Show Transparent HugePage (THP) across processes sorted by usage

thp_usage.bash

#!/usr/bin/env bash

set -eu

usage_error() { echo >&2 "$(basename "$0"): $1"; exit 2; }

show_options() {
    cat << 'EOF'
Show Transparent HugePage (THP) across processes sorted by usage

Usage: hptop [OPTIONS]
Options:
-h     / --help         - Show this help
-f     / --full         - Show full command line of the processes
-t [n] / --top [n]      - Show top [n] processes by THP usage
EOF
}

full=0
top=0

while [[ $# -gt 0 ]]; do
    case "$1" in
        -h|--help) show_options; exit 0 ;;
        -f|--full) full=1 ;;
        -t|--top) top="${2:-0}"; shift ;;
        *) usage_error "Unknown option: $1" ;;
    esac
    shift
done

tmp_file=$(mktemp)
trap 'rm -f "$tmp_file"' EXIT

# Use parallel processing for THP usage calculation
# shellcheck disable=SC2016
find /proc -maxdepth 1 -regex '/proc/[0-9]+' -printf '%P\n' | xargs -P "$(nproc)" -I {} bash -c '
    thp=$(awk "/AnonHugePages/ {sum+=\$2} END {print sum}" /proc/{}/smaps 2>/dev/null)
    [[ -n "$thp" && "$thp" -gt 0 ]] && echo "$thp {}"
' | sort -rn > "${tmp_file}"

[[ ${top} -gt 0 ]] && sed -i "$((top+1)),\$d" "${tmp_file}"

if ((full)); then
    printf "%-10s %-10s %-10s %-15s %-40s %s\n" "PID" "THP (kB)" "UID" "START_TIME" "PROCESS_NAME" "COMMAND"
    format='%-10s %-10s %-10s %-15s %-40s %s\n'
else
    printf "%-10s %-10s %-10s %-15s %s\n" "PID" "THP (kB)" "UID" "START_TIME" "PROCESS_NAME"
    format='%-10s %-10s %-10s %-15s %s\n'
fi

while read -r thp_usage pid; do
    [[ -d "/proc/${pid}" ]] || continue
    read -r ps_output < <(sudo ps -o pid=,uid=,start_time=,cmd= -p "${pid}" 2>/dev/null)
    [[ -z "${ps_output}" ]] && continue
    if ((full)); then
        read -r _ uid start_time process_name cmdline <<< "${ps_output}"
        # shellcheck disable=SC2059
        printf "${format}" "${pid}" "${thp_usage}" "${uid}" "${start_time}" "${process_name}" "${cmdline}"
    else
        read -r _ uid start_time process_name _ <<< "${ps_output}"
        # shellcheck disable=SC2059
        printf "${format}" "${pid}" "${thp_usage}" "${uid}" "${start_time}" "${process_name}"
    fi
done < "${tmp_file}"