Netdev Archive on lore.kernel.org
help / color / mirror / Atom feed
From: Eric Dumazet <eric.dumazet@gmail.com>
To: Vasily Averin <vvs@virtuozzo.com>,
Eric Dumazet <eric.dumazet@gmail.com>,
Christoph Paasch <christoph.paasch@gmail.com>,
"David S. Miller" <davem@davemloft.net>
Cc: Hideaki YOSHIFUJI <yoshfuji@linux-ipv6.org>,
David Ahern <dsahern@kernel.org>,
Jakub Kicinski <kuba@kernel.org>, netdev <netdev@vger.kernel.org>,
linux-kernel@vger.kernel.org, kernel@openvz.org,
Alexey Kuznetsov <kuznet@ms2.inr.ac.ru>,
Julian Wiedmann <jwi@linux.ibm.com>
Subject: Re: [PATCH net-next v4] skb_expand_head() adjust skb->truesize incorrectly
Date: Wed, 1 Sep 2021 21:32:49 -0700 [thread overview]
Message-ID: <ef7ccff8-700b-79c2-9a82-199b9ed3d95b@gmail.com> (raw)
In-Reply-To: <b7c2cb05-7307-f04e-530e-89fc466aa83f@virtuozzo.com>
On 9/1/21 8:59 PM, Vasily Averin wrote:
> On 9/1/21 10:17 PM, Eric Dumazet wrote:
>>
>>
>> On 9/1/21 1:11 AM, Vasily Averin wrote:
>>> Christoph Paasch reports [1] about incorrect skb->truesize
>>> after skb_expand_head() call in ip6_xmit.
>>> This may happen because of two reasons:
>>> - skb_set_owner_w() for newly cloned skb is called too early,
>>> before pskb_expand_head() where truesize is adjusted for (!skb-sk) case.
>>> - pskb_expand_head() does not adjust truesize in (skb->sk) case.
>>> In this case sk->sk_wmem_alloc should be adjusted too.
>>>
>>> [1] https://lkml.org/lkml/2021/8/20/1082
>>>
>>> Fixes: f1260ff15a71 ("skbuff: introduce skb_expand_head()")
>>> Reported-by: Christoph Paasch <christoph.paasch@gmail.com>
>>> Signed-off-by: Vasily Averin <vvs@virtuozzo.com>
>>> ---
>>> v4: decided to use is_skb_wmem() after pskb_expand_head() call
>>> fixed 'return (EXPRESSION);' in os_skb_wmem according to Eric Dumazet
>>> v3: removed __pskb_expand_head(),
>>> added is_skb_wmem() helper for skb with wmem-compatible destructors
>>> there are 2 ways to use it:
>>> - before pskb_expand_head(), to create skb clones
>>> - after successfull pskb_expand_head() to change owner on extended skb.
>>> v2: based on patch version from Eric Dumazet,
>>> added __pskb_expand_head() function, which can be forced
>>> to adjust skb->truesize and sk->sk_wmem_alloc.
>>> ---
>>> include/net/sock.h | 1 +
>>> net/core/skbuff.c | 35 ++++++++++++++++++++++++++---------
>>> net/core/sock.c | 8 ++++++++
>>> 3 files changed, 35 insertions(+), 9 deletions(-)
>>>
>>> diff --git a/include/net/sock.h b/include/net/sock.h
>>> index 95b2577..173d58c 100644
>>> --- a/include/net/sock.h
>>> +++ b/include/net/sock.h
>>> @@ -1695,6 +1695,7 @@ struct sk_buff *sock_wmalloc(struct sock *sk, unsigned long size, int force,
>>> gfp_t priority);
>>> void __sock_wfree(struct sk_buff *skb);
>>> void sock_wfree(struct sk_buff *skb);
>>> +bool is_skb_wmem(const struct sk_buff *skb);
>>> struct sk_buff *sock_omalloc(struct sock *sk, unsigned long size,
>>> gfp_t priority);
>>> void skb_orphan_partial(struct sk_buff *skb);
>>> diff --git a/net/core/skbuff.c b/net/core/skbuff.c
>>> index f931176..09991cb 100644
>>> --- a/net/core/skbuff.c
>>> +++ b/net/core/skbuff.c
>>> @@ -1804,28 +1804,45 @@ struct sk_buff *skb_realloc_headroom(struct sk_buff *skb, unsigned int headroom)
>>> struct sk_buff *skb_expand_head(struct sk_buff *skb, unsigned int headroom)
>>> {
>>> int delta = headroom - skb_headroom(skb);
>>> + int osize = skb_end_offset(skb);
>>> + struct sk_buff *oskb = NULL;
>>> + struct sock *sk = skb->sk;
>>>
>>> if (WARN_ONCE(delta <= 0,
>>> "%s is expecting an increase in the headroom", __func__))
>>> return skb;
>>>
>>> - /* pskb_expand_head() might crash, if skb is shared */
>>> + delta = SKB_DATA_ALIGN(delta);
>>> + /* pskb_expand_head() might crash, if skb is shared.
>>> + * Also we should clone skb if its destructor does
>>> + * not adjust skb->truesize and sk->sk_wmem_alloc
>>> + */
>>> if (skb_shared(skb)) {
>>> struct sk_buff *nskb = skb_clone(skb, GFP_ATOMIC);
>>>
>>> - if (likely(nskb)) {
>>> - if (skb->sk)
>>> - skb_set_owner_w(nskb, skb->sk);
>>> - consume_skb(skb);
>>> - } else {
>>> + if (unlikely(!nskb)) {
>>> kfree_skb(skb);
>>> + return NULL;
>>> }
>>> + oskb = skb;
>>> skb = nskb;
>>> }
>>> - if (skb &&
>>> - pskb_expand_head(skb, SKB_DATA_ALIGN(delta), 0, GFP_ATOMIC)) {
>>> + if (pskb_expand_head(skb, delta, 0, GFP_ATOMIC)) {
>>> kfree_skb(skb);
>>> - skb = NULL;
>>> + kfree_skb(oskb);
>>> + return NULL;
>>> + }
>>> + if (oskb) {
>>> + if (sk)
>>
>> if (is_skb_wmem(oskb))
>> Again, it is not valid to call skb_set_owner_w(skb, sk) on all kind of sockets.
>
> I'm disagree.
:/ :/ :/
>
> In this particular case we have new skb with skb->sk = NULL,
> In this case skb_orphan() called inside skb_set_owner_w(() will do nothing,
> we just properly set destructor to sock_wfree and adjust sk->sk_wmem_alloc,
>
We can not adjust sk_wmem_alloc if this is already 0
The only way you can guarantee this is :
to look at is_skb_wmem(oskb)
Because then you are certain that _at_ least this skb owns a reference on sk->sk_wmem_alloc
If another kind of destructor is held by oskb, then you can not assume this.
Otherwise we need a new refcount_add_if_not_zero() function, and make skb_set_owner_w()
more expensive for a very corner case.
> It is 100% equivalent of code used with skb_realloc_headroom(),
> and there was no claims on this.
> Cristoph's reproducer do not use shared skb and to not check this path,
> so it cannot be the reason of troubles in his experiments.
>
> Old destructor (sock_edemux?) can be calleda bit later, for old skb, inside consume_skb().
> It can decrement last refcount and can trigger sk_free(). However in this case
> adjusted sk_wmem_alloc did not allow to free sk.
>
> So I'm sure it is safe.
It is not safe.
>
>>> + skb_set_owner_w(skb, sk);
>>> + consume_skb(oskb);
>>> + } else if (sk) {
>>
>> && (skb->destructor != sock_edemux)
>> (Because in this case , pskb_expand_head() already adjusted skb->truesize)
>
> Agree, thank you, my fault, I've missed it.
> I think it was the reason of the troubles in last Cristoph's experiment.
>
>>> + delta = osize - skb_end_offset(skb);
>>
>>> + if (!is_skb_wmem(skb))
>>> + skb_set_owner_w(skb, sk);
>>
>> This is dangerous, even if a socket is there, its sk->sk_wmem_alloc could be zero.
>> We can not add skb->truesize to a refcount_t that already reached 0 (sk_free())
>>
>> If is_skb_wmem() is false, you probably should do nothing, and leave
>> current destructor as it is.
>
> I;m still not sure and think it is tricky too.
> I've found few destructors called sock_wfree inside, they require sk_wmem_alloc adjustement.
> sctp_wfree, unix_destruct_scm and tpacket_destruct_skb
>
> In the same time another ones do not use sk_wmem_alloc and I do not know how to detect proper ones.
> Potentially there are some 3rd party protocols out-of-tree, and I cannot list all of them here.
I think you missed netem case, in particular
skb_orphan_partial() which I already pointed out.
You can setup a stack of virtual devices (tunnels),
with a qdisc on them, before ip6_xmit() is finally called...
Socket might have been closed already.
To test your patch, you could force a skb_orphan_partial() at the beginning
of skb_expand_head() (extending code coverage)
>
> However I think I can use the same trick as one described above:
> I can increase sk_wmem_alloc before skb_orphan(), so sk_free() called by old destuctor
> cannot call __sk_free() and release sk.
You can not change sk_wmem_alloc if this is already 0.
refcount_add() will trigger a warning (panic under KASAN)
>
> I hope this should work,
> otherwise we'll need to clone skb for !is_skb_wmem(skb) before pskb_expand_head() call.
>
> Thank you,
> Vasily Averin
>
next prev parent reply other threads:[~2021-09-02 4:40 UTC|newest]
Thread overview: 106+ messages / expand[flat|nested] mbox.gz Atom feed top
[not found] <cover.1625665132.git.vvs@virtuozzo.com>
2021-07-07 14:04 ` [PATCH IPV6 1/1] ipv6: allocate enough headroom in ip6_finish_output2() Vasily Averin
2021-07-07 14:45 ` David Ahern
2021-07-07 16:42 ` Jakub Kicinski
2021-07-07 17:41 ` Eric Dumazet
2021-07-07 17:53 ` Vasily Averin
2021-07-07 18:30 ` Jakub Kicinski
2021-07-07 18:50 ` Eric Dumazet
2021-07-09 9:04 ` [PATCH IPV6 v2 0/4] " Vasily Averin
2021-07-12 6:44 ` [PATCH IPV6 v3 0/1] " Vasily Averin
[not found] ` <cover.1626069562.git.vvs@virtuozzo.com>
2021-07-12 6:45 ` [PATCH IPV6 v3 1/1] " Vasily Averin
2021-07-12 18:30 ` patchwork-bot+netdevbpf
2021-07-13 7:46 ` Vasily Averin
2021-07-13 12:01 ` [PATCH NET v4 0/1] " Vasily Averin
[not found] ` <cover.1626177047.git.vvs@virtuozzo.com>
2021-07-13 12:01 ` [PATCH NET v4 1/1] " Vasily Averin
2021-07-18 10:44 ` Vasily Averin
2021-07-18 15:22 ` David Ahern
2021-07-18 17:04 ` David Miller
2021-07-19 7:55 ` [PATCH NET] ipv6: ip6_finish_output2: set sk into newly allocated nskb Vasily Averin
2021-07-20 10:10 ` patchwork-bot+netdevbpf
2021-07-13 12:31 ` [PATCH IPV6 v3 1/1] ipv6: allocate enough headroom in ip6_finish_output2() Vasily Averin
2021-07-12 13:26 ` [PATCH NET 0/7] skbuff: introduce pskb_realloc_headroom() Vasily Averin
[not found] ` <cover.1626093470.git.vvs@virtuozzo.com>
2021-07-12 13:26 ` [PATCH NET 1/7] " Vasily Averin
2021-07-12 17:53 ` Jakub Kicinski
2021-07-12 18:45 ` Vasily Averin
2021-07-13 20:57 ` [PATCH NET v2 0/7] skbuff: introduce skb_expand_head() Vasily Averin
2021-08-02 8:52 ` [PATCH NET v3 " Vasily Averin
[not found] ` <cover.1627891754.git.vvs@virtuozzo.com>
2021-08-02 8:52 ` [PATCH NET v3 1/7] " Vasily Averin
2021-08-02 8:52 ` [PATCH NET v3 2/7] ipv6: use skb_expand_head in ip6_finish_output2 Vasily Averin
2021-08-02 8:52 ` [PATCH NET v3 3/7] ipv6: use skb_expand_head in ip6_xmit Vasily Averin
2021-08-02 8:52 ` [PATCH NET v3 4/7] ipv4: use skb_expand_head in ip_finish_output2 Vasily Averin
2021-08-02 8:52 ` [PATCH NET v3 5/7] vrf: use skb_expand_head in vrf_finish_output Vasily Averin
2021-08-05 11:55 ` Julian Wiedmann
2021-08-05 12:55 ` Vasily Averin
2021-08-06 7:49 ` [PATCH NET v4 0/7] skbuff: introduce skb_expand_head() Vasily Averin
2021-08-06 10:14 ` David Miller
2021-08-06 12:53 ` [PATCH NET] vrf: fix null pointer dereference in vrf_finish_output() Vasily Averin
2021-08-06 22:42 ` Jakub Kicinski
2021-08-07 6:41 ` Vasily Averin
[not found] ` <cover.1628235065.git.vvs@virtuozzo.com>
2021-08-06 7:49 ` [PATCH NET v4 1/7] skbuff: introduce skb_expand_head() Vasily Averin
2021-08-06 7:50 ` [PATCH NET v4 2/7] ipv6: use skb_expand_head in ip6_finish_output2 Vasily Averin
2021-08-06 7:50 ` [PATCH NET v4 3/7] ipv6: use skb_expand_head in ip6_xmit Vasily Averin
[not found] ` <CALMXkpaay1y=0tkbnskr4gf-HTMjJJsVryh4Prnej_ws-hJvBg@mail.gmail.com>
2021-08-20 22:44 ` Christoph Paasch
2021-08-21 6:21 ` Vasily Averin
2021-08-22 17:04 ` Christoph Paasch
2021-08-22 17:13 ` Christoph Paasch
2021-08-23 5:44 ` Vasily Averin
2021-08-23 5:59 ` Vasily Averin
2021-08-23 7:56 ` [PATCH NET-NEXT] ipv6: skb_expand_head() adjust skb->truesize incorrectly Vasily Averin
2021-08-23 17:25 ` Christoph Paasch
2021-08-23 21:45 ` Eric Dumazet
2021-08-23 21:51 ` Eric Dumazet
2021-08-23 22:23 ` Eric Dumazet
2021-08-24 8:50 ` Vasily Averin
2021-08-24 17:21 ` Vasily Averin
2021-08-25 17:49 ` Christoph Paasch
2021-08-29 12:59 ` [PATCH v2] " Vasily Averin
2021-08-30 5:52 ` [PATCH net-next " Vasily Averin
2021-08-30 16:01 ` [PATCH " Eric Dumazet
2021-08-30 18:09 ` Vasily Averin
2021-08-30 18:37 ` Vasily Averin
2021-08-30 19:58 ` Eric Dumazet
2021-08-31 14:34 ` [PATCH net-next v3 RFC] " Vasily Averin
2021-08-31 19:38 ` Eric Dumazet
2021-09-01 6:20 ` Vasily Averin
2021-09-01 8:11 ` [PATCH net-next v4] " Vasily Averin
2021-09-01 16:58 ` Christoph Paasch
2021-09-01 19:17 ` Eric Dumazet
2021-09-02 3:59 ` Vasily Averin
2021-09-02 4:32 ` Eric Dumazet [this message]
2021-09-02 4:48 ` Eric Dumazet
2021-09-02 7:13 ` Vasily Averin
2021-09-02 7:33 ` Vasily Averin
2021-09-02 8:31 ` Vasily Averin
2021-09-02 11:12 ` [PATCH net-next v5] " Vasily Averin
2021-09-02 15:53 ` Christoph Paasch
2021-09-02 16:32 ` Vasily Averin
2021-09-06 18:01 ` [PATCH net v6] " Vasily Averin
2021-09-06 18:03 ` Vasily Averin
2021-08-27 15:23 ` [PATCH NET-NEXT] ipv6: " Vasily Averin
2021-08-27 16:47 ` Eric Dumazet
2021-08-28 8:01 ` Vasily Averin
2021-08-06 7:50 ` [PATCH NET v4 4/7] ipv4: use skb_expand_head in ip_finish_output2 Vasily Averin
2021-08-06 7:50 ` [PATCH NET v4 5/7] vrf: use skb_expand_head in vrf_finish_output Vasily Averin
2021-08-06 7:50 ` [PATCH NET v4 6/7] ax25: use skb_expand_head Vasily Averin
2021-08-06 7:50 ` [PATCH NET v4 7/7] bpf: use skb_expand_head in bpf_out_neigh_v4/6 Vasily Averin
2021-08-02 8:52 ` [PATCH NET v3 6/7] ax25: use skb_expand_head Vasily Averin
2021-08-02 8:52 ` [PATCH NET v3 7/7] bpf: use skb_expand_head in bpf_out_neigh_v4/6 Vasily Averin
[not found] ` <cover.1626206993.git.vvs@virtuozzo.com>
2021-07-13 20:57 ` [PATCH NET v2 1/7] skbuff: introduce skb_expand_head() Vasily Averin
2021-07-13 20:58 ` [PATCH NET v2 2/7] ipv6: use skb_expand_head in ip6_finish_output2 Vasily Averin
2021-07-13 20:58 ` [PATCH NET v2 3/7] ipv6: use skb_expand_head in ip6_xmit Vasily Averin
2021-07-13 20:58 ` [PATCH NET v2 4/7] ipv4: use skb_expand_head in ip_finish_output2 Vasily Averin
2021-07-13 20:58 ` [PATCH NET v2 5/7] vrf: use skb_expand_head in vrf_finish_output Vasily Averin
2021-07-13 20:58 ` [PATCH NET v2 6/7] ax25: use skb_expand_head Vasily Averin
2021-07-13 20:58 ` [PATCH NET v2 7/7] bpf: use skb_expand_head in bpf_out_neigh_v4/6 Vasily Averin
2021-07-12 13:26 ` [PATCH NET 2/7] ipv6: use pskb_realloc_headroom in ip6_finish_output2 Vasily Averin
2021-07-12 13:26 ` [PATCH NET 3/7] ipv6: use pskb_realloc_headroom in ip6_xmit refactoring Vasily Averin
2021-07-12 13:27 ` [PATCH NET 4/7] ipv4: use pskb_realloc_headroom in ip_finish_output2 Vasily Averin
2021-07-12 13:27 ` [PATCH NET 5/7] vrf: use pskb_realloc_headroom in vrf_finish_output Vasily Averin
2021-07-12 13:27 ` [PATCH NET 6/7] ax25: use pskb_realloc_headroom Vasily Averin
2021-07-12 13:27 ` [PATCH NET 7/7] bpf: use pskb_realloc_headroom in bpf_out_neigh_v4/6 Vasily Averin
[not found] ` <cover.1625818825.git.vvs@virtuozzo.com>
2021-07-09 9:04 ` [PATCH IPV6 v2 1/4] ipv6: allocate enough headroom in ip6_finish_output2() Vasily Averin
2021-07-09 17:58 ` David Miller
2021-07-10 2:53 ` Vasily Averin
2021-07-09 9:04 ` [PATCH IPV6 v2 2/4] ipv6: use new helper skb_expand_head() in ip6_xmit() Vasily Averin
2021-07-09 9:05 ` [PATCH IPV6 v2 3/4] ipv6: ip6_finish_output2 refactoring Vasily Averin
2021-07-09 9:05 ` [PATCH IPV6 v2 4/4] ipv6: ip6_xmit refactoring Vasily Averin
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=ef7ccff8-700b-79c2-9a82-199b9ed3d95b@gmail.com \
--to=eric.dumazet@gmail.com \
--cc=christoph.paasch@gmail.com \
--cc=davem@davemloft.net \
--cc=dsahern@kernel.org \
--cc=jwi@linux.ibm.com \
--cc=kernel@openvz.org \
--cc=kuba@kernel.org \
--cc=kuznet@ms2.inr.ac.ru \
--cc=linux-kernel@vger.kernel.org \
--cc=netdev@vger.kernel.org \
--cc=vvs@virtuozzo.com \
--cc=yoshfuji@linux-ipv6.org \
--subject='Re: [PATCH net-next v4] skb_expand_head() adjust skb->truesize incorrectly' \
/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
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).