#include<iostream>
#include<vector>
#include<algorithm>
#include<iomanip>
using namespace std;
#define OVERLOAD_REP(_1, _2, _3, name, ...) name
#define REP1(i, n) for (auto i = std::decay_t<decltype(n)>{}; (i) != (n); ++(i))
#define REP2(i, l, r) for (auto i = (l); (i) != (r); ++(i))
#define rep(...) OVERLOAD_REP(__VA_ARGS__, REP2, REP1)(__VA_ARGS__)
#define REP(i, l, r) rep(i, l, r+1)
#define all(x) (x).begin(),(x).end()
#define inf 1000000000
using ll = unsigned long long;
using P = pair<int,int>;
using Graph = vector<vector<int> >;

int main() {
    // 高速化
    ios::sync_with_stdio(false);
    cin.tie(nullptr);

    // 小数点の出力桁数を指定
    cout << fixed << setprecision(10);

    // メイン
    ll M, N;
    cin >> M >> N;
    vector<ll> X(N);
    vector<ll> wa(N+1);
    rep(i, 0, N) {
        cin >> X[i];
        wa[i] = X[i] - 1;
        if(i != 0) wa[i] -= X[i-1];
    }
    wa[N] = M - (N >= 1 ? X[N-1] : 0);
    ll const mod = 998244353;
    ll ans = 0;
    REP(i, 0, N) {
        ll n = wa[i];
        ans += n%mod*(n+1)%mod*(2*n+1)%mod;
    }
    cout << ans/6 << endl;

    return 0;
}