On Wed, 2018-05-02 at 06:42 -0700, Dennis Dalessandro wrote:
Hi Doug and Jason,
Here are some patches to go to for-next. These include the couple patches that needed rework that were posted before the OFA conf. Well actually those patches that had issues were just dropped with the exception of the one from Alex, to add handling of kernel restart to hfi1 and qib. Patch 8 is his V2.
Nothing else too scary or exciting in here. Well OK so that's not quite right the CQ completion vector patch is rather interesting. This adds support for compeltion vectors for hfi1 and helps improve performance in things like IPoIB.
There is a signifianct patch from Mitko that redoes a lof our fault injection stuff. It's a big patch but I'm not sure it lends itself to being broken up further.
One other thing of note is the "Create common functions" patch from Sebastian depends on one of the patches that I sent for the -rc. It won't apply cleanly without that.
Alex Estrin (2): IB/hfi1: Complete check for locally terminated smp IB/{hfi1,qib}: Add handling of kernel restart
Brian Welty (1): IB/{hfi1,qib,rdmavt}: Move logic to allocate receive WQE into rdmavt
Kamenee Arumugam (1): IB/Hfi1: Read CCE Revision register to verify the device is responsive
Michael J. Ruhl (4): IB/hfi1: Return actual error value from program_rcvarray() IB/hfi1: Use after free race condition in send context error path IB/hfi1: Return correct value for device state IB/hfi1: Reorder incorrect send context disable
Mike Marciniszyn (1): IB/hfi1: Fix fault injection init/exit issues
Mitko Haralanov (1): IB/hfi1: Rework fault injection machinery
Sebastian Sanchez (4): IB/hfi1: Prevent LNI hang when LCB can't obtain lanes IB/hfi1: Optimize kthread pointer locking when queuing CQ entries IB/hfi1: Create common functions for affinity CPU mask operations IB/{hfi1,rdmavt,qib}: Implement CQ completion vector support
Hi Denny,
As mentioned, I pulled patch 5 of this series into for-rc, and I pulled the rest into for-next last week. Thanks.