You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
The following changes are done as part of this change
1) destroy_workqueue() calls flush_workqueue() internally. Hence
removed calling flush_workqueue() explicitly from the Resilient
RDMAIP module.
2) Moved unregister_ib_client() to after calling rdmaip_restore_ips().
Resilient RDMAIP module IB client callback will be invoked by the
IB core informs about the device removal for all the IB core clients.
In the callback, rdmaip_dev structure will be deallocated. It is
good to move the deallocation after rdmaip_restpore_ips() to
avoid crashed due to access to the deallocated memory.
3) In rdmaip_init() function, if creating rdmaip_garps_wq fails,
destroy_workqueue(rdmaip_wq) is called in the failure path.
This is not needed as rdmaip_cleanup() takes care of this.
Orabug: 29781216
Signed-off-by: Sudhakar Dindukurti <[email protected]>
Reviewed-by: Yuval Shaia <[email protected]>
Signed-off-by: Somasundaram Krishnasamy <[email protected]>
0 commit comments