On 09/06/2018 08:21 PM, Ben Hutchings wrote:
> On Sat, 2018-08-04 at 11:01 +0200, Greg Kroah-Hartman wrote:
>> 4.4-stable review patch. If anyone has any objections, please let me know.
>>
>> ------------------
>>
>> From: Xiao Liang <xiliang@xxxxxxxxxx>
>>
>> [ Upstream commit 822fb18a82abaf4ee7058793d95d340f5dab7bfc ]
>>
>> When loading module manually, after call xenbus_switch_state to initializes
>> the state of the netfront device, the driver state did not change so fast
>> that may lead no dev created in latest kernel. This patch adds wait to make
>> sure xenbus knows the driver is not in closed/unknown state.
> [...]
>> --- a/drivers/net/xen-netfront.c
>> +++ b/drivers/net/xen-netfront.c
>> @@ -86,6 +86,7 @@ struct netfront_cb {
>> /* IRQ name is queue name with "-tx" or "-rx" appended */
>> #define IRQ_NAME_SIZE (QUEUE_NAME_SIZE + 3)
>>
>> +static DECLARE_WAIT_QUEUE_HEAD(module_load_q);
>> static DECLARE_WAIT_QUEUE_HEAD(module_unload_q);
>>
>> struct netfront_stats {
>> @@ -1335,6 +1336,11 @@ static struct net_device *xennet_create_
>> netif_carrier_off(netdev);
>>
>> xenbus_switch_state(dev, XenbusStateInitialising);
>> + wait_event(module_load_q,
>> + xenbus_read_driver_state(dev->otherend) !=
>> + XenbusStateClosed &&
>> + xenbus_read_driver_state(dev->otherend) !=
>> + XenbusStateUnknown);
>> return netdev;
>>
>> exit:
> This won't work; it will hang. Something (maybe netback_changed()?)
> needs to wake up tasks on the module_load_q.
https://lkml.org/lkml/2018/9/7/691
-boris
