If any discard requests fails, the mmc driver should not cause kernel
panic. The error handling and informing block layer about error will be
done by completion context.
Change-Id: I0c88ad851ad12c392a557bcb94d3b21dfc62d1f8
Signed-off-by: Vijay Viswanath <vviswana@codeaurora.org>
ret = mq->cmdq_issue_fn(mq, mq->cmdq_req_peeked);
/*
- * Don't requeue if issue_fn fails, just bug on.
- * We don't expect failure here and there is no recovery other
- * than fixing the actual issue if there is any.
+ * Don't requeue if issue_fn fails.
+ * Recovery will be come by completion softirq
* Also we end the request if there is a partition switch error,
* so we should not requeue the request here.
*/
- if (ret)
- BUG_ON(1);
} /* loop */
return 0;