#pragma GCC target ("avx")
#pragma GCC optimize("Ofast")
#pragma GCC optimize("unroll-loops")
//#pragma GCC target("sse,sse2,sse3,ssse3,sse4,popcnt,abm,mmx,avx,tune=native")
#define _USE_MATH_DEFINES
#include<iostream>
#include<string>
#include<queue>
#include<cmath>
#include<map>
#include<set>
#include<list>
#include<iomanip>
#include<vector>
#include<random>
#include<functional>
#include<algorithm>
#include<stack>
#include<cstdio>
#include<cstring>
#include<bitset>
#include<unordered_map>
#include<climits>
#include<fstream>
#include<complex>
#include<time.h>
#include<cassert>
#include<functional>
#include<numeric>
#include<tuple>
using namespace std;
using ll = long long;
using ld = long double;
#define int long long
#define all(a) (a).begin(),(a).end()
#define fs first
#define sc second
#define xx first
#define yy second.first
#define zz second.second
#define H pair<int, int>
#define P pair<int, pair<int, int>>
#define Q(i,j,k) mkp(i,mkp(j,k))
#define rng(i,s,n) for(int i = (s) ; i < (n) ; i++)
#define rep(i,n) rng(i, 0, (n))
#define mkp make_pair
#define vec vector
#define vi vec<int>
#define pb emplace_back
#define siz(a) (int)(a).size()
#define crdcomp(b) sort(all((b)));(b).erase(unique(all((b))),(b).end())
#define getidx(b,i) (lower_bound(all(b),(i))-(b).begin())
#define ssp(i,n) (i==(int)(n)-1?"\n":" ")
#define ctoi(c) (int)(c-'0')
#define itoc(c) (char)(c+'0')
#define cyes printf("Yes\n")
#define cno printf("No\n")
#define cdf(n) int quetimes_=(n);rep(qq123_,quetimes_)
#define gcj printf("Case #%lld: ",qq123_+1)
#define readv(a,n) a.resize(n,0);rep(i,(n)) a[i]=read()
#define found(a,x) (a.find(x)!=a.end())
//#define endl "\n"
constexpr int mod = (ll)1e9 + 7;
constexpr int Mod = 998244353;
constexpr ld EPS = 1e-10;
constexpr ll inf = (ll)3 * 1e18;
constexpr int Inf = (ll)15 * 1e8;
constexpr int dx[] = { -1,1,0,0 }, dy[] = { 0,0,-1,1 };
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 (b < a) { a = b; return 1; } return 0; }
ll read() { ll u, k = scanf("%lld", &u); return u; }
string reads() { string s; cin >> s; return s; }
H readh(short g = 0) { H u; int k = scanf("%lld %lld", &u.fs, &u.sc); if (g == 1) u.fs--, u.sc--; if (g == 2) u.fs--; return u; }
bool ina(H t, int h, int w) { return 0 <= t.fs && t.fs < h && 0 <= t.sc && t.sc < w; }
bool ina(int t, int l, int r) { return l <= t && t < r; }
ll gcd(ll i, ll j) { return j ? gcd(j, i % j) : i; }
ll popcount(ll x) {
    int sum = 0; for (int i = 0; i < 60; i++)if ((1ll << i) & x) sum++;
    return sum;
}
template<typename T>
class csum {
    vec<T> v;
public:
    csum(vec<T>& a) :v(a) { build(); }
    csum(){}
    void init(vec<T>& a) { v = a; build(); }
    void build() {
        for (int i = 1; i < v.size(); i++) v[i] += v[i - 1];
    }
    T a(int l, int r) {
        if (r < l) return 0;
        return v[r] - (l == 0 ? 0 : v[l - 1]);
    }//[l,r]
    T b(int l, int r) {
        return a(l, r - 1);
    }//[l,r)
    T a(pair<int, int>t) {
        return a(t.first, t.second);
    }
    T b(pair<int, int>t) {
        return b(t.first, t.second);
    }
};
class mint {
public:ll v;
      mint(ll v = 0) { s(v % mod + mod); }
      constexpr static int mod = (ll)1e9 + 7;
      constexpr static int fn_ = (ll)2e6 + 5;
      static mint fact[fn_], comp[fn_];
      mint pow(int x) const {
          mint b(v), c(1);
          while (x) {
              if (x & 1) c *= b;
              b *= b;
              x >>= 1;
          }
          return c;
      }
      inline mint& s(int vv) {
          v = vv < mod ? vv : vv - mod;
          return *this;
      }
      inline mint inv()const { return pow(mod - 2); }
      inline mint operator-()const { return mint() - *this; }
      inline mint& operator+=(const mint b) { return s(v + b.v); }
      inline mint& operator-=(const mint b) { return s(v + mod - b.v); }
      inline mint& operator*=(const mint b) { v = v * b.v % mod; return *this; }
      inline mint& operator/=(const mint b) { v = v * b.inv().v % mod; return *this; }
      inline mint operator+(const mint b) const { return mint(v) += b; }
      inline mint operator-(const mint b) const { return mint(v) -= b; }
      inline mint operator*(const mint b) const { return mint(v) *= b; }
      inline mint operator/(const mint b) const { return mint(v) /= b; }
      friend ostream& operator<<(ostream& os, const mint& m) {
          return os << m.v;
      }
      friend istream& operator>>(istream& is, mint& m) {
          int x; is >> x; m = mint(x);
          return is;
      }
      bool operator<(const mint& r)const { return v < r.v; }
      bool operator>(const mint& r)const { return v > r.v; }
      bool operator<=(const mint& r)const { return v <= r.v; }
      bool operator>=(const mint& r)const { return v >= r.v; }
      bool operator==(const mint& r)const { return v == r.v; }
      bool operator!=(const mint& r)const { return v != r.v; }
      explicit operator bool()const { return v; }
      explicit operator int()const { return v; }
      mint comb(mint k) {
          if (k > * this) return mint();
          if (!fact[0]) combinit();
          if (v >= fn_) {
              if (k > * this - k) k = *this - k;
              mint tmp(1);
              for (int i = v; i >= v - k.v + 1; i--) tmp *= mint(i);
              return tmp * comp[k.v];
          }
          return fact[v] * comp[k.v] * comp[v - k.v];
      }//nCk
      mint perm(mint k) {
          if (k > * this) return mint();
          if (!fact[0]) combinit();
          if (v >= fn_) {
              mint tmp(1);
              for (int i = v; i >= v - k.v + 1; i--) tmp *= mint(i);
              return tmp;
          }
          return fact[v] * comp[v - k.v];
      }//nPk
      static void combinit() {
          fact[0] = 1;
          for (int i = 1; i < fn_; i++) fact[i] = fact[i - 1] * mint(i);
          comp[fn_ - 1] = fact[fn_ - 1].inv();
          for (int i = fn_ - 2; i >= 0; i--) comp[i] = comp[i + 1] * mint(i + 1);
      }
}; mint mint::fact[fn_], mint::comp[fn_];
//--------------------------------------------------------------
class LCA {
#define H pair<int, int>
#define fs first
#define sc second
    int n;
    H table[400000][25];
    vec<H>e[400000];
    int dep[400000];
public:
    void init(const int& size) {
        n = size;
        for (int i = 0; i <= n; i++) {
            e[i].clear(); dep[i] = 0;
            for (int j = 0; j < 25; j++)
                table[i][j] = H{ -1,-1 };
        }
    }
    void add_edge(int u, int v, int r) {
        e[u].pb(H{ v,r });
        e[v].pb(H{ u,r });
    }
    void add_edge(int u, int v) {
        e[u].pb(H{ v,1 });
        e[v].pb(H{ u,1 });
    }
    void build(const int st) {
        dfs(st, -1, 0);
        for (int j = 0; j < 24; j++)for (int i = 0; i <= n; i++) {
            if (table[i][j].fs > -1) table[i][j + 1] = H{ table[table[i][j].fs][j].fs, table[i][j].sc + table[table[i][j].fs][j].sc };
        }
    }
    H get(int x, int y) {
        if (dep[x] > dep[y]) swap(x, y);
        int sum = 0;
        for (int i = 24; i >= 0; i--) {
            if (((dep[y] - dep[x]) >> i) & 1) {
                sum += table[y][i].sc;
                y = table[y][i].fs;
            }
        }
        if (x == y) return H{ x,sum };
        for (int i = 24; i >= 0; i--) {
            if (table[x][i].fs != table[y][i].fs) {
                sum += table[x][i].sc + table[y][i].sc;
                x = table[x][i].fs, y = table[y][i].fs;
            }
        }
        return H{ table[x][0].fs, sum + table[x][0].sc + table[y][0].sc };
    }
    int operator[](const int& x) const {
        return dep[x];
    }
private:
    void dfs(int v, int p, int d) {
        table[v][0] = H{ p,-1 };
        dep[v] = d;
        for (auto& to : e[v]) {
            if (to.fs != p) dfs(to.fs, v, d + 1);
            else table[v][0].sc = to.sc;
        }
    }
};
auto RUQ = [](int& num, int x, int width) {num = x; };
auto RAQ = [](int& num, int x, int width) {num += x; };
auto RCMXQ = [](int& num, int x, int width) {num = max(num, x); };
auto RCMNQ = [](int& num, int x, int width) {num = min(num, x); };
auto RASQ = [](int& num, int x, int width) {num += (x * width); };
auto RUSQ = [](int& num, int x, int width) {num = (x * width); };
auto RSQ = [](int x, int y)->int {return x + y; };
auto RMXQ = [](int x, int y)->int {return max(x, y); };
auto RMNQ = [](int x, int y)->int {return min(x, y); };
class Segtree {
#define SEG_SIZE 900000
    using F = function<void(int&, int, int)>;
    using T = function<int(int, int)>;
    int siz, rr, zer, zer2;
    int dat[SEG_SIZE], lazy[SEG_SIZE];
    bool updated[SEG_SIZE];
    F upd; T qur;
public:
    //for update, for query
    void init(int size, F update, T query, int zero, int zero2) {
        siz = size, upd = update, qur = query, zer = zero2, zer2 = zero;
        rr = 1; while (rr < size) rr *= 2;
        for (int i = 0; i < SEG_SIZE; i++) dat[i] = zer, lazy[i] = zer2, updated[i] = 0;
    }
    void rmnq(int n) { init(n, RUQ, RMNQ, 0, inf); }
    void rmxq(int n) { init(n, RUQ, RMXQ, 0, -inf); }
    template<class Iterator>
    void build(const Iterator st, const Iterator ed) {
        Iterator it = st; int cur = rr - 1;
        while (it != ed) dat[cur++] = (*it++);
        for (int i = rr - 2; i >= 0; i--)
            dat[i] = qur(dat[i * 2 + 1], dat[i * 2 + 2]);
    }
    void build(vector<int>v) {
        for (int i = 0; i < min((int)v.size(), siz); i++)
            dat[i + rr - 1] = v[i];
        for (int i = rr - 2; i >= 0; i--)
            dat[i] = qur(dat[i * 2 + 1], dat[i * 2 + 2]);
    }
    void update(int a, int b, int x) {
        update(0, a, b, 0, rr, x);
    }
    void change(int a, int x) {
        change2(a, x);
    }//一点更新
    int query(int a, int b) {
        return query(0, a, b, 0, rr);
    }
    int lower_bound(int a, int b, function<bool(int)>comp) {
        return lower_bound(0, a, b, 0, rr, comp);
    }
    int upper_bound(int a, int b, function<bool(int)>comp) {
        return upper_bound(0, a, b, 0, rr, comp);
    }
    int operator[](const int i) {
        return query(i, i + 1);
    }
private:
    void eval(int i, int l, int r) {
        if (!updated[i]) return;
        if (r - l > 1) {
            upd(lazy[i * 2 + 1], lazy[i], 1);
            upd(lazy[i * 2 + 2], lazy[i], 1);
            updated[i * 2 + 1] = updated[i * 2 + 2] = 1;
        }
        upd(dat[i], lazy[i], min(r, siz) - l);
        lazy[i] = zer2;
        updated[i] = 0;
    }
    void update(int i, int a, int b, int l, int r, int x) {
        eval(i, l, r);
        if (b <= l || r <= a) return;
        if (a <= l && r <= b) {
            upd(lazy[i], x, 1); updated[i] = 1;
            eval(i, l, r);
            return;
        }
        update(i * 2 + 1, a, b, l, (l + r) / 2, x);
        update(i * 2 + 2, a, b, (l + r) / 2, r, x);
        dat[i] = qur(dat[i * 2 + 1], dat[i * 2 + 2]);
    }
    void change2(int a, int x) {
        query(a, a + 1);
        int t = a + rr - 1;
        dat[t] = x;
        while (t > 0) {
            t = (t - 1) / 2;
            dat[t] = qur(dat[t * 2 + 1], dat[t * 2 + 2]);
        }
    }
    int query(int i, int a, int b, int l, int r) {
        eval(i, l, r);
        if (b <= l || r <= a) return zer;
        if (a <= l && r <= b) return dat[i];
        return qur(query(i * 2 + 1, a, b, l, (l + r) / 2),
            query(i * 2 + 2, a, b, (l + r) / 2, r));
    }
    int lower_bound(int i, int a, int b, int l, int r, function<bool(int)>comp) {
        eval(i, l, r);
        if (b <= l || r <= a || !comp(dat[i])) return siz;
        if (r - l == 1) return l;
        int tmp = lower_bound(i * 2 + 1, a, b, l, (l + r) / 2, comp);
        if (tmp < siz) return tmp;
        return lower_bound(i * 2 + 2, a, b, (l + r) / 2, r, comp);
    }
    int upper_bound(int i, int a, int b, int l, int r, function<bool(int)>comp) {
        eval(i, l, r);
        if (b <= l || r <= a || !comp(dat[i])) return 0;
        if (r - l == 1) return r;
        int tmp = upper_bound(i * 2 + 2, a, b, (l + r) / 2, r, comp);
        if (tmp > 0) return tmp;
        return upper_bound(i * 2 + 1, a, b, l, (l + r) / 2, comp);
    }
} seg;

//---------------------------------------------------------------------

int n, k;
vi a;
vec<H>e;
LCA lca;

mt19937 rnd(314159);

void generate(int mode, int num) {
    a.clear();
    e.clear();
    if (mode == 0) {
        cin >> n >> k;
        rep(i, k) a.pb(read() - 1);
        rep(i, n - 1) {
            e.pb(readh(1));
        }
    }
    else {

        n = rnd() % 100000 + 1, k = rnd() % n + 1;

        if (num <= 7) {
            n = n % (num + 1) + 1;
            k = k % n + 1;
        }
        else if (num <= 13) {
            n = n % 100 + 1;
            k = k % n + 1;
        }
        else if (num <= 34) {

        }
        else if (num <= 38) {
            n = 99800;
            k = 99800;
        }
        else {
            n = 100000, k = 100000;
        }
        vi tmp;
        rep(i, n) tmp.pb(i);
        shuffle(all(tmp), rnd);
        rep(i, k) {
            a.pb(tmp[i]);
        }

        rng(i, 1, n) {
            int t = rnd() % i;
            if (num == 14) {
                t = i - 1;
            }
            H g = H{ i + 1,t + 1 };
            e.pb(H{ i,t });
        }
    }
}
int guchoku() {
    lca.init(n);
    rep(i, n - 1) lca.add_edge(e[i].fs, e[i].sc);
    lca.build(0);
    int ans = k;
    int tmp = a[0];
    rep(i, k) tmp = lca.get(a[i], tmp).fs;
    ans += lca[tmp];

    rep(i, k)rng(j, i, k) { //[i,j]
        if (i == 0 && j == k - 1) continue;
        int t;
        if (i > 0) t = a[0];
        else t = a[j + 1];
        rep(r, i) t = lca.get(t, a[r]).fs;
        rng(r, j + 1, k) t = lca.get(t, a[r]).fs;
        chmax(ans, lca[t] + (k - (j - i + 1)));
    }
    return ans;
}

int solve() {
    //前半を固定すると、LCAは右端を伸ばすときに徐々に登っていき、最終的に右端と一致するようになる
    //一致した後の最大値は、RMQをすればよいから、セグ木
    lca.init(n);
    rep(i, n - 1) lca.add_edge(e[i].fs, e[i].sc);
    lca.build(0);
    seg.init(k, RUQ, RMXQ, 0, -inf);

    int ans = 0, ance = a[0];

    vi la(k, a[k - 1]);
    seg.update(k - 1, k, lca[a[k - 1]] + 1);
    ans = lca[a[k - 1]] + 1;

    for (int i = k - 2; i >= 0; i--) {
        la[i] = lca.get(la[i + 1], a[i]).fs;
        seg.update(i, i + 1, lca[la[i]] + k - i);
        chmax(ans, lca[la[i]] + k - i);
    }

    rep(i, k - 1) {
        ance = lca.get(ance, a[i]).fs;
        chmax(ans, i + 1 + lca[ance]);

        int tmp = lca.get(ance, a[k - 1]).fs;
        int ok = k - 1, ng = i, mid;
        while (ok - ng > 1) {
            mid = (ok + ng) / 2;
            if (lca.get(tmp, la[mid]).fs == tmp) ok = mid;
            else ng = mid;
        }
        chmax(ans, (k - ok) + i + 1 + lca[tmp]);

        chmax(ans, i + 1 + seg.query(i + 1, ok));

        chmax(ans, i + 1 + lca[ance]);
        //segは、lca[]+人数
    }
    return ans;
}
void edit(int num) {
    ofstream ofs("input/" + to_string(num) + ".txt");
    ofs << n << " " << k << endl;
    rep(i, k) ofs << a[i] + 1 << ssp(i, k);
    rep(i, n - 1) {
        ofs << e[i].fs + 1 << " " << e[i].sc + 1 << endl;
    }
}

signed main() {
    generate(0, 20);
    /*cout << n << " " << k << " " << q << endl;
    rep(i, k) cout << a[i]+1 << ssp(i, k);
    rep(i, n - 1) {
        cout << e[i].fs+1 << " " << e[i].sc+1 << endl;
    }
    rep(i, q) {
        if (queries[i].xx == 1) {
            cout << 1 << " " << queries[i].yy+1 << " " << queries[i].zz+1 << endl;
        }
        else {
            cout << 2 << " " << queries[i].yy+1 << endl;
        }
    }*/
    int ans = solve();
    cout << ans << endl;
    /*rep(i, 40) {
        generate(1, i + 1);
        //if (i == 0) continue;
        edit(i + 1);
        /*int u = guchoku();
        int v = solve();
        //ofstream ofs("output/" + to_string(i + 1) + ".txt");
        //for (auto g : v) ofs << g << endl;
        if (u != v) {
            cout << n << " " << k << endl;
            rep(i, k) cout << a[i] << ssp(i, k);
            rep(i, n - 1) cout << e[i].fs << " " << e[i].sc << endl;
            cout << u << endl;
            cout << endl;
            cout << v << endl;
            return 0;
        }
        if (i % 10 == 0) cout << i << endl;
    }*/
}