From: Konstantin Ananyev <konstantin.ananyev@huawei.com>
To: "Morten Brørup" <mb@smartsharesystems.com>,
"dev@dpdk.org" <dev@dpdk.org>,
"Marat Khalili" <marat.khalili@huawei.com>
Cc: Stephen Hemminger <stephen@networkplumber.org>,
Christophe Fontaine <cfontain@redhat.com>,
Konstantin Ananyev <konstantin.v.ananyev@yandex.ru>,
Wathsala Vithanage <wathsala.vithanage@arm.com>,
"stable@dpdk.org" <stable@dpdk.org>
Subject: RE: [PATCH v2] mbuf: fix read packet data
Date: Thu, 19 Mar 2026 13:15:40 +0000 [thread overview]
Message-ID: <71be4d5ba80847d3994012a082fc42d6@huawei.com> (raw)
In-Reply-To: <20260319124329.729034-1-mb@smartsharesystems.com>
> The sum of offset + length, both 32 bit unsigned integers, could wrap
> around, causing comparisons to give the wrong result.
> This was fixed by using 64 bit instead of 32 bit for calculating the sum.
>
> Note:
> When the branch is not taken for the initial "if ((uint64_t)off + len >
> rte_pktmbuf_pkt_len(m))" comparison, the sum is known to not exceed the
> maximum possible value of rte_pktmbuf_pkt_len(m), UINT32_MAX, and
> following sum calculations can proceed using 32 bit.
>
> Also, fixed a related bug in an mbuf test case:
> It expected reading a length of UINT_MAX from a non-zero offset to not
> fail.
> And due to the offset+length wraparound bug, the read operation did not
> fail.
> This test case was updated to expect the read operation to fail.
>
> Fixes: b84110e7baa2 ("mbuf: add function to read packet data")
> Fixes: 7b295dceea07 ("test/mbuf: add unit test cases")
> Cc: stable@dpdk.org
>
> Signed-off-by: Morten Brørup <mb@smartsharesystems.com>
> ---
> v2:
> * Fixed mbuf test case. (Marat Khalili)
> ---
> app/test/test_mbuf.c | 18 +++++++++---------
> lib/mbuf/rte_mbuf.c | 2 +-
> lib/mbuf/rte_mbuf.h | 2 +-
> 3 files changed, 11 insertions(+), 11 deletions(-)
>
> diff --git a/app/test/test_mbuf.c b/app/test/test_mbuf.c
> index a41d2d0f97..db23259745 100644
> --- a/app/test/test_mbuf.c
> +++ b/app/test/test_mbuf.c
> @@ -2037,13 +2037,13 @@ test_pktmbuf_read_from_offset(struct
> rte_mempool *pktmbuf_pool)
> /* read length greater than mbuf data_len */
> if (rte_pktmbuf_read(m, hdr_len, rte_pktmbuf_data_len(m) + 1,
> NULL) != NULL)
> - GOTO_FAIL("%s: Requested len is larger than mbuf data len!\n",
> + GOTO_FAIL("%s: Requested offset + len is larger than mbuf data
> len!\n",
> __func__);
>
> /* read length greater than mbuf pkt_len */
> if (rte_pktmbuf_read(m, hdr_len, rte_pktmbuf_pkt_len(m) + 1,
> NULL) != NULL)
> - GOTO_FAIL("%s: Requested len is larger than mbuf pkt len!\n",
> + GOTO_FAIL("%s: Requested offset + len is larger than mbuf pkt
> len!\n",
> __func__);
>
> /* read data of zero len from valid offset */
> @@ -2065,21 +2065,21 @@ test_pktmbuf_read_from_offset(struct
> rte_mempool *pktmbuf_pool)
>
> /* read data of max length from valid offset */
> data_copy = rte_pktmbuf_read(m, hdr_len, UINT_MAX, NULL);
> - if (data_copy == NULL)
> - GOTO_FAIL("%s: Error in reading packet data!\n", __func__);
> - /* check if the received address is the beginning of data segment */
> - if (data_copy != data)
> - GOTO_FAIL("%s: Corrupted data address!\n", __func__);
> + if (data_copy != NULL)
> + GOTO_FAIL("%s: Requested offset + max len is larger than mbuf
> pkt len!\n",
> + __func__);
>
> /* try to read from mbuf with max size offset */
> data_copy = rte_pktmbuf_read(m, UINT_MAX, 0, NULL);
> if (data_copy != NULL)
> - GOTO_FAIL("%s: Error in reading packet data!\n", __func__);
> + GOTO_FAIL("%s: Requested max offset is larger than mbuf pkt
> len!\n",
> + __func__);
>
> /* try to read from mbuf with max size offset and len */
> data_copy = rte_pktmbuf_read(m, UINT_MAX, UINT_MAX, NULL);
> if (data_copy != NULL)
> - GOTO_FAIL("%s: Error in reading packet data!\n", __func__);
> + GOTO_FAIL("%s: Requested max offset + max len is larger than
> mbuf pkt len!\n",
> + __func__);
>
> rte_pktmbuf_dump(stdout, m, rte_pktmbuf_pkt_len(m));
>
> diff --git a/lib/mbuf/rte_mbuf.c b/lib/mbuf/rte_mbuf.c
> index a5d16e4c97..c2476e7704 100644
> --- a/lib/mbuf/rte_mbuf.c
> +++ b/lib/mbuf/rte_mbuf.c
> @@ -795,7 +795,7 @@ const void *__rte_pktmbuf_read(const struct rte_mbuf
> *m, uint32_t off,
> const struct rte_mbuf *seg = m;
> uint32_t buf_off = 0, copy_len;
>
> - if (off + len > rte_pktmbuf_pkt_len(m))
> + if ((uint64_t)off + len > rte_pktmbuf_pkt_len(m))
Thanks Morten for addressing it.
Just as a nit, for 32-bit systems, wouldn't it be more friendly:
If (off + len < off || off + len > rte_pktmbuf_pkt_len(m))
?
Same for comparison in rte_pktmbuf_read().
With or without suggested change:
Acked-by: Konstantin Ananyev <konstantin.ananyev@huawei.com>
> return NULL;
>
> while (off >= rte_pktmbuf_data_len(seg)) {
> diff --git a/lib/mbuf/rte_mbuf.h b/lib/mbuf/rte_mbuf.h
> index 592af2388c..d6602f74bc 100644
> --- a/lib/mbuf/rte_mbuf.h
> +++ b/lib/mbuf/rte_mbuf.h
> @@ -1843,7 +1843,7 @@ const void *__rte_pktmbuf_read(const struct rte_mbuf
> *m, uint32_t off,
> static inline const void *rte_pktmbuf_read(const struct rte_mbuf *m,
> uint32_t off, uint32_t len, void *buf)
> {
> - if (likely(off + len <= rte_pktmbuf_data_len(m)))
> + if (likely((uint64_t)off + len <= rte_pktmbuf_data_len(m)))
> return rte_pktmbuf_mtod_offset(m, char *, off);
> else
> return __rte_pktmbuf_read(m, off, len, buf);
> --
> 2.43.0
next prev parent reply other threads:[~2026-03-19 13:15 UTC|newest]
Thread overview: 8+ messages / expand[flat|nested] mbox.gz Atom feed top
2026-03-19 8:40 [PATCH] mbuf: fix read packet data Morten Brørup
2026-03-19 9:36 ` Marat Khalili
2026-03-19 12:43 ` [PATCH v2] " Morten Brørup
2026-03-19 13:15 ` Konstantin Ananyev [this message]
2026-03-19 13:55 ` Morten Brørup
2026-03-19 14:21 ` Marat Khalili
2026-03-19 14:31 ` Morten Brørup
2026-03-25 22:33 ` Thomas Monjalon
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=71be4d5ba80847d3994012a082fc42d6@huawei.com \
--to=konstantin.ananyev@huawei.com \
--cc=cfontain@redhat.com \
--cc=dev@dpdk.org \
--cc=konstantin.v.ananyev@yandex.ru \
--cc=marat.khalili@huawei.com \
--cc=mb@smartsharesystems.com \
--cc=stable@dpdk.org \
--cc=stephen@networkplumber.org \
--cc=wathsala.vithanage@arm.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox