class segt:
  def __init__(self, n, ele, calc):
    self.num = 2 ** (n - 1).bit_length()
    self.el = ele
    self.data = [ele] * (2 * self.num)
    self.calc = calc

  def update(self, idx, x):
    idx += self.num - 1
    self.data[idx] = x
    while idx > 0:
      idx = (idx - 1) // 2
      self.data[idx] = self.calc(self.data[2*idx+1], self.data[2*idx+2])
  
  def renew(self, idx, x):
    self.update(idx, self.calc(self.get(idx), x))

  def prod(self, left, right):
    l = left + self.num
    r = right + self.num
    res = self.el
    while l < r:
      if l % 2:
        res = self.calc(res, self.data[l-1])
        l += 1
      if r % 2:
        r -= 1
        res = self.calc(res, self.data[r-1])
      l //= 2
      r //= 2
    return res

  def get(self, idx):
    return self.data[idx+self.num-1]

def add(x, y):
  return x + y


n, m = map(int, input().split())
pk = [list(map(lambda x: int(x)-1, input().split())) for _ in range(m)]
def solve(n, m, pk):
  mod = 998244353
  fac = [1]
  for i in range(n):
    fac.append(fac[i]*(i+1)%mod)
  idx = [-1] * n
  st1 = segt(n, 0, add)
  for p, k in pk:
    st1.update(k, 1)
    idx[p] = k
  st2 = segt(n, 0, add)
  st3 = segt(n, 0, add)
  s = 0
  ans = 0
  for i in range(n):
    #print(i)
    if idx[i] == -1:
      ans += s * fac[n-m-1] % mod
      #print(ans, end = " ")
      cnd = i - st2.prod(0, i)
      ans += cnd * ((n - m) * (n - m - 1) // 2) * fac[n-m-2] % mod
      #print(ans)
    else:
      st2.update(i, 1)
      st3.update(idx[i], 1)
      ans += st3.prod(idx[i]+1, n) * fac[n-m] % mod
      #print(ans, end = " ")
      cnd = i - st2.prod(0, i)
      ans += cnd * ((n - idx[i]) - st1.prod(idx[i], n)) * fac[n-m-1] % mod
      s += idx[i] - st1.prod(0, idx[i])
      #print(ans, s)
    ans %= mod
  return ans


from itertools import permutations
def naive(n, m, pk):
  P = [-1] * n
  for pi, ki in pk:
    P[ki] = pi
  ans = 0
  for p in permutations(range(n)):
    flg = 0
    for i in range(n):
      if P[i] != -1 and p[i] != P[i]:
        flg = 1
    if flg:
      continue
    for i in range(n):
      for j in range(i+1, n):
        ans += p[i] > p[j]
  return ans

print(solve(n, m, pk))

exit()
print(naive(n, m, pk))
print("=" * 10)

import random
for _ in range(100):
  n = random.randint(1, 10)
  usedp = [0] * n
  usedk = [0] * n
  pk = []
  for _ in range(random.randint(0, n*2)):
    p, k = random.randint(0, n-1), random.randint(0, n-1)
    if usedp[p] == 0 and usedk[k] == 0:
      pk.append([p, k])
      usedp[p] = 1
      usedk[k] = 1
  m = len(pk)
  if solve(n, m, pk) != naive(n, m, pk):
    print(n, m)
    for p, k in pk:
      print(p+1, k+1)
    break