Linux kernel -stable discussions
 help / color / mirror / Atom feed
* FAILED: patch "[PATCH] nvmet-auth: serialize access to DH-HMAC-CHAP state to fix UAF" failed to apply to 6.6-stable tree
@ 2026-11-04 14:29 gregkh
  2026-11-05  1:05 ` [PATCH 6.6.y 1/2] nvmet: replace use of system_wq with system_percpu_wq Sasha Levin
  0 siblings, 1 reply; 3+ messages in thread
From: gregkh @ 2026-11-04 14:29 UTC (permalink / raw)
  To: Jeremy.Jean, hch, kbusch; +Cc: stable


The patch below does not apply to the 6.6-stable tree.
If someone wants it applied there, or to any other stable or longterm
tree, then please email the backport, including the original git commit
id to <stable@vger.kernel.org>.

To reproduce the conflict and resubmit, you may use the following commands:

git fetch https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/ linux-6.6.y
git checkout FETCH_HEAD
git cherry-pick -x 53f6efe00cea0c0cf7bd99debe3cdac8567170b5
# <resolve conflicts, build, test, etc.>
git commit -s
git send-email --to '<stable@vger.kernel.org>' --in-reply-to '2026110458-chariot-kinfolk-545d@gregkh' --subject-prefix 'PATCH 6.6.y' 'HEAD^..'

Possible dependencies:



thanks,

greg k-h

------------------ original commit in Linus's tree ------------------

From 53f6efe00cea0c0cf7bd99debe3cdac8567170b5 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?J=C3=A9r=C3=A9my=20Jean?= <Jeremy.Jean@oss.cyber.gouv.fr>
Date: Sat, 3 Oct 2026 21:01:55 +0000
Subject: [PATCH] nvmet-auth: serialize access to DH-HMAC-CHAP state to fix UAF
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

A REPLY AUTH_SEND command can derive a dhchap_skey key while a
concurrent AUTH_RECEIVE command observing this REPLY state executes
the failure path and frees that key. The REPLY command then continues
and writes into freed memory, causing a use-after-free. KASAN reports
it as:

  BUG: KASAN: slab-use-after-free in __sha256_final
  Write of size 4 at addr ffff88800359c540
  ...
  nvmet_auth_sq_free (drivers/nvme/target/auth.c:237)
  nvmet_execute_auth_receive (drivers/nvme/target/fabrics-cmd-auth.c:621)

Add a mutex to each submission queue to serialize authentication
commands and timeout handling. Unlock the mutex before completing
requests and ignore outdated timeout callbacks. Limit timeout delays
to prevent immediate expiry on 32-bit kernels.

Fixes: 7a277c37d352 ("nvmet-auth: Diffie-Hellman key exchange support")
Cc: stable@vger.kernel.org
Assisted-by: LLM
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Jérémy Jean <Jeremy.Jean@oss.cyber.gouv.fr>
Signed-off-by: Keith Busch <kbusch@kernel.org>

diff --git a/drivers/nvme/target/auth.c b/drivers/nvme/target/auth.c
index a55319bcdbd1..1680e5e6839c 100644
--- a/drivers/nvme/target/auth.c
+++ b/drivers/nvme/target/auth.c
@@ -229,6 +229,7 @@ out_unlock:
 
 void nvmet_auth_sq_free(struct nvmet_sq *sq)
 {
+	sq->auth_expiry_active = false;
 	cancel_delayed_work(&sq->auth_expired_work);
 	kfree(sq->dhchap_c1);
 	sq->dhchap_c1 = NULL;
diff --git a/drivers/nvme/target/fabrics-cmd-auth.c b/drivers/nvme/target/fabrics-cmd-auth.c
index 92f8a76f10ff..e94d8bb8a934 100644
--- a/drivers/nvme/target/fabrics-cmd-auth.c
+++ b/drivers/nvme/target/fabrics-cmd-auth.c
@@ -6,6 +6,7 @@
  */
 #define pr_fmt(fmt) KBUILD_MODNAME ": " fmt
 #include <linux/blkdev.h>
+#include <linux/jiffies.h>
 #include <linux/random.h>
 #include <linux/nvme-auth.h>
 #include <crypto/kpp.h>
@@ -17,16 +18,27 @@ static void nvmet_auth_expired_work(struct work_struct *work)
 	struct nvmet_sq *sq = container_of(to_delayed_work(work),
 			struct nvmet_sq, auth_expired_work);
 
+	mutex_lock(&sq->auth_lock);
+	/* A command may have cancelled or rearmed this timeout. */
+	if (!sq->auth_expiry_active ||
+	    delayed_work_pending(&sq->auth_expired_work))
+		goto out_unlock;
+	sq->auth_expiry_active = false;
+
 	pr_debug("%s: ctrl %d qid %d transaction %u expired, resetting\n",
 		 __func__, sq->ctrl->cntlid, sq->qid, sq->dhchap_tid);
 	sq->dhchap_step = NVME_AUTH_DHCHAP_MESSAGE_NEGOTIATE;
 	sq->dhchap_tid = -1;
+out_unlock:
+	mutex_unlock(&sq->auth_lock);
 }
 
 void nvmet_auth_sq_init(struct nvmet_sq *sq)
 {
 	/* Initialize in-band authentication */
+	mutex_init(&sq->auth_lock);
 	INIT_DELAYED_WORK(&sq->auth_expired_work, nvmet_auth_expired_work);
+	sq->auth_expiry_active = false;
 	sq->authenticated = false;
 	sq->dhchap_step = NVME_AUTH_DHCHAP_MESSAGE_NEGOTIATE;
 }
@@ -256,6 +268,8 @@ void nvmet_execute_auth_send(struct nvmet_req *req)
 	u16 status = 0;
 	u8 dhchap_status;
 
+	mutex_lock(&req->sq->auth_lock);
+
 	if (req->cmd->auth_send.secp != NVME_AUTH_DHCHAP_PROTOCOL_IDENTIFIER) {
 		status = NVME_SC_INVALID_FIELD | NVME_STATUS_DNR;
 		req->error_loc =
@@ -281,8 +295,10 @@ void nvmet_execute_auth_send(struct nvmet_req *req)
 			offsetof(struct nvmf_auth_send_command, tl);
 		goto done;
 	}
-	if (!nvmet_check_transfer_len(req, tl)) {
+	if (unlikely(tl != req->transfer_len)) {
 		pr_debug("%s: transfer length mismatch (%u)\n", __func__, tl);
+		mutex_unlock(&req->sq->auth_lock);
+		nvmet_check_transfer_len(req, tl);
 		return;
 	}
 
@@ -406,10 +422,12 @@ done:
 			 status, req->error_loc);
 	if (req->sq->dhchap_step != NVME_AUTH_DHCHAP_MESSAGE_SUCCESS2 &&
 	    req->sq->dhchap_step != NVME_AUTH_DHCHAP_MESSAGE_FAILURE2) {
-		unsigned long auth_expire_secs = ctrl->kato ? ctrl->kato : 120;
+		unsigned long auth_expire = min_t(u64, MAX_JIFFY_OFFSET,
+			(u64)(ctrl->kato ? ctrl->kato : 120) * HZ);
 
+		req->sq->auth_expiry_active = true;
 		mod_delayed_work(system_percpu_wq, &req->sq->auth_expired_work,
-				 auth_expire_secs * HZ);
+				 auth_expire);
 		goto complete;
 	}
 	/* Final states, clear up variables */
@@ -418,6 +436,7 @@ done:
 		nvmet_ctrl_fatal_error(ctrl);
 
 complete:
+	mutex_unlock(&req->sq->auth_lock);
 	nvmet_req_complete(req, status);
 }
 
@@ -536,6 +555,8 @@ void nvmet_execute_auth_receive(struct nvmet_req *req)
 	u32 al;
 	u16 status = 0;
 
+	mutex_lock(&req->sq->auth_lock);
+
 	if (req->cmd->auth_receive.secp != NVME_AUTH_DHCHAP_PROTOCOL_IDENTIFIER) {
 		status = NVME_SC_INVALID_FIELD | NVME_STATUS_DNR;
 		req->error_loc =
@@ -561,8 +582,10 @@ void nvmet_execute_auth_receive(struct nvmet_req *req)
 			offsetof(struct nvmf_auth_receive_command, al);
 		goto done;
 	}
-	if (!nvmet_check_transfer_len(req, al)) {
+	if (unlikely(al != req->transfer_len)) {
 		pr_debug("%s: transfer length mismatch (%u)\n", __func__, al);
+		mutex_unlock(&req->sq->auth_lock);
+		nvmet_check_transfer_len(req, al);
 		return;
 	}
 
@@ -621,5 +644,6 @@ done:
 		nvmet_auth_sq_free(req->sq);
 		nvmet_ctrl_fatal_error(ctrl);
 	}
+	mutex_unlock(&req->sq->auth_lock);
 	nvmet_req_complete(req, status);
 }
diff --git a/drivers/nvme/target/nvmet.h b/drivers/nvme/target/nvmet.h
index a0bd05603271..e315f1c24e73 100644
--- a/drivers/nvme/target/nvmet.h
+++ b/drivers/nvme/target/nvmet.h
@@ -159,8 +159,10 @@ struct nvmet_sq {
 	u32			sqhd;
 	bool			sqhd_disabled;
 #ifdef CONFIG_NVME_TARGET_AUTH
+	struct mutex		auth_lock;
 	bool			authenticated;
 	struct delayed_work	auth_expired_work;
+	bool			auth_expiry_active;
 	u16			dhchap_tid;
 	u8			sc_c;
 	u8			dhchap_status;


^ permalink raw reply related	[flat|nested] 3+ messages in thread
end of thread, other threads:[~2026-11-05  1:05 UTC | newest]

Thread overview: 3+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2026-11-04 14:29 FAILED: patch "[PATCH] nvmet-auth: serialize access to DH-HMAC-CHAP state to fix UAF" failed to apply to 6.6-stable tree gregkh
2026-11-05  1:05 ` [PATCH 6.6.y 1/2] nvmet: replace use of system_wq with system_percpu_wq Sasha Levin
2026-11-05  1:05   ` [PATCH 6.6.y 2/2] nvmet-auth: serialize access to DH-HMAC-CHAP state to fix UAF Sasha Levin

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox