| Message ID | 20240529071736.224973-1-karen.ostrowska@intel.com (mailing list archive) |
|---|---|
| State | Awaiting Upstream |
| Delegated to: | Netdev Maintainers |
| Headers | show |
| Series | [iwl-net,v1] ice: Rebuild TC queues on VSI queue reconfiguration | expand |
Dear Karen, dear Jan, Thank you for the patch. Am 29.05.24 um 09:17 schrieb Karen Ostrowska: > From: Jan Sokolowski <jan.sokolowski@intel.com> > > TC queues needs to be correctly updated when the number of queues on need > a VSI is reconfigured, so netdev's queue and TC settings will be > dynamically adjusted and could accurately represent the underlying > hardware state after changes to the VSI queue counts. Please document the test configuration, and how to test/verify your change. > Fixes: 0754d65bd4be ("ice: Add infrastructure for mqprio support via ndo_setup_tc") > Reviewed-by: Wojciech Drewek <wojciech.drewek@intel.com> > Signed-off-by: Jan Sokolowski <jan.sokolowski@intel.com> > Signed-off-by: Karen Ostrowska <karen.ostrowska@intel.com> > --- > drivers/net/ethernet/intel/ice/ice_main.c | 10 +++++++++- > 1 file changed, 9 insertions(+), 1 deletion(-) > > diff --git a/drivers/net/ethernet/intel/ice/ice_main.c b/drivers/net/ethernet/intel/ice/ice_main.c > index 1b61ca3a6eb6..a1798ec4d904 100644 > --- a/drivers/net/ethernet/intel/ice/ice_main.c > +++ b/drivers/net/ethernet/intel/ice/ice_main.c > @@ -4136,7 +4136,7 @@ bool ice_is_wol_supported(struct ice_hw *hw) > int ice_vsi_recfg_qs(struct ice_vsi *vsi, int new_rx, int new_tx, bool locked) > { > struct ice_pf *pf = vsi->back; > - int err = 0, timeout = 50; > + int i, err = 0, timeout = 50; unsigned int > if (!new_rx && !new_tx) > return -EINVAL; > @@ -4162,6 +4162,14 @@ int ice_vsi_recfg_qs(struct ice_vsi *vsi, int new_rx, int new_tx, bool locked) > > ice_vsi_close(vsi); > ice_vsi_rebuild(vsi, ICE_VSI_FLAG_NO_INIT); > + > + ice_for_each_traffic_class(i) { > + if (vsi->tc_cfg.ena_tc & BIT(i)) > + netdev_set_tc_queue(vsi->netdev, > + vsi->tc_cfg.tc_info[i].netdev_tc, > + vsi->tc_cfg.tc_info[i].qcount_tx, > + vsi->tc_cfg.tc_info[i].qoffset); > + } > ice_pf_dcb_recfg(pf, locked); > ice_vsi_open(vsi); > done: Kind regards, Paul
> -----Original Message----- > From: Intel-wired-lan <intel-wired-lan-bounces@osuosl.org> On Behalf Of Karen Ostrowska > Sent: Wednesday, May 29, 2024 12:48 PM > To: intel-wired-lan@lists.osuosl.org > Cc: Ostrowska, Karen <karen.ostrowska@intel.com>; netdev@vger.kernel.org; Sokolowski, Jan <jan.sokolowski@intel.com>; Drewek, Wojciech <wojciech.drewek@intel.com> > Subject: [Intel-wired-lan] [PATCH iwl-net v1] ice: Rebuild TC queues on VSI queue reconfiguration > > From: Jan Sokolowski <jan.sokolowski@intel.com> > > TC queues needs to be correctly updated when the number of queues on a VSI is reconfigured, so netdev's queue and TC settings will be dynamically adjusted and could accurately represent the underlying hardware state after changes to the VSI queue counts. > > Fixes: 0754d65bd4be ("ice: Add infrastructure for mqprio support via ndo_setup_tc") > Reviewed-by: Wojciech Drewek <wojciech.drewek@intel.com> > Signed-off-by: Jan Sokolowski <jan.sokolowski@intel.com> > Signed-off-by: Karen Ostrowska <karen.ostrowska@intel.com> > --- > drivers/net/ethernet/intel/ice/ice_main.c | 10 +++++++++- > 1 file changed, 9 insertions(+), 1 deletion(-) > Tested-by: Pucha Himasekhar Reddy <himasekharx.reddy.pucha@intel.com> (A Contingent worker at Intel)
diff --git a/drivers/net/ethernet/intel/ice/ice_main.c b/drivers/net/ethernet/intel/ice/ice_main.c index 1b61ca3a6eb6..a1798ec4d904 100644 --- a/drivers/net/ethernet/intel/ice/ice_main.c +++ b/drivers/net/ethernet/intel/ice/ice_main.c @@ -4136,7 +4136,7 @@ bool ice_is_wol_supported(struct ice_hw *hw) int ice_vsi_recfg_qs(struct ice_vsi *vsi, int new_rx, int new_tx, bool locked) { struct ice_pf *pf = vsi->back; - int err = 0, timeout = 50; + int i, err = 0, timeout = 50; if (!new_rx && !new_tx) return -EINVAL; @@ -4162,6 +4162,14 @@ int ice_vsi_recfg_qs(struct ice_vsi *vsi, int new_rx, int new_tx, bool locked) ice_vsi_close(vsi); ice_vsi_rebuild(vsi, ICE_VSI_FLAG_NO_INIT); + + ice_for_each_traffic_class(i) { + if (vsi->tc_cfg.ena_tc & BIT(i)) + netdev_set_tc_queue(vsi->netdev, + vsi->tc_cfg.tc_info[i].netdev_tc, + vsi->tc_cfg.tc_info[i].qcount_tx, + vsi->tc_cfg.tc_info[i].qoffset); + } ice_pf_dcb_recfg(pf, locked); ice_vsi_open(vsi); done: