| Message ID | 1616589523-32024-1-git-send-email-zhouchuangao@vivo.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
| Series | fs/fuse/virtio_fs: Fix a potential memory allocation failure | expand |
On 3/24/21 7:38 AM, zhouchuangao wrote: > Allocate memory for struct fuse_conn may fail, we should not jump to > out_err to kfree(fc). Why not? If fc's allocation fails then it is NULL and calling kfree() on a NULL pointer is a noop[1]. Connor [1] https://www.kernel.org/doc/html/latest/core-api/mm-api.html?highlight=kfree#c.kfree
diff --git a/fs/fuse/virtio_fs.c b/fs/fuse/virtio_fs.c index 4ee6f73..1f333c6 100644 --- a/fs/fuse/virtio_fs.c +++ b/fs/fuse/virtio_fs.c @@ -1430,11 +1430,11 @@ static int virtio_fs_get_tree(struct fs_context *fsc) err = -ENOMEM; fc = kzalloc(sizeof(struct fuse_conn), GFP_KERNEL); if (!fc) - goto out_err; + goto out_err_fc; fm = kzalloc(sizeof(struct fuse_mount), GFP_KERNEL); if (!fm) - goto out_err; + goto out_err_fm; fuse_conn_init(fc, fm, get_user_ns(current_user_ns()), &virtio_fs_fiq_ops, fs); @@ -1468,8 +1468,9 @@ static int virtio_fs_get_tree(struct fs_context *fsc) fsc->root = dget(sb->s_root); return 0; -out_err: +out_err_fm: kfree(fc); +out_err_fc: mutex_lock(&virtio_fs_mutex); virtio_fs_put(fs); mutex_unlock(&virtio_fs_mutex);
Allocate memory for struct fuse_conn may fail, we should not jump to out_err to kfree(fc). Signed-off-by: zhouchuangao <zhouchuangao@vivo.com> --- fs/fuse/virtio_fs.c | 7 ++++--- 1 file changed, 4 insertions(+), 3 deletions(-)