[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <878rhpfcc4.fsf@nvidia.com>
Date: Thu, 26 Jan 2023 19:12:56 +0200
From: Vlad Buslov <vladbu@...dia.com>
To: Jakub Kicinski <kuba@...nel.org>
CC: <davem@...emloft.net>, <pabeni@...hat.com>, <pablo@...filter.org>,
<netdev@...r.kernel.org>, <netfilter-devel@...r.kernel.org>,
<jhs@...atatu.com>, <xiyou.wangcong@...il.com>, <jiri@...nulli.us>,
<ozsh@...dia.com>, <marcelo.leitner@...il.com>,
<simon.horman@...igine.com>
Subject: Re: [PATCH net-next v4 6/7] net/sched: act_ct: offload UDP NEW
connections
On Tue 24 Jan 2023 at 21:42, Jakub Kicinski <kuba@...nel.org> wrote:
> On Tue, 24 Jan 2023 15:02:06 +0100 Vlad Buslov wrote:
>> Modify the offload algorithm of UDP connections to the following:
>>
>> - Offload NEW connection as unidirectional.
>>
>> - When connection state changes to ESTABLISHED also update the hardware
>> flow. However, in order to prevent act_ct from spamming offload add wq for
>> every packet coming in reply direction in this state verify whether
>> connection has already been updated to ESTABLISHED in the drivers. If that
>> it the case, then skip flow_table and let conntrack handle such packets
>> which will also allow conntrack to potentially promote the connection to
>> ASSURED.
>>
>> - When connection state changes to ASSURED set the flow_table flow
>> NF_FLOW_HW_BIDIRECTIONAL flag which will cause refresh mechanism to offload
>> the reply direction.
>>
>> All other protocols have their offload algorithm preserved and are always
>> offloaded as bidirectional.
>>
>> Note that this change tries to minimize the load on flow_table add
>> workqueue. First, it tracks the last ctinfo that was offloaded by using new
>> flow 'ext_data' field and doesn't schedule the refresh for reply direction
>> packets when the offloads have already been updated with current ctinfo.
>> Second, when 'add' task executes on workqueue it always update the offload
>> with current flow state (by checking 'bidirectional' flow flag and
>> obtaining actual ctinfo/cookie through meta action instead of caching any
>> of these from the moment of scheduling the 'add' work) preventing the need
>> from scheduling more updates if state changed concurrently while the 'add'
>> work was pending on workqueue.
>
> Clang is not happy:
>
> net/sched/act_ct.c:677:12: warning: cast to smaller integer type 'enum ip_conntrack_info' from 'typeof (_Generic((flow->ext_data), char: (char)0, unsigned char: (unsigned char)0, signed char: (signed char)0, unsigned short: (unsigned short)0, short: (short)0, unsigned int: (unsigned int)0, int: (int)0, unsigned long: (unsigned long)0, long: (long)0, unsigned long long: (unsigned long long)0, long long: (long long)0, default: (flow->ext_data)))' (aka 'void *') [-Wvoid-pointer-to-enum-cast]
> else if ((enum ip_conntrack_info)READ_ONCE(flow->ext_data) ==
> ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Thanks, will fix it in v5.
Powered by blists - more mailing lists