mbox series

[v2,for-next,0/4] IB/hfi1: Updates and fixes for user SDMA

Message ID 20180910163707.23326.1348.stgit@scvm10.sc.intel.com (mailing list archive)
Headers show
Series IB/hfi1: Updates and fixes for user SDMA | expand

Message

Dennis Dalessandro Sept. 10, 2018, 4:38 p.m. UTC
Hi Doug and Jason,

Here is a series from Mike that fixes some issues with user SDMA. These are bit
more complicated than we'd like for RC so we can go ahead and send these to
next. The real goal of this series is to address some race conditions in user
SDMA, the other two patches are less serious issues but still need to be
addressed.

---
Changes since v1: Shuffled the order

---

Michael J. Ruhl (4):
      IB/hfi1: Eliminate races in the SDMA send error path
      IB/hfi1: Remove race conditions in user_sdma send path
      IB/hfi1: Right size user_sdma sequence numbers and related variables
      IB/hfi1: Missing return value in error path for user sdma


 drivers/infiniband/hw/hfi1/sdma.c      |    4 +
 drivers/infiniband/hw/hfi1/sdma.h      |    2 -
 drivers/infiniband/hw/hfi1/user_sdma.c |  123 +++++++++++++++-----------------
 drivers/infiniband/hw/hfi1/user_sdma.h |   20 ++---
 4 files changed, 68 insertions(+), 81 deletions(-)

--
-Denny

Comments

Jason Gunthorpe Sept. 11, 2018, 4:12 p.m. UTC | #1
On Mon, Sep 10, 2018 at 09:38:52AM -0700, Dennis Dalessandro wrote:
> Hi Doug and Jason,
> 
> Here is a series from Mike that fixes some issues with user SDMA. These are bit
> more complicated than we'd like for RC so we can go ahead and send these to
> next. The real goal of this series is to address some race conditions in user
> SDMA, the other two patches are less serious issues but still need to be
> addressed.
> 
> ---
> Changes since v1: Shuffled the order
> 
> ---
> 
> Michael J. Ruhl (4):
>       IB/hfi1: Eliminate races in the SDMA send error path
>       IB/hfi1: Remove race conditions in user_sdma send path
>       IB/hfi1: Right size user_sdma sequence numbers and related variables
>       IB/hfi1: Missing return value in error path for user sdma

Applied to for-next

Thanks,
Jason