Re: [PATCH] net/mlx5e: TC, release meter if rule offload fails

[Jianbo Liu <jianbol@nvidia.com>]

On 11/7/2026 9:05 PM, netdev-bot+sashiko@kernel.org wrote:
> Thank you for your contribution! Sashiko AI review found 1 potential
> issue(s) to consider.
> 
> Critical: 0 · High: 0 · Medium: 1 · Low: 0

[ ... ]

>> +	rule = mlx5_eswitch_add_offloaded_rule(esw, spec, attr);
>> +	if (IS_ERR(rule) && meter_added)
>> +		mlx5e_tc_del_flow_meter(esw, attr);
> 
> [Severity: Medium]
> Can this lead to a double free of the post meter and an extra meter put?

[ ... ]

> Would it help to set attr->meter_attr.meter and
> attr->meter_attr.post_meter to NULL, and undo the attr->dest_ft change,
> after mlx5e_tc_del_flow_meter() in this failure path? A later unoffload
> would then skip the meter release.

Hi Wentao,

Thanks for the fix, the leak is real. This comment is valid too:
mlx5e_tc_del_flow_meter() leaves the pointers set, and
mlx5e_tc_rule_unoffload() releases the meter whenever
attr->meter_attr.meter is non-NULL. After a failed re-offload in
mlx5e_reoffload_encap() the flow is still marked OFFLOADED, so the next
unoffload releases the meter and post meter a second time.

Please clear both pointers in mlx5e_tc_del_flow_meter() in v2. The
dest_ft part is not needed, the next mlx5e_tc_add_flow_meter() sets it
again before it is used.

Thanks,
Jianbo