On 12/1/2023 3:00 AM, Chaoyong He wrote: >> On 11/30/2023 8:52 AM, Chaoyong He wrote: >>> Fix the resource leak problem in the exit logic of CoreNIC firmware. >>> >>> Fixes: 646ea79ce481 ("net/nfp: move PF functions into its own file") >>> Cc: sta...@dpdk.org >>> >>> Signed-off-by: Chaoyong He <chaoyong...@corigine.com> >>> Reviewed-by: Long Wu <long...@corigine.com> >>> Reviewed-by: Peng Zhang <peng.zh...@corigine.com> >> >> <...> >> >>> +static int >>> +nfp_pf_secondary_uninit(struct nfp_pf_dev *pf_dev) { >>> + free(pf_dev->sym_tbl); >>> + rte_free(pf_dev); >>> + >>> + return 0; >>> +} >>> + >>> /* Reset and stop device. The device can not be restarted. */ static >>> int nfp_net_close(struct rte_eth_dev *dev) @@ -333,14 +381,25 @@ >>> nfp_net_close(struct rte_eth_dev *dev) >>> struct rte_pci_device *pci_dev; >>> struct nfp_app_fw_nic *app_fw_nic; >>> >>> - if (rte_eal_process_type() != RTE_PROC_PRIMARY) >>> - return 0; >>> - >>> hw = dev->data->dev_private; >>> pf_dev = hw->pf_dev; >>> pci_dev = RTE_ETH_DEV_TO_PCI(dev); >>> app_fw_nic = NFP_PRIV_TO_APP_FW_NIC(pf_dev->app_fw_priv); >>> >>> + /* >>> + * In secondary process, a released eth device can be found by its >> name >>> + * in shared memory. >>> + * If the state of the eth device is RTE_ETH_DEV_UNUSED, it means the >>> + * eth device has been released. >>> + */ >>> + if (rte_eal_process_type() == RTE_PROC_SECONDARY) { >>> + if (dev->state == RTE_ETH_DEV_UNUSED) >>> + return 0; >>> + >>> + nfp_pf_secondary_uninit(pf_dev); >>> + return 0; >>> + } >>> + >>> >> >> Mostly expectation is secondary process doesn't free shared resources, but >> init and free done by primary process. > > I agree. > Maybe the comment here make reader a little confused. > But the `nfp_pf_secondary_uninit()` does not free any shared resources, it > only free two memory which private to each secondary process. >
What freed is not process private, it is in the shared memory: hw = dev->data->dev_private; pf_dev = hw->pf_dev; rte_free(pf_dev); And when there are multiple secondaries, one of them frees `pf_dev`, how this is not effecting others that may use `pf_dev`? >> When there are multiple secondaries active, and if one of them closes the >> port, >> will system behave properly? Can you please double check above logic? > > Yes, the system behave properly. >