[PATCH 5.10.y v2] net/sched: act_mirred: don't override retval if we already lost the skb

[PATCH 5.10.y v2] net/sched: act_mirred: don't override retval if we already lost the skb

[jianqi.ren.cn]

From: Jakub Kicinski <kuba@kernel.org>

[ Upstream commit 166c2c8a6a4dc2e4ceba9e10cfe81c3e469e3210 ]

If we're redirecting the skb, and haven't called tcf_mirred_forward(),
yet, we need to tell the core to drop the skb by setting the retcode
to SHOT. If we have called tcf_mirred_forward(), however, the skb
is out of our hands and returning SHOT will lead to UaF.

Move the retval override to the error path which actually need it.

Reviewed-by: Michal Swiatkowski <michal.swiatkowski@linux.intel.com>
Fixes: e5cf1baf92cb ("act_mirred: use TC_ACT_REINSERT when possible")
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Acked-by: Jamal Hadi Salim <jhs@mojatatu.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
[Minor conflict resolved due to code context change.]
Signed-off-by: Jianqi Ren <jianqi.ren.cn@windriver.com>
Signed-off-by: He Zhe <zhe.he@windriver.com>
---
v2: Fix the following issue
net/sched/act_mirred.c:265:6: error: variable 'is_redirect' is used
uninitialized whenever 'if' condition is true
found by the following tuxmake
(https://lore.kernel.org/stable/CA+G9fYu+FEZ-3ye30Hk2sk1+LFsw7iO5AHueUa9H1Ub=JO-k2g@mail.gmail.com/)
tuxmake --runtime podman --target-arch arm --toolchain clang-20 --kconfig allmodconfig LLVM=1 LLVM_IAS=1
The above command line couldn't pass for some other issue unrelated
to this patch, but we're sure that the above "is_redirect" issue
has been fixed.
---
 net/sched/act_mirred.c | 22 +++++++++++++---------
 1 file changed, 13 insertions(+), 9 deletions(-)

diff --git a/net/sched/act_mirred.c b/net/sched/act_mirred.c
index 91a19460cb57..07b9f8335f05 100644
--- a/net/sched/act_mirred.c
+++ b/net/sched/act_mirred.c
@@ -256,31 +256,31 @@ static int tcf_mirred_act(struct sk_buff *skb, const struct tc_action *a,
 
 	m_mac_header_xmit = READ_ONCE(m->tcfm_mac_header_xmit);
 	m_eaction = READ_ONCE(m->tcfm_eaction);
+	is_redirect = tcf_mirred_is_act_redirect(m_eaction);
 	retval = READ_ONCE(m->tcf_action);
 	dev = rcu_dereference_bh(m->tcfm_dev);
 	if (unlikely(!dev)) {
 		pr_notice_once("tc mirred: target device is gone\n");
-		goto out;
+		goto err_cant_do;
 	}
 
 	if (unlikely(!(dev->flags & IFF_UP)) || !netif_carrier_ok(dev)) {
 		net_notice_ratelimited("tc mirred to Houston: device %s is down\n",
 				       dev->name);
-		goto out;
+		goto err_cant_do;
 	}
 
 	/* we could easily avoid the clone only if called by ingress and clsact;
 	 * since we can't easily detect the clsact caller, skip clone only for
 	 * ingress - that covers the TC S/W datapath.
 	 */
-	is_redirect = tcf_mirred_is_act_redirect(m_eaction);
 	at_ingress = skb_at_tc_ingress(skb);
 	use_reinsert = at_ingress && is_redirect &&
 		       tcf_mirred_can_reinsert(retval);
 	if (!use_reinsert) {
 		skb2 = skb_clone(skb, GFP_ATOMIC);
 		if (!skb2)
-			goto out;
+			goto err_cant_do;
 	}
 
 	want_ingress = tcf_mirred_act_wants_ingress(m_eaction);
@@ -323,12 +323,16 @@ static int tcf_mirred_act(struct sk_buff *skb, const struct tc_action *a,
 	}
 
 	err = tcf_mirred_forward(want_ingress, skb2);
-	if (err) {
-out:
+	if (err)
 		tcf_action_inc_overlimit_qstats(&m->common);
-		if (tcf_mirred_is_act_redirect(m_eaction))
-			retval = TC_ACT_SHOT;
-	}
+	__this_cpu_dec(mirred_nest_level);
+
+	return retval;
+
+err_cant_do:
+	if (is_redirect)
+		retval = TC_ACT_SHOT;
+	tcf_action_inc_overlimit_qstats(&m->common);
 	__this_cpu_dec(mirred_nest_level);
 
 	return retval;
-- 
2.34.1

Re: [PATCH 5.10.y v2] net/sched: act_mirred: don't override retval if we already lost the skb

[Sasha Levin]

[ Sasha's backport helper bot ]

Hi,

✅ All tests passed successfully. No issues detected.
No action required from the submitter.

The upstream commit SHA1 provided is correct: 166c2c8a6a4dc2e4ceba9e10cfe81c3e469e3210

WARNING: Author mismatch between patch and upstream commit:
Backport author: <jianqi.ren.cn@windriver.com>
Commit author: Jakub Kicinski<kuba@kernel.org>

Status in newer kernel trees:
6.14.y | Present (exact SHA1)
6.12.y | Present (exact SHA1)
6.6.y | Present (different SHA1: 28cdbbd38a44)
6.1.y | Not found
5.15.y | Not found

Note: The patch differs from the upstream commit:
---
1:  166c2c8a6a4dc < -:  ------------- net/sched: act_mirred: don't override retval if we already lost the skb
-:  ------------- > 1:  9fc922291d44b net/sched: act_mirred: don't override retval if we already lost the skb
---

Results of testing on various branches:

| Branch                    | Patch Apply | Build Test |
|---------------------------|-------------|------------|
| stable/linux-5.10.y       |  Success    |  Success   |