* [PATCH v2] net/nfc: fix skb ownership on data exchange errors
@ 2026-11-18 8:15 Bin Luo
2026-11-18 8:18 ` netdev-bot+sinfo
2026-11-22 8:39 ` netdev-bot+sashiko
0 siblings, 2 replies; 4+ messages in thread
From: Bin Luo @ 2026-11-18 8:15 UTC (permalink / raw)
To: oe-linux-nfc; +Cc: johan, david+nfc, netdev, Bin Luo
NFC data exchange callbacks may receive a non-NULL response skb together
with an error. The ownership of the skb is not documented, causing the
rawsock and LLCP error paths to return without freeing it. Digital DEP
also frees the response after invoking the callback, which conflicts
with callback ownership.
Document that the callback takes ownership of a non-NULL skb regardless
of err. Free it on the rawsock and LLCP error paths, clear the Digital
DEP response pointer after the callback, and release the skb when the
NCI loopback connection lookup fails.
Signed-off-by: Bin Luo <luob87709@gmail.com>
---
Built on Linux v7.4-rc1 commit 6628e7d4c6b25c8d18f1345e3a20d21829aae7e4
with NFC_VIRTUAL_NCI, KASAN, and kmemleak enabled. A QEMU TCG guest ran
32 NCI 1.0 MIFARE exchanges with a nonzero data status. The root harness
provisioned the virtual controller and polling session; each rawsock
exchange ran with effective UID 65534. The baseline reported 48 kmemleak
suspects; the fixed run reported no unreferenced objects and no KASAN
report.
The runtime test covers NCI-to-rawsock only. It did not trigger the LLCP,
Digital DEP, or NCI loopback connection-lookup failure paths.
---
include/net/nfc/nfc.h | 2 ++
net/nfc/digital_dep.c | 1 +
net/nfc/llcp_core.c | 2 ++
net/nfc/nci/core.c | 1 +
net/nfc/rawsock.c | 5 ++++-
5 files changed, 10 insertions(+), 1 deletion(-)
diff --git a/include/net/nfc/nfc.h b/include/net/nfc/nfc.h
index bcafab5c5..d3c2d6735 100644
--- a/include/net/nfc/nfc.h
+++ b/include/net/nfc/nfc.h
@@ -37,6 +37,8 @@ struct nfc_dev;
*
* When a rx or tx package is lost or corrupted or the target gets out
* of the operating field, err is -EIO.
+ * If skb is non-NULL, ownership is transferred to the callback regardless
+ * of err. Providers must not access or free skb after invoking the callback.
*/
typedef void (*data_exchange_cb_t)(void *context, struct sk_buff *skb,
int err);
diff --git a/net/nfc/digital_dep.c b/net/nfc/digital_dep.c
index 968547c30..c7f5378cd 100644
--- a/net/nfc/digital_dep.c
+++ b/net/nfc/digital_dep.c
@@ -882,6 +882,7 @@ static void digital_in_recv_dep_res(struct nfc_digital_dev *ddev, void *arg,
exit:
data_exch->cb(data_exch->cb_context, resp, rc);
+ resp = NULL;
error:
kfree(data_exch);
diff --git a/net/nfc/llcp_core.c b/net/nfc/llcp_core.c
index 74bf81700..3619e12b3 100644
--- a/net/nfc/llcp_core.c
+++ b/net/nfc/llcp_core.c
@@ -1638,6 +1638,8 @@ void nfc_llcp_recv(void *data, struct sk_buff *skb, int err)
if (err < 0) {
pr_err("LLCP PDU receive err %d\n", err);
+ if (skb)
+ kfree_skb(skb);
return;
}
diff --git a/net/nfc/nci/core.c b/net/nfc/nci/core.c
index 73e3a9647..678f744a2 100644
--- a/net/nfc/nci/core.c
+++ b/net/nfc/nci/core.c
@@ -413,6 +413,7 @@ static void nci_nfcc_loopback_cb(void *context, struct sk_buff *skb, int err)
conn_info = nci_get_conn_info_by_conn_id(ndev, ndev->cur_conn_id);
if (!conn_info) {
+ kfree_skb(skb);
nci_req_complete(ndev, NCI_STATUS_REJECTED);
return;
}
diff --git a/net/nfc/rawsock.c b/net/nfc/rawsock.c
index f7d7a599f..3bf65b1fb 100644
--- a/net/nfc/rawsock.c
+++ b/net/nfc/rawsock.c
@@ -156,8 +156,11 @@ static void rawsock_data_exchange_complete(void *context, struct sk_buff *skb,
pr_debug("sk=%p err=%d\n", sk, err);
- if (err)
+ if (err) {
+ if (skb)
+ kfree_skb(skb);
goto error;
+ }
err = rawsock_add_header(skb);
if (err)
--
2.51.0.windows.1
^ permalink raw reply related [flat|nested] 4+ messages in thread* Re: [PATCH v2] net/nfc: fix skb ownership on data exchange errors
2026-11-18 8:15 [PATCH v2] net/nfc: fix skb ownership on data exchange errors Bin Luo
@ 2026-11-18 8:18 ` netdev-bot+sinfo
2026-11-18 8:23 ` bin luo
2026-11-22 8:39 ` netdev-bot+sashiko
1 sibling, 1 reply; 4+ messages in thread
From: netdev-bot+sinfo @ 2026-11-18 8:18 UTC (permalink / raw)
To: Bin Luo; +Cc: oe-linux-nfc, johan, david+nfc, netdev
Hi!
This is an automated message. This series looks like a fix, but its
commit messages seem to be missing some information:
- How the issue was discovered, e.g. hit in production, hit during
development, syzbot report, manual code inspection, LLM or static
analysis tool scan.
Please do not repost the series just to address the above. Instead,
reply to this email with the missing information, so that reviewers
can take it into account.
*IMPORTANT*: If you do repost (for other reasons) - please include
the relevant info in the cover letter or commit messages.
--
The evaluation is done by an LLM so it may be wrong, if you think
that is the case please reply and explain.
^ permalink raw reply [flat|nested] 4+ messages in thread
* Re: [PATCH v2] net/nfc: fix skb ownership on data exchange errors
2026-11-18 8:15 [PATCH v2] net/nfc: fix skb ownership on data exchange errors Bin Luo
2026-11-18 8:18 ` netdev-bot+sinfo
@ 2026-11-22 8:39 ` netdev-bot+sashiko
1 sibling, 0 replies; 4+ messages in thread
From: netdev-bot+sashiko @ 2026-11-22 8:39 UTC (permalink / raw)
To: luob87709; +Cc: oe-linux-nfc, johan, david+nfc, netdev, kuba
Thank you for your contribution! Sashiko AI review found 2 potential
issue(s) to consider.
Critical: 0 · High: 1 · Medium: 0 · Low: 1
If you believe none of the issues are legitimate, please reply to each
issue of Medium severity or higher and explain why the AI is wrong.
If any of them are legitimate and you plan to respin, please let patchwork
know by including "pw-bot: cr" as a separate line at the end of your reply
(one such reply per series is enough).
- [High] The fix is incomplete.
- [Low] The commit fixes a user-reachable skb leak in
rawsock_data_exchange_complete() (net/nfc/rawsock.c) and changes…
--
net/nfc: fix skb ownership on data exchange errors
The commit documents that a data_exchange_cb_t callback takes ownership
of a non-NULL skb regardless of err. It frees the skb on the rawsock,
LLCP and NCI loopback error paths. It also stops digital DEP from
freeing the response after the callback has been invoked.
> diff --git a/include/net/nfc/nfc.h b/include/net/nfc/nfc.h
> index bcafab5c53e51..d3c2d67357f1b 100644
> --- a/include/net/nfc/nfc.h
> +++ b/include/net/nfc/nfc.h
> @@ -37,6 +37,8 @@ struct nfc_dev;
> *
> * When a rx or tx package is lost or corrupted or the target gets out
> * of the operating field, err is -EIO.
> + * If skb is non-NULL, ownership is transferred to the callback regardless
> + * of err. Providers must not access or free skb after invoking the callback.
> */
> typedef void (*data_exchange_cb_t)(void *context, struct sk_buff *skb,
> int err);
[Severity: High]
This rule now covers every data_exchange_cb_t implementer. Several
in-tree callbacks still return on error without freeing the skb. Is the
fix complete without them?
In net/nfc/nci/hci.c, nci_hci_data_received_cb() is installed as
conn_info->data_exchange_cb and does:
if (err) {
nci_req_complete(ndev, err);
return;
}
nci_rx_data_packet() strips a status byte for every conn_id, based on
the device-wide ndev->target_active_prot. That status ends up in:
nci_rx_data_packet()
nci_add_rx_data_frag(..., nci_to_errno(status))
nci_data_exchange_complete(ndev, skb, conn_id, err)
cb(cb_context, skb, err);
Suppose an NFCC sends an error-status data packet on the HCI conn_id
while a frame-I/F target is active and an RF transceive is pending.
Wouldn't that skb leak? This is the same provider path that was tested
for rawsock.
On the HCI side, nfc_hci_resp_received() calls:
__nfc_hci_cmd_completion(hdev, nfc_hci_result_to_errno(result), skb);
As a result, every non-OK HCI response reaches the callback registered
by nfc_hci_send_cmd_async() with both a non-NULL skb and a nonzero err.
In drivers/nfc/st21nfca/dep.c, st21nfca_im_recv_atr_res_cb() and
st21nfca_im_recv_dep_res_cb() are registered directly as that callback.
Both start with:
if (err != 0)
return;
Would every HCI error response during NFC-DEP leak the response skb
here? For example, a timeout when the peer leaves the field.
Separately, the ST21NFCA_CB_TYPE_READER_F success branch of
st21nfca_im_recv_atr_res_cb() never frees skb at all.
Four default branches only free on success:
default:
if (err == 0)
kfree_skb(skb);
break;
They are hci_transceive_cb() in net/nfc/hci/core.c,
pn544_hci_data_exchange_cb() in drivers/nfc/pn544/pn544.c,
st21nfca_hci_data_exchange_cb() in drivers/nfc/st21nfca/core.c and
microread_im_transceive_cb() in drivers/nfc/microread/microread.c. Under
the new contract, don't these leak skb when err != 0?
nfc_hci_execute_cb() already frees skb on error, so the in-tree
callbacks end up inconsistent with each other and with the documented
rule. Should these callbacks be updated in the same patch, or at least
mentioned in the commit message?
> diff --git a/net/nfc/digital_dep.c b/net/nfc/digital_dep.c
> index 968547c306a5d..c7f5378cd12e2 100644
> --- a/net/nfc/digital_dep.c
> +++ b/net/nfc/digital_dep.c
> @@ -882,6 +882,7 @@ static void digital_in_recv_dep_res(struct nfc_digital_dev *ddev, void *arg,
>
> exit:
> data_exch->cb(data_exch->cb_context, resp, rc);
> + resp = NULL;
>
> error:
> kfree(data_exch);
[Severity: Low]
This isn't a bug, but should this carry a Fixes: tag and Cc: stable? The
rawsock_data_exchange_complete() leak can be reached by an unprivileged
user through an NFC raw socket.
The hunks also depend on each other. The resp = NULL in
digital_in_recv_dep_res() stops the provider from freeing resp, and the
rawsock and LLCP hunks make the consumers free it.
A partial backport would either leak or double free the skb. Fixes:
tags would help stable take the change as one unit.
[ ... ]
--
Sashiko AI review · https://netdev-ai.bots.linux.dev/sashiko/#/patchset/20261118081530.32-1-luob87709%40gmail.com
^ permalink raw reply [flat|nested] 4+ messages in thread
end of thread, other threads:[~2026-11-22 8:39 UTC | newest]
Thread overview: 4+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2026-11-18 8:15 [PATCH v2] net/nfc: fix skb ownership on data exchange errors Bin Luo
2026-11-18 8:18 ` netdev-bot+sinfo
2026-11-18 8:23 ` bin luo
2026-11-22 8:39 ` netdev-bot+sashiko
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox