#include <bits/stdc++.h>
#pragma GCC target("avx2")
#pragma GCC optimize("O3")
#pragma GCC optimize("unroll-loops")
using namespace std;
using ll = long long;
const ll mod = 998244353;
#define fi first
#define se second
#define rep(i,n) for(ll i=0;i<n;i++)
#define all(x) x.begin(),x.end()
#define faster ios::sync_with_stdio(false);cin.tie(nullptr)
ll f(ll x,ll a,ll va,ll b,ll vb){
ll ans=0;
if(va<vb){
swap(a,b);
swap(va,vb);
}
ans+=va*min(x,a);
x-=min(x,a);
ans+=vb*min(x,b);
x-=min(x,b);
return ans;
}
int main() {
ll N,K,X;
cin >> N >> K >> X;
vector<ll> l(N),t(N),r(N);
vector<ll> x;
rep(i,N){
cin >> l[i] >> t[i] >> r[i];
x.push_back(l[i]);
x.push_back(r[i]+1);
if(l[i]+t[i]<=r[i]){
x.push_back(l[i]+t[i]);
}
}
x.push_back(0);
x.push_back(1LL<<30);
sort(all(x));
x.erase(unique(all(x)),x.end());
ll M=x.size()-1;
vector<ll> A(M,0);
vector<ll> B(M,0);
rep(i,N){
ll a=lower_bound(all(x),l[i])-x.begin();
ll b=lower_bound(all(x),l[i]+t[i])-x.begin();
ll c=lower_bound(all(x),r[i]+1)-x.begin();
A[a]++;
A[c]--;
if(l[i]+t[i]<=r[i]){
B[b]++;
B[c]--;
}
}
rep(i,M-1) A[i+1]+=A[i],B[i+1]+=B[i];
vector<ll> cum(M,0);
if(A[0]>=K) cum[0]+=B[0]*(x[1]-x[0]);
for(ll i=1;i<M;i++){
cum[i]=cum[i-1];
if(A[i]>=K) cum[i]+=B[i]*(x[i+1]-x[i]);
}
ll ans=0;
for(ll i=1;i<M;i++){
ll ok=i,ng=M;
while(ng-ok>1){
ll mid=(ok+ng)/2;
if(x[mid]>x[i]+X) ng=mid;
else ok=mid;
}
ll len=X-(x[ok]-x[i]);
ll l=x[i]-x[i-1];
ll val_l=0;
if(A[i-1]>=K) val_l=B[i-1];
ll r=x[ok+1]-x[ok];
ll val_r=0;
if(A[ok]>=K) val_r=B[ok];
ll cnt=cum[ok-1]-cum[i-1]+f(len,l,val_l,r,val_r);
ans=max(ans,cnt);
if(i==ok) continue;
i++;
len=X-(x[ok]-x[i]);
l=x[i]-x[i-1];
val_l=0;
if(A[i-1]>=K) val_l=B[i-1];
r=x[ok+1]-x[ok];
val_r=0;
if(A[ok]>=K) val_r=B[ok];
cnt=cum[ok-1]-cum[i-1]+f(len,l,val_l,r,val_r);
ans=max(ans,cnt);
i--;
}
cout << ans << endl;
return 0;
}
