Re: [PATCH] loop: aio inherit the ioprio of original request
From: Bart Van Assche
Date: Fri Apr 11 2025 - 15:37:21 EST
On 4/11/25 12:28 AM, Yunlong Xing wrote:
diff --git a/drivers/block/loop.c b/drivers/block/loop.c
index 674527d770dc..ed73627af3ff 100644
--- a/drivers/block/loop.c
+++ b/drivers/block/loop.c
@@ -447,7 +447,6 @@ static int lo_rw_aio(struct loop_device *lo, struct loop_cmd *cmd,
cmd->iocb.ki_filp = file;
cmd->iocb.ki_complete = lo_rw_aio_complete;
cmd->iocb.ki_flags = IOCB_DIRECT;
- cmd->iocb.ki_ioprio = IOPRIO_PRIO_VALUE(IOPRIO_CLASS_NONE, 0);
if (rw == ITER_SOURCE)
ret = file->f_op->write_iter(&cmd->iocb, &iter);
@@ -1892,6 +1891,9 @@ static blk_status_t loop_queue_rq(struct blk_mq_hw_ctx *hctx,
break;
}
+ /* get original request ioprio */
+ cmd->iocb.ki_ioprio = req_get_ioprio(rq);
+
/* always use the first bio's css */
cmd->blkcg_css = NULL;
cmd->memcg_css = NULL;
Shouldn't the cmd->iocb.ki_ioprio assignment occur in lo_rw_aio() rather
than in loop_queue_rq() since the iocb is only used by lo_rw_aio()?
Bart.