From: "Renyong Wan" <wanry@yunsilicon.com>
To: "David Marchand" <david.marchand@redhat.com>, <dev@dpdk.org>
Cc: <thomas@monjalon.net>, "Na Na" <nana@yunsilicon.com>,
"Rong Qian" <qianr@yunsilicon.com>,
"Xiaoxiong Zhang" <zhangxx@yunsilicon.com>,
"Dongwei Xu" <xudw@yunsilicon.com>
Subject: Re: [PATCH] net/xsc: fix use after free in some RXQ cleanup
Date: Tue, 9 Sep 2025 16:08:14 +0800 [thread overview]
Message-ID: <f50263c8-895e-4878-962e-f888f599bd90@yunsilicon.com> (raw)
In-Reply-To: <20250909070427.2711048-1-david.marchand@redhat.com>
Thanks David for catching this issue.
We'll address it in the 25.11 release.
On 2025/9/9 15:04, David Marchand wrote:
> Debian 12 gcc complains about a use after free in this cleanup section.
>
> [7/11] Compiling C object drivers/libtmp_rte_net_xsc.a.p/net_xsc_xsc_rx.c.o
> In function 'xsc_rss_qp_create',
> inlined from 'xsc_rxq_rss_obj_new' at ../drivers/net/xsc/xsc_rx.c:565:8:
> ../drivers/net/xsc/xsc_rx.c:501:9: warning: pointer 'req' may be used after
> 'free' [-Wuse-after-free]
> 501 | free(req);
> | ^~~~~~~~~
> ../drivers/net/xsc/xsc_rx.c:501:9: note: call to 'free' here
>
> Indeed, req may be free'd twice, as an error in the cleanup loop may
> jump back to the set_qp_fail label.
>
> Instead, skip the erroneous rxq and don't touch errno since all the code
> jumping to set_qp_fail already sets it.
>
> Fixes: 3991c890fb4c ("net/xsc: optimize RSS queue creation")
>
> Signed-off-by: David Marchand <david.marchand@redhat.com>
> ---
> drivers/net/xsc/xsc_rx.c | 6 ++----
> 1 file changed, 2 insertions(+), 4 deletions(-)
>
> diff --git a/drivers/net/xsc/xsc_rx.c b/drivers/net/xsc/xsc_rx.c
> index 5f8003a1f6..5ff3f818c2 100644
> --- a/drivers/net/xsc/xsc_rx.c
> +++ b/drivers/net/xsc/xsc_rx.c
> @@ -502,10 +502,8 @@ xsc_rss_qp_create(struct xsc_ethdev_priv *priv, int port_id)
> for (i = 0; i < set_last_no; i++) {
> xsc_unset_qp_info(xdev, rqn_base + i);
> rxq_data = xsc_rxq_get(priv, i);
> - if (rxq_data == NULL) {
> - rte_errno = EINVAL;
> - goto set_qp_fail;
> - }
> + if (rxq_data == NULL)
> + continue;
> rte_memzone_free(rxq_data->rq_pas);
> rxq_data->rq_pas = NULL;
> }
--
Best regards,
Renyong Wan
next prev parent reply other threads:[~2025-09-09 8:08 UTC|newest]
Thread overview: 5+ messages / expand[flat|nested] mbox.gz Atom feed top
2025-09-09 7:04 David Marchand
2025-09-09 8:08 ` Renyong Wan [this message]
2025-09-09 9:22 ` Thomas Monjalon
2025-09-09 13:04 ` Renyong Wan
2025-09-09 14:45 ` David Marchand
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=f50263c8-895e-4878-962e-f888f599bd90@yunsilicon.com \
--to=wanry@yunsilicon.com \
--cc=david.marchand@redhat.com \
--cc=dev@dpdk.org \
--cc=nana@yunsilicon.com \
--cc=qianr@yunsilicon.com \
--cc=thomas@monjalon.net \
--cc=xudw@yunsilicon.com \
--cc=zhangxx@yunsilicon.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;
as well as URLs for NNTP newsgroup(s).