Tested-by: Xie,WeiX < weix....@intel.com> Regards, Xie Wei
> -----Original Message----- > From: Zhang,Alvin [mailto:alvinx.zh...@intel.com] > Sent: Monday, January 18, 2021 1:04 PM > To: Guo, Jia <jia....@intel.com>; Xie, WeiX <weix....@intel.com>; Wang, > Haiyue <haiyue.w...@intel.com> > Cc: dev@dpdk.org; Zhang, AlvinX <alvinx.zh...@intel.com>; > sta...@dpdk.org > Subject: [PATCH v5] net/ixgbe: fix configuration of max frame size > > From: Alvin Zhang <alvinx.zh...@intel.com> > > For some types of NIC, jumbo frame is not supported in IOV mode, so if a VF > requests to configure the frame size to not bigger than > RTE_ETHER_MAX_LEN, the kernel driver returns 0, but the DPDK ixgbe PMD > returens -1, this will cause the VF to fail to start when the PF driven by > DPDK > ixgbe PMD. > > This patch keeps ixgbe PMD's handling mode consistent with kernel driver in > above situation. > > In addition, the value set by the command IXGBE_VF_SET_LPE represents > the max frame size, not the mtu. > > Fixes: 1b9ea09c067b ("ixgbe: support X550") > Fixes: 95a27b3ba5f5 ("net/ixgbe: enable jumbo frame for VF") > Cc: sta...@dpdk.org > > Signed-off-by: Alvin Zhang <alvinx.zh...@intel.com> > --- > > V3: Restore variable name from cur_frame_size to max_frs. > V4: Update git log and add notes. > V5: Add check for VF api version. > --- > drivers/net/ixgbe/ixgbe_pf.c | 43 > +++++++++++++++++++++++++++++++++++-------- > 1 file changed, 35 insertions(+), 8 deletions(-) > > diff --git a/drivers/net/ixgbe/ixgbe_pf.c b/drivers/net/ixgbe/ixgbe_pf.c > index 833863a..793f970 100644 > --- a/drivers/net/ixgbe/ixgbe_pf.c > +++ b/drivers/net/ixgbe/ixgbe_pf.c > @@ -552,20 +552,47 @@ int ixgbe_pf_host_configure(struct rte_eth_dev > *eth_dev) } > > static int > -ixgbe_set_vf_lpe(struct rte_eth_dev *dev, __rte_unused uint32_t vf, > uint32_t *msgbuf) > +ixgbe_set_vf_lpe(struct rte_eth_dev *dev, uint32_t vf, uint32_t > +*msgbuf) > { > struct ixgbe_hw *hw = IXGBE_DEV_PRIVATE_TO_HW(dev->data- > >dev_private); > - uint32_t new_mtu = msgbuf[1]; > + uint32_t max_frame = msgbuf[1]; > uint32_t max_frs; > uint32_t hlreg0; > - int max_frame = new_mtu + RTE_ETHER_HDR_LEN + > RTE_ETHER_CRC_LEN; > > /* X540 and X550 support jumbo frames in IOV mode */ > if (hw->mac.type != ixgbe_mac_X540 && > hw->mac.type != ixgbe_mac_X550 && > hw->mac.type != ixgbe_mac_X550EM_x && > - hw->mac.type != ixgbe_mac_X550EM_a) > - return -1; > + hw->mac.type != ixgbe_mac_X550EM_a) { > + struct ixgbe_vf_info *vfinfo = > + *IXGBE_DEV_PRIVATE_TO_P_VFDATA(dev->data- > >dev_private); > + > + switch (vfinfo[vf].api_version) { > + case ixgbe_mbox_api_11: > + case ixgbe_mbox_api_12: > + case ixgbe_mbox_api_13: > + /** > + * Version 1.1 supports jumbo frames on VFs if PF > has > + * jumbo frames enabled which means legacy VFs > are > + * disabled > + */ > + if (dev->data->dev_conf.rxmode.max_rx_pkt_len > > + RTE_ETHER_MAX_LEN) > + break; > + /* fall through */ > + default: > + /** > + * If the PF or VF are running jumbo frames enabled, > + * return -1 since cannot support jumbo frames on > + * legacy VFs. > + */ > + if (max_frame > RTE_ETHER_MAX_LEN || > + dev->data->dev_conf.rxmode.max_rx_pkt_len > > + RTE_ETHER_MAX_LEN) > + return -1; > + break; > + } > + } > > if (max_frame < RTE_ETHER_MIN_LEN || > max_frame > > RTE_ETHER_MAX_JUMBO_FRAME_LEN) @@ -573,9 +600,9 @@ int > ixgbe_pf_host_configure(struct rte_eth_dev *eth_dev) > > max_frs = (IXGBE_READ_REG(hw, IXGBE_MAXFRS) & > IXGBE_MHADD_MFS_MASK) >> > IXGBE_MHADD_MFS_SHIFT; > - if (max_frs < new_mtu) { > + if (max_frs < max_frame) { > hlreg0 = IXGBE_READ_REG(hw, IXGBE_HLREG0); > - if (new_mtu > RTE_ETHER_MAX_LEN) { > + if (max_frame > RTE_ETHER_MAX_LEN) { > dev->data->dev_conf.rxmode.offloads |= > DEV_RX_OFFLOAD_JUMBO_FRAME; > hlreg0 |= IXGBE_HLREG0_JUMBOEN; > @@ -586,7 +613,7 @@ int ixgbe_pf_host_configure(struct rte_eth_dev > *eth_dev) > } > IXGBE_WRITE_REG(hw, IXGBE_HLREG0, hlreg0); > > - max_frs = new_mtu << IXGBE_MHADD_MFS_SHIFT; > + max_frs = max_frame << IXGBE_MHADD_MFS_SHIFT; > IXGBE_WRITE_REG(hw, IXGBE_MAXFRS, max_frs); > } > > -- > 1.8.3.1
