[PATCH] sched: fix for tc_ov_enable flag position in subport structure.

David Marchand david.marchand at redhat.com
Tue Jan 10 12:27:33 CET 2023


On Mon, Jan 9, 2023 at 3:59 PM Megha Ajmera <megha.ajmera at intel.com> wrote:
>
> Current position of "tv_ov_enable" variable in

tc_ov_enabled*


> rte_sched_subport structure makes the "memory" variable unused.

I did not enter the beast... but my understanding is that some object
internal to rte_sched_subport currently shares location with this
tc_ov_enabled field.
So please find a better title and describe the impact.


>
> Fixes: f5e60154ade ("sched: enable traffic class oversubscription conditionally")
> CC: marcinx.danilewicz at intel.com

This is stable at dpdk.org material, isn't it?


> Signed-off-by: Megha Ajmera <megha.ajmera at intel.com>
> ---
>  lib/sched/rte_sched.c | 5 +++--
>  1 file changed, 3 insertions(+), 2 deletions(-)
>
> diff --git a/lib/sched/rte_sched.c b/lib/sched/rte_sched.c
> index c91697131d..eaecd7ceb4 100644
> --- a/lib/sched/rte_sched.c
> +++ b/lib/sched/rte_sched.c
> @@ -202,6 +202,9 @@ struct rte_sched_subport {
>         uint32_t qsize_add[RTE_SCHED_QUEUES_PER_PIPE];
>         uint32_t qsize_sum;
>
> +       /* TC oversubscription activation */
> +       int tc_ov_enabled;
> +
>         struct rte_sched_pipe *pipe;
>         struct rte_sched_queue *queue;
>         struct rte_sched_queue_extra *queue_extra;
> @@ -210,8 +213,6 @@ struct rte_sched_subport {
>         struct rte_mbuf **queue_array;
>         uint8_t memory[0] __rte_cache_aligned;
>
> -       /* TC oversubscription activation */
> -       int tc_ov_enabled;
>  } __rte_cache_aligned;
>
>  struct rte_sched_port {



-- 
David Marchand



More information about the stable mailing list