diff options
author | Jason Gunthorpe <jgg@mellanox.com> | 2018-07-10 20:55:15 -0600 |
---|---|---|
committer | Jason Gunthorpe <jgg@mellanox.com> | 2018-07-25 14:21:21 -0600 |
commit | c561c288463102b12c9089a42c6c2a9f55c4fb53 (patch) | |
tree | 4264c61e2b4327f89977b9c555f5814a636fb9db /drivers/infiniband/core/cgroup.c | |
parent | IB/uverbs: Handle IDR and FD types without truncation (diff) | |
download | linux-dev-c561c288463102b12c9089a42c6c2a9f55c4fb53.tar.xz linux-dev-c561c288463102b12c9089a42c6c2a9f55c4fb53.zip |
IB/uverbs: Clarify the kref'ing ordering for alloc_commit
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>
Diffstat (limited to 'drivers/infiniband/core/cgroup.c')
0 files changed, 0 insertions, 0 deletions