> -----Original Message-----
> From: Richardson, Bruce <bruce.richard...@intel.com>
> Sent: Tuesday, June 17, 2025 7:34 PM
> To: Ye, MingjinX <mingjinx...@intel.com>
> Cc: dev@dpdk.org; sta...@dpdk.org; Burakov, Anatoly
> <anatoly.bura...@intel.com>
> Subject: Re: [PATCH] net/ice: fix DCF RSS hash update
>
> On Wed, Jun 11, 2025 at 09:50:57AM +0000, Mingjin Ye wrote:
> > Get rss hash configuration, rss_hf is always the default value.
> > The driver does nothing if the rss key is invalid during the rss hash
> > update.
> >
> > This patch is get the current configuration of rss_hf. Extract the
> > update rss key code from ice_dcf_dev_rss_hash_update to
> > ice_dcf_set_rss_key and make it consistent with the pf behaviour.
> >
> > Fixes: c223cadc9e5f ("net/ice: support RSS hash configuration in DCF
> > mode")
> > Cc: sta...@dpdk.org
> >
> > Signed-off-by: Mingjin Ye <mingjinx...@intel.com>
>
> Hi,
>
> some comments inline below.
>
> /Bruce
>
> > ---
> > drivers/net/intel/ice/ice_dcf_ethdev.c | 38
> > +++++++++++++++-----------
> > 1 file changed, 22 insertions(+), 16 deletions(-)
> >
> > diff --git a/drivers/net/intel/ice/ice_dcf_ethdev.c
> > b/drivers/net/intel/ice/ice_dcf_ethdev.c
> > index efff76afa8..accabd0ab9 100644
> > --- a/drivers/net/intel/ice/ice_dcf_ethdev.c
> > +++ b/drivers/net/intel/ice/ice_dcf_ethdev.c
> > @@ -1394,31 +1394,38 @@ ice_dcf_dev_rss_reta_query(struct
> rte_eth_dev
> > *dev, }
> >
> > static int
> > -ice_dcf_dev_rss_hash_update(struct rte_eth_dev *dev,
> > - struct rte_eth_rss_conf *rss_conf)
> > +ice_dcf_set_rss_key(struct ice_dcf_hw *hw, uint8_t *key, uint8_t
> > +key_len)
> > {
> > - struct ice_dcf_adapter *adapter = dev->data->dev_private;
> > - struct ice_dcf_hw *hw = &adapter->real_hw;
> > - int ret;
> > -
> > - if (!(hw->vf_res->vf_cap_flags & VIRTCHNL_VF_OFFLOAD_RSS_PF))
> > - return -ENOTSUP;
> > -
> > /* HENA setting, it is enabled by default, no change */
> > - if (!rss_conf->rss_key || rss_conf->rss_key_len == 0) {
> > + if (!key || key_len == 0) {
> > PMD_DRV_LOG(DEBUG, "No key to be configured");
> > return 0;
> > - } else if (rss_conf->rss_key_len != hw->vf_res->rss_key_size) {
> > + } else if (key_len != hw->vf_res->rss_key_size) {
> > PMD_DRV_LOG(ERR, "The size of hash key configured "
> > "(%d) doesn't match the size of hardware can "
> > - "support (%d)", rss_conf->rss_key_len,
> > + "support (%d)", key_len,
> > hw->vf_res->rss_key_size);
> > return -EINVAL;
> > }
> >
> > - rte_memcpy(hw->rss_key, rss_conf->rss_key, rss_conf-
> >rss_key_len);
> > + rte_memcpy(hw->rss_key, key, key_len);
> > +
> > + return ice_dcf_configure_rss_key(hw); }
> > +
> > +static int
> > +ice_dcf_dev_rss_hash_update(struct rte_eth_dev *dev,
> > + struct rte_eth_rss_conf *rss_conf) {
> > + struct ice_dcf_adapter *adapter = dev->data->dev_private;
> > + struct ice_dcf_hw *hw = &adapter->real_hw;
> > + int ret;
> > +
> > + if (!(hw->vf_res->vf_cap_flags & VIRTCHNL_VF_OFFLOAD_RSS_PF))
> > + return -ENOTSUP;
> >
> > - ret = ice_dcf_configure_rss_key(hw);
> > + /* set hash key */
> > + ret = ice_dcf_set_rss_key(hw, rss_conf->rss_key,
> > +rss_conf->rss_key_len);
> > if (ret)
> > return ret;
> >
>
> This part of the diff (from start of the patch to here) is just extracting the
> existing code into a new function, right? There is no change to behaviour
> here?
No logic was changed. There is a difference in behavior. Previously, when the
rss_key is null, noting will happen. But now it continues to perform the next
steps. This patch makes the behavior of PF and VF consistent.
>
> > @@ -1452,8 +1459,7 @@ ice_dcf_dev_rss_hash_conf_get(struct
> rte_eth_dev *dev,
> > if (!(hw->vf_res->vf_cap_flags & VIRTCHNL_VF_OFFLOAD_RSS_PF))
> > return -ENOTSUP;
> >
> > - /* Just set it to default value now. */
> > - rss_conf->rss_hf = ICE_RSS_OFFLOAD_ALL;
> > + rss_conf->rss_hf = dev->data-
> >dev_conf.rx_adv_conf.rss_conf.rss_hf;
> >
>
> This line seems to be the main fix in this patch. Is it worth splitting the
> patch
> in two - having this fix and the code refactor above in separate patches?
It is okay to split the patch as these two are for separate purpose.
>
> > if (!rss_conf->rss_key)
> > return 0;
> > --
> > 2.25.1
> >
