From: Hannes Reinecke <hare@suse.de>
To: Sagi Grimberg <sagi@grimberg.me>
Cc: Christoph Hellwig <hch@lst.de>, Keith Busch <kbusch@kernel.org>,
linux-nvme@lists.infradead.org,
Chuck Lever <chuck.lever@oracle.com>,
kernel-tls-handshake@lists.linux.dev,
Hannes Reinecke <hare@suse.de>
Subject: [PATCH 08/17] nvme-tcp: fixup MSG_SENDPAGE_NOTLAST
Date: Wed, 19 Apr 2023 08:57:05 +0200 [thread overview]
Message-ID: <20230419065714.52076-9-hare@suse.de> (raw)
In-Reply-To: <20230419065714.52076-1-hare@suse.de>
We can only set MSG_SENDPAGE_NOTLAST when sending the command PDU if the
actual data is transferred via sendpage(), too.
If we use sendmsg() to transfer the data we end up with a TX stall on TLS
as it implements different code paths for sendmsg() and sendpage().
So check in nvme_tcp_try_send_pdu() if inline data is present and revert
to sendmsg if the inline data cannot be sent via sendpage().
Signed-off-by: Hannes Reinecke <hare@suse.de>
---
drivers/nvme/host/tcp.c | 18 ++++++++++++++----
1 file changed, 14 insertions(+), 4 deletions(-)
diff --git a/drivers/nvme/host/tcp.c b/drivers/nvme/host/tcp.c
index 273c1f2760a4..9ecd7db2cd98 100644
--- a/drivers/nvme/host/tcp.c
+++ b/drivers/nvme/host/tcp.c
@@ -995,9 +995,11 @@ static int nvme_tcp_try_send_data(struct nvme_tcp_request *req)
if (last && !queue->data_digest && !nvme_tcp_queue_more(queue))
flags |= MSG_EOR;
else
- flags |= MSG_MORE | MSG_SENDPAGE_NOTLAST;
+ flags |= MSG_MORE;
if (sendpage_ok(page)) {
+ if (flags & MSG_MORE)
+ flags |= MSG_SENDPAGE_NOTLAST;
ret = kernel_sendpage(queue->sock, page, offset, len,
flags);
} else {
@@ -1049,15 +1051,23 @@ static int nvme_tcp_try_send_cmd_pdu(struct nvme_tcp_request *req)
int ret;
if (inline_data || nvme_tcp_queue_more(queue))
- flags |= MSG_MORE | MSG_SENDPAGE_NOTLAST;
+ flags |= MSG_MORE;
else
flags |= MSG_EOR;
if (queue->hdr_digest && !req->offset)
nvme_tcp_hdgst(queue->snd_hash, pdu, sizeof(*pdu));
- ret = kernel_sendpage(queue->sock, virt_to_page(pdu),
- offset_in_page(pdu) + req->offset, len, flags);
+ if (!inline_data || sendpage_ok(nvme_tcp_req_cur_page(req))) {
+ if (flags & MSG_MORE)
+ flags |= MSG_SENDPAGE_NOTLAST;
+ ret = kernel_sendpage(queue->sock, virt_to_page(pdu),
+ offset_in_page(pdu) + req->offset, len, flags);
+ } else {
+ ret = sock_no_sendpage(queue->sock, virt_to_page(pdu),
+ offset_in_page(pdu) + req->offset,
+ len, flags);
+ }
if (unlikely(ret <= 0))
return ret;
--
2.35.3
next prev parent reply other threads:[~2023-04-19 6:57 UTC|newest]
Thread overview: 46+ messages / expand[flat|nested] mbox.gz Atom feed top
2023-04-19 6:56 [PATCHv4 00/17] nvme: In-kernel TLS support for TCP Hannes Reinecke
2023-04-19 6:56 ` [PATCH 01/17] nvme-keyring: register '.nvme' keyring Hannes Reinecke
2023-05-02 14:17 ` Aurelien Aptel
2023-04-19 6:56 ` [PATCH 02/17] nvme-keyring: define a 'psk' keytype Hannes Reinecke
2023-05-08 8:59 ` Max Gurtovoy
2023-05-08 13:56 ` Hannes Reinecke
2023-04-19 6:57 ` [PATCH 03/17] nvme: add TCP TSAS definitions Hannes Reinecke
2023-05-08 16:36 ` Max Gurtovoy
2023-05-08 21:01 ` Hannes Reinecke
2023-04-19 6:57 ` [PATCH 04/17] nvme-tcp: add definitions for TLS cipher suites Hannes Reinecke
2023-04-19 6:57 ` [PATCH 05/17] nvme-keyring: implement nvme_tls_psk_default() Hannes Reinecke
2023-05-09 7:31 ` Max Gurtovoy
2023-05-09 14:13 ` Hannes Reinecke
2023-04-19 6:57 ` [PATCH 06/17] net/tls: implement ->read_sock() Hannes Reinecke
2023-04-19 9:09 ` Sagi Grimberg
2023-05-09 23:30 ` Sagi Grimberg
2023-05-10 0:47 ` Jakub Kicinski
2023-05-17 6:43 ` Sagi Grimberg
2023-05-17 7:53 ` Hannes Reinecke
2023-05-17 7:54 ` Christoph Hellwig
2023-05-17 7:56 ` Hannes Reinecke
2023-05-17 9:36 ` Sagi Grimberg
2023-04-19 6:57 ` [PATCH 07/17] net/tls: handle MSG_EOR for tls_sw TX flow Hannes Reinecke
2023-05-09 9:19 ` Max Gurtovoy
2023-05-09 14:18 ` Hannes Reinecke
2023-05-09 15:13 ` Jakub Kicinski
2023-05-09 23:02 ` Max Gurtovoy
2023-05-09 23:07 ` Hannes Reinecke
2023-04-19 6:57 ` Hannes Reinecke [this message]
2023-04-19 9:08 ` [PATCH 08/17] nvme-tcp: fixup MSG_SENDPAGE_NOTLAST Sagi Grimberg
2023-04-19 6:57 ` [PATCH 09/17] security/keys: export key_lookup() Hannes Reinecke
2023-04-19 6:57 ` [PATCH 10/17] nvme/tcp: allocate socket file Hannes Reinecke
2023-04-19 6:57 ` [PATCH 11/17] nvme-tcp: enable TLS handshake upcall Hannes Reinecke
2023-05-09 9:48 ` Max Gurtovoy
2023-05-09 14:22 ` Hannes Reinecke
2023-05-09 23:16 ` Max Gurtovoy
2023-05-10 7:37 ` Hannes Reinecke
2023-04-19 6:57 ` [PATCH 12/17] nvme-tcp: control message handling for recvmsg() Hannes Reinecke
2023-04-19 6:57 ` [PATCH 13/17] nvme-fabrics: parse options 'keyring' and 'tls_key' Hannes Reinecke
2023-04-21 6:32 ` Daniel Wagner
2023-05-09 10:00 ` Max Gurtovoy
2023-05-09 14:24 ` Hannes Reinecke
2023-04-19 6:57 ` [PATCH 14/17] nvmet: make TCP sectype settable via configfs Hannes Reinecke
2023-04-19 6:57 ` [PATCH 15/17] nvmet-tcp: allocate socket file Hannes Reinecke
2023-04-19 6:57 ` [PATCH 16/17] nvmet-tcp: enable TLS handshake upcall Hannes Reinecke
2023-04-19 6:57 ` [PATCH 17/17] nvmet-tcp: control messages for recvmsg() Hannes Reinecke
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20230419065714.52076-9-hare@suse.de \
--to=hare@suse.de \
--cc=chuck.lever@oracle.com \
--cc=hch@lst.de \
--cc=kbusch@kernel.org \
--cc=kernel-tls-handshake@lists.linux.dev \
--cc=linux-nvme@lists.infradead.org \
--cc=sagi@grimberg.me \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox