linux-kernel - Re: [PATCH net 4/6] net/mlx5: Destroy vport QoS element when no configuration remains

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives
Hash Suite: Windows password security audit tool. GUI, reports in PDF.
[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <3a5673b7-47d2-4a59-a515-95b396364994@nvidia.com>
Date: Sun, 17 Aug 2025 11:36:43 +0300
From: Carolina Jubran <cjubran@...dia.com>
To: Przemek Kitszel <przemyslaw.kitszel@...el.com>,
 Tariq Toukan <tariqt@...dia.com>
Cc: Saeed Mahameed <saeedm@...dia.com>, Leon Romanovsky <leon@...nel.org>,
 Mark Bloch <mbloch@...dia.com>, netdev@...r.kernel.org,
 linux-rdma@...r.kernel.org, linux-kernel@...r.kernel.org,
 Gal Pressman <gal@...dia.com>, Dragos Tatulea <dtatulea@...dia.com>,
 Eric Dumazet <edumazet@...gle.com>, Jakub Kicinski <kuba@...nel.org>,
 Paolo Abeni <pabeni@...hat.com>, Andrew Lunn <andrew+netdev@...n.ch>,
 "David S. Miller" <davem@...emloft.net>
Subject: Re: [PATCH net 4/6] net/mlx5: Destroy vport QoS element when no
 configuration remains



On 14/08/2025 12:55, Przemek Kitszel wrote:
> On 8/13/25 16:31, Tariq Toukan wrote:
>> From: Carolina Jubran <cjubran@...dia.com>
>>
>> If a VF has been configured and the user later clears all QoS settings,
>> the vport element remains in the firmware QoS tree. This leads to
>> inconsistent behavior compared to VFs that were never configured, since
>> the FW assumes that unconfigured VFs are outside the QoS hierarchy.
>> As a result, the bandwidth share across VFs may differ, even though
>> none of them appear to have any configuration.
>>
>> Align the driver behavior with the FW expectation by destroying the
>> vport QoS element when all configurations are removed.
>>
>> Fixes: c9497c98901c ("net/mlx5: Add support for setting VF min rate")
>> Fixes: cf7e73770d1b ("net/mlx5: Manage TC arbiter nodes and implement 
>> full support for tc-bw")
>> Signed-off-by: Carolina Jubran <cjubran@...dia.com>
>> Reviewed-by: Cosmin Ratiu <cratiu@...dia.com>
>> Signed-off-by: Tariq Toukan <tariqt@...dia.com>
>> ---
>>   .../net/ethernet/mellanox/mlx5/core/esw/qos.c | 54 +++++++++++++++++--
>>   1 file changed, 51 insertions(+), 3 deletions(-)
> 
> 
>> +static bool esw_vport_qos_check_and_disable(struct mlx5_vport *vport,
>> +                        struct devlink_rate *parent,
>> +                        u64 tx_max, u64 tx_share,
>> +                        u32 *tc_bw)
>> +{
>> +    struct mlx5_eswitch *esw = vport->dev->priv.eswitch;
>> +
>> +    if (parent || tx_max || tx_share || !esw_qos_tc_bw_disabled(tc_bw))
>> +        return false;
>> +
>> +    esw_qos_lock(esw);
>> +    mlx5_esw_qos_vport_disable_locked(vport);
>> +    esw_qos_unlock(esw);
>> +
>> +    return true;
>> +}
>> +
>>   int mlx5_esw_qos_init(struct mlx5_eswitch *esw)
>>   {
>>       if (esw->qos.domain)
>> @@ -1703,6 +1731,11 @@ int 
>> mlx5_esw_devlink_rate_leaf_tx_share_set(struct devlink_rate 
>> *rate_leaf, void
>>       if (!mlx5_esw_allowed(esw))
>>           return -EPERM;
>> +    if (esw_vport_qos_check_and_disable(vport, rate_leaf->parent,
>> +                        rate_leaf->tx_max, tx_share,
>> +                        rate_leaf->tc_bw))
>> +        return 0;
>> +
> 
> I would rather keep executing the code that "sets tx_share to 0 and
> propagates the info", and only then prune all-0 nodes.
> Same for other params (tx_max, ...)
> 
> That would be less risky and more future-proof, and also would let your
> check&disable function to take less params.
> 
> Finally, the name is poor, what about:?
>      esw_vport_qos_prune_empty(vport, rate_leaf);
> (after applying my prev suggestion the above line will be at the
> bottom of function).
> 
> Also a note, that if you apply the above, it would be also good to
> keep the "esw_qos_lock() just once" (as it is now)
> 

Thanks for the review!
My initial thought was to reduce the amount of FW commands, but I agree 
with your points.
Also, thanks for the name suggestion.
I’ll fix and send v2.

Carolina

>>       err = esw_qos_devlink_rate_to_mbps(vport->dev, "tx_share", 
>> &tx_share, extack);
>>       if (err)
>>           return err;
>> @@ -1724,6 +1757,11 @@ int 
>> mlx5_esw_devlink_rate_leaf_tx_max_set(struct devlink_rate *rate_leaf, 
>> void *
>>       if (!mlx5_esw_allowed(esw))
>>           return -EPERM;
>> +    if (esw_vport_qos_check_and_disable(vport, rate_leaf->parent, 
>> tx_max,
>> +                        rate_leaf->tx_share,
>> +                        rate_leaf->tc_bw))
>> +        return 0;
>> +
>>       err = esw_qos_devlink_rate_to_mbps(vport->dev, "tx_max", 
>> &tx_max, extack);
>>       if (err)
>>           return err;
>> @@ -1749,6 +1787,11 @@ int mlx5_esw_devlink_rate_leaf_tc_bw_set(struct 
>> devlink_rate *rate_leaf,
>>       if (!mlx5_esw_allowed(esw))
>>           return -EPERM;
>> +    if (esw_vport_qos_check_and_disable(vport, rate_leaf->parent,
>> +                        rate_leaf->tx_max,
>> +                        rate_leaf->tx_share, tc_bw))
>> +        return 0;
>> +
>>       disable = esw_qos_tc_bw_disabled(tc_bw);
>>       esw_qos_lock(esw);
>> @@ -1930,6 +1973,11 @@ int 
>> mlx5_esw_devlink_rate_leaf_parent_set(struct devlink_rate *devlink_rate,
>>       struct mlx5_esw_sched_node *node;
>>       struct mlx5_vport *vport = priv;
>> +    if (esw_vport_qos_check_and_disable(vport, parent, devlink_rate- 
>> >tx_max,
>> +                        devlink_rate->tx_share,
>> +                        devlink_rate->tc_bw))
>> +        return 0;
>> +
>>       if (!parent)
>>           return mlx5_esw_qos_vport_update_parent(vport, NULL, extack);
>