[PATCH] nvmet-tcp: move send/recv error handling in the send/recv methods instead of call-sites

[PATCH] nvmet-tcp: move send/recv error handling in the send/recv methods instead of call-sites

[Sagi Grimberg]

Have routines handle errors and just bail out of the poll loop. This simplifies
the code and will help as we may enhance the poll loop logic and these are somewhat
in the way.

Signed-off-by: Sagi Grimberg <sagi@grimberg.me>
---
 drivers/nvme/target/tcp.c | 33 ++++++++++++++++++---------------
 1 file changed, 18 insertions(+), 15 deletions(-)

diff --git a/drivers/nvme/target/tcp.c b/drivers/nvme/target/tcp.c
index f8dd69f3b657..9527cdc19611 100644
--- a/drivers/nvme/target/tcp.c
+++ b/drivers/nvme/target/tcp.c
@@ -723,6 +723,13 @@ static int nvmet_tcp_try_send(struct nvmet_tcp_queue *queue,
 		(*sends)++;
 	}
 
+	if (ret < 0) {
+		if (ret == -EPIPE || ret == -ECONNRESET)
+			kernel_sock_shutdown(queue->sock, SHUT_RDWR);
+		else
+			nvmet_tcp_fatal_error(queue);
+	}
+
 	return ret;
 }
 
@@ -1164,6 +1171,13 @@ static int nvmet_tcp_try_recv(struct nvmet_tcp_queue *queue,
 		(*recvs)++;
 	}
 
+	if (ret < 0) {
+		if (ret == -EPIPE || ret == -ECONNRESET)
+			kernel_sock_shutdown(queue->sock, SHUT_RDWR);
+		else
+			nvmet_tcp_fatal_error(queue);
+	}
+
 	return ret;
 }
 
@@ -1188,27 +1202,16 @@ static void nvmet_tcp_io_work(struct work_struct *w)
 		pending = false;
 
 		ret = nvmet_tcp_try_recv(queue, NVMET_TCP_RECV_BUDGET, &ops);
-		if (ret > 0) {
+		if (ret > 0)
 			pending = true;
-		} else if (ret < 0) {
-			if (ret == -EPIPE || ret == -ECONNRESET)
-				kernel_sock_shutdown(queue->sock, SHUT_RDWR);
-			else
-				nvmet_tcp_fatal_error(queue);
+		else if (ret < 0)
 			return;
-		}
 
 		ret = nvmet_tcp_try_send(queue, NVMET_TCP_SEND_BUDGET, &ops);
-		if (ret > 0) {
-			/* transmitted message/data */
+		if (ret > 0)
 			pending = true;
-		} else if (ret < 0) {
-			if (ret == -EPIPE || ret == -ECONNRESET)
-				kernel_sock_shutdown(queue->sock, SHUT_RDWR);
-			else
-				nvmet_tcp_fatal_error(queue);
+		else if (ret < 0)
 			return;
-		}
 
 	} while (pending && ops < NVMET_TCP_IO_WORK_BUDGET);
 
-- 
2.25.1


_______________________________________________
linux-nvme mailing list
linux-nvme@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-nvme

Re: [PATCH] nvmet-tcp: move send/recv error handling in the send/recv methods instead of call-sites

[Christoph Hellwig]

On Sun, May 17, 2020 at 04:15:22AM -0700, Sagi Grimberg wrote:
> Have routines handle errors and just bail out of the poll loop. This simplifies
> the code and will help as we may enhance the poll loop logic and these are somewhat
> in the way.

Way too long lines in the commit message.

> diff --git a/drivers/nvme/target/tcp.c b/drivers/nvme/target/tcp.c
> index f8dd69f3b657..9527cdc19611 100644
> --- a/drivers/nvme/target/tcp.c
> +++ b/drivers/nvme/target/tcp.c
> @@ -723,6 +723,13 @@ static int nvmet_tcp_try_send(struct nvmet_tcp_queue *queue,
>  		(*sends)++;
>  	}
>  
> +	if (ret < 0) {
> +		if (ret == -EPIPE || ret == -ECONNRESET)
> +			kernel_sock_shutdown(queue->sock, SHUT_RDWR);
> +		else
> +			nvmet_tcp_fatal_error(queue);
> +	}

Not really important, but how about a nvmet_tcp_handle_error helper
to consolidate this duplicated code snipplet?

_______________________________________________
linux-nvme mailing list
linux-nvme@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-nvme