#include <bits/stdc++.h>
using namespace std;

struct iofast_t {
    iofast_t() {
        ios::sync_with_stdio(false);
        cin.tie(nullptr);
    }
} iofast;

struct uns_t {} uns;
template <typename Element, typename Head, typename ...Args>
auto vec(Element init, Head arg, Args ...args) {
    if constexpr (sizeof...(Args) == 0) return vector(arg, init);
    else return vector(arg, vec(init, args...));
}
template <typename Element, typename Head, typename ...Args>
auto vec(uns_t, Head arg, Args ...args) {
    return vec(Element(), arg, args...);
}

template <typename Container>
auto distance(const Container &c, decltype(begin(c)) iter) {
    return distance(begin(c), iter);
}

template <typename RIter, typename Compare = less<typename iterator_traits<RIter>::value_type>>
auto isort(RIter first, RIter last, Compare comp = Compare()) {
    vector<int> i(distance(first, last));
    iota(begin(i), end(i), 0);
    sort(begin(i), end(i), [&](auto x, auto y) {
        return comp(*(first + x), *(first + y));
    });
    return i;
}

template <typename, template <typename> typename, typename = void_t<>>
struct detect : false_type {};
template <typename T, template <typename> typename Check>
struct detect<T, Check, void_t<Check<T>>> : true_type {};
template <typename T, template <typename> typename Check>
constexpr inline bool detect_v = detect<T, Check>::value;

template <typename T>
using has_member_sort = decltype(declval<T>().sort());

template <typename Container, typename Compare = less<typename Container::value_type>>
auto sorted(Container c, Compare comp = Compare()) {
    if constexpr (detect_v<Container, has_member_sort>) {
        c.sort(comp);
        return c;
    }
    else {
        sort(begin(c), end(c), comp);
        return c;
    }
}

template <typename Container, typename Compare = equal_to<typename Container::value_type>>
auto uniqued(Container c, Compare comp = Compare()) {
    c.erase(unique(begin(c), end(c), comp), end(c));
    return c;
}

template <typename T, typename Compare = less<T>>
T &chmin(T &l, T r, Compare &&f = less<T>()) { return l = min(l, r, f); }
template <typename T, typename Compare = less<T>>
T &chmax(T &l, T r, Compare &&f = less<T>()) { return l = max(l, r, f); }

template <typename F>
constexpr auto fix(F &&f) noexcept {
    return [f = std::tuple<F>(std::forward<F>(f))](auto &&...args) mutable {
        return std::get<0>(f)(fix(std::get<0>(f)), std::forward<decltype(args)>(args)...);
    };
}

#include <atcoder/segtree>

int op(int x, int y) {
    return x + y;
}
int elem() {
    return 0;
}

using segtree = atcoder::segtree<int, op, elem>;

#include <atcoder/modint>
using mint = atcoder::modint998244353;

int main() {
    int n, m; cin >> n >> m;
    auto l = vec<array<int, 2>>(uns, m);
    for (auto &[p, k] : l) {
        cin >> p >> k; --p; --k;
    }

    sort(begin(l), end(l), [](auto x, auto y) {
        return x[1] < y[1];
    });

    auto f = vec<mint>(uns, n + 1);

    f[0] = 1;
    for (int i = 1; i <= n; ++i) {
        f[i] = f[i - 1] * i;
    }

    mint ans = [&]() {
        mint sum = 0;

        segtree seg(n);
        for (auto [p, k] : l) {
            sum += seg.prod(p, n);
            seg.set(p, 1);
        }

        return sum * f[n - m];
    }();

    auto used = vec<char>(false, n);
    for (auto [p, k] : l) {
        used[p] = true;
    }

    segtree unused(n);
    for (int i = 0; i < n; ++i) {
        if (!used[i]) {
            unused.set(i, 1);
        }
    }

    int used1 = 0;
    for (auto [p, k] : l) {
        mint rem = k - used1++;
        mint upper = unused.prod(p, n);
        ans += upper * rem * f[max(0, n - m - 1)];
    }

    reverse(begin(l), end(l));

    int used2 = 0;
    for (auto [p, k] : l) {
        mint rem = (n - k - 1) - used2++;
        mint lower = unused.prod(0, p);
        ans += lower * rem * f[max(0, n - m - 1)];
    }

    mint count = n - m;

    ans += (count * (count - 1) / 2) * f[max(0, n - m - 2)] * (mint(max(1, n - m)) * max(0, n - m - 1) / 2);

    cout << ans.val() << endl;
}