *vdev = VIRTIO_DEVICE(dev);
     VHostUserSCSI *s = VHOST_USER_SCSI(dev);
     VHostSCSICommon *vsc = VHOST_SCSI_COMMON(s);
+    struct vhost_virtqueue *vqs = vsc->dev.vqs;
 
     /* This will stop the vhost backend. */
     vhost_user_scsi_set_status(vdev, 0);
 
     vhost_dev_cleanup(&vsc->dev);
-    g_free(vsc->dev.vqs);
+    g_free(vqs);
 
     virtio_scsi_common_unrealize(dev, errp);
 
-- 
1.8.3.1

Re: [Qemu-devel] [PATCH v2] qemu: avoid memory leak while remove disk

[Stefan Hajnoczi]

[-- Attachment #1: Type: text/plain, Size: 904 bytes --]

On Sat, Dec 22, 2018 at 06:27:28PM +0800, w00426999 wrote:
> Memset vhost_dev to zero in the vhost_dev_cleanup function.
> This causes dev.vqs to be NULL, so that
> vqs does not free up space when calling the g_free function.
> This will result in a memory leak. But you can't release vqs
> directly in the vhost_dev_cleanup function, because vhost_net
> will also call this function, and vhost_net's vqs is assigned by array.
> In order to solve this problem, we first save the pointer of vqs,
> and release the space of vqs after vhost_dev_cleanup is called.
> 
> Signed-off-by: Jian Wang <wangjian161@huawei.com>
> ---
>  hw/block/vhost-user-blk.c | 7 +++++--
>  hw/scsi/vhost-scsi.c      | 3 ++-
>  hw/scsi/vhost-user-scsi.c | 3 ++-
>  3 files changed, 9 insertions(+), 4 deletions(-)

This patch can go through Michael's tree.

Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 455 bytes --]

Re: [Qemu-devel] [PATCH v2] qemu: avoid memory leak while remove disk

[Eric Blake]

[-- Attachment #1: Type: text/plain, Size: 1409 bytes --]

On 12/22/18 4:27 AM, w00426999 wrote:
> Memset vhost_dev to zero in the vhost_dev_cleanup function.
> This causes dev.vqs to be NULL, so that
> vqs does not free up space when calling the g_free function.
> This will result in a memory leak. But you can't release vqs
> directly in the vhost_dev_cleanup function, because vhost_net
> will also call this function, and vhost_net's vqs is assigned by array.
> In order to solve this problem, we first save the pointer of vqs,
> and release the space of vqs after vhost_dev_cleanup is called.
> 
> Signed-off-by: Jian Wang <wangjian161@huawei.com>

It appears that this is your first qemu contribution - welcome!

Your S-o-b tag uses a different spelling "Jian Wang" than your From:
line "w00426999".  You may want to address your git config settings so
that outgoing mail sets the From: line correctly (I'm not sure which
config settings are involved; I also know that sometimes you can't
control the actual mail From: such as when sending from an address that
undergoes list munging due to SPF settings, but in that case git can be
told to insert a From: in the message body that will override any From:
header).  Fixing it now will avoid having to add a .mailmap entry to tie
the two spellings together.

-- 
Eric Blake, Principal Software Engineer
Red Hat, Inc.           +1-919-301-3226
Virtualization:  qemu.org | libvirt.org


[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 488 bytes --]