[for-rc,0/4] RDMA/bnxt_re: Bug Fixes

Message

Selvin Xavier Sept. 20, 2018, 3:02 p.m. UTC

Few bug fixes for bnxt_re

Patch 1 and 2 are fixes for two smatch warnings.
Patch 3 fix some of the resource leaks in error path.
Patch 4 changes the rtnl locking during registration to avoid system crash

Please review and apply

Selvin Xavier (4):
  RDMA/bnxt_re: Avoid null check after accessing the pointer
  RDMA/bnxt_re: Avoid accessing nq->bar_reg_iomem in failure case
  RDMA/bnxt_re: Avoid resource leak in case the NQ registration fails
  RDMA/bnxt_re: Handle all RDMA resource initialization under rtnl_lock

 drivers/infiniband/hw/bnxt_re/bnxt_re.h    |   2 +
 drivers/infiniband/hw/bnxt_re/main.c       | 122 ++++++++++++++---------------
 drivers/infiniband/hw/bnxt_re/qplib_fp.c   |   3 +-
 drivers/infiniband/hw/bnxt_re/qplib_rcfw.c |  13 ++-
 4 files changed, 69 insertions(+), 71 deletions(-)

Comments

Jason Gunthorpe Sept. 20, 2018, 3:58 p.m. UTC | #1

On Thu, Sep 20, 2018 at 08:02:31AM -0700, Selvin Xavier wrote:
> Few bug fixes for bnxt_re
> 
> Patch 1 and 2 are fixes for two smatch warnings.
> Patch 3 fix some of the resource leaks in error path.
> Patch 4 changes the rtnl locking during registration to avoid system crash
> 
> Please review and apply

These patches all need better commit messages for the rc4 timeframe,
especially since some of them are so long.

rc patches need to be fixing real bug, include a oops report or
clearly explain how the bug can be triggered. Read the

 Documentation/process/stable-kernel-rules.rst

File for guidlines.

also you should cc Dan on the patches you credited to him..

Jason

Selvin Xavier Sept. 20, 2018, 4:26 p.m. UTC | #2

On Thu, Sep 20, 2018 at 9:28 PM Jason Gunthorpe <[e-mail]> wrote:
>
> On Thu, Sep 20, 2018 at 08:02:31AM -0700, Selvin Xavier wrote:
> > Few bug fixes for bnxt_re
> >
> > Patch 1 and 2 are fixes for two smatch warnings.
> > Patch 3 fix some of the resource leaks in error path.
> > Patch 4 changes the rtnl locking during registration to avoid system crash
> >
> > Please review and apply
>
> These patches all need better commit messages for the rc4 timeframe,
> especially since some of them are so long.
>
> rc patches need to be fixing real bug, include a oops report or
> clearly explain how the bug can be triggered. Read the
>
>  Documentation/process/stable-kernel-rules.rst
>
> File for guidlines.
>
 Patch 4 is a fix for oops. I feel this is needed for RC, even though
its bit longer.
I am ok with moving Patches 1, 2 and 3 to for-next as they are not
seen n the normal execution.
I will post them for for-next.
I will drop this patch series and post a single patch for for-rc.
Rework required for this patch
as it is dependent on patch 3 now. I will update the commit message also.
> also you should cc Dan on the patches you credited to him..
Sure.. i missed it
>
> Jason

Jason Gunthorpe Sept. 20, 2018, 4:29 p.m. UTC | #3

On Thu, Sep 20, 2018 at 09:56:09PM +0530, Selvin Xavier wrote:
> On Thu, Sep 20, 2018 at 9:28 PM Jason Gunthorpe <[e-mail]> wrote:
> >
> > On Thu, Sep 20, 2018 at 08:02:31AM -0700, Selvin Xavier wrote:
> > > Few bug fixes for bnxt_re
> > >
> > > Patch 1 and 2 are fixes for two smatch warnings.
> > > Patch 3 fix some of the resource leaks in error path.
> > > Patch 4 changes the rtnl locking during registration to avoid system crash
> > >
> > > Please review and apply
> >
> > These patches all need better commit messages for the rc4 timeframe,
> > especially since some of them are so long.
> >
> > rc patches need to be fixing real bug, include a oops report or
> > clearly explain how the bug can be triggered. Read the
> >
> >  Documentation/process/stable-kernel-rules.rst
> >
> > File for guidlines.
> >
>  Patch 4 is a fix for oops. I feel this is needed for RC, even though
> its bit longer.

Include the oops message, it is much easier to justify that way.

Jason