On Mon, Dec 4, 2017 at 1:48 PM, Yuval Mintz <yuv...@mellanox.com> wrote:
>> Advertise NETIF_F_GRO_HW and turn on or off hardware GRO based on
>> NETIF_F_GRO_FW flag.
>>
>> Cc: Ariel Elior <ariel.el...@cavium.com>
>> Cc: everest-linux...@cavium.com
>> Signed-off-by: Michael Chan <michael.c...@broadcom.com>
>> ---
>> drivers/net/ethernet/qlogic/qede/qede_filter.c | 9 ++-------
>> drivers/net/ethernet/qlogic/qede/qede_main.c | 4 ++--
>> 2 files changed, 4 insertions(+), 9 deletions(-)
>>
>> diff --git a/drivers/net/ethernet/qlogic/qede/qede_filter.c
>> b/drivers/net/ethernet/qlogic/qede/qede_filter.c
>> index c1a0708..7ee49b4 100644
>> --- a/drivers/net/ethernet/qlogic/qede/qede_filter.c
>> +++ b/drivers/net/ethernet/qlogic/qede/qede_filter.c
>> @@ -901,13 +901,8 @@ int qede_set_features(struct net_device *dev,
>> netdev_features_t features)
>> netdev_features_t changes = features ^ dev->features;
>> bool need_reload = false;
>>
>> - /* No action needed if hardware GRO is disabled during driver load */
>> - if (changes & NETIF_F_GRO) {
>> - if (dev->features & NETIF_F_GRO)
>> - need_reload = !edev->gro_disable;
>> - else
>> - need_reload = edev->gro_disable;
>> - }
>> + if (changes & NETIF_F_GRO_HW)
>> + need_reload = true;
>
> This doesn't look right; edev->gro_disable can change due to other
> conditions as well - otherwise, it would have been synonymous with
> (dev->features & NETIF_F_GRO).
>
Feel free to modify the patch. The idea is for the driver to
determine in ndo_fix_features()/ndo_set_features() if GRO_HW can be
enabled. For things like generic XDP, I think we can add code to
centrally disable GRO_HW and not have to worry about that in the
driver.
