On Sun, Sep 06, 2020 at 07:58:12PM -0700, Bart Van Assche wrote:
> The following appeared:
>
> WARNING: CPU: 5 PID: 1760 at drivers/infiniband/core/device.c:335 ib_device_put+0xf2/0x100 [ib_core]
>
> Call Trace:
> rxe_elem_release+0x76/0x90 [rdma_rxe]
> rxe_destroy_cq+0x4f/0x70 [rdma_rxe]
> ib_free_cq_user+0x12b/0x2b0 [ib_core]
> ib_cq_pool_destroy+0xa8/0x140 [ib_core]
> __ib_unregister_device+0x9c/0x1c0 [ib_core]
> ib_unregister_driver+0x181/0x1a0 [ib_core]
> rxe_module_exit+0x31/0x50 [rdma_rxe]
> __x64_sys_delete_module+0x22a/0x310
> do_syscall_64+0x36/0x80
> entry_SYSCALL_64_after_hwframe+0x44/0xa9
Oh interesting..
> Do you agree that the above proves that the hang-on-unload is a
> regression that has been introduced by the cq-pool patches? Is the patch
> below a good way to fix this?
> +++ b/drivers/infiniband/core/device.c
> @@ -1454,8 +1454,8 @@ static void __ib_unregister_device(struct ib_device *ib_dev)
> if (!refcount_read(&ib_dev->refcount))
> goto out;
>
> - disable_device(ib_dev);
> ib_cq_pool_destroy(ib_dev);
> + disable_device(ib_dev);
>
> /* Expedite removing unregistered pointers from the hash table */
> free_netdevs(ib_dev);
Hum. Not quite..
disable_device() ensures that all ib_clients have disconnected from
the device, up until the clients have disconnected the cq_pool must
remain operational.
It is reasonable to consider the cq_pool as a built-in client, so I
would suggest moving it to right around the time the dynamic clients
are handled. Something like this:
diff --git a/drivers/infiniband/core/device.c b/drivers/infiniband/core/device.c
index c36b4d2b61e0c0..e3651dacad1da6 100644
--- a/drivers/infiniband/core/device.c
+++ b/drivers/infiniband/core/device.c
@@ -1285,6 +1285,8 @@ static void disable_device(struct ib_device *device)
remove_client_context(device, cid);
}
+ ib_cq_pool_destroy(ib_dev);
+
/* Pairs with refcount_set in enable_device */
ib_device_put(device);
wait_for_completion(&device->unreg_completion);
@@ -1328,6 +1330,8 @@ static int enable_device_and_get(struct ib_device *device)
goto out;
}
+ ib_cq_pool_init(device);
+
down_read(&clients_rwsem);
xa_for_each_marked (&clients, index, client, CLIENT_REGISTERED) {
ret = add_client_context(device, client);
@@ -1400,7 +1404,6 @@ int ib_register_device(struct ib_device *device, const char *name)
goto dev_cleanup;
}
- ib_cq_pool_init(device);
ret = enable_device_and_get(device);
dev_set_uevent_suppress(&device->dev, false);
/* Mark for userspace that device is ready */
@@ -1455,7 +1458,6 @@ static void __ib_unregister_device(struct ib_device *ib_dev)
goto out;
disable_device(ib_dev);
- ib_cq_pool_destroy(ib_dev);
/* Expedite removing unregistered pointers from the hash table */
free_netdevs(ib_dev);
