Re: [PATCH v3] xprtrdma: serialize unmap_sync with xprt_disconnect

On Fri, Aug 21, 2026, at 7:48 PM, [email protected] wrote:

> diff --git a/net/sunrpc/xprtrdma/verbs.c b/net/sunrpc/xprtrdma/verbs.c
> index 04b286223b24..b9ebfcbcabc5 100644
> --- a/net/sunrpc/xprtrdma/verbs.c
> +++ b/net/sunrpc/xprtrdma/verbs.c

> @@ -388,6 +387,14 @@ static noinline int rpcrdma_ep_put(struct rpcrdma_ep *ep)
>  	return kref_put(&ep->re_kref, rpcrdma_ep_destroy);
>  }
> 
> +void rpcrdma_ep_release(struct rpcrdma_ep *ep)
> +{
> +	struct rdma_cm_id *id = ep->re_id;
> +
> +	if (rpcrdma_ep_put(ep))
> +		rdma_destroy_id(id);
> +}
> +

Nit: The new externally-visible functions need kdoc comments.
rpcrdma_ep_put() is now static, so the kdoc comment there can
be converted to non-kdoc-style.


> @@ -601,6 +614,9 @@ void rpcrdma_xprt_disconnect(struct rpcrdma_xprt *r_xprt)
>  		rdma_destroy_id(id);
> 
>  	r_xprt->rx_ep = NULL;
> +
> +out_unlock:
> +	up_write(&r_xprt->rx_unmap_rwsem);
>  }

Nit: rpcrdma_xprt_disconnect() retains

    if (rpcrdma_ep_put(ep)) 
        rdma_destroy_id(id);

but should call the new rpcrdma_ep_release() helper.

The new block comments are wordy, and could be tightened. But no
show-stoppers here:

Reviewed-by: Chuck Lever <[email protected]>

-- 
Chuck Lever