2018-05-10 09:43:50

by jianchao.wang

Subject: [PATCH V2] nvme-rdma: stop queue first before free it in config admin queue

When any of cases after nvme_rdma_start_queue in
nvme_rdma_configure_admin_queue fails, the ctrl->queues[0] will be
freed but the NVME_RDMA_Q_LIVE is still set. If nvme_rdma_stop_queue
is invoked, we will incur use-after-free which will cause memory
corruption.
BUG: KASAN: use-after-free in rdma_disconnect+0x1f/0xe0 [rdma_cm]
Read of size 8 at addr ffff8801dc3969c0 by task kworker/u16:3/9304

CPU: 3 PID: 9304 Comm: kworker/u16:3 Kdump: loaded Tainted: G W 4.17.0-rc3+ #20
Workqueue: nvme-delete-wq nvme_delete_ctrl_work
Call Trace:
dump_stack+0x91/0xeb
print_address_description+0x6b/0x290
kasan_report+0x261/0x360
rdma_disconnect+0x1f/0xe0 [rdma_cm]
nvme_rdma_stop_queue+0x25/0x40 [nvme_rdma]
nvme_rdma_shutdown_ctrl+0xf3/0x150 [nvme_rdma]
nvme_delete_ctrl_work+0x98/0xe0
process_one_work+0x3ca/0xaa0
worker_thread+0x4e2/0x6c0
kthread+0x18d/0x1e0
ret_from_fork+0x24/0x30

To fix it, call nvme_rdma_stop_queue for all the failed cases after
nvme_rdma_start_queue.

Signed-off-by: Jianchao Wang <[email protected]>
---

V2:
based on Sagi's suggestion, add out_stop_queue lable and invoke
nvme_rdma_stop_queue in all the failed cases after nvme_rdma_start_queue

drivers/nvme/host/rdma.c | 10 ++++++----
1 file changed, 6 insertions(+), 4 deletions(-)

diff --git a/drivers/nvme/host/rdma.c b/drivers/nvme/host/rdma.c
index a0ead1d..966e0dd 100644
--- a/drivers/nvme/host/rdma.c
+++ b/drivers/nvme/host/rdma.c
@@ -777,7 +777,7 @@ static int nvme_rdma_configure_admin_queue(struct nvme_rdma_ctrl *ctrl,
if (error) {
dev_err(ctrl->ctrl.device,
"prop_get NVME_REG_CAP failed\n");
- goto out_cleanup_queue;
+ goto out_stop_queue;
}

ctrl->ctrl.sqsize =
@@ -785,23 +785,25 @@ static int nvme_rdma_configure_admin_queue(struct nvme_rdma_ctrl *ctrl,

error = nvme_enable_ctrl(&ctrl->ctrl, ctrl->ctrl.cap);
if (error)
- goto out_cleanup_queue;
+ goto out_stop_queue;

ctrl->ctrl.max_hw_sectors =
(ctrl->max_fr_pages - 1) << (ilog2(SZ_4K) - 9);

error = nvme_init_identify(&ctrl->ctrl);
if (error)
- goto out_cleanup_queue;
+ goto out_stop_queue;

error = nvme_rdma_alloc_qe(ctrl->queues[0].device->dev,
&ctrl->async_event_sqe, sizeof(struct nvme_command),
DMA_TO_DEVICE);
if (error)
- goto out_cleanup_queue;
+ goto out_stop_queue;

return 0;

+out_stop_queue:
+ nvme_rdma_stop_queue(&ctrl->queues[0]);
out_cleanup_queue:
if (new)
blk_cleanup_queue(ctrl->ctrl.admin_q);
--
2.7.4

2018-05-16 12:07:29

by Max Gurtovoy

Subject: Re: [PATCH V2] nvme-rdma: stop queue first before free it in config admin queue

On 5/10/2018 12:42 PM, Jianchao Wang wrote:
> When any of cases after nvme_rdma_start_queue in
> nvme_rdma_configure_admin_queue fails, the ctrl->queues[0] will be
> freed but the NVME_RDMA_Q_LIVE is still set. If nvme_rdma_stop_queue
> is invoked, we will incur use-after-free which will cause memory
> corruption.
> BUG: KASAN: use-after-free in rdma_disconnect+0x1f/0xe0 [rdma_cm]
> Read of size 8 at addr ffff8801dc3969c0 by task kworker/u16:3/9304
>
> CPU: 3 PID: 9304 Comm: kworker/u16:3 Kdump: loaded Tainted: G W 4.17.0-rc3+ #20
> Workqueue: nvme-delete-wq nvme_delete_ctrl_work
> Call Trace:
> dump_stack+0x91/0xeb
> print_address_description+0x6b/0x290
> kasan_report+0x261/0x360
> rdma_disconnect+0x1f/0xe0 [rdma_cm]
> nvme_rdma_stop_queue+0x25/0x40 [nvme_rdma]
> nvme_rdma_shutdown_ctrl+0xf3/0x150 [nvme_rdma]
> nvme_delete_ctrl_work+0x98/0xe0
> process_one_work+0x3ca/0xaa0
> worker_thread+0x4e2/0x6c0
> kthread+0x18d/0x1e0
> ret_from_fork+0x24/0x30
>
> To fix it, call nvme_rdma_stop_queue for all the failed cases after
> nvme_rdma_start_queue.
>
> Signed-off-by: Jianchao Wang <[email protected]>
> ---
>
> V2:
> based on Sagi's suggestion, add out_stop_queue lable and invoke
> nvme_rdma_stop_queue in all the failed cases after nvme_rdma_start_queue
>

Looks good,

Reviewed-by: Max Gurtovoy <[email protected]>