Revision be4b499323bf7291b491c6df51baae62f45b8404 authored by Erez Shitrit on 25 June 2015, 14:13:22 UTC, committed by Doug Ledford on 14 July 2015, 17:20:09 UTC
Whenever ib_cm gets remove_one call, like when there is a hot-unplug event, the driver should mark itself as going_down and confirm that no new works are going to be queued for that device. so, the order of the actions are: 1. mark the going_down bit. 2. flush the wq. 3. [make sure no new works for that device.] 4. unregister mad agent. otherwise, works that are already queued can be scheduled after the mad agent was freed. Signed-off-by: Erez Shitrit <erezsh@mellanox.com> Signed-off-by: Doug Ledford <dledford@redhat.com>
1 parent 3fdf70a
| File | Mode | Size |
|---|---|---|
| cache-policies.txt | -rw-r--r-- | 6.6 KB |
| cache.txt | -rw-r--r-- | 12.1 KB |
| delay.txt | -rw-r--r-- | 694 bytes |
| dm-crypt.txt | -rw-r--r-- | 3.4 KB |
| dm-flakey.txt | -rw-r--r-- | 1.9 KB |
| dm-io.txt | -rw-r--r-- | 3.2 KB |
| dm-log.txt | -rw-r--r-- | 2.3 KB |
| dm-queue-length.txt | -rw-r--r-- | 1.2 KB |
| dm-raid.txt | -rw-r--r-- | 9.7 KB |
| dm-service-time.txt | -rw-r--r-- | 3.2 KB |
| dm-uevent.txt | -rw-r--r-- | 2.6 KB |
| era.txt | -rw-r--r-- | 3.1 KB |
| kcopyd.txt | -rw-r--r-- | 1.9 KB |
| linear.txt | -rw-r--r-- | 1.4 KB |
| log-writes.txt | -rw-r--r-- | 4.7 KB |
| persistent-data.txt | -rw-r--r-- | 2.9 KB |
| snapshot.txt | -rw-r--r-- | 6.4 KB |
| statistics.txt | -rw-r--r-- | 7.4 KB |
| striped.txt | -rw-r--r-- | 1.8 KB |
| switch.txt | -rw-r--r-- | 5.4 KB |
| thin-provisioning.txt | -rw-r--r-- | 14.0 KB |
| verity.txt | -rw-r--r-- | 6.4 KB |
| zero.txt | -rw-r--r-- | 1.6 KB |
Computing file changes ...
