[PATCH] net: sched: act_ife: initialize struct tc_ife to fix KMSAN kernel-infoleak

[PATCH] net: sched: act_ife: initialize struct tc_ife to fix KMSAN kernel-infoleak

[Ranganath V N]

Fix a KMSAN kernel-infoleak detected  by the syzbot .

[net?] KMSAN: kernel-infoleak in __skb_datagram_iter

In tcf_ife_dump(), the variable 'opt' was partially initialized using a
designatied initializer. While the padding bytes are reamined
uninitialized. nla_put() copies the entire structure into a
netlink message, these uninitialized bytes leaked to userspace.

Initialize the structure with memset before assigning its fields
to ensure all members and padding are cleared prior to beign copied.

This change silences the KMSAN report and prevents potential information
leaks from the kernel memory.

This fix has been tested and validated by syzbot. This patch closes the
bug reported at the following syzkaller link and ensures no infoleak.

Reported-by: syzbot+0c85cae3350b7d486aee@syzkaller.appspotmail.com
Closes: https://syzkaller.appspot.com/bug?extid=0c85cae3350b7d486aee
Tested-by: syzbot+0c85cae3350b7d486aee@syzkaller.appspotmail.com
Fixes: ef6980b6becb ("introduce IFE action")
Signed-off-by: Ranganath V N <vnranganath.20@gmail.com>
---
Fix a KMSAN kernel-infoleak detected  by the syzbot .

[net?] KMSAN: kernel-infoleak in __skb_datagram_iter

In tcf_ife_dump(), the variable 'opt' was partially initialized using a
designatied initializer. While the padding bytes are reamined
uninitialized. nla_put() copies the entire structure into a
netlink message, these uninitialized bytes leaked to userspace.

Initialize the structure with memset before assigning its fields
to ensure all members and padding are cleared prior to beign copied.
---
 net/sched/act_ife.c | 13 ++++++++-----
 1 file changed, 8 insertions(+), 5 deletions(-)

diff --git a/net/sched/act_ife.c b/net/sched/act_ife.c
index 107c6d83dc5c..608ef6cc2224 100644
--- a/net/sched/act_ife.c
+++ b/net/sched/act_ife.c
@@ -644,13 +644,16 @@ static int tcf_ife_dump(struct sk_buff *skb, struct tc_action *a, int bind,
 	unsigned char *b = skb_tail_pointer(skb);
 	struct tcf_ife_info *ife = to_ife(a);
 	struct tcf_ife_params *p;
-	struct tc_ife opt = {
-		.index = ife->tcf_index,
-		.refcnt = refcount_read(&ife->tcf_refcnt) - ref,
-		.bindcnt = atomic_read(&ife->tcf_bindcnt) - bind,
-	};
+	struct tc_ife opt;
 	struct tcf_t t;
 
+	memset(&opt, 0, sizeof(opt));
+	memset(&t, 0, sizeof(t));
+
+	opt.index = ife->tcf_index,
+	opt.refcnt = refcount_read(&ife->tcf_refcnt) - ref,
+	opt.bindcnt = atomic_read(&ife->tcf_bindcnt) - bind,
+
 	spin_lock_bh(&ife->tcf_lock);
 	opt.action = ife->tcf_action;
 	p = rcu_dereference_protected(ife->params,

---
base-commit: d127176862a93c4b3216bda533d2bee170af5e71
change-id: 20251031-infoleak-8a7de6afc987

Best regards,
-- 
Ranganath V N <vnranganath.20@gmail.com>

Re: [PATCH] net: sched: act_ife: initialize struct tc_ife to fix KMSAN kernel-infoleak

[Eric Dumazet]

On Fri, Oct 31, 2025 at 5:24 AM Ranganath V N <vnranganath.20@gmail.com> wrote:
>
> Fix a KMSAN kernel-infoleak detected  by the syzbot .
>
> [net?] KMSAN: kernel-infoleak in __skb_datagram_iter
>
> In tcf_ife_dump(), the variable 'opt' was partially initialized using a
> designatied initializer. While the padding bytes are reamined
> uninitialized. nla_put() copies the entire structure into a
> netlink message, these uninitialized bytes leaked to userspace.
>
> Initialize the structure with memset before assigning its fields
> to ensure all members and padding are cleared prior to beign copied.
>
> This change silences the KMSAN report and prevents potential information
> leaks from the kernel memory.
>
> This fix has been tested and validated by syzbot. This patch closes the
> bug reported at the following syzkaller link and ensures no infoleak.
>
> Reported-by: syzbot+0c85cae3350b7d486aee@syzkaller.appspotmail.com
> Closes: https://syzkaller.appspot.com/bug?extid=0c85cae3350b7d486aee
> Tested-by: syzbot+0c85cae3350b7d486aee@syzkaller.appspotmail.com
> Fixes: ef6980b6becb ("introduce IFE action")
> Signed-off-by: Ranganath V N <vnranganath.20@gmail.com>
> ---
> Fix a KMSAN kernel-infoleak detected  by the syzbot .
>
> [net?] KMSAN: kernel-infoleak in __skb_datagram_iter
>
> In tcf_ife_dump(), the variable 'opt' was partially initialized using a
> designatied initializer. While the padding bytes are reamined
> uninitialized. nla_put() copies the entire structure into a
> netlink message, these uninitialized bytes leaked to userspace.
>
> Initialize the structure with memset before assigning its fields
> to ensure all members and padding are cleared prior to beign copied.
> ---
>  net/sched/act_ife.c | 13 ++++++++-----
>  1 file changed, 8 insertions(+), 5 deletions(-)
>
> diff --git a/net/sched/act_ife.c b/net/sched/act_ife.c
> index 107c6d83dc5c..608ef6cc2224 100644
> --- a/net/sched/act_ife.c
> +++ b/net/sched/act_ife.c
> @@ -644,13 +644,16 @@ static int tcf_ife_dump(struct sk_buff *skb, struct tc_action *a, int bind,
>         unsigned char *b = skb_tail_pointer(skb);
>         struct tcf_ife_info *ife = to_ife(a);
>         struct tcf_ife_params *p;
> -       struct tc_ife opt = {
> -               .index = ife->tcf_index,
> -               .refcnt = refcount_read(&ife->tcf_refcnt) - ref,
> -               .bindcnt = atomic_read(&ife->tcf_bindcnt) - bind,
> -       };
> +       struct tc_ife opt;
>         struct tcf_t t;
>
> +       memset(&opt, 0, sizeof(opt));
> +       memset(&t, 0, sizeof(t));

Why is the second memset() needed ? Please do not add unrelated changes.

Also I would also fix tcf_connmark_dump()

> +
> +       opt.index = ife->tcf_index,
> +       opt.refcnt = refcount_read(&ife->tcf_refcnt) - ref,
> +       opt.bindcnt = atomic_read(&ife->tcf_bindcnt) - bind,
> +
>         spin_lock_bh(&ife->tcf_lock);
>         opt.action = ife->tcf_action;
>         p = rcu_dereference_protected(ife->params,
>
> ---
> base-commit: d127176862a93c4b3216bda533d2bee170af5e71
> change-id: 20251031-infoleak-8a7de6afc987
>
> Best regards,
> --
> Ranganath V N <vnranganath.20@gmail.com>
>

Re: [PATCH] net: sctp: fix KMSAN uninit-value in sctp_inq_pop

[Ranganath V N]

Hi Xin,

Thank you for the feedback and response to the patch.
I would like to know that above analysis is valid or not.
And do you want me to test this suggestion with the syzbot?

regards,
Ranganath

Re: [PATCH] net: sched: act_ife: initialize struct tc_ife to fix KMSAN kernel-infoleak

[Ranganath V N]

>> Fix a KMSAN kernel-infoleak detected  by the syzbot .
>>
>> [net?] KMSAN: kernel-infoleak in __skb_datagram_iter
>>
>> In tcf_ife_dump(), the variable 'opt' was partially initialized using a
>> designatied initializer. While the padding bytes are reamined
>> uninitialized. nla_put() copies the entire structure into a
>> netlink message, these uninitialized bytes leaked to userspace.
>>
>> Initialize the structure with memset before assigning its fields
>> to ensure all members and padding are cleared prior to beign copied.
>>
>> This change silences the KMSAN report and prevents potential information
>> leaks from the kernel memory.
>>
>> This fix has been tested and validated by syzbot. This patch closes the
>> bug reported at the following syzkaller link and ensures no infoleak.
>>
>> Reported-by: syzbot+0c85cae3350b7d486aee@syzkaller.appspotmail.com
>> Closes: https://syzkaller.appspot.com/bug?extid=0c85cae3350b7d486aee
>> Tested-by: syzbot+0c85cae3350b7d486aee@syzkaller.appspotmail.com
>> Fixes: ef6980b6becb ("introduce IFE action")
>> Signed-off-by: Ranganath V N <vnranganath.20@gmail.com>
>> ---
>> Fix a KMSAN kernel-infoleak detected  by the syzbot .
>>
>> [net?] KMSAN: kernel-infoleak in __skb_datagram_iter
>>
>> In tcf_ife_dump(), the variable 'opt' was partially initialized using a
>> designatied initializer. While the padding bytes are reamined
>> uninitialized. nla_put() copies the entire structure into a
>> netlink message, these uninitialized bytes leaked to userspace.
>>
>> Initialize the structure with memset before assigning its fields
>> to ensure all members and padding are cleared prior to beign copied.
>> ---
>>  net/sched/act_ife.c | 13 ++++++++-----
>>  1 file changed, 8 insertions(+), 5 deletions(-)
>>
>> diff --git a/net/sched/act_ife.c b/net/sched/act_ife.c
>> index 107c6d83dc5c..608ef6cc2224 100644
>> --- a/net/sched/act_ife.c
>> +++ b/net/sched/act_ife.c
>> @@ -644,13 +644,16 @@ static int tcf_ife_dump(struct sk_buff *skb, struct tc_action *a, int bind,
>>         unsigned char *b = skb_tail_pointer(skb);
>>         struct tcf_ife_info *ife = to_ife(a);
>>         struct tcf_ife_params *p;
>> -       struct tc_ife opt = {
>> -               .index = ife->tcf_index,
>> -               .refcnt = refcount_read(&ife->tcf_refcnt) - ref,
>> -               .bindcnt = atomic_read(&ife->tcf_bindcnt) - bind,
>> -       };
>> +       struct tc_ife opt;
>>         struct tcf_t t;
>>
>> +       memset(&opt, 0, sizeof(opt));
>> +       memset(&t, 0, sizeof(t));
>
>Why is the second memset() needed ? Please do not add unrelated changes.
>
>Also I would also fix tcf_connmark_dump()

Hi Eric,
Do you want me fix tcf_connmark_dump() in this patch or new?

Regards,
Ranganath

Re: [PATCH] net: sched: act_ife: initialize struct tc_ife to fix KMSAN kernel-infoleak

[Eric Dumazet]

On Fri, Oct 31, 2025 at 8:56 AM Ranganath V N <vnranganath.20@gmail.com> wrote:
>
> >> Fix a KMSAN kernel-infoleak detected  by the syzbot .
> >>
> >> [net?] KMSAN: kernel-infoleak in __skb_datagram_iter
> >>
> >> In tcf_ife_dump(), the variable 'opt' was partially initialized using a
> >> designatied initializer. While the padding bytes are reamined
> >> uninitialized. nla_put() copies the entire structure into a
> >> netlink message, these uninitialized bytes leaked to userspace.
> >>
> >> Initialize the structure with memset before assigning its fields
> >> to ensure all members and padding are cleared prior to beign copied.
> >>
> >> This change silences the KMSAN report and prevents potential information
> >> leaks from the kernel memory.
> >>
> >> This fix has been tested and validated by syzbot. This patch closes the
> >> bug reported at the following syzkaller link and ensures no infoleak.
> >>
> >> Reported-by: syzbot+0c85cae3350b7d486aee@syzkaller.appspotmail.com
> >> Closes: https://syzkaller.appspot.com/bug?extid=0c85cae3350b7d486aee
> >> Tested-by: syzbot+0c85cae3350b7d486aee@syzkaller.appspotmail.com
> >> Fixes: ef6980b6becb ("introduce IFE action")
> >> Signed-off-by: Ranganath V N <vnranganath.20@gmail.com>
> >> ---
> >> Fix a KMSAN kernel-infoleak detected  by the syzbot .
> >>
> >> [net?] KMSAN: kernel-infoleak in __skb_datagram_iter
> >>
> >> In tcf_ife_dump(), the variable 'opt' was partially initialized using a
> >> designatied initializer. While the padding bytes are reamined
> >> uninitialized. nla_put() copies the entire structure into a
> >> netlink message, these uninitialized bytes leaked to userspace.
> >>
> >> Initialize the structure with memset before assigning its fields
> >> to ensure all members and padding are cleared prior to beign copied.
> >> ---
> >>  net/sched/act_ife.c | 13 ++++++++-----
> >>  1 file changed, 8 insertions(+), 5 deletions(-)
> >>
> >> diff --git a/net/sched/act_ife.c b/net/sched/act_ife.c
> >> index 107c6d83dc5c..608ef6cc2224 100644
> >> --- a/net/sched/act_ife.c
> >> +++ b/net/sched/act_ife.c
> >> @@ -644,13 +644,16 @@ static int tcf_ife_dump(struct sk_buff *skb, struct tc_action *a, int bind,
> >>         unsigned char *b = skb_tail_pointer(skb);
> >>         struct tcf_ife_info *ife = to_ife(a);
> >>         struct tcf_ife_params *p;
> >> -       struct tc_ife opt = {
> >> -               .index = ife->tcf_index,
> >> -               .refcnt = refcount_read(&ife->tcf_refcnt) - ref,
> >> -               .bindcnt = atomic_read(&ife->tcf_bindcnt) - bind,
> >> -       };
> >> +       struct tc_ife opt;
> >>         struct tcf_t t;
> >>
> >> +       memset(&opt, 0, sizeof(opt));
> >> +       memset(&t, 0, sizeof(t));
> >
> >Why is the second memset() needed ? Please do not add unrelated changes.
> >
> >Also I would also fix tcf_connmark_dump()
>
> Hi Eric,
> Do you want me fix tcf_connmark_dump() in this patch or new?

Send a series of two patches, thank you.

[PATCH] net: sctp: fix KMSAN uninit-value in sctp_inq_pop

[Ranganath V N]

Fix an issue detected by syzbot:

KMSAN reported an uninitialized-value access in sctp_inq_pop
while parsing an SCTP chunk header received frma a locally transmitted packet.

BUG: KMSAN: uninit-value in sctp_inq_pop

skb allocated in sctp_packet_transmit() contain uninitialized bytes.
sctp transmit path writes only the necessary header and chunk data,
the receive path read from uinitialized parts of the skb, triggering KMSAN.

Fix this by explicitly zeroing the skb payload area after allocation
and reservation, ensuring all future reads from this region are fully
initialized.

Reported-by: syzbot+d101e12bccd4095460e7@syzkaller.appspotmail.com
Tested-by: syzbot+d101e12bccd4095460e7@syzkaller.appspotmail.com
Fixes: https://syzkaller.appspot.com/bug?extid=d101e12bccd4095460e7
Signed-off-by: Ranganath V N <vnranganath.20@gmail.com>
---
KMSAN reported an uninitialized-value access in sctp_inq_pop
while parsing an SCTP chunk header received frma a locally transmitted packet.
---
 net/sctp/output.c | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/net/sctp/output.c b/net/sctp/output.c
index 23e96305cad7..e76413741faf 100644
--- a/net/sctp/output.c
+++ b/net/sctp/output.c
@@ -602,6 +602,8 @@ int sctp_packet_transmit(struct sctp_packet *packet, gfp_t gfp)
 	skb_reserve(head, packet->overhead + MAX_HEADER);
 	skb_set_owner_w(head, sk);
 
+	memset(head->data, 0, skb_tailroom(head));
+
 	/* set sctp header */
 	sh = skb_push(head, sizeof(struct sctphdr));
 	skb_reset_transport_header(head);

---
base-commit: 43e9ad0c55a369ecc84a4788d06a8a6bfa634f1c
change-id: 20251023-kmsan_fix-78d527b9960b

Best regards,
-- 
Ranganath V N <vnranganath.20@gmail.com>

Re: [PATCH] net: sctp: fix KMSAN uninit-value in sctp_inq_pop

[Xin Long]

On Thu, Oct 23, 2025 at 5:52 AM Ranganath V N <vnranganath.20@gmail.com> wrote:
>
> Fix an issue detected by syzbot:
>
> KMSAN reported an uninitialized-value access in sctp_inq_pop
Hi, Ranganath,

The issue is actually caused by skb trimming via sk_filter() in sctp_rcv().
In the reproducer, skb->len becomes 1 after sk_filter(), which bypassed the
original check:

        if (skb->len < sizeof(struct sctphdr) + sizeof(struct sctp_chunkhdr) +
                       skb_transport_offset(skb))

(TBH, I didn't expect it would allow BPF to trim skb in sk_filter().)

To handle this safely, a new check should be performed after sk_filter() like:

+       if (sk_filter(sk, skb) || skb->len < sizeof(struct sctp_chunkhdr))
                goto discard_release;

Could you please proceed with this change in sctp_rcv()?

Thanks.

Re: [PATCH] net: sctp: fix KMSAN uninit-value in sctp_inq_pop

[Ranganath V N]

Hi Xin,

Thank you for the feedback and response to the patch.
I would like to know that above analysis is valid or not.
And do you want me to test this suggestion with the syzbot?

regards,
Ranganath

Re: [PATCH] net: sctp: fix KMSAN uninit-value in sctp_inq_pop

[Xin Long]

On Thu, Oct 23, 2025 at 1:38 PM Ranganath V N <vnranganath.20@gmail.com> wrote:
>
> Hi Xin,
>
> Thank you for the feedback and response to the patch.
> I would like to know that above analysis is valid or not.
> And do you want me to test this suggestion with the syzbot?
>
Yes, if it's possible.