lf_interop_port_reset_test.py: count driver-initiated scans on Linux

[Narayana-CT]

The Linux branch counted only "<3>CTRL-EVENT-SCAN-STARTED", which
wpa_supplicant emits for scans it runs itself. Laptops whose driver
handles SME scan without going through wpa_supplicant, so those scans
left no trace beyond the kernel event and the client reported
Scanning = 0 on every iteration.

Fall back to the kernel "scan started" event when the wpa_supplicant
one is absent, mirroring the fallback the disconnect count already
uses. Replayed against previously captured wifi-msgs: an affected
laptop goes from 0 to 1 scan across all five resets, while one that
already worked stays at 1. The latter logs both strings, so summing
the two counts instead of falling back would double it.

Also rename linux_mac_* to linux_* in the same branch. That name dates
from when the branch handled Linux and macOS together; macOS got its
own branch in "count wifi events on macOS" and can no longer reach
this code, so the mac half of the name is dead.

Verified CLI:

python3 lf_interop_port_reset_test.py
--lanforge_ip 192.168.245.117
--upstream_port eth2
--dut 'Testing_wpa2 AP'
--ssid Testing_wpa2
--encryp psk2
--passwd lanforge
--iterations 2
--reset_interval 5
--android_releases 12 16
--device_list R9ZY30RD0KW,user-ThinkPad-T450,DesktopLatitude5490,Macs-MacBook-Air,User,lin-16,user-HP-ProBook-445R-G6,DESKTOP-0P0OM22,DESKTOP-6UC6DNN