[PATCH net v2] xfrm: esp: avoid in-place decrypt on shared skb frags

[PATCH net v2] xfrm: esp: avoid in-place decrypt on shared skb frags

[HexRabbit]

From: Kuan-Ting Chen <h3xrabbit@gmail.com>

MSG_SPLICE_PAGES can attach pages from a pipe directly to an skb. TCP
marks such skbs with SKBFL_SHARED_FRAG after skb_splice_from_iter(),
so later paths that may modify packet data can first make a private
copy. The IPv4/IPv6 datagram append paths did not set this flag when
splicing pages into UDP skbs.

That leaves an ESP-in-UDP packet made from shared pipe pages looking
like an ordinary uncloned nonlinear skb. ESP input then takes the no-COW
fast path for uncloned skbs without a frag_list and decrypts in place
over data that is not owned privately by the skb.

Mark IPv4/IPv6 datagram splice frags with SKBFL_SHARED_FRAG, matching
TCP. Also make ESP input fall back to skb_cow_data() when the flag is
present, so ESP does not decrypt externally backed frags in place.
Private nonlinear skb frags still use the existing fast path.

This intentionally does not change ESP output. In esp_output_head(),
the path that appends the ESP trailer to existing skb tailroom without
calling skb_cow_data() is not reachable for nonlinear skbs:
skb_tailroom() returns zero when skb->data_len is nonzero, while ESP
tailen is positive. Thus ESP output will either use the separate
destination-frag path or fall back to skb_cow_data().

Fixes: cac2661c53f3 ("esp4: Avoid skb_cow_data whenever possible")
Fixes: 03e2a30f6a27 ("esp6: Avoid skb_cow_data whenever possible")
Fixes: 7da0dde68486 ("ip, udp: Support MSG_SPLICE_PAGES")
Fixes: 6d8192bd69bb ("ip6, udp6: Support MSG_SPLICE_PAGES")
Reported-by: Hyunwoo Kim <imv4bel@gmail.com>
Reported-by: Kuan-Ting Chen <h3xrabbit@gmail.com>
Cc: stable@vger.kernel.org
Signed-off-by: Kuan-Ting Chen <h3xrabbit@gmail.com>
---
v2:
- Add Fixes tags
- Add stable Cc and Reported-by trailers.

 net/ipv4/esp4.c       | 3 ++-
 net/ipv4/ip_output.c  | 2 ++
 net/ipv6/esp6.c       | 3 ++-
 net/ipv6/ip6_output.c | 2 ++
 4 files changed, 8 insertions(+), 2 deletions(-)

diff --git a/net/ipv4/esp4.c b/net/ipv4/esp4.c
index 6dfc0bcde..6a5febbdb 100644
--- a/net/ipv4/esp4.c
+++ b/net/ipv4/esp4.c
@@ -873,7 +873,8 @@ static int esp_input(struct xfrm_state *x, struct sk_buff *skb)
 			nfrags = 1;
 
 			goto skip_cow;
-		} else if (!skb_has_frag_list(skb)) {
+		} else if (!skb_has_frag_list(skb) &&
+			   !skb_has_shared_frag(skb)) {
 			nfrags = skb_shinfo(skb)->nr_frags;
 			nfrags++;
 
diff --git a/net/ipv4/ip_output.c b/net/ipv4/ip_output.c
index e4790cc7b..5bcd73cbd 100644
--- a/net/ipv4/ip_output.c
+++ b/net/ipv4/ip_output.c
@@ -1233,6 +1233,8 @@ static int __ip_append_data(struct sock *sk,
 			if (err < 0)
 				goto error;
 			copy = err;
+			if (!(flags & MSG_NO_SHARED_FRAGS))
+				skb_shinfo(skb)->flags |= SKBFL_SHARED_FRAG;
 			wmem_alloc_delta += copy;
 		} else if (!zc) {
 			int i = skb_shinfo(skb)->nr_frags;
diff --git a/net/ipv6/esp6.c b/net/ipv6/esp6.c
index 9f7531373..9c06c5a14 100644
--- a/net/ipv6/esp6.c
+++ b/net/ipv6/esp6.c
@@ -915,7 +915,8 @@ static int esp6_input(struct xfrm_state *x, struct sk_buff *skb)
 			nfrags = 1;
 
 			goto skip_cow;
-		} else if (!skb_has_frag_list(skb)) {
+		} else if (!skb_has_frag_list(skb) &&
+			   !skb_has_shared_frag(skb)) {
 			nfrags = skb_shinfo(skb)->nr_frags;
 			nfrags++;
 
diff --git a/net/ipv6/ip6_output.c b/net/ipv6/ip6_output.c
index 7e92909ab..1f2a33fbe 100644
--- a/net/ipv6/ip6_output.c
+++ b/net/ipv6/ip6_output.c
@@ -1794,6 +1794,8 @@ static int __ip6_append_data(struct sock *sk,
 			if (err < 0)
 				goto error;
 			copy = err;
+			if (!(flags & MSG_NO_SHARED_FRAGS))
+				skb_shinfo(skb)->flags |= SKBFL_SHARED_FRAG;
 			wmem_alloc_delta += copy;
 		} else if (!zc) {
 			int i = skb_shinfo(skb)->nr_frags;
-- 
2.43.0

Re: [PATCH net v2] xfrm: esp: avoid in-place decrypt on shared skb frags

[Hyunwoo Kim]

On Mon, May 04, 2026 at 11:27:12PM +0800, HexRabbit wrote:
> From: Kuan-Ting Chen <h3xrabbit@gmail.com>
> 
> MSG_SPLICE_PAGES can attach pages from a pipe directly to an skb. TCP
> marks such skbs with SKBFL_SHARED_FRAG after skb_splice_from_iter(),
> so later paths that may modify packet data can first make a private
> copy. The IPv4/IPv6 datagram append paths did not set this flag when
> splicing pages into UDP skbs.
> 
> That leaves an ESP-in-UDP packet made from shared pipe pages looking
> like an ordinary uncloned nonlinear skb. ESP input then takes the no-COW
> fast path for uncloned skbs without a frag_list and decrypts in place
> over data that is not owned privately by the skb.
> 
> Mark IPv4/IPv6 datagram splice frags with SKBFL_SHARED_FRAG, matching
> TCP. Also make ESP input fall back to skb_cow_data() when the flag is
> present, so ESP does not decrypt externally backed frags in place.
> Private nonlinear skb frags still use the existing fast path.
> 
> This intentionally does not change ESP output. In esp_output_head(),
> the path that appends the ESP trailer to existing skb tailroom without
> calling skb_cow_data() is not reachable for nonlinear skbs:
> skb_tailroom() returns zero when skb->data_len is nonzero, while ESP
> tailen is positive. Thus ESP output will either use the separate
> destination-frag path or fall back to skb_cow_data().
> 
> Fixes: cac2661c53f3 ("esp4: Avoid skb_cow_data whenever possible")
> Fixes: 03e2a30f6a27 ("esp6: Avoid skb_cow_data whenever possible")
> Fixes: 7da0dde68486 ("ip, udp: Support MSG_SPLICE_PAGES")
> Fixes: 6d8192bd69bb ("ip6, udp6: Support MSG_SPLICE_PAGES")
> Reported-by: Hyunwoo Kim <imv4bel@gmail.com>
> Reported-by: Kuan-Ting Chen <h3xrabbit@gmail.com>

I dynamically tested this patch and confirm it resolves the
issue. Clean work.

One correction request before merge -- please drop the
second Reported-by tag (your own) from the trailer.

The report and patch for this issue were already posted on
the public netdev ML 6 days ago, i.e., the bug was already
publicly reported:

  https://lore.kernel.org/all/afLDKSvAvMwGh7Fy@v4bel/

Credit for patch authorship is adequately covered by
Signed-off-by alone. Setting aside that your work proceeded
independently rather than as a review of my earlier
submission, the trailer should conform to convention to
avoid future misunderstanding.

No objections to the patch itself.


Best regards,
Hyunwoo Kim

Re: [PATCH net v2] xfrm: esp: avoid in-place decrypt on shared skb frags

[Steffen Klassert]

On Tue, May 05, 2026 at 03:01:15AM +0900, Hyunwoo Kim wrote:
> On Mon, May 04, 2026 at 11:27:12PM +0800, HexRabbit wrote:
> > From: Kuan-Ting Chen <h3xrabbit@gmail.com>
> > 
> > MSG_SPLICE_PAGES can attach pages from a pipe directly to an skb. TCP
> > marks such skbs with SKBFL_SHARED_FRAG after skb_splice_from_iter(),
> > so later paths that may modify packet data can first make a private
> > copy. The IPv4/IPv6 datagram append paths did not set this flag when
> > splicing pages into UDP skbs.
> > 
> > That leaves an ESP-in-UDP packet made from shared pipe pages looking
> > like an ordinary uncloned nonlinear skb. ESP input then takes the no-COW
> > fast path for uncloned skbs without a frag_list and decrypts in place
> > over data that is not owned privately by the skb.
> > 
> > Mark IPv4/IPv6 datagram splice frags with SKBFL_SHARED_FRAG, matching
> > TCP. Also make ESP input fall back to skb_cow_data() when the flag is
> > present, so ESP does not decrypt externally backed frags in place.
> > Private nonlinear skb frags still use the existing fast path.
> > 
> > This intentionally does not change ESP output. In esp_output_head(),
> > the path that appends the ESP trailer to existing skb tailroom without
> > calling skb_cow_data() is not reachable for nonlinear skbs:
> > skb_tailroom() returns zero when skb->data_len is nonzero, while ESP
> > tailen is positive. Thus ESP output will either use the separate
> > destination-frag path or fall back to skb_cow_data().
> > 
> > Fixes: cac2661c53f3 ("esp4: Avoid skb_cow_data whenever possible")
> > Fixes: 03e2a30f6a27 ("esp6: Avoid skb_cow_data whenever possible")
> > Fixes: 7da0dde68486 ("ip, udp: Support MSG_SPLICE_PAGES")
> > Fixes: 6d8192bd69bb ("ip6, udp6: Support MSG_SPLICE_PAGES")
> > Reported-by: Hyunwoo Kim <imv4bel@gmail.com>
> > Reported-by: Kuan-Ting Chen <h3xrabbit@gmail.com>
> 
> I dynamically tested this patch and confirm it resolves the
> issue. Clean work.

Feel free to add a Tested-by: tag.

> One correction request before merge -- please drop the
> second Reported-by tag (your own) from the trailer.
> 
> The report and patch for this issue were already posted on
> the public netdev ML 6 days ago, i.e., the bug was already
> publicly reported:
> 
>   https://lore.kernel.org/all/afLDKSvAvMwGh7Fy@v4bel/
> 
> Credit for patch authorship is adequately covered by
> Signed-off-by alone. Setting aside that your work proceeded
> independently rather than as a review of my earlier
> submission, the trailer should conform to convention to
> avoid future misunderstanding.

The issue was reported independently, so both Reported-by tags
are valid. But indeed the Signed-off-by tag should cover the
second one. I've applied it to the testing branch of the ipsec
tree to make it available to our test systems. I can still fix
the tags on request, no need for a v3.

> No objections to the patch itself.

Thanks a lot for your effort!

Re: [PATCH net v2] xfrm: esp: avoid in-place decrypt on shared skb frags

[Hyunwoo Kim]

On Mon, May 04, 2026 at 08:47:04PM +0200, Steffen Klassert wrote:
> On Tue, May 05, 2026 at 03:01:15AM +0900, Hyunwoo Kim wrote:
> > On Mon, May 04, 2026 at 11:27:12PM +0800, HexRabbit wrote:
> > > From: Kuan-Ting Chen <h3xrabbit@gmail.com>
> > > 
> > > MSG_SPLICE_PAGES can attach pages from a pipe directly to an skb. TCP
> > > marks such skbs with SKBFL_SHARED_FRAG after skb_splice_from_iter(),
> > > so later paths that may modify packet data can first make a private
> > > copy. The IPv4/IPv6 datagram append paths did not set this flag when
> > > splicing pages into UDP skbs.
> > > 
> > > That leaves an ESP-in-UDP packet made from shared pipe pages looking
> > > like an ordinary uncloned nonlinear skb. ESP input then takes the no-COW
> > > fast path for uncloned skbs without a frag_list and decrypts in place
> > > over data that is not owned privately by the skb.
> > > 
> > > Mark IPv4/IPv6 datagram splice frags with SKBFL_SHARED_FRAG, matching
> > > TCP. Also make ESP input fall back to skb_cow_data() when the flag is
> > > present, so ESP does not decrypt externally backed frags in place.
> > > Private nonlinear skb frags still use the existing fast path.
> > > 
> > > This intentionally does not change ESP output. In esp_output_head(),
> > > the path that appends the ESP trailer to existing skb tailroom without
> > > calling skb_cow_data() is not reachable for nonlinear skbs:
> > > skb_tailroom() returns zero when skb->data_len is nonzero, while ESP
> > > tailen is positive. Thus ESP output will either use the separate
> > > destination-frag path or fall back to skb_cow_data().
> > > 
> > > Fixes: cac2661c53f3 ("esp4: Avoid skb_cow_data whenever possible")
> > > Fixes: 03e2a30f6a27 ("esp6: Avoid skb_cow_data whenever possible")
> > > Fixes: 7da0dde68486 ("ip, udp: Support MSG_SPLICE_PAGES")
> > > Fixes: 6d8192bd69bb ("ip6, udp6: Support MSG_SPLICE_PAGES")
> > > Reported-by: Hyunwoo Kim <imv4bel@gmail.com>
> > > Reported-by: Kuan-Ting Chen <h3xrabbit@gmail.com>
> > 
> > I dynamically tested this patch and confirm it resolves the
> > issue. Clean work.
> 
> Feel free to add a Tested-by: tag.
> 
> > One correction request before merge -- please drop the
> > second Reported-by tag (your own) from the trailer.
> > 
> > The report and patch for this issue were already posted on
> > the public netdev ML 6 days ago, i.e., the bug was already
> > publicly reported:
> > 
> >   https://lore.kernel.org/all/afLDKSvAvMwGh7Fy@v4bel/
> > 
> > Credit for patch authorship is adequately covered by
> > Signed-off-by alone. Setting aside that your work proceeded
> > independently rather than as a review of my earlier
> > submission, the trailer should conform to convention to
> > avoid future misunderstanding.
> 
> The issue was reported independently, so both Reported-by tags
> are valid. But indeed the Signed-off-by tag should cover the
> second one. I've applied it to the testing branch of the ipsec
> tree to make it available to our test systems. I can still fix
> the tags on request, no need for a v3.
> 
> > No objections to the patch itself.
> 
> Thanks a lot for your effort!

Hi Steffen,

Thanks for the Tested-by tag offer -- please add:

  Tested-by: Hyunwoo Kim <imv4bel@gmail.com>

And please drop the second Reported-by: line as well.
Appreciate the careful handling.


Best regards,
Hyunwoo Kim

Re: [PATCH net v2] xfrm: esp: avoid in-place decrypt on shared skb frags

[Hex Rabbit]

Hi Hyunwoo, Steffen,

> The report and patch for this issue were already posted on
> the public netdev ML 6 days ago, i.e., the bug was already
> publicly reported:
>
>   https://lore.kernel.org/all/afLDKSvAvMwGh7Fy@v4bel/
>
> Credit for patch authorship is adequately covered by
> Signed-off-by alone. Setting aside that your work proceeded
> independently rather than as a review of my earlier
> submission, the trailer should conform to convention to
> avoid future misunderstanding.

For clarity, I also found and reported the issue independently to
security@kernel.org on the same day, with my own reproducer and
root-cause analysis. At that time I was not aware of your report or
patch; otherwise I would have referenced it earlier.

I am still not fully familiar with the exact kernel trailer convention
here, so I added both Reported-by tags because the reports were
independent.

Steffen, either trailer form is fine with me. If you decide to drop my
Reported-by because the patch already has my Signed-off-by, I have no
objection.

Thanks,
Kuan-Ting