* [PATCH net 0/2] net/smc: fix potential UAF in smc_pnet_add_ib
@ 2026-03-25 11:03 Li Xiasong
2026-03-25 11:03 ` [PATCH net 1/2] net/smc: fix potential UAF in smc_pnet_add_ib for ib device Li Xiasong
2026-03-25 11:03 ` [PATCH net 2/2] net/smc: fix potential UAF in smc_pnet_add_ib for smcd device Li Xiasong
0 siblings, 2 replies; 5+ messages in thread
From: Li Xiasong @ 2026-03-25 11:03 UTC (permalink / raw)
To: D. Wythe, Dust Li, Sidraya Jayagond, Wenjia Zhang,
Mahanta Jambigi, Tony Lu, Wen Gu, David S. Miller, Eric Dumazet,
Jakub Kicinski, Paolo Abeni, Simon Horman
Cc: linux-rdma, linux-s390, netdev, linux-kernel, yuehaibing,
zhangchangzhong, weiyongjun1
This series fixes potential use-after-free issues in smc_pnet_add_ib()
where the device pointer could be freed between find and apply operations.
The race occurs because smc_pnet_find_ib() and smc_pnet_find_smcd()
release the mutex before returning the device pointer. If the device is
removed (e.g., via smc_ib_remove_dev() or smcd_unregister_dev()) before
smc_pnet_apply_ib() or smc_pnet_apply_smcd() is called, the freed
pointer will be accessed.
Patch 1 fixes the issue for ib device, and patch 2 fixes the same issue
for smcd device.
Li Xiasong (2):
net/smc: fix potential UAF in smc_pnet_add_ib for ib device
net/smc: fix potential UAF in smc_pnet_add_ib for smcd device
net/smc/smc_pnet.c | 121 ++++++++++++++++++++++++++++-----------------
1 file changed, 75 insertions(+), 46 deletions(-)
--
2.34.1
^ permalink raw reply [flat|nested] 5+ messages in thread
* [PATCH net 1/2] net/smc: fix potential UAF in smc_pnet_add_ib for ib device
2026-03-25 11:03 [PATCH net 0/2] net/smc: fix potential UAF in smc_pnet_add_ib Li Xiasong
@ 2026-03-25 11:03 ` Li Xiasong
2026-03-29 20:35 ` Jakub Kicinski
2026-03-25 11:03 ` [PATCH net 2/2] net/smc: fix potential UAF in smc_pnet_add_ib for smcd device Li Xiasong
1 sibling, 1 reply; 5+ messages in thread
From: Li Xiasong @ 2026-03-25 11:03 UTC (permalink / raw)
To: D. Wythe, Dust Li, Sidraya Jayagond, Wenjia Zhang,
Mahanta Jambigi, Tony Lu, Wen Gu, David S. Miller, Eric Dumazet,
Jakub Kicinski, Paolo Abeni, Simon Horman
Cc: linux-rdma, linux-s390, netdev, linux-kernel, yuehaibing,
zhangchangzhong, weiyongjun1
smc_pnet_find_ib() returns an ib device pointer and releases the lock,
then smc_pnet_apply_ib() is called to use this pointer. The device could
be removed between these two calls, leading to a potential use-after-free
when accessing the freed ib_dev pointer.
CPU 0 CPU 1
---- ----
smc_pnet_add_ib()
ib_dev = smc_pnet_find_ib()
mutex_lock(&smc_ib_devices.mutex)
list_for_each_entry() ...
mutex_unlock(&smc_ib_devices.mutex)
smc_ib_remove_dev()
mutex_lock(&smc_ib_devices.mutex)
list_del_init(&smcibdev->list)
mutex_unlock(&smc_ib_devices.mutex)
kfree(smcibdev)
smc_pnet_apply_ib(ib_dev)
ib_dev->pnetid[ib_port - 1] <- UAF (ib_dev already freed)
Fix this by introducing smc_pnet_find_ib_apply() which performs both
find and apply under the same lock, preventing the device from being
removed in between.
Also refactor smc_pnet_apply_ib() into __smc_pnet_apply_ib() (without
lock) and smc_pnet_apply_ib() (with lock) for reuse.
Fixes: 890a2cb4a966 ("net/smc: rework pnet table")
Signed-off-by: Li Xiasong <lixiasong1@huawei.com>
---
net/smc/smc_pnet.c | 63 ++++++++++++++++++++++++++++------------------
1 file changed, 39 insertions(+), 24 deletions(-)
diff --git a/net/smc/smc_pnet.c b/net/smc/smc_pnet.c
index 63e286e2dfaa..91c0b1c473b2 100644
--- a/net/smc/smc_pnet.c
+++ b/net/smc/smc_pnet.c
@@ -245,18 +245,25 @@ static int smc_pnet_remove_by_ndev(struct net_device *ndev)
/* Apply pnetid to ib device when no pnetid is set.
*/
-static bool smc_pnet_apply_ib(struct smc_ib_device *ib_dev, u8 ib_port,
- char *pnet_name)
+static bool __smc_pnet_apply_ib(struct smc_ib_device *ib_dev, u8 ib_port,
+ char *pnet_name)
{
- bool applied = false;
-
- mutex_lock(&smc_ib_devices.mutex);
if (!smc_pnet_is_pnetid_set(ib_dev->pnetid[ib_port - 1])) {
memcpy(ib_dev->pnetid[ib_port - 1], pnet_name,
SMC_MAX_PNETID_LEN);
ib_dev->pnetid_by_user[ib_port - 1] = true;
- applied = true;
+ return true;
}
+ return false;
+}
+
+static bool smc_pnet_apply_ib(struct smc_ib_device *ib_dev, u8 ib_port,
+ char *pnet_name)
+{
+ bool applied;
+
+ mutex_lock(&smc_ib_devices.mutex);
+ applied = __smc_pnet_apply_ib(ib_dev, ib_port, pnet_name);
mutex_unlock(&smc_ib_devices.mutex);
return applied;
}
@@ -305,24 +312,42 @@ static bool smc_pnetid_valid(const char *pnet_name, char *pnetid)
}
/* Find an infiniband device by a given name. The device might not exist. */
-static struct smc_ib_device *smc_pnet_find_ib(char *ib_name)
+static struct smc_ib_device *__smc_pnet_find_ib(char *ib_name)
{
struct smc_ib_device *ibdev;
- mutex_lock(&smc_ib_devices.mutex);
list_for_each_entry(ibdev, &smc_ib_devices.list, list) {
if (!strncmp(ibdev->ibdev->name, ib_name,
sizeof(ibdev->ibdev->name)) ||
(ibdev->ibdev->dev.parent &&
!strncmp(dev_name(ibdev->ibdev->dev.parent), ib_name,
IB_DEVICE_NAME_MAX - 1))) {
- goto out;
+ return ibdev;
}
}
- ibdev = NULL;
-out:
+ return NULL;
+}
+
+/* Find an ib device by name and apply pnetid under lock. */
+static bool smc_pnet_find_ib_apply(char *ib_name, u8 ib_port, char *pnet_name)
+{
+ struct smc_ib_device *ibdev;
+ bool rc = true;
+
+ mutex_lock(&smc_ib_devices.mutex);
+ ibdev = __smc_pnet_find_ib(ib_name);
+ if (ibdev) {
+ if (!__smc_pnet_apply_ib(ibdev, ib_port, pnet_name))
+ rc = false;
+ else
+ pr_warn_ratelimited("smc: ib device %s ibport %d "
+ "applied user defined pnetid "
+ "%.16s\n", ibdev->ibdev->name,
+ ib_port,
+ ibdev->pnetid[ib_port - 1]);
+ }
mutex_unlock(&smc_ib_devices.mutex);
- return ibdev;
+ return rc;
}
/* Find an smcd device by a given name. The device might not exist. */
@@ -412,23 +437,13 @@ static int smc_pnet_add_ib(struct smc_pnettable *pnettable, char *ib_name,
u8 ib_port, char *pnet_name)
{
struct smc_pnetentry *tmp_pe, *new_pe;
- struct smc_ib_device *ib_dev;
bool smcddev_applied = true;
- bool ibdev_applied = true;
+ bool ibdev_applied;
struct smcd_dev *smcd;
bool new_ibdev;
/* try to apply the pnetid to active devices */
- ib_dev = smc_pnet_find_ib(ib_name);
- if (ib_dev) {
- ibdev_applied = smc_pnet_apply_ib(ib_dev, ib_port, pnet_name);
- if (ibdev_applied)
- pr_warn_ratelimited("smc: ib device %s ibport %d "
- "applied user defined pnetid "
- "%.16s\n", ib_dev->ibdev->name,
- ib_port,
- ib_dev->pnetid[ib_port - 1]);
- }
+ ibdev_applied = smc_pnet_find_ib_apply(ib_name, ib_port, pnet_name);
smcd = smc_pnet_find_smcd(ib_name);
if (smcd) {
smcddev_applied = smc_pnet_apply_smcd(smcd, pnet_name);
--
2.34.1
^ permalink raw reply related [flat|nested] 5+ messages in thread
* [PATCH net 2/2] net/smc: fix potential UAF in smc_pnet_add_ib for smcd device
2026-03-25 11:03 [PATCH net 0/2] net/smc: fix potential UAF in smc_pnet_add_ib Li Xiasong
2026-03-25 11:03 ` [PATCH net 1/2] net/smc: fix potential UAF in smc_pnet_add_ib for ib device Li Xiasong
@ 2026-03-25 11:03 ` Li Xiasong
2026-03-29 20:35 ` Jakub Kicinski
1 sibling, 1 reply; 5+ messages in thread
From: Li Xiasong @ 2026-03-25 11:03 UTC (permalink / raw)
To: D. Wythe, Dust Li, Sidraya Jayagond, Wenjia Zhang,
Mahanta Jambigi, Tony Lu, Wen Gu, David S. Miller, Eric Dumazet,
Jakub Kicinski, Paolo Abeni, Simon Horman
Cc: linux-rdma, linux-s390, netdev, linux-kernel, yuehaibing,
zhangchangzhong, weiyongjun1
smc_pnet_find_smcd() returns an smcd device pointer and releases the
lock, then smc_pnet_apply_smcd() is called to use this pointer. The
device could be removed between these two calls, leading to a potential
use-after-free when accessing the freed smcd pointer.
CPU 0 CPU 1
---- ----
smc_pnet_add_ib()
smcd = smc_pnet_find_smcd()
mutex_lock(&smcd_dev_list.mutex)
list_for_each_entry() ...
mutex_unlock(&smcd_dev_list.mutex)
smcd_unregister_dev()
mutex_lock(&smcd_dev_list.mutex)
list_del_init(&smcd->list)
mutex_unlock(&smcd_dev_list.mutex)
kfree(smcd)
smc_pnet_apply_smcd(smcd)
smcd->pnetid <- UAF (smcd already freed)
Fix this by introducing smc_pnet_find_smcd_apply() which performs both
find and apply under the same lock, preventing the device from being
removed in between.
Also refactor smc_pnet_apply_smcd() into __smc_pnet_apply_smcd()
(without lock) and smc_pnet_apply_smcd() (with lock) for reuse.
Fixes: 69baaac9361e ("dibs: Define dibs_client_ops and dibs_dev_ops")
Signed-off-by: Li Xiasong <lixiasong1@huawei.com>
---
net/smc/smc_pnet.c | 58 ++++++++++++++++++++++++++++------------------
1 file changed, 36 insertions(+), 22 deletions(-)
diff --git a/net/smc/smc_pnet.c b/net/smc/smc_pnet.c
index 91c0b1c473b2..a1603740167b 100644
--- a/net/smc/smc_pnet.c
+++ b/net/smc/smc_pnet.c
@@ -270,16 +270,22 @@ static bool smc_pnet_apply_ib(struct smc_ib_device *ib_dev, u8 ib_port,
/* Apply pnetid to smcd device when no pnetid is set.
*/
-static bool smc_pnet_apply_smcd(struct smcd_dev *smcd_dev, char *pnet_name)
+static bool __smc_pnet_apply_smcd(struct smcd_dev *smcd_dev, char *pnet_name)
{
- bool applied = false;
-
- mutex_lock(&smcd_dev_list.mutex);
if (!smc_pnet_is_pnetid_set(smcd_dev->pnetid)) {
memcpy(smcd_dev->pnetid, pnet_name, SMC_MAX_PNETID_LEN);
smcd_dev->pnetid_by_user = true;
- applied = true;
+ return true;
}
+ return false;
+}
+
+static bool smc_pnet_apply_smcd(struct smcd_dev *smcd_dev, char *pnet_name)
+{
+ bool applied;
+
+ mutex_lock(&smcd_dev_list.mutex);
+ applied = __smc_pnet_apply_smcd(smcd_dev, pnet_name);
mutex_unlock(&smcd_dev_list.mutex);
return applied;
}
@@ -351,23 +357,39 @@ static bool smc_pnet_find_ib_apply(char *ib_name, u8 ib_port, char *pnet_name)
}
/* Find an smcd device by a given name. The device might not exist. */
-static struct smcd_dev *smc_pnet_find_smcd(char *smcd_name)
+static struct smcd_dev *__smc_pnet_find_smcd(char *smcd_name)
{
struct smcd_dev *smcd_dev;
- mutex_lock(&smcd_dev_list.mutex);
list_for_each_entry(smcd_dev, &smcd_dev_list.list, list) {
if (!strncmp(dev_name(&smcd_dev->dibs->dev), smcd_name,
IB_DEVICE_NAME_MAX - 1) ||
(smcd_dev->dibs->dev.parent &&
!strncmp(dev_name(smcd_dev->dibs->dev.parent), smcd_name,
IB_DEVICE_NAME_MAX - 1)))
- goto out;
+ return smcd_dev;
+ }
+ return NULL;
+}
+
+/* Find an smcd device by name and apply pnetid under lock. */
+static bool smc_pnet_find_smcd_apply(char *smcd_name, char *pnet_name)
+{
+ struct smcd_dev *smcd_dev;
+ bool rc = true;
+
+ mutex_lock(&smcd_dev_list.mutex);
+ smcd_dev = __smc_pnet_find_smcd(smcd_name);
+ if (smcd_dev) {
+ if (!__smc_pnet_apply_smcd(smcd_dev, pnet_name))
+ rc = false;
+ else
+ pr_warn_ratelimited("smc: smcd device %s applied user defined pnetid %.16s\n",
+ dev_name(&smcd_dev->dibs->dev),
+ smcd_dev->pnetid);
}
- smcd_dev = NULL;
-out:
mutex_unlock(&smcd_dev_list.mutex);
- return smcd_dev;
+ return rc;
}
static int smc_pnet_add_eth(struct smc_pnettable *pnettable, struct net *net,
@@ -437,22 +459,14 @@ static int smc_pnet_add_ib(struct smc_pnettable *pnettable, char *ib_name,
u8 ib_port, char *pnet_name)
{
struct smc_pnetentry *tmp_pe, *new_pe;
- bool smcddev_applied = true;
+ bool smcddev_applied;
bool ibdev_applied;
- struct smcd_dev *smcd;
bool new_ibdev;
/* try to apply the pnetid to active devices */
ibdev_applied = smc_pnet_find_ib_apply(ib_name, ib_port, pnet_name);
- smcd = smc_pnet_find_smcd(ib_name);
- if (smcd) {
- smcddev_applied = smc_pnet_apply_smcd(smcd, pnet_name);
- if (smcddev_applied) {
- pr_warn_ratelimited("smc: smcd device %s applied user defined pnetid %.16s\n",
- dev_name(&smcd->dibs->dev),
- smcd->pnetid);
- }
- }
+ smcddev_applied = smc_pnet_find_smcd_apply(ib_name, pnet_name);
+
/* Apply fails when a device has a hardware-defined pnetid set, do not
* add a pnet table entry in that case.
*/
--
2.34.1
^ permalink raw reply related [flat|nested] 5+ messages in thread
* Re: [PATCH net 1/2] net/smc: fix potential UAF in smc_pnet_add_ib for ib device
2026-03-25 11:03 ` [PATCH net 1/2] net/smc: fix potential UAF in smc_pnet_add_ib for ib device Li Xiasong
@ 2026-03-29 20:35 ` Jakub Kicinski
0 siblings, 0 replies; 5+ messages in thread
From: Jakub Kicinski @ 2026-03-29 20:35 UTC (permalink / raw)
To: lixiasong1
Cc: Jakub Kicinski, alibuda, dust.li, sidraya, wenjia, mjambigi,
tonylu, guwen, davem, edumazet, pabeni, horms, linux-rdma,
linux-s390, netdev, linux-kernel, yuehaibing, zhangchangzhong,
weiyongjun1
This is an AI-generated review of your patch. The human sending this
email has considered the AI review valid, or at least plausible.
---
No issues found.
--
pw-bot: cr
^ permalink raw reply [flat|nested] 5+ messages in thread
* Re: [PATCH net 2/2] net/smc: fix potential UAF in smc_pnet_add_ib for smcd device
2026-03-25 11:03 ` [PATCH net 2/2] net/smc: fix potential UAF in smc_pnet_add_ib for smcd device Li Xiasong
@ 2026-03-29 20:35 ` Jakub Kicinski
0 siblings, 0 replies; 5+ messages in thread
From: Jakub Kicinski @ 2026-03-29 20:35 UTC (permalink / raw)
To: lixiasong1
Cc: Jakub Kicinski, alibuda, dust.li, sidraya, wenjia, mjambigi,
tonylu, guwen, davem, edumazet, pabeni, horms, linux-rdma,
linux-s390, netdev, linux-kernel, yuehaibing, zhangchangzhong,
weiyongjun1
This is an AI-generated review of your patch. The human sending this
email has considered the AI review valid, or at least plausible.
---
net/smc: fix potential UAF in smc_pnet_add_ib for smcd device
This commit fixes a potential use-after-free when accessing an smcd device
in smc_pnet_add_ib(). It resolves the issue by combining the lookup and
apply operations into a single function that holds the smcd_dev_list.mutex
for the duration of both operations.
> diff --git a/net/smc/smc_pnet.c b/net/smc/smc_pnet.c
> index 91c0b1c473b2..a1603740167b 100644
> --- a/net/smc/smc_pnet.c
> +++ b/net/smc/smc_pnet.c
[ ... ]
> @@ -351,23 +357,39 @@ static bool smc_pnet_find_ib_apply(char *ib_name, u8 ib_port, char *pnet_name)
> }
>
> /* Find an smcd device by a given name. The device might not exist. */
> -static struct smcd_dev *smc_pnet_find_smcd(char *smcd_name)
> +static struct smcd_dev *__smc_pnet_find_smcd(char *smcd_name)
> {
> struct smcd_dev *smcd_dev;
>
> - mutex_lock(&smcd_dev_list.mutex);
> list_for_each_entry(smcd_dev, &smcd_dev_list.list, list) {
> if (!strncmp(dev_name(&smcd_dev->dibs->dev), smcd_name,
> IB_DEVICE_NAME_MAX - 1) ||
> (smcd_dev->dibs->dev.parent &&
> !strncmp(dev_name(smcd_dev->dibs->dev.parent), smcd_name,
> IB_DEVICE_NAME_MAX - 1)))
> - goto out;
> + return smcd_dev;
> + }
> + return NULL;
> +}
> +
> +/* Find an smcd device by name and apply pnetid under lock. */
> +static bool smc_pnet_find_smcd_apply(char *smcd_name, char *pnet_name)
> +{
> + struct smcd_dev *smcd_dev;
> + bool rc = true;
> +
> + mutex_lock(&smcd_dev_list.mutex);
> + smcd_dev = __smc_pnet_find_smcd(smcd_name);
> + if (smcd_dev) {
> + if (!__smc_pnet_apply_smcd(smcd_dev, pnet_name))
While this prevents the use-after-free in smc_pnet_add_ib() by expanding the
mutex coverage, does the same vulnerability pattern still exist in the
connection handshake paths?
Neither struct smcd_dev nor struct smc_ib_device appear to use reference
counting. During a connection handshake, functions like
smc_pnet_find_ism_resource() and _smc_pnet_find_roce_by_pnetid() find a
device under the mutex, store the raw pointer (e.g., into ini->ism_dev[0]),
and immediately release the mutex.
If a concurrent hot-unplug event triggers smcd_unregister_dev() or
smc_ib_remove_dev() while the handshake process is accessing the device
without holding a lock or reference, could it lead to another use-after-free?
Would a more complete fix require introducing reference counting for both
device structures to ensure they remain valid while pointed to by the
handshake structures?
> + rc = false;
> + else
> + pr_warn_ratelimited("smc: smcd device %s applied user defined pnetid %.16s\n",
> + dev_name(&smcd_dev->dibs->dev),
> + smcd_dev->pnetid);
> }
> - smcd_dev = NULL;
> -out:
> mutex_unlock(&smcd_dev_list.mutex);
> - return smcd_dev;
> + return rc;
> }
^ permalink raw reply [flat|nested] 5+ messages in thread
end of thread, other threads:[~2026-03-29 20:35 UTC | newest]
Thread overview: 5+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2026-03-25 11:03 [PATCH net 0/2] net/smc: fix potential UAF in smc_pnet_add_ib Li Xiasong
2026-03-25 11:03 ` [PATCH net 1/2] net/smc: fix potential UAF in smc_pnet_add_ib for ib device Li Xiasong
2026-03-29 20:35 ` Jakub Kicinski
2026-03-25 11:03 ` [PATCH net 2/2] net/smc: fix potential UAF in smc_pnet_add_ib for smcd device Li Xiasong
2026-03-29 20:35 ` Jakub Kicinski
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox