[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <39446ac7-9ce1-1c81-427b-a9821145fd1d@pensando.io>
Date: Wed, 4 Mar 2020 16:24:01 -0800
From: Shannon Nelson <snelson@...sando.io>
To: Jakub Kicinski <kuba@...nel.org>
Cc: davem@...emloft.net, netdev@...r.kernel.org
Subject: Re: [PATCH v2 net-next 5/8] ionic: support ethtool rxhash disable
On 3/4/20 11:59 AM, Jakub Kicinski wrote:
> On Tue, 3 Mar 2020 20:20:10 -0800 Shannon Nelson wrote:
>> We can disable rxhashing by setting rss_types to 0. The user
>> can toggle this with "ethtool -K <ethX> rxhash off|on",
>> which calls into the .ndo_set_features callback with the
>> NETIF_F_RXHASH feature bit set or cleared. This patch adds
>> a check for that bit and updates the FW if necessary.
>>
>> Signed-off-by: Shannon Nelson <snelson@...sando.io>
>> ---
>> drivers/net/ethernet/pensando/ionic/ionic_lif.c | 11 +++++++++--
>> 1 file changed, 9 insertions(+), 2 deletions(-)
>>
>> diff --git a/drivers/net/ethernet/pensando/ionic/ionic_lif.c b/drivers/net/ethernet/pensando/ionic/ionic_lif.c
>> index d1567e477b1f..4b953f9e9084 100644
>> --- a/drivers/net/ethernet/pensando/ionic/ionic_lif.c
>> +++ b/drivers/net/ethernet/pensando/ionic/ionic_lif.c
>> @@ -1094,6 +1094,7 @@ static int ionic_set_nic_features(struct ionic_lif *lif,
>> u64 vlan_flags = IONIC_ETH_HW_VLAN_TX_TAG |
>> IONIC_ETH_HW_VLAN_RX_STRIP |
>> IONIC_ETH_HW_VLAN_RX_FILTER;
>> + u64 old_hw_features;
>> int err;
>>
>> ctx.cmd.lif_setattr.features = ionic_netdev_features_to_nic(features);
>> @@ -1101,9 +1102,13 @@ static int ionic_set_nic_features(struct ionic_lif *lif,
>> if (err)
>> return err;
>>
>> + old_hw_features = lif->hw_features;
>> lif->hw_features = le64_to_cpu(ctx.cmd.lif_setattr.features &
>> ctx.comp.lif_setattr.features);
>>
>> + if ((old_hw_features ^ lif->hw_features) & IONIC_ETH_HW_RX_HASH)
>> + ionic_lif_rss_config(lif, lif->rss_types, NULL, NULL);
> Is this change coming from the HW or from ethtool? AFAIK hw_features
> are what's supported, features is what's enabled..
This is looking at the feature bits coming in from ndo_set_features - if
the RX_HASH bit has been turned off in the incoming features bitmask,
then I need to disable the hw hashing.
I believe the confusion is between lif->hw_features, describing what is
currently enabled in the hw, versus the netdev->hw_features, that is
what we've told the the stack we have available.
sln
Powered by blists - more mailing lists