IB/uverbs: Clarify the kref'ing ordering for alloc_commit

Browse source

The alloc_commit callback makes the uobj visible to other threads,
and it does so using a 'move' semantic of the uobj kref on the stack
into the public storage (eg the IDR, uobject list and file_private_data)

Once this is done another thread could start up and trigger deletion
of the kref. Fortunately cleanup_rwsem happens to prevent this from
being a bug, but that is a fantastically unclear side effect.

Re-organize things so that alloc_commit is that last thing to touch
the uobj, get rid of the sneaky implicit dependency on cleanup_rwsem,
and add a comment reminding that uobj is no longer kref'd after
alloc_commit.

Signed-off-by: Jason Gunthorpe <jgg@mellanox.com>

...

This commit is contained in:

Jason Gunthorpe 2018-07-10 20:55:15 -06:00

parent 1250c3048c

commit c561c28846

1 changed files with 22 additions and 4 deletions

Show all changes Ignore whitespace when comparing lines Ignore changes in amount of whitespace Ignore changes in whitespace at EOL

Download patch file Download diff file Expand all files Collapse all files

26

drivers/infiniband/core/rdma_core.c

View file