[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <8EA496BE-669B-44C1-A3D7-AF7BD7E866ED@nutanix.com>
Date: Wed, 26 Nov 2025 16:21:33 +0000
From: Jon Kohler <jon@...anix.com>
To: Ido Schimmel <idosch@...dia.com>
CC: "David S. Miller" <davem@...emloft.net>,
Eric Dumazet
<edumazet@...gle.com>, Jakub Kicinski <kuba@...nel.org>,
Paolo Abeni
<pabeni@...hat.com>, Simon Horman <horms@...nel.org>,
Cong Wang
<xiyou.wangcong@...il.com>,
Nicolas Dichtel <nicolas.dichtel@...nd.com>,
Jason Wang <jasowang@...hat.com>,
"netdev@...r.kernel.org"
<netdev@...r.kernel.org>,
"linux-kernel@...r.kernel.org"
<linux-kernel@...r.kernel.org>
Subject: Re: [PATCH net-next] flow_dissector: save computed hash in
__skb_get_hash_symmetric_net()
> On Nov 26, 2025, at 2:45 AM, Ido Schimmel <idosch@...dia.com> wrote:
>
> On Tue, Nov 25, 2025 at 11:19:27AM -0700, Jon Kohler wrote:
>> tun.c changed from skb_get_hash() to __skb_get_hash_symmetric() on
>> commit feec084a7cf4 ("tun: use symmetric hash"), which exposes an
>> overhead for OVS datapath, where ovs_dp_process_packet() has to
>> calculate the hash again because __skb_get_hash_symmetric() does not
>> retain the hash that it calculates.
>
> This seems to be intentional according to commit eb70db875671 ("packet:
> Use symmetric hash for PACKET_FANOUT_HASH."): "This hash does not get
> installed into and override the normal skb hash, so this change has no
> effect whatsoever on the rest of the stack."
Ah! Good eye
What about a variant of this patch that had an arg like:
__skb_get_hash_symmetric(struct sk_buff *skb, bool save_hash)
Then we just make calls (like tun) opt in?
>
>>
>> Save the computed hash in __skb_get_hash_symmetric_net() so that the
>> calcuation work does not go to waste.
>>
>> Fixes: feec084a7cf4 ("tun: use symmetric hash")
>> Signed-off-by: Jon Kohler <jon@...anix.com>
>> ---
>> include/linux/skbuff.h | 4 ++--
>> net/core/flow_dissector.c | 7 +++++--
>> 2 files changed, 7 insertions(+), 4 deletions(-)
>>
>> diff --git a/include/linux/skbuff.h b/include/linux/skbuff.h
>> index ff90281ddf90..f58afa49a50e 100644
>> --- a/include/linux/skbuff.h
>> +++ b/include/linux/skbuff.h
>> @@ -1568,9 +1568,9 @@ __skb_set_sw_hash(struct sk_buff *skb, __u32 hash, bool is_l4)
>> __skb_set_hash(skb, hash, true, is_l4);
>> }
>>
>> -u32 __skb_get_hash_symmetric_net(const struct net *net, const struct sk_buff *skb);
>> +u32 __skb_get_hash_symmetric_net(const struct net *net, struct sk_buff *skb);
>>
>> -static inline u32 __skb_get_hash_symmetric(const struct sk_buff *skb)
>> +static inline u32 __skb_get_hash_symmetric(struct sk_buff *skb)
>> {
>> return __skb_get_hash_symmetric_net(NULL, skb);
>> }
>> diff --git a/net/core/flow_dissector.c b/net/core/flow_dissector.c
>> index 1b61bb25ba0e..4a74dcc4799c 100644
>> --- a/net/core/flow_dissector.c
>> +++ b/net/core/flow_dissector.c
>> @@ -1864,9 +1864,10 @@ EXPORT_SYMBOL(make_flow_keys_digest);
>>
>> static struct flow_dissector flow_keys_dissector_symmetric __read_mostly;
>>
>> -u32 __skb_get_hash_symmetric_net(const struct net *net, const struct sk_buff *skb)
>> +u32 __skb_get_hash_symmetric_net(const struct net *net, struct sk_buff *skb)
>> {
>> struct flow_keys keys;
>> + u32 flow_hash;
>>
>> __flow_hash_secret_init();
>>
>> @@ -1874,7 +1875,9 @@ u32 __skb_get_hash_symmetric_net(const struct net *net, const struct sk_buff *sk
>> __skb_flow_dissect(net, skb, &flow_keys_dissector_symmetric,
>> &keys, NULL, 0, 0, 0, 0);
>>
>> - return __flow_hash_from_keys(&keys, &hashrnd);
>> + flow_hash = __flow_hash_from_keys(&keys, &hashrnd);
>> + __skb_set_sw_hash(skb, flow_hash, flow_keys_have_l4(&keys));
>> + return flow_hash;
>> }
>> EXPORT_SYMBOL_GPL(__skb_get_hash_symmetric_net);
>>
>> --
>> 2.43.0
>>
Powered by blists - more mailing lists