/*

 * Copyright (c) 2012, 2018, Oracle and/or its affiliates. All rights reserved.

 * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.

 *

 * This code is free software; you can redistribute it and/or modify it

 * under the terms of the GNU General Public License version 2 only, as

 * published by the Free Software Foundation.

 *

 * This code is distributed in the hope that it will be useful, but WITHOUT

 * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or

 * FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License

 * version 2 for more details (a copy is included in the LICENSE file that

 * accompanied this code).

 *

 * You should have received a copy of the GNU General Public License version

 * 2 along with this work; if not, write to the Free Software Foundation,

 * Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA.

 *

 * Please contact Oracle, 500 Oracle Parkway, Redwood Shores, CA 94065 USA

 * or visit www.oracle.com if you need additional information or have any

 * questions.

 *

 */

#include "precompiled.hpp"

#include "jvm.h"

#include "memory/allocation.inline.hpp"

#include "runtime/os.hpp"

#include "runtime/os_perf.hpp"

#include "os_solaris.inline.hpp"

#include "utilities/macros.hpp"

#include CPU_HEADER(vm_version_ext)

#include <sys/types.h>

#include <procfs.h>

#include <dirent.h>

#include <errno.h>

#include <stdio.h>

#include <stdlib.h>

#include <strings.h>

#include <unistd.h>

#include <fcntl.h>

#include <kstat.h>

#include <unistd.h>

#include <string.h>

#include <sys/sysinfo.h>

#include <sys/lwp.h>

#include <pthread.h>

#include <time.h>

#include <utmpx.h>

#include <dlfcn.h>

#include <sys/loadavg.h>

#include <limits.h>

static const double NANOS_PER_SEC = 1000000000.0;

struct CPUPerfTicks {

  kstat_t* kstat;

  uint64_t last_idle;

  uint64_t last_total;

  double   last_ratio;

};

struct CPUPerfCounters {

  int           nProcs;

  CPUPerfTicks* jvmTicks;

  kstat_ctl_t*  kstat_ctrl;

};

static int get_info(const char* path, void* info, size_t s, off_t o) {

  assert(path != NULL, "path is NULL!");

  assert(info != NULL, "info is NULL!");

  int fd = -1;

  if ((fd = open(path, O_RDONLY)) < 0) {

    return OS_ERR;

  }

  if (pread(fd, info, s, o) != s) {

    close(fd);

    return OS_ERR;

  }

  close(fd);

  return OS_OK;

}

static int get_psinfo2(void* info, size_t s, off_t o) {

  return get_info("/proc/self/psinfo", info, s, o);

}

static int get_psinfo(psinfo_t* info) {

  return get_psinfo2(info, sizeof(*info), 0);

}

static int get_psinfo(char* file, psinfo_t* info) {

  assert(file != NULL, "file is NULL!");

  assert(info != NULL, "info is NULL!");

  return get_info(file, info, sizeof(*info), 0);

}

static int get_usage(prusage_t* usage) {

  assert(usage != NULL, "usage is NULL!");

  return get_info("/proc/self/usage", usage, sizeof(*usage), 0);

}

static int read_cpustat(kstat_ctl_t* kstat_ctrl, CPUPerfTicks* load, cpu_stat_t* cpu_stat) {

  assert(kstat_ctrl != NULL, "kstat_ctrl pointer is NULL!");

  assert(load != NULL, "load pointer is NULL!");

  assert(cpu_stat != NULL, "cpu_stat pointer is NULL!");

  if (load->kstat == NULL) {

    // no handle.

    return OS_ERR;

  }

  if (kstat_read(kstat_ctrl, load->kstat, cpu_stat) == OS_ERR) {

    // disable handle for this CPU

     load->kstat = NULL;

     return OS_ERR;

  }

  return OS_OK;

}

static double get_cpu_load(int which_logical_cpu, CPUPerfCounters* counters) {

  assert(counters != NULL, "counters pointer is NULL!");

  cpu_stat_t  cpu_stat = {0};

  if (which_logical_cpu >= counters->nProcs) {

    return .0;

  }

  CPUPerfTicks load = counters->jvmTicks[which_logical_cpu];

  if (read_cpustat(counters->kstat_ctrl, &load, &cpu_stat) != OS_OK) {

    return .0;

  }

  uint_t* usage = cpu_stat.cpu_sysinfo.cpu;

  if (usage == NULL) {

    return .0;

  }

  uint64_t c_idle  = usage[CPU_IDLE];

  uint64_t c_total = 0;

  for (int i = 0; i < CPU_STATES; i++) {

    c_total += usage[i];

  }

  // Calculate diff against previous snapshot

  uint64_t d_idle  = c_idle - load.last_idle;

  uint64_t d_total = c_total - load.last_total;

  /** update if weve moved */

  if (d_total > 0) {

    // Save current values for next time around

    load.last_idle  = c_idle;

    load.last_total = c_total;

    load.last_ratio = (double) (d_total - d_idle) / d_total;

  }

  return load.last_ratio;

}

static int get_boot_time(uint64_t* time) {

  assert(time != NULL, "time pointer is NULL!");

  setutxent();

  for(;;) {

    struct utmpx* u;

    if ((u = getutxent()) == NULL) {

      break;

    }

    if (u->ut_type == BOOT_TIME) {

      *time = u->ut_xtime;

      endutxent();

      return OS_OK;

    }

  }

  endutxent();

  return OS_ERR;

}

static int get_noof_context_switches(CPUPerfCounters* counters, uint64_t* switches) {

  assert(switches != NULL, "switches pointer is NULL!");

  assert(counters != NULL, "counter pointer is NULL!");

  *switches = 0;

  uint64_t s = 0;

  // Collect data from all CPUs

  for (int i = 0; i < counters->nProcs; i++) {

    cpu_stat_t cpu_stat = {0};

    CPUPerfTicks load = counters->jvmTicks[i];

    if (read_cpustat(counters->kstat_ctrl, &load, &cpu_stat) == OS_OK) {

      s += cpu_stat.cpu_sysinfo.pswitch;

    } else {

      //fail fast...

      return OS_ERR;

    }

  }

  *switches = s;

  return OS_OK;

}

static int perf_context_switch_rate(CPUPerfCounters* counters, double* rate) {

  assert(counters != NULL, "counters is NULL!");

  assert(rate != NULL, "rate pointer is NULL!");

  static pthread_mutex_t contextSwitchLock = PTHREAD_MUTEX_INITIALIZER;

  static uint64_t lastTime = 0;

  static uint64_t lastSwitches = 0;

  static double   lastRate = 0.0;

  uint64_t lt = 0;

  int res = 0;

  if (lastTime == 0) {

    uint64_t tmp;

    if (get_boot_time(&tmp) < 0) {

      return OS_ERR;

    }

    lt = tmp * 1000;

  }

  res = OS_OK;

  pthread_mutex_lock(&contextSwitchLock);

  {

    uint64_t sw = 0;

    clock_t t, d;

    if (lastTime == 0) {

      lastTime = lt;

    }

    t = clock();

    d = t - lastTime;

    if (d == 0) {

      *rate = lastRate;

    } else if (get_noof_context_switches(counters, &sw)== OS_OK) {

      *rate      = ((double)(sw - lastSwitches) / d) * 1000;

      lastRate     = *rate;

      lastSwitches = sw;

      lastTime     = t;

    } else {

      *rate = 0.0;

      res   = OS_ERR;

    }

    if (*rate < 0.0) {

      *rate = 0.0;

      lastRate = 0.0;

    }

  }

  pthread_mutex_unlock(&contextSwitchLock);

  return res;

 }

class CPUPerformanceInterface::CPUPerformance : public CHeapObj<mtInternal> {

   friend class CPUPerformanceInterface;

 private:

  CPUPerfCounters _counters;

  int cpu_load(int which_logical_cpu, double* cpu_load);

  int context_switch_rate(double* rate);

  int cpu_load_total_process(double* cpu_load);

  int cpu_loads_process(double* pjvmUserLoad, double* pjvmKernelLoad, double* psystemTotalLoad);

  CPUPerformance();

  ~CPUPerformance();

  bool initialize();

};

CPUPerformanceInterface::CPUPerformance::CPUPerformance() {

  _counters.nProcs = 0;

  _counters.jvmTicks = NULL;

  _counters.kstat_ctrl = NULL;

}

bool CPUPerformanceInterface::CPUPerformance::initialize() {

  // initialize kstat control structure,

  _counters.kstat_ctrl = kstat_open();

  assert(_counters.kstat_ctrl != NULL, "error initializing kstat control structure!");

  if (NULL == _counters.kstat_ctrl) {

    return false;

  }

  // Get number of CPU(s)

  if ((_counters.nProcs = sysconf(_SC_NPROCESSORS_ONLN)) == OS_ERR) {

    // ignore error?

    _counters.nProcs = 1;

  }

  assert(_counters.nProcs > 0, "no CPUs detected in sysconf call!");

  if (_counters.nProcs == 0) {

    return false;

  }

  // Data structure(s) for saving CPU load (one per CPU)

  size_t tick_array_size = _counters.nProcs * sizeof(CPUPerfTicks);

  _counters.jvmTicks = (CPUPerfTicks*)NEW_C_HEAP_ARRAY(char, tick_array_size, mtInternal);

  if (NULL == _counters.jvmTicks) {

    return false;

  }

  memset(_counters.jvmTicks, 0, tick_array_size);

  // Get kstat cpu_stat counters for every CPU

  // loop over kstat to find our cpu_stat(s)

  int i = 0;

  for (kstat_t* kstat = _counters.kstat_ctrl->kc_chain; kstat != NULL; kstat = kstat->ks_next) {

    if (strncmp(kstat->ks_module, "cpu_stat", 8) == 0) {

      if (kstat_read(_counters.kstat_ctrl, kstat, NULL) == OS_ERR) {

        continue;

      }

      if (i == _counters.nProcs) {

        // more cpu_stats than reported CPUs

        break;

      }

      _counters.jvmTicks[i++].kstat = kstat;

    }

  }

  return true;

}

CPUPerformanceInterface::CPUPerformance::~CPUPerformance() {

  if (_counters.jvmTicks != NULL) {

    FREE_C_HEAP_ARRAY(char, _counters.jvmTicks);

  }

  if (_counters.kstat_ctrl != NULL) {

    kstat_close(_counters.kstat_ctrl);

  }

}

int CPUPerformanceInterface::CPUPerformance::cpu_load(int which_logical_cpu, double* cpu_load) {

  assert(cpu_load != NULL, "cpu_load pointer is NULL!");

  double t = .0;

  if (-1 == which_logical_cpu) {

    for (int i = 0; i < _counters.nProcs; i++) {

      t += get_cpu_load(i, &_counters);

    }

    // Cap total systemload to 1.0

    t = MIN2<double>((t / _counters.nProcs), 1.0);

  } else {

    t = MIN2<double>(get_cpu_load(which_logical_cpu, &_counters), 1.0);

  }

  *cpu_load = t;

  return OS_OK;

}

int CPUPerformanceInterface::CPUPerformance::cpu_load_total_process(double* cpu_load) {

  assert(cpu_load != NULL, "cpu_load pointer is NULL!");

  psinfo_t info;

  // Get the percentage of "recent cpu usage" from all the lwp:s in the JVM:s

  // process. This is returned as a value between 0.0 and 1.0 multiplied by 0x8000.

  if (get_psinfo2(&info.pr_pctcpu, sizeof(info.pr_pctcpu), offsetof(psinfo_t, pr_pctcpu)) != 0) {

    *cpu_load = 0.0;

    return OS_ERR;

  }

  *cpu_load = (double) info.pr_pctcpu / 0x8000;

  return OS_OK;

}

int CPUPerformanceInterface::CPUPerformance::cpu_loads_process(double* pjvmUserLoad, double* pjvmKernelLoad, double* psystemTotalLoad) {

  assert(pjvmUserLoad != NULL, "pjvmUserLoad not inited");

  assert(pjvmKernelLoad != NULL, "pjvmKernelLoad not inited");

  assert(psystemTotalLoad != NULL, "psystemTotalLoad not inited");

  static uint64_t lastTime;

  static uint64_t lastUser, lastKernel;

  static double lastUserRes, lastKernelRes;

  pstatus_t pss;

  psinfo_t  info;

  *pjvmKernelLoad = *pjvmUserLoad = *psystemTotalLoad = 0;

  if (get_info("/proc/self/status", &pss.pr_utime, sizeof(timestruc_t)*2, offsetof(pstatus_t, pr_utime)) != 0) {

    return OS_ERR;

  }

  if (get_psinfo(&info) != 0) {

    return OS_ERR;

  }

  // get the total time in user, kernel and total time

  // check ratios for 'lately' and multiply the 'recent load'.

  uint64_t time   = (info.pr_time.tv_sec * NANOS_PER_SEC) + info.pr_time.tv_nsec;

  uint64_t user   = (pss.pr_utime.tv_sec * NANOS_PER_SEC) + pss.pr_utime.tv_nsec;

  uint64_t kernel = (pss.pr_stime.tv_sec * NANOS_PER_SEC) + pss.pr_stime.tv_nsec;

  uint64_t diff   = time - lastTime;

  double load     = (double) info.pr_pctcpu / 0x8000;

  if (diff > 0) {

    lastUserRes = (load * (user - lastUser)) / diff;

    lastKernelRes = (load * (kernel - lastKernel)) / diff;

    // BUG9182835 - patch for clamping these values to sane ones.

    lastUserRes   = MIN2<double>(1, lastUserRes);

    lastUserRes   = MAX2<double>(0, lastUserRes);

    lastKernelRes = MIN2<double>(1, lastKernelRes);

    lastKernelRes = MAX2<double>(0, lastKernelRes);

  }

  double t = .0;

  cpu_load(-1, &t);

  // clamp at user+system and 1.0

  if (lastUserRes + lastKernelRes > t) {

    t = MIN2<double>(lastUserRes + lastKernelRes, 1.0);

  }

  *pjvmUserLoad   = lastUserRes;

  *pjvmKernelLoad = lastKernelRes;

  *psystemTotalLoad = t;

  lastTime   = time;

  lastUser   = user;

  lastKernel = kernel;

  return OS_OK;

}

int CPUPerformanceInterface::CPUPerformance::context_switch_rate(double* rate) {

  return perf_context_switch_rate(&_counters, rate);

}

CPUPerformanceInterface::CPUPerformanceInterface() {

  _impl = NULL;

}

bool CPUPerformanceInterface::initialize() {

  _impl = new CPUPerformanceInterface::CPUPerformance();

  return _impl != NULL && _impl->initialize();

}

CPUPerformanceInterface::~CPUPerformanceInterface(void) {

  if (_impl != NULL) {

    delete _impl;

  }

}

int CPUPerformanceInterface::cpu_load(int which_logical_cpu, double* cpu_load) const {

  return _impl->cpu_load(which_logical_cpu, cpu_load);

}

int CPUPerformanceInterface::cpu_load_total_process(double* cpu_load) const {

  return _impl->cpu_load_total_process(cpu_load);

}

int CPUPerformanceInterface::cpu_loads_process(double* pjvmUserLoad, double* pjvmKernelLoad, double* psystemTotalLoad) const {

  return _impl->cpu_loads_process(pjvmUserLoad, pjvmKernelLoad, psystemTotalLoad);

}

int CPUPerformanceInterface::context_switch_rate(double* rate) const {

  return _impl->context_switch_rate(rate);

}

class SystemProcessInterface::SystemProcesses : public CHeapObj<mtInternal> {

  friend class SystemProcessInterface;

 private:

  class ProcessIterator : public CHeapObj<mtInternal> {

    friend class SystemProcessInterface::SystemProcesses;

   private:

    DIR*           _dir;

    struct dirent* _entry;

    bool           _valid;

    ProcessIterator();

    ~ProcessIterator();

    bool initialize();

    bool is_valid() const { return _valid; }

    bool is_valid_entry(struct dirent* const entry) const;

    bool is_dir(const char* const name) const;

    char* allocate_string(const char* const str) const;

    int current(SystemProcess* const process_info);

    int next_process();

  };

  ProcessIterator* _iterator;

  SystemProcesses();

  bool initialize();

  ~SystemProcesses();

  //information about system processes

  int system_processes(SystemProcess** system_processes, int* no_of_sys_processes) const;

};

bool SystemProcessInterface::SystemProcesses::ProcessIterator::is_dir(const char* name) const {

  struct stat64 mystat;

  int ret_val = 0;

  ret_val = ::stat64(name, &mystat);

  if (ret_val < 0) {

    return false;

  }

  ret_val = S_ISDIR(mystat.st_mode);

  return ret_val > 0;

}

// if it has a numeric name, is a directory and has a 'psinfo' file in it

bool SystemProcessInterface::SystemProcesses::ProcessIterator::is_valid_entry(struct dirent* entry) const {

  // ignore the "." and ".." directories

  if ((strcmp(entry->d_name, ".") == 0) ||

      (strcmp(entry->d_name, "..") == 0)) {

    return false;

  }

  char buffer[PATH_MAX] = {0};

  uint64_t size = 0;

  bool result = false;

  FILE *fp = NULL;

  if (atoi(entry->d_name) != 0) {

    jio_snprintf(buffer, PATH_MAX, "/proc/%s", entry->d_name);

    if (is_dir(buffer)) {

      memset(buffer, 0, PATH_MAX);

      jio_snprintf(buffer, PATH_MAX, "/proc/%s/psinfo", entry->d_name);

      if ((fp = fopen(buffer, "r")) != NULL) {

        int nread = 0;

        psinfo_t psinfo_data;

        if ((nread = fread(&psinfo_data, 1, sizeof(psinfo_t), fp)) != -1) {

          // only considering system process owned by root