All of lore.kernel.org
 help / color / mirror / Atom feed
* [PATCH] nvmet-tcp: Fix receive data digest calculation for multiple h2cdata PDUs
@ 2021-02-03 23:00 Sagi Grimberg
  2021-02-07 16:16 ` Christoph Hellwig
  2021-02-21 19:37 ` Grupi, Elad
  0 siblings, 2 replies; 5+ messages in thread
From: Sagi Grimberg @ 2021-02-03 23:00 UTC (permalink / raw)
  To: linux-nvme, Christoph Hellwig, Keith Busch, Chaitanya Kulkarni

When a host sends multiple h2cdata PDUs for a single command, we
should verify the data digest calculation per PDU and not
per command.

Fixes: 872d26a391da ("nvmet-tcp: add NVMe over TCP target driver")
Reported-by: Narayan Ayalasomayajula <Narayan.Ayalasomayajula@wdc.com>
Tested-by: Narayan Ayalasomayajula <Narayan.Ayalasomayajula@wdc.com>
Signed-off-by: Sagi Grimberg <sagi@grimberg.me>
---
 drivers/nvme/target/tcp.c | 31 ++++++++++++++++++++++++-------
 1 file changed, 24 insertions(+), 7 deletions(-)

diff --git a/drivers/nvme/target/tcp.c b/drivers/nvme/target/tcp.c
index aacf06f0b431..577ce7d403ae 100644
--- a/drivers/nvme/target/tcp.c
+++ b/drivers/nvme/target/tcp.c
@@ -379,7 +379,7 @@ static int nvmet_tcp_map_data(struct nvmet_tcp_cmd *cmd)
 	return NVME_SC_INTERNAL;
 }
 
-static void nvmet_tcp_ddgst(struct ahash_request *hash,
+static void nvmet_tcp_send_ddgst(struct ahash_request *hash,
 		struct nvmet_tcp_cmd *cmd)
 {
 	ahash_request_set_crypt(hash, cmd->req.sg,
@@ -387,6 +387,23 @@ static void nvmet_tcp_ddgst(struct ahash_request *hash,
 	crypto_ahash_digest(hash);
 }
 
+static void nvmet_tcp_recv_ddgst(struct ahash_request *hash,
+		struct nvmet_tcp_cmd *cmd)
+{
+	struct scatterlist sg;
+	struct kvec *iov;
+	int i;
+
+	crypto_ahash_init(hash);
+	for (i = 0, iov = cmd->iov; i < cmd->nr_mapped; i++, iov++) {
+		sg_init_one(&sg, iov->iov_base, iov->iov_len);
+		ahash_request_set_crypt(hash, &sg, NULL, iov->iov_len);
+		crypto_ahash_update(hash);
+	}
+	ahash_request_set_crypt(hash, NULL, (void *)&cmd->exp_ddgst, 0);
+	crypto_ahash_final(hash);
+}
+
 static void nvmet_setup_c2h_data_pdu(struct nvmet_tcp_cmd *cmd)
 {
 	struct nvme_tcp_data_pdu *pdu = cmd->data_pdu;
@@ -411,7 +428,7 @@ static void nvmet_setup_c2h_data_pdu(struct nvmet_tcp_cmd *cmd)
 
 	if (queue->data_digest) {
 		pdu->hdr.flags |= NVME_TCP_F_DDGST;
-		nvmet_tcp_ddgst(queue->snd_hash, cmd);
+		nvmet_tcp_send_ddgst(queue->snd_hash, cmd);
 	}
 
 	if (cmd->queue->hdr_digest) {
@@ -1060,7 +1077,7 @@ static void nvmet_tcp_prep_recv_ddgst(struct nvmet_tcp_cmd *cmd)
 {
 	struct nvmet_tcp_queue *queue = cmd->queue;
 
-	nvmet_tcp_ddgst(queue->rcv_hash, cmd);
+	nvmet_tcp_recv_ddgst(queue->rcv_hash, cmd);
 	queue->offset = 0;
 	queue->left = NVME_TCP_DIGEST_LENGTH;
 	queue->rcv_state = NVMET_TCP_RECV_DDGST;
@@ -1081,14 +1098,14 @@ static int nvmet_tcp_try_recv_data(struct nvmet_tcp_queue *queue)
 		cmd->rbytes_done += ret;
 	}
 
+	if (queue->data_digest) {
+		nvmet_tcp_prep_recv_ddgst(cmd);
+		return 0;
+	}
 	nvmet_tcp_unmap_pdu_iovec(cmd);
 
 	if (!(cmd->flags & NVMET_TCP_F_INIT_FAILED) &&
 	    cmd->rbytes_done == cmd->req.transfer_len) {
-		if (queue->data_digest) {
-			nvmet_tcp_prep_recv_ddgst(cmd);
-			return 0;
-		}
 		cmd->req.execute(&cmd->req);
 	}
 
-- 
2.27.0


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

^ permalink raw reply related	[flat|nested] 5+ messages in thread
* Re: [PATCH] nvmet-tcp: Fix receive data digest calculation for multiple h2cdata PDUs
  2021-02-03 23:00 [PATCH] nvmet-tcp: Fix receive data digest calculation for multiple h2cdata PDUs Sagi Grimberg
@ 2021-02-07 16:16 ` Christoph Hellwig
  2021-02-21 19:37 ` Grupi, Elad
  1 sibling, 0 replies; 5+ messages in thread
From: Christoph Hellwig @ 2021-02-07 16:16 UTC (permalink / raw)
  To: Sagi Grimberg
  Cc: Keith Busch, Chaitanya Kulkarni, Christoph Hellwig, linux-nvme

Thanks,

applied to nvme-5.12.

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

^ permalink raw reply	[flat|nested] 5+ messages in thread
* RE: [PATCH] nvmet-tcp: Fix receive data digest calculation for multiple h2cdata PDUs
  2021-02-03 23:00 [PATCH] nvmet-tcp: Fix receive data digest calculation for multiple h2cdata PDUs Sagi Grimberg
  2021-02-07 16:16 ` Christoph Hellwig
@ 2021-02-21 19:37 ` Grupi, Elad
  2021-03-16  6:29   ` Sagi Grimberg
  1 sibling, 1 reply; 5+ messages in thread
From: Grupi, Elad @ 2021-02-21 19:37 UTC (permalink / raw)
  To: Sagi Grimberg, linux-nvme, Christoph Hellwig, Keith Busch,
	Chaitanya Kulkarni

Hi Sagi,

Can you please explain why you avoid calling nvmet_tcp_unmap_pdu_iovec on data_digest PDU?
I don't recall any other case that unmap is being called on good path.

Regards,
Elad

-----Original Message-----
From: Linux-nvme <linux-nvme-bounces@lists.infradead.org> On Behalf Of Sagi Grimberg
Sent: Thursday, 4 February 2021 1:00
To: linux-nvme@lists.infradead.org; Christoph Hellwig; Keith Busch; Chaitanya Kulkarni
Subject: [PATCH] nvmet-tcp: Fix receive data digest calculation for multiple h2cdata PDUs


[EXTERNAL EMAIL] 

When a host sends multiple h2cdata PDUs for a single command, we should verify the data digest calculation per PDU and not per command.

Fixes: 872d26a391da ("nvmet-tcp: add NVMe over TCP target driver")
Reported-by: Narayan Ayalasomayajula <Narayan.Ayalasomayajula@wdc.com>
Tested-by: Narayan Ayalasomayajula <Narayan.Ayalasomayajula@wdc.com>
Signed-off-by: Sagi Grimberg <sagi@grimberg.me>
---
 drivers/nvme/target/tcp.c | 31 ++++++++++++++++++++++++-------
 1 file changed, 24 insertions(+), 7 deletions(-)

diff --git a/drivers/nvme/target/tcp.c b/drivers/nvme/target/tcp.c index aacf06f0b431..577ce7d403ae 100644
--- a/drivers/nvme/target/tcp.c
+++ b/drivers/nvme/target/tcp.c
@@ -379,7 +379,7 @@ static int nvmet_tcp_map_data(struct nvmet_tcp_cmd *cmd)
 	return NVME_SC_INTERNAL;
 }
 
-static void nvmet_tcp_ddgst(struct ahash_request *hash,
+static void nvmet_tcp_send_ddgst(struct ahash_request *hash,
 		struct nvmet_tcp_cmd *cmd)
 {
 	ahash_request_set_crypt(hash, cmd->req.sg, @@ -387,6 +387,23 @@ static void nvmet_tcp_ddgst(struct ahash_request *hash,
 	crypto_ahash_digest(hash);
 }
 
+static void nvmet_tcp_recv_ddgst(struct ahash_request *hash,
+		struct nvmet_tcp_cmd *cmd)
+{
+	struct scatterlist sg;
+	struct kvec *iov;
+	int i;
+
+	crypto_ahash_init(hash);
+	for (i = 0, iov = cmd->iov; i < cmd->nr_mapped; i++, iov++) {
+		sg_init_one(&sg, iov->iov_base, iov->iov_len);
+		ahash_request_set_crypt(hash, &sg, NULL, iov->iov_len);
+		crypto_ahash_update(hash);
+	}
+	ahash_request_set_crypt(hash, NULL, (void *)&cmd->exp_ddgst, 0);
+	crypto_ahash_final(hash);
+}
+
 static void nvmet_setup_c2h_data_pdu(struct nvmet_tcp_cmd *cmd)  {
 	struct nvme_tcp_data_pdu *pdu = cmd->data_pdu; @@ -411,7 +428,7 @@ static void nvmet_setup_c2h_data_pdu(struct nvmet_tcp_cmd *cmd)
 
 	if (queue->data_digest) {
 		pdu->hdr.flags |= NVME_TCP_F_DDGST;
-		nvmet_tcp_ddgst(queue->snd_hash, cmd);
+		nvmet_tcp_send_ddgst(queue->snd_hash, cmd);
 	}
 
 	if (cmd->queue->hdr_digest) {
@@ -1060,7 +1077,7 @@ static void nvmet_tcp_prep_recv_ddgst(struct nvmet_tcp_cmd *cmd)  {
 	struct nvmet_tcp_queue *queue = cmd->queue;
 
-	nvmet_tcp_ddgst(queue->rcv_hash, cmd);
+	nvmet_tcp_recv_ddgst(queue->rcv_hash, cmd);
 	queue->offset = 0;
 	queue->left = NVME_TCP_DIGEST_LENGTH;
 	queue->rcv_state = NVMET_TCP_RECV_DDGST; @@ -1081,14 +1098,14 @@ static int nvmet_tcp_try_recv_data(struct nvmet_tcp_queue *queue)
 		cmd->rbytes_done += ret;
 	}
 
+	if (queue->data_digest) {
+		nvmet_tcp_prep_recv_ddgst(cmd);
+		return 0;
+	}
 	nvmet_tcp_unmap_pdu_iovec(cmd);
 
 	if (!(cmd->flags & NVMET_TCP_F_INIT_FAILED) &&
 	    cmd->rbytes_done == cmd->req.transfer_len) {
-		if (queue->data_digest) {
-			nvmet_tcp_prep_recv_ddgst(cmd);
-			return 0;
-		}
 		cmd->req.execute(&cmd->req);
 	}
 
--
2.27.0


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

^ permalink raw reply	[flat|nested] 5+ messages in thread
* Re: [PATCH] nvmet-tcp: Fix receive data digest calculation for multiple h2cdata PDUs
  2021-02-21 19:37 ` Grupi, Elad
@ 2021-03-16  6:29   ` Sagi Grimberg
  2021-03-16 15:24     ` Grupi, Elad
  0 siblings, 1 reply; 5+ messages in thread
From: Sagi Grimberg @ 2021-03-16  6:29 UTC (permalink / raw)
  To: Grupi, Elad, linux-nvme, Christoph Hellwig, Keith Busch,
	Chaitanya Kulkarni


> Hi Sagi,

Hey Elad,

> Can you please explain why you avoid calling nvmet_tcp_unmap_pdu_iovec on data_digest PDU?
> I don't recall any other case that unmap is being called on good path.

You are correct. We need to either add unmap in the data_digest path
here or when completing the data digest reception.. Probably better
here.

Care to send a patch or should I?

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

^ permalink raw reply	[flat|nested] 5+ messages in thread
* RE: [PATCH] nvmet-tcp: Fix receive data digest calculation for multiple h2cdata PDUs
  2021-03-16  6:29   ` Sagi Grimberg
@ 2021-03-16 15:24     ` Grupi, Elad
  0 siblings, 0 replies; 5+ messages in thread
From: Grupi, Elad @ 2021-03-16 15:24 UTC (permalink / raw)
  To: Sagi Grimberg, linux-nvme, Christoph Hellwig, Keith Busch,
	Chaitanya Kulkarni

Sure. I will send new patch shortly.

Elad

-----Original Message-----
From: Sagi Grimberg <sagi@grimberg.me> 
Sent: Tuesday, 16 March 2021 8:30
To: Grupi, Elad; linux-nvme@lists.infradead.org; Christoph Hellwig; Keith Busch; Chaitanya Kulkarni
Subject: Re: [PATCH] nvmet-tcp: Fix receive data digest calculation for multiple h2cdata PDUs


[EXTERNAL EMAIL] 


> Hi Sagi,

Hey Elad,

> Can you please explain why you avoid calling nvmet_tcp_unmap_pdu_iovec on data_digest PDU?
> I don't recall any other case that unmap is being called on good path.

You are correct. We need to either add unmap in the data_digest path here or when completing the data digest reception.. Probably better here.

Care to send a patch or should I?
_______________________________________________
Linux-nvme mailing list
Linux-nvme@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-nvme

^ permalink raw reply	[flat|nested] 5+ messages in thread
end of thread, other threads:[~2021-03-16 15:25 UTC | newest]

Thread overview: 5+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2021-02-03 23:00 [PATCH] nvmet-tcp: Fix receive data digest calculation for multiple h2cdata PDUs Sagi Grimberg
2021-02-07 16:16 ` Christoph Hellwig
2021-02-21 19:37 ` Grupi, Elad
2021-03-16  6:29   ` Sagi Grimberg
2021-03-16 15:24     ` Grupi, Elad

This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.