Fix issue: wrong teamd link watch state after warm reboot (#14084)

#### Why I did it

Fix issue: wrong teamd link watch state after warm reboot due to TEAM_ATTR_PORT_CHANGED lost

The flag TEAM_ATTR_PORT_CHANGED is maintained by kernel team driver:
- a flag "changed" is maintained in struct team_port struct
- the flag is set by __team_port_change_send once relevant information is updated, including port linkup (together with speed, duplex), adding or removing
- the flag is cleared by team_nl_fill_one_port_get once the updated information has been notified to user space via RTNL

In the userspace, the change flag is maintained by libteam in struct team_port.
The team daemon calls port_priv_change_handler_func on receiving port change event.
The logic in port_priv_change_handler_func
1. creates the port if it did not exist, which triggers port add event and eventually calls lacp_port_added callback.
2. triggers port change event if team_port->changed is true, which eventually calls lw_ethtool_event_watch_port_changed to update port state for link watch ethtool.
3. removes the port if team_port->removed is removed

In lacp_port_added, it calls team_refresh to refresh ifinfo, port info, and option info from the kernel via RTNL.
In this step, port_priv_change_handler_func is called recursively.
- In the inner call, it won't get TEAM_ATTR_PORT_CHANGED flag because kernel has already notified that.
- As a result, team_port->changed flag is cleared in the libteam.
- The port change event won't be triggered from either inner or outer call of port_priv_change_handler_func.

If the port has been up when the port is being added to the team device, the "port up" information is carried in the outer call but will be lost.

In case the flag TEAM_ATTR_PORT_CHANGED is set only in the inner call, function port_priv_change_handler_func can be called in the inner call.
However, it will fail to fetch "enable" options because option_list_init has not be called.

Signed-off-by: Stephen Sun <stephens@nvidia.com>

#### How I did it

Fix:
Do not call check_call_change_handlers when parsing RTNL function is called from another check_call_change_handlers recursively.

#### How to verify it

- Manually test
- Regression test
  - warm reboot
  - warm reboot sad lag
  - warm reboot sad lag member
  - warm reboot sad (partial)

src/libteam/patch/0008-libteam-Add-warm_reboot-mode.patch

@@ -5,12 +5,13 @@ Subject: [PATCH] [libteam]: Reimplement Warm-Reboot procedure'
 
 ---
  libteam/ifinfo.c          |   6 +-
+ libteam/ports.c           |  19 +-
  teamd/teamd.c             |  51 +++-
  teamd/teamd.h             |   6 +
  teamd/teamd_events.c      |  13 ++
  teamd/teamd_per_port.c    |   6 +
  teamd/teamd_runner_lacp.c | 475 +++++++++++++++++++++++++++++++++++---
- 6 files changed, 513 insertions(+), 44 deletions(-)
+ 7 files changed, 530 insertions(+), 46 deletions(-)
 
 diff --git a/libteam/ifinfo.c b/libteam/ifinfo.c
 index 46d56a2..b86d34c 100644
@@ -34,6 +35,58 @@ index 46d56a2..b86d34c 100644
  		set_changed(ifinfo, CHANGED_HWADDR);
  	}
  }
+diff --git a/libteam/ports.c b/libteam/ports.c
+index 9ebf30f..0bd7cc0 100644
+--- a/libteam/ports.c
++++ b/libteam/ports.c
+@@ -128,6 +128,12 @@ int get_port_list_handler(struct nl_msg *msg, void *arg)
+ 	struct nlattr *port_attrs[TEAM_ATTR_PORT_MAX + 1];
+ 	int i;
+ 	uint32_t team_ifindex = 0;
++	/*
++	 * In case get_port_list is being called from check_call_change_handlers recursively,
++	 * there can be some attributes which have not been consumed by callbacks.
++	 * In this case, we should only merge the new attributes into the existing ones without clearing them
++	 */
++	bool recursive = (th->change_handler.pending_type_mask & TEAM_PORT_CHANGE) ? true : false;
+ 
+ 	genlmsg_parse(nlh, 0, attrs, TEAM_ATTR_MAX, NULL);
+ 	if (attrs[TEAM_ATTR_TEAM_IFINDEX])
+@@ -140,7 +146,8 @@ int get_port_list_handler(struct nl_msg *msg, void *arg)
+ 		return NL_SKIP;
+ 
+ 	if (!th->msg_recv_started) {
+-		port_list_cleanup_last_state(th);
++		if (!recursive)
++			port_list_cleanup_last_state(th);
+ 		th->msg_recv_started = true;
+ 	}
+ 	nla_for_each_nested(nl_port, attrs[TEAM_ATTR_LIST_PORT], i) {
+@@ -165,7 +172,9 @@ int get_port_list_handler(struct nl_msg *msg, void *arg)
+ 			if (!port)
+ 				return NL_SKIP;
+ 		}
+-		port->changed = port_attrs[TEAM_ATTR_PORT_CHANGED] ? true : false;
++
++		if (!port->changed || !recursive)
++			port->changed = port_attrs[TEAM_ATTR_PORT_CHANGED] ? true : false;
+ 		port->linkup = port_attrs[TEAM_ATTR_PORT_LINKUP] ? true : false;
+ 		port->removed = port_attrs[TEAM_ATTR_PORT_REMOVED] ? true : false;
+ 		if (port_attrs[TEAM_ATTR_PORT_SPEED])
+@@ -196,6 +204,13 @@ static int get_port_list(struct team_handle *th)
+ 	if (err)
+ 		return err;
+ 
++	/*
++	 * Do not call check_call_change_handlers if this is called recursively to avoid racing conditions
++	 * It will be called by the outer call
++	 */
++	if (th->change_handler.pending_type_mask & TEAM_PORT_CHANGE)
++		return 0;
++
+ 	return check_call_change_handlers(th, TEAM_PORT_CHANGE);
+ 
+ nla_put_failure:
 diff --git a/teamd/teamd.c b/teamd/teamd.c
 index 421e34d..33512a6 100644
 --- a/teamd/teamd.c
@@ -881,5 +934,5 @@ index 955ef0c..782fc05 100644
  
  const struct teamd_runner teamd_runner_lacp = {
 -- 
-2.17.1
+2.30.2