Am 17.08.19 um 13:39 schrieb Chris Wilson:
> Rearrange the couple of 32-bit atomics hidden amongst the field of
> pointers that unnecessarily caused the compiler to insert some padding,
> shrinks the size of the base struct dma_fence from 80 to 72 bytes on
> x86-64.
>
> Signed-off-by: Chris Wilson <chris@xxxxxxxxxxxxxxxxxx>
> Cc: Christian König <christian.koenig@xxxxxxx>
Reviewed-by: Christian König <christian.koenig@xxxxxxx>
BTW: We could also put the timestamp in the union if we want.
E.g. the cb_list should only be used while the fence is unsignaled, the
timestamp while it is signaled and the rcu while it is freed.
Would save another 8 bytes, bringing us down to 64.
Christian.
> ---
> include/linux/dma-fence.h | 6 +++---
> 1 file changed, 3 insertions(+), 3 deletions(-)
>
> diff --git a/include/linux/dma-fence.h b/include/linux/dma-fence.h
> index 404aa748eda6..2ce4d877d33e 100644
> --- a/include/linux/dma-fence.h
> +++ b/include/linux/dma-fence.h
> @@ -63,7 +63,7 @@ struct dma_fence_cb;
> * been completed, or never called at all.
> */
> struct dma_fence {
> - struct kref refcount;
> + spinlock_t *lock;
> const struct dma_fence_ops *ops;
> /* We clear the callback list on kref_put so that by the time we
> * release the fence it is unused. No one should be adding to the cb_list
> @@ -73,11 +73,11 @@ struct dma_fence {
> struct rcu_head rcu;
> struct list_head cb_list;
> };
> - spinlock_t *lock;
> u64 context;
> u64 seqno;
> - unsigned long flags;
> ktime_t timestamp;
> + unsigned long flags;
> + struct kref refcount;
> int error;
> };
>
_______________________________________________
Intel-gfx mailing list
Intel-gfx@xxxxxxxxxxxxxxxxxxxxx
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
