|
|
@ -6503,10 +6503,28 @@ static void hash_sole_work(struct thr_info *mythr) |
|
|
|
static void fill_queue(struct thr_info *mythr, struct cgpu_info *cgpu, struct device_drv *drv, const int thr_id) |
|
|
|
static void fill_queue(struct thr_info *mythr, struct cgpu_info *cgpu, struct device_drv *drv, const int thr_id) |
|
|
|
{ |
|
|
|
{ |
|
|
|
do { |
|
|
|
do { |
|
|
|
|
|
|
|
bool need_work; |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
/* Do this lockless just to know if we need more unqueued work. */ |
|
|
|
|
|
|
|
need_work = (!cgpu->unqueued_work); |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
/* get_work is a blocking function so do it outside of lock
|
|
|
|
|
|
|
|
* to prevent deadlocks with other locks. */ |
|
|
|
|
|
|
|
if (need_work) { |
|
|
|
|
|
|
|
struct work *work = get_work(mythr, thr_id); |
|
|
|
|
|
|
|
|
|
|
|
wr_lock(&cgpu->qlock); |
|
|
|
wr_lock(&cgpu->qlock); |
|
|
|
if (!cgpu->unqueued_work) |
|
|
|
/* Check we haven't grabbed work somehow between
|
|
|
|
cgpu->unqueued_work = get_work(mythr, thr_id); |
|
|
|
* checking and picking up the lock. */ |
|
|
|
|
|
|
|
if (likely(!cgpu->unqueued_work)) |
|
|
|
|
|
|
|
cgpu->unqueued_work = work; |
|
|
|
|
|
|
|
else |
|
|
|
|
|
|
|
need_work = false; |
|
|
|
wr_unlock(&cgpu->qlock); |
|
|
|
wr_unlock(&cgpu->qlock); |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
if (unlikely(!need_work)) |
|
|
|
|
|
|
|
discard_work(work); |
|
|
|
|
|
|
|
} |
|
|
|
/* The queue_full function should be used by the driver to
|
|
|
|
/* The queue_full function should be used by the driver to
|
|
|
|
* actually place work items on the physical device if it |
|
|
|
* actually place work items on the physical device if it |
|
|
|
* does have a queue. */ |
|
|
|
* does have a queue. */ |
|
|
|