Skip to contentSkip to conversation

nvme-auth: update sc_c in host response

PATCH In review Martin George 3 participants 6 messages Sep 8, 2025 – Oct 21, 2025

Subsystems (2)
lore
Share & tools
lore
#0 lore raw

PATCH v2 1/1 nvme-auth: update sc_c in host response

The sc_c field is currently not updated in the host response to the controller challenge leading to failures while attempting secure channel concatenation. Fix this by adding a new sc_c variable to the dhchap queue context structure which is appropriately set during negotiate and then used in the host response.

Fixes e88a7595b57f ("nvme-tcp: request secure channel concatenation")

---
Changes in v2: set *buf to sc_c as suggested by Hannes

---
 drivers/nvme/host/auth.c | 6 +++++-
 1 file changed, 5 insertions(+), 1 deletion(-)
drivers/nvme/host/auth.c +5 -1
index 201fc8809a62..c70db1bc3fda 100644
@@ -36,6 +36,7 @@ struct nvme_dhchap_queue_context {
 	u8 status;
 	u8 dhgroup_id;
 	u8 hash_id;
+	u8 sc_c;
 	size_t hash_len;
 	u8 c1[64];
 	u8 c2[64];
@@ -154,6 +155,8 @@ static int nvme_auth_set_dhchap_negotiate_data(struct nvme_ctrl *ctrl,
 	data->auth_protocol[0].dhchap.idlist[34] = NVME_AUTH_DHGROUP_6144;
 	data->auth_protocol[0].dhchap.idlist[35] = NVME_AUTH_DHGROUP_8192;
 
+	chap->sc_c = data->sc_c;
+
 	return size;
 }
 
@@ -488,7 +491,7 @@ static int nvme_auth_dhchap_setup_host_response(struct nvme_ctrl *ctrl,
 	ret = crypto_shash_update(shash, buf, 2);
 	if (ret)
 		goto out;
-	memset(buf, 0, sizeof(buf));
+	*buf = chap->sc_c;
 	ret = crypto_shash_update(shash, buf, 1);
 	if (ret)
 		goto out;
@@ -499,6 +502,7 @@ static int nvme_auth_dhchap_setup_host_response(struct nvme_ctrl *ctrl,
 				  strlen(ctrl->opts->host->nqn));
 	if (ret)
 		goto out;
+	memset(buf, 0, sizeof(buf));
 	ret = crypto_shash_update(shash, buf, 1);
 	if (ret)
 		goto out;
Signature · 1 line
2.43.0
maintainer Reviewed-by on v2 1/1
#1 lore raw
Quoted 6 lines · Martin George

The sc_c field is currently not updated in the host response to the controller challenge leading to failures while attempting secure channel concatenation. Fix this by adding a new sc_c variable to the dhchap queue context structure which is appropriately set during negotiate and then used in the host response.

Quoted 9 lines
Fixes e88a7595b57f ("nvme-tcp: request secure channel concatenation")

---
Changes in v2: set *buf to sc_c as suggested by Hannes

---
  drivers/nvme/host/auth.c | 6 +++++-
  1 file changed, 5 insertions(+), 1 deletion(-)

Cheers,

Hannes

Signature · 4 lines
Dr. Hannes Reinecke                  Kernel Storage Architect
[email protected]                                +49 911 74053 688
SUSE Software Solutions GmbH, Frankenstr. 146, 90461 Nürnberg
HRB 36809 (AG Nürnberg), GF: I. Totev, A. McDonald, W. Knoblich
#2 lore raw
Quoted 18 lines · Hannes Reinecke
Quoted 8 lines · Martin George

The sc_c field is currently not updated in the host response to the controller challenge leading to failures while attempting secure channel concatenation. Fix this by adding a new sc_c variable to the
dhchap queue context structure which is appropriately set during negotiate and then used in the host response.

Fixes e88a7595b57f ("nvme-tcp: request secure channel
Quoted 9 lines

concatenation")

---
Changes in v2: set *buf to sc_c as suggested by Hannes
---
  drivers/nvme/host/auth.c | 6 +++++-

  1 file changed, 5 insertions(+), 1 deletion(-)
Quoted 5 lines

Cheers,

Hannes

Gentle ping

#3 lore raw
Quoted 20 lines · Martin George
Quoted 19 lines · Hannes Reinecke
Quoted 18 lines · Martin George

The sc_c field is currently not updated in the host response to the
controller challenge leading to failures while attempting secure channel concatenation. Fix this by adding a new sc_c variable to the
dhchap queue context structure which is appropriately set during negotiate and then used in the host response.

Fixes e88a7595b57f ("nvme-tcp: request secure channel

concatenation")

---
Changes in v2: set *buf to sc_c as suggested by Hannes
---
  drivers/nvme/host/auth.c | 6 +++++-

  1 file changed, 5 insertions(+), 1 deletion(-)
Quoted 7 lines
Quoted 5 lines

Cheers,

Hannes

Gentle ping

Pinging again. Please review and provide feedback if any. Thanks.

-Martin

#4 lore raw
In reply to Martin George
Quoted 11 lines · Martin George

The sc_c field is currently not updated in the host response to the controller challenge leading to failures while attempting secure channel concatenation. Fix this by adding a new sc_c variable to the dhchap queue context structure which is appropriately set during negotiate and then used in the host response.

Fixes e88a7595b57f ("nvme-tcp: request secure channel concatenation")

---
Changes in v2: set *buf to sc_c as suggested by Hannes

I ran blktests with v6.18-rc2 kernel, and found the test case nvme/063 shows a new failure symptom. I bisected and found this patch in the v6.18-rc2 tag triggers the failrue. Actions for fix will be appreciated. FYI, I attach the blktests console output [1] and the kernel messages [2].

To be precise, nvme/063 fails even without this patch as I reported [3]. But that failure was due to lockdep WARN. This patch introduced a new failure symptom. I think both the two failure symptoms should be fixed.

[1] blktests run console output

# NVMET_TRTYPES=tcp ./check nvme/063
nvme/063 (tr=tcp) (Create authenticated TCP connections with secure concatenation)

    runtime  2.922s  ...

WARNING: Test did not clean up port: 0
WARNING: Test did not clean up subsystem: blktests-subsystem-1
rmdir: failed to remove '/sys/kernel/config/nvmet//subsystems/blktests-subsystem-1': Directory not empty
WARNING: Test did not clean up host: nqn.2014-08.org.nvmexpress:uuid:0f01fb42-9f7f-4856-b0b3-51e60b8de349
nvme/063 (tr=tcp) (Create authenticated TCP connections with secure concatenation) [failed]01fb42-9f7f-48 runtime 2.922s ... 3.908s resource busy
--- tests/nvme/063.out 2025-08-26 21:28:52.805848674 +0900
+++ /home/shin/Blktests/blktests/results/nodev_tr_tcp/nvme/063.out.bad 2025-10-20 17:42:12.097662643 +0900
@@ -1,7 +1,3 @@
Running nvme/063
Test secure concatenation with SHA256
-Reset controller
-disconnected 1 controller(s)
-Test secure concatenation with SHA384
-disconnected 1 controller(s)
-Test complete
...
(Run 'diff -u tests/nvme/063.out /home/shin/Blktests/blktests/results/nodev_tr_tcp/nvme/063.out.bad' to see the entire diff)
WARNING: Test did not clean up subsystem: blktests-subsystem-1
rmdir: failed to remove '/sys/kernel/config/nvmet//subsystems/blktests-subsystem-1': Directory not empty
WARNING: Test did not clean up host: nqn.2014-08.org.nvmexpress:uuid:0f01fb42-9f7f-4856-b0b3-51e60b8de349
rmdir: failed to remove '/sys/kernel/config/nvmet//hosts/nqn.2014-08.org.nvmexpress:uuid:0f01fb42-9f7f-4856-b0b3-51e60b8de349': Device or resource busy

[2] kernel message log

[   59.957999] [    T950] run blktests nvme/063 at 2025-10-20 17:42:08
[   60.146451] [   T1059] nvmet: adding nsid 1 to subsystem blktests-subsystem-1
[   60.168784] [   T1060] nvmet: Allow non-TLS connections while TLS1.3 is enabled
[   60.179760] [   T1063] nvmet_tcp: enabling port 0 (127.0.0.1:4420)
[   60.321134] [     T80] nvmet: Created nvm controller 1 for subsystem blktests-subsystem-1 for NQN nqn.2014-08.org.nvmexpress:uuid:0f01fb42-9f7f-4856-b0b3-51e60b8de349 with DH-HMAC-CHAP.
[   60.329310] [     T89] nvmet: ctrl 1 qid 0 host response mismatch
[   60.329914] [     T89] nvmet: ctrl 1 qid 0 failure1 (1)
[   60.330650] [     T82] nvmet: ctrl 1 fatal error occurred!
[   60.331254] [   T1073] nvme nvme5: qid 0: authentication failed, error -129
[   60.333179] [   T1073] nvme nvme5: failed to connect queue: 0 ret=-129

[3] https://lore.kernel.org/linux-nvme/ynmi72x5wt5ooljjafebhcarit3pvu6axkslqenikb2p5txe57@ldytqa2t4i2x/T/#u

maintainer
#5 lore raw
Quoted 65 lines · Shinichiro Kawasaki
Quoted 11 lines · Martin George

The sc_c field is currently not updated in the host response to the controller challenge leading to failures while attempting secure channel concatenation. Fix this by adding a new sc_c variable to the dhchap queue context structure which is appropriately set during negotiate and then used in the host response.

Fixes e88a7595b57f ("nvme-tcp: request secure channel concatenation")

---
Changes in v2: set *buf to sc_c as suggested by Hannes

I ran blktests with v6.18-rc2 kernel, and found the test case nvme/063 shows a new failure symptom. I bisected and found this patch in the v6.18-rc2 tag triggers the failrue. Actions for fix will be appreciated. FYI, I attach the blktests console output [1] and the kernel messages [2].

To be precise, nvme/063 fails even without this patch as I reported [3]. But that failure was due to lockdep WARN. This patch introduced a new failure symptom. I think both the two failure symptoms should be fixed.

[1] blktests run console output

# NVMET_TRTYPES=tcp ./check nvme/063
nvme/063 (tr=tcp) (Create authenticated TCP connections with secure concatenation)

     runtime  2.922s  ...

WARNING: Test did not clean up port: 0
WARNING: Test did not clean up subsystem: blktests-subsystem-1
rmdir: failed to remove '/sys/kernel/config/nvmet//subsystems/blktests-subsystem-1': Directory not empty
WARNING: Test did not clean up host: nqn.2014-08.org.nvmexpress:uuid:0f01fb42-9f7f-4856-b0b3-51e60b8de349
nvme/063 (tr=tcp) (Create authenticated TCP connections with secure concatenation) [failed]01fb42-9f7f-48 runtime 2.922s ... 3.908s resource busy
--- tests/nvme/063.out 2025-08-26 21:28:52.805848674 +0900
+++ /home/shin/Blktests/blktests/results/nodev_tr_tcp/nvme/063.out.bad 2025-10-20 17:42:12.097662643 +0900
@@ -1,7 +1,3 @@
Running nvme/063
Test secure concatenation with SHA256
-Reset controller
-disconnected 1 controller(s)
-Test secure concatenation with SHA384
-disconnected 1 controller(s)
-Test complete
...
(Run 'diff -u tests/nvme/063.out /home/shin/Blktests/blktests/results/nodev_tr_tcp/nvme/063.out.bad' to see the entire diff)
WARNING: Test did not clean up subsystem: blktests-subsystem-1
rmdir: failed to remove '/sys/kernel/config/nvmet//subsystems/blktests-subsystem-1': Directory not empty
WARNING: Test did not clean up host: nqn.2014-08.org.nvmexpress:uuid:0f01fb42-9f7f-4856-b0b3-51e60b8de349
rmdir: failed to remove '/sys/kernel/config/nvmet//hosts/nqn.2014-08.org.nvmexpress:uuid:0f01fb42-9f7f-4856-b0b3-51e60b8de349': Device or resource busy

[2] kernel message log

[   59.957999] [    T950] run blktests nvme/063 at 2025-10-20 17:42:08
[   60.146451] [   T1059] nvmet: adding nsid 1 to subsystem blktests-subsystem-1
[   60.168784] [   T1060] nvmet: Allow non-TLS connections while TLS1.3 is enabled
[   60.179760] [   T1063] nvmet_tcp: enabling port 0 (127.0.0.1:4420)
[   60.321134] [     T80] nvmet: Created nvm controller 1 for subsystem blktests-subsystem-1 for NQN nqn.2014-08.org.nvmexpress:uuid:0f01fb42-9f7f-4856-b0b3-51e60b8de349 with DH-HMAC-CHAP.
[   60.329310] [     T89] nvmet: ctrl 1 qid 0 host response mismatch
[   60.329914] [     T89] nvmet: ctrl 1 qid 0 failure1 (1)
[   60.330650] [     T82] nvmet: ctrl 1 fatal error occurred!
[   60.331254] [   T1073] nvme nvme5: qid 0: authentication failed, error -129
[   60.333179] [   T1073] nvme nvme5: failed to connect queue: 0 ret=-129
Quoted 1 line

[3] https://lore.kernel.org/linux-nvme/ynmi72x5wt5ooljjafebhcarit3pvu6axkslqenikb2p5txe57@ldytqa2t4i2x/T/#u

That's because the target side needs the same fix. I'll send a patch.

Cheers,

Hannes

Signature · 4 lines
Dr. Hannes Reinecke                  Kernel Storage Architect
[email protected]                                +49 911 74053 688
SUSE Software Solutions GmbH, Frankenstr. 146, 90461 Nürnberg
HRB 36809 (AG Nürnberg), GF: I. Totev, A. McDonald, W. Knoblich