From: Brett Creeley <brett.creeley@xxxxxxxxx> commit e6ba5273d4ede03d075d7a116b8edad1f6115f4d upstream. The VF can be configured via the PF's ndo ops at the same time the PF is receiving/handling virtchnl messages. This has many issues, with one of them being the ndo op could be actively resetting a VF (i.e. resetting it to the default state and deleting/re-adding the VF's VSI) while a virtchnl message is being handled. The following error was seen because a VF ndo op was used to change a VF's trust setting while the VIRTCHNL_OP_CONFIG_VSI_QUEUES was ongoing: [35274.192484] ice 0000:88:00.0: Failed to set LAN Tx queue context, error: ICE_ERR_PARAM [35274.193074] ice 0000:88:00.0: VF 0 failed opcode 6, retval: -5 [35274.193640] iavf 0000:88:01.0: PF returned error -5 (IAVF_ERR_PARAM) to our request 6 Fix this by making sure the virtchnl handling and VF ndo ops that trigger VF resets cannot run concurrently. This is done by adding a struct mutex cfg_lock to each VF structure. For VF ndo ops, the mutex will be locked around the critical operations and VFR. Since the ndo ops will trigger a VFR, the virtchnl thread will use mutex_trylock(). This is done because if any other thread (i.e. VF ndo op) has the mutex, then that means the current VF message being handled is no longer valid, so just ignore it. This issue can be seen using the following commands: for i in {0..50}; do rmmod ice modprobe ice sleep 1 echo 1 > /sys/class/net/ens785f0/device/sriov_numvfs echo 1 > /sys/class/net/ens785f1/device/sriov_numvfs ip link set ens785f1 vf 0 trust on ip link set ens785f0 vf 0 trust on sleep 2 echo 0 > /sys/class/net/ens785f0/device/sriov_numvfs echo 0 > /sys/class/net/ens785f1/device/sriov_numvfs sleep 1 echo 1 > /sys/class/net/ens785f0/device/sriov_numvfs echo 1 > /sys/class/net/ens785f1/device/sriov_numvfs ip link set ens785f1 vf 0 trust on ip link set ens785f0 vf 0 trust on done Fixes: 7c710869d64e ("ice: Add handlers for VF netdevice operations") Cc: <stable@xxxxxxxxxxxxxxx> Signed-off-by: Brett Creeley <brett.creeley@xxxxxxxxx> Tested-by: Konrad Jankowski <konrad0.jankowski@xxxxxxxxx> Signed-off-by: Tony Nguyen <anthony.l.nguyen@xxxxxxxxx> [I had to fix the cherry-pick manually as the patch added a line around some context that was missing.] Signed-off-by: Jacob Keller <jacob.e.keller@xxxxxxxxx> Signed-off-by: Greg Kroah-Hartman <gregkh@xxxxxxxxxxxxxxxxxxx> --- drivers/net/ethernet/intel/ice/ice_virtchnl_pf.c | 25 +++++++++++++++++++++++ drivers/net/ethernet/intel/ice/ice_virtchnl_pf.h | 5 ++++ 2 files changed, 30 insertions(+) --- a/drivers/net/ethernet/intel/ice/ice_virtchnl_pf.c +++ b/drivers/net/ethernet/intel/ice/ice_virtchnl_pf.c @@ -646,6 +646,8 @@ void ice_free_vfs(struct ice_pf *pf) set_bit(ICE_VF_STATE_DIS, pf->vf[i].vf_states); ice_free_vf_res(&pf->vf[i]); } + + mutex_destroy(&pf->vf[i].cfg_lock); } if (ice_sriov_free_msix_res(pf)) @@ -1894,6 +1896,8 @@ static void ice_set_dflt_settings_vfs(st */ ice_vf_ctrl_invalidate_vsi(vf); ice_vf_fdir_init(vf); + + mutex_init(&vf->cfg_lock); } } @@ -4082,6 +4086,8 @@ ice_set_vf_port_vlan(struct net_device * return 0; } + mutex_lock(&vf->cfg_lock); + vf->port_vlan_info = vlanprio; if (vf->port_vlan_info) @@ -4091,6 +4097,7 @@ ice_set_vf_port_vlan(struct net_device * dev_info(dev, "Clearing port VLAN on VF %d\n", vf_id); ice_vc_reset_vf(vf); + mutex_unlock(&vf->cfg_lock); return 0; } @@ -4465,6 +4472,15 @@ error_handler: return; } + /* VF is being configured in another context that triggers a VFR, so no + * need to process this message + */ + if (!mutex_trylock(&vf->cfg_lock)) { + dev_info(dev, "VF %u is being configured in another context that will trigger a VFR, so there is no need to handle this message\n", + vf->vf_id); + return; + } + switch (v_opcode) { case VIRTCHNL_OP_VERSION: err = ice_vc_get_ver_msg(vf, msg); @@ -4553,6 +4569,8 @@ error_handler: dev_info(dev, "PF failed to honor VF %d, opcode %d, error %d\n", vf_id, v_opcode, err); } + + mutex_unlock(&vf->cfg_lock); } /** @@ -4668,6 +4686,8 @@ int ice_set_vf_mac(struct net_device *ne return -EINVAL; } + mutex_lock(&vf->cfg_lock); + /* VF is notified of its new MAC via the PF's response to the * VIRTCHNL_OP_GET_VF_RESOURCES message after the VF has been reset */ @@ -4686,6 +4706,7 @@ int ice_set_vf_mac(struct net_device *ne } ice_vc_reset_vf(vf); + mutex_unlock(&vf->cfg_lock); return 0; } @@ -4715,11 +4736,15 @@ int ice_set_vf_trust(struct net_device * if (trusted == vf->trusted) return 0; + mutex_lock(&vf->cfg_lock); + vf->trusted = trusted; ice_vc_reset_vf(vf); dev_info(ice_pf_to_dev(pf), "VF %u is now %strusted\n", vf_id, trusted ? "" : "un"); + mutex_unlock(&vf->cfg_lock); + return 0; } --- a/drivers/net/ethernet/intel/ice/ice_virtchnl_pf.h +++ b/drivers/net/ethernet/intel/ice/ice_virtchnl_pf.h @@ -74,6 +74,11 @@ struct ice_mdd_vf_events { struct ice_vf { struct ice_pf *pf; + /* Used during virtchnl message handling and NDO ops against the VF + * that will trigger a VFR + */ + struct mutex cfg_lock; + u16 vf_id; /* VF ID in the PF space */ u16 lan_vsi_idx; /* index into PF struct */ u16 ctrl_vsi_idx; Patches currently in stable-queue which might be from jacob.e.keller@xxxxxxxxx are queue-5.15/ice-fix-race-conditions-between-virtchnl-handling-and-vf-ndo-ops.patch queue-5.15/ice-fix-concurrent-reset-and-removal-of-vfs.patch