From: Huisong Li <lihuisong@huawei.com> When stopping a bonded port, all slaves should be deactivated. But only active slaves are stopped. So fix it and do "deactivae_slave()" for active slaves. Fixes: 0911d4ec0183 ("net/bonding: fix crash when stopping mode 4 port") Cc: stable@dpdk.org Signed-off-by: Huisong Li <lihuisong@huawei.com> Signed-off-by: Min Hu (Connor) <humin29@huawei.com> --- drivers/net/bonding/rte_eth_bond_pmd.c | 20 +++++++++++--------- 1 file changed, 11 insertions(+), 9 deletions(-) diff --git a/drivers/net/bonding/rte_eth_bond_pmd.c b/drivers/net/bonding/rte_eth_bond_pmd.c index b305b6a35b..469dc71170 100644 --- a/drivers/net/bonding/rte_eth_bond_pmd.c +++ b/drivers/net/bonding/rte_eth_bond_pmd.c @@ -2118,18 +2118,20 @@ bond_ethdev_stop(struct rte_eth_dev *eth_dev) internals->link_status_polling_enabled = 0; for (i = 0; i < internals->slave_count; i++) { uint16_t slave_id = internals->slaves[i].port_id; + + internals->slaves[i].last_link_status = 0; + ret = rte_eth_dev_stop(slave_id); + if (ret != 0) { + RTE_BOND_LOG(ERR, "Failed to stop device on port %u", + slave_id); + return ret; + } + + /* active slaves need to deactivate. */ if (find_slave_by_id(internals->active_slaves, internals->active_slave_count, slave_id) != - internals->active_slave_count) { - internals->slaves[i].last_link_status = 0; - ret = rte_eth_dev_stop(slave_id); - if (ret != 0) { - RTE_BOND_LOG(ERR, "Failed to stop device on port %u", - slave_id); - return ret; - } + internals->active_slave_count) deactivate_slave(eth_dev, slave_id); - } } return 0; -- 2.33.0
From: Huisong Li <lihuisong@huawei.com> All slaves will be stopped and removed when closing a bonded port. But the while loop can not stop if both rte_eth_dev_stop and rte_eth_bond_slave_remove fail to run. Fixes: fb0379bc5db3 ("net/bonding: check stop call status") Cc: stable@dpdk.org Signed-off-by: Huisong Li <lihuisong@huawei.com> Signed-off-by: Min Hu (Connor) <humin29@huawei.com> --- drivers/net/bonding/rte_eth_bond_pmd.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/drivers/net/bonding/rte_eth_bond_pmd.c b/drivers/net/bonding/rte_eth_bond_pmd.c index 469dc71170..00d4deda44 100644 --- a/drivers/net/bonding/rte_eth_bond_pmd.c +++ b/drivers/net/bonding/rte_eth_bond_pmd.c @@ -2149,13 +2149,14 @@ bond_ethdev_close(struct rte_eth_dev *dev) return 0; RTE_BOND_LOG(INFO, "Closing bonded device %s", dev->device->name); - while (internals->slave_count != skipped) { + while (skipped < internals->slave_count) { uint16_t port_id = internals->slaves[skipped].port_id; if (rte_eth_dev_stop(port_id) != 0) { RTE_BOND_LOG(ERR, "Failed to stop device on port %u", port_id); skipped++; + continue; } if (rte_eth_bond_slave_remove(bond_port_id, port_id) != 0) { -- 2.33.0
From: Huisong Li <lihuisong@huawei.com> Starting or stopping a bonded port also starts or stops all active slaves under the bonded port. If this port is a bonded device, we need to modify the port status of all slaves. Fixes: 0e545d3047fe ("app/testpmd: check stopping port is not in bonding") Cc: stable@dpdk.org Signed-off-by: Huisong Li <lihuisong@huawei.com> Signed-off-by: Min Hu (Connor) <humin29@huawei.com> Acked-by: Aman Singh <aman.deep.singh@intel.com> --- app/test-pmd/cmdline.c | 1 + app/test-pmd/testpmd.c | 74 +++++++++++++++++++++++++++++++++++++++--- app/test-pmd/testpmd.h | 3 +- 3 files changed, 73 insertions(+), 5 deletions(-) diff --git a/app/test-pmd/cmdline.c b/app/test-pmd/cmdline.c index 7ab0575e64..f0efcf09f0 100644 --- a/app/test-pmd/cmdline.c +++ b/app/test-pmd/cmdline.c @@ -6671,6 +6671,7 @@ static void cmd_create_bonded_device_parsed(void *parsed_result, "Failed to enable promiscuous mode for port %u: %s - ignore\n", port_id, rte_strerror(-ret)); + ports[port_id].bond_flag = 1; ports[port_id].need_setup = 0; ports[port_id].port_status = RTE_PORT_STOPPED; } diff --git a/app/test-pmd/testpmd.c b/app/test-pmd/testpmd.c index fe2ce19f99..dc90600787 100644 --- a/app/test-pmd/testpmd.c +++ b/app/test-pmd/testpmd.c @@ -66,6 +66,9 @@ #ifdef RTE_EXEC_ENV_WINDOWS #include <process.h> #endif +#ifdef RTE_NET_BOND +#include <rte_eth_bond.h> +#endif #include "testpmd.h" @@ -597,11 +600,57 @@ eth_dev_configure_mp(uint16_t port_id, uint16_t nb_rx_q, uint16_t nb_tx_q, return 0; } +#ifdef RTE_NET_BOND +static int +change_bonding_slave_port_status(portid_t bond_pid, bool is_stop) +{ + portid_t slave_pids[RTE_MAX_ETHPORTS]; + struct rte_port *port; + int num_slaves; + portid_t slave_pid; + int i; + + num_slaves = rte_eth_bond_slaves_get(bond_pid, slave_pids, + RTE_MAX_ETHPORTS); + if (num_slaves < 0) { + fprintf(stderr, "Failed to get slave list for port = %u\n", + bond_pid); + return num_slaves; + } + + for (i = 0; i < num_slaves; i++) { + slave_pid = slave_pids[i]; + port = &ports[slave_pid]; + port->port_status = + is_stop ? RTE_PORT_STOPPED : RTE_PORT_STARTED; + } + + return 0; +} +#endif + static int eth_dev_start_mp(uint16_t port_id) { - if (is_proc_primary()) - return rte_eth_dev_start(port_id); + int ret; + + if (is_proc_primary()) { + ret = rte_eth_dev_start(port_id); + if (ret != 0) + return ret; + +#ifdef RTE_NET_BOND + struct rte_port *port = &ports[port_id]; + + /* + * Starting a bonded port also starts all slaves under the bonded + * device. So if this port is bond device, we need to modify the + * port status of these slaves. + */ + if (port->bond_flag == 1) + return change_bonding_slave_port_status(port_id, false); +#endif + } return 0; } @@ -609,8 +658,25 @@ eth_dev_start_mp(uint16_t port_id) static int eth_dev_stop_mp(uint16_t port_id) { - if (is_proc_primary()) - return rte_eth_dev_stop(port_id); + int ret; + + if (is_proc_primary()) { + ret = rte_eth_dev_stop(port_id); + if (ret != 0) + return ret; + +#ifdef RTE_NET_BOND + struct rte_port *port = &ports[port_id]; + + /* + * Stopping a bonded port also stops all slaves under the bonded + * device. So if this port is bond device, we need to modify the + * port status of these slaves. + */ + if (port->bond_flag == 1) + return change_bonding_slave_port_status(port_id, true); +#endif + } return 0; } diff --git a/app/test-pmd/testpmd.h b/app/test-pmd/testpmd.h index 31f766c965..67f253b30e 100644 --- a/app/test-pmd/testpmd.h +++ b/app/test-pmd/testpmd.h @@ -266,7 +266,8 @@ struct rte_port { uint32_t mc_addr_nb; /**< nb. of addr. in mc_addr_pool */ queueid_t queue_nb; /**< nb. of queues for flow rules */ uint32_t queue_sz; /**< size of a queue for flow rules */ - uint8_t slave_flag; /**< bonding slave port */ + uint8_t slave_flag : 1, /**< bonding slave port */ + bond_flag : 1; /**< port is bond device */ struct port_template *pattern_templ_list; /**< Pattern templates. */ struct port_template *actions_templ_list; /**< Actions templates. */ struct port_table *table_list; /**< Flow tables. */ -- 2.33.0
From: Huisong Li <lihuisong@huawei.com> Currently, some eth devices are added to bond device, these devices are not released when the quit command is executed in testpmd. This patch adds the release operation for all active slaves under a bond device. Fixes: 0e545d3047fe ("app/testpmd: check stopping port is not in bonding") Cc: stable@dpdk.org Signed-off-by: Huisong Li <lihuisong@huawei.com> Signed-off-by: Min Hu (Connor) <humin29@huawei.com> --- app/test-pmd/cmdline.c | 1 + app/test-pmd/testpmd.c | 48 ++++++++++++++++++++++++++++++++++++++++++ app/test-pmd/testpmd.h | 2 ++ 3 files changed, 51 insertions(+) diff --git a/app/test-pmd/cmdline.c b/app/test-pmd/cmdline.c index f0efcf09f0..cd3873e1e0 100644 --- a/app/test-pmd/cmdline.c +++ b/app/test-pmd/cmdline.c @@ -8891,6 +8891,7 @@ static void cmd_quit_parsed(__rte_unused void *parsed_result, __rte_unused void *data) { cmdline_quit(cl); + cl_quit = 1; } cmdline_parse_token_string_t cmd_quit_quit = diff --git a/app/test-pmd/testpmd.c b/app/test-pmd/testpmd.c index dc90600787..22700e073d 100644 --- a/app/test-pmd/testpmd.c +++ b/app/test-pmd/testpmd.c @@ -229,6 +229,7 @@ unsigned int xstats_display_num; /**< Size of extended statistics to show */ * option. Set flag to exit stats period loop after received SIGINT/SIGTERM. */ uint8_t f_quit; +uint8_t cl_quit; /* Quit testpmd from cmdline. */ /* * Max Rx frame size, set by '--max-pkt-len' parameter. @@ -3167,11 +3168,43 @@ remove_invalid_ports(void) nb_cfg_ports = nb_fwd_ports; } +#ifdef RTE_NET_BOND +static void +clear_bonding_slave_device(portid_t *slave_pids, uint16_t num_slaves) +{ + struct rte_port *port; + portid_t slave_pid; + uint16_t i; + + for (i = 0; i < num_slaves; i++) { + slave_pid = slave_pids[i]; + if (port_is_started(slave_pid) == 1) { + if (rte_eth_dev_stop(slave_pid) != 0) + fprintf(stderr, "rte_eth_dev_stop failed for port %u\n", + slave_pid); + + port = &ports[slave_pid]; + port->port_status = RTE_PORT_STOPPED; + } + + clear_port_slave_flag(slave_pid); + + /* Close slave device when testpmd quit or is killed. */ + if (cl_quit == 1 || f_quit == 1) + rte_eth_dev_close(slave_pid); + } +} +#endif + void close_port(portid_t pid) { portid_t pi; struct rte_port *port; +#ifdef RTE_NET_BOND + portid_t slave_pids[RTE_MAX_ETHPORTS]; + int num_slaves = 0; +#endif if (port_id_is_invalid(pid, ENABLED_WARN)) return; @@ -3205,7 +3238,22 @@ close_port(portid_t pid) if (is_proc_primary()) { port_flow_flush(pi); port_flex_item_flush(pi); +#ifdef RTE_NET_BOND + if (port->bond_flag == 1) + num_slaves = rte_eth_bond_slaves_get(pi, + slave_pids, + RTE_MAX_ETHPORTS); +#endif rte_eth_dev_close(pi); +#ifdef RTE_NET_BOND + /* + * If this port is bonded device, all slaves under the + * device need to be removed or closed. + */ + if (port->bond_flag == 1 && num_slaves > 0) + clear_bonding_slave_device(slave_pids, + num_slaves); +#endif } free_xstats_display_info(pi); diff --git a/app/test-pmd/testpmd.h b/app/test-pmd/testpmd.h index 67f253b30e..5af9455012 100644 --- a/app/test-pmd/testpmd.h +++ b/app/test-pmd/testpmd.h @@ -32,6 +32,8 @@ #define RTE_PORT_CLOSED (uint16_t)2 #define RTE_PORT_HANDLING (uint16_t)3 +extern uint8_t cl_quit; + /* * It is used to allocate the memory for hash key. * The hash key size is NIC dependent. -- 2.33.0
On 3/24/2022 3:00 AM, Min Hu (Connor) wrote: > From: Huisong Li <lihuisong@huawei.com> > > When stopping a bonded port, all slaves should be deactivated. But only s/deactivated/stopped/ ? > active slaves are stopped. So fix it and do "deactivae_slave()" for active s/deactivae_slave()/deactivate_slave()/ > slaves. Hi Connor, When a bonding port is closed, is it clear if all slave ports or active slave ports should be stopped? > > Fixes: 0911d4ec0183 ("net/bonding: fix crash when stopping mode 4 port") > Cc: stable@dpdk.org > > Signed-off-by: Huisong Li <lihuisong@huawei.com> > Signed-off-by: Min Hu (Connor) <humin29@huawei.com> > --- > drivers/net/bonding/rte_eth_bond_pmd.c | 20 +++++++++++--------- > 1 file changed, 11 insertions(+), 9 deletions(-) > > diff --git a/drivers/net/bonding/rte_eth_bond_pmd.c b/drivers/net/bonding/rte_eth_bond_pmd.c > index b305b6a35b..469dc71170 100644 > --- a/drivers/net/bonding/rte_eth_bond_pmd.c > +++ b/drivers/net/bonding/rte_eth_bond_pmd.c > @@ -2118,18 +2118,20 @@ bond_ethdev_stop(struct rte_eth_dev *eth_dev) > internals->link_status_polling_enabled = 0; > for (i = 0; i < internals->slave_count; i++) { > uint16_t slave_id = internals->slaves[i].port_id; > + > + internals->slaves[i].last_link_status = 0; > + ret = rte_eth_dev_stop(slave_id); > + if (ret != 0) { > + RTE_BOND_LOG(ERR, "Failed to stop device on port %u", > + slave_id); > + return ret; Should it return here or try to stop all ports? What about to record the return status, but keep continue to stop all ports. And return error if any of the stop failed? > + } > + > + /* active slaves need to deactivate. */ " active slaves need to be deactivated. " ? > if (find_slave_by_id(internals->active_slaves, > internals->active_slave_count, slave_id) != > - internals->active_slave_count) { > - internals->slaves[i].last_link_status = 0; > - ret = rte_eth_dev_stop(slave_id); > - if (ret != 0) { > - RTE_BOND_LOG(ERR, "Failed to stop device on port %u", > - slave_id); > - return ret; > - } > + internals->active_slave_count) I think original indentation for this line is better. > deactivate_slave(eth_dev, slave_id); > - } > } > > return 0;
On 3/24/2022 3:00 AM, Min Hu (Connor) wrote: > From: Huisong Li <lihuisong@huawei.com> > > All slaves will be stopped and removed when closing a bonded port. But the > while loop can not stop if both rte_eth_dev_stop and > rte_eth_bond_slave_remove fail to run. > Agree that this is a defect introduced in below commit. Thanks for the fix. > Fixes: fb0379bc5db3 ("net/bonding: check stop call status") > Cc: stable@dpdk.org > > Signed-off-by: Huisong Li <lihuisong@huawei.com> > Signed-off-by: Min Hu (Connor) <humin29@huawei.com> > --- > drivers/net/bonding/rte_eth_bond_pmd.c | 3 ++- > 1 file changed, 2 insertions(+), 1 deletion(-) > > diff --git a/drivers/net/bonding/rte_eth_bond_pmd.c b/drivers/net/bonding/rte_eth_bond_pmd.c > index 469dc71170..00d4deda44 100644 > --- a/drivers/net/bonding/rte_eth_bond_pmd.c > +++ b/drivers/net/bonding/rte_eth_bond_pmd.c > @@ -2149,13 +2149,14 @@ bond_ethdev_close(struct rte_eth_dev *dev) > return 0; > > RTE_BOND_LOG(INFO, "Closing bonded device %s", dev->device->name); > - while (internals->slave_count != skipped) { > + while (skipped < internals->slave_count) { When below fixed with adding 'continue', no need to change the check, right? Although new one is also correct. > uint16_t port_id = internals->slaves[skipped].port_id; > > if (rte_eth_dev_stop(port_id) != 0) { > RTE_BOND_LOG(ERR, "Failed to stop device on port %u", > port_id); > skipped++; > + continue; Can't we remove the slave even if 'stop()' failed? If so I think better to just log the error and keep continue in that case, what do you think? > } > > if (rte_eth_bond_slave_remove(bond_port_id, port_id) != 0) {
Hi, Ferruh, 在 2022/4/27 2:19, Ferruh Yigit 写道: > On 3/24/2022 3:00 AM, Min Hu (Connor) wrote: >> From: Huisong Li <lihuisong@huawei.com> >> >> When stopping a bonded port, all slaves should be deactivated. But only > > s/deactivated/stopped/ ? not agreed. deactivated and stopped are different state for slave. > >> active slaves are stopped. So fix it and do "deactivae_slave()" for >> active > > s/deactivae_slave()/deactivate_slave()/ > agreed. >> slaves. > > Hi Connor, > > When a bonding port is closed, is it clear if all slave ports or active > slave ports should be stopped? Yes, I think all the slave ports should be stopped(or try to be stopped). > >> >> Fixes: 0911d4ec0183 ("net/bonding: fix crash when stopping mode 4 port") >> Cc: stable@dpdk.org >> >> Signed-off-by: Huisong Li <lihuisong@huawei.com> >> Signed-off-by: Min Hu (Connor) <humin29@huawei.com> >> --- >> drivers/net/bonding/rte_eth_bond_pmd.c | 20 +++++++++++--------- >> 1 file changed, 11 insertions(+), 9 deletions(-) >> >> diff --git a/drivers/net/bonding/rte_eth_bond_pmd.c >> b/drivers/net/bonding/rte_eth_bond_pmd.c >> index b305b6a35b..469dc71170 100644 >> --- a/drivers/net/bonding/rte_eth_bond_pmd.c >> +++ b/drivers/net/bonding/rte_eth_bond_pmd.c >> @@ -2118,18 +2118,20 @@ bond_ethdev_stop(struct rte_eth_dev *eth_dev) >> internals->link_status_polling_enabled = 0; >> for (i = 0; i < internals->slave_count; i++) { >> uint16_t slave_id = internals->slaves[i].port_id; >> + >> + internals->slaves[i].last_link_status = 0; >> + ret = rte_eth_dev_stop(slave_id); >> + if (ret != 0) { >> + RTE_BOND_LOG(ERR, "Failed to stop device on port %u", >> + slave_id); >> + return ret; > > Should it return here or try to stop all ports? > What about to record the return status, but keep continue to stop all > ports. And return error if any of the stop failed? I think no need to do this. APP only see the bonded device. If bonded device stop failed, it means it works failed. And the number of "stopped" successfully slave does not make any sense. > >> + } >> + >> + /* active slaves need to deactivate. */ > > " active slaves need to be deactivated. " ? agreed. > >> if (find_slave_by_id(internals->active_slaves, >> internals->active_slave_count, slave_id) != >> - internals->active_slave_count) { >> - internals->slaves[i].last_link_status = 0; >> - ret = rte_eth_dev_stop(slave_id); >> - if (ret != 0) { >> - RTE_BOND_LOG(ERR, "Failed to stop device on port %u", >> - slave_id); >> - return ret; >> - } >> + internals->active_slave_count) > > I think original indentation for this line is better. > agreed. >> deactivate_slave(eth_dev, slave_id); >> - } >> } >> return 0; > > .
Hi, Ferruh, 在 2022/4/27 2:19, Ferruh Yigit 写道: > On 3/24/2022 3:00 AM, Min Hu (Connor) wrote: >> From: Huisong Li <lihuisong@huawei.com> >> >> All slaves will be stopped and removed when closing a bonded port. But >> the >> while loop can not stop if both rte_eth_dev_stop and >> rte_eth_bond_slave_remove fail to run. >> > > Agree that this is a defect introduced in below commit. Thanks for the fix. thanks. > >> Fixes: fb0379bc5db3 ("net/bonding: check stop call status") >> Cc: stable@dpdk.org >> >> Signed-off-by: Huisong Li <lihuisong@huawei.com> >> Signed-off-by: Min Hu (Connor) <humin29@huawei.com> >> --- >> drivers/net/bonding/rte_eth_bond_pmd.c | 3 ++- >> 1 file changed, 2 insertions(+), 1 deletion(-) >> >> diff --git a/drivers/net/bonding/rte_eth_bond_pmd.c >> b/drivers/net/bonding/rte_eth_bond_pmd.c >> index 469dc71170..00d4deda44 100644 >> --- a/drivers/net/bonding/rte_eth_bond_pmd.c >> +++ b/drivers/net/bonding/rte_eth_bond_pmd.c >> @@ -2149,13 +2149,14 @@ bond_ethdev_close(struct rte_eth_dev *dev) >> return 0; >> RTE_BOND_LOG(INFO, "Closing bonded device %s", dev->device->name); >> - while (internals->slave_count != skipped) { >> + while (skipped < internals->slave_count) { > > When below fixed with adding 'continue', no need to change the check, > right? Although new one is also correct. Agreed. > >> uint16_t port_id = internals->slaves[skipped].port_id; >> if (rte_eth_dev_stop(port_id) != 0) { >> RTE_BOND_LOG(ERR, "Failed to stop device on port %u", >> port_id); >> skipped++; >> + continue; > > Can't we remove the slave even if 'stop()' failed? If so I think better > to just log the error and keep continue in that case, what do you think? NO, if slave stop failed, we cannot remove the slave. just see the function stack: rte_eth_bond_slave_remove __eth_bond_slave_remove_lock_free slave_remove rte_eth_dev_internal_reset if (dev->data->dev_started) { RTE_ETHDEV_LOG(ERR, "Port %u must be stopped to allow reset\n", dev->data->port_id); return; } > >> } >> if (rte_eth_bond_slave_remove(bond_port_id, port_id) != 0) { > > .
On 4/29/2022 7:45 AM, Min Hu (Connor) wrote: > Hi, Ferruh, > > 在 2022/4/27 2:19, Ferruh Yigit 写道: >> On 3/24/2022 3:00 AM, Min Hu (Connor) wrote: >>> From: Huisong Li <lihuisong@huawei.com> >>> >>> When stopping a bonded port, all slaves should be deactivated. But only >> >> s/deactivated/stopped/ ? > not agreed. deactivated and stopped are different state for slave. > Just to clarify the sentences, otherwise I see the 'stopped' and 'deactivated' states are different. Next sentences complains that not all ports are stopped: "But only active slaves are stopped.", so I thought intention in this sentences to claim that all slaves should be stopped (but it mentions all slaves should be 'deactivated'). As long as you address the disconnection between two sentences, I don't mind the wording. >> >>> active slaves are stopped. So fix it and do "deactivae_slave()" for >>> active >> >> s/deactivae_slave()/deactivate_slave()/ >> > agreed. > >>> slaves. >> >> Hi Connor, >> >> When a bonding port is closed, is it clear if all slave ports or >> active slave ports should be stopped? > Yes, I think all the slave ports should be stopped(or try to be stopped). >> >>> >>> Fixes: 0911d4ec0183 ("net/bonding: fix crash when stopping mode 4 port") >>> Cc: stable@dpdk.org >>> >>> Signed-off-by: Huisong Li <lihuisong@huawei.com> >>> Signed-off-by: Min Hu (Connor) <humin29@huawei.com> >>> --- >>> drivers/net/bonding/rte_eth_bond_pmd.c | 20 +++++++++++--------- >>> 1 file changed, 11 insertions(+), 9 deletions(-) >>> >>> diff --git a/drivers/net/bonding/rte_eth_bond_pmd.c >>> b/drivers/net/bonding/rte_eth_bond_pmd.c >>> index b305b6a35b..469dc71170 100644 >>> --- a/drivers/net/bonding/rte_eth_bond_pmd.c >>> +++ b/drivers/net/bonding/rte_eth_bond_pmd.c >>> @@ -2118,18 +2118,20 @@ bond_ethdev_stop(struct rte_eth_dev *eth_dev) >>> internals->link_status_polling_enabled = 0; >>> for (i = 0; i < internals->slave_count; i++) { >>> uint16_t slave_id = internals->slaves[i].port_id; >>> + >>> + internals->slaves[i].last_link_status = 0; >>> + ret = rte_eth_dev_stop(slave_id); >>> + if (ret != 0) { >>> + RTE_BOND_LOG(ERR, "Failed to stop device on port %u", >>> + slave_id); >>> + return ret; >> >> Should it return here or try to stop all ports? >> What about to record the return status, but keep continue to stop all >> ports. And return error if any of the stop failed? > I think no need to do this. APP only see the bonded device. If bonded > device stop failed, it means it works failed. And the number of > "stopped" successfully slave does not make any sense. > OK if trying to stop as much as possible 'slave' devices doesn't make sense, we can keep as it is. Btw, when functions fails at this point, bonding device itself already marked as stopped, right? And some of the slave devices may be stopped already before failure. I don't know how confusing this is for the user, that stop() function is failed but bonding device state is 'stopped'. I don't know if function should recover at least bonding device status (back to started) on failure, what do you think? >> >>> + } >>> + >>> + /* active slaves need to deactivate. */ >> >> " active slaves need to be deactivated. " ? > agreed. >> >>> if (find_slave_by_id(internals->active_slaves, >>> internals->active_slave_count, slave_id) != >>> - internals->active_slave_count) { >>> - internals->slaves[i].last_link_status = 0; >>> - ret = rte_eth_dev_stop(slave_id); >>> - if (ret != 0) { >>> - RTE_BOND_LOG(ERR, "Failed to stop device on port %u", >>> - slave_id); >>> - return ret; >>> - } >>> + internals->active_slave_count) >> >> I think original indentation for this line is better. >> > agreed. >>> deactivate_slave(eth_dev, slave_id); >>> - } >>> } >>> return 0; >> >> .
On 4/29/2022 7:52 AM, Min Hu (Connor) wrote: > Hi, Ferruh, > > 在 2022/4/27 2:19, Ferruh Yigit 写道: >> On 3/24/2022 3:00 AM, Min Hu (Connor) wrote: >>> From: Huisong Li <lihuisong@huawei.com> >>> >>> All slaves will be stopped and removed when closing a bonded port. >>> But the >>> while loop can not stop if both rte_eth_dev_stop and >>> rte_eth_bond_slave_remove fail to run. >>> >> >> Agree that this is a defect introduced in below commit. Thanks for the >> fix. > thanks. >> >>> Fixes: fb0379bc5db3 ("net/bonding: check stop call status") >>> Cc: stable@dpdk.org >>> >>> Signed-off-by: Huisong Li <lihuisong@huawei.com> >>> Signed-off-by: Min Hu (Connor) <humin29@huawei.com> >>> --- >>> drivers/net/bonding/rte_eth_bond_pmd.c | 3 ++- >>> 1 file changed, 2 insertions(+), 1 deletion(-) >>> >>> diff --git a/drivers/net/bonding/rte_eth_bond_pmd.c >>> b/drivers/net/bonding/rte_eth_bond_pmd.c >>> index 469dc71170..00d4deda44 100644 >>> --- a/drivers/net/bonding/rte_eth_bond_pmd.c >>> +++ b/drivers/net/bonding/rte_eth_bond_pmd.c >>> @@ -2149,13 +2149,14 @@ bond_ethdev_close(struct rte_eth_dev *dev) >>> return 0; >>> RTE_BOND_LOG(INFO, "Closing bonded device %s", dev->device->name); >>> - while (internals->slave_count != skipped) { >>> + while (skipped < internals->slave_count) { >> >> When below fixed with adding 'continue', no need to change the check, >> right? Although new one is also correct. > Agreed. >> >>> uint16_t port_id = internals->slaves[skipped].port_id; >>> if (rte_eth_dev_stop(port_id) != 0) { >>> RTE_BOND_LOG(ERR, "Failed to stop device on port %u", >>> port_id); >>> skipped++; >>> + continue; >> >> Can't we remove the slave even if 'stop()' failed? If so I think >> better to just log the error and keep continue in that case, what do >> you think? > NO, if slave stop failed, we cannot remove the slave. Got it, thanks for clarification. > just see the function stack: > rte_eth_bond_slave_remove > __eth_bond_slave_remove_lock_free > slave_remove > rte_eth_dev_internal_reset > if (dev->data->dev_started) { > RTE_ETHDEV_LOG(ERR, "Port %u must be stopped to allow reset\n", > dev->data->port_id); > return; > } > >> >>> } >>> if (rte_eth_bond_slave_remove(bond_port_id, port_id) != 0) { >> >> .
Hi, Ferruh, 在 2022/4/29 21:31, Ferruh Yigit 写道: > On 4/29/2022 7:45 AM, Min Hu (Connor) wrote: >> Hi, Ferruh, >> >> 在 2022/4/27 2:19, Ferruh Yigit 写道: >>> On 3/24/2022 3:00 AM, Min Hu (Connor) wrote: >>>> From: Huisong Li <lihuisong@huawei.com> >>>> >>>> When stopping a bonded port, all slaves should be deactivated. But only >>> >>> s/deactivated/stopped/ ? >> not agreed. deactivated and stopped are different state for slave. >> > > Just to clarify the sentences, otherwise I see the 'stopped' and > 'deactivated' states are different. > Next sentences complains that not all ports are stopped: "But only > active slaves are stopped.", so I thought intention in this sentences to > claim that all slaves should be stopped (but it mentions all slaves > should be 'deactivated'). > As long as you address the disconnection between two sentences, I don't > mind the wording. Actually, there is something wrong with the wording. Yes, I should take your advice. > >>> >>>> active slaves are stopped. So fix it and do "deactivae_slave()" for >>>> active >>> >>> s/deactivae_slave()/deactivate_slave()/ >>> >> agreed. >> >>>> slaves. >>> >>> Hi Connor, >>> >>> When a bonding port is closed, is it clear if all slave ports or >>> active slave ports should be stopped? >> Yes, I think all the slave ports should be stopped(or try to be stopped). >>> >>>> >>>> Fixes: 0911d4ec0183 ("net/bonding: fix crash when stopping mode 4 >>>> port") >>>> Cc: stable@dpdk.org >>>> >>>> Signed-off-by: Huisong Li <lihuisong@huawei.com> >>>> Signed-off-by: Min Hu (Connor) <humin29@huawei.com> >>>> --- >>>> drivers/net/bonding/rte_eth_bond_pmd.c | 20 +++++++++++--------- >>>> 1 file changed, 11 insertions(+), 9 deletions(-) >>>> >>>> diff --git a/drivers/net/bonding/rte_eth_bond_pmd.c >>>> b/drivers/net/bonding/rte_eth_bond_pmd.c >>>> index b305b6a35b..469dc71170 100644 >>>> --- a/drivers/net/bonding/rte_eth_bond_pmd.c >>>> +++ b/drivers/net/bonding/rte_eth_bond_pmd.c >>>> @@ -2118,18 +2118,20 @@ bond_ethdev_stop(struct rte_eth_dev *eth_dev) >>>> internals->link_status_polling_enabled = 0; >>>> for (i = 0; i < internals->slave_count; i++) { >>>> uint16_t slave_id = internals->slaves[i].port_id; >>>> + >>>> + internals->slaves[i].last_link_status = 0; >>>> + ret = rte_eth_dev_stop(slave_id); >>>> + if (ret != 0) { >>>> + RTE_BOND_LOG(ERR, "Failed to stop device on port %u", >>>> + slave_id); >>>> + return ret; >>> >>> Should it return here or try to stop all ports? >>> What about to record the return status, but keep continue to stop all >>> ports. And return error if any of the stop failed? Well, I am glad you have found something unreasaonable about 'stop'. Let us see API 'rte_eth_dev_stop' rte_eth_dev_stop(dev) { .... dev->data->dev_started = 0; ret = (*dev->dev_ops->dev_stop)(dev) retur ret; } This is unreasaonable. No matter 'dev_ops->dev_stop' succeed or fail, the state 'dev_started ' will always set to be '0'. But this does not only influence bonding device but other devices like eth dev or vdev. This is the bug in rte ethdev level. I will send another patch to fix it. >> I think no need to do this. APP only see the bonded device. If bonded >> device stop failed, it means it works failed. And the number of >> "stopped" successfully slave does not make any sense. >> > > OK if trying to stop as much as possible 'slave' devices doesn't make > sense, we can keep as it is. > > Btw, when functions fails at this point, bonding device itself already > marked as stopped, right? And some of the slave devices may be stopped > already before failure. > I don't know how confusing this is for the user, that stop() function is > failed but bonding device state is 'stopped'. I don't know if function > should recover at least bonding device status (back to started) on > failure, what do you think? > >>> >>>> + } >>>> + >>>> + /* active slaves need to deactivate. */ >>> >>> " active slaves need to be deactivated. " ? >> agreed. >>> >>>> if (find_slave_by_id(internals->active_slaves, >>>> internals->active_slave_count, slave_id) != >>>> - internals->active_slave_count) { >>>> - internals->slaves[i].last_link_status = 0; >>>> - ret = rte_eth_dev_stop(slave_id); >>>> - if (ret != 0) { >>>> - RTE_BOND_LOG(ERR, "Failed to stop device on port %u", >>>> - slave_id); >>>> - return ret; >>>> - } >>>> + internals->active_slave_count) >>> >>> I think original indentation for this line is better. >>> >> agreed. >>>> deactivate_slave(eth_dev, slave_id); >>>> - } >>>> } >>>> return 0; >>> >>> . > > .
From: Huisong Li <lihuisong@huawei.com> When stopping a bonded port, all slaves should be stopped. But only active slaves are stopped. So fix it and do "deactivate_slave()" for active slaves. Fixes: 0911d4ec0183 ("net/bonding: fix crash when stopping mode 4 port") Cc: stable@dpdk.org Signed-off-by: Huisong Li <lihuisong@huawei.com> Signed-off-by: Min Hu (Connor) <humin29@huawei.com> --- drivers/net/bonding/rte_eth_bond_pmd.c | 20 +++++++++++--------- 1 file changed, 11 insertions(+), 9 deletions(-) diff --git a/drivers/net/bonding/rte_eth_bond_pmd.c b/drivers/net/bonding/rte_eth_bond_pmd.c index b305b6a35b..92caf24f4b 100644 --- a/drivers/net/bonding/rte_eth_bond_pmd.c +++ b/drivers/net/bonding/rte_eth_bond_pmd.c @@ -2118,18 +2118,20 @@ bond_ethdev_stop(struct rte_eth_dev *eth_dev) internals->link_status_polling_enabled = 0; for (i = 0; i < internals->slave_count; i++) { uint16_t slave_id = internals->slaves[i].port_id; + + internals->slaves[i].last_link_status = 0; + ret = rte_eth_dev_stop(slave_id); + if (ret != 0) { + RTE_BOND_LOG(ERR, "Failed to stop device on port %u", + slave_id); + return ret; + } + + /* active slaves need to be deactivated. */ if (find_slave_by_id(internals->active_slaves, internals->active_slave_count, slave_id) != - internals->active_slave_count) { - internals->slaves[i].last_link_status = 0; - ret = rte_eth_dev_stop(slave_id); - if (ret != 0) { - RTE_BOND_LOG(ERR, "Failed to stop device on port %u", - slave_id); - return ret; - } + internals->active_slave_count) deactivate_slave(eth_dev, slave_id); - } } return 0; -- 2.33.0
From: Huisong Li <lihuisong@huawei.com> All slaves will be stopped and removed when closing a bonded port. But the while loop can not stop if both rte_eth_dev_stop and rte_eth_bond_slave_remove fail to run. Fixes: fb0379bc5db3 ("net/bonding: check stop call status") Cc: stable@dpdk.org Signed-off-by: Huisong Li <lihuisong@huawei.com> Signed-off-by: Min Hu (Connor) <humin29@huawei.com> --- drivers/net/bonding/rte_eth_bond_pmd.c | 1 + 1 file changed, 1 insertion(+) diff --git a/drivers/net/bonding/rte_eth_bond_pmd.c b/drivers/net/bonding/rte_eth_bond_pmd.c index 92caf24f4b..8f48ba7d23 100644 --- a/drivers/net/bonding/rte_eth_bond_pmd.c +++ b/drivers/net/bonding/rte_eth_bond_pmd.c @@ -2156,6 +2156,7 @@ bond_ethdev_close(struct rte_eth_dev *dev) RTE_BOND_LOG(ERR, "Failed to stop device on port %u", port_id); skipped++; + continue; } if (rte_eth_bond_slave_remove(bond_port_id, port_id) != 0) { -- 2.33.0
From: Huisong Li <lihuisong@huawei.com> Starting or stopping a bonded port also starts or stops all active slaves under the bonded port. If this port is a bonded device, we need to modify the port status of all slaves. Fixes: 0e545d3047fe ("app/testpmd: check stopping port is not in bonding") Cc: stable@dpdk.org Signed-off-by: Huisong Li <lihuisong@huawei.com> Signed-off-by: Min Hu (Connor) <humin29@huawei.com> Acked-by: Aman Singh <aman.deep.singh@intel.com> --- app/test-pmd/cmdline.c | 1 + app/test-pmd/testpmd.c | 74 +++++++++++++++++++++++++++++++++++++++--- app/test-pmd/testpmd.h | 3 +- 3 files changed, 73 insertions(+), 5 deletions(-) diff --git a/app/test-pmd/cmdline.c b/app/test-pmd/cmdline.c index 6ffea8e21a..d9fc7a88bd 100644 --- a/app/test-pmd/cmdline.c +++ b/app/test-pmd/cmdline.c @@ -6671,6 +6671,7 @@ static void cmd_create_bonded_device_parsed(void *parsed_result, "Failed to enable promiscuous mode for port %u: %s - ignore\n", port_id, rte_strerror(-ret)); + ports[port_id].bond_flag = 1; ports[port_id].need_setup = 0; ports[port_id].port_status = RTE_PORT_STOPPED; } diff --git a/app/test-pmd/testpmd.c b/app/test-pmd/testpmd.c index fe2ce19f99..dc90600787 100644 --- a/app/test-pmd/testpmd.c +++ b/app/test-pmd/testpmd.c @@ -66,6 +66,9 @@ #ifdef RTE_EXEC_ENV_WINDOWS #include <process.h> #endif +#ifdef RTE_NET_BOND +#include <rte_eth_bond.h> +#endif #include "testpmd.h" @@ -597,11 +600,57 @@ eth_dev_configure_mp(uint16_t port_id, uint16_t nb_rx_q, uint16_t nb_tx_q, return 0; } +#ifdef RTE_NET_BOND +static int +change_bonding_slave_port_status(portid_t bond_pid, bool is_stop) +{ + portid_t slave_pids[RTE_MAX_ETHPORTS]; + struct rte_port *port; + int num_slaves; + portid_t slave_pid; + int i; + + num_slaves = rte_eth_bond_slaves_get(bond_pid, slave_pids, + RTE_MAX_ETHPORTS); + if (num_slaves < 0) { + fprintf(stderr, "Failed to get slave list for port = %u\n", + bond_pid); + return num_slaves; + } + + for (i = 0; i < num_slaves; i++) { + slave_pid = slave_pids[i]; + port = &ports[slave_pid]; + port->port_status = + is_stop ? RTE_PORT_STOPPED : RTE_PORT_STARTED; + } + + return 0; +} +#endif + static int eth_dev_start_mp(uint16_t port_id) { - if (is_proc_primary()) - return rte_eth_dev_start(port_id); + int ret; + + if (is_proc_primary()) { + ret = rte_eth_dev_start(port_id); + if (ret != 0) + return ret; + +#ifdef RTE_NET_BOND + struct rte_port *port = &ports[port_id]; + + /* + * Starting a bonded port also starts all slaves under the bonded + * device. So if this port is bond device, we need to modify the + * port status of these slaves. + */ + if (port->bond_flag == 1) + return change_bonding_slave_port_status(port_id, false); +#endif + } return 0; } @@ -609,8 +658,25 @@ eth_dev_start_mp(uint16_t port_id) static int eth_dev_stop_mp(uint16_t port_id) { - if (is_proc_primary()) - return rte_eth_dev_stop(port_id); + int ret; + + if (is_proc_primary()) { + ret = rte_eth_dev_stop(port_id); + if (ret != 0) + return ret; + +#ifdef RTE_NET_BOND + struct rte_port *port = &ports[port_id]; + + /* + * Stopping a bonded port also stops all slaves under the bonded + * device. So if this port is bond device, we need to modify the + * port status of these slaves. + */ + if (port->bond_flag == 1) + return change_bonding_slave_port_status(port_id, true); +#endif + } return 0; } diff --git a/app/test-pmd/testpmd.h b/app/test-pmd/testpmd.h index 31f766c965..67f253b30e 100644 --- a/app/test-pmd/testpmd.h +++ b/app/test-pmd/testpmd.h @@ -266,7 +266,8 @@ struct rte_port { uint32_t mc_addr_nb; /**< nb. of addr. in mc_addr_pool */ queueid_t queue_nb; /**< nb. of queues for flow rules */ uint32_t queue_sz; /**< size of a queue for flow rules */ - uint8_t slave_flag; /**< bonding slave port */ + uint8_t slave_flag : 1, /**< bonding slave port */ + bond_flag : 1; /**< port is bond device */ struct port_template *pattern_templ_list; /**< Pattern templates. */ struct port_template *actions_templ_list; /**< Actions templates. */ struct port_table *table_list; /**< Flow tables. */ -- 2.33.0
From: Huisong Li <lihuisong@huawei.com> Currently, some eth devices are added to bond device, these devices are not released when the quit command is executed in testpmd. This patch adds the release operation for all active slaves under a bond device. Fixes: 0e545d3047fe ("app/testpmd: check stopping port is not in bonding") Cc: stable@dpdk.org Signed-off-by: Huisong Li <lihuisong@huawei.com> Signed-off-by: Min Hu (Connor) <humin29@huawei.com> --- app/test-pmd/cmdline.c | 1 + app/test-pmd/testpmd.c | 48 ++++++++++++++++++++++++++++++++++++++++++ app/test-pmd/testpmd.h | 2 ++ 3 files changed, 51 insertions(+) diff --git a/app/test-pmd/cmdline.c b/app/test-pmd/cmdline.c index d9fc7a88bd..f3c1d9ab79 100644 --- a/app/test-pmd/cmdline.c +++ b/app/test-pmd/cmdline.c @@ -8891,6 +8891,7 @@ static void cmd_quit_parsed(__rte_unused void *parsed_result, __rte_unused void *data) { cmdline_quit(cl); + cl_quit = 1; } cmdline_parse_token_string_t cmd_quit_quit = diff --git a/app/test-pmd/testpmd.c b/app/test-pmd/testpmd.c index dc90600787..22700e073d 100644 --- a/app/test-pmd/testpmd.c +++ b/app/test-pmd/testpmd.c @@ -229,6 +229,7 @@ unsigned int xstats_display_num; /**< Size of extended statistics to show */ * option. Set flag to exit stats period loop after received SIGINT/SIGTERM. */ uint8_t f_quit; +uint8_t cl_quit; /* Quit testpmd from cmdline. */ /* * Max Rx frame size, set by '--max-pkt-len' parameter. @@ -3167,11 +3168,43 @@ remove_invalid_ports(void) nb_cfg_ports = nb_fwd_ports; } +#ifdef RTE_NET_BOND +static void +clear_bonding_slave_device(portid_t *slave_pids, uint16_t num_slaves) +{ + struct rte_port *port; + portid_t slave_pid; + uint16_t i; + + for (i = 0; i < num_slaves; i++) { + slave_pid = slave_pids[i]; + if (port_is_started(slave_pid) == 1) { + if (rte_eth_dev_stop(slave_pid) != 0) + fprintf(stderr, "rte_eth_dev_stop failed for port %u\n", + slave_pid); + + port = &ports[slave_pid]; + port->port_status = RTE_PORT_STOPPED; + } + + clear_port_slave_flag(slave_pid); + + /* Close slave device when testpmd quit or is killed. */ + if (cl_quit == 1 || f_quit == 1) + rte_eth_dev_close(slave_pid); + } +} +#endif + void close_port(portid_t pid) { portid_t pi; struct rte_port *port; +#ifdef RTE_NET_BOND + portid_t slave_pids[RTE_MAX_ETHPORTS]; + int num_slaves = 0; +#endif if (port_id_is_invalid(pid, ENABLED_WARN)) return; @@ -3205,7 +3238,22 @@ close_port(portid_t pid) if (is_proc_primary()) { port_flow_flush(pi); port_flex_item_flush(pi); +#ifdef RTE_NET_BOND + if (port->bond_flag == 1) + num_slaves = rte_eth_bond_slaves_get(pi, + slave_pids, + RTE_MAX_ETHPORTS); +#endif rte_eth_dev_close(pi); +#ifdef RTE_NET_BOND + /* + * If this port is bonded device, all slaves under the + * device need to be removed or closed. + */ + if (port->bond_flag == 1 && num_slaves > 0) + clear_bonding_slave_device(slave_pids, + num_slaves); +#endif } free_xstats_display_info(pi); diff --git a/app/test-pmd/testpmd.h b/app/test-pmd/testpmd.h index 67f253b30e..5af9455012 100644 --- a/app/test-pmd/testpmd.h +++ b/app/test-pmd/testpmd.h @@ -32,6 +32,8 @@ #define RTE_PORT_CLOSED (uint16_t)2 #define RTE_PORT_HANDLING (uint16_t)3 +extern uint8_t cl_quit; + /* * It is used to allocate the memory for hash key. * The hash key size is NIC dependent. -- 2.33.0
Currently, 'dev_started' is always set to be 0 when dev stop, whether it succeeded or failed. This is unreasonable and this patch fixed it. Fixes: 62024eb82756 ("ethdev: change stop operation callback to return int") Cc: stable@dpdk.org Signed-off-by: Min Hu (Connor) <humin29@huawei.com> --- lib/ethdev/rte_ethdev.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/lib/ethdev/rte_ethdev.c b/lib/ethdev/rte_ethdev.c index 29a3d80466..e0011372aa 100644 --- a/lib/ethdev/rte_ethdev.c +++ b/lib/ethdev/rte_ethdev.c @@ -1533,8 +1533,9 @@ rte_eth_dev_stop(uint16_t port_id) /* point fast-path functions to dummy ones */ eth_dev_fp_ops_reset(rte_eth_fp_ops + port_id); - dev->data->dev_started = 0; ret = (*dev->dev_ops->dev_stop)(dev); + if (ret == 0) + dev->data->dev_started = 0; rte_ethdev_trace_stop(port_id, ret); return ret; -- 2.33.0
On 5/3/2022 7:54 AM, Min Hu (Connor) wrote: > Hi, Ferruh, > > 在 2022/4/29 21:31, Ferruh Yigit 写道: >> On 4/29/2022 7:45 AM, Min Hu (Connor) wrote: >>> Hi, Ferruh, >>> >>> 在 2022/4/27 2:19, Ferruh Yigit 写道: >>>> On 3/24/2022 3:00 AM, Min Hu (Connor) wrote: >>>>> From: Huisong Li <lihuisong@huawei.com> >>>>> >>>>> When stopping a bonded port, all slaves should be deactivated. But >>>>> only >>>> >>>> s/deactivated/stopped/ ? >>> not agreed. deactivated and stopped are different state for slave. >>> >> >> Just to clarify the sentences, otherwise I see the 'stopped' and >> 'deactivated' states are different. >> Next sentences complains that not all ports are stopped: "But only >> active slaves are stopped.", so I thought intention in this sentences >> to claim that all slaves should be stopped (but it mentions all slaves >> should be 'deactivated'). >> As long as you address the disconnection between two sentences, I >> don't mind the wording. > Actually, there is something wrong with the wording. > Yes, I should take your advice. > >> >>>> >>>>> active slaves are stopped. So fix it and do "deactivae_slave()" for >>>>> active >>>> >>>> s/deactivae_slave()/deactivate_slave()/ >>>> >>> agreed. >>> >>>>> slaves. >>>> >>>> Hi Connor, >>>> >>>> When a bonding port is closed, is it clear if all slave ports or >>>> active slave ports should be stopped? >>> Yes, I think all the slave ports should be stopped(or try to be >>> stopped). >>>> >>>>> >>>>> Fixes: 0911d4ec0183 ("net/bonding: fix crash when stopping mode 4 >>>>> port") >>>>> Cc: stable@dpdk.org >>>>> >>>>> Signed-off-by: Huisong Li <lihuisong@huawei.com> >>>>> Signed-off-by: Min Hu (Connor) <humin29@huawei.com> >>>>> --- >>>>> drivers/net/bonding/rte_eth_bond_pmd.c | 20 +++++++++++--------- >>>>> 1 file changed, 11 insertions(+), 9 deletions(-) >>>>> >>>>> diff --git a/drivers/net/bonding/rte_eth_bond_pmd.c >>>>> b/drivers/net/bonding/rte_eth_bond_pmd.c >>>>> index b305b6a35b..469dc71170 100644 >>>>> --- a/drivers/net/bonding/rte_eth_bond_pmd.c >>>>> +++ b/drivers/net/bonding/rte_eth_bond_pmd.c >>>>> @@ -2118,18 +2118,20 @@ bond_ethdev_stop(struct rte_eth_dev *eth_dev) >>>>> internals->link_status_polling_enabled = 0; >>>>> for (i = 0; i < internals->slave_count; i++) { >>>>> uint16_t slave_id = internals->slaves[i].port_id; >>>>> + >>>>> + internals->slaves[i].last_link_status = 0; >>>>> + ret = rte_eth_dev_stop(slave_id); >>>>> + if (ret != 0) { >>>>> + RTE_BOND_LOG(ERR, "Failed to stop device on port %u", >>>>> + slave_id); >>>>> + return ret; >>>> >>>> Should it return here or try to stop all ports? >>>> What about to record the return status, but keep continue to stop >>>> all ports. And return error if any of the stop failed? > Well, I am glad you have found something unreasaonable about 'stop'. > Let us see API 'rte_eth_dev_stop' > > rte_eth_dev_stop(dev) > { > .... > dev->data->dev_started = 0; > ret = (*dev->dev_ops->dev_stop)(dev) > retur ret; > } > This is unreasaonable. No matter 'dev_ops->dev_stop' succeed or fail, > the state 'dev_started ' will always set to be '0'. > > But this does not only influence bonding device but other devices like > eth dev or vdev. > This is the bug in rte ethdev level. I will send another patch to fix > it. > Hi Connor, I agree this is an issue in the API, cc'ed Andrew and Thomas. I vaguely remember that "dev_started = 0" was required for some dev_ops, but not quite sure, let me check this. At worst we can do as following to be sure: dev->data->dev_started = 0; ret = (*dev->dev_ops->dev_stop)(dev) if (ret) dev->data->dev_started = 1; Also we need to clarify in the API documentation (.h file), what is the status of the device if 'rte_eth_dev_stop()' returned error. Btw, would you be OK to separate this ethdev patch from your bonding patch, to not stuck your series because of ethdev one. > >>> I think no need to do this. APP only see the bonded device. If bonded >>> device stop failed, it means it works failed. And the number of >>> "stopped" successfully slave does not make any sense. >>> >> >> OK if trying to stop as much as possible 'slave' devices doesn't make >> sense, we can keep as it is. >> >> Btw, when functions fails at this point, bonding device itself already >> marked as stopped, right? And some of the slave devices may be stopped >> already before failure. >> I don't know how confusing this is for the user, that stop() function >> is failed but bonding device state is 'stopped'. I don't know if >> function should recover at least bonding device status (back to >> started) on failure, what do you think? >> >>>> >>>>> + } >>>>> + >>>>> + /* active slaves need to deactivate. */ >>>> >>>> " active slaves need to be deactivated. " ? >>> agreed. >>>> >>>>> if (find_slave_by_id(internals->active_slaves, >>>>> internals->active_slave_count, slave_id) != >>>>> - internals->active_slave_count) { >>>>> - internals->slaves[i].last_link_status = 0; >>>>> - ret = rte_eth_dev_stop(slave_id); >>>>> - if (ret != 0) { >>>>> - RTE_BOND_LOG(ERR, "Failed to stop device on port %u", >>>>> - slave_id); >>>>> - return ret; >>>>> - } >>>>> + internals->active_slave_count) >>>> >>>> I think original indentation for this line is better. >>>> >>> agreed. >>>>> deactivate_slave(eth_dev, slave_id); >>>>> - } >>>>> } >>>>> return 0; >>>> >>>> . >> >> .
03/05/2022 11:02, Min Hu (Connor) пишет: > From: Huisong Li <lihuisong@huawei.com> > > Starting or stopping a bonded port also starts or stops all active slaves > under the bonded port. If this port is a bonded device, we need to modify > the port status of all slaves. > > Fixes: 0e545d3047fe ("app/testpmd: check stopping port is not in bonding") > Cc: stable@dpdk.org > > Signed-off-by: Huisong Li <lihuisong@huawei.com> > Signed-off-by: Min Hu (Connor) <humin29@huawei.com> > Acked-by: Aman Singh <aman.deep.singh@intel.com> > --- > app/test-pmd/cmdline.c | 1 + > app/test-pmd/testpmd.c | 74 +++++++++++++++++++++++++++++++++++++++--- > app/test-pmd/testpmd.h | 3 +- > 3 files changed, 73 insertions(+), 5 deletions(-) > > diff --git a/app/test-pmd/cmdline.c b/app/test-pmd/cmdline.c > index 6ffea8e21a..d9fc7a88bd 100644 > --- a/app/test-pmd/cmdline.c > +++ b/app/test-pmd/cmdline.c > @@ -6671,6 +6671,7 @@ static void cmd_create_bonded_device_parsed(void *parsed_result, > "Failed to enable promiscuous mode for port %u: %s - ignore\n", > port_id, rte_strerror(-ret)); > > + ports[port_id].bond_flag = 1; > ports[port_id].need_setup = 0; > ports[port_id].port_status = RTE_PORT_STOPPED; > } > diff --git a/app/test-pmd/testpmd.c b/app/test-pmd/testpmd.c > index fe2ce19f99..dc90600787 100644 > --- a/app/test-pmd/testpmd.c > +++ b/app/test-pmd/testpmd.c > @@ -66,6 +66,9 @@ > #ifdef RTE_EXEC_ENV_WINDOWS > #include <process.h> > #endif > +#ifdef RTE_NET_BOND > +#include <rte_eth_bond.h> > +#endif > > #include "testpmd.h" > > @@ -597,11 +600,57 @@ eth_dev_configure_mp(uint16_t port_id, uint16_t nb_rx_q, uint16_t nb_tx_q, > return 0; > } > > +#ifdef RTE_NET_BOND > +static int > +change_bonding_slave_port_status(portid_t bond_pid, bool is_stop) > +{ > + portid_t slave_pids[RTE_MAX_ETHPORTS]; > + struct rte_port *port; > + int num_slaves; > + portid_t slave_pid; > + int i; > + > + num_slaves = rte_eth_bond_slaves_get(bond_pid, slave_pids, > + RTE_MAX_ETHPORTS); > + if (num_slaves < 0) { > + fprintf(stderr, "Failed to get slave list for port = %u\n", > + bond_pid); > + return num_slaves; > + } > + > + for (i = 0; i < num_slaves; i++) { > + slave_pid = slave_pids[i]; > + port = &ports[slave_pid]; > + port->port_status = > + is_stop ? RTE_PORT_STOPPED : RTE_PORT_STARTED; > + } > + > + return 0; > +} > +#endif > + > static int > eth_dev_start_mp(uint16_t port_id) > { > - if (is_proc_primary()) > - return rte_eth_dev_start(port_id); > + int ret; > + > + if (is_proc_primary()) { > + ret = rte_eth_dev_start(port_id); > + if (ret != 0) > + return ret; > + > +#ifdef RTE_NET_BOND > + struct rte_port *port = &ports[port_id]; > + > + /* > + * Starting a bonded port also starts all slaves under the bonded > + * device. So if this port is bond device, we need to modify the > + * port status of these slaves. > + */ > + if (port->bond_flag == 1) > + return change_bonding_slave_port_status(port_id, false); > +#endif > + } > > return 0; > } > @@ -609,8 +658,25 @@ eth_dev_start_mp(uint16_t port_id) > static int > eth_dev_stop_mp(uint16_t port_id) > { > - if (is_proc_primary()) > - return rte_eth_dev_stop(port_id); > + int ret; > + > + if (is_proc_primary()) { > + ret = rte_eth_dev_stop(port_id); > + if (ret != 0) > + return ret; > + > +#ifdef RTE_NET_BOND Here and in other places - probably no need to pollute the code with all these 'ifdef RTE_NET_BOND'. I suppose this logic (for checking is bonding API present or not) can be hidden inside change_bonding_slave_port_status() itself. > + struct rte_port *port = &ports[port_id]; > + > + /* > + * Stopping a bonded port also stops all slaves under the bonded > + * device. So if this port is bond device, we need to modify the > + * port status of these slaves. > + */ > + if (port->bond_flag == 1) > + return change_bonding_slave_port_status(port_id, true); > +#endif > + } > > return 0; > } > diff --git a/app/test-pmd/testpmd.h b/app/test-pmd/testpmd.h > index 31f766c965..67f253b30e 100644 > --- a/app/test-pmd/testpmd.h > +++ b/app/test-pmd/testpmd.h > @@ -266,7 +266,8 @@ struct rte_port { > uint32_t mc_addr_nb; /**< nb. of addr. in mc_addr_pool */ > queueid_t queue_nb; /**< nb. of queues for flow rules */ > uint32_t queue_sz; /**< size of a queue for flow rules */ > - uint8_t slave_flag; /**< bonding slave port */ > + uint8_t slave_flag : 1, /**< bonding slave port */ > + bond_flag : 1; /**< port is bond device */ > struct port_template *pattern_templ_list; /**< Pattern templates. */ > struct port_template *actions_templ_list; /**< Actions templates. */ > struct port_table *table_list; /**< Flow tables. */
Hi, Ferruh, 在 2022/5/4 3:04, Ferruh Yigit 写道: > On 5/3/2022 7:54 AM, Min Hu (Connor) wrote: >> Hi, Ferruh, >> >> 在 2022/4/29 21:31, Ferruh Yigit 写道: >>> On 4/29/2022 7:45 AM, Min Hu (Connor) wrote: >>>> Hi, Ferruh, >>>> >>>> 在 2022/4/27 2:19, Ferruh Yigit 写道: >>>>> On 3/24/2022 3:00 AM, Min Hu (Connor) wrote: >>>>>> From: Huisong Li <lihuisong@huawei.com> >>>>>> >>>>>> When stopping a bonded port, all slaves should be deactivated. But >>>>>> only >>>>> >>>>> s/deactivated/stopped/ ? >>>> not agreed. deactivated and stopped are different state for slave. >>>> >>> >>> Just to clarify the sentences, otherwise I see the 'stopped' and >>> 'deactivated' states are different. >>> Next sentences complains that not all ports are stopped: "But only >>> active slaves are stopped.", so I thought intention in this sentences >>> to claim that all slaves should be stopped (but it mentions all >>> slaves should be 'deactivated'). >>> As long as you address the disconnection between two sentences, I >>> don't mind the wording. >> Actually, there is something wrong with the wording. >> Yes, I should take your advice. >> >>> >>>>> >>>>>> active slaves are stopped. So fix it and do "deactivae_slave()" >>>>>> for active >>>>> >>>>> s/deactivae_slave()/deactivate_slave()/ >>>>> >>>> agreed. >>>> >>>>>> slaves. >>>>> >>>>> Hi Connor, >>>>> >>>>> When a bonding port is closed, is it clear if all slave ports or >>>>> active slave ports should be stopped? >>>> Yes, I think all the slave ports should be stopped(or try to be >>>> stopped). >>>>> >>>>>> >>>>>> Fixes: 0911d4ec0183 ("net/bonding: fix crash when stopping mode 4 >>>>>> port") >>>>>> Cc: stable@dpdk.org >>>>>> >>>>>> Signed-off-by: Huisong Li <lihuisong@huawei.com> >>>>>> Signed-off-by: Min Hu (Connor) <humin29@huawei.com> >>>>>> --- >>>>>> drivers/net/bonding/rte_eth_bond_pmd.c | 20 +++++++++++--------- >>>>>> 1 file changed, 11 insertions(+), 9 deletions(-) >>>>>> >>>>>> diff --git a/drivers/net/bonding/rte_eth_bond_pmd.c >>>>>> b/drivers/net/bonding/rte_eth_bond_pmd.c >>>>>> index b305b6a35b..469dc71170 100644 >>>>>> --- a/drivers/net/bonding/rte_eth_bond_pmd.c >>>>>> +++ b/drivers/net/bonding/rte_eth_bond_pmd.c >>>>>> @@ -2118,18 +2118,20 @@ bond_ethdev_stop(struct rte_eth_dev *eth_dev) >>>>>> internals->link_status_polling_enabled = 0; >>>>>> for (i = 0; i < internals->slave_count; i++) { >>>>>> uint16_t slave_id = internals->slaves[i].port_id; >>>>>> + >>>>>> + internals->slaves[i].last_link_status = 0; >>>>>> + ret = rte_eth_dev_stop(slave_id); >>>>>> + if (ret != 0) { >>>>>> + RTE_BOND_LOG(ERR, "Failed to stop device on port %u", >>>>>> + slave_id); >>>>>> + return ret; >>>>> >>>>> Should it return here or try to stop all ports? >>>>> What about to record the return status, but keep continue to stop >>>>> all ports. And return error if any of the stop failed? >> Well, I am glad you have found something unreasaonable about 'stop'. >> Let us see API 'rte_eth_dev_stop' >> >> rte_eth_dev_stop(dev) >> { >> .... >> dev->data->dev_started = 0; >> ret = (*dev->dev_ops->dev_stop)(dev) >> retur ret; >> } >> This is unreasaonable. No matter 'dev_ops->dev_stop' succeed or fail, >> the state 'dev_started ' will always set to be '0'. >> >> But this does not only influence bonding device but other devices like >> eth dev or vdev. >> This is the bug in rte ethdev level. I will send another patch to fix >> it. >> > > Hi Connor, > > I agree this is an issue in the API, cc'ed Andrew and Thomas. > > I vaguely remember that "dev_started = 0" was required for some dev_ops, > but not quite sure, let me check this. > At worst we can do as following to be sure: > > dev->data->dev_started = 0; > ret = (*dev->dev_ops->dev_stop)(dev) > if (ret) > dev->data->dev_started = 1; > > Also we need to clarify in the API documentation (.h file), what is the > status of the device if 'rte_eth_dev_stop()' returned error. > > > Btw, would you be OK to separate this ethdev patch from your bonding > patch, to not stuck your series because of ethdev one. Yes, this patch can be abandoned from this set. > > >> >>>> I think no need to do this. APP only see the bonded device. If bonded >>>> device stop failed, it means it works failed. And the number of >>>> "stopped" successfully slave does not make any sense. >>>> >>> >>> OK if trying to stop as much as possible 'slave' devices doesn't make >>> sense, we can keep as it is. >>> >>> Btw, when functions fails at this point, bonding device itself >>> already marked as stopped, right? And some of the slave devices may >>> be stopped already before failure. >>> I don't know how confusing this is for the user, that stop() function >>> is failed but bonding device state is 'stopped'. I don't know if >>> function should recover at least bonding device status (back to >>> started) on failure, what do you think? >>> >>>>> >>>>>> + } >>>>>> + >>>>>> + /* active slaves need to deactivate. */ >>>>> >>>>> " active slaves need to be deactivated. " ? >>>> agreed. >>>>> >>>>>> if (find_slave_by_id(internals->active_slaves, >>>>>> internals->active_slave_count, slave_id) != >>>>>> - internals->active_slave_count) { >>>>>> - internals->slaves[i].last_link_status = 0; >>>>>> - ret = rte_eth_dev_stop(slave_id); >>>>>> - if (ret != 0) { >>>>>> - RTE_BOND_LOG(ERR, "Failed to stop device on port >>>>>> %u", >>>>>> - slave_id); >>>>>> - return ret; >>>>>> - } >>>>>> + internals->active_slave_count) >>>>> >>>>> I think original indentation for this line is better. >>>>> >>>> agreed. >>>>>> deactivate_slave(eth_dev, slave_id); >>>>>> - } >>>>>> } >>>>>> return 0; >>>>> >>>>> . >>> >>> . > > .
Hi, Konstantin, 在 2022/5/4 7:39, Konstantin Ananyev 写道: > 03/05/2022 11:02, Min Hu (Connor) пишет: >> From: Huisong Li <lihuisong@huawei.com> >> >> Starting or stopping a bonded port also starts or stops all active slaves >> under the bonded port. If this port is a bonded device, we need to modify >> the port status of all slaves. >> >> Fixes: 0e545d3047fe ("app/testpmd: check stopping port is not in >> bonding") >> Cc: stable@dpdk.org >> >> Signed-off-by: Huisong Li <lihuisong@huawei.com> >> Signed-off-by: Min Hu (Connor) <humin29@huawei.com> >> Acked-by: Aman Singh <aman.deep.singh@intel.com> >> --- >> app/test-pmd/cmdline.c | 1 + >> app/test-pmd/testpmd.c | 74 +++++++++++++++++++++++++++++++++++++++--- >> app/test-pmd/testpmd.h | 3 +- >> 3 files changed, 73 insertions(+), 5 deletions(-) >> >> diff --git a/app/test-pmd/cmdline.c b/app/test-pmd/cmdline.c >> index 6ffea8e21a..d9fc7a88bd 100644 >> --- a/app/test-pmd/cmdline.c >> +++ b/app/test-pmd/cmdline.c >> @@ -6671,6 +6671,7 @@ static void cmd_create_bonded_device_parsed(void >> *parsed_result, >> "Failed to enable promiscuous mode for port %u: %s - >> ignore\n", >> port_id, rte_strerror(-ret)); >> + ports[port_id].bond_flag = 1; >> ports[port_id].need_setup = 0; >> ports[port_id].port_status = RTE_PORT_STOPPED; >> } >> diff --git a/app/test-pmd/testpmd.c b/app/test-pmd/testpmd.c >> index fe2ce19f99..dc90600787 100644 >> --- a/app/test-pmd/testpmd.c >> +++ b/app/test-pmd/testpmd.c >> @@ -66,6 +66,9 @@ >> #ifdef RTE_EXEC_ENV_WINDOWS >> #include <process.h> >> #endif >> +#ifdef RTE_NET_BOND >> +#include <rte_eth_bond.h> >> +#endif >> #include "testpmd.h" >> @@ -597,11 +600,57 @@ eth_dev_configure_mp(uint16_t port_id, uint16_t >> nb_rx_q, uint16_t nb_tx_q, >> return 0; >> } >> +#ifdef RTE_NET_BOND >> +static int >> +change_bonding_slave_port_status(portid_t bond_pid, bool is_stop) >> +{ >> + portid_t slave_pids[RTE_MAX_ETHPORTS]; >> + struct rte_port *port; >> + int num_slaves; >> + portid_t slave_pid; >> + int i; >> + >> + num_slaves = rte_eth_bond_slaves_get(bond_pid, slave_pids, >> + RTE_MAX_ETHPORTS); >> + if (num_slaves < 0) { >> + fprintf(stderr, "Failed to get slave list for port = %u\n", >> + bond_pid); >> + return num_slaves; >> + } >> + >> + for (i = 0; i < num_slaves; i++) { >> + slave_pid = slave_pids[i]; >> + port = &ports[slave_pid]; >> + port->port_status = >> + is_stop ? RTE_PORT_STOPPED : RTE_PORT_STARTED; >> + } >> + >> + return 0; >> +} >> +#endif >> + >> static int >> eth_dev_start_mp(uint16_t port_id) >> { >> - if (is_proc_primary()) >> - return rte_eth_dev_start(port_id); >> + int ret; >> + >> + if (is_proc_primary()) { >> + ret = rte_eth_dev_start(port_id); >> + if (ret != 0) >> + return ret; >> + >> +#ifdef RTE_NET_BOND >> + struct rte_port *port = &ports[port_id]; >> + >> + /* >> + * Starting a bonded port also starts all slaves under the >> bonded >> + * device. So if this port is bond device, we need to modify the >> + * port status of these slaves. >> + */ >> + if (port->bond_flag == 1) >> + return change_bonding_slave_port_status(port_id, false); >> +#endif >> + } >> return 0; >> } >> @@ -609,8 +658,25 @@ eth_dev_start_mp(uint16_t port_id) >> static int >> eth_dev_stop_mp(uint16_t port_id) >> { >> - if (is_proc_primary()) >> - return rte_eth_dev_stop(port_id); >> + int ret; >> + >> + if (is_proc_primary()) { >> + ret = rte_eth_dev_stop(port_id); >> + if (ret != 0) >> + return ret; >> + >> +#ifdef RTE_NET_BOND > > Here and in other places - probably no need to pollute the code > with all these 'ifdef RTE_NET_BOND'. > I suppose this logic (for checking is bonding API present or not) > can be hidden inside change_bonding_slave_port_status() itself. > I think it does not pollute the code. anyone can tell according to the flag 'ifdef RTE_NET_BOND'. if hiddle inside 'change_bonding_slave_port_status', it will be weird. For example, if the port is not bonding port, It will also invoke 'change_bonding_slave_port_status'. That is unreasonable. > >> + struct rte_port *port = &ports[port_id]; >> + >> + /* >> + * Stopping a bonded port also stops all slaves under the bonded >> + * device. So if this port is bond device, we need to modify the >> + * port status of these slaves. >> + */ >> + if (port->bond_flag == 1) >> + return change_bonding_slave_port_status(port_id, true); >> +#endif >> + } >> return 0; >> } >> diff --git a/app/test-pmd/testpmd.h b/app/test-pmd/testpmd.h >> index 31f766c965..67f253b30e 100644 >> --- a/app/test-pmd/testpmd.h >> +++ b/app/test-pmd/testpmd.h >> @@ -266,7 +266,8 @@ struct rte_port { >> uint32_t mc_addr_nb; /**< nb. of addr. in >> mc_addr_pool */ >> queueid_t queue_nb; /**< nb. of queues for flow >> rules */ >> uint32_t queue_sz; /**< size of a queue for flow >> rules */ >> - uint8_t slave_flag; /**< bonding slave port */ >> + uint8_t slave_flag : 1, /**< bonding slave port */ >> + bond_flag : 1; /**< port is bond device */ >> struct port_template *pattern_templ_list; /**< Pattern >> templates. */ >> struct port_template *actions_templ_list; /**< Actions >> templates. */ >> struct port_table *table_list; /**< Flow tables. */ > > .
Hi Conor, > 在 2022/5/4 7:39, Konstantin Ananyev 写道: >> 03/05/2022 11:02, Min Hu (Connor) пишет: >>> From: Huisong Li <lihuisong@huawei.com> >>> >>> Starting or stopping a bonded port also starts or stops all active >>> slaves >>> under the bonded port. If this port is a bonded device, we need to >>> modify >>> the port status of all slaves. >>> >>> Fixes: 0e545d3047fe ("app/testpmd: check stopping port is not in >>> bonding") >>> Cc: stable@dpdk.org >>> >>> Signed-off-by: Huisong Li <lihuisong@huawei.com> >>> Signed-off-by: Min Hu (Connor) <humin29@huawei.com> >>> Acked-by: Aman Singh <aman.deep.singh@intel.com> >>> --- >>> app/test-pmd/cmdline.c | 1 + >>> app/test-pmd/testpmd.c | 74 +++++++++++++++++++++++++++++++++++++++--- >>> app/test-pmd/testpmd.h | 3 +- >>> 3 files changed, 73 insertions(+), 5 deletions(-) >>> >>> diff --git a/app/test-pmd/cmdline.c b/app/test-pmd/cmdline.c >>> index 6ffea8e21a..d9fc7a88bd 100644 >>> --- a/app/test-pmd/cmdline.c >>> +++ b/app/test-pmd/cmdline.c >>> @@ -6671,6 +6671,7 @@ static void >>> cmd_create_bonded_device_parsed(void *parsed_result, >>> "Failed to enable promiscuous mode for port %u: %s >>> - ignore\n", >>> port_id, rte_strerror(-ret)); >>> + ports[port_id].bond_flag = 1; >>> ports[port_id].need_setup = 0; >>> ports[port_id].port_status = RTE_PORT_STOPPED; >>> } >>> diff --git a/app/test-pmd/testpmd.c b/app/test-pmd/testpmd.c >>> index fe2ce19f99..dc90600787 100644 >>> --- a/app/test-pmd/testpmd.c >>> +++ b/app/test-pmd/testpmd.c >>> @@ -66,6 +66,9 @@ >>> #ifdef RTE_EXEC_ENV_WINDOWS >>> #include <process.h> >>> #endif >>> +#ifdef RTE_NET_BOND >>> +#include <rte_eth_bond.h> >>> +#endif >>> #include "testpmd.h" >>> @@ -597,11 +600,57 @@ eth_dev_configure_mp(uint16_t port_id, uint16_t >>> nb_rx_q, uint16_t nb_tx_q, >>> return 0; >>> } >>> +#ifdef RTE_NET_BOND >>> +static int >>> +change_bonding_slave_port_status(portid_t bond_pid, bool is_stop) >>> +{ >>> + portid_t slave_pids[RTE_MAX_ETHPORTS]; >>> + struct rte_port *port; >>> + int num_slaves; >>> + portid_t slave_pid; >>> + int i; >>> + >>> + num_slaves = rte_eth_bond_slaves_get(bond_pid, slave_pids, >>> + RTE_MAX_ETHPORTS); >>> + if (num_slaves < 0) { >>> + fprintf(stderr, "Failed to get slave list for port = %u\n", >>> + bond_pid); >>> + return num_slaves; >>> + } >>> + >>> + for (i = 0; i < num_slaves; i++) { >>> + slave_pid = slave_pids[i]; >>> + port = &ports[slave_pid]; >>> + port->port_status = >>> + is_stop ? RTE_PORT_STOPPED : RTE_PORT_STARTED; >>> + } >>> + >>> + return 0; >>> +} >>> +#endif >>> + >>> static int >>> eth_dev_start_mp(uint16_t port_id) >>> { >>> - if (is_proc_primary()) >>> - return rte_eth_dev_start(port_id); >>> + int ret; >>> + >>> + if (is_proc_primary()) { >>> + ret = rte_eth_dev_start(port_id); >>> + if (ret != 0) >>> + return ret; >>> + >>> +#ifdef RTE_NET_BOND >>> + struct rte_port *port = &ports[port_id]; >>> + >>> + /* >>> + * Starting a bonded port also starts all slaves under the >>> bonded >>> + * device. So if this port is bond device, we need to modify >>> the >>> + * port status of these slaves. >>> + */ >>> + if (port->bond_flag == 1) >>> + return change_bonding_slave_port_status(port_id, false); >>> +#endif >>> + } >>> return 0; >>> } >>> @@ -609,8 +658,25 @@ eth_dev_start_mp(uint16_t port_id) >>> static int >>> eth_dev_stop_mp(uint16_t port_id) >>> { >>> - if (is_proc_primary()) >>> - return rte_eth_dev_stop(port_id); >>> + int ret; >>> + >>> + if (is_proc_primary()) { >>> + ret = rte_eth_dev_stop(port_id); >>> + if (ret != 0) >>> + return ret; >>> + >>> +#ifdef RTE_NET_BOND >> >> Here and in other places - probably no need to pollute the code >> with all these 'ifdef RTE_NET_BOND'. >> I suppose this logic (for checking is bonding API present or not) >> can be hidden inside change_bonding_slave_port_status() itself. >> > I think it does not pollute the code. anyone can tell according to > the flag 'ifdef RTE_NET_BOND'. That what I am talking about. Spreading ifdefed code all around the palce is not a good thing. It makes it harder to read, understand and maintain. Much more plausible is to hide that logic in one place whenever possible. > if hiddle inside 'change_bonding_slave_port_status', it will be weird. I don't see why is that. Let say if bonding is disabled that function can do nothing or even return an error to avoid it's misuse. > For example, if the port is not bonding port, It will also invoke > 'change_bonding_slave_port_status'. That is unreasonable. As I can read the code, hange_bonding_slave_port_status() will be invoked only when port->bond_flag is set. Which implies that port is a bonded one, no? > > >> >>> + struct rte_port *port = &ports[port_id]; >>> + >>> + /* >>> + * Stopping a bonded port also stops all slaves under the >>> bonded >>> + * device. So if this port is bond device, we need to modify >>> the >>> + * port status of these slaves. >>> + */ >>> + if (port->bond_flag == 1) >>> + return change_bonding_slave_port_status(port_id, true); >>> +#endif >>> + } >>> return 0; >>> } >>> diff --git a/app/test-pmd/testpmd.h b/app/test-pmd/testpmd.h >>> index 31f766c965..67f253b30e 100644 >>> --- a/app/test-pmd/testpmd.h >>> +++ b/app/test-pmd/testpmd.h >>> @@ -266,7 +266,8 @@ struct rte_port { >>> uint32_t mc_addr_nb; /**< nb. of addr. in >>> mc_addr_pool */ >>> queueid_t queue_nb; /**< nb. of queues for flow >>> rules */ >>> uint32_t queue_sz; /**< size of a queue for flow >>> rules */ >>> - uint8_t slave_flag; /**< bonding slave port */ >>> + uint8_t slave_flag : 1, /**< bonding slave port */ >>> + bond_flag : 1; /**< port is bond device */ >>> struct port_template *pattern_templ_list; /**< Pattern >>> templates. */ >>> struct port_template *actions_templ_list; /**< Actions >>> templates. */ >>> struct port_table *table_list; /**< Flow tables. */ >> >> .
On 5/6/2022 9:16 AM, Min Hu (Connor) wrote: > Hi, Konstantin, > > 在 2022/5/4 7:39, Konstantin Ananyev 写道: >> 03/05/2022 11:02, Min Hu (Connor) пишет: >>> From: Huisong Li <lihuisong@huawei.com> >>> >>> Starting or stopping a bonded port also starts or stops all active >>> slaves >>> under the bonded port. If this port is a bonded device, we need to >>> modify >>> the port status of all slaves. >>> >>> Fixes: 0e545d3047fe ("app/testpmd: check stopping port is not in >>> bonding") >>> Cc: stable@dpdk.org >>> >>> Signed-off-by: Huisong Li <lihuisong@huawei.com> >>> Signed-off-by: Min Hu (Connor) <humin29@huawei.com> >>> Acked-by: Aman Singh <aman.deep.singh@intel.com> >>> --- >>> app/test-pmd/cmdline.c | 1 + >>> app/test-pmd/testpmd.c | 74 +++++++++++++++++++++++++++++++++++++++--- >>> app/test-pmd/testpmd.h | 3 +- >>> 3 files changed, 73 insertions(+), 5 deletions(-) >>> >>> diff --git a/app/test-pmd/cmdline.c b/app/test-pmd/cmdline.c >>> index 6ffea8e21a..d9fc7a88bd 100644 >>> --- a/app/test-pmd/cmdline.c >>> +++ b/app/test-pmd/cmdline.c >>> @@ -6671,6 +6671,7 @@ static void >>> cmd_create_bonded_device_parsed(void *parsed_result, >>> "Failed to enable promiscuous mode for port %u: %s >>> - ignore\n", >>> port_id, rte_strerror(-ret)); >>> + ports[port_id].bond_flag = 1; >>> ports[port_id].need_setup = 0; >>> ports[port_id].port_status = RTE_PORT_STOPPED; >>> } >>> diff --git a/app/test-pmd/testpmd.c b/app/test-pmd/testpmd.c >>> index fe2ce19f99..dc90600787 100644 >>> --- a/app/test-pmd/testpmd.c >>> +++ b/app/test-pmd/testpmd.c >>> @@ -66,6 +66,9 @@ >>> #ifdef RTE_EXEC_ENV_WINDOWS >>> #include <process.h> >>> #endif >>> +#ifdef RTE_NET_BOND >>> +#include <rte_eth_bond.h> >>> +#endif >>> #include "testpmd.h" >>> @@ -597,11 +600,57 @@ eth_dev_configure_mp(uint16_t port_id, uint16_t >>> nb_rx_q, uint16_t nb_tx_q, >>> return 0; >>> } >>> +#ifdef RTE_NET_BOND >>> +static int >>> +change_bonding_slave_port_status(portid_t bond_pid, bool is_stop) >>> +{ >>> + portid_t slave_pids[RTE_MAX_ETHPORTS]; >>> + struct rte_port *port; >>> + int num_slaves; >>> + portid_t slave_pid; >>> + int i; >>> + >>> + num_slaves = rte_eth_bond_slaves_get(bond_pid, slave_pids, >>> + RTE_MAX_ETHPORTS); >>> + if (num_slaves < 0) { >>> + fprintf(stderr, "Failed to get slave list for port = %u\n", >>> + bond_pid); >>> + return num_slaves; >>> + } >>> + >>> + for (i = 0; i < num_slaves; i++) { >>> + slave_pid = slave_pids[i]; >>> + port = &ports[slave_pid]; >>> + port->port_status = >>> + is_stop ? RTE_PORT_STOPPED : RTE_PORT_STARTED; >>> + } >>> + >>> + return 0; >>> +} >>> +#endif >>> + >>> static int >>> eth_dev_start_mp(uint16_t port_id) >>> { >>> - if (is_proc_primary()) >>> - return rte_eth_dev_start(port_id); >>> + int ret; >>> + >>> + if (is_proc_primary()) { >>> + ret = rte_eth_dev_start(port_id); >>> + if (ret != 0) >>> + return ret; >>> + >>> +#ifdef RTE_NET_BOND >>> + struct rte_port *port = &ports[port_id]; >>> + >>> + /* >>> + * Starting a bonded port also starts all slaves under the >>> bonded >>> + * device. So if this port is bond device, we need to modify >>> the >>> + * port status of these slaves. >>> + */ >>> + if (port->bond_flag == 1) >>> + return change_bonding_slave_port_status(port_id, false); >>> +#endif >>> + } >>> return 0; >>> } >>> @@ -609,8 +658,25 @@ eth_dev_start_mp(uint16_t port_id) >>> static int >>> eth_dev_stop_mp(uint16_t port_id) >>> { >>> - if (is_proc_primary()) >>> - return rte_eth_dev_stop(port_id); >>> + int ret; >>> + >>> + if (is_proc_primary()) { >>> + ret = rte_eth_dev_stop(port_id); >>> + if (ret != 0) >>> + return ret; >>> + >>> +#ifdef RTE_NET_BOND >> >> Here and in other places - probably no need to pollute the code >> with all these 'ifdef RTE_NET_BOND'. >> I suppose this logic (for checking is bonding API present or not) >> can be hidden inside change_bonding_slave_port_status() itself. >> > I think it does not pollute the code. anyone can tell according to > the flag 'ifdef RTE_NET_BOND'. > if hiddle inside 'change_bonding_slave_port_status', it will be weird. > For example, if the port is not bonding port, It will also invoke > 'change_bonding_slave_port_status'. That is unreasonable. > Hi Konstantin, I also was not happy to have bonding (or any PMD) ifdef in the generic (start()/stop()) functions, but the ifdef blocks updates testpmd (application) level status. So that can't be handled in the PMD and need to be in the application level. Which is enforcing to have same PMD specific code in the testpmd level, if you have any suggestion to prevent this, I am for it. I will proceed with first two patch of this set, which fixes bonding PMD issues, I will hold the testpmd ones for more comments. Thanks, ferruh > >> >>> + struct rte_port *port = &ports[port_id]; >>> + >>> + /* >>> + * Stopping a bonded port also stops all slaves under the >>> bonded >>> + * device. So if this port is bond device, we need to modify >>> the >>> + * port status of these slaves. >>> + */ >>> + if (port->bond_flag == 1) >>> + return change_bonding_slave_port_status(port_id, true); >>> +#endif >>> + } >>> return 0; >>> } >>> diff --git a/app/test-pmd/testpmd.h b/app/test-pmd/testpmd.h >>> index 31f766c965..67f253b30e 100644 >>> --- a/app/test-pmd/testpmd.h >>> +++ b/app/test-pmd/testpmd.h >>> @@ -266,7 +266,8 @@ struct rte_port { >>> uint32_t mc_addr_nb; /**< nb. of addr. in >>> mc_addr_pool */ >>> queueid_t queue_nb; /**< nb. of queues for flow >>> rules */ >>> uint32_t queue_sz; /**< size of a queue for flow >>> rules */ >>> - uint8_t slave_flag; /**< bonding slave port */ >>> + uint8_t slave_flag : 1, /**< bonding slave port */ >>> + bond_flag : 1; /**< port is bond device */ >>> struct port_template *pattern_templ_list; /**< Pattern >>> templates. */ >>> struct port_template *actions_templ_list; /**< Actions >>> templates. */ >>> struct port_table *table_list; /**< Flow tables. */ >> >> .
Hi Ferruh, > On 5/6/2022 9:16 AM, Min Hu (Connor) wrote: >> Hi, Konstantin, >> >> 在 2022/5/4 7:39, Konstantin Ananyev 写道: >>> 03/05/2022 11:02, Min Hu (Connor) пишет: >>>> From: Huisong Li <lihuisong@huawei.com> >>>> >>>> Starting or stopping a bonded port also starts or stops all active >>>> slaves >>>> under the bonded port. If this port is a bonded device, we need to >>>> modify >>>> the port status of all slaves. >>>> >>>> Fixes: 0e545d3047fe ("app/testpmd: check stopping port is not in >>>> bonding") >>>> Cc: stable@dpdk.org >>>> >>>> Signed-off-by: Huisong Li <lihuisong@huawei.com> >>>> Signed-off-by: Min Hu (Connor) <humin29@huawei.com> >>>> Acked-by: Aman Singh <aman.deep.singh@intel.com> >>>> --- >>>> app/test-pmd/cmdline.c | 1 + >>>> app/test-pmd/testpmd.c | 74 >>>> +++++++++++++++++++++++++++++++++++++++--- >>>> app/test-pmd/testpmd.h | 3 +- >>>> 3 files changed, 73 insertions(+), 5 deletions(-) >>>> >>>> diff --git a/app/test-pmd/cmdline.c b/app/test-pmd/cmdline.c >>>> index 6ffea8e21a..d9fc7a88bd 100644 >>>> --- a/app/test-pmd/cmdline.c >>>> +++ b/app/test-pmd/cmdline.c >>>> @@ -6671,6 +6671,7 @@ static void >>>> cmd_create_bonded_device_parsed(void *parsed_result, >>>> "Failed to enable promiscuous mode for port %u: %s >>>> - ignore\n", >>>> port_id, rte_strerror(-ret)); >>>> + ports[port_id].bond_flag = 1; >>>> ports[port_id].need_setup = 0; >>>> ports[port_id].port_status = RTE_PORT_STOPPED; >>>> } >>>> diff --git a/app/test-pmd/testpmd.c b/app/test-pmd/testpmd.c >>>> index fe2ce19f99..dc90600787 100644 >>>> --- a/app/test-pmd/testpmd.c >>>> +++ b/app/test-pmd/testpmd.c >>>> @@ -66,6 +66,9 @@ >>>> #ifdef RTE_EXEC_ENV_WINDOWS >>>> #include <process.h> >>>> #endif >>>> +#ifdef RTE_NET_BOND >>>> +#include <rte_eth_bond.h> >>>> +#endif >>>> #include "testpmd.h" >>>> @@ -597,11 +600,57 @@ eth_dev_configure_mp(uint16_t port_id, >>>> uint16_t nb_rx_q, uint16_t nb_tx_q, >>>> return 0; >>>> } >>>> +#ifdef RTE_NET_BOND >>>> +static int >>>> +change_bonding_slave_port_status(portid_t bond_pid, bool is_stop) >>>> +{ >>>> + portid_t slave_pids[RTE_MAX_ETHPORTS]; >>>> + struct rte_port *port; >>>> + int num_slaves; >>>> + portid_t slave_pid; >>>> + int i; >>>> + >>>> + num_slaves = rte_eth_bond_slaves_get(bond_pid, slave_pids, >>>> + RTE_MAX_ETHPORTS); >>>> + if (num_slaves < 0) { >>>> + fprintf(stderr, "Failed to get slave list for port = %u\n", >>>> + bond_pid); >>>> + return num_slaves; >>>> + } >>>> + >>>> + for (i = 0; i < num_slaves; i++) { >>>> + slave_pid = slave_pids[i]; >>>> + port = &ports[slave_pid]; >>>> + port->port_status = >>>> + is_stop ? RTE_PORT_STOPPED : RTE_PORT_STARTED; >>>> + } >>>> + >>>> + return 0; >>>> +} >>>> +#endif >>>> + >>>> static int >>>> eth_dev_start_mp(uint16_t port_id) >>>> { >>>> - if (is_proc_primary()) >>>> - return rte_eth_dev_start(port_id); >>>> + int ret; >>>> + >>>> + if (is_proc_primary()) { >>>> + ret = rte_eth_dev_start(port_id); >>>> + if (ret != 0) >>>> + return ret; >>>> + >>>> +#ifdef RTE_NET_BOND >>>> + struct rte_port *port = &ports[port_id]; >>>> + >>>> + /* >>>> + * Starting a bonded port also starts all slaves under the >>>> bonded >>>> + * device. So if this port is bond device, we need to >>>> modify the >>>> + * port status of these slaves. >>>> + */ >>>> + if (port->bond_flag == 1) >>>> + return change_bonding_slave_port_status(port_id, false); >>>> +#endif >>>> + } >>>> return 0; >>>> } >>>> @@ -609,8 +658,25 @@ eth_dev_start_mp(uint16_t port_id) >>>> static int >>>> eth_dev_stop_mp(uint16_t port_id) >>>> { >>>> - if (is_proc_primary()) >>>> - return rte_eth_dev_stop(port_id); >>>> + int ret; >>>> + >>>> + if (is_proc_primary()) { >>>> + ret = rte_eth_dev_stop(port_id); >>>> + if (ret != 0) >>>> + return ret; >>>> + >>>> +#ifdef RTE_NET_BOND >>> >>> Here and in other places - probably no need to pollute the code >>> with all these 'ifdef RTE_NET_BOND'. >>> I suppose this logic (for checking is bonding API present or not) >>> can be hidden inside change_bonding_slave_port_status() itself. >>> >> I think it does not pollute the code. anyone can tell according to >> the flag 'ifdef RTE_NET_BOND'. >> if hiddle inside 'change_bonding_slave_port_status', it will be weird. >> For example, if the port is not bonding port, It will also invoke >> 'change_bonding_slave_port_status'. That is unreasonable. >> > > Hi Konstantin, > > I also was not happy to have bonding (or any PMD) ifdef in the generic > (start()/stop()) functions, but the ifdef blocks updates testpmd > (application) level status. So that can't be handled in the PMD and need > to be in the application level. > Which is enforcing to have same PMD specific code in the testpmd level, > if you have any suggestion to prevent this, I am for it. I am not aking to move it to PMD. What I am saying that this ifdef logic could be grouped in one place (inside change_bonding_slave_port_status()) instead of spreading it around multiple places. > > I will proceed with first two patch of this set, which fixes bonding PMD > issues, I will hold the testpmd ones for more comments. > > Thanks, > ferruh > >> >>> >>>> + struct rte_port *port = &ports[port_id]; >>>> + >>>> + /* >>>> + * Stopping a bonded port also stops all slaves under the >>>> bonded >>>> + * device. So if this port is bond device, we need to >>>> modify the >>>> + * port status of these slaves. >>>> + */ >>>> + if (port->bond_flag == 1) >>>> + return change_bonding_slave_port_status(port_id, true); >>>> +#endif >>>> + } >>>> return 0; >>>> } >>>> diff --git a/app/test-pmd/testpmd.h b/app/test-pmd/testpmd.h >>>> index 31f766c965..67f253b30e 100644 >>>> --- a/app/test-pmd/testpmd.h >>>> +++ b/app/test-pmd/testpmd.h >>>> @@ -266,7 +266,8 @@ struct rte_port { >>>> uint32_t mc_addr_nb; /**< nb. of addr. in >>>> mc_addr_pool */ >>>> queueid_t queue_nb; /**< nb. of queues for flow >>>> rules */ >>>> uint32_t queue_sz; /**< size of a queue for >>>> flow rules */ >>>> - uint8_t slave_flag; /**< bonding slave port */ >>>> + uint8_t slave_flag : 1, /**< bonding slave port */ >>>> + bond_flag : 1; /**< port is bond device */ >>>> struct port_template *pattern_templ_list; /**< Pattern >>>> templates. */ >>>> struct port_template *actions_templ_list; /**< Actions >>>> templates. */ >>>> struct port_table *table_list; /**< Flow tables. */ >>> >>> . >
From: Huisong Li <lihuisong@huawei.com> Starting or stopping a bonded port also starts or stops all active slaves under the bonded port. If this port is a bonded device, we need to modify the port status of all slaves. Fixes: 0e545d3047fe ("app/testpmd: check stopping port is not in bonding") Cc: stable@dpdk.org Signed-off-by: Huisong Li <lihuisong@huawei.com> Signed-off-by: Min Hu (Connor) <humin29@huawei.com> Acked-by: Aman Singh <aman.deep.singh@intel.com> --- v4: * set 'ifdef logic' grouped inside change_bonding_slave_port_status. --- app/test-pmd/cmdline.c | 1 + app/test-pmd/testpmd.c | 73 +++++++++++++++++++++++++++++++++++++++--- app/test-pmd/testpmd.h | 3 +- 3 files changed, 72 insertions(+), 5 deletions(-) diff --git a/app/test-pmd/cmdline.c b/app/test-pmd/cmdline.c index 6ffea8e21a..d9fc7a88bd 100644 --- a/app/test-pmd/cmdline.c +++ b/app/test-pmd/cmdline.c @@ -6671,6 +6671,7 @@ static void cmd_create_bonded_device_parsed(void *parsed_result, "Failed to enable promiscuous mode for port %u: %s - ignore\n", port_id, rte_strerror(-ret)); + ports[port_id].bond_flag = 1; ports[port_id].need_setup = 0; ports[port_id].port_status = RTE_PORT_STOPPED; } diff --git a/app/test-pmd/testpmd.c b/app/test-pmd/testpmd.c index fe2ce19f99..51fa344bb2 100644 --- a/app/test-pmd/testpmd.c +++ b/app/test-pmd/testpmd.c @@ -66,6 +66,9 @@ #ifdef RTE_EXEC_ENV_WINDOWS #include <process.h> #endif +#ifdef RTE_NET_BOND +#include <rte_eth_bond.h> +#endif #include "testpmd.h" @@ -597,11 +600,58 @@ eth_dev_configure_mp(uint16_t port_id, uint16_t nb_rx_q, uint16_t nb_tx_q, return 0; } +static int +change_bonding_slave_port_status(portid_t bond_pid, bool is_stop) +{ +#ifdef RTE_NET_BOND + + portid_t slave_pids[RTE_MAX_ETHPORTS]; + struct rte_port *port; + int num_slaves; + portid_t slave_pid; + int i; + + num_slaves = rte_eth_bond_slaves_get(bond_pid, slave_pids, + RTE_MAX_ETHPORTS); + if (num_slaves < 0) { + fprintf(stderr, "Failed to get slave list for port = %u\n", + bond_pid); + return num_slaves; + } + + for (i = 0; i < num_slaves; i++) { + slave_pid = slave_pids[i]; + port = &ports[slave_pid]; + port->port_status = + is_stop ? RTE_PORT_STOPPED : RTE_PORT_STARTED; + } +#else + RTE_SET_USED(bond_pid); + RTE_SET_USED(is_stop); +#endif + return 0; +} + static int eth_dev_start_mp(uint16_t port_id) { - if (is_proc_primary()) - return rte_eth_dev_start(port_id); + int ret; + + if (is_proc_primary()) { + ret = rte_eth_dev_start(port_id); + if (ret != 0) + return ret; + + struct rte_port *port = &ports[port_id]; + + /* + * Starting a bonded port also starts all slaves under the bonded + * device. So if this port is bond device, we need to modify the + * port status of these slaves. + */ + if (port->bond_flag == 1) + return change_bonding_slave_port_status(port_id, false); + } return 0; } @@ -609,8 +659,23 @@ eth_dev_start_mp(uint16_t port_id) static int eth_dev_stop_mp(uint16_t port_id) { - if (is_proc_primary()) - return rte_eth_dev_stop(port_id); + int ret; + + if (is_proc_primary()) { + ret = rte_eth_dev_stop(port_id); + if (ret != 0) + return ret; + + struct rte_port *port = &ports[port_id]; + + /* + * Stopping a bonded port also stops all slaves under the bonded + * device. So if this port is bond device, we need to modify the + * port status of these slaves. + */ + if (port->bond_flag == 1) + return change_bonding_slave_port_status(port_id, true); + } return 0; } diff --git a/app/test-pmd/testpmd.h b/app/test-pmd/testpmd.h index 31f766c965..67f253b30e 100644 --- a/app/test-pmd/testpmd.h +++ b/app/test-pmd/testpmd.h @@ -266,7 +266,8 @@ struct rte_port { uint32_t mc_addr_nb; /**< nb. of addr. in mc_addr_pool */ queueid_t queue_nb; /**< nb. of queues for flow rules */ uint32_t queue_sz; /**< size of a queue for flow rules */ - uint8_t slave_flag; /**< bonding slave port */ + uint8_t slave_flag : 1, /**< bonding slave port */ + bond_flag : 1; /**< port is bond device */ struct port_template *pattern_templ_list; /**< Pattern templates. */ struct port_template *actions_templ_list; /**< Actions templates. */ struct port_table *table_list; /**< Flow tables. */ -- 2.33.0
Hi, Konstantin,
fixed in v4, thanks.
在 2022/5/11 5:48, Konstantin Ananyev 写道:
>
> Hi Ferruh,
>
>> On 5/6/2022 9:16 AM, Min Hu (Connor) wrote:
>>> Hi, Konstantin,
>>>
>>> 在 2022/5/4 7:39, Konstantin Ananyev 写道:
>>>> 03/05/2022 11:02, Min Hu (Connor) пишет:
>>>>> From: Huisong Li <lihuisong@huawei.com>
>>>>>
>>>>> Starting or stopping a bonded port also starts or stops all active
>>>>> slaves
>>>>> under the bonded port. If this port is a bonded device, we need to
>>>>> modify
>>>>> the port status of all slaves.
>>>>>
>>>>> Fixes: 0e545d3047fe ("app/testpmd: check stopping port is not in
>>>>> bonding")
>>>>> Cc: stable@dpdk.org
>>>>>
>>>>> Signed-off-by: Huisong Li <lihuisong@huawei.com>
>>>>> Signed-off-by: Min Hu (Connor) <humin29@huawei.com>
>>>>> Acked-by: Aman Singh <aman.deep.singh@intel.com>
>>>>> ---
>>>>> app/test-pmd/cmdline.c | 1 +
>>>>> app/test-pmd/testpmd.c | 74
>>>>> +++++++++++++++++++++++++++++++++++++++---
>>>>> app/test-pmd/testpmd.h | 3 +-
>>>>> 3 files changed, 73 insertions(+), 5 deletions(-)
>>>>>
>>>>> diff --git a/app/test-pmd/cmdline.c b/app/test-pmd/cmdline.c
>>>>> index 6ffea8e21a..d9fc7a88bd 100644
>>>>> --- a/app/test-pmd/cmdline.c
>>>>> +++ b/app/test-pmd/cmdline.c
>>>>> @@ -6671,6 +6671,7 @@ static void
>>>>> cmd_create_bonded_device_parsed(void *parsed_result,
>>>>> "Failed to enable promiscuous mode for port %u:
>>>>> %s - ignore\n",
>>>>> port_id, rte_strerror(-ret));
>>>>> + ports[port_id].bond_flag = 1;
>>>>> ports[port_id].need_setup = 0;
>>>>> ports[port_id].port_status = RTE_PORT_STOPPED;
>>>>> }
>>>>> diff --git a/app/test-pmd/testpmd.c b/app/test-pmd/testpmd.c
>>>>> index fe2ce19f99..dc90600787 100644
>>>>> --- a/app/test-pmd/testpmd.c
>>>>> +++ b/app/test-pmd/testpmd.c
>>>>> @@ -66,6 +66,9 @@
>>>>> #ifdef RTE_EXEC_ENV_WINDOWS
>>>>> #include <process.h>
>>>>> #endif
>>>>> +#ifdef RTE_NET_BOND
>>>>> +#include <rte_eth_bond.h>
>>>>> +#endif
>>>>> #include "testpmd.h"
>>>>> @@ -597,11 +600,57 @@ eth_dev_configure_mp(uint16_t port_id,
>>>>> uint16_t nb_rx_q, uint16_t nb_tx_q,
>>>>> return 0;
>>>>> }
>>>>> +#ifdef RTE_NET_BOND
>>>>> +static int
>>>>> +change_bonding_slave_port_status(portid_t bond_pid, bool is_stop)
>>>>> +{
>>>>> + portid_t slave_pids[RTE_MAX_ETHPORTS];
>>>>> + struct rte_port *port;
>>>>> + int num_slaves;
>>>>> + portid_t slave_pid;
>>>>> + int i;
>>>>> +
>>>>> + num_slaves = rte_eth_bond_slaves_get(bond_pid, slave_pids,
>>>>> + RTE_MAX_ETHPORTS);
>>>>> + if (num_slaves < 0) {
>>>>> + fprintf(stderr, "Failed to get slave list for port = %u\n",
>>>>> + bond_pid);
>>>>> + return num_slaves;
>>>>> + }
>>>>> +
>>>>> + for (i = 0; i < num_slaves; i++) {
>>>>> + slave_pid = slave_pids[i];
>>>>> + port = &ports[slave_pid];
>>>>> + port->port_status =
>>>>> + is_stop ? RTE_PORT_STOPPED : RTE_PORT_STARTED;
>>>>> + }
>>>>> +
>>>>> + return 0;
>>>>> +}
>>>>> +#endif
>>>>> +
>>>>> static int
>>>>> eth_dev_start_mp(uint16_t port_id)
>>>>> {
>>>>> - if (is_proc_primary())
>>>>> - return rte_eth_dev_start(port_id);
>>>>> + int ret;
>>>>> +
>>>>> + if (is_proc_primary()) {
>>>>> + ret = rte_eth_dev_start(port_id);
>>>>> + if (ret != 0)
>>>>> + return ret;
>>>>> +
>>>>> +#ifdef RTE_NET_BOND
>>>>> + struct rte_port *port = &ports[port_id];
>>>>> +
>>>>> + /*
>>>>> + * Starting a bonded port also starts all slaves under the
>>>>> bonded
>>>>> + * device. So if this port is bond device, we need to
>>>>> modify the
>>>>> + * port status of these slaves.
>>>>> + */
>>>>> + if (port->bond_flag == 1)
>>>>> + return change_bonding_slave_port_status(port_id, false);
>>>>> +#endif
>>>>> + }
>>>>> return 0;
>>>>> }
>>>>> @@ -609,8 +658,25 @@ eth_dev_start_mp(uint16_t port_id)
>>>>> static int
>>>>> eth_dev_stop_mp(uint16_t port_id)
>>>>> {
>>>>> - if (is_proc_primary())
>>>>> - return rte_eth_dev_stop(port_id);
>>>>> + int ret;
>>>>> +
>>>>> + if (is_proc_primary()) {
>>>>> + ret = rte_eth_dev_stop(port_id);
>>>>> + if (ret != 0)
>>>>> + return ret;
>>>>> +
>>>>> +#ifdef RTE_NET_BOND
>>>>
>>>> Here and in other places - probably no need to pollute the code
>>>> with all these 'ifdef RTE_NET_BOND'.
>>>> I suppose this logic (for checking is bonding API present or not)
>>>> can be hidden inside change_bonding_slave_port_status() itself.
>>>>
>>> I think it does not pollute the code. anyone can tell according to
>>> the flag 'ifdef RTE_NET_BOND'.
>>> if hiddle inside 'change_bonding_slave_port_status', it will be weird.
>>> For example, if the port is not bonding port, It will also invoke
>>> 'change_bonding_slave_port_status'. That is unreasonable.
>>>
>>
>> Hi Konstantin,
>>
>> I also was not happy to have bonding (or any PMD) ifdef in the generic
>> (start()/stop()) functions, but the ifdef blocks updates testpmd
>> (application) level status. So that can't be handled in the PMD and
>> need to be in the application level.
>> Which is enforcing to have same PMD specific code in the testpmd level,
>> if you have any suggestion to prevent this, I am for it.
>
>
> I am not aking to move it to PMD.
> What I am saying that this ifdef logic could be grouped in one place
> (inside change_bonding_slave_port_status()) instead of spreading it
> around multiple places.
>
>>
>> I will proceed with first two patch of this set, which fixes bonding
>> PMD issues, I will hold the testpmd ones for more comments.
>>
>> Thanks,
>> ferruh
>>
>>>
>>>>
>>>>> + struct rte_port *port = &ports[port_id];
>>>>> +
>>>>> + /*
>>>>> + * Stopping a bonded port also stops all slaves under the
>>>>> bonded
>>>>> + * device. So if this port is bond device, we need to
>>>>> modify the
>>>>> + * port status of these slaves.
>>>>> + */
>>>>> + if (port->bond_flag == 1)
>>>>> + return change_bonding_slave_port_status(port_id, true);
>>>>> +#endif
>>>>> + }
>>>>> return 0;
>>>>> }
>>>>> diff --git a/app/test-pmd/testpmd.h b/app/test-pmd/testpmd.h
>>>>> index 31f766c965..67f253b30e 100644
>>>>> --- a/app/test-pmd/testpmd.h
>>>>> +++ b/app/test-pmd/testpmd.h
>>>>> @@ -266,7 +266,8 @@ struct rte_port {
>>>>> uint32_t mc_addr_nb; /**< nb. of addr. in
>>>>> mc_addr_pool */
>>>>> queueid_t queue_nb; /**< nb. of queues for flow
>>>>> rules */
>>>>> uint32_t queue_sz; /**< size of a queue for
>>>>> flow rules */
>>>>> - uint8_t slave_flag; /**< bonding slave port */
>>>>> + uint8_t slave_flag : 1, /**< bonding slave
>>>>> port */
>>>>> + bond_flag : 1; /**< port is bond device */
>>>>> struct port_template *pattern_templ_list; /**< Pattern
>>>>> templates. */
>>>>> struct port_template *actions_templ_list; /**< Actions
>>>>> templates. */
>>>>> struct port_table *table_list; /**< Flow tables. */
>>>>
>>>> .
>>
>
> .
On 5/11/2022 3:16 AM, Min Hu (Connor) wrote:
<...>
>>>>>> @@ -609,8 +658,25 @@ eth_dev_start_mp(uint16_t port_id)
>>>>>> static int
>>>>>> eth_dev_stop_mp(uint16_t port_id)
>>>>>> {
>>>>>> - if (is_proc_primary())
>>>>>> - return rte_eth_dev_stop(port_id);
>>>>>> + int ret;
>>>>>> +
>>>>>> + if (is_proc_primary()) {
>>>>>> + ret = rte_eth_dev_stop(port_id);
>>>>>> + if (ret != 0)
>>>>>> + return ret;
>>>>>> +
>>>>>> +#ifdef RTE_NET_BOND
>>>>>
>>>>> Here and in other places - probably no need to pollute the code
>>>>> with all these 'ifdef RTE_NET_BOND'.
>>>>> I suppose this logic (for checking is bonding API present or not)
>>>>> can be hidden inside change_bonding_slave_port_status() itself.
>>>>>
>>>> I think it does not pollute the code. anyone can tell according to
>>>> the flag 'ifdef RTE_NET_BOND'.
>>>> if hiddle inside 'change_bonding_slave_port_status', it will be weird.
>>>> For example, if the port is not bonding port, It will also invoke
>>>> 'change_bonding_slave_port_status'. That is unreasonable.
>>>>
>>>
>>> Hi Konstantin,
>>>
>>> I also was not happy to have bonding (or any PMD) ifdef in the
>>> generic (start()/stop()) functions, but the ifdef blocks updates
>>> testpmd (application) level status. So that can't be handled in the
>>> PMD and need to be in the application level.
>>> Which is enforcing to have same PMD specific code in the testpmd level,
>>> if you have any suggestion to prevent this, I am for it.
>>
>>
>> I am not aking to move it to PMD.
>> What I am saying that this ifdef logic could be grouped in one place
>> (inside change_bonding_slave_port_status()) instead of spreading it
>> around multiple places.
>>
>
> Hi, Konstantin,
> fixed in v4, thanks.
Hi Conor,
What do you think to apply same on patch 4/5?
> From: Huisong Li <lihuisong@huawei.com>
>
> Starting or stopping a bonded port also starts or stops all active slaves
> under the bonded port. If this port is a bonded device, we need to modify
> the port status of all slaves.
>
> Fixes: 0e545d3047fe ("app/testpmd: check stopping port is not in bonding")
> Cc: stable@dpdk.org
>
> Signed-off-by: Huisong Li <lihuisong@huawei.com>
> Signed-off-by: Min Hu (Connor) <humin29@huawei.com>
> Acked-by: Aman Singh <aman.deep.singh@intel.com>
> ---
> v4:
> * set 'ifdef logic' grouped inside change_bonding_slave_port_status.
> ---
> app/test-pmd/cmdline.c | 1 +
> app/test-pmd/testpmd.c | 73 +++++++++++++++++++++++++++++++++++++++---
> app/test-pmd/testpmd.h | 3 +-
> 3 files changed, 72 insertions(+), 5 deletions(-)
>
> diff --git a/app/test-pmd/cmdline.c b/app/test-pmd/cmdline.c
> index 6ffea8e21a..d9fc7a88bd 100644
> --- a/app/test-pmd/cmdline.c
> +++ b/app/test-pmd/cmdline.c
> @@ -6671,6 +6671,7 @@ static void cmd_create_bonded_device_parsed(void *parsed_result,
> "Failed to enable promiscuous mode for port %u: %s - ignore\n",
> port_id, rte_strerror(-ret));
>
> + ports[port_id].bond_flag = 1;
> ports[port_id].need_setup = 0;
> ports[port_id].port_status = RTE_PORT_STOPPED;
> }
> diff --git a/app/test-pmd/testpmd.c b/app/test-pmd/testpmd.c
> index fe2ce19f99..51fa344bb2 100644
> --- a/app/test-pmd/testpmd.c
> +++ b/app/test-pmd/testpmd.c
> @@ -66,6 +66,9 @@
> #ifdef RTE_EXEC_ENV_WINDOWS
> #include <process.h>
> #endif
> +#ifdef RTE_NET_BOND
> +#include <rte_eth_bond.h>
> +#endif
>
> #include "testpmd.h"
>
> @@ -597,11 +600,58 @@ eth_dev_configure_mp(uint16_t port_id, uint16_t nb_rx_q, uint16_t nb_tx_q,
> return 0;
> }
>
> +static int
> +change_bonding_slave_port_status(portid_t bond_pid, bool is_stop)
> +{
> +#ifdef RTE_NET_BOND
> +
> + portid_t slave_pids[RTE_MAX_ETHPORTS];
> + struct rte_port *port;
> + int num_slaves;
> + portid_t slave_pid;
> + int i;
> +
> + num_slaves = rte_eth_bond_slaves_get(bond_pid, slave_pids,
> + RTE_MAX_ETHPORTS);
> + if (num_slaves < 0) {
> + fprintf(stderr, "Failed to get slave list for port = %u\n",
> + bond_pid);
> + return num_slaves;
> + }
> +
> + for (i = 0; i < num_slaves; i++) {
> + slave_pid = slave_pids[i];
> + port = &ports[slave_pid];
> + port->port_status =
> + is_stop ? RTE_PORT_STOPPED : RTE_PORT_STARTED;
> + }
> +#else
> + RTE_SET_USED(bond_pid);
> + RTE_SET_USED(is_stop);
> +#endif
> + return 0;
> +}
> +
> static int
> eth_dev_start_mp(uint16_t port_id)
> {
> - if (is_proc_primary())
> - return rte_eth_dev_start(port_id);
> + int ret;
> +
> + if (is_proc_primary()) {
> + ret = rte_eth_dev_start(port_id);
> + if (ret != 0)
> + return ret;
> +
> + struct rte_port *port = &ports[port_id];
> +
> + /*
> + * Starting a bonded port also starts all slaves under the bonded
> + * device. So if this port is bond device, we need to modify the
> + * port status of these slaves.
> + */
> + if (port->bond_flag == 1)
> + return change_bonding_slave_port_status(port_id, false);
> + }
>
> return 0;
> }
> @@ -609,8 +659,23 @@ eth_dev_start_mp(uint16_t port_id)
> static int
> eth_dev_stop_mp(uint16_t port_id)
> {
> - if (is_proc_primary())
> - return rte_eth_dev_stop(port_id);
> + int ret;
> +
> + if (is_proc_primary()) {
> + ret = rte_eth_dev_stop(port_id);
> + if (ret != 0)
> + return ret;
> +
> + struct rte_port *port = &ports[port_id];
> +
> + /*
> + * Stopping a bonded port also stops all slaves under the bonded
> + * device. So if this port is bond device, we need to modify the
> + * port status of these slaves.
> + */
> + if (port->bond_flag == 1)
> + return change_bonding_slave_port_status(port_id, true);
> + }
>
> return 0;
> }
> diff --git a/app/test-pmd/testpmd.h b/app/test-pmd/testpmd.h
> index 31f766c965..67f253b30e 100644
> --- a/app/test-pmd/testpmd.h
> +++ b/app/test-pmd/testpmd.h
> @@ -266,7 +266,8 @@ struct rte_port {
> uint32_t mc_addr_nb; /**< nb. of addr. in mc_addr_pool */
> queueid_t queue_nb; /**< nb. of queues for flow rules */
> uint32_t queue_sz; /**< size of a queue for flow rules */
> - uint8_t slave_flag; /**< bonding slave port */
> + uint8_t slave_flag : 1, /**< bonding slave port */
> + bond_flag : 1; /**< port is bond device */
> struct port_template *pattern_templ_list; /**< Pattern templates. */
> struct port_template *actions_templ_list; /**< Actions templates. */
> struct port_table *table_list; /**< Flow tables. */
Acked-by: Konstantin Ananyev <konstantin.v.ananyev@yandex.ru>
On 5/12/22 01:08, Konstantin Ananyev wrote:
>
>> From: Huisong Li <lihuisong@huawei.com>
>>
>> Starting or stopping a bonded port also starts or stops all active slaves
>> under the bonded port. If this port is a bonded device, we need to modify
>> the port status of all slaves.
>>
>> Fixes: 0e545d3047fe ("app/testpmd: check stopping port is not in
>> bonding")
>> Cc: stable@dpdk.org
>>
>> Signed-off-by: Huisong Li <lihuisong@huawei.com>
>> Signed-off-by: Min Hu (Connor) <humin29@huawei.com>
>> Acked-by: Aman Singh <aman.deep.singh@intel.com>
>
> Acked-by: Konstantin Ananyev <konstantin.v.ananyev@yandex.ru>
Applied, thanks.
On 5/3/2022 11:02 AM, Min Hu (Connor) wrote:
> Currently, 'dev_started' is always set to be 0 when dev stop, whether
> it succeeded or failed. This is unreasonable and this patch fixed it.
>
> Fixes: 62024eb82756 ("ethdev: change stop operation callback to return int")
> Cc: stable@dpdk.org
>
> Signed-off-by: Min Hu (Connor) <humin29@huawei.com>
> ---
> lib/ethdev/rte_ethdev.c | 3 ++-
> 1 file changed, 2 insertions(+), 1 deletion(-)
>
> diff --git a/lib/ethdev/rte_ethdev.c b/lib/ethdev/rte_ethdev.c
> index 29a3d80466..e0011372aa 100644
> --- a/lib/ethdev/rte_ethdev.c
> +++ b/lib/ethdev/rte_ethdev.c
> @@ -1533,8 +1533,9 @@ rte_eth_dev_stop(uint16_t port_id)
> /* point fast-path functions to dummy ones */
> eth_dev_fp_ops_reset(rte_eth_fp_ops + port_id);
>
> - dev->data->dev_started = 0;
> ret = (*dev->dev_ops->dev_stop)(dev);
> + if (ret == 0)
> + dev->data->dev_started = 0;
> rte_ethdev_trace_stop(port_id, ret);
>
> return ret;
Change looks good to me, I checked for possible unexpected side effect
but I did not see any.
@Andrew, @Thomas, if you also don't see/remember any issue related
change, I will push it soon.
25/05/2022 19:44, Ferruh Yigit:
> On 5/3/2022 11:02 AM, Min Hu (Connor) wrote:
> > Currently, 'dev_started' is always set to be 0 when dev stop, whether
> > it succeeded or failed. This is unreasonable and this patch fixed it.
> >
> > Fixes: 62024eb82756 ("ethdev: change stop operation callback to return int")
> > Cc: stable@dpdk.org
> >
> > Signed-off-by: Min Hu (Connor) <humin29@huawei.com>
> > ---
> > lib/ethdev/rte_ethdev.c | 3 ++-
> > 1 file changed, 2 insertions(+), 1 deletion(-)
> >
> > diff --git a/lib/ethdev/rte_ethdev.c b/lib/ethdev/rte_ethdev.c
> > index 29a3d80466..e0011372aa 100644
> > --- a/lib/ethdev/rte_ethdev.c
> > +++ b/lib/ethdev/rte_ethdev.c
> > @@ -1533,8 +1533,9 @@ rte_eth_dev_stop(uint16_t port_id)
> > /* point fast-path functions to dummy ones */
> > eth_dev_fp_ops_reset(rte_eth_fp_ops + port_id);
> >
> > - dev->data->dev_started = 0;
> > ret = (*dev->dev_ops->dev_stop)(dev);
> > + if (ret == 0)
> > + dev->data->dev_started = 0;
> > rte_ethdev_trace_stop(port_id, ret);
> >
> > return ret;
>
> Change looks good to me, I checked for possible unexpected side effect
> but I did not see any.
>
> @Andrew, @Thomas, if you also don't see/remember any issue related
> change, I will push it soon.
Acked-by: Thomas Monjalon <thomas@monjalon.net>
Hi, all,
any comments for this patch?
在 2022/3/24 11:00, Min Hu (Connor) 写道:
> From: Huisong Li <lihuisong@huawei.com>
>
> Currently, some eth devices are added to bond device, these devices are not
> released when the quit command is executed in testpmd. This patch adds the
> release operation for all active slaves under a bond device.
>
> Fixes: 0e545d3047fe ("app/testpmd: check stopping port is not in bonding")
> Cc: stable@dpdk.org
>
> Signed-off-by: Huisong Li <lihuisong@huawei.com>
> Signed-off-by: Min Hu (Connor) <humin29@huawei.com>
> ---
> app/test-pmd/cmdline.c | 1 +
> app/test-pmd/testpmd.c | 48 ++++++++++++++++++++++++++++++++++++++++++
> app/test-pmd/testpmd.h | 2 ++
> 3 files changed, 51 insertions(+)
>
> diff --git a/app/test-pmd/cmdline.c b/app/test-pmd/cmdline.c
> index f0efcf09f0..cd3873e1e0 100644
> --- a/app/test-pmd/cmdline.c
> +++ b/app/test-pmd/cmdline.c
> @@ -8891,6 +8891,7 @@ static void cmd_quit_parsed(__rte_unused void *parsed_result,
> __rte_unused void *data)
> {
> cmdline_quit(cl);
> + cl_quit = 1;
> }
>
> cmdline_parse_token_string_t cmd_quit_quit =
> diff --git a/app/test-pmd/testpmd.c b/app/test-pmd/testpmd.c
> index dc90600787..22700e073d 100644
> --- a/app/test-pmd/testpmd.c
> +++ b/app/test-pmd/testpmd.c
> @@ -229,6 +229,7 @@ unsigned int xstats_display_num; /**< Size of extended statistics to show */
> * option. Set flag to exit stats period loop after received SIGINT/SIGTERM.
> */
> uint8_t f_quit;
> +uint8_t cl_quit; /* Quit testpmd from cmdline. */
>
> /*
> * Max Rx frame size, set by '--max-pkt-len' parameter.
> @@ -3167,11 +3168,43 @@ remove_invalid_ports(void)
> nb_cfg_ports = nb_fwd_ports;
> }
>
> +#ifdef RTE_NET_BOND
> +static void
> +clear_bonding_slave_device(portid_t *slave_pids, uint16_t num_slaves)
> +{
> + struct rte_port *port;
> + portid_t slave_pid;
> + uint16_t i;
> +
> + for (i = 0; i < num_slaves; i++) {
> + slave_pid = slave_pids[i];
> + if (port_is_started(slave_pid) == 1) {
> + if (rte_eth_dev_stop(slave_pid) != 0)
> + fprintf(stderr, "rte_eth_dev_stop failed for port %u\n",
> + slave_pid);
> +
> + port = &ports[slave_pid];
> + port->port_status = RTE_PORT_STOPPED;
> + }
> +
> + clear_port_slave_flag(slave_pid);
> +
> + /* Close slave device when testpmd quit or is killed. */
> + if (cl_quit == 1 || f_quit == 1)
> + rte_eth_dev_close(slave_pid);
> + }
> +}
> +#endif
> +
> void
> close_port(portid_t pid)
> {
> portid_t pi;
> struct rte_port *port;
> +#ifdef RTE_NET_BOND
> + portid_t slave_pids[RTE_MAX_ETHPORTS];
> + int num_slaves = 0;
> +#endif
>
> if (port_id_is_invalid(pid, ENABLED_WARN))
> return;
> @@ -3205,7 +3238,22 @@ close_port(portid_t pid)
> if (is_proc_primary()) {
> port_flow_flush(pi);
> port_flex_item_flush(pi);
> +#ifdef RTE_NET_BOND
> + if (port->bond_flag == 1)
> + num_slaves = rte_eth_bond_slaves_get(pi,
> + slave_pids,
> + RTE_MAX_ETHPORTS);
> +#endif
> rte_eth_dev_close(pi);
> +#ifdef RTE_NET_BOND
> + /*
> + * If this port is bonded device, all slaves under the
> + * device need to be removed or closed.
> + */
> + if (port->bond_flag == 1 && num_slaves > 0)
> + clear_bonding_slave_device(slave_pids,
> + num_slaves);
> +#endif
> }
>
> free_xstats_display_info(pi);
> diff --git a/app/test-pmd/testpmd.h b/app/test-pmd/testpmd.h
> index 67f253b30e..5af9455012 100644
> --- a/app/test-pmd/testpmd.h
> +++ b/app/test-pmd/testpmd.h
> @@ -32,6 +32,8 @@
> #define RTE_PORT_CLOSED (uint16_t)2
> #define RTE_PORT_HANDLING (uint16_t)3
>
> +extern uint8_t cl_quit;
> +
> /*
> * It is used to allocate the memory for hash key.
> * The hash key size is NIC dependent.
>
Hi Connor, On 5/30/2022 11:31 AM, Min Hu (Connor) wrote: > Hi, all, > any comments for this patch? > > 在 2022/3/24 11:00, Min Hu (Connor) 写道: >> From: Huisong Li <lihuisong@huawei.com> >> >> Currently, some eth devices are added to bond device, these devices >> are not >> released when the quit command is executed in testpmd. This patch >> adds the >> release operation for all active slaves under a bond device. >> >> Fixes: 0e545d3047fe ("app/testpmd: check stopping port is not in >> bonding") >> Cc: stable@dpdk.org >> >> Signed-off-by: Huisong Li <lihuisong@huawei.com> >> Signed-off-by: Min Hu (Connor) <humin29@huawei.com> >> --- >> app/test-pmd/cmdline.c | 1 + >> app/test-pmd/testpmd.c | 48 ++++++++++++++++++++++++++++++++++++++++++ >> app/test-pmd/testpmd.h | 2 ++ >> 3 files changed, 51 insertions(+) >> >> diff --git a/app/test-pmd/cmdline.c b/app/test-pmd/cmdline.c >> index f0efcf09f0..cd3873e1e0 100644 >> --- a/app/test-pmd/cmdline.c >> +++ b/app/test-pmd/cmdline.c >> @@ -8891,6 +8891,7 @@ static void cmd_quit_parsed(__rte_unused void >> *parsed_result, >> __rte_unused void *data) >> { >> cmdline_quit(cl); >> + cl_quit = 1; >> } >> cmdline_parse_token_string_t cmd_quit_quit = >> diff --git a/app/test-pmd/testpmd.c b/app/test-pmd/testpmd.c >> index dc90600787..22700e073d 100644 >> --- a/app/test-pmd/testpmd.c >> +++ b/app/test-pmd/testpmd.c >> @@ -229,6 +229,7 @@ unsigned int xstats_display_num; /**< Size of >> extended statistics to show */ >> * option. Set flag to exit stats period loop after received >> SIGINT/SIGTERM. >> */ >> uint8_t f_quit; >> +uint8_t cl_quit; /* Quit testpmd from cmdline. */ >> /* >> * Max Rx frame size, set by '--max-pkt-len' parameter. >> @@ -3167,11 +3168,43 @@ remove_invalid_ports(void) >> nb_cfg_ports = nb_fwd_ports; >> } >> +#ifdef RTE_NET_BOND >> +static void >> +clear_bonding_slave_device(portid_t *slave_pids, uint16_t num_slaves) >> +{ >> + struct rte_port *port; >> + portid_t slave_pid; >> + uint16_t i; >> + >> + for (i = 0; i < num_slaves; i++) { >> + slave_pid = slave_pids[i]; >> + if (port_is_started(slave_pid) == 1) { >> + if (rte_eth_dev_stop(slave_pid) != 0) >> + fprintf(stderr, "rte_eth_dev_stop failed for port >> %u\n", >> + slave_pid); >> + >> + port = &ports[slave_pid]; >> + port->port_status = RTE_PORT_STOPPED; >> + } >> + >> + clear_port_slave_flag(slave_pid); >> + >> + /* Close slave device when testpmd quit or is killed. */ >> + if (cl_quit == 1 || f_quit == 1) >> + rte_eth_dev_close(slave_pid); When we close the bond device, shouldn't we close all slave ports also with it. Just to avoid usage of global flag "cl_quit" >> + } >> +} >> +#endif >> + >> void >> close_port(portid_t pid) >> { >> portid_t pi; >> struct rte_port *port; >> +#ifdef RTE_NET_BOND >> + portid_t slave_pids[RTE_MAX_ETHPORTS]; >> + int num_slaves = 0; >> +#endif >> if (port_id_is_invalid(pid, ENABLED_WARN)) >> return; >> @@ -3205,7 +3238,22 @@ close_port(portid_t pid) >> if (is_proc_primary()) { >> port_flow_flush(pi); >> port_flex_item_flush(pi); >> +#ifdef RTE_NET_BOND >> + if (port->bond_flag == 1) >> + num_slaves = rte_eth_bond_slaves_get(pi, >> + slave_pids, >> + RTE_MAX_ETHPORTS); >> +#endif >> rte_eth_dev_close(pi); >> +#ifdef RTE_NET_BOND >> + /* >> + * If this port is bonded device, all slaves under the >> + * device need to be removed or closed. >> + */ >> + if (port->bond_flag == 1 && num_slaves > 0) >> + clear_bonding_slave_device(slave_pids, >> + num_slaves); >> +#endif Can we merge these two #ifdef sections, like move _close() below. >> } >> free_xstats_display_info(pi); >> diff --git a/app/test-pmd/testpmd.h b/app/test-pmd/testpmd.h >> index 67f253b30e..5af9455012 100644 >> --- a/app/test-pmd/testpmd.h >> +++ b/app/test-pmd/testpmd.h >> @@ -32,6 +32,8 @@ >> #define RTE_PORT_CLOSED (uint16_t)2 >> #define RTE_PORT_HANDLING (uint16_t)3 >> +extern uint8_t cl_quit; >> + >> /* >> * It is used to allocate the memory for hash key. >> * The hash key size is NIC dependent. >>
On 5/26/2022 11:21 AM, Thomas Monjalon wrote:
> [CAUTION: External Email]
>
> 25/05/2022 19:44, Ferruh Yigit:
>> On 5/3/2022 11:02 AM, Min Hu (Connor) wrote:
>>> Currently, 'dev_started' is always set to be 0 when dev stop, whether
>>> it succeeded or failed. This is unreasonable and this patch fixed it.
>>>
>>> Fixes: 62024eb82756 ("ethdev: change stop operation callback to return int")
>>> Cc: stable@dpdk.org
>>>
>>> Signed-off-by: Min Hu (Connor) <humin29@huawei.com>
>>> ---
>>> lib/ethdev/rte_ethdev.c | 3 ++-
>>> 1 file changed, 2 insertions(+), 1 deletion(-)
>>>
>>> diff --git a/lib/ethdev/rte_ethdev.c b/lib/ethdev/rte_ethdev.c
>>> index 29a3d80466..e0011372aa 100644
>>> --- a/lib/ethdev/rte_ethdev.c
>>> +++ b/lib/ethdev/rte_ethdev.c
>>> @@ -1533,8 +1533,9 @@ rte_eth_dev_stop(uint16_t port_id)
>>> /* point fast-path functions to dummy ones */
>>> eth_dev_fp_ops_reset(rte_eth_fp_ops + port_id);
>>>
>>> - dev->data->dev_started = 0;
>>> ret = (*dev->dev_ops->dev_stop)(dev);
>>> + if (ret == 0)
>>> + dev->data->dev_started = 0;
>>> rte_ethdev_trace_stop(port_id, ret);
>>>
>>> return ret;
>>
>> Change looks good to me, I checked for possible unexpected side effect
>> but I did not see any.
>>
>> @Andrew, @Thomas, if you also don't see/remember any issue related
>> change, I will push it soon.
>
> Acked-by: Thomas Monjalon <thomas@monjalon.net>
>
Acked-by: Ferruh Yigit <ferruh.yigit@xilinx.com>
Applied to dpdk-next-net/main, thanks.
On 5/3/2022 11:02 AM, Min Hu (Connor) wrote:
> From: Huisong Li<lihuisong@huawei.com>
>
> Currently, some eth devices are added to bond device, these devices are not
> released when the quit command is executed in testpmd. This patch adds the
> release operation for all active slaves under a bond device.
>
> Fixes: 0e545d3047fe ("app/testpmd: check stopping port is not in bonding")
> Cc:stable@dpdk.org
>
> Signed-off-by: Huisong Li<lihuisong@huawei.com>
> Signed-off-by: Min Hu (Connor)<humin29@huawei.com>
Hi Connor,
Can you please do the similar #ifdef grouping done in 3/5 patch?
From: Huisong Li <lihuisong@huawei.com> Currently, some eth devices are added to bond device, these devices are not released when the quit command is executed in testpmd. This patch adds the release operation for all active slaves under a bond device. Fixes: 0e545d3047fe ("app/testpmd: check stopping port is not in bonding") Cc: stable@dpdk.org Signed-off-by: Huisong Li <lihuisong@huawei.com> Signed-off-by: Min Hu (Connor) <humin29@huawei.com> Signed-off-by: Dongdong Liu <liudongdong3@huawei.com> --- app/test-pmd/cmdline.c | 1 + app/test-pmd/testpmd.c | 46 ++++++++++++++++++++++++++++++++++++++++++ app/test-pmd/testpmd.h | 2 ++ 3 files changed, 49 insertions(+) diff --git a/app/test-pmd/cmdline.c b/app/test-pmd/cmdline.c index fdd0cada3b..6c58b02650 100644 --- a/app/test-pmd/cmdline.c +++ b/app/test-pmd/cmdline.c @@ -8915,6 +8915,7 @@ static void cmd_quit_parsed(__rte_unused void *parsed_result, __rte_unused void *data) { cmdline_quit(cl); + cl_quit = 1; } static cmdline_parse_token_string_t cmd_quit_quit = diff --git a/app/test-pmd/testpmd.c b/app/test-pmd/testpmd.c index fc1b64b60d..2b16551a26 100644 --- a/app/test-pmd/testpmd.c +++ b/app/test-pmd/testpmd.c @@ -229,6 +229,7 @@ unsigned int xstats_display_num; /**< Size of extended statistics to show */ * option. Set flag to exit stats period loop after received SIGINT/SIGTERM. */ uint8_t f_quit; +uint8_t cl_quit; /* Quit testpmd from cmdline. */ /* * Max Rx frame size, set by '--max-pkt-len' parameter. @@ -3201,11 +3202,44 @@ remove_invalid_ports(void) nb_cfg_ports = nb_fwd_ports; } +static void +clear_bonding_slave_device(portid_t *slave_pids, uint16_t num_slaves) +{ +#ifdef RTE_NET_BOND + struct rte_port *port; + portid_t slave_pid; + uint16_t i; + + for (i = 0; i < num_slaves; i++) { + slave_pid = slave_pids[i]; + if (port_is_started(slave_pid) == 1) { + if (rte_eth_dev_stop(slave_pid) != 0) + fprintf(stderr, "rte_eth_dev_stop failed for port %u\n", + slave_pid); + + port = &ports[slave_pid]; + port->port_status = RTE_PORT_STOPPED; + } + + clear_port_slave_flag(slave_pid); + + /* Close slave device when testpmd quit or is killed. */ + if (cl_quit == 1 || f_quit == 1) + rte_eth_dev_close(slave_pid); + } +#else + RTE_SET_USED(slave_pids); + RTE_SET_USED(num_slaves); +#endif +} + void close_port(portid_t pid) { portid_t pi; struct rte_port *port; + portid_t slave_pids[RTE_MAX_ETHPORTS]; + int num_slaves = 0; if (port_id_is_invalid(pid, ENABLED_WARN)) return; @@ -3240,7 +3274,19 @@ close_port(portid_t pid) port_flow_flush(pi); port_flex_item_flush(pi); port_action_handle_flush(pi); +#ifdef RTE_NET_BOND + if (port->bond_flag == 1) + num_slaves = rte_eth_bond_slaves_get(pi, + slave_pids, RTE_MAX_ETHPORTS); +#endif rte_eth_dev_close(pi); + /* + * If this port is bonded device, all slaves under the + * device need to be removed or closed. + */ + if (port->bond_flag == 1 && num_slaves > 0) + clear_bonding_slave_device(slave_pids, + num_slaves); } free_xstats_display_info(pi); diff --git a/app/test-pmd/testpmd.h b/app/test-pmd/testpmd.h index 6693813dda..b44d5dd5ac 100644 --- a/app/test-pmd/testpmd.h +++ b/app/test-pmd/testpmd.h @@ -38,6 +38,8 @@ #define RTE_PORT_CLOSED (uint16_t)2 #define RTE_PORT_HANDLING (uint16_t)3 +extern uint8_t cl_quit; + /* * It is used to allocate the memory for hash key. * The hash key size is NIC dependent. -- 2.33.0
Hi Ferruh
"[PATCH v4] app/testpmd: fix slave device isn't released" has been sent
out, help to review.
Thanks,
Dongdong
On 2022/6/2 1:54, Ferruh Yigit wrote:
> On 5/3/2022 11:02 AM, Min Hu (Connor) wrote:
>> From: Huisong Li<lihuisong@huawei.com>
>>
>> Currently, some eth devices are added to bond device, these devices
>> are not
>> released when the quit command is executed in testpmd. This patch adds
>> the
>> release operation for all active slaves under a bond device.
>>
>> Fixes: 0e545d3047fe ("app/testpmd: check stopping port is not in
>> bonding")
>> Cc:stable@dpdk.org
>>
>> Signed-off-by: Huisong Li<lihuisong@huawei.com>
>> Signed-off-by: Min Hu (Connor)<humin29@huawei.com>
>
> Hi Connor,
>
> Can you please do the similar #ifdef grouping done in 3/5 patch?
> .
>
On 6/7/2022 9:10 AM, Dongdong Liu wrote: > From: Huisong Li <lihuisong@huawei.com> > > Currently, some eth devices are added to bond device, these devices are not > released when the quit command is executed in testpmd. This patch adds the > release operation for all active slaves under a bond device. > > Fixes: 0e545d3047fe ("app/testpmd: check stopping port is not in bonding") > Cc: stable@dpdk.org > > Signed-off-by: Huisong Li <lihuisong@huawei.com> > Signed-off-by: Min Hu (Connor) <humin29@huawei.com> > Signed-off-by: Dongdong Liu <liudongdong3@huawei.com> > --- > app/test-pmd/cmdline.c | 1 + > app/test-pmd/testpmd.c | 46 ++++++++++++++++++++++++++++++++++++++++++ > app/test-pmd/testpmd.h | 2 ++ > 3 files changed, 49 insertions(+) > > diff --git a/app/test-pmd/cmdline.c b/app/test-pmd/cmdline.c > index fdd0cada3b..6c58b02650 100644 > --- a/app/test-pmd/cmdline.c > +++ b/app/test-pmd/cmdline.c > @@ -8915,6 +8915,7 @@ static void cmd_quit_parsed(__rte_unused void *parsed_result, > __rte_unused void *data) > { > cmdline_quit(cl); > + cl_quit = 1; > } > > static cmdline_parse_token_string_t cmd_quit_quit = > diff --git a/app/test-pmd/testpmd.c b/app/test-pmd/testpmd.c > index fc1b64b60d..2b16551a26 100644 > --- a/app/test-pmd/testpmd.c > +++ b/app/test-pmd/testpmd.c > @@ -229,6 +229,7 @@ unsigned int xstats_display_num; /**< Size of extended statistics to show */ > * option. Set flag to exit stats period loop after received SIGINT/SIGTERM. > */ > uint8_t f_quit; > +uint8_t cl_quit; /* Quit testpmd from cmdline. */ > > /* > * Max Rx frame size, set by '--max-pkt-len' parameter. > @@ -3201,11 +3202,44 @@ remove_invalid_ports(void) > nb_cfg_ports = nb_fwd_ports; > } > > +static void > +clear_bonding_slave_device(portid_t *slave_pids, uint16_t num_slaves) > +{ > +#ifdef RTE_NET_BOND > + struct rte_port *port; > + portid_t slave_pid; > + uint16_t i; > + > + for (i = 0; i < num_slaves; i++) { > + slave_pid = slave_pids[i]; > + if (port_is_started(slave_pid) == 1) { > + if (rte_eth_dev_stop(slave_pid) != 0) > + fprintf(stderr, "rte_eth_dev_stop failed for port %u\n", > + slave_pid); > + > + port = &ports[slave_pid]; > + port->port_status = RTE_PORT_STOPPED; > + } > + > + clear_port_slave_flag(slave_pid); > + > + /* Close slave device when testpmd quit or is killed. */ > + if (cl_quit == 1 || f_quit == 1) > + rte_eth_dev_close(slave_pid); > + } > +#else > + RTE_SET_USED(slave_pids); > + RTE_SET_USED(num_slaves); > +#endif Do we need this #ifdef, everything is already available in compile time. > +} > + > void > close_port(portid_t pid) > { > portid_t pi; > struct rte_port *port; > + portid_t slave_pids[RTE_MAX_ETHPORTS]; > + int num_slaves = 0; > > if (port_id_is_invalid(pid, ENABLED_WARN)) > return; > @@ -3240,7 +3274,19 @@ close_port(portid_t pid) > port_flow_flush(pi); > port_flex_item_flush(pi); > port_action_handle_flush(pi); > +#ifdef RTE_NET_BOND > + if (port->bond_flag == 1) > + num_slaves = rte_eth_bond_slaves_get(pi, > + slave_pids, RTE_MAX_ETHPORTS); > +#endif > rte_eth_dev_close(pi); > + /* > + * If this port is bonded device, all slaves under the > + * device need to be removed or closed. > + */ > + if (port->bond_flag == 1 && num_slaves > 0) > + clear_bonding_slave_device(slave_pids, > + num_slaves); > } > > free_xstats_display_info(pi); > diff --git a/app/test-pmd/testpmd.h b/app/test-pmd/testpmd.h > index 6693813dda..b44d5dd5ac 100644 > --- a/app/test-pmd/testpmd.h > +++ b/app/test-pmd/testpmd.h > @@ -38,6 +38,8 @@ > #define RTE_PORT_CLOSED (uint16_t)2 > #define RTE_PORT_HANDLING (uint16_t)3 > > +extern uint8_t cl_quit; > + > /* > * It is used to allocate the memory for hash key. > * The hash key size is NIC dependent.
Hi Ferruh Many thanks for your review. On 2022/6/7 22:31, Ferruh Yigit wrote: > On 6/7/2022 9:10 AM, Dongdong Liu wrote: >> From: Huisong Li <lihuisong@huawei.com> >> >> Currently, some eth devices are added to bond device, these devices >> are not >> released when the quit command is executed in testpmd. This patch adds >> the >> release operation for all active slaves under a bond device. >> >> Fixes: 0e545d3047fe ("app/testpmd: check stopping port is not in >> bonding") >> Cc: stable@dpdk.org >> >> Signed-off-by: Huisong Li <lihuisong@huawei.com> >> Signed-off-by: Min Hu (Connor) <humin29@huawei.com> >> Signed-off-by: Dongdong Liu <liudongdong3@huawei.com> >> --- >> app/test-pmd/cmdline.c | 1 + >> app/test-pmd/testpmd.c | 46 ++++++++++++++++++++++++++++++++++++++++++ >> app/test-pmd/testpmd.h | 2 ++ >> 3 files changed, 49 insertions(+) >> >> diff --git a/app/test-pmd/cmdline.c b/app/test-pmd/cmdline.c >> index fdd0cada3b..6c58b02650 100644 >> --- a/app/test-pmd/cmdline.c >> +++ b/app/test-pmd/cmdline.c >> @@ -8915,6 +8915,7 @@ static void cmd_quit_parsed(__rte_unused void >> *parsed_result, >> __rte_unused void *data) >> { >> cmdline_quit(cl); >> + cl_quit = 1; >> } >> static cmdline_parse_token_string_t cmd_quit_quit = >> diff --git a/app/test-pmd/testpmd.c b/app/test-pmd/testpmd.c >> index fc1b64b60d..2b16551a26 100644 >> --- a/app/test-pmd/testpmd.c >> +++ b/app/test-pmd/testpmd.c >> @@ -229,6 +229,7 @@ unsigned int xstats_display_num; /**< Size of >> extended statistics to show */ >> * option. Set flag to exit stats period loop after received >> SIGINT/SIGTERM. >> */ >> uint8_t f_quit; >> +uint8_t cl_quit; /* Quit testpmd from cmdline. */ >> /* >> * Max Rx frame size, set by '--max-pkt-len' parameter. >> @@ -3201,11 +3202,44 @@ remove_invalid_ports(void) >> nb_cfg_ports = nb_fwd_ports; >> } >> +static void >> +clear_bonding_slave_device(portid_t *slave_pids, uint16_t num_slaves) >> +{ >> +#ifdef RTE_NET_BOND >> + struct rte_port *port; >> + portid_t slave_pid; >> + uint16_t i; >> + >> + for (i = 0; i < num_slaves; i++) { >> + slave_pid = slave_pids[i]; >> + if (port_is_started(slave_pid) == 1) { >> + if (rte_eth_dev_stop(slave_pid) != 0) >> + fprintf(stderr, "rte_eth_dev_stop failed for port %u\n", >> + slave_pid); >> + >> + port = &ports[slave_pid]; >> + port->port_status = RTE_PORT_STOPPED; >> + } >> + >> + clear_port_slave_flag(slave_pid); >> + >> + /* Close slave device when testpmd quit or is killed. */ >> + if (cl_quit == 1 || f_quit == 1) >> + rte_eth_dev_close(slave_pid); >> + } >> +#else >> + RTE_SET_USED(slave_pids); >> + RTE_SET_USED(num_slaves); >> +#endif > > Do we need this #ifdef, everything is already available in compile time. Although it can be compiled ok without this #ifdef, I think we still need this #ifdef as this is bond related implementation. This patch does the similar #ifdef grouping done in 3/5 patch. Thanks, Dongdong > >> +} >> + >> void >> close_port(portid_t pid) >> { >> portid_t pi; >> struct rte_port *port; >> + portid_t slave_pids[RTE_MAX_ETHPORTS]; >> + int num_slaves = 0; >> if (port_id_is_invalid(pid, ENABLED_WARN)) >> return; >> @@ -3240,7 +3274,19 @@ close_port(portid_t pid) >> port_flow_flush(pi); >> port_flex_item_flush(pi); >> port_action_handle_flush(pi); >> +#ifdef RTE_NET_BOND >> + if (port->bond_flag == 1) >> + num_slaves = rte_eth_bond_slaves_get(pi, >> + slave_pids, RTE_MAX_ETHPORTS); >> +#endif >> rte_eth_dev_close(pi); >> + /* >> + * If this port is bonded device, all slaves under the >> + * device need to be removed or closed. >> + */ >> + if (port->bond_flag == 1 && num_slaves > 0) >> + clear_bonding_slave_device(slave_pids, >> + num_slaves); >> } >> free_xstats_display_info(pi); >> diff --git a/app/test-pmd/testpmd.h b/app/test-pmd/testpmd.h >> index 6693813dda..b44d5dd5ac 100644 >> --- a/app/test-pmd/testpmd.h >> +++ b/app/test-pmd/testpmd.h >> @@ -38,6 +38,8 @@ >> #define RTE_PORT_CLOSED (uint16_t)2 >> #define RTE_PORT_HANDLING (uint16_t)3 >> +extern uint8_t cl_quit; >> + >> /* >> * It is used to allocate the memory for hash key. >> * The hash key size is NIC dependent. > > . >
On 6/9/2022 8:50 AM, Dongdong Liu wrote: > Hi Ferruh > > Many thanks for your review. > On 2022/6/7 22:31, Ferruh Yigit wrote: >> On 6/7/2022 9:10 AM, Dongdong Liu wrote: >>> From: Huisong Li <lihuisong@huawei.com> >>> >>> Currently, some eth devices are added to bond device, these devices >>> are not >>> released when the quit command is executed in testpmd. This patch adds >>> the >>> release operation for all active slaves under a bond device. >>> >>> Fixes: 0e545d3047fe ("app/testpmd: check stopping port is not in >>> bonding") >>> Cc: stable@dpdk.org >>> >>> Signed-off-by: Huisong Li <lihuisong@huawei.com> >>> Signed-off-by: Min Hu (Connor) <humin29@huawei.com> >>> Signed-off-by: Dongdong Liu <liudongdong3@huawei.com> >>> --- >>> app/test-pmd/cmdline.c | 1 + >>> app/test-pmd/testpmd.c | 46 ++++++++++++++++++++++++++++++++++++++++++ >>> app/test-pmd/testpmd.h | 2 ++ >>> 3 files changed, 49 insertions(+) >>> >>> diff --git a/app/test-pmd/cmdline.c b/app/test-pmd/cmdline.c >>> index fdd0cada3b..6c58b02650 100644 >>> --- a/app/test-pmd/cmdline.c >>> +++ b/app/test-pmd/cmdline.c >>> @@ -8915,6 +8915,7 @@ static void cmd_quit_parsed(__rte_unused void >>> *parsed_result, >>> __rte_unused void *data) >>> { >>> cmdline_quit(cl); >>> + cl_quit = 1; >>> } >>> static cmdline_parse_token_string_t cmd_quit_quit = >>> diff --git a/app/test-pmd/testpmd.c b/app/test-pmd/testpmd.c >>> index fc1b64b60d..2b16551a26 100644 >>> --- a/app/test-pmd/testpmd.c >>> +++ b/app/test-pmd/testpmd.c >>> @@ -229,6 +229,7 @@ unsigned int xstats_display_num; /**< Size of >>> extended statistics to show */ >>> * option. Set flag to exit stats period loop after received >>> SIGINT/SIGTERM. >>> */ >>> uint8_t f_quit; >>> +uint8_t cl_quit; /* Quit testpmd from cmdline. */ >>> /* >>> * Max Rx frame size, set by '--max-pkt-len' parameter. >>> @@ -3201,11 +3202,44 @@ remove_invalid_ports(void) >>> nb_cfg_ports = nb_fwd_ports; >>> } >>> +static void >>> +clear_bonding_slave_device(portid_t *slave_pids, uint16_t num_slaves) >>> +{ >>> +#ifdef RTE_NET_BOND >>> + struct rte_port *port; >>> + portid_t slave_pid; >>> + uint16_t i; >>> + >>> + for (i = 0; i < num_slaves; i++) { >>> + slave_pid = slave_pids[i]; >>> + if (port_is_started(slave_pid) == 1) { >>> + if (rte_eth_dev_stop(slave_pid) != 0) >>> + fprintf(stderr, "rte_eth_dev_stop failed for port >>> %u\n", >>> + slave_pid); >>> + >>> + port = &ports[slave_pid]; >>> + port->port_status = RTE_PORT_STOPPED; >>> + } >>> + >>> + clear_port_slave_flag(slave_pid); >>> + >>> + /* Close slave device when testpmd quit or is killed. */ >>> + if (cl_quit == 1 || f_quit == 1) >>> + rte_eth_dev_close(slave_pid); >>> + } >>> +#else >>> + RTE_SET_USED(slave_pids); >>> + RTE_SET_USED(num_slaves); >>> +#endif >> >> Do we need this #ifdef, everything is already available in compile time. > > Although it can be compiled ok without this #ifdef, > I think we still need this #ifdef as this is bond related > implementation. > Code will compile and function as expected even if bonding PMD is disable, right? If so why #ifdef is needed. > This patch does the similar #ifdef grouping done in 3/5 patch. > > Thanks, > Dongdong >> >>> +} >>> + >>> void >>> close_port(portid_t pid) >>> { >>> portid_t pi; >>> struct rte_port *port; >>> + portid_t slave_pids[RTE_MAX_ETHPORTS]; >>> + int num_slaves = 0; >>> if (port_id_is_invalid(pid, ENABLED_WARN)) >>> return; >>> @@ -3240,7 +3274,19 @@ close_port(portid_t pid) >>> port_flow_flush(pi); >>> port_flex_item_flush(pi); >>> port_action_handle_flush(pi); >>> +#ifdef RTE_NET_BOND >>> + if (port->bond_flag == 1) >>> + num_slaves = rte_eth_bond_slaves_get(pi, >>> + slave_pids, RTE_MAX_ETHPORTS); >>> +#endif >>> rte_eth_dev_close(pi); >>> + /* >>> + * If this port is bonded device, all slaves under the >>> + * device need to be removed or closed. >>> + */ >>> + if (port->bond_flag == 1 && num_slaves > 0) >>> + clear_bonding_slave_device(slave_pids, >>> + num_slaves); >>> } >>> free_xstats_display_info(pi); >>> diff --git a/app/test-pmd/testpmd.h b/app/test-pmd/testpmd.h >>> index 6693813dda..b44d5dd5ac 100644 >>> --- a/app/test-pmd/testpmd.h >>> +++ b/app/test-pmd/testpmd.h >>> @@ -38,6 +38,8 @@ >>> #define RTE_PORT_CLOSED (uint16_t)2 >>> #define RTE_PORT_HANDLING (uint16_t)3 >>> +extern uint8_t cl_quit; >>> + >>> /* >>> * It is used to allocate the memory for hash key. >>> * The hash key size is NIC dependent. >> >> . >>
Hi Ferruh Many thanks for your review. On 2022/6/9 16:50, Ferruh Yigit wrote: > On 6/9/2022 8:50 AM, Dongdong Liu wrote: >> Hi Ferruh >> >> Many thanks for your review. >> On 2022/6/7 22:31, Ferruh Yigit wrote: >>> On 6/7/2022 9:10 AM, Dongdong Liu wrote: >>>> From: Huisong Li <lihuisong@huawei.com> >>>> >>>> Currently, some eth devices are added to bond device, these devices >>>> are not >>>> released when the quit command is executed in testpmd. This patch adds >>>> the >>>> release operation for all active slaves under a bond device. >>>> >>>> Fixes: 0e545d3047fe ("app/testpmd: check stopping port is not in >>>> bonding") >>>> Cc: stable@dpdk.org >>>> >>>> Signed-off-by: Huisong Li <lihuisong@huawei.com> >>>> Signed-off-by: Min Hu (Connor) <humin29@huawei.com> >>>> Signed-off-by: Dongdong Liu <liudongdong3@huawei.com> >>>> --- >>>> app/test-pmd/cmdline.c | 1 + >>>> app/test-pmd/testpmd.c | 46 >>>> ++++++++++++++++++++++++++++++++++++++++++ >>>> app/test-pmd/testpmd.h | 2 ++ >>>> 3 files changed, 49 insertions(+) >>>> >>>> diff --git a/app/test-pmd/cmdline.c b/app/test-pmd/cmdline.c >>>> index fdd0cada3b..6c58b02650 100644 >>>> --- a/app/test-pmd/cmdline.c >>>> +++ b/app/test-pmd/cmdline.c >>>> @@ -8915,6 +8915,7 @@ static void cmd_quit_parsed(__rte_unused void >>>> *parsed_result, >>>> __rte_unused void *data) >>>> { >>>> cmdline_quit(cl); >>>> + cl_quit = 1; >>>> } >>>> static cmdline_parse_token_string_t cmd_quit_quit = >>>> diff --git a/app/test-pmd/testpmd.c b/app/test-pmd/testpmd.c >>>> index fc1b64b60d..2b16551a26 100644 >>>> --- a/app/test-pmd/testpmd.c >>>> +++ b/app/test-pmd/testpmd.c >>>> @@ -229,6 +229,7 @@ unsigned int xstats_display_num; /**< Size of >>>> extended statistics to show */ >>>> * option. Set flag to exit stats period loop after received >>>> SIGINT/SIGTERM. >>>> */ >>>> uint8_t f_quit; >>>> +uint8_t cl_quit; /* Quit testpmd from cmdline. */ >>>> /* >>>> * Max Rx frame size, set by '--max-pkt-len' parameter. >>>> @@ -3201,11 +3202,44 @@ remove_invalid_ports(void) >>>> nb_cfg_ports = nb_fwd_ports; >>>> } >>>> +static void >>>> +clear_bonding_slave_device(portid_t *slave_pids, uint16_t num_slaves) >>>> +{ >>>> +#ifdef RTE_NET_BOND >>>> + struct rte_port *port; >>>> + portid_t slave_pid; >>>> + uint16_t i; >>>> + >>>> + for (i = 0; i < num_slaves; i++) { >>>> + slave_pid = slave_pids[i]; >>>> + if (port_is_started(slave_pid) == 1) { >>>> + if (rte_eth_dev_stop(slave_pid) != 0) >>>> + fprintf(stderr, "rte_eth_dev_stop failed for port >>>> %u\n", >>>> + slave_pid); >>>> + >>>> + port = &ports[slave_pid]; >>>> + port->port_status = RTE_PORT_STOPPED; >>>> + } >>>> + >>>> + clear_port_slave_flag(slave_pid); >>>> + >>>> + /* Close slave device when testpmd quit or is killed. */ >>>> + if (cl_quit == 1 || f_quit == 1) >>>> + rte_eth_dev_close(slave_pid); >>>> + } >>>> +#else >>>> + RTE_SET_USED(slave_pids); >>>> + RTE_SET_USED(num_slaves); >>>> +#endif >>> >>> Do we need this #ifdef, everything is already available in compile time. >> >> Although it can be compiled ok without this #ifdef, >> I think we still need this #ifdef as this is bond related >> implementation. >> > > Code will compile and function as expected even if bonding PMD is > disable, right? If so why #ifdef is needed. Yes, will delete. Thanks, Dongdong > >> This patch does the similar #ifdef grouping done in 3/5 patch. >> >> Thanks, >> Dongdong >>> >>>> +} >>>> + >>>> void >>>> close_port(portid_t pid) >>>> { >>>> portid_t pi; >>>> struct rte_port *port; >>>> + portid_t slave_pids[RTE_MAX_ETHPORTS]; >>>> + int num_slaves = 0; >>>> if (port_id_is_invalid(pid, ENABLED_WARN)) >>>> return; >>>> @@ -3240,7 +3274,19 @@ close_port(portid_t pid) >>>> port_flow_flush(pi); >>>> port_flex_item_flush(pi); >>>> port_action_handle_flush(pi); >>>> +#ifdef RTE_NET_BOND >>>> + if (port->bond_flag == 1) >>>> + num_slaves = rte_eth_bond_slaves_get(pi, >>>> + slave_pids, RTE_MAX_ETHPORTS); >>>> +#endif >>>> rte_eth_dev_close(pi); >>>> + /* >>>> + * If this port is bonded device, all slaves under the >>>> + * device need to be removed or closed. >>>> + */ >>>> + if (port->bond_flag == 1 && num_slaves > 0) >>>> + clear_bonding_slave_device(slave_pids, >>>> + num_slaves); >>>> } >>>> free_xstats_display_info(pi); >>>> diff --git a/app/test-pmd/testpmd.h b/app/test-pmd/testpmd.h >>>> index 6693813dda..b44d5dd5ac 100644 >>>> --- a/app/test-pmd/testpmd.h >>>> +++ b/app/test-pmd/testpmd.h >>>> @@ -38,6 +38,8 @@ >>>> #define RTE_PORT_CLOSED (uint16_t)2 >>>> #define RTE_PORT_HANDLING (uint16_t)3 >>>> +extern uint8_t cl_quit; >>>> + >>>> /* >>>> * It is used to allocate the memory for hash key. >>>> * The hash key size is NIC dependent. >>> >>> . >>> > > . >
From: Huisong Li <lihuisong@huawei.com> Currently, some eth devices are added to bond device, these devices are not released when the quit command is executed in testpmd. This patch adds the release operation for all active slaves under a bond device. Fixes: 0e545d3047fe ("app/testpmd: check stopping port is not in bonding") Cc: stable@dpdk.org Signed-off-by: Huisong Li <lihuisong@huawei.com> Signed-off-by: Min Hu (Connor) <humin29@huawei.com> Signed-off-by: Dongdong Liu <liudongdong3@huawei.com> --- v4->v5: - Delete the no needed #ifdef --- app/test-pmd/cmdline.c | 1 + app/test-pmd/testpmd.c | 41 +++++++++++++++++++++++++++++++++++++++++ app/test-pmd/testpmd.h | 2 ++ 3 files changed, 44 insertions(+) diff --git a/app/test-pmd/cmdline.c b/app/test-pmd/cmdline.c index fdd0cada3b..6c58b02650 100644 --- a/app/test-pmd/cmdline.c +++ b/app/test-pmd/cmdline.c @@ -8915,6 +8915,7 @@ static void cmd_quit_parsed(__rte_unused void *parsed_result, __rte_unused void *data) { cmdline_quit(cl); + cl_quit = 1; } static cmdline_parse_token_string_t cmd_quit_quit = diff --git a/app/test-pmd/testpmd.c b/app/test-pmd/testpmd.c index fc1b64b60d..8bcb488d33 100644 --- a/app/test-pmd/testpmd.c +++ b/app/test-pmd/testpmd.c @@ -229,6 +229,7 @@ unsigned int xstats_display_num; /**< Size of extended statistics to show */ * option. Set flag to exit stats period loop after received SIGINT/SIGTERM. */ uint8_t f_quit; +uint8_t cl_quit; /* Quit testpmd from cmdline. */ /* * Max Rx frame size, set by '--max-pkt-len' parameter. @@ -3201,11 +3202,39 @@ remove_invalid_ports(void) nb_cfg_ports = nb_fwd_ports; } +static void +clear_bonding_slave_device(portid_t *slave_pids, uint16_t num_slaves) +{ + struct rte_port *port; + portid_t slave_pid; + uint16_t i; + + for (i = 0; i < num_slaves; i++) { + slave_pid = slave_pids[i]; + if (port_is_started(slave_pid) == 1) { + if (rte_eth_dev_stop(slave_pid) != 0) + fprintf(stderr, "rte_eth_dev_stop failed for port %u\n", + slave_pid); + + port = &ports[slave_pid]; + port->port_status = RTE_PORT_STOPPED; + } + + clear_port_slave_flag(slave_pid); + + /* Close slave device when testpmd quit or is killed. */ + if (cl_quit == 1 || f_quit == 1) + rte_eth_dev_close(slave_pid); + } +} + void close_port(portid_t pid) { portid_t pi; struct rte_port *port; + portid_t slave_pids[RTE_MAX_ETHPORTS]; + int num_slaves = 0; if (port_id_is_invalid(pid, ENABLED_WARN)) return; @@ -3240,7 +3269,19 @@ close_port(portid_t pid) port_flow_flush(pi); port_flex_item_flush(pi); port_action_handle_flush(pi); +#ifdef RTE_NET_BOND + if (port->bond_flag == 1) + num_slaves = rte_eth_bond_slaves_get(pi, + slave_pids, RTE_MAX_ETHPORTS); +#endif rte_eth_dev_close(pi); + /* + * If this port is bonded device, all slaves under the + * device need to be removed or closed. + */ + if (port->bond_flag == 1 && num_slaves > 0) + clear_bonding_slave_device(slave_pids, + num_slaves); } free_xstats_display_info(pi); diff --git a/app/test-pmd/testpmd.h b/app/test-pmd/testpmd.h index 6693813dda..b44d5dd5ac 100644 --- a/app/test-pmd/testpmd.h +++ b/app/test-pmd/testpmd.h @@ -38,6 +38,8 @@ #define RTE_PORT_CLOSED (uint16_t)2 #define RTE_PORT_HANDLING (uint16_t)3 +extern uint8_t cl_quit; + /* * It is used to allocate the memory for hash key. * The hash key size is NIC dependent. -- 2.33.0
On 6/9/2022 12:49 PM, Dongdong Liu wrote:
> From: Huisong Li<lihuisong@huawei.com>
>
> Currently, some eth devices are added to bond device, these devices are not
> released when the quit command is executed in testpmd. This patch adds the
> release operation for all active slaves under a bond device.
>
> Fixes: 0e545d3047fe ("app/testpmd: check stopping port is not in bonding")
> Cc:stable@dpdk.org
>
> Signed-off-by: Huisong Li<lihuisong@huawei.com>
> Signed-off-by: Min Hu (Connor)<humin29@huawei.com>
> Signed-off-by: Dongdong Liu<liudongdong3@huawei.com>
Acked-by: Ferruh Yigit <ferruh.yigit@xilinx.com>
Applied to dpdk-next-net/main, thanks.