#!/bin/bash

#Zeb H.
#Dell MD1200 DAS fan control - read temps, set fan speed to match.
#Drives the curve off drive temperature, falls back to the enclosure backplane sensors.
#Usage: das_fanctl.sh [loop|once|temps|disks|selftest]   (default: loop)

source /root/fan_speed/functions.sh

PORT=/dev/ttyS1
LOG=/root/fan_speed/log/das_fan.log
CHECK_INTERVAL=30              # seconds between temperature readings
KEEPALIVE_INTERVAL=5           # re-send _shutup this often, the DAS forgets it otherwise
LOW_FAN_SPEED=10
HIGH_FAN_SPEED=60

# Disks in the enclosure - anything behind the DAS SAS expander. The server's own
# drives hang off pci-...-scsi-* instead, so this glob does not pick them up.
DISK_GLOB='/dev/disk/by-path/*-sas-exp*-lun-0'
DISK_LOW_TEMPERATURE_THRESHOLD=40   # at/below this -> LOW_FAN_SPEED  (idle spread is 32-39c)
DISK_HIGH_TEMPERATURE_THRESHOLD=50  # at/above this -> HIGH_FAN_SPEED

# Backplane air sensors, only used when no disk will answer (all spun down, smartctl gone)
BP_LOW_TEMPERATURE_THRESHOLD=30
BP_HIGH_TEMPERATURE_THRESHOLD=45

SPEED=$LOW_FAN_SPEED

# Only one talker on the serial line (detach screen/attachDAS.sh first)
exec 9>/run/das_fanctl.lock
flock -n 9 || { echo "another das_fanctl is running (or /dev/ttyS1 is busy)" >&2; exit 1; }

stty -F $PORT 38400 raw -echo -echoe -echok -echoctl -echoke

# Fire and forget - the DAS answers with a bare prompt, nothing worth reading
set_fan() { printf '_shutup %s\r' "$1" > $PORT; }

# Send a command, return whatever the DAS says back
das_cmd() {
  local out; out=$(mktemp)
  timeout 3 cat $PORT > "$out" &
  local pid=$!
  sleep 0.2
  printf '%s\r' "$1" > $PORT
  wait $pid
  cat "$out"; rm -f "$out"
}

# Hottest drive-backplane sensor from a _temp_rd readout on stdin.
# ponytail: BP_* only - EXP0 (SAS expander) idles ~59c and would peg the fans,
# SIM0 is the enclosure module. Add them here if drives ever run cool but the DAS cooks.
hottest_bp() {
  awk -F'[=c]' '/BP_[0-9]+\[/ {gsub(/ /,"",$2); if ($2+0 > m) m = $2+0} END {print m+0}'
}

# Hottest drive in the enclosure, 0 if nothing answered.
# -n standby so a sleeping drive is skipped rather than spun up just to be measured.
# SATA reports attribute 194, SAS reports "Current Drive Temperature".
hottest_disk() {
  local link dev temp max=0
  for link in $DISK_GLOB; do
    [ -e "$link" ] || continue
    dev=$(readlink -f "$link")
    temp=$(smartctl -n standby -A "$dev" 2>/dev/null |
           awk '/Temperature_Celsius/ {print $10; exit} /Current Drive Temperature/ {print $4; exit}')
    [ -n "$temp" ] && [ "$temp" -gt "$max" ] 2>/dev/null && max=$temp
  done
  echo $max
}

# Read temps, recalculate SPEED, apply it
one_pass() {
  local disk bp source temp low high
  disk=$(hottest_disk)
  if [ "$disk" -gt 0 ]; then
    source=disk; temp=$disk
    low=$DISK_LOW_TEMPERATURE_THRESHOLD; high=$DISK_HIGH_TEMPERATURE_THRESHOLD
  else
    bp=$(das_cmd _temp_rd | hottest_bp)
    if [ "$bp" -le 0 ] 2>/dev/null; then
      echo "$(date +'%Y-%m-%d %H:%M:%S') || no disk or backplane reading, holding fan at ${SPEED}%" >> $LOG
      return 1
    fi
    source=backplane; temp=$bp
    low=$BP_LOW_TEMPERATURE_THRESHOLD; high=$BP_HIGH_TEMPERATURE_THRESHOLD
  fi

  SPEED=$(calculate_interpolated_fan_speed "$temp" $low $high $LOW_FAN_SPEED $HIGH_FAN_SPEED)
  set_fan "$SPEED"
  echo "$(date +'%Y-%m-%d %H:%M:%S') || DAS ${source}:${temp}c | Fan Speed:$SPEED" >> $LOG
  echo "DAS ${source}:${temp}c -> fan ${SPEED}%"
}

selftest() {
  local canned='  BP_1[2] = 30c
  BP_2[3] = 44c
  SIM0[0] = 36c
  EXP0[4] = 59c
  AVG = 39c'
  [ "$(printf '%s' "$canned" | hottest_bp)" = 44 ] || { echo "FAIL: bp parser"; exit 1; }
  [ "$(printf '%s' 'garbage' | hottest_bp)" = 0 ] || { echo "FAIL: bp parser on junk"; exit 1; }
  [ "$(calculate_interpolated_fan_speed 40 40 50 10 60)" = 10 ] || { echo "FAIL: low clamp"; exit 1; }
  [ "$(calculate_interpolated_fan_speed 55 40 50 10 60)" = 60 ] || { echo "FAIL: high clamp"; exit 1; }
  [ "$(calculate_interpolated_fan_speed 45 40 50 10 60)" = 35 ] || { echo "FAIL: midpoint"; exit 1; }
  [ "$(hottest_disk)" -gt 0 ] || { echo "FAIL: no disk temperatures readable"; exit 1; }
  echo "selftest OK"
}

case "${1:-loop}" in
  temps)    das_cmd _temp_rd ;;
  disks)    for l in $DISK_GLOB; do
              d=$(readlink -f "$l")
              printf '%-12s %s\n' "$d" "$(smartctl -n standby -A "$d" 2>/dev/null |
                awk '/Temperature_Celsius/ {print $10"c"; exit} /Current Drive Temperature/ {print $4"c"; exit}')"
            done; echo "hottest: $(hottest_disk)c" ;;
  once)     one_pass ;;
  selftest) selftest ;;
  loop)     while true; do
              one_pass
              # keepalive between readings - matches the old 5s timer cadence
              for _ in $(seq $((CHECK_INTERVAL / KEEPALIVE_INTERVAL))); do
                sleep $KEEPALIVE_INTERVAL
                set_fan "$SPEED"
              done
            done ;;
  *)        echo "usage: $0 [loop|once|temps|disks|selftest]" >&2; exit 1 ;;
esac