[bpf] xsk: Free skb when TX metadata options are invalid

Commit Message

Felix Maurer Nov. 14, 2024, 11:30 a.m. UTC

When a new skb is allocated for transmitting an xsk descriptor, i.e., for
every non-multibuf descriptor or the first frag of a multibuf descriptor,
but the descriptor is later found to have invalid options set for the TX
metadata, the new skb is never freed. This can leak skbs until the send
buffer is full which makes sending more packets impossible.

Fix this by freeing the skb in the error path if we are currently dealing
with the first frag, i.e., an skb allocated in this iteration of
xsk_build_skb.

Fixes: 48eb03dd2630 ("xsk: Add TX timestamp and TX checksum offload support")
Reported-by: Michal Schmidt <[e-mail]>
Signed-off-by: Felix Maurer <[e-mail]>
---
 net/xdp/xsk.c | 11 ++++++-----
 1 file changed, 6 insertions(+), 5 deletions(-)

Comments

Toke Høiland-Jørgensen Nov. 14, 2024, 1:13 p.m. UTC | #1

Felix Maurer <[e-mail]> writes:

> When a new skb is allocated for transmitting an xsk descriptor, i.e., for
> every non-multibuf descriptor or the first frag of a multibuf descriptor,
> but the descriptor is later found to have invalid options set for the TX
> metadata, the new skb is never freed. This can leak skbs until the send
> buffer is full which makes sending more packets impossible.
>
> Fix this by freeing the skb in the error path if we are currently dealing
> with the first frag, i.e., an skb allocated in this iteration of
> xsk_build_skb.
>
> Fixes: 48eb03dd2630 ("xsk: Add TX timestamp and TX checksum offload support")
> Reported-by: Michal Schmidt <[e-mail]>
> Signed-off-by: Felix Maurer <[e-mail]>

Reviewed-by: Toke Høiland-Jørgensen <[e-mail]>

Stanislav Fomichev Nov. 14, 2024, 6:39 p.m. UTC | #2

On 11/14, Felix Maurer wrote:
> When a new skb is allocated for transmitting an xsk descriptor, i.e., for
> every non-multibuf descriptor or the first frag of a multibuf descriptor,
> but the descriptor is later found to have invalid options set for the TX
> metadata, the new skb is never freed. This can leak skbs until the send
> buffer is full which makes sending more packets impossible.
> 
> Fix this by freeing the skb in the error path if we are currently dealing
> with the first frag, i.e., an skb allocated in this iteration of
> xsk_build_skb.
> 
> Fixes: 48eb03dd2630 ("xsk: Add TX timestamp and TX checksum offload support")
> Reported-by: Michal Schmidt <[e-mail]>
> Signed-off-by: Felix Maurer <[e-mail]>
> ---

Acked-by: Stanislav Fomichev <[e-mail]>

Reminds me of the following:
https://lore.kernel.org/netdev/[e-mail]/#t

Maybe I need to try to cleanup this path. Too many corner cases so it's
impossible to get right :-(

Martin KaFai Lau Nov. 14, 2024, 9:43 p.m. UTC | #3

On 11/14/24 3:30 AM, Felix Maurer wrote:
> When a new skb is allocated for transmitting an xsk descriptor, i.e., for
> every non-multibuf descriptor or the first frag of a multibuf descriptor,
> but the descriptor is later found to have invalid options set for the TX
> metadata, the new skb is never freed. This can leak skbs until the send
> buffer is full which makes sending more packets impossible.
> 
> Fix this by freeing the skb in the error path if we are currently dealing
> with the first frag, i.e., an skb allocated in this iteration of
> xsk_build_skb.

Acked-by: Martin KaFai Lau <[e-mail]>

Jakub, can you help to take it directly to the net tree? Thanks!

Jakub Kicinski Nov. 15, 2024, 2:22 a.m. UTC | #4

On Thu, 14 Nov 2024 13:43:22 -0800 Martin KaFai Lau wrote:
> Jakub, can you help to take it directly to the net tree? Thanks!

Ack, will do!

[e-mail] Nov. 15, 2024, 11:10 p.m. UTC | #5

Hello:

This patch was applied to netdev/net.git (main)
by Jakub Kicinski <[e-mail]>:

On Thu, 14 Nov 2024 12:30:05 +0100 you wrote:
> When a new skb is allocated for transmitting an xsk descriptor, i.e., for
> every non-multibuf descriptor or the first frag of a multibuf descriptor,
> but the descriptor is later found to have invalid options set for the TX
> metadata, the new skb is never freed. This can leak skbs until the send
> buffer is full which makes sending more packets impossible.
> 
> Fix this by freeing the skb in the error path if we are currently dealing
> with the first frag, i.e., an skb allocated in this iteration of
> xsk_build_skb.
> 
> [...]

Here is the summary with links:
  - [bpf] xsk: Free skb when TX metadata options are invalid
    https://git.kernel.org/netdev/net/c/0c0d0f42ffa6

You are awesome, thank you!

Patch

diff --git a/net/xdp/xsk.c b/net/xdp/xsk.c
index 1140b2a120ca..b57d5d2904eb 100644
--- a/net/xdp/xsk.c
+++ b/net/xdp/xsk.c
@@ -675,6 +675,8 @@  static struct sk_buff *xsk_build_skb(struct xdp_sock *xs,
 		len = desc->len;
 
 		if (!skb) {
+			first_frag = true;
+
 			hr = max(NET_SKB_PAD, L1_CACHE_ALIGN(dev->needed_headroom));
 			tr = dev->needed_tailroom;
 			skb = sock_alloc_send_skb(&xs->sk, hr + len + tr, 1, &err);
@@ -685,12 +687,8 @@  static struct sk_buff *xsk_build_skb(struct xdp_sock *xs,
 			skb_put(skb, len);
 
 			err = skb_store_bits(skb, 0, buffer, len);
-			if (unlikely(err)) {
-				kfree_skb(skb);
+			if (unlikely(err))
 				goto free_err;
-			}
-
-			first_frag = true;
 		} else {
 			int nr_frags = skb_shinfo(skb)->nr_frags;
 			struct page *page;
@@ -758,6 +756,9 @@  static struct sk_buff *xsk_build_skb(struct xdp_sock *xs,
 	return skb;
 
 free_err:
+	if (first_frag && skb)
+		kfree_skb(skb);
+
 	if (err == -EOVERFLOW) {
 		/* Drop the packet */
 		xsk_set_destructor_arg(xs->skb);