Re: [PATCH net-next v5 2/3] net: prevent mixing frags from different memory providers or bindings

From
Stanislav Fomichev <hidden>
Date
2026-11-16 22:40:24
Also in
lkml
Reply to
Mina Almasry, 2026-11-16
On 11/16, Mina Almasry wrote:
On Mon, Nov 16, 2026 at 11:05 AM Stanislav Fomichev
[off-list ref] wrote:
quoted
On 11/15, Mina Almasry wrote:
quoted
Prevent coalescing, shifting, or appending skb fragments from
incompatible memory providers or different devmem bindings.

Add skb_net_iov_provider(), skb_mem_provider_compat(), and
skb_frags_can_coalesce() to check readability and fragment provider
compatibility (net_iov_owner(niov) for devmem, niov->desc.pp for other
net_iov providers such as io_uring zcrx, and NULL for page-backed skbs)
across skb_try_coalesce(), skb_shift(), skb_gro_receive(),
skb_gro_receive_list(), tcp_skb_can_collapse(), and
tcp_skb_can_collapse_rx().

Also add net_devmem_skb_binding() and use it in tcp_sendmsg_locked(),
zerocopy_fill_skb_from_devmem(), and validate_xmit_unreadable_skb() so
TX skbs cannot mix different devmem bindings or transmit non-devmem
unreadable frags on DMA TX devices. Reject converting linear head_frag
payloads into page frags on unreadable skbs in skb_try_coalesce() and
skb_gro_receive(), and propagate skb->unreadable and compare
skb_frag_netmem() in tcp_clone_payload().

Update netmem_ref_test_net_iov and netmem_ref_test_try_coalesce so their
target skbs carry a compatible unreadable net_iov frag before shifting
or coalescing.

This issue was found by manual code inspection and LLM review; it was
not triggered in production. Verified with the net_core KUnit test
suite; not tested on real hardware.

Fixes: 65249feb6b3d ("net: add support for skbs with unreadable frags")
Fixes: bd61848900bf ("net: devmem: Implement TX path")
Signed-off-by: Mina Almasry <redacted>
---
 include/linux/skbuff.h | 55 ++++++++++++++++++++++++++++++++++++++++++
 include/net/tcp.h      |  5 ++--
 net/core/datagram.c    |  4 +--
 net/core/dev.c         | 14 ++++-------
 net/core/devmem.h      | 22 +++++++++++++++++
 net/core/gro.c         | 16 ++++--------
 net/core/net_test.c    | 42 +++++++++++++++++++++++++++-----
 net/core/skbuff.c      | 24 +++++++-----------
 net/ipv4/tcp.c         |  3 ++-
 net/ipv4/tcp_output.c  |  4 ++-
 10 files changed, 142 insertions(+), 47 deletions(-)
diff --git a/include/linux/skbuff.h b/include/linux/skbuff.h
index 1d6832878930..867e3ab17c17 100644
--- a/include/linux/skbuff.h
+++ b/include/linux/skbuff.h
@@ -3770,6 +3770,61 @@ static inline void *skb_frag_address(const skb_frag_t *frag)
      return page_address(skb_frag_page(frag)) + skb_frag_off(frag);
 }

+/* Return an opaque pointer identifying the memory provider instance of @skb's
+ * net_iov fragments:
+ * - For devmem (NET_IOV_DMABUF), net_iov_owner(niov) (&binding->area, 1:1 with
+ *   the binding across RX queues and TX).
+ * - For other net_iov providers (e.g., io_uring zcrx), niov->desc.pp (shared
+ *   across all areas of an io_zcrx_ifq).
+ */
+static inline void *skb_net_iov_provider(const struct sk_buff *skb)
+{
+     struct net_iov *niov;
+
+     if (skb_frags_readable(skb) || !skb_shinfo(skb)->nr_frags)
+             return NULL;
+
+     niov = skb_frag_net_iov(&skb_shinfo(skb)->frags[0]);
+     if (!niov)
+             return NULL;
+
+     if (net_is_devmem_iov(niov))
+             return net_iov_owner(niov);
+
+     return niov->desc.pp;
+}
+
+/* Return true if the memory providers of @skb1 and @skb2 are compatible
+ * for coalescing. For any struct page-backed provider (the buddy allocator
+ * or a future provider handing out struct pages), coalescing is assumed
+ * feasible. For net_iov-backed providers, the provider instance pointer
+ * must match.
+ */
+static inline bool skb_mem_provider_compat(const struct sk_buff *skb1,
+                                        const struct sk_buff *skb2)
+{
+     /* Check skb_frags_readable() first to reject readable/unreadable
+      * mismatches and avoid chasing frags[0] and struct net_iov when
+      * only one skb has unreadable frags.
+      */
+     if (skb_frags_readable(skb1) != skb_frags_readable(skb2))
+             return false;
+
+     return skb_net_iov_provider(skb1) == skb_net_iov_provider(skb2);
+}
+
+static inline bool skb_frags_can_coalesce(struct sk_buff *to,
+                                       struct sk_buff *from)
+{
+     if (to->pp_recycle != from->pp_recycle)
+             return false;
+
+     if (skb_zcopy(to) || skb_zcopy(from))
+             return false;
+
+     return skb_mem_provider_compat(to, from);
+}
+
 /**
  * skb_frag_address_safe - gets the address of the data contained in a paged fragment
  * @frag: the paged fragment buffer
diff --git a/include/net/tcp.h b/include/net/tcp.h
index 25d11fab1b03..bf5f28e52fca 100644
--- a/include/net/tcp.h
+++ b/include/net/tcp.h
@@ -1275,14 +1275,15 @@ static inline bool tcp_skb_can_collapse(const struct sk_buff *to,
                    mptcp_skb_can_collapse(to, from) &&
                    !skb_cmp_decrypted(to, from) &&
                    skb_pure_zcopy_same(to, from) &&
-                   skb_frags_readable(to) == skb_frags_readable(from));
+                   skb_mem_provider_compat(to, from));
 }

 static inline bool tcp_skb_can_collapse_rx(const struct sk_buff *to,
                                         const struct sk_buff *from)
 {
      return likely(mptcp_skb_can_collapse(to, from) &&
-                   !skb_cmp_decrypted(to, from));
+                   !skb_cmp_decrypted(to, from) &&
+                   skb_mem_provider_compat(to, from));
 }

 /* Events passed to congestion control interface */
diff --git a/net/core/datagram.c b/net/core/datagram.c
index ecc30a6f8302..88cc0a6c62f6 100644
--- a/net/core/datagram.c
+++ b/net/core/datagram.c
@@ -712,8 +712,8 @@ zerocopy_fill_skb_from_devmem(struct sk_buff *skb, struct iov_iter *from,
      size_t virt_addr, size, off;
      struct net_iov *niov;

-     if (i && skb_frags_readable(skb))
-             return -EFAULT;
+     if (i && net_devmem_skb_binding(skb) != binding)
+             return -EMSGSIZE;

      /* Devmem filling works by taking an IOVEC from the user where the
       * iov_addrs are interpreted as an offset in bytes into the dma-buf to
diff --git a/net/core/dev.c b/net/core/dev.c
index a230ecce9327..51f5e4800eee 100644
--- a/net/core/dev.c
+++ b/net/core/dev.c
[..]
quoted
@@ -4175,8 +4175,7 @@ static struct sk_buff *sk_validate_xmit_skb(struct sk_buff *skb,
 static struct sk_buff *validate_xmit_unreadable_skb(struct sk_buff *skb,
                                                  struct net_device *dev)
 {
-     struct skb_shared_info *shinfo;
-     struct net_iov *niov;
+     struct net_devmem_dmabuf_binding *binding;

      if (likely(skb_frags_readable(skb) ||
                 dev->netmem_tx == NETMEM_TX_NO_DMA))
@@ -4185,14 +4184,11 @@ static struct sk_buff *validate_xmit_unreadable_skb(struct sk_buff *skb,
      if (dev->netmem_tx == NETMEM_TX_NONE)
              goto out_free;

-     shinfo = skb_shinfo(skb);
-
      /* All net_iovs in an unreadable skb share the same binding. */
-     if (shinfo->nr_frags > 0) {
-             niov = netmem_to_net_iov(skb_frag_netmem(&shinfo->frags[0]));
-             if (net_is_devmem_iov(niov) &&
-                 (net_devmem_iov_binding(niov)->direction != DMA_TO_DEVICE ||
-                  READ_ONCE(net_devmem_iov_binding(niov)->dev) != dev))
+     if (skb_shinfo(skb)->nr_frags > 0) {
+             binding = net_devmem_skb_binding(skb);
+             if (!binding || binding->direction != DMA_TO_DEVICE ||
+                 READ_ONCE(binding->dev) != dev)
                      goto out_free;
      }
Should we limit this patch to the real fixes only? This part does
not look essential. validate_xmit_unreadable_skb already does skb_frags_readable
check above, it already does nr_frags > 0 check here...
so why go through the same stuff again via net_devmem_skb_binding?
So this is a real fix actually, although I agree it's kinda clobbered
into a patch with an unrelated fix. The existing checks do not catch
io_uring zcrx RX packets that have been received and the stack is
trying to forward them over a netdev where the memory is not mapped,
and this patch fixes that.
quoted
If you want, maybe move this to a separate patch at least? And do
something like:

binding = net_devmem_skb_binding(skb);
/* I hope likely() below properly unwraps net_devmem_skb_binding similar
 * to existing likely(skb_frags_readable) case... */
if (likely(!binding)) goto out;
if (dev->netmem_tx == NETMEM_TX_NO_DMA) goto out;
if (dev->netmem_tx == NETMEM_TX_NONE) goto out_free;
if (READ_ONCE(binding->dev) != dev) goto out_free;
return skb;

And call it cleanup with no fixes tags.
I could break it into a separate patch and reduce some of the
duplicated checks, but it's still a fix. It may be good as-is once you
realize it's a necessary change?
I see, so that desc.pp part is missing currently :-/

Acked-by: Stanislav Fomichev <sdf@fomichev.me>

If you happen to repost, I'd remove unrelated `struct skb_shared_info *shinfo`
removal/etc to make it more clear.
Thread (11 messages) 11 messages, 3 authors, last reply 2d ago
Keyboard shortcuts
hback out one level
jnext message in thread / next series-diff position
kprevious message in thread / previous series-diff position
ldrill in
Escclose help / fold thread tree
?toggle this help