[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <490eac5e-d429-427e-823f-556504428f9d@blackwall.org>
Date: Tue, 18 Mar 2025 09:24:18 +0200
From: Nikolay Aleksandrov <razor@...ckwall.org>
To: Wang Liang <wangliang74@...wei.com>,
Toke Høiland-Jørgensen <toke@...hat.com>,
jv@...sburgh.net, andrew+netdev@...n.ch, davem@...emloft.net,
edumazet@...gle.com, kuba@...nel.org, pabeni@...hat.com, ast@...nel.org,
daniel@...earbox.net, hawk@...nel.org, john.fastabend@...il.com,
joamaki@...il.com
Cc: yuehaibing@...wei.com, zhangchangzhong@...wei.com,
netdev@...r.kernel.org, linux-kernel@...r.kernel.org, bpf@...r.kernel.org
Subject: Re: [PATCH net] bonding: check xdp prog when set bond mode
On 3/18/25 08:50, Wang Liang wrote:
>
> 在 2025/3/17 15:39, Nikolay Aleksandrov 写道:
>> On 3/17/25 06:07, Wang Liang wrote:
>>> 在 2025/3/14 18:44, Nikolay Aleksandrov 写道:
>>>> On 3/14/25 12:22 PM, Nikolay Aleksandrov wrote:
>>>>> On 3/14/25 12:13 PM, Toke Høiland-Jørgensen wrote:
>>>>>> Wang Liang <wangliang74@...wei.com> writes:
>>>>>>
>>>>>>> Following operations can trigger a warning[1]:
>>>>>>>
>>>>>>> ip netns add ns1
>>>>>>> ip netns exec ns1 ip link add bond0 type bond mode balance-rr
>>>>>>> ip netns exec ns1 ip link set dev bond0 xdp obj af_xdp_kern.o sec xdp
>>>>>>> ip netns exec ns1 ip link set bond0 type bond mode broadcast
>>>>>>> ip netns del ns1
>>>>>>>
>>>>>>> When delete the namespace, dev_xdp_uninstall() is called to remove xdp
>>>>>>> program on bond dev, and bond_xdp_set() will check the bond mode. If bond
>>>>>>> mode is changed after attaching xdp program, the warning may occur.
>>>>>>>
>>>>>>> Some bond modes (broadcast, etc.) do not support native xdp. Set bond mode
>>>>>>> with xdp program attached is not good. Add check for xdp program when set
>>>>>>> bond mode.
>>>>>>>
>>>>>>> [1]
>>>>>>> ------------[ cut here ]------------
>>>>>>> WARNING: CPU: 0 PID: 11 at net/core/dev.c:9912 unregister_netdevice_many_notify+0x8d9/0x930
>>>>>>> Modules linked in:
>>>>>>> CPU: 0 UID: 0 PID: 11 Comm: kworker/u4:0 Not tainted 6.14.0-rc4 #107
>>>>>>> Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS rel-1.15.0-0-g2dd4b9b3f840-prebuilt.qemu.org 04/01/2014
>>>>>>> Workqueue: netns cleanup_net
>>>>>>> RIP: 0010:unregister_netdevice_many_notify+0x8d9/0x930
>>>>>>> Code: 00 00 48 c7 c6 6f e3 a2 82 48 c7 c7 d0 b3 96 82 e8 9c 10 3e ...
>>>>>>> RSP: 0018:ffffc90000063d80 EFLAGS: 00000282
>>>>>>> RAX: 00000000ffffffa1 RBX: ffff888004959000 RCX: 00000000ffffdfff
>>>>>>> RDX: 0000000000000000 RSI: 00000000ffffffea RDI: ffffc90000063b48
>>>>>>> RBP: ffffc90000063e28 R08: ffffffff82d39b28 R09: 0000000000009ffb
>>>>>>> R10: 0000000000000175 R11: ffffffff82d09b40 R12: ffff8880049598e8
>>>>>>> R13: 0000000000000001 R14: dead000000000100 R15: ffffc90000045000
>>>>>>> FS: 0000000000000000(0000) GS:ffff888007a00000(0000) knlGS:0000000000000000
>>>>>>> CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033
>>>>>>> CR2: 000000000d406b60 CR3: 000000000483e000 CR4: 00000000000006f0
>>>>>>> Call Trace:
>>>>>>> <TASK>
>>>>>>> ? __warn+0x83/0x130
>>>>>>> ? unregister_netdevice_many_notify+0x8d9/0x930
>>>>>>> ? report_bug+0x18e/0x1a0
>>>>>>> ? handle_bug+0x54/0x90
>>>>>>> ? exc_invalid_op+0x18/0x70
>>>>>>> ? asm_exc_invalid_op+0x1a/0x20
>>>>>>> ? unregister_netdevice_many_notify+0x8d9/0x930
>>>>>>> ? bond_net_exit_batch_rtnl+0x5c/0x90
>>>>>>> cleanup_net+0x237/0x3d0
>>>>>>> process_one_work+0x163/0x390
>>>>>>> worker_thread+0x293/0x3b0
>>>>>>> ? __pfx_worker_thread+0x10/0x10
>>>>>>> kthread+0xec/0x1e0
>>>>>>> ? __pfx_kthread+0x10/0x10
>>>>>>> ? __pfx_kthread+0x10/0x10
>>>>>>> ret_from_fork+0x2f/0x50
>>>>>>> ? __pfx_kthread+0x10/0x10
>>>>>>> ret_from_fork_asm+0x1a/0x30
>>>>>>> </TASK>
>>>>>>> ---[ end trace 0000000000000000 ]---
>>>>>>>
>>>>>>> Fixes: 9e2ee5c7e7c3 ("net, bonding: Add XDP support to the bonding driver")
>>>>>>> Signed-off-by: Wang Liang <wangliang74@...wei.com>
>>>>>>> ---
>>>>>>> drivers/net/bonding/bond_options.c | 3 +++
>>>>>>> 1 file changed, 3 insertions(+)
>>>>>>>
>>>>>>> diff --git a/drivers/net/bonding/bond_options.c b/drivers/net/bonding/bond_options.c
>>>>>>> index 327b6ecdc77e..127181866829 100644
>>>>>>> --- a/drivers/net/bonding/bond_options.c
>>>>>>> +++ b/drivers/net/bonding/bond_options.c
>>>>>>> @@ -868,6 +868,9 @@ static bool bond_set_xfrm_features(struct bonding *bond)
>>>>>>> static int bond_option_mode_set(struct bonding *bond,
>>>>>>> const struct bond_opt_value *newval)
>>>>>>> {
>>>>>>> + if (bond->xdp_prog)
>>>>>>> + return -EOPNOTSUPP;
>>>>>>> +
>>>>>> Should we allow changing as long as the new mode also supports XDP?
>>>>>>
>>>>>> -Toke
>>>>>>
>>>>>>
>>>>> +1
>>>>> I think we should allow it, the best way probably is to add a new option
>>>>> BOND_VALFLAG_XDP_UNSUPP (for example) as a bond option flag and to set
>>>>> it in bond_options.c for each mode that doesn't support XDP, then you
>>>>> can do the check in a generic way (for any option) in
>>>>> bond_opt_check_deps. Any bond option that can't be changed with XDP prog
>>>> err, I meant any bond option's value that isn't supported with XDP, for
>>>> a whole option it would be a bit different
>>> Thanks for your suggestions!
>>>
>>> When install xdp prog, bond_xdp_set() use bond_xdp_check() to check whether the bond mode support xdp.
>>>
>>> When uninstall xdp prog, the paramter prog of bond_xdp_set() is NULL. How about not call bond_xdp_check() to avoid the warning when the prog is NULL, like:
>>>
>>> static int bond_xdp_set(struct net_device *dev, struct bpf_prog *prog,
>>> struct netlink_ext_ack *extack)
>>> ...
>>> if (prog && !bond_xdp_check(bond))
>> No, this could cause other problems. Actually, for -net I think the best would be to stick to
>> a simpler fix and just do bond_xdp_check() if there's a XDP program attached when changing
>> the mode so it can be backported easier. The option value flag can be done in the future
>> if more option values (or options) need to be disabled for XDP.
>>
>> Cheers,
>> Nik
> Excuse me again. I hope that I haven't misunderstood your suggestions.
>
> When I try to do bond_xdp_check() if there's a XDP program attached when
> changing the mode in function bond_option_mode_set(). bond_xdp_check()
> visit bond->params.mode and bond->params.xmit_policy, should I create a
> new function, whose paramter is int mode?
>
> Looking forward to getting more replies. Thanks.
No need to create a new function, you can modify the existing and pass
mode as an argument, then use it to do this check when changing the mode.
>>>>> should have that flag set.
>>>>>
>>>>> Cheers,
>>>>> Nik
>>>>>
>>
Powered by blists - more mailing lists