[MCLAG] Fix iccpd crash bug

[jianjundong]

Signed-off-by: dongjianjun@inspur.com

Why I did it

1. If the peer warm reboot, 'time(&csm->warm_reboot_disconn_time)' is used in function 'mlacp_peer_disconn_handler()' to record the time when the peer enters the warm reboot state; If the neighbor relationship has not been reestablished after 90s, the local switch will migrate from warm reboot disconnected to normal disconnected. However, the 'iccp_csm_status_reset()' function is called immediately after the 'mlacp_peer_disconn_handler()' function, in which 'csm->warm_reboot_disconn_time' will be set to 0. As a result, the local switch cannot migrate to normal disconnected after the 90s timeout.
2. If 'csm->peer_link_if' is set to NULL, 'csm->peer_link_if->is_peer_link' also needs to be set to 0 at the same time; Otherwise, it will lead to the judgment 'if (local_if->is_peer_link)' error, and then use 'csm->peer_link_if' incorrectly, which will lead to iccpd crash.

How I did it

1. In the function 'iccp_csm_status_reset()', 'csm->warm_reboot_disconn_time' is not set to 0.
2. If 'csm->peer_link_if' is set to NULL, 'csm->peer_link_if->is_peer_link' also be set to 0 at the same time.

How to verify it

Test on the switch.

Which release branch to backport (provide reason below if selected)

• 201811
• 201911
• 202006
• 202012

Description for the changelog

A picture of a cute animal (not mandatory but encouraged)