* [PATCH] nvmet: fix heap out-of-bounds read in nvmet_auth_negotiate()
@ 2026-08-04 3:38 Guixin Liu
2026-08-04 7:24 ` Hannes Reinecke
` (2 more replies)
0 siblings, 3 replies; 5+ messages in thread
From: Guixin Liu @ 2026-08-04 3:38 UTC (permalink / raw)
To: Hannes Reinecke, Christoph Hellwig, Sagi Grimberg,
Chaitanya Kulkarni, Jens Axboe
Cc: linux-nvme
nvmet_execute_auth_send() allocates the DH-HMAC-CHAP message buffer with
the host-supplied transfer length (tl) and hands it to
nvmet_auth_negotiate() without passing tl along. nvmet_auth_negotiate()
then reads the negotiate header and, for each of the halen hash
identifiers and dhlen DH group identifiers, indexes into the fixed
idlist[60] array (hashes at idlist[0..halen), groups at idlist[30..]).
Neither the transfer length nor halen/dhlen is validated. A malicious or
non-conformant host can report a tl smaller than the negotiate structure,
or a halen/dhlen larger than the array (both are u8, up to 255), making
the loops read past the end of the allocated buffer (heap out-of-bounds
read). The sibling nvmet_auth_reply() already validates tl against the
structure size; the negotiate path did not.
Pass tl into nvmet_auth_negotiate(), reject a tl that does not cover the
negotiate data plus one full protocol descriptor, and reject halen/dhlen
larger than NVME_AUTH_DHCHAP_MAX_DH_IDS.
Fixes: db1312dd9548 ("nvmet: implement basic In-Band Authentication")
Signed-off-by: Guixin Liu <kanie@linux.alibaba.com>
---
drivers/nvme/target/fabrics-cmd-auth.c | 12 ++++++++++--
1 file changed, 10 insertions(+), 2 deletions(-)
diff --git a/drivers/nvme/target/fabrics-cmd-auth.c b/drivers/nvme/target/fabrics-cmd-auth.c
index d1b39e64d877..d015968bae2d 100644
--- a/drivers/nvme/target/fabrics-cmd-auth.c
+++ b/drivers/nvme/target/fabrics-cmd-auth.c
@@ -31,12 +31,16 @@ void nvmet_auth_sq_init(struct nvmet_sq *sq)
sq->dhchap_step = NVME_AUTH_DHCHAP_MESSAGE_NEGOTIATE;
}
-static u8 nvmet_auth_negotiate(struct nvmet_req *req, void *d)
+static u8 nvmet_auth_negotiate(struct nvmet_req *req, void *d, u32 tl)
{
struct nvmet_ctrl *ctrl = req->sq->ctrl;
struct nvmf_auth_dhchap_negotiate_data *data = d;
int i, hash_id = 0, fallback_hash_id = 0, dhgid, fallback_dhgid;
+ if (tl < sizeof(*data) +
+ sizeof(struct nvmf_auth_dhchap_protocol_descriptor))
+ return NVME_AUTH_DHCHAP_FAILURE_INCORRECT_PAYLOAD;
+
pr_debug("%s: ctrl %d qid %d: data sc_d %d napd %d authid %d halen %d dhlen %d\n",
__func__, ctrl->cntlid, req->sq->qid,
data->sc_c, data->napd, data->auth_protocol[0].dhchap.authid,
@@ -72,6 +76,10 @@ static u8 nvmet_auth_negotiate(struct nvmet_req *req, void *d)
NVME_AUTH_DHCHAP_AUTH_ID)
return NVME_AUTH_DHCHAP_FAILURE_INCORRECT_PAYLOAD;
+ if (data->auth_protocol[0].dhchap.halen > NVME_AUTH_DHCHAP_MAX_DH_IDS ||
+ data->auth_protocol[0].dhchap.dhlen > NVME_AUTH_DHCHAP_MAX_DH_IDS)
+ return NVME_AUTH_DHCHAP_FAILURE_INCORRECT_PAYLOAD;
+
for (i = 0; i < data->auth_protocol[0].dhchap.halen; i++) {
u8 host_hmac_id = data->auth_protocol[0].dhchap.idlist[i];
@@ -317,7 +325,7 @@ void nvmet_execute_auth_send(struct nvmet_req *req)
} else if (data->auth_id != req->sq->dhchap_step)
goto done_failure1;
/* Validate negotiation parameters */
- dhchap_status = nvmet_auth_negotiate(req, d);
+ dhchap_status = nvmet_auth_negotiate(req, d, tl);
if (dhchap_status == 0)
req->sq->dhchap_step =
NVME_AUTH_DHCHAP_MESSAGE_CHALLENGE;
--
2.43.7
^ permalink raw reply related [flat|nested] 5+ messages in thread* Re: [PATCH] nvmet: fix heap out-of-bounds read in nvmet_auth_negotiate()
2026-08-04 3:38 [PATCH] nvmet: fix heap out-of-bounds read in nvmet_auth_negotiate() Guixin Liu
@ 2026-08-04 7:24 ` Hannes Reinecke
2026-08-04 16:19 ` Christoph Hellwig
2026-08-10 19:37 ` Keith Busch
2 siblings, 0 replies; 5+ messages in thread
From: Hannes Reinecke @ 2026-08-04 7:24 UTC (permalink / raw)
To: Guixin Liu, Christoph Hellwig, Sagi Grimberg, Chaitanya Kulkarni,
Jens Axboe
Cc: linux-nvme
On 8/4/26 5:38 AM, Guixin Liu wrote:
> nvmet_execute_auth_send() allocates the DH-HMAC-CHAP message buffer with
> the host-supplied transfer length (tl) and hands it to
> nvmet_auth_negotiate() without passing tl along. nvmet_auth_negotiate()
> then reads the negotiate header and, for each of the halen hash
> identifiers and dhlen DH group identifiers, indexes into the fixed
> idlist[60] array (hashes at idlist[0..halen), groups at idlist[30..]).
>
> Neither the transfer length nor halen/dhlen is validated. A malicious or
> non-conformant host can report a tl smaller than the negotiate structure,
> or a halen/dhlen larger than the array (both are u8, up to 255), making
> the loops read past the end of the allocated buffer (heap out-of-bounds
> read). The sibling nvmet_auth_reply() already validates tl against the
> structure size; the negotiate path did not.
>
> Pass tl into nvmet_auth_negotiate(), reject a tl that does not cover the
> negotiate data plus one full protocol descriptor, and reject halen/dhlen
> larger than NVME_AUTH_DHCHAP_MAX_DH_IDS.
>
> Fixes: db1312dd9548 ("nvmet: implement basic In-Band Authentication")
> Signed-off-by: Guixin Liu <kanie@linux.alibaba.com>
> ---
> drivers/nvme/target/fabrics-cmd-auth.c | 12 ++++++++++--
> 1 file changed, 10 insertions(+), 2 deletions(-)
>
Reviewed-by: Hannes Reinecke <hare@kernel.org>
Cheers,
Hannes
--
Dr. Hannes Reinecke Kernel Storage Architect
hare@suse.de +49 911 74053 688
SUSE Software Solutions GmbH, Frankenstr. 146, 90461 Nürnberg
HRB 36809 (AG Nürnberg), GF: I. Totev, A. McDonald, W. Knoblich
^ permalink raw reply [flat|nested] 5+ messages in thread* Re: [PATCH] nvmet: fix heap out-of-bounds read in nvmet_auth_negotiate()
2026-08-04 3:38 [PATCH] nvmet: fix heap out-of-bounds read in nvmet_auth_negotiate() Guixin Liu
2026-08-04 7:24 ` Hannes Reinecke
@ 2026-08-04 16:19 ` Christoph Hellwig
2026-08-05 1:48 ` Guixin Liu
2026-08-10 19:37 ` Keith Busch
2 siblings, 1 reply; 5+ messages in thread
From: Christoph Hellwig @ 2026-08-04 16:19 UTC (permalink / raw)
To: Guixin Liu
Cc: Hannes Reinecke, Christoph Hellwig, Sagi Grimberg,
Chaitanya Kulkarni, Jens Axboe, linux-nvme
On Tue, Aug 04, 2026 at 11:38:00AM +0800, Guixin Liu wrote:
> nvmet_execute_auth_send() allocates the DH-HMAC-CHAP message buffer with
> the host-supplied transfer length (tl) and hands it to
> nvmet_auth_negotiate() without passing tl along. nvmet_auth_negotiate()
> then reads the negotiate header and, for each of the halen hash
> identifiers and dhlen DH group identifiers, indexes into the fixed
> idlist[60] array (hashes at idlist[0..halen), groups at idlist[30..]).
>
> Neither the transfer length nor halen/dhlen is validated. A malicious or
> non-conformant host can report a tl smaller than the negotiate structure,
> or a halen/dhlen larger than the array (both are u8, up to 255), making
> the loops read past the end of the allocated buffer (heap out-of-bounds
> read). The sibling nvmet_auth_reply() already validates tl against the
> structure size; the negotiate path did not.
>
> Pass tl into nvmet_auth_negotiate(), reject a tl that does not cover the
> negotiate data plus one full protocol descriptor, and reject halen/dhlen
> larger than NVME_AUTH_DHCHAP_MAX_DH_IDS.
>
> Fixes: db1312dd9548 ("nvmet: implement basic In-Band Authentication")
> Signed-off-by: Guixin Liu <kanie@linux.alibaba.com>
> ---
> drivers/nvme/target/fabrics-cmd-auth.c | 12 ++++++++++--
> 1 file changed, 10 insertions(+), 2 deletions(-)
>
> diff --git a/drivers/nvme/target/fabrics-cmd-auth.c b/drivers/nvme/target/fabrics-cmd-auth.c
> index d1b39e64d877..d015968bae2d 100644
> --- a/drivers/nvme/target/fabrics-cmd-auth.c
> +++ b/drivers/nvme/target/fabrics-cmd-auth.c
> @@ -31,12 +31,16 @@ void nvmet_auth_sq_init(struct nvmet_sq *sq)
> sq->dhchap_step = NVME_AUTH_DHCHAP_MESSAGE_NEGOTIATE;
> }
>
> -static u8 nvmet_auth_negotiate(struct nvmet_req *req, void *d)
> +static u8 nvmet_auth_negotiate(struct nvmet_req *req, void *d, u32 tl)
> {
> struct nvmet_ctrl *ctrl = req->sq->ctrl;
> struct nvmf_auth_dhchap_negotiate_data *data = d;
> int i, hash_id = 0, fallback_hash_id = 0, dhgid, fallback_dhgid;
>
> + if (tl < sizeof(*data) +
> + sizeof(struct nvmf_auth_dhchap_protocol_descriptor))
Odd formatting here. Condition continuations either get two-tab indents
or after the brace:
if (tl < sizeof(*data) +
sizeof(struct nvmf_auth_dhchap_protocol_descriptor))
if (tl <
sizeof(*data) + sizeof(struct nvmf_auth_dhchap_protocol_descriptor))
Otherwise looks good:
Reviewed-by: Christoph Hellwig <hch@lst.de>
^ permalink raw reply [flat|nested] 5+ messages in thread* Re: [PATCH] nvmet: fix heap out-of-bounds read in nvmet_auth_negotiate()
2026-08-04 16:19 ` Christoph Hellwig
@ 2026-08-05 1:48 ` Guixin Liu
0 siblings, 0 replies; 5+ messages in thread
From: Guixin Liu @ 2026-08-05 1:48 UTC (permalink / raw)
To: Christoph Hellwig, Keith Busch
Cc: Hannes Reinecke, Sagi Grimberg, Chaitanya Kulkarni, Jens Axboe,
linux-nvme
在 2026/8/5 00:19, Christoph Hellwig 写道:
> On Tue, Aug 04, 2026 at 11:38:00AM +0800, Guixin Liu wrote:
>> nvmet_execute_auth_send() allocates the DH-HMAC-CHAP message buffer with
>> the host-supplied transfer length (tl) and hands it to
>> nvmet_auth_negotiate() without passing tl along. nvmet_auth_negotiate()
>> then reads the negotiate header and, for each of the halen hash
>> identifiers and dhlen DH group identifiers, indexes into the fixed
>> idlist[60] array (hashes at idlist[0..halen), groups at idlist[30..]).
>>
>> Neither the transfer length nor halen/dhlen is validated. A malicious or
>> non-conformant host can report a tl smaller than the negotiate structure,
>> or a halen/dhlen larger than the array (both are u8, up to 255), making
>> the loops read past the end of the allocated buffer (heap out-of-bounds
>> read). The sibling nvmet_auth_reply() already validates tl against the
>> structure size; the negotiate path did not.
>>
>> Pass tl into nvmet_auth_negotiate(), reject a tl that does not cover the
>> negotiate data plus one full protocol descriptor, and reject halen/dhlen
>> larger than NVME_AUTH_DHCHAP_MAX_DH_IDS.
>>
>> Fixes: db1312dd9548 ("nvmet: implement basic In-Band Authentication")
>> Signed-off-by: Guixin Liu <kanie@linux.alibaba.com>
>> ---
>> drivers/nvme/target/fabrics-cmd-auth.c | 12 ++++++++++--
>> 1 file changed, 10 insertions(+), 2 deletions(-)
>>
>> diff --git a/drivers/nvme/target/fabrics-cmd-auth.c b/drivers/nvme/target/fabrics-cmd-auth.c
>> index d1b39e64d877..d015968bae2d 100644
>> --- a/drivers/nvme/target/fabrics-cmd-auth.c
>> +++ b/drivers/nvme/target/fabrics-cmd-auth.c
>> @@ -31,12 +31,16 @@ void nvmet_auth_sq_init(struct nvmet_sq *sq)
>> sq->dhchap_step = NVME_AUTH_DHCHAP_MESSAGE_NEGOTIATE;
>> }
>>
>> -static u8 nvmet_auth_negotiate(struct nvmet_req *req, void *d)
>> +static u8 nvmet_auth_negotiate(struct nvmet_req *req, void *d, u32 tl)
>> {
>> struct nvmet_ctrl *ctrl = req->sq->ctrl;
>> struct nvmf_auth_dhchap_negotiate_data *data = d;
>> int i, hash_id = 0, fallback_hash_id = 0, dhgid, fallback_dhgid;
>>
>> + if (tl < sizeof(*data) +
>> + sizeof(struct nvmf_auth_dhchap_protocol_descriptor))
> Odd formatting here. Condition continuations either get two-tab indents
> or after the brace:
>
> if (tl < sizeof(*data) +
> sizeof(struct nvmf_auth_dhchap_protocol_descriptor))
>
> if (tl <
> sizeof(*data) + sizeof(struct nvmf_auth_dhchap_protocol_descriptor))
Well, this is an oversight.
Hi Keith, could you please fix this when you apply? Or I can send a v2
patch.
Best Regards,
Guixin Liu
> Otherwise looks good:
>
> Reviewed-by: Christoph Hellwig <hch@lst.de>
^ permalink raw reply [flat|nested] 5+ messages in thread
* Re: [PATCH] nvmet: fix heap out-of-bounds read in nvmet_auth_negotiate()
2026-08-04 3:38 [PATCH] nvmet: fix heap out-of-bounds read in nvmet_auth_negotiate() Guixin Liu
2026-08-04 7:24 ` Hannes Reinecke
2026-08-04 16:19 ` Christoph Hellwig
@ 2026-08-10 19:37 ` Keith Busch
2 siblings, 0 replies; 5+ messages in thread
From: Keith Busch @ 2026-08-10 19:37 UTC (permalink / raw)
To: Guixin Liu
Cc: Hannes Reinecke, Christoph Hellwig, Sagi Grimberg,
Chaitanya Kulkarni, Jens Axboe, linux-nvme
On Tue, Aug 04, 2026 at 11:38:00AM +0800, Guixin Liu wrote:
> nvmet_execute_auth_send() allocates the DH-HMAC-CHAP message buffer with
> the host-supplied transfer length (tl) and hands it to
> nvmet_auth_negotiate() without passing tl along. nvmet_auth_negotiate()
> then reads the negotiate header and, for each of the halen hash
> identifiers and dhlen DH group identifiers, indexes into the fixed
> idlist[60] array (hashes at idlist[0..halen), groups at idlist[30..]).
Thanks, applied to nvme-7.3 with the formatting updated.
^ permalink raw reply [flat|nested] 5+ messages in thread
end of thread, other threads:[~2026-08-10 19:37 UTC | newest]
Thread overview: 5+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2026-08-04 3:38 [PATCH] nvmet: fix heap out-of-bounds read in nvmet_auth_negotiate() Guixin Liu
2026-08-04 7:24 ` Hannes Reinecke
2026-08-04 16:19 ` Christoph Hellwig
2026-08-05 1:48 ` Guixin Liu
2026-08-10 19:37 ` Keith Busch
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.