#ifdef ONLINE_JUDGE
#pragma GCC optimize("Ofast,unroll-loops")
#pragma GCC target("avx2,bmi,bmi2,lzcnt,popcnt")
#endif
#include <bits/stdc++.h>
#include <ext/rope>
#include <ext/pb_ds/assoc_container.hpp>
#include <ext/pb_ds/hash_policy.hpp>
#include <ext/pb_ds/tree_policy.hpp>
#include <ext/pb_ds/trie_policy.hpp>
#include <ext/pb_ds/priority_queue.hpp>
using namespace std;
using namespace __gnu_cxx;
using namespace __gnu_pbds;
template <class T> using pbds_set = tree<T, null_type, less_equal<T>, rb_tree_tag,tree_order_statistics_node_update>;
using Trie = trie<string, null_type, trie_string_access_traits<>, pat_trie_tag, trie_prefix_search_node_update>;
// template <class T> using heapq = __gnu_pbds::priority_queue<T, greater<T>, pairing_heap_tag>;
template <class T> using heapq = std::priority_queue<T, vector<T>, greater<T>>;
using ll   =                long long;
using u32  =                unsigned int;
using u64  =                unsigned long long;
using i128 =                __int128;
using u128 =                __uint128_t;
using f128 =                __float128;
using ld   =                long double;
using ui   =                unsigned int;
using ull  =                unsigned long long;
using pii  =                pair<int, int>;
using pll  =                pair<ll, ll>;
using pdd  =                pair<ld, ld>;
using vi   =                vector<int>;
using vvi  =                vector<vector<int>>;
using vll  =                vector<ll>;
using vvll =                vector<vector<ll>>;
using vpii =                vector<pii>;
using vpll =                vector<pll>;
template <class T>
constexpr T infty = 0;
template <>
constexpr int infty<int> = 1'000'000'000;
template <>
constexpr ll infty<ll> = ll(infty<int>) * infty<int> * 2;
template <>
constexpr u32 infty<u32> = infty<int>;
template <>
constexpr u64 infty<u64> = infty<ll>;
template <>
constexpr i128 infty<i128> = i128(infty<ll>) * infty<ll>;
template <>
constexpr double infty<double> = infty<ll>;
template <>
constexpr long double infty<long double> = infty<ll>;
template <class T>
using vc = vector<T>;
template <class T>
using vvc = vector<vc<T>>;
template <class T>
using vvvc = vector<vvc<T>>;
template <class T>
using vvvvc = vector<vvvc<T>>;
template <class T>
using vvvvvc = vector<vvvvc<T>>;
template <class T>
using pq = std::priority_queue<T>;
template <class T>
using pqg = std::priority_queue<T, vector<T>, greater<T>>;
#define vv(type, name, h, ...) \
  vector<vector<type>> name(h, vector<type>(__VA_ARGS__))
#define vvv(type, name, h, w, ...)   \
  vector<vector<vector<type>>> name( \
      h, vector<vector<type>>(w, vector<type>(__VA_ARGS__)))
#define vvvv(type, name, a, b, c, ...)       \
  vector<vector<vector<vector<type>>>> name( \
      a, vector<vector<vector<type>>>(       \
             b, vector<vector<type>>(c, vector<type>(__VA_ARGS__))))
#define lb                  lower_bound
#define ub                  upper_bound
#define pb                  push_back
#define pf                  push_front
#define eb                  emplace_back
#define fi                  first
#define se                  second
#define overload4(_1, _2, _3, _4, name, ...) name
#define overload3(_1, _2, _3, name, ...) name
#define rep1(n)             for(ll _ = 0; _ < n; ++_)
#define rep2(i, n)          for(ll i = 0; i < n; ++i)
#define rep3(i, a, b)       for(ll i = a; i < b; ++i)
#define rep4(i, a, b, c)    for(int i = a; i < b; i += c)
#define rep(...)            overload4(__VA_ARGS__, rep4, rep3, rep2, rep1) (__VA_ARGS__)
#define rrep1(n)            for(ll i = n; i--; )
#define rrep2(i, n)         for(ll i = n; i--; )
#define rrep3(i, a, b)      for(ll i = a; i > b; i--)
#define rrep4(i, a, b, c)   for(ll i = a; i > b; i -= c)
#define rrep(...)           overload4(__VA_ARGS__, rrep4, rrep3, rrep2, rrep1) (__VA_ARGS__)
#define each1(i, a)         for(auto&& i : a)
#define each2(x, y, a)      for(auto&& [x, y] : a)
#define each3(x, y, z, a)   for(auto&& [x, y, z] : a)
#define each(...)           overload4(__VA_ARGS__, each3, each2, each1) (__VA_ARGS__)
#define FOR1(a)             for (ll _ = 0; _ < ll(a); ++_)
#define FOR2(i, a)          for (ll i = 0; i < ll(a); ++i)
#define FOR3(i, a, b)       for (ll i = a; i < ll(b); ++i)
#define FOR4(i, a, b, c)    for (ll i = a; i < ll(b); i += (c))
#define FOR1_R(a)           for (ll i = (a)-1; i >= ll(0); --i)
#define FOR2_R(i, a)        for (ll i = (a)-1; i >= ll(0); --i)
#define FOR3_R(i, a, b)     for (ll i = (b)-1; i >= ll(a); --i)
#define FOR(...)            overload4(__VA_ARGS__, FOR4, FOR3, FOR2, FOR1) (__VA_ARGS__)
#define FOR_R(...)          overload3(__VA_ARGS__, FOR3_R, FOR2_R, FOR1_R) (__VA_ARGS__)
#define FOR_subset(t, s)    for (ll t = (s); t >= 0; t = (t == 0 ? -1 : (t - 1) & (s)))
#define len(x)              (int)x.size()
#define elif                else if
#define all1(i)             begin(i), end(i)
#define all2(i, a)          begin(i), begin(i) + a
#define all3(i, a, b)       begin(i) + a, begin(i) + b
#define all(...)            overload3(__VA_ARGS__, all3, all2, all1) (__VA_ARGS__)
#define rall1(i)            rbegin(i), rend(i)
#define rall2(i, a)         rbegin(i), rbegin(i) + a
#define rall3(i, a, b)      rbegin(i) + a, rbegin(i) + b
#define rall(...)           overload3(__VA_ARGS__, rall3, rall2, rall1) (__VA_ARGS__)
#define mst(x, a)           memset(x, a, sizeof(x))
#define bitcnt(x)           (__builtin_popcountll(x))
#define endl                "\n"
#define MIN(v)              *min_element(all(v))
#define MAX(v)              *max_element(all(v))
#define LB(c, x)            distance((c).begin(), lower_bound(all(c), (x)))
#define UB(c, x)            distance((c).begin(), upper_bound(all(c), (x)))
#define UNIQUE(x)           sort(all(x)), x.erase(unique(all(x)), x.end()), x.shrink_to_fit()
#define SORT(a)             sort(all(a))
#define REV(a)              reverse(all(a))
int popcnt(int x) { return __builtin_popcount(x); }
int popcnt(u32 x) { return __builtin_popcount(x); }
int popcnt(ll x) { return __builtin_popcountll(x); }
int popcnt(u64 x) { return __builtin_popcountll(x); }
// (0, 1, 2, 3, 4) -> (-1, 0, 1, 1, 2)
int topbit(int x) { return (x == 0 ? -1 : 31 - __builtin_clz(x)); }
int topbit(u32 x) { return (x == 0 ? -1 : 31 - __builtin_clz(x)); }
int topbit(ll x) { return (x == 0 ? -1 : 63 - __builtin_clzll(x)); }
int topbit(u64 x) { return (x == 0 ? -1 : 63 - __builtin_clzll(x)); }
// (0, 1, 2, 3, 4) -> (-1, 0, 1, 0, 2)
int lowbit(int x) { return (x == 0 ? -1 : __builtin_ctz(x)); }
int lowbit(u32 x) { return (x == 0 ? -1 : __builtin_ctz(x)); }
int lowbit(ll x) { return (x == 0 ? -1 : __builtin_ctzll(x)); }
int lowbit(u64 x) { return (x == 0 ? -1 : __builtin_ctzll(x)); }
template<class T> auto max(const T& a){ return *max_element(all(a)); }
template<class T> auto min(const T& a){ return *min_element(all(a)); }
template <typename T, typename U>
T ceil(T x, U y) {
    return (x > 0 ? (x + y - 1) / y : x / y);
}
template <typename T, typename U>
T floor(T x, U y) {
    return (x > 0 ? x / y : (x - y + 1) / y);
}
template <typename T, typename U>
pair<T, T> divmod(T x, U y) {
    T q = floor(x, y);
    return {q, x - q * y};
}
template <typename T, typename U>
T SUM(const vector<U> &A) {
    T sum = 0;
    for (auto &&a: A) sum += a;
    return sum;
}
template <typename T, typename U>
vector<T> cumsum(vector<U> &A, int off = 1) {
    int N = A.size();
    vector<T> B(N + 1);
    for (int i = 0; i < N; i++) B[i + 1] = B[i] + A[i];
    if (off == 0) B.erase(B.begin());
    return B;
}
template <typename T>
vector<int> argsort(const vector<T> &A) {
  vector<int> ids(len(A));
  iota(all(ids), 0);
  sort(all(ids),
       [&](int i, int j) { return (A[i] == A[j] ? i < j : A[i] < A[j]); });
  return ids;
}
template <typename T>
vc<T> rearrange(const vc<T> &A, const vc<int> &I) {
  vc<T> B(len(I));
  FOR(i, len(I)) B[i] = A[I[i]];
  return B;
}
template <typename T>
T POP(deque<T> &que) {
  T a = que.front();
  que.pop_front();
  return a;
}
template <typename T>
T POP(pq<T> &que) {
  T a = que.top();
  que.pop();
  return a;
}
template <typename T>
T POP(pqg<T> &que) {
  assert(!que.empty());
  T a = que.top();
  que.pop();
  return a;
}
template <typename T>
T POP(vc<T> &que) {
  assert(!que.empty());
  T a = que.back();
  que.pop_back();
  return a;
}
template <typename F>
ll binary_search(F check, ll ok, ll ng, bool check_ok = true) {
  if (check_ok) assert(check(ok));
  while (abs(ok - ng) > 1) {
    auto x = (ng + ok) / 2;
    (check(x) ? ok : ng) = x;
  }
  return ok;
}
template <typename F>
double binary_search_real(F check, double ok, double ng, int iter = 100) {
  while (iter--) {
    double x = (ok + ng) / 2;
    (check(x) ? ok : ng) = x;
  }
  return (ok + ng) / 2;
}
template <class T, class S> inline bool chmax(T &a, const S &b) {
    return (a < b ? a = b, 1 : 0);
}
template <class T, class S> inline bool chmin(T &a, const S &b) {
    return (a > b ? a = b, 1 : 0);
}
mt19937 rng( chrono::steady_clock::now().time_since_epoch().count() );
#define Ran(a, b) rng() % ( (b) - (a) + 1 ) + (a)
struct custom_hash {
    static uint64_t splitmix64(uint64_t x) {
        // http://xorshift.di.unimi.it/splitmix64.c
        x += 0x9e3779b97f4a7c15;
        x = (x ^ (x >> 30)) * 0xbf58476d1ce4e5b9;
        x = (x ^ (x >> 27)) * 0x94d049bb133111eb;
        return x ^ (x >> 31);
    }

    size_t operator()(uint64_t x) const {
        static const uint64_t FIXED_RANDOM = chrono::steady_clock::now().time_since_epoch().count();
        return splitmix64(x + FIXED_RANDOM);
    }

    size_t operator()(pair<uint64_t,uint64_t> x) const {
        static const uint64_t FIXED_RANDOM = chrono::steady_clock::now().time_since_epoch().count();
        return splitmix64(x.first + FIXED_RANDOM) ^ (splitmix64(x.second + FIXED_RANDOM) >> 1);
    }
};
#define FASTIO
#include <unistd.h>

// https://judge.yosupo.jp/submission/21623

namespace fastio {
static constexpr uint32_t SZ = 1 << 17;
char ibuf[SZ];
char obuf[SZ];
char out[100];
// pointer of ibuf, obuf

uint32_t pil = 0, pir = 0, por = 0;

struct Pre {
  char num[10000][4];
  constexpr Pre() : num() {
    for (int i = 0; i < 10000; i++) {
      int n = i;
      for (int j = 3; j >= 0; j--) {
        num[i][j] = n % 10 | '0';
        n /= 10;
      }
    }
  }
} constexpr pre;

inline void load() {
  memcpy(ibuf, ibuf + pil, pir - pil);
  pir = pir - pil + fread(ibuf + pir - pil, 1, SZ - pir + pil, stdin);
  pil = 0;
  if (pir < SZ) ibuf[pir++] = '\n';
}

inline void flush() {
  fwrite(obuf, 1, por, stdout);
  por = 0;
}

void rd(char &c) {
  do {
    if (pil + 1 > pir) load();
    c = ibuf[pil++];
  } while (isspace(c));
}

void rd(string &x) {
  x.clear();
  char c;
  do {
    if (pil + 1 > pir) load();
    c = ibuf[pil++];
  } while (isspace(c));
  do {
    x += c;
    if (pil == pir) load();
    c = ibuf[pil++];
  } while (!isspace(c));
}

template <typename T>
void rd_real(T &x) {
  string s;
  rd(s);
  x = stod(s);
}

template <typename T>
void rd_integer(T &x) {
  if (pil + 100 > pir) load();
  char c;
  do
    c = ibuf[pil++];
  while (c < '-');
  bool minus = 0;
  if constexpr (is_signed<T>::value || is_same_v<T, i128>) {
    if (c == '-') { minus = 1, c = ibuf[pil++]; }
  }
  x = 0;
  while ('0' <= c) { x = x * 10 + (c & 15), c = ibuf[pil++]; }
  if constexpr (is_signed<T>::value || is_same_v<T, i128>) {
    if (minus) x = -x;
  }
}

void rd(int &x) { rd_integer(x); }
void rd(ll &x) { rd_integer(x); }
void rd(i128 &x) { rd_integer(x); }
void rd(u32 &x) { rd_integer(x); }
void rd(u64 &x) { rd_integer(x); }
void rd(u128 &x) { rd_integer(x); }
void rd(double &x) { rd_real(x); }
void rd(long double &x) { rd_real(x); }
void rd(f128 &x) { rd_real(x); }

template <class T, class U>
void rd(pair<T, U> &p) {
  return rd(p.first), rd(p.second);
}
template <size_t N = 0, typename T>
void rd_tuple(T &t) {
  if constexpr (N < std::tuple_size<T>::value) {
    auto &x = std::get<N>(t);
    rd(x);
    rd_tuple<N + 1>(t);
  }
}
template <class... T>
void rd(tuple<T...> &tpl) {
  rd_tuple(tpl);
}

template <size_t N = 0, typename T>
void rd(array<T, N> &x) {
  for (auto &d: x) rd(d);
}
template <class T>
void rd(vc<T> &x) {
  for (auto &d: x) rd(d);
}

void read() {}
template <class H, class... T>
void read(H &h, T &... t) {
  rd(h), read(t...);
}

void wt(const char c) {
  if (por == SZ) flush();
  obuf[por++] = c;
}
void wt(const string s) {
  for (char c: s) wt(c);
}
void wt(const char *s) {
  size_t len = strlen(s);
  for (size_t i = 0; i < len; i++) wt(s[i]);
}

template <typename T>
void wt_integer(T x) {
  if (por > SZ - 100) flush();
  if (x < 0) { obuf[por++] = '-', x = -x; }
  int outi;
  for (outi = 96; x >= 10000; outi -= 4) {
    memcpy(out + outi, pre.num[x % 10000], 4);
    x /= 10000;
  }
  if (x >= 1000) {
    memcpy(obuf + por, pre.num[x], 4);
    por += 4;
  } else if (x >= 100) {
    memcpy(obuf + por, pre.num[x] + 1, 3);
    por += 3;
  } else if (x >= 10) {
    int q = (x * 103) >> 10;
    obuf[por] = q | '0';
    obuf[por + 1] = (x - q * 10) | '0';
    por += 2;
  } else
    obuf[por++] = x | '0';
  memcpy(obuf + por, out + outi + 4, 96 - outi);
  por += 96 - outi;
}

template <typename T>
void wt_real(T x) {
  ostringstream oss;
  oss << fixed << setprecision(15) << double(x);
  string s = oss.str();
  wt(s);
}

void wt(int x) { wt_integer(x); }
void wt(ll x) { wt_integer(x); }
void wt(i128 x) { wt_integer(x); }
void wt(u32 x) { wt_integer(x); }
void wt(u64 x) { wt_integer(x); }
void wt(u128 x) { wt_integer(x); }
void wt(double x) { wt_real(x); }
void wt(long double x) { wt_real(x); }
void wt(f128 x) { wt_real(x); }

template <class T, class U>
void wt(const pair<T, U> val) {
  wt(val.first);
  wt(' ');
  wt(val.second);
}
template <size_t N = 0, typename T>
void wt_tuple(const T t) {
  if constexpr (N < std::tuple_size<T>::value) {
    if constexpr (N > 0) { wt(' '); }
    const auto x = std::get<N>(t);
    wt(x);
    wt_tuple<N + 1>(t);
  }
}
template <class... T>
void wt(tuple<T...> tpl) {
  wt_tuple(tpl);
}
template <class T, size_t S>
void wt(const array<T, S> val) {
  auto n = val.size();
  for (size_t i = 0; i < n; i++) {
    if (i) wt(' ');
    wt(val[i]);
  }
}
template <class T>
void wt(const vector<T> val) {
  auto n = val.size();
  for (size_t i = 0; i < n; i++) {
    if (i) wt(' ');
    wt(val[i]);
  }
}

void print() { wt('\n'); }
template <class Head, class... Tail>
void print(Head &&head, Tail &&... tail) {
  wt(head);
  if (sizeof...(Tail)) wt(' ');
  print(forward<Tail>(tail)...);
}

// gcc expansion. called automaticall after main.

void __attribute__((destructor)) _d() { flush(); }
} // namespace fastio

using fastio::read;
using fastio::print;
using fastio::flush;

#define INT(...)   \
  int __VA_ARGS__; \
  read(__VA_ARGS__)
#define LL(...)   \
  ll __VA_ARGS__; \
  read(__VA_ARGS__)
#define U32(...)   \
  u32 __VA_ARGS__; \
  read(__VA_ARGS__)
#define U64(...)   \
  u64 __VA_ARGS__; \
  read(__VA_ARGS__)
#define STR(...)      \
  string __VA_ARGS__; \
  read(__VA_ARGS__)
#define CHAR(...)   \
  char __VA_ARGS__; \
  read(__VA_ARGS__)
#define DBL(...)      \
  double __VA_ARGS__; \
  read(__VA_ARGS__)

#define VEC(type, name, size) \
  vector<type> name(size);    \
  read(name)
#define VV(type, name, h, w)                     \
  vector<vector<type>> name(h, vector<type>(w)); \
  read(name)

void YES(bool t = 1) { print(t ? "YES" : "NO"); }
void NO(bool t = 1) { YES(!t); }
void Yes(bool t = 1) { print(t ? "Yes" : "No"); }
void No(bool t = 1) { Yes(!t); }
void yes(bool t = 1) { print(t ? "yes" : "no"); }
void no(bool t = 1) { yes(!t); }
const i128 ONE = 1;
template <typename Iterable>
auto print_all(const Iterable& v, std::string sep = " ", std::string end = "\n") -> decltype(fastio::wt(*v.begin())) {
    for (auto it = v.begin(); it != v.end();) {
        fastio::wt(*it);
        if (++it != v.end()) fastio::wt(sep);
    }
    fastio::wt(end);
}
ll gcd(ll x, ll y) {
    if(!x) return y;
    if(!y) return x;
    int t = __builtin_ctzll(x | y);
    x >>= __builtin_ctzll(x);
    do {
        y >>= __builtin_ctzll(y);
        if (x > y) swap(x, y);
        y -= x;
    } while (y);
    return x << t;
}
ll lcm(ll x, ll y) { return x * y / gcd(x, y); }
ll exgcd(ll a, ll b, ll &x, ll &y) {
    if(!b) return x = 1, y = 0, a;
    ll d = exgcd(b, a % b, x, y);
    ll t = x;
    x = y;
    y = t - a / b * x;
    return d;
}
ll max(ll x, ll y) { return x > y ? x : y; }
ll min(ll x, ll y) { return x < y ? x : y; }
ll Mod(ll x, int mod) { return (x % mod + mod) % mod; }
ll pow(ll x, ll y, ll mod){
    ll res = 1, cur = x;
    while (y) {
        if (y & 1) res = res * cur % mod;
        cur = ONE * cur * cur % mod;
        y >>= 1;
    }
    return res % mod;
}
ll probabilityMod(ll x, ll y, ll mod) {
    return x * pow(y, mod-2, mod) % mod;
}
vvi getGraph(int n, int m, bool directed = false) {
    vvi res(n);
    rep(_, 0, m) {
        INT(u, v);
        u--, v--;
        res[u].emplace_back(v);
        if(!directed) res[v].emplace_back(u);
    }
    return res;
}
vector<vpii> getWeightedGraph(int n, int m, bool directed = false) {
    vector<vpii> res(n);
    rep(_, 0, m) {
        INT(u, v, w);
        u--, v--;
        res[u].emplace_back(v, w);
        if(!directed) res[v].emplace_back(u, w);
    }
    return res;
}
template <class... Args> auto ndvector(size_t n, Args &&...args) {
    if constexpr (sizeof...(args) == 1) {
        return vector(n, args...);
    } else {
        return vector(n, ndvector(args...));
    }
}
const ll LINF = 0x1fffffffffffffff;
const ll MINF = 0x7fffffffffff;
const int INF = 0x3fffffff;
const int MOD = 1000000007;
const int MODD = 998244353;
const int N = 1e6 + 10;

#line 2 "mod/dynamic_modint.hpp"

#line 2 "mod/modint_common.hpp"

struct has_mod_impl {
  template <class T>
  static auto check(T &&x) -> decltype(x.get_mod(), std::true_type{});
  template <class T>
  static auto check(...) -> std::false_type;
};

template <class T>
class has_mod : public decltype(has_mod_impl::check<T>(std::declval<T>())) {};

template <typename mint>
mint inv(int n) {
  static const int mod = mint::get_mod();
  static vector<mint> dat = {0, 1};
  assert(0 <= n);
  if (n >= mod) n %= mod;
  while (len(dat) <= n) {
    int k = len(dat);
    int q = (mod + k - 1) / k;
    dat.eb(dat[k * q - mod] * mint::raw(q));
  }
  return dat[n];
}

template <typename mint>
mint fact(int n) {
  static const int mod = mint::get_mod();
  assert(0 <= n && n < mod);
  static vector<mint> dat = {1, 1};
  while (len(dat) <= n) dat.eb(dat[len(dat) - 1] * mint::raw(len(dat)));
  return dat[n];
}

template <typename mint>
mint fact_inv(int n) {
  static vector<mint> dat = {1, 1};
  if (n < 0) return mint(0);
  while (len(dat) <= n) dat.eb(dat[len(dat) - 1] * inv<mint>(len(dat)));
  return dat[n];
}

template <class mint, class... Ts>
mint fact_invs(Ts... xs) {
  return (mint(1) * ... * fact_inv<mint>(xs));
}

template <typename mint, class Head, class... Tail>
mint multinomial(Head &&head, Tail &&... tail) {
  return fact<mint>(head) * fact_invs<mint>(std::forward<Tail>(tail)...);
}

template <typename mint>
mint C_dense(int n, int k) {
  static vvc<mint> C;
  static int H = 0, W = 0;
  auto calc = [&](int i, int j) -> mint {
    if (i == 0) return (j == 0 ? mint(1) : mint(0));
    return C[i - 1][j] + (j ? C[i - 1][j - 1] : 0);
  };
  if (W <= k) {
    FOR(i, H) {
      C[i].resize(k + 1);
      FOR(j, W, k + 1) { C[i][j] = calc(i, j); }
    }
    W = k + 1;
  }
  if (H <= n) {
    C.resize(n + 1);
    FOR(i, H, n + 1) {
      C[i].resize(W);
      FOR(j, W) { C[i][j] = calc(i, j); }
    }
    H = n + 1;
  }
  return C[n][k];
}

template <typename mint, bool large = false, bool dense = false>
mint C(ll n, ll k) {
  assert(n >= 0);
  if (k < 0 || n < k) return 0;
  if constexpr (dense) return C_dense<mint>(n, k);
  if constexpr (!large) return multinomial<mint>(n, k, n - k);
  k = min(k, n - k);
  mint x(1);
  FOR(i, k) x *= mint(n - i);
  return x * fact_inv<mint>(k);
}

template <typename mint, bool large = false>
mint C_inv(ll n, ll k) {
  assert(n >= 0);
  assert(0 <= k && k <= n);
  if (!large) return fact_inv<mint>(n) * fact<mint>(k) * fact<mint>(n - k);
  return mint(1) / C<mint, 1>(n, k);
}

// [x^d](1-x)^{-n}
template <typename mint, bool large = false, bool dense = false>
mint C_negative(ll n, ll d) {
  assert(n >= 0);
  if (d < 0) return mint(0);
  if (n == 0) { return (d == 0 ? mint(1) : mint(0)); }
  return C<mint, large, dense>(n + d - 1, d);
}
#line 2 "mod/primitive_root.hpp"

#line 2 "nt/factor.hpp"

#line 2 "random/base.hpp"

u64 RNG_64() {
  static uint64_t x_
      = uint64_t(chrono::duration_cast<chrono::nanoseconds>(
                     chrono::high_resolution_clock::now().time_since_epoch())
                     .count())
        * 10150724397891781847ULL;
  x_ ^= x_ << 7;
  return x_ ^= x_ >> 9;
}

u64 RNG(u64 lim) { return RNG_64() % lim; }

ll RNG(ll l, ll r) { return l + RNG_64() % (r - l); }
#line 2 "mod/mongomery_modint.hpp"

// odd mod.
// x の代わりに rx を持つ
template <int id, typename U1, typename U2>
struct Mongomery_modint {
  using mint = Mongomery_modint;
  inline static U1 m, r, n2;
  static constexpr int W = numeric_limits<U1>::digits;

  static void set_mod(U1 mod) {
    assert(mod & 1 && mod <= U1(1) << (W - 2));
    m = mod, n2 = -U2(m) % m, r = m;
    FOR(5) r *= 2 - m * r;
    r = -r;
    assert(r * m == U1(-1));
  }
  static U1 reduce(U2 b) { return (b + U2(U1(b) * r) * m) >> W; }

  U1 x;
  Mongomery_modint() : x(0) {}
  Mongomery_modint(U1 x) : x(reduce(U2(x) * n2)){};
  U1 val() const {
    U1 y = reduce(x);
    return y >= m ? y - m : y;
  }
  mint &operator+=(mint y) {
    x = ((x += y.x) >= m ? x - m : x);
    return *this;
  }
  mint &operator-=(mint y) {
    x -= (x >= y.x ? y.x : y.x - m);
    return *this;
  }
  mint &operator*=(mint y) {
    x = reduce(U2(x) * y.x);
    return *this;
  }
  mint operator+(mint y) const { return mint(*this) += y; }
  mint operator-(mint y) const { return mint(*this) -= y; }
  mint operator*(mint y) const { return mint(*this) *= y; }
  bool operator==(mint y) const {
    return (x >= m ? x - m : x) == (y.x >= m ? y.x - m : y.x);
  }
  bool operator!=(mint y) const { return not operator==(y); }
  mint pow(ll n) const {
    assert(n >= 0);
    mint y = 1, z = *this;
    for (; n; n >>= 1, z *= z)
      if (n & 1) y *= z;
    return y;
  }
};

template <int id>
using Mongomery_modint_32 = Mongomery_modint<id, u32, u64>;
template <int id>
using Mongomery_modint_64 = Mongomery_modint<id, u64, u128>;
#line 3 "nt/primetest.hpp"

bool primetest(const u64 x) {
  assert(x < u64(1) << 62);
  if (x == 2 or x == 3 or x == 5 or x == 7) return true;
  if (x % 2 == 0 or x % 3 == 0 or x % 5 == 0 or x % 7 == 0) return false;
  if (x < 121) return x > 1;
  const u64 d = (x - 1) >> lowbit(x - 1);

  using mint = Mongomery_modint_64<202311020>;

  mint::set_mod(x);
  const mint one(u64(1)), minus_one(x - 1);
  auto ok = [&](u64 a) -> bool {
    auto y = mint(a).pow(d);
    u64 t = d;
    while (y != one && y != minus_one && t != x - 1) y *= y, t <<= 1;
    if (y != minus_one && t % 2 == 0) return false;
    return true;
  };
  if (x < (u64(1) << 32)) {
    for (u64 a: {2, 7, 61})
      if (!ok(a)) return false;
  } else {
    for (u64 a: {2, 325, 9375, 28178, 450775, 9780504, 1795265022}) {
      if (!ok(a)) return false;
    }
  }
  return true;
}
#line 5 "nt/factor.hpp"

template <typename mint>
ll rho(ll n, ll c) {
  assert(n > 1);
  const mint cc(c);
  auto f = [&](mint x) { return x * x + cc; };
  mint x = 1, y = 2, z = 1, q = 1;
  ll g = 1;
  const ll m = 1LL << (__lg(n) / 5);
  for (ll r = 1; g == 1; r <<= 1) {
    x = y;
    FOR(r) y = f(y);
    for (ll k = 0; k < r && g == 1; k += m) {
      z = y;
      FOR(min(m, r - k)) y = f(y), q *= x - y;
      g = gcd(q.val(), n);
    }
  }
  if (g == n) do {
      z = f(z);
      g = gcd((x - z).val(), n);
    } while (g == 1);
  return g;
}

ll find_prime_factor(ll n) {
  assert(n > 1);
  if (primetest(n)) return n;
  FOR(100) {
    ll m = 0;
    if (n < (1 << 30)) {
      using mint = Mongomery_modint_32<20231025>;
      mint::set_mod(n);
      m = rho<mint>(n, RNG(0, n));
    } else {
      using mint = Mongomery_modint_64<20231025>;
      mint::set_mod(n);
      m = rho<mint>(n, RNG(0, n));
    }
    if (primetest(m)) return m;
    n = m;
  }
  assert(0);
  return -1;
}

// ソートしてくれる
vc<pair<ll, int>> factor(ll n) {
  assert(n >= 1);
  vc<pair<ll, int>> pf;
  FOR(p, 2, 100) {
    if (p * p > n) break;
    if (n % p == 0) {
      ll e = 0;
      do { n /= p, e += 1; } while (n % p == 0);
      pf.eb(p, e);
    }
  }
  while (n > 1) {
    ll p = find_prime_factor(n);
    ll e = 0;
    do { n /= p, e += 1; } while (n % p == 0);
    pf.eb(p, e);
  }
  sort(all(pf));
  return pf;
}

vc<pair<ll, int>> factor_by_lpf(ll n, vc<int>& lpf) {
  vc<pair<ll, int>> res;
  while (n > 1) {
    int p = lpf[n];
    int e = 0;
    while (n % p == 0) {
      n /= p;
      ++e;
    }
    res.eb(p, e);
  }
  return res;
}
#line 2 "mod/mod_pow.hpp"

#line 2 "mod/barrett.hpp"

// https://github.com/atcoder/ac-library/blob/master/atcoder/internal_math.hpp
struct Barrett {
  u32 m;
  u64 im;
  explicit Barrett(u32 m = 1) : m(m), im(u64(-1) / m + 1) {}
  u32 umod() const { return m; }
  u32 modulo(u64 z) {
    if (m == 1) return 0;
    u64 x = (u64)(((unsigned __int128)(z)*im) >> 64);
    u64 y = x * m;
    return (z - y + (z < y ? m : 0));
  }
  u64 floor(u64 z) {
    if (m == 1) return z;
    u64 x = (u64)(((unsigned __int128)(z)*im) >> 64);
    u64 y = x * m;
    return (z < y ? x - 1 : x);
  }
  pair<u64, u32> divmod(u64 z) {
    if (m == 1) return {z, 0};
    u64 x = (u64)(((unsigned __int128)(z)*im) >> 64);
    u64 y = x * m;
    if (z < y) return {x - 1, z - y + m};
    return {x, z - y};
  }
  u32 mul(u32 a, u32 b) { return modulo(u64(a) * b); }
};

struct Barrett_64 {
  u128 mod, mh, ml;

  explicit Barrett_64(u64 mod = 1) : mod(mod) {
    u128 m = u128(-1) / mod;
    if (m * mod + mod == u128(0)) ++m;
    mh = m >> 64;
    ml = m & u64(-1);
  }

  u64 umod() const { return mod; }

  u64 modulo(u128 x) {
    u128 z = (x & u64(-1)) * ml;
    z = (x & u64(-1)) * mh + (x >> 64) * ml + (z >> 64);
    z = (x >> 64) * mh + (z >> 64);
    x -= z * mod;
    return x < mod ? x : x - mod;
  }

  u64 mul(u64 a, u64 b) { return modulo(u128(a) * b); }
};
#line 5 "mod/mod_pow.hpp"

u32 mod_pow(int a, ll n, int mod) {
  assert(n >= 0);
  a = ((a %= mod) < 0 ? a + mod : a);
  if ((mod & 1) && (mod < (1 << 30))) {
    using mint = Mongomery_modint_32<202311021>;
    mint::set_mod(mod);
    return mint(a).pow(n).val();
  }
  Barrett bt(mod);
  int r = 1;
  while (n) {
    if (n & 1) r = bt.mul(r, a);
    a = bt.mul(a, a), n >>= 1;
  }
  return r;
}

u64 mod_pow_64(ll a, ll n, u64 mod) {
  assert(n >= 0);
  a = ((a %= mod) < 0 ? a + mod : a);
  if ((mod & 1) && (mod < (u64(1) << 62))) {
    using mint = Mongomery_modint_64<202311021>;
    mint::set_mod(mod);
    return mint(a).pow(n).val();
  }
  Barrett_64 bt(mod);
  ll r = 1;
  while (n) {
    if (n & 1) r = bt.mul(r, a);
    a = bt.mul(a, a), n >>= 1;
  }
  return r;
}
#line 6 "mod/primitive_root.hpp"

// int

int primitive_root(int p) {
  auto pf = factor(p - 1);
  auto is_ok = [&](int g) -> bool {
    for (auto&& [q, e]: pf)
      if (mod_pow(g, (p - 1) / q, p) == 1) return false;
    return true;
  };
  while (1) {
    int x = RNG(1, p);
    if (is_ok(x)) return x;
  }
  return -1;
}

ll primitive_root_64(ll p) {
  auto pf = factor(p - 1);
  auto is_ok = [&](ll g) -> bool {
    for (auto&& [q, e]: pf)
      if (mod_pow_64(g, (p - 1) / q, p) == 1) return false;
    return true;
  };
  while (1) {
    ll x = RNG(1, p);
    if (is_ok(x)) return x;
  }
  return -1;
}
#line 6 "mod/dynamic_modint.hpp"

template <int id>
struct Dynamic_Modint {
  static constexpr bool is_modint = true;
  using mint = Dynamic_Modint;
  u32 val;
  static Barrett bt;
  static u32 umod() { return bt.umod(); }

  static int get_mod() { return (int)(bt.umod()); }
  static void set_mod(int m) {
    assert(1 <= m);
    bt = Barrett(m);
  }

  static Dynamic_Modint raw(u32 v) {
    Dynamic_Modint x;
    x.val = v;
    return x;
  }
  Dynamic_Modint() : val(0) {}
  Dynamic_Modint(u32 x) : val(bt.modulo(x)) {}
  Dynamic_Modint(u64 x) : val(bt.modulo(x)) {}
  Dynamic_Modint(int x) : val((x %= get_mod()) < 0 ? x + get_mod() : x) {}
  Dynamic_Modint(ll x) : val((x %= get_mod()) < 0 ? x + get_mod() : x) {}

  mint& operator+=(const mint& rhs) {
    val = (val += rhs.val) < umod() ? val : val - umod();
    return *this;
  }
  mint& operator-=(const mint& rhs) {
    val = (val += umod() - rhs.val) < umod() ? val : val - umod();
    return *this;
  }
  mint& operator*=(const mint& rhs) {
    val = bt.mul(val, rhs.val);
    return *this;
  }
  mint& operator/=(const mint& rhs) { return *this = *this * rhs.inverse(); }
  mint operator-() const { return mint() - *this; }
  mint pow(ll n) const {
    assert(0 <= n);
    mint x = *this, r = 1;
    while (n) {
      if (n & 1) r *= x;
      x *= x, n >>= 1;
    }
    return r;
  }
  mint inverse() const {
    int x = val, mod = get_mod();
    int a = x, b = mod, u = 1, v = 0, t;
    while (b > 0) {
      t = a / b;
      swap(a -= t * b, b), swap(u -= t * v, v);
    }
    if (u < 0) u += mod;
    return u;
  }

  friend mint operator+(const mint& lhs, const mint& rhs) {
    return mint(lhs) += rhs;
  }
  friend mint operator-(const mint& lhs, const mint& rhs) {
    return mint(lhs) -= rhs;
  }
  friend mint operator*(const mint& lhs, const mint& rhs) {
    return mint(lhs) *= rhs;
  }
  friend mint operator/(const mint& lhs, const mint& rhs) {
    return mint(lhs) /= rhs;
  }
  friend bool operator==(const mint& lhs, const mint& rhs) {
    return lhs.val == rhs.val;
  }
  friend bool operator!=(const mint& lhs, const mint& rhs) {
    return lhs.val != rhs.val;
  }
  static pair<int, int>& get_ntt() {
    static pair<int, int> p = {-1, -1};
    return p;
  }
  static void set_ntt_info() {
    int mod = get_mod();
    int k = lowbit(mod - 1);
    int r = primitive_root(mod);
    r = mod_pow(r, (mod - 1) >> k, mod);
    get_ntt() = {k, r};
  }
  static pair<int, int> ntt_info() { return get_ntt(); }
  static bool can_ntt() { return ntt_info().fi != -1; }
};

#ifdef FASTIO
template <int id>
void rd(Dynamic_Modint<id>& x) {
  fastio::rd(x.val);
  x.val %= Dynamic_Modint<id>::umod();
}
template <int id>
void wt(Dynamic_Modint<id> x) {
  fastio::wt(x.val);
}
#endif

using dmint = Dynamic_Modint<-1>;
template <int id>
Barrett Dynamic_Modint<id>::bt;

#line 3 "linalg/matrix_mul.hpp"

template <class T, typename enable_if<has_mod<T>::value>::type* = nullptr>
vc<vc<T>> matrix_mul(const vc<vc<T>>& A, const vc<vc<T>>& B, int N1 = -1,
                     int N2 = -1, int N3 = -1) {
  if (N1 == -1) { N1 = len(A), N2 = len(B), N3 = len(B[0]); }
  vv(u32, b, N3, N2);
  FOR(i, N2) FOR(j, N3) b[j][i] = B[i][j].val;
  vv(T, C, N1, N3);

  if ((T::get_mod() < (1 << 30)) && N2 <= 16) {
    FOR(i, N1) FOR(j, N3) {
      u64 sm = 0;
      FOR(m, N2) sm += u64(A[i][m].val) * b[j][m];
      C[i][j] = sm;
    }
  } else {
    FOR(i, N1) FOR(j, N3) {
      u128 sm = 0;
      FOR(m, N2) sm += u64(A[i][m].val) * b[j][m];
      C[i][j] = T::raw(sm % (T::get_mod()));
    }
  }
  return C;
}

template <class T, typename enable_if<!has_mod<T>::value>::type* = nullptr>
vc<vc<T>> matrix_mul(const vc<vc<T>>& A, const vc<vc<T>>& B, int N1 = -1,
                     int N2 = -1, int N3 = -1) {
  if (N1 == -1) { N1 = len(A), N2 = len(B), N3 = len(B[0]); }
  vv(T, b, N2, N3);
  FOR(i, N2) FOR(j, N3) b[j][i] = B[i][j];
  vv(T, C, N1, N3);
  FOR(n, N1) FOR(m, N2) FOR(k, N3) C[n][k] += A[n][m] * b[k][m];
  return C;
}

// square-matrix defined as array

template <class T, int N>
array<array<T, N>, N> matrix_mul(const array<array<T, N>, N>& A,
                                 const array<array<T, N>, N>& B) {
  array<array<T, N>, N> C{};

  if ((T::get_mod() < (1 << 30)) && N <= 16) {
    FOR(i, N) FOR(k, N) {
      u64 sm = 0;
      FOR(j, N) sm += u64(A[i][j].val) * (B[j][k].val);
      C[i][k] = sm;
    }
  } else {
    FOR(i, N) FOR(k, N) {
      u128 sm = 0;
      FOR(j, N) sm += u64(A[i][j].val) * (B[j][k].val);
      C[i][k] = sm;
    }
  }
  return C;
}

#line 2 "ds/segtree/lazy_segtree.hpp"

template <typename ActedMonoid>
struct Lazy_SegTree {
  using AM = ActedMonoid;
  using MX = typename AM::Monoid_X;
  using MA = typename AM::Monoid_A;
  using X = typename MX::value_type;
  using A = typename MA::value_type;
  int n, log, size;
  vc<X> dat;
  vc<A> laz;

  Lazy_SegTree() {}
  Lazy_SegTree(int n) { build(n); }
  template <typename F>
  Lazy_SegTree(int n, F f) {
    build(n, f);
  }
  Lazy_SegTree(const vc<X>& v) { build(v); }

  void build(int m) {
    build(m, [](int i) -> X { return MX::unit(); });
  }
  void build(const vc<X>& v) {
    build(len(v), [&](int i) -> X { return v[i]; });
  }
  template <typename F>
  void build(int m, F f) {
    n = m, log = 1;
    while ((1 << log) < n) ++log;
    size = 1 << log;
    dat.assign(size << 1, MX::unit());
    laz.assign(size, MA::unit());
    FOR(i, n) dat[size + i] = f(i);
    FOR_R(i, 1, size) update(i);
  }

  void update(int k) { dat[k] = MX::op(dat[2 * k], dat[2 * k + 1]); }
  void set(int p, X x) {
    assert(0 <= p && p < n);
    p += size;
    for (int i = log; i >= 1; i--) push(p >> i);
    dat[p] = x;
    for (int i = 1; i <= log; i++) update(p >> i);
  }
  void multiply(int p, const X& x) {
    assert(0 <= p && p < n);
    p += size;
    for (int i = log; i >= 1; i--) push(p >> i);
    dat[p] = MX::op(dat[p], x);
    for (int i = 1; i <= log; i++) update(p >> i);
  }

  X get(int p) {
    assert(0 <= p && p < n);
    p += size;
    for (int i = log; i >= 1; i--) push(p >> i);
    return dat[p];
  }

  vc<X> get_all() {
    FOR(k, 1, size) { push(k); }
    return {dat.begin() + size, dat.begin() + size + n};
  }

  X prod(int l, int r) {
    assert(0 <= l && l <= r && r <= n);
    if (l == r) return MX::unit();
    l += size, r += size;
    for (int i = log; i >= 1; i--) {
      if (((l >> i) << i) != l) push(l >> i);
      if (((r >> i) << i) != r) push((r - 1) >> i);
    }
    X xl = MX::unit(), xr = MX::unit();
    while (l < r) {
      if (l & 1) xl = MX::op(xl, dat[l++]);
      if (r & 1) xr = MX::op(dat[--r], xr);
      l >>= 1, r >>= 1;
    }
    return MX::op(xl, xr);
  }

  X prod_all() { return dat[1]; }

  void apply(int l, int r, A a) {
    assert(0 <= l && l <= r && r <= n);
    if (l == r) return;
    l += size, r += size;
    for (int i = log; i >= 1; i--) {
      if (((l >> i) << i) != l) push(l >> i);
      if (((r >> i) << i) != r) push((r - 1) >> i);
    }
    int l2 = l, r2 = r;
    while (l < r) {
      if (l & 1) apply_at(l++, a);
      if (r & 1) apply_at(--r, a);
      l >>= 1, r >>= 1;
    }
    l = l2, r = r2;
    for (int i = 1; i <= log; i++) {
      if (((l >> i) << i) != l) update(l >> i);
      if (((r >> i) << i) != r) update((r - 1) >> i);
    }
  }

  template <typename F>
  int max_right(const F check, int l) {
    assert(0 <= l && l <= n);
    assert(check(MX::unit()));
    if (l == n) return n;
    l += size;
    for (int i = log; i >= 1; i--) push(l >> i);
    X sm = MX::unit();
    do {
      while (l % 2 == 0) l >>= 1;
      if (!check(MX::op(sm, dat[l]))) {
        while (l < size) {
          push(l);
          l = (2 * l);
          if (check(MX::op(sm, dat[l]))) { sm = MX::op(sm, dat[l++]); }
        }
        return l - size;
      }
      sm = MX::op(sm, dat[l++]);
    } while ((l & -l) != l);
    return n;
  }

  template <typename F>
  int min_left(const F check, int r) {
    assert(0 <= r && r <= n);
    assert(check(MX::unit()));
    if (r == 0) return 0;
    r += size;
    for (int i = log; i >= 1; i--) push((r - 1) >> i);
    X sm = MX::unit();
    do {
      r--;
      while (r > 1 && (r % 2)) r >>= 1;
      if (!check(MX::op(dat[r], sm))) {
        while (r < size) {
          push(r);
          r = (2 * r + 1);
          if (check(MX::op(dat[r], sm))) { sm = MX::op(dat[r--], sm); }
        }
        return r + 1 - size;
      }
      sm = MX::op(dat[r], sm);
    } while ((r & -r) != r);
    return 0;
  }

private:
  void apply_at(int k, A a) {
    ll sz = 1 << (log - topbit(k));
    dat[k] = AM::act(dat[k], a, sz);
    if (k < size) laz[k] = MA::op(laz[k], a);
  }
  void push(int k) {
    if (laz[k] == MA::unit()) return;
    apply_at(2 * k, laz[k]), apply_at(2 * k + 1, laz[k]);
    laz[k] = MA::unit();
  }
};

#line 2 "alg/monoid/add.hpp"

template <typename X>
struct Monoid_Add {
  using value_type = X;
  static constexpr X op(const X &x, const X &y) noexcept { return x + y; }
  static constexpr X inverse(const X &x) noexcept { return -x; }
  static constexpr X power(const X &x, ll n) noexcept { return X(n) * x; }
  static constexpr X unit() { return X(0); }
  static constexpr bool commute = true;
};
#line 2 "alg/acted_monoid/sum_add.hpp"

template <typename E>
struct ActedMonoid_Sum_Add {
  using Monoid_X = Monoid_Add<E>;
  using Monoid_A = Monoid_Add<E>;
  using X = typename Monoid_X::value_type;
  using A = typename Monoid_A::value_type;
  static constexpr X act(const X &x, const A &a, const ll &size) {
    return x + a * E(size);
  }
};

using mint = dmint;

void solve() {
    INT(n, B, q);
    mint::set_mod(B);
    vc<tuple<mint, mint, mint>> XYZ = {{1, 1, 1}};
    rep(i, 1e5) {
        auto [x, y, z] = XYZ.back();
        x += 1;
        y = 3 * y + 2 * x * z;
        z = 3 * z;
        XYZ.eb(x, y, z);
    }
    Lazy_SegTree<ActedMonoid_Sum_Add<int>> seg(n);
    rep(q) {
        INT(L, M, R);
        L--, M--;
        seg.apply(L, R, 1);
        int idx = seg.get(M);
        print(XYZ[idx]);
    }
    
}

signed main() {
    int T = 1;
    // read(T);
    while (T--) {
        solve();
    }
    return 0;
}