| Message ID | 1447178205-99510-1-git-send-email-anjali.singhai@intel.com |
|---|---|
| State | Changes Requested |
| Delegated to: | Jeff Kirsher |
| Headers | show |
Hi Anjali, On Tue, 2015-11-10 at 09:56 -0800, Anjali Singhai Jain wrote: > If the kernel flag CONFIG_VXLAN is true or CONFIG_VXLAN_MODULE is > true, > enable VXLAN offload in the driver. > > v2: Fix bisection error for this patch series. > > Signed-off-by: Kiran Patil <kiran.patil@intel.com> > Signed-off-by: Anjali Singhai Jain <anjali.singhai@intel.com> > --- > drivers/net/ethernet/intel/i40e/i40e.h | 2 -- > drivers/net/ethernet/intel/i40e/i40e_main.c | 10 ++-------- > 2 files changed, 2 insertions(+), 10 deletions(-) > > diff --git a/drivers/net/ethernet/intel/i40e/i40e.h > b/drivers/net/ethernet/intel/i40e/i40e.h > index 8ed759e..6c4d154 100644 > --- a/drivers/net/ethernet/intel/i40e/i40e.h > +++ b/drivers/net/ethernet/intel/i40e/i40e.h > @@ -282,11 +282,9 @@ struct i40e_pf { > u32 fd_atr_cnt; > u32 fd_tcp_rule; > > -#if IS_ENABLED(CONFIG_VXLAN) > __be16 vxlan_ports[I40E_MAX_PF_UDP_OFFLOAD_PORTS]; > u16 pending_vxlan_bitmap; > > -#endif > enum i40e_interrupt_policy int_policy; > u16 rx_itr_default; > u16 tx_itr_default; > diff --git a/drivers/net/ethernet/intel/i40e/i40e_main.c > b/drivers/net/ethernet/intel/i40e/i40e_main.c > index b447af6..0235a8a 100644 > --- a/drivers/net/ethernet/intel/i40e/i40e_main.c > +++ b/drivers/net/ethernet/intel/i40e/i40e_main.c > @@ -6990,13 +6990,13 @@ static void i40e_handle_mdd_event(struct > i40e_pf *pf) > i40e_flush(hw); > } > > -#if IS_ENABLED(CONFIG_VXLAN) > /** > * i40e_sync_vxlan_filters_subtask - Sync the VSI filter list with > HW > * @pf: board private structure > **/ > static void i40e_sync_vxlan_filters_subtask(struct i40e_pf *pf) > { > +#if IS_ENABLED(CONFIG_VXLAN) Direct use of CONFIG_VXLAN is not going to work. What happens in the case where CONFIG_VXLAN=m and CONFIG_I40E=y? The kernel will load i40e as it loads itself at boot and vxlan module will not get included (because i40e isn't a module so it doesn't run depmod etc) and then it will run and try to call vxlan code and fail because of an unknown symbol since vxlan won't have been started yet. This is the entire reason why CONFIG_XXX_VXLAN exists. The alternative is to use select VXLAN and force VXLAN on when we enable I40E. That's the primary reason why we can't directly use CONFIG_VXLAN as a check in driver code. Regards, Jake > struct i40e_hw *hw = &pf->hw; > i40e_status ret; > __be16 port; > @@ -7030,9 +7030,9 @@ static void > i40e_sync_vxlan_filters_subtask(struct i40e_pf *pf) > } > } > } > +#endif > } > > -#endif > /** > * i40e_service_task - Run the driver's async subtasks > * @work: pointer to work_struct containing our data > @@ -7057,9 +7057,7 @@ static void i40e_service_task(struct > work_struct *work) > i40e_watchdog_subtask(pf); > i40e_fdir_reinit_subtask(pf); > i40e_sync_filters_subtask(pf); > -#if IS_ENABLED(CONFIG_VXLAN) > i40e_sync_vxlan_filters_subtask(pf); > -#endif > i40e_clean_adminq_subtask(pf); > > i40e_service_event_complete(pf); > @@ -8433,7 +8431,6 @@ static int i40e_set_features(struct net_device > *netdev, > return 0; > } > > -#if IS_ENABLED(CONFIG_VXLAN) > /** > * i40e_get_vxlan_port_idx - Lookup a possibly offloaded for Rx UDP > port > * @pf: board private structure > @@ -8528,7 +8525,6 @@ static void i40e_del_vxlan_port(struct > net_device *netdev, > } > } > > -#endif > static int i40e_get_phys_port_id(struct net_device *netdev, > struct netdev_phys_item_id *ppid) > { > @@ -8753,10 +8749,8 @@ static const struct net_device_ops > i40e_netdev_ops = { > .ndo_get_vf_config = i40e_ndo_get_vf_config, > .ndo_set_vf_link_state = i40e_ndo_set_vf_link_state, > .ndo_set_vf_spoofchk = i40e_ndo_set_vf_spoofchk, > -#if IS_ENABLED(CONFIG_VXLAN) > .ndo_add_vxlan_port = i40e_add_vxlan_port, > .ndo_del_vxlan_port = i40e_del_vxlan_port, > -#endif > .ndo_get_phys_port_id = i40e_get_phys_port_id, > .ndo_fdb_add = i40e_ndo_fdb_add, > .ndo_features_check = i40e_features_check,
Jake this patch doesn't change the use of CONFIG_VXLAN in our driver to guard against a potential issue that you pointed out. Our driver had already changed to using CONFIG_VXLAN in a previous patch. But I believe that we will still be okay in the scenario that VXLAN module is not loaded because of two reasons 1) We check a flag in sync_vsi_filters if a filter was added or deleted before proceeding to do anything and that will return very quickly from the function when VXLAN or GENEVE module is not loaded. 2) udp_offload_get_port function has empty definitions in the base kernel that will always exist even when the module is not loaded. Having said that, I am still going to make sure that it's not an issue with a previous patch for our driver. Although the shift to using udp_offload_get_port with this patch series will help alleviate that issue. Anjali > -----Original Message----- > From: Keller, Jacob E > Sent: Tuesday, November 10, 2015 10:01 AM > To: intel-wired-lan@lists.osuosl.org; Singhai, Anjali > Subject: Re: [Intel-wired-lan] [PATCH v4 1/6] i40e: Remove > CONFIG_I40E_VXLAN > > Hi Anjali, > > On Tue, 2015-11-10 at 09:56 -0800, Anjali Singhai Jain wrote: > > If the kernel flag CONFIG_VXLAN is true or CONFIG_VXLAN_MODULE is > > true, enable VXLAN offload in the driver. > > > > v2: Fix bisection error for this patch series. > > > > Signed-off-by: Kiran Patil <kiran.patil@intel.com> > > Signed-off-by: Anjali Singhai Jain <anjali.singhai@intel.com> > > --- > > drivers/net/ethernet/intel/i40e/i40e.h | 2 -- > > drivers/net/ethernet/intel/i40e/i40e_main.c | 10 ++-------- > > 2 files changed, 2 insertions(+), 10 deletions(-) > > > > diff --git a/drivers/net/ethernet/intel/i40e/i40e.h > > b/drivers/net/ethernet/intel/i40e/i40e.h > > index 8ed759e..6c4d154 100644 > > --- a/drivers/net/ethernet/intel/i40e/i40e.h > > +++ b/drivers/net/ethernet/intel/i40e/i40e.h > > @@ -282,11 +282,9 @@ struct i40e_pf { > > u32 fd_atr_cnt; > > u32 fd_tcp_rule; > > > > -#if IS_ENABLED(CONFIG_VXLAN) > > __be16 vxlan_ports[I40E_MAX_PF_UDP_OFFLOAD_PORTS]; > > u16 pending_vxlan_bitmap; > > > > -#endif > > enum i40e_interrupt_policy int_policy; > > u16 rx_itr_default; > > u16 tx_itr_default; > > diff --git a/drivers/net/ethernet/intel/i40e/i40e_main.c > > b/drivers/net/ethernet/intel/i40e/i40e_main.c > > index b447af6..0235a8a 100644 > > --- a/drivers/net/ethernet/intel/i40e/i40e_main.c > > +++ b/drivers/net/ethernet/intel/i40e/i40e_main.c > > @@ -6990,13 +6990,13 @@ static void i40e_handle_mdd_event(struct > > i40e_pf *pf) > > i40e_flush(hw); > > } > > > > -#if IS_ENABLED(CONFIG_VXLAN) > > /** > > * i40e_sync_vxlan_filters_subtask - Sync the VSI filter list with HW > > * @pf: board private structure > > **/ > > static void i40e_sync_vxlan_filters_subtask(struct i40e_pf *pf) > > { > > +#if IS_ENABLED(CONFIG_VXLAN) > > Direct use of CONFIG_VXLAN is not going to work. What happens in the case > where CONFIG_VXLAN=m and CONFIG_I40E=y? > > The kernel will load i40e as it loads itself at boot and vxlan module will not get > included (because i40e isn't a module so it doesn't run depmod etc) and then > it will run and try to call vxlan code and fail because of an unknown symbol > since vxlan won't have been started yet. > > This is the entire reason why CONFIG_XXX_VXLAN exists. The alternative is > to use select VXLAN and force VXLAN on when we enable I40E. > > That's the primary reason why we can't directly use CONFIG_VXLAN as a > check in driver code. > > Regards, > Jake > > > struct i40e_hw *hw = &pf->hw; > > i40e_status ret; > > __be16 port; > > @@ -7030,9 +7030,9 @@ static void > > i40e_sync_vxlan_filters_subtask(struct i40e_pf *pf) > > } > > } > > } > > +#endif > > } > > > > -#endif > > /** > > * i40e_service_task - Run the driver's async subtasks > > * @work: pointer to work_struct containing our data @@ -7057,9 > > +7057,7 @@ static void i40e_service_task(struct work_struct *work) > > i40e_watchdog_subtask(pf); > > i40e_fdir_reinit_subtask(pf); > > i40e_sync_filters_subtask(pf); > > -#if IS_ENABLED(CONFIG_VXLAN) > > i40e_sync_vxlan_filters_subtask(pf); > > -#endif > > i40e_clean_adminq_subtask(pf); > > > > i40e_service_event_complete(pf); > > @@ -8433,7 +8431,6 @@ static int i40e_set_features(struct net_device > > *netdev, > > return 0; > > } > > > > -#if IS_ENABLED(CONFIG_VXLAN) > > /** > > * i40e_get_vxlan_port_idx - Lookup a possibly offloaded for Rx UDP > > port > > * @pf: board private structure > > @@ -8528,7 +8525,6 @@ static void i40e_del_vxlan_port(struct > > net_device *netdev, > > } > > } > > > > -#endif > > static int i40e_get_phys_port_id(struct net_device *netdev, > > struct netdev_phys_item_id *ppid) > > { > > @@ -8753,10 +8749,8 @@ static const struct net_device_ops > > i40e_netdev_ops = { > > .ndo_get_vf_config = i40e_ndo_get_vf_config, > > .ndo_set_vf_link_state = i40e_ndo_set_vf_link_state, > > .ndo_set_vf_spoofchk = i40e_ndo_set_vf_spoofchk, > > -#if IS_ENABLED(CONFIG_VXLAN) > > .ndo_add_vxlan_port = i40e_add_vxlan_port, > > .ndo_del_vxlan_port = i40e_del_vxlan_port, > > -#endif > > .ndo_get_phys_port_id = i40e_get_phys_port_id, > > .ndo_fdb_add = i40e_ndo_fdb_add, > > .ndo_features_check = i40e_features_check,
Ok. That should work, but a quick check would be helpful. Regards, Jake On Tue, 2015-11-10 at 19:08 +0000, Singhai, Anjali wrote: > Jake this patch doesn't change the use of CONFIG_VXLAN in our driver > to guard against a potential issue that you pointed out. > Our driver had already changed to using CONFIG_VXLAN in a previous > patch. > > But I believe that we will still be okay in the scenario that VXLAN > module is not loaded because of two reasons > 1) We check a flag in sync_vsi_filters if a filter was added or > deleted before proceeding to do anything and that will return very > quickly from the function when VXLAN or GENEVE module is not loaded. > 2) udp_offload_get_port function has empty definitions in the base > kernel that will always exist even when the module is not loaded. > > Having said that, I am still going to make sure that it's not an > issue with a previous patch for our driver. Although the shift to > using udp_offload_get_port with this patch series will help alleviate > that issue. > > Anjali > > > -----Original Message----- > > From: Keller, Jacob E > > Sent: Tuesday, November 10, 2015 10:01 AM > > To: intel-wired-lan@lists.osuosl.org; Singhai, Anjali > > Subject: Re: [Intel-wired-lan] [PATCH v4 1/6] i40e: Remove > > CONFIG_I40E_VXLAN > > > > Hi Anjali, > > > > On Tue, 2015-11-10 at 09:56 -0800, Anjali Singhai Jain wrote: > > > If the kernel flag CONFIG_VXLAN is true or CONFIG_VXLAN_MODULE is > > > true, enable VXLAN offload in the driver. > > > > > > v2: Fix bisection error for this patch series. > > > > > > Signed-off-by: Kiran Patil <kiran.patil@intel.com> > > > Signed-off-by: Anjali Singhai Jain <anjali.singhai@intel.com> > > > --- > > > drivers/net/ethernet/intel/i40e/i40e.h | 2 -- > > > drivers/net/ethernet/intel/i40e/i40e_main.c | 10 ++-------- > > > 2 files changed, 2 insertions(+), 10 deletions(-) > > > > > > diff --git a/drivers/net/ethernet/intel/i40e/i40e.h > > > b/drivers/net/ethernet/intel/i40e/i40e.h > > > index 8ed759e..6c4d154 100644 > > > --- a/drivers/net/ethernet/intel/i40e/i40e.h > > > +++ b/drivers/net/ethernet/intel/i40e/i40e.h > > > @@ -282,11 +282,9 @@ struct i40e_pf { > > > u32 fd_atr_cnt; > > > u32 fd_tcp_rule; > > > > > > -#if IS_ENABLED(CONFIG_VXLAN) > > > __be16 vxlan_ports[I40E_MAX_PF_UDP_OFFLOAD_PORTS]; > > > u16 pending_vxlan_bitmap; > > > > > > -#endif > > > enum i40e_interrupt_policy int_policy; > > > u16 rx_itr_default; > > > u16 tx_itr_default; > > > diff --git a/drivers/net/ethernet/intel/i40e/i40e_main.c > > > b/drivers/net/ethernet/intel/i40e/i40e_main.c > > > index b447af6..0235a8a 100644 > > > --- a/drivers/net/ethernet/intel/i40e/i40e_main.c > > > +++ b/drivers/net/ethernet/intel/i40e/i40e_main.c > > > @@ -6990,13 +6990,13 @@ static void i40e_handle_mdd_event(struct > > > i40e_pf *pf) > > > i40e_flush(hw); > > > } > > > > > > -#if IS_ENABLED(CONFIG_VXLAN) > > > /** > > > * i40e_sync_vxlan_filters_subtask - Sync the VSI filter list > > > with HW > > > * @pf: board private structure > > > **/ > > > static void i40e_sync_vxlan_filters_subtask(struct i40e_pf *pf) > > > { > > > +#if IS_ENABLED(CONFIG_VXLAN) > > > > Direct use of CONFIG_VXLAN is not going to work. What happens in > > the case > > where CONFIG_VXLAN=m and CONFIG_I40E=y? > > > > The kernel will load i40e as it loads itself at boot and vxlan > > module will not get > > included (because i40e isn't a module so it doesn't run depmod etc) > > and then > > it will run and try to call vxlan code and fail because of an > > unknown symbol > > since vxlan won't have been started yet. > > > > This is the entire reason why CONFIG_XXX_VXLAN exists. The > > alternative is > > to use select VXLAN and force VXLAN on when we enable I40E. > > > > That's the primary reason why we can't directly use CONFIG_VXLAN as > > a > > check in driver code. > > > > Regards, > > Jake > > > > > struct i40e_hw *hw = &pf->hw; > > > i40e_status ret; > > > __be16 port; > > > @@ -7030,9 +7030,9 @@ static void > > > i40e_sync_vxlan_filters_subtask(struct i40e_pf *pf) > > > } > > > } > > > } > > > +#endif > > > } > > > > > > -#endif > > > /** > > > * i40e_service_task - Run the driver's async subtasks > > > * @work: pointer to work_struct containing our data @@ -7057,9 > > > +7057,7 @@ static void i40e_service_task(struct work_struct > > > *work) > > > i40e_watchdog_subtask(pf); > > > i40e_fdir_reinit_subtask(pf); > > > i40e_sync_filters_subtask(pf); > > > -#if IS_ENABLED(CONFIG_VXLAN) > > > i40e_sync_vxlan_filters_subtask(pf); > > > -#endif > > > i40e_clean_adminq_subtask(pf); > > > > > > i40e_service_event_complete(pf); > > > @@ -8433,7 +8431,6 @@ static int i40e_set_features(struct > > > net_device > > > *netdev, > > > return 0; > > > } > > > > > > -#if IS_ENABLED(CONFIG_VXLAN) > > > /** > > > * i40e_get_vxlan_port_idx - Lookup a possibly offloaded for Rx > > > UDP > > > port > > > * @pf: board private structure > > > @@ -8528,7 +8525,6 @@ static void i40e_del_vxlan_port(struct > > > net_device *netdev, > > > } > > > } > > > > > > -#endif > > > static int i40e_get_phys_port_id(struct net_device *netdev, > > > struct netdev_phys_item_id > > > *ppid) > > > { > > > @@ -8753,10 +8749,8 @@ static const struct net_device_ops > > > i40e_netdev_ops = { > > > .ndo_get_vf_config = i40e_ndo_get_vf_config, > > > .ndo_set_vf_link_state = > > > i40e_ndo_set_vf_link_state, > > > .ndo_set_vf_spoofchk = i40e_ndo_set_vf_spoofchk, > > > -#if IS_ENABLED(CONFIG_VXLAN) > > > .ndo_add_vxlan_port = i40e_add_vxlan_port, > > > .ndo_del_vxlan_port = i40e_del_vxlan_port, > > > -#endif > > > .ndo_get_phys_port_id = i40e_get_phys_port_id, > > > .ndo_fdb_add = i40e_ndo_fdb_add, > > > .ndo_features_check = i40e_features_check,
> -----Original Message----- > From: Intel-wired-lan [mailto:intel-wired-lan-bounces@lists.osuosl.org] On > Behalf Of Anjali Singhai Jain > Sent: Tuesday, November 10, 2015 9:57 AM > To: intel-wired-lan@lists.osuosl.org > Subject: [Intel-wired-lan] [PATCH v4 1/6] i40e: Remove CONFIG_I40E_VXLAN > > If the kernel flag CONFIG_VXLAN is true or CONFIG_VXLAN_MODULE is true, > enable VXLAN offload in the driver. > > v2: Fix bisection error for this patch series. > > Signed-off-by: Kiran Patil <kiran.patil@intel.com> > Signed-off-by: Anjali Singhai Jain <anjali.singhai@intel.com> > --- > drivers/net/ethernet/intel/i40e/i40e.h | 2 -- > drivers/net/ethernet/intel/i40e/i40e_main.c | 10 ++-------- > 2 files changed, 2 insertions(+), 10 deletions(-) Tested-by: Andrew Bowers <andrewx.bowers@intel.com> Patch code changes altered by subesquent patch
> -----Original Message----- > From: Intel-wired-lan [mailto:intel-wired-lan-bounces@lists.osuosl.org] On > Behalf Of Anjali Singhai Jain > Sent: Tuesday, November 10, 2015 9:57 AM > To: intel-wired-lan@lists.osuosl.org > Subject: [Intel-wired-lan] [PATCH v4 1/6] i40e: Remove CONFIG_I40E_VXLAN > > If the kernel flag CONFIG_VXLAN is true or CONFIG_VXLAN_MODULE is true, > enable VXLAN offload in the driver. > > v2: Fix bisection error for this patch series. > > Signed-off-by: Kiran Patil <kiran.patil@intel.com> > Signed-off-by: Anjali Singhai Jain <anjali.singhai@intel.com> > --- > drivers/net/ethernet/intel/i40e/i40e.h | 2 -- > drivers/net/ethernet/intel/i40e/i40e_main.c | 10 ++-------- > 2 files changed, 2 insertions(+), 10 deletions(-) Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Ok, So I did do a check. This series does fix the issue where we no more have to worry about whether VXLAN or GENEVE is a module and the driver is in-built since udp_offload_get_port will always be defined.. But we will have to actually drop the previous patch that changes the driver to start using CONFIG_VXLAN, and rebase this patch series. I will go ahead and do that. Anjali > -----Original Message----- > From: Keller, Jacob E > Sent: Tuesday, November 10, 2015 11:23 AM > To: intel-wired-lan@lists.osuosl.org; Singhai, Anjali > Subject: Re: [Intel-wired-lan] [PATCH v4 1/6] i40e: Remove > CONFIG_I40E_VXLAN > > Ok. That should work, but a quick check would be helpful. > > Regards, > Jake > > On Tue, 2015-11-10 at 19:08 +0000, Singhai, Anjali wrote: > > Jake this patch doesn't change the use of CONFIG_VXLAN in our driver > > to guard against a potential issue that you pointed out. > > Our driver had already changed to using CONFIG_VXLAN in a previous > > patch. > > > > But I believe that we will still be okay in the scenario that VXLAN > > module is not loaded because of two reasons > > 1) We check a flag in sync_vsi_filters if a filter was added or > > deleted before proceeding to do anything and that will return very > > quickly from the function when VXLAN or GENEVE module is not loaded. > > 2) udp_offload_get_port function has empty definitions in the base > > kernel that will always exist even when the module is not loaded. > > > > Having said that, I am still going to make sure that it's not an issue > > with a previous patch for our driver. Although the shift to using > > udp_offload_get_port with this patch series will help alleviate that > > issue. > > > > Anjali > > > > > -----Original Message----- > > > From: Keller, Jacob E > > > Sent: Tuesday, November 10, 2015 10:01 AM > > > To: intel-wired-lan@lists.osuosl.org; Singhai, Anjali > > > Subject: Re: [Intel-wired-lan] [PATCH v4 1/6] i40e: Remove > > > CONFIG_I40E_VXLAN > > > > > > Hi Anjali, > > > > > > On Tue, 2015-11-10 at 09:56 -0800, Anjali Singhai Jain wrote: > > > > If the kernel flag CONFIG_VXLAN is true or CONFIG_VXLAN_MODULE is > > > > true, enable VXLAN offload in the driver. > > > > > > > > v2: Fix bisection error for this patch series. > > > > > > > > Signed-off-by: Kiran Patil <kiran.patil@intel.com> > > > > Signed-off-by: Anjali Singhai Jain <anjali.singhai@intel.com> > > > > --- > > > > drivers/net/ethernet/intel/i40e/i40e.h | 2 -- > > > > drivers/net/ethernet/intel/i40e/i40e_main.c | 10 ++-------- > > > > 2 files changed, 2 insertions(+), 10 deletions(-) > > > > > > > > diff --git a/drivers/net/ethernet/intel/i40e/i40e.h > > > > b/drivers/net/ethernet/intel/i40e/i40e.h > > > > index 8ed759e..6c4d154 100644 > > > > --- a/drivers/net/ethernet/intel/i40e/i40e.h > > > > +++ b/drivers/net/ethernet/intel/i40e/i40e.h > > > > @@ -282,11 +282,9 @@ struct i40e_pf { > > > > u32 fd_atr_cnt; > > > > u32 fd_tcp_rule; > > > > > > > > -#if IS_ENABLED(CONFIG_VXLAN) > > > > __be16 vxlan_ports[I40E_MAX_PF_UDP_OFFLOAD_PORTS]; > > > > u16 pending_vxlan_bitmap; > > > > > > > > -#endif > > > > enum i40e_interrupt_policy int_policy; > > > > u16 rx_itr_default; > > > > u16 tx_itr_default; > > > > diff --git a/drivers/net/ethernet/intel/i40e/i40e_main.c > > > > b/drivers/net/ethernet/intel/i40e/i40e_main.c > > > > index b447af6..0235a8a 100644 > > > > --- a/drivers/net/ethernet/intel/i40e/i40e_main.c > > > > +++ b/drivers/net/ethernet/intel/i40e/i40e_main.c > > > > @@ -6990,13 +6990,13 @@ static void i40e_handle_mdd_event(struct > > > > i40e_pf *pf) > > > > i40e_flush(hw); > > > > } > > > > > > > > -#if IS_ENABLED(CONFIG_VXLAN) > > > > /** > > > > * i40e_sync_vxlan_filters_subtask - Sync the VSI filter list > > > > with HW > > > > * @pf: board private structure > > > > **/ > > > > static void i40e_sync_vxlan_filters_subtask(struct i40e_pf *pf) > > > > { > > > > +#if IS_ENABLED(CONFIG_VXLAN) > > > > > > Direct use of CONFIG_VXLAN is not going to work. What happens in the > > > case where CONFIG_VXLAN=m and CONFIG_I40E=y? > > > > > > The kernel will load i40e as it loads itself at boot and vxlan > > > module will not get included (because i40e isn't a module so it > > > doesn't run depmod etc) and then it will run and try to call vxlan > > > code and fail because of an unknown symbol since vxlan won't have > > > been started yet. > > > > > > This is the entire reason why CONFIG_XXX_VXLAN exists. The > > > alternative is to use select VXLAN and force VXLAN on when we enable > > > I40E. > > > > > > That's the primary reason why we can't directly use CONFIG_VXLAN as > > > a check in driver code. > > > > > > Regards, > > > Jake > > > > > > > struct i40e_hw *hw = &pf->hw; > > > > i40e_status ret; > > > > __be16 port; > > > > @@ -7030,9 +7030,9 @@ static void > > > > i40e_sync_vxlan_filters_subtask(struct i40e_pf *pf) > > > > } > > > > } > > > > } > > > > +#endif > > > > } > > > > > > > > -#endif > > > > /** > > > > * i40e_service_task - Run the driver's async subtasks > > > > * @work: pointer to work_struct containing our data @@ -7057,9 > > > > +7057,7 @@ static void i40e_service_task(struct work_struct > > > > *work) > > > > i40e_watchdog_subtask(pf); > > > > i40e_fdir_reinit_subtask(pf); > > > > i40e_sync_filters_subtask(pf); > > > > -#if IS_ENABLED(CONFIG_VXLAN) > > > > i40e_sync_vxlan_filters_subtask(pf); > > > > -#endif > > > > i40e_clean_adminq_subtask(pf); > > > > > > > > i40e_service_event_complete(pf); @@ -8433,7 +8431,6 @@ static > > > > int i40e_set_features(struct net_device *netdev, > > > > return 0; > > > > } > > > > > > > > -#if IS_ENABLED(CONFIG_VXLAN) > > > > /** > > > > * i40e_get_vxlan_port_idx - Lookup a possibly offloaded for Rx > > > > UDP port > > > > * @pf: board private structure > > > > @@ -8528,7 +8525,6 @@ static void i40e_del_vxlan_port(struct > > > > net_device *netdev, > > > > } > > > > } > > > > > > > > -#endif > > > > static int i40e_get_phys_port_id(struct net_device *netdev, > > > > struct netdev_phys_item_id > > > > *ppid) > > > > { > > > > @@ -8753,10 +8749,8 @@ static const struct net_device_ops > > > > i40e_netdev_ops = { > > > > .ndo_get_vf_config = i40e_ndo_get_vf_config, > > > > .ndo_set_vf_link_state = > > > > i40e_ndo_set_vf_link_state, > > > > .ndo_set_vf_spoofchk = i40e_ndo_set_vf_spoofchk, > > > > -#if IS_ENABLED(CONFIG_VXLAN) > > > > .ndo_add_vxlan_port = i40e_add_vxlan_port, > > > > .ndo_del_vxlan_port = i40e_del_vxlan_port, > > > > -#endif > > > > .ndo_get_phys_port_id = i40e_get_phys_port_id, > > > > .ndo_fdb_add = i40e_ndo_fdb_add, > > > > .ndo_features_check = i40e_features_check,
diff --git a/drivers/net/ethernet/intel/i40e/i40e.h b/drivers/net/ethernet/intel/i40e/i40e.h index 8ed759e..6c4d154 100644 --- a/drivers/net/ethernet/intel/i40e/i40e.h +++ b/drivers/net/ethernet/intel/i40e/i40e.h @@ -282,11 +282,9 @@ struct i40e_pf { u32 fd_atr_cnt; u32 fd_tcp_rule; -#if IS_ENABLED(CONFIG_VXLAN) __be16 vxlan_ports[I40E_MAX_PF_UDP_OFFLOAD_PORTS]; u16 pending_vxlan_bitmap; -#endif enum i40e_interrupt_policy int_policy; u16 rx_itr_default; u16 tx_itr_default; diff --git a/drivers/net/ethernet/intel/i40e/i40e_main.c b/drivers/net/ethernet/intel/i40e/i40e_main.c index b447af6..0235a8a 100644 --- a/drivers/net/ethernet/intel/i40e/i40e_main.c +++ b/drivers/net/ethernet/intel/i40e/i40e_main.c @@ -6990,13 +6990,13 @@ static void i40e_handle_mdd_event(struct i40e_pf *pf) i40e_flush(hw); } -#if IS_ENABLED(CONFIG_VXLAN) /** * i40e_sync_vxlan_filters_subtask - Sync the VSI filter list with HW * @pf: board private structure **/ static void i40e_sync_vxlan_filters_subtask(struct i40e_pf *pf) { +#if IS_ENABLED(CONFIG_VXLAN) struct i40e_hw *hw = &pf->hw; i40e_status ret; __be16 port; @@ -7030,9 +7030,9 @@ static void i40e_sync_vxlan_filters_subtask(struct i40e_pf *pf) } } } +#endif } -#endif /** * i40e_service_task - Run the driver's async subtasks * @work: pointer to work_struct containing our data @@ -7057,9 +7057,7 @@ static void i40e_service_task(struct work_struct *work) i40e_watchdog_subtask(pf); i40e_fdir_reinit_subtask(pf); i40e_sync_filters_subtask(pf); -#if IS_ENABLED(CONFIG_VXLAN) i40e_sync_vxlan_filters_subtask(pf); -#endif i40e_clean_adminq_subtask(pf); i40e_service_event_complete(pf); @@ -8433,7 +8431,6 @@ static int i40e_set_features(struct net_device *netdev, return 0; } -#if IS_ENABLED(CONFIG_VXLAN) /** * i40e_get_vxlan_port_idx - Lookup a possibly offloaded for Rx UDP port * @pf: board private structure @@ -8528,7 +8525,6 @@ static void i40e_del_vxlan_port(struct net_device *netdev, } } -#endif static int i40e_get_phys_port_id(struct net_device *netdev, struct netdev_phys_item_id *ppid) { @@ -8753,10 +8749,8 @@ static const struct net_device_ops i40e_netdev_ops = { .ndo_get_vf_config = i40e_ndo_get_vf_config, .ndo_set_vf_link_state = i40e_ndo_set_vf_link_state, .ndo_set_vf_spoofchk = i40e_ndo_set_vf_spoofchk, -#if IS_ENABLED(CONFIG_VXLAN) .ndo_add_vxlan_port = i40e_add_vxlan_port, .ndo_del_vxlan_port = i40e_del_vxlan_port, -#endif .ndo_get_phys_port_id = i40e_get_phys_port_id, .ndo_fdb_add = i40e_ndo_fdb_add, .ndo_features_check = i40e_features_check,