#include <bits/stdc++.h>
using namespace std;
typedef signed long long ll;

#define _P(...) (void)printf(__VA_ARGS__)
#define FOR(x,to) for(x=0;x<(to);x++)
#define FORR(x,arr) for(auto& x:arr)
#define FORR2(x,y,arr) for(auto& [x,y]:arr)
#define ALL(a) (a.begin()),(a.end())
#define ZERO(a) memset(a,0,sizeof(a))
#define MINUS(a) memset(a,0xff,sizeof(a))
template<class T> bool chmax(T &a, const T &b) { if(a<b){a=b;return 1;}return 0;}
template<class T> bool chmin(T &a, const T &b) { if(a>b){a=b;return 1;}return 0;}
//-------------------------------------------------------

int T,N,P;
int A[202020];

vector<int> LP[202020];
vector<int> cand[202020];
int vis[202020];
int ret[202020];

int rev(int a,int m) {  // a*x=1 mod m
	int i;
	for(int i=1;i<m;i++) if(a*i%m==1) return i;
	return 0;
}

void solve() {
	int i,j,k,l,r,x,y; string s;
	
	cin>>T;
	while(T--) {
		cin>>N>>P;
		FOR(i,N) {
			cin>>A[i];
			A[i]--;
			LP[i].clear();
			vis[i]=0;
		}
		FOR(i,N) if(vis[i]==0) {
			
			x=i;
			while(1) {
				vis[x]=1;
				LP[i].push_back(x);
				if(A[x]==i) break;
				x=A[x];
			}
			cand[LP[i].size()].push_back(i);
		}
		
		for(i=1;i<=N;i++) {
			if(i%2==1) {
				int d=rev(2*P,i);
				FORR(v,cand[i]) {
					auto p=LP[v];
					FOR(j,i) ret[p[j]]=p[(j-d+i)%i];
				}
			}
			else {
				int d=rev(P,i);
				while(cand[i].size()>=2) {
					auto p=LP[cand[i].back()];
					cand[i].pop_back();
					auto q=LP[cand[i].back()];
					cand[i].pop_back();
					FOR(j,i) {
						ret[p[j]]=q[j];
						ret[q[j]]=p[(j-d+i)%i];
					}
				}
				if(cand[i].size()) {
					auto p=LP[cand[i].back()];
					int d=rev(2*P,i-1);
					FOR(j,i) ret[p[j]]=p[(j-d+i-1)%(i-1)];
				}
			}
			cand[i].clear();
		}
		
		FOR(i,N) cout<<ret[i]+1<<" ";
		cout<<endl;
		
	}
}


int main(int argc,char** argv){
	string s;int i;
	if(argc==1) ios::sync_with_stdio(false), cin.tie(0);
	FOR(i,argc-1) s+=argv[i+1],s+='\n'; FOR(i,s.size()) ungetc(s[s.size()-1-i],stdin);
	cout.tie(0); solve(); return 0;
}