I tested the patch I sent to Paolo yesterday, and it seems to work fine. I broke it up into two pieces, so the functional change is restricted to patch #2.
Basically this fixes the case where we can invoke the blk-mq-sched put request functions in an inconsistent state. Most of the time we invoke them without any locks held, but for the case where we get a successful request-to-request merge on the back of a bio-to-request merge, we can invoke it with whatever lock the scheduler held when it called blk_mq_sched_try_merge(). -- Jens Axboe -- To unsubscribe from this list: send the line "unsubscribe linux-block" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
