On 02/21/2017 11:59 AM, Jon Derrick wrote: > v3->v4: > Passes in 'lock_held' into opal_lock_unlock() so we don't need to > mutex_trylock(). I wasn't totally confident in that approach anyways. > > v2->v3: > Squashed 5/5 into 4/5 > Changed opal_step structs back to const > Cleaned up opal_lock_unlock > > v1->v2: > Moved misplaced code from 5/5 to 4/5 Guys, where are we at with this? I'm shipping off some more fixes for -rc1 soon, and I'd really like to get this wrapped up sooner rather than later. -- Jens Axboe
