Re: [PATCH net v2] xsk: reject undersized tx_metadata at runtime validation

On 07/24, Alexander Lobakin wrote:

From: Stanislav Fomichev <redacted>
Date: Thu, 23 Jul 2026 10:15:15 -0700

quoted

On 07/23, Alexander Lobakin wrote:

quoted

From: Stanislav Fomichev <redacted>
Date: Wed, 22 Jul 2026 10:55:17 -0700

quoted

On 07/22, Cen Zhang (Microsoft) wrote:

quoted

There is no metadata length check before accessing struct xsk_tx_metadata
fields in both xdp_umem_reg() (registration) and
xsk_buff_valid_tx_metadata() (runtime).  Thus, an invalid small
tx_metadata_len (e.g. 8 bytes) passes registration and later causes an
out-of-bounds read when xsk_skb_metadata() accesses fields beyond the
declared metadata region.

[...]

quoted

@@ -773,7 +773,7 @@ struct xdp_desc_ctx xp_raw_get_ctx(const struct xsk_buff_pool *pool, u64 addr)
 	addr = __xp_raw_get_addr(pool, addr);
 
 	ret.dma = __xp_raw_get_dma(pool, addr);
-	ret.meta = __xsk_buff_get_metadata(pool, __xp_raw_get_data(pool, addr));
+	xsk_buff_read_metadata(pool, addr, &ret.meta);

That effectively disables some micro-optimizations that were the reason
I created this helper -- __xp_raw_get_data() is inline, while
xsk_buff_read_metadata() uses the external version, probably some other
stuff (note the underscored __xsk_buff_get_metadata()) =\ This function
expanded to have no external calls on clang -O2, with this patch I'm not
sure it would.

Any better ideas on how to solve it? I can convert all non libeth stuff
to these new xsk_buff_read_metadata helpers and keep xp_raw_get_ctx
to you? Will that work?

Sure, I can take a look.

Do you want me to pick this patch as a base or you'd like to send a new rev?

I have an idea on how to keep the existing pointer (instead of introducing a
copy). Let me play with it today, will send something hopefully Monday..