Skip to contentSkip to conversation

nvme-tcp: always set valid seq_num in dhchap reply

PATCH In review Mark O'Donovan 3 participants 7 messages Oct 25, 2023 – Nov 6, 2023

Subsystems (3)
lore
Share & tools
lore
#0 lore raw

v3 cover nvme-tcp: always set valid seq_num in dhchap reply

The first patch is a small unrelated fix which makes it clear that the response data section of the Success1 message is not optional.

The second patch removes use of the host sequence number (S2) as an indicator of bi-directional authentication.

The third patch causes us to always set the host sequence number (S2) to a non-zero value instead of the 0 value reserved for the secure channel feature.

v1: original submission
v2: rebased + added dedicated variable for bi-directional auth
v3: nvme-target code now reads seq-num

Mark O'Donovan (3):
nvme-auth: auth success1 msg always includes resp
nvme-auth: add flag for bi-directional auth
nvme-auth: always set valid seq_num in dhchap reply

 drivers/nvme/host/auth.c               | 13 ++++++-------
 drivers/nvme/target/fabrics-cmd-auth.c |  2 +-
 include/linux/nvme.h                   |  2 +-
 3 files changed, 8 insertions(+), 9 deletions(-)

base-commit: 4f82870119a46b0d04d91ef4697ac4977a255a9d

Signature · 1 line
2.39.2
#1 lore raw

PATCH v3 1/3 nvme-auth: auth success1 msg always includes resp

In cases where RVALID is false, the response is still transmitted, but is cleared to zero.

Relevant extract from the spec:
Response R2, if valid (i.e., if the RVALID field is set to 01h), cleared to 0h otherwise

---
v1: used incorrect prefix nvme-tcp
v2: rebase on latest git

 drivers/nvme/host/auth.c | 5 +----
 include/linux/nvme.h     | 2 +-
 2 files changed, 2 insertions(+), 5 deletions(-)
drivers/nvme/host/auth.c +1 -4
index 064592a5d546..e0fc33d41baa 100644
@@ -339,10 +339,7 @@ static int nvme_auth_process_dhchap_success1(struct nvme_ctrl *ctrl,
 		struct nvme_dhchap_queue_context *chap)
 {
 	struct nvmf_auth_dhchap_success1_data *data = chap->buf;
-	size_t size = sizeof(*data);
-
-	if (chap->s2)
-		size += chap->hash_len;
+	size_t size = sizeof(*data) + chap->hash_len;
 
 	if (size > CHAP_BUF_SIZE) {
 		chap->status = NVME_AUTH_DHCHAP_FAILURE_INCORRECT_PAYLOAD;
include/linux/nvme.h +1 -1
index 26dd3f859d9d..8d8df9c15b74 100644
@@ -1722,7 +1722,7 @@ struct nvmf_auth_dhchap_success1_data {
 	__u8		rsvd2;
 	__u8		rvalid;
 	__u8		rsvd3[7];
-	/* 'hl' bytes of response value if 'rvalid' is set */
+	/* 'hl' bytes of response value */
 	__u8		rval[];
 };
 
Signature · 1 line
2.39.2
#2 lore raw

PATCH v3 2/3 nvme-auth: add flag for bi-directional auth

In reply to Mark O'Donovan

Introduces an explicit variable for bi-directional auth. The currently used variable chap->s2 is incorrectly zeroed for uni-directional auth. That will be fixed in the next patch so this needs to change to avoid sending unexpected success2 messages

---
 drivers/nvme/host/auth.c | 5 ++++-
 1 file changed, 4 insertions(+), 1 deletion(-)
drivers/nvme/host/auth.c +4 -1
index e0fc33d41baa..8558a02865ac 100644
@@ -28,6 +28,7 @@ struct nvme_dhchap_queue_context {
 	int error;
 	u32 s1;
 	u32 s2;
+	bool bi_directional;
 	u16 transaction;
 	u8 status;
 	u8 dhgroup_id;
@@ -312,6 +313,7 @@ static int nvme_auth_set_dhchap_reply_data(struct nvme_ctrl *ctrl,
 	data->dhvlen = cpu_to_le16(chap->host_key_len);
 	memcpy(data->rval, chap->response, chap->hash_len);
 	if (ctrl->ctrl_key) {
+		chap->bi_directional = true;
 		get_random_bytes(chap->c2, chap->hash_len);
 		data->cvalid = 1;
 		chap->s2 = nvme_auth_get_seqnum();
@@ -660,6 +662,7 @@ static void nvme_auth_reset_dhchap(struct nvme_dhchap_queue_context *chap)
 	chap->error = 0;
 	chap->s1 = 0;
 	chap->s2 = 0;
+	chap->bi_directional = false;
 	chap->transaction = 0;
 	memset(chap->c1, 0, sizeof(chap->c1));
 	memset(chap->c2, 0, sizeof(chap->c2));
@@ -822,7 +825,7 @@ static void nvme_queue_auth_work(struct work_struct *work)
 		goto fail2;
 	}
 
-	if (chap->s2) {
+	if (chap->bi_directional) {
 		/* DH-HMAC-CHAP Step 5: send success2 */
 		dev_dbg(ctrl->device, "%s: qid %d send success2\n",
 			__func__, chap->qid);
Signature · 1 line
2.39.2
maintainer Reviewed-by on v3 2/3
#3 lore raw

Re: [PATCH v3 2/3] nvme-auth: add flag for bi-directional auth

Quoted 11 lines · Mark O'Donovan

Introduces an explicit variable for bi-directional auth. The currently used variable chap->s2 is incorrectly zeroed for uni-directional auth. That will be fixed in the next patch so this needs to change to avoid sending unexpected success2 messages

---
  drivers/nvme/host/auth.c | 5 ++++-
  1 file changed, 4 insertions(+), 1 deletion(-)

Cheers,

Hannes

#4 lore raw

PATCH v3 3/3 nvme-auth: always set valid seq_num in dhchap reply

In reply to Mark O'Donovan

Currently a seqnum of zero is sent during uni-directional authentication. The zero value is reserved for the secure channel feature which is not yet implemented.

Relevant extract from the spec:
The value 0h is used to indicate that bidirectional authentication is not performed, but a challenge value C2 is carried in order to generate a pre-shared key (PSK) for subsequent establishment of a secure channel

---
v1: used incorrect prefix nvme-tcp
v2: added spec extract to commit message
v3: read the seq-num in nvme-target code

 drivers/nvme/host/auth.c               | 3 +--
 drivers/nvme/target/fabrics-cmd-auth.c | 2 +-
 2 files changed, 2 insertions(+), 3 deletions(-)
drivers/nvme/host/auth.c +1 -2
index 8558a02865ac..7f6b2e99a78c 100644
@@ -316,15 +316,14 @@ static int nvme_auth_set_dhchap_reply_data(struct nvme_ctrl *ctrl,
 		chap->bi_directional = true;
 		get_random_bytes(chap->c2, chap->hash_len);
 		data->cvalid = 1;
-		chap->s2 = nvme_auth_get_seqnum();
 		memcpy(data->rval + chap->hash_len, chap->c2,
 		       chap->hash_len);
 		dev_dbg(ctrl->device, "%s: qid %d ctrl challenge %*ph\n",
 			__func__, chap->qid, (int)chap->hash_len, chap->c2);
 	} else {
 		memset(chap->c2, 0, chap->hash_len);
-		chap->s2 = 0;
 	}
+	chap->s2 = nvme_auth_get_seqnum();
 	data->seqnum = cpu_to_le32(chap->s2);
 	if (chap->host_key_len) {
 		dev_dbg(ctrl->device, "%s: qid %d host public key %*ph\n",
drivers/nvme/target/fabrics-cmd-auth.c +1 -1
index 1d9854484e2e..eb7785be0ca7 100644
@@ -163,11 +163,11 @@ static u16 nvmet_auth_reply(struct nvmet_req *req, void *d)
 		pr_debug("%s: ctrl %d qid %d challenge %*ph\n",
 			 __func__, ctrl->cntlid, req->sq->qid, data->hl,
 			 req->sq->dhchap_c2);
-		req->sq->dhchap_s2 = le32_to_cpu(data->seqnum);
 	} else {
 		req->sq->authenticated = true;
 		req->sq->dhchap_c2 = NULL;
 	}
+	req->sq->dhchap_s2 = le32_to_cpu(data->seqnum);
 
 	return 0;
 }
Signature · 1 line
2.39.2
maintainer Reviewed-by on v3 3/3
#5 lore raw

Re: [PATCH v3 3/3] nvme-auth: always set valid seq_num in dhchap reply

Quoted 10 lines · Mark O'Donovan

Currently a seqnum of zero is sent during uni-directional authentication. The zero value is reserved for the secure channel feature which is not yet implemented.

Relevant extract from the spec:
The value 0h is used to indicate that bidirectional authentication is not performed, but a challenge value C2 is carried in order to generate a pre-shared key (PSK) for subsequent establishment of a secure channel

Quoted 8 lines

---
v1: used incorrect prefix nvme-tcp
v2: added spec extract to commit message
v3: read the seq-num in nvme-target code

Cheers,

Hannes

maintainer
#6 lore raw
In reply to Mark O'Donovan
Quoted 9 lines · Mark O'Donovan

The first patch is a small unrelated fix which makes it clear that the response data section of the Success1 message is not optional.

The second patch removes use of the host sequence number (S2) as an indicator of bi-directional authentication.

The third patch causes us to always set the host sequence number (S2) to a non-zero value instead of the 0 value reserved for the secure channel feature.

Thanks, queued up for nvme-6.7.