// It looks like nk^2 but is not. Code is very similar to Benq's because I was analyzing his code to understand why the idea I have thought is not nk^2. The reason is that it is not very probable to have two children.
#include "bits/stdc++.h"
#define ll long long
using namespace std;
const int sz = 2e5 + 10;
vector<int> aj[sz];
int vl[sz];
int m;
vector<int> comb(vector<int> a, vector<int> b) {
vector<int> c(min(m, (int)a.size() + (int)b.size()));
for (int i = 0; i < (int) a.size(); i ++)
for (int j = 0; j < (int) b.size() && i + j <= m; j ++)
c[i + j] = max(c[i + j], a[i] + b[j]);
return c;
}
vector<int> dfs(int v) {
vector<int> r = {0};
for (int u : aj[v])
r = comb(r, dfs(u));
if (aj[v].empty())
r.push_back(0);
r[1] = max(r[1], vl[v]);
return r;
}
int main() {
int n;
cin >> n >> m;
for (int i = 1; i <= n; i ++) {
int p;
cin >> p >> vl[i];
aj[p].push_back(i);
}
auto r = dfs(aj[0][0]);
cout << r[m] << '\n';
}
