Hi, chenxu On 08/27, chenxux...@intel.com wrote: >From: Di ChenxuX <chenxux...@intel.com> > >Set RTE_ETH_DEV_CLOSE_REMOVE upon probe so all the private >resources for the port can be freed by rte_eth_dev_close(). >This patch cover all the intel drivers >
Just setting RTE_ETH_DEV_CLOSE_REMOVE in probe stage is not enough, you can see the advice in http://git.dpdk.org/dpdk/commit/?id=23ea57a2a "When enabling RTE_ETH_DEV_CLOSE_REMOVE, the PMD must free all its private resources for the port in its dev_close function, it's advised to call the dev_close function in the remove function in order to support removing a device without closing its ports". You can refer to other PMD which has finished this transition, like 696202ca5396 ("net/mvpp2: remove resources when port is closed") Also, I'd suggest to split it to several patches, one patch for one PMD. Thanks, Xiaolong >Signed-off-by: Di ChenxuX <chenxux...@intel.com> >--- > drivers/net/e1000/igb_ethdev.c | 5 +++++ > drivers/net/fm10k/fm10k_ethdev.c | 5 +++++ > drivers/net/i40e/i40e_ethdev.c | 5 +++++ > drivers/net/ice/ice_ethdev.c | 5 +++++ > drivers/net/ixgbe/ixgbe_ethdev.c | 5 +++++ > 5 files changed, 25 insertions(+) > >diff --git a/drivers/net/e1000/igb_ethdev.c b/drivers/net/e1000/igb_ethdev.c >index fec2b4289..fe785d1d7 100644 >--- a/drivers/net/e1000/igb_ethdev.c >+++ b/drivers/net/e1000/igb_ethdev.c >@@ -843,6 +843,11 @@ eth_igb_dev_init(struct rte_eth_dev *eth_dev) > rte_ether_addr_copy((struct rte_ether_addr *)hw->mac.addr, > ð_dev->data->mac_addrs[0]); > >+ /* Pass the information to the rte_eth_dev_close() that it should also >+ * release the private port resources. >+ */ >+ dev->data->dev_flags |= RTE_ETH_DEV_CLOSE_REMOVE; >+ > /* initialize the vfta */ > memset(shadow_vfta, 0, sizeof(*shadow_vfta)); > >diff --git a/drivers/net/fm10k/fm10k_ethdev.c >b/drivers/net/fm10k/fm10k_ethdev.c >index db4d72129..b5a427c7b 100644 >--- a/drivers/net/fm10k/fm10k_ethdev.c >+++ b/drivers/net/fm10k/fm10k_ethdev.c >@@ -3103,6 +3103,11 @@ eth_fm10k_dev_init(struct rte_eth_dev *dev) > &dev->data->mac_addrs[0]); > } > >+ /* Pass the information to the rte_eth_dev_close() that it should also >+ * release the private port resources. >+ */ >+ dev->data->dev_flags |= RTE_ETH_DEV_CLOSE_REMOVE; >+ > /* Reset the hw statistics */ > fm10k_stats_reset(dev); > >diff --git a/drivers/net/i40e/i40e_ethdev.c b/drivers/net/i40e/i40e_ethdev.c >index 4e40b7ab5..a0536f14c 100644 >--- a/drivers/net/i40e/i40e_ethdev.c >+++ b/drivers/net/i40e/i40e_ethdev.c >@@ -1520,6 +1520,11 @@ eth_i40e_dev_init(struct rte_eth_dev *dev, void >*init_params __rte_unused) > rte_ether_addr_copy((struct rte_ether_addr *)hw->mac.perm_addr, > &dev->data->mac_addrs[0]); > >+ /* Pass the information to the rte_eth_dev_close() that it should also >+ * release the private port resources. >+ */ >+ dev->data->dev_flags |= RTE_ETH_DEV_CLOSE_REMOVE; >+ > /* Init dcb to sw mode by default */ > ret = i40e_dcb_init_configure(dev, TRUE); > if (ret != I40E_SUCCESS) { >diff --git a/drivers/net/ice/ice_ethdev.c b/drivers/net/ice/ice_ethdev.c >index 44a14cb8a..476c768ee 100644 >--- a/drivers/net/ice/ice_ethdev.c >+++ b/drivers/net/ice/ice_ethdev.c >@@ -1488,6 +1488,11 @@ ice_dev_init(struct rte_eth_dev *dev) > goto err_init_mac; > } > >+ /* Pass the information to the rte_eth_dev_close() that it should also >+ * release the private port resources. >+ */ >+ dev->data->dev_flags |= RTE_ETH_DEV_CLOSE_REMOVE; >+ > ret = ice_res_pool_init(&pf->msix_pool, 1, > hw->func_caps.common_cap.num_msix_vectors - 1); > if (ret) { >diff --git a/drivers/net/ixgbe/ixgbe_ethdev.c >b/drivers/net/ixgbe/ixgbe_ethdev.c >index 03fc1f717..3e2b537ec 100644 >--- a/drivers/net/ixgbe/ixgbe_ethdev.c >+++ b/drivers/net/ixgbe/ixgbe_ethdev.c >@@ -1245,6 +1245,11 @@ eth_ixgbe_dev_init(struct rte_eth_dev *eth_dev, void >*init_params __rte_unused) > return -ENOMEM; > } > >+ /* Pass the information to the rte_eth_dev_close() that it should also >+ * release the private port resources. >+ */ >+ dev->data->dev_flags |= RTE_ETH_DEV_CLOSE_REMOVE; >+ > /* initialize the vfta */ > memset(shadow_vfta, 0, sizeof(*shadow_vfta)); > >-- >2.17.1 >
