Blob Blame History Raw
/*
 * BSD LICENSE
 *
 * Copyright(c) 2014-2018 Intel Corporation. All rights reserved.
 * All rights reserved.
 *
 * Redistribution and use in source and binary forms, with or without
 * modification, are permitted provided that the following conditions
 * are met:
 *
 *   * Redistributions of source code must retain the above copyright
 *     notice, this list of conditions and the following disclaimer.
 *   * Redistributions in binary form must reproduce the above copyright
 *     notice, this list of conditions and the following disclaimer in
 *     the documentation and/or other materials provided with the
 *     distribution.
 *   * Neither the name of Intel Corporation nor the names of its
 *     contributors may be used to endorse or promote products derived
 *     from this software without specific prior written permission.
 *
 * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
 * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
 * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
 * A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
 * OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
 * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
 * LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
 * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
 * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
 * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
 * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 */

/**
 * @brief Host implementation of PQoS API / capabilities.
 *
 * This module is responsible for PQoS management and capability
 * functionalities.
 *
 * Management functions include:
 * - management includes initializing and shutting down all other sub-modules
 *   including: monitoring, allocation, log, cpuinfo and machine
 * - provide functions for safe access to PQoS API - this is required for
 *   allocation and monitoring modules which also implement PQoS API
 *
 * Capability functions:
 * - monitoring detection, this is to discover all monitoring event types.
 *   LLC occupancy is only supported now.
 * - LLC allocation detection, this is to discover last level cache
 *   allocation feature.
 * - A new targeted function has to be implemented to discover new allocation
 *   technology.
 */

#include <stdlib.h>
#include <string.h>
#include <fcntl.h>     /* O_CREAT */
#include <unistd.h>    /* usleep(), lockf() */
#include <sys/stat.h>  /* S_Ixxx */
#include <pthread.h>

#include "pqos.h"

#include "cap.h"
#include "allocation.h"
#include "monitoring.h"

#include "cpuinfo.h"
#include "machine.h"
#include "types.h"
#include "log.h"
#include "api.h"
#include "utils.h"
#include "resctrl.h"
#include "resctrl_alloc.h"
#include "perf_monitoring.h"

/**
 * ---------------------------------------
 * Local macros
 * ---------------------------------------
 */

/**
 * Available types of allocation resource ID's.
 * (matches CPUID enumeration)
 */
#define PQOS_RES_ID_L3_ALLOCATION    1       /**< L3 cache allocation */
#define PQOS_RES_ID_L2_ALLOCATION    2       /**< L2 cache allocation */
#define PQOS_RES_ID_MB_ALLOCATION    3       /**< Memory BW allocation */

#define PQOS_CPUID_CAT_CDP_BIT       2       /**< CDP supported bit */

#define PQOS_MSR_L3_QOS_CFG          0xC81   /**< L3 CAT config register */
#define PQOS_MSR_L3_QOS_CFG_CDP_EN   1ULL    /**< L3 CDP enable bit */

#define PQOS_MSR_L3CA_MASK_START     0xC90   /**< L3 CAT class 0 register */
#define PQOS_MSR_L3CA_MASK_END       0xD0F   /**< L3 CAT class 127 register */
#define PQOS_MSR_ASSOC               0xC8F   /**< CAT class to core association
                                                register */
#define PQOS_MSR_ASSOC_QECOS_SHIFT   32
#define PQOS_MSR_ASSOC_QECOS_MASK    0xffffffff00000000ULL

#define PQOS_MSR_L2_QOS_CFG          0xC82   /**< L2 CAT config register */
#define PQOS_MSR_L2_QOS_CFG_CDP_EN   1ULL    /**< L2 CDP enable bit */

#define PQOS_MSR_L2CA_MASK_START     0xC10   /**< L2 CAT class 0 register */
#define PQOS_MSR_L2CA_MASK_END       0xD8F   /**< L2 CAT class 127 register */

#ifndef LOCKFILE
#ifdef __linux__
#define LOCKFILE "/var/lock/libpqos"
#endif
#ifdef __FreeBSD__
#define LOCKFILE "/var/tmp/libpqos.lockfile"
#endif
#endif /*!LOCKFILE*/

#define PROC_CPUINFO "/proc/cpuinfo"

/**
 * ---------------------------------------
 * Local data types
 * ---------------------------------------
 */

/**
 * ---------------------------------------
 * Local data structures
 * ---------------------------------------
 */

/**
 * This pointer is allocated and initialized in this module.
 * Then other sub-modules get this pointer in order to retrieve
 * capability information.
 */
static struct pqos_cap *m_cap = NULL;

/**
 * This gets allocated and initialized in this module.
 * This hold information about CPU topology in PQoS format.
 */
static const struct pqos_cpuinfo *m_cpu = NULL;

/**
 * Library initialization status.
 */
static int m_init_done = 0;

/**
 * API thread/process safe access is secured through these locks.
 */
static int m_apilock = -1;
static pthread_mutex_t m_apilock_mutex;

/**
 * Interface status
 *   0  PQOS_INTER_MSR
 *   1  PQOS_INTER_OS
 *   2  PQOS_INTER_OS_RESCTRL_MON
 */
#ifdef __linux__
static int m_interface = PQOS_INTER_MSR;
#endif
/**
 * ---------------------------------------
 * Functions for safe multi-threading
 * ---------------------------------------
 */

/**
 * @brief Initalizes API locks
 *
 * @return Operation status
 * @retval 0 success
 * @retval -1 error
 */
static int
_pqos_api_init(void)
{

        const char *lock_filename = LOCKFILE;

        if (m_apilock != -1)
                return -1;

        m_apilock = open(lock_filename, O_WRONLY | O_CREAT,
                         S_IRUSR | S_IWUSR | S_IRGRP | S_IROTH);
        if (m_apilock == -1)
                return -1;

        if (pthread_mutex_init(&m_apilock_mutex, NULL) != 0) {
                close(m_apilock);
                m_apilock = -1;
                return -1;
        }

        return 0;
}

/**
 * @brief Uninitializes API locks
 *
 * @return Operation status
 * @retval 0 success
 * @retval -1 error
 */
static int
_pqos_api_exit(void)
{
        int ret = 0;

        if (close(m_apilock) != 0)
                ret = -1;

        if (pthread_mutex_destroy(&m_apilock_mutex) != 0)
                ret = -1;

        m_apilock = -1;

        return ret;
}

void
_pqos_api_lock(void)
{
        int err = 0;

        if (lockf(m_apilock, F_LOCK, 0) != 0)
                err = 1;

        if (pthread_mutex_lock(&m_apilock_mutex) != 0)
                err = 1;

        if (err)
                LOG_ERROR("API lock error!\n");
}

void
_pqos_api_unlock(void)
{
        int err = 0;

        if (lockf(m_apilock, F_ULOCK, 0) != 0)
                err = 1;

        if (pthread_mutex_unlock(&m_apilock_mutex) != 0)
                err = 1;

        if (err)
                LOG_ERROR("API unlock error!\n");
}

/**
 * ---------------------------------------
 * Function for library initialization
 * ---------------------------------------
 */

int
_pqos_check_init(const int expect)
{
        if (m_init_done && (!expect)) {
                LOG_ERROR("PQoS library already initialized\n");
                return PQOS_RETVAL_INIT;
        }

        if ((!m_init_done) && expect) {
                LOG_ERROR("PQoS library not initialized\n");
                return PQOS_RETVAL_INIT;
        }

        return PQOS_RETVAL_OK;
}

/*
 * =======================================
 * =======================================
 *
 * Capability discovery routines
 *
 * =======================================
 * =======================================
 */

/**
 * @brief Retrieves cache size and number of ways
 *
 * Retrieves information about cache from \a cache_info structure.
 *
 * @param cache_info cache information structure
 * @param num_ways place to store number of cache ways
 * @param size place to store cache size in bytes
 *
 * @return Operation status
 * @retval PQOS_RETVAL_OK success
 * @retval PQOS_RETVAL_PARAM incorrect parameters
 * @retval PQOS_RETVAL_RESOURCE cache not detected
 */
static int
get_cache_info(const struct pqos_cacheinfo *cache_info,
               unsigned *num_ways,
               unsigned *size)
{
        if (num_ways == NULL && size == NULL)
                return PQOS_RETVAL_PARAM;
        if (cache_info == NULL)
                return PQOS_RETVAL_PARAM;
        if (!cache_info->detected)
                return PQOS_RETVAL_RESOURCE;
        if (num_ways != NULL)
                *num_ways = cache_info->num_ways;
        if (size != NULL)
                *size = cache_info->total_size;
        return PQOS_RETVAL_OK;
}

/**
 * @brief Adds new event type to \a mon monitoring structure
 *
 * @param mon Monitoring structure which is to be updated with the new
 *        event type
 * @param res_id resource id
 * @param event_type event type
 * @param max_rmid max RMID for the event
 * @param scale_factor event specific scale factor
 * @param max_num_events maximum number of events that \a mon can accommodate
 */
static void
add_monitoring_event(struct pqos_cap_mon *mon,
                     const unsigned res_id,
                     const int event_type,
                     const unsigned max_rmid,
                     const uint32_t scale_factor,
                     const unsigned max_num_events)
{
        if (mon->num_events >= max_num_events) {
                LOG_WARN("%s() no space for event type %d (resource id %u)!\n",
                         __func__, event_type, res_id);
                return;
        }

        LOG_DEBUG("Adding monitoring event: resource ID %u, "
                  "type %d to table index %u\n",
                  res_id, event_type, mon->num_events);

        mon->events[mon->num_events].type = (enum pqos_mon_event) event_type;
        mon->events[mon->num_events].max_rmid = max_rmid;
        mon->events[mon->num_events].scale_factor = scale_factor;
        mon->events[mon->num_events].os_support = PQOS_OS_MON_UNSUPPORTED;
        mon->num_events++;
}

/**
 * @brief Discovers monitoring capabilities
 *
 * Runs series of CPUID instructions to discover system CMT
 * capabilities.
 * Allocates memory for monitoring structure and
 * returns it through \a r_mon to the caller.
 *
 * @param r_mon place to store created monitoring structure
 * @param cpu CPU topology structure
 *
 * @return Operation status
 * @retval PQOS_RETVAL_OK success
 * @retval PQOS_RETVAL_RESOURCE monitoring not supported
 * @retval PQOS_RETVAL_ERROR enumeration error
 */
static int
discover_monitoring(struct pqos_cap_mon **r_mon,
                    const struct pqos_cpuinfo *cpu)
{
        struct cpuid_out res, cpuid_0xa;
        struct cpuid_out cpuid_0xf_1;
        int ret = PQOS_RETVAL_OK;
        unsigned sz = 0, max_rmid = 0,
                l3_size = 0, num_events = 0;
        struct pqos_cap_mon *mon = NULL;

        ASSERT(r_mon != NULL && cpu != NULL);

        /**
         * Run CPUID.0x7.0 to check
         * for quality monitoring capability (bit 12 of ebx)
         */
        lcpuid(0x7, 0x0, &res);
        if (!(res.ebx & (1 << 12))) {
                LOG_WARN("CPUID.0x7.0: Monitoring capability not supported!\n");
                return PQOS_RETVAL_RESOURCE;
        }

        /**
         * We can go to CPUID.0xf.0 for further
         * exploration of monitoring capabilities
         */
        lcpuid(0xf, 0x0, &res);
        if (!(res.edx & (1 << 1))) {
                LOG_WARN("CPUID.0xf.0: Monitoring capability not supported!\n");
                return PQOS_RETVAL_RESOURCE;
        }

        /**
         * MAX_RMID for the socket
         */
        max_rmid = (unsigned) res.ebx + 1;
        ret = get_cache_info(&cpu->l3, NULL, &l3_size);  /**< L3 cache size */
        if (ret != PQOS_RETVAL_OK) {
                LOG_ERROR("Error reading L3 information!\n");
                return PQOS_RETVAL_ERROR;
        }

        /**
         * Check number of monitoring events to allocate memory for
         * Sub-leaf 1 provides information on monitoring.
         */
        lcpuid(0xf, 1, &cpuid_0xf_1); /**< query resource monitoring */

        if (cpuid_0xf_1.edx & 1)
                num_events++; /**< LLC occupancy */
        if (cpuid_0xf_1.edx & 2)
                num_events++; /**< total memory bandwidth event */
        if (cpuid_0xf_1.edx & 4)
                num_events++; /**< local memory bandwidth event */
        if ((cpuid_0xf_1.edx & 2) && (cpuid_0xf_1.edx & 4))
                num_events++; /**< remote memory bandwidth virtual event */

        if (!num_events)
                return PQOS_RETVAL_ERROR;

        /**
         * Check if IPC can be calculated & supported
         */
        lcpuid(0xa, 0x0, &cpuid_0xa);
        if (((cpuid_0xa.ebx & 3) == 0) && ((cpuid_0xa.edx & 31) > 1))
                num_events++;

        /**
         * This means we can program LLC misses too
         */
        if (((cpuid_0xa.eax >> 8) & 0xff) > 1)
                num_events++;

        /**
         * Allocate memory for detected events and
         * fill the events in.
         */
        sz = (num_events * sizeof(struct pqos_monitor)) + sizeof(*mon);
        mon = (struct pqos_cap_mon *) malloc(sz);
        if (mon == NULL)
                return PQOS_RETVAL_RESOURCE;

        memset(mon, 0, sz);
        mon->mem_size = sz;
        mon->max_rmid = max_rmid;
        mon->l3_size = l3_size;

        if (cpuid_0xf_1.edx & 1)
                add_monitoring_event(mon, 1, PQOS_MON_EVENT_L3_OCCUP,
                                     cpuid_0xf_1.ecx + 1, cpuid_0xf_1.ebx,
                                     num_events);
        if (cpuid_0xf_1.edx & 2)
                add_monitoring_event(mon, 1, PQOS_MON_EVENT_TMEM_BW,
                                     cpuid_0xf_1.ecx + 1, cpuid_0xf_1.ebx,
                                     num_events);
        if (cpuid_0xf_1.edx & 4)
                add_monitoring_event(mon, 1, PQOS_MON_EVENT_LMEM_BW,
                                     cpuid_0xf_1.ecx + 1, cpuid_0xf_1.ebx,
                                     num_events);

        if ((cpuid_0xf_1.edx & 2) && (cpuid_0xf_1.edx & 4))
                add_monitoring_event(mon, 1, PQOS_MON_EVENT_RMEM_BW,
                                     cpuid_0xf_1.ecx + 1, cpuid_0xf_1.ebx,
                                     num_events);

        if (((cpuid_0xa.ebx & 3) == 0) && ((cpuid_0xa.edx & 31) > 1))
                add_monitoring_event(mon, 0, PQOS_PERF_EVENT_IPC,
                                     0, 0, num_events);

        if (((cpuid_0xa.eax >> 8) & 0xff) > 1)
                add_monitoring_event(mon, 0, PQOS_PERF_EVENT_LLC_MISS,
                                     0, 0, num_events);

        (*r_mon) = mon;
        return PQOS_RETVAL_OK;
}

/**
 * @brief Checks L3 CDP enable status across all CPU sockets
 *
 * It also validates if L3 CDP enabling is consistent across
 * CPU sockets.
 * At the moment, such scenario is considered as error
 * that requires CAT reset.
 *
 * @param cpu detected CPU topology
 * @param enabled place to store L3 CDP enabling status
 *
 * @return Operations status
 * @retval PQOS_RETVAL_OK on success
 */
static int
l3cdp_is_enabled(const struct pqos_cpuinfo *cpu,
               int *enabled)
{
        unsigned *sockets = NULL;
        unsigned sockets_num = 0, j = 0;
        unsigned enabled_num = 0, disabled_num = 0;
        int ret = PQOS_RETVAL_OK;

        ASSERT(enabled != NULL && cpu != NULL);
        if (enabled == NULL || cpu == NULL)
                return PQOS_RETVAL_PARAM;

        *enabled = 0;

        /**
         * Get list of socket id's
         */
	sockets = pqos_cpu_get_sockets(cpu, &sockets_num);
        if (sockets == NULL)
                return PQOS_RETVAL_RESOURCE;

        for (j = 0; j < sockets_num; j++) {
                uint64_t reg = 0;
                unsigned core = 0;

                ret = pqos_cpu_get_one_core(cpu, sockets[j], &core);
                if (ret != PQOS_RETVAL_OK)
                        goto l3cdp_is_enabled_exit;

                if (msr_read(core, PQOS_MSR_L3_QOS_CFG, &reg) !=
                    MACHINE_RETVAL_OK) {
                        ret = PQOS_RETVAL_ERROR;
                        goto l3cdp_is_enabled_exit;
                }

                if (reg & PQOS_MSR_L3_QOS_CFG_CDP_EN)
                        enabled_num++;
                else
                        disabled_num++;
        }

        if (disabled_num > 0 && enabled_num > 0) {
                LOG_ERROR("Inconsistent L3 CDP settings across sockets."
                          "Please reset CAT or reboot your system!\n");
                ret = PQOS_RETVAL_ERROR;
                goto l3cdp_is_enabled_exit;
        }

        if (enabled_num > 0)
                *enabled = 1;

        LOG_INFO("L3 CDP is %s\n", (*enabled) ? "enabled" : "disabled");

 l3cdp_is_enabled_exit:
        if (sockets != NULL)
                free(sockets);

        return ret;
}

/**
 * @brief Checks L2 CDP enable status across all CPU clusters
 *
 * It also validates if L2 CDP enabling is consistent across
 * CPU clusters.
 * At the moment, such scenario is considered as error
 * that requires CAT reset.
 *
 * @param cpu detected CPU topology
 * @param enabled place to store L2 CDP enabling status
 *
 * @return Operations status
 * @retval PQOS_RETVAL_OK on success
 */
static int
l2cdp_is_enabled(const struct pqos_cpuinfo *cpu, int *enabled)
{
        unsigned *l2ids = NULL;
        unsigned l2id_num = 0;
        unsigned enabled_num = 0, disabled_num = 0;
        unsigned i;
        int ret = PQOS_RETVAL_OK;

        /**
         * Get list of L2 clusters id's
         */
        l2ids = pqos_cpu_get_l2ids(cpu, &l2id_num);
        if (l2ids == NULL || l2id_num == 0) {
                ret = PQOS_RETVAL_RESOURCE;
                goto l2cdp_is_enabled_exit;
        }

        for (i = 0; i < l2id_num; i++) {
                uint64_t reg = 0;
                unsigned core = 0;

                ret = pqos_cpu_get_one_by_l2id(cpu, l2ids[i], &core);
                if (ret != PQOS_RETVAL_OK)
                        goto l2cdp_is_enabled_exit;

                if (msr_read(core, PQOS_MSR_L2_QOS_CFG, &reg) !=
                    MACHINE_RETVAL_OK) {
                        ret = PQOS_RETVAL_ERROR;
                        goto l2cdp_is_enabled_exit;
                }

                if (reg & PQOS_MSR_L2_QOS_CFG_CDP_EN)
                        enabled_num++;
                else
                        disabled_num++;
        }

        if (disabled_num > 0 && enabled_num > 0) {
                LOG_ERROR("Inconsistent L2 CDP settings across clusters."
                          "Please reset CAT or reboot your system!\n");
                ret = PQOS_RETVAL_ERROR;
                goto l2cdp_is_enabled_exit;
        }

        if (enabled_num > 0)
                *enabled = 1;

        LOG_INFO("L2 CDP is %s\n", (*enabled) ? "enabled" : "disabled");

 l2cdp_is_enabled_exit:
        if (l2ids != NULL)
                free(l2ids);

        return ret;
}

/**
 * @brief Detects presence of L3 CAT based on register probing.
 *
 * This method of detecting CAT does the following steps.
 * - probe COS registers one by one and exit on first error
 * - if procedure fails on COS0 then CAT is not supported
 * - use CPUID.0x4.0x3 to get number of cache ways
 *
 * @param cap CAT structure to be initialized
 * @param cpu CPU topology structure
 *
 * @return Operation status
 * @retval PQOS_RETVAL_OK success
 * @retval PQOS_RETVAL_PARAM invalid input configuration/parameters
 * @retval PQOS_RETVAL_RESOURCE technology not supported
 */
static int
discover_alloc_l3_probe(struct pqos_cap_l3ca *cap,
                        const struct pqos_cpuinfo *cpu)
{
        unsigned i = 0, lcore;
        const unsigned max_classes =
                PQOS_MSR_L3CA_MASK_END - PQOS_MSR_L3CA_MASK_START + 1;

        ASSERT(cap != NULL && cpu != NULL);

        /**
         * Pick a valid core and run series of MSR reads on it
         */
        lcore = cpu->cores[0].lcore;
        for (i = 0; i < max_classes; i++) {
                int msr_ret;
                uint64_t value;

                msr_ret = msr_read(lcore, PQOS_MSR_L3CA_MASK_START + i, &value);
                if (msr_ret != MACHINE_RETVAL_OK)
                        break;
        }

        if (i == 0) {
		LOG_WARN("Error probing COS0 on core %u\n", lcore);
                return PQOS_RETVAL_RESOURCE;
        }

        /**
         * Number of ways and CBM is detected with CPUID.0x4.0x3 later on
         */
        cap->num_classes = i;
        return PQOS_RETVAL_OK;
}

/**
 * @brief Detects presence of L3 CAT based on brand string.
 *
 * If CPUID.0x7.0 doesn't report CAT feature
 * platform may still support it:
 * - we need to check brand string vs known ones
 * - use CPUID.0x4.0x3 to get number of cache ways
 *
 * @param cap CAT structure to be initialized
 *
 * @return Operation status
 * @retval PQOS_RETVAL_OK success
 * @retval PQOS_RETVAL_RESOURCE technology not supported
 */
static int
discover_alloc_l3_brandstr(struct pqos_cap_l3ca *cap)
{
#define CPUID_LEAF_BRAND_START 0x80000002UL
#define CPUID_LEAF_BRAND_END   0x80000004UL
#define CPUID_LEAF_BRAND_NUM   (CPUID_LEAF_BRAND_END-CPUID_LEAF_BRAND_START+1)
#define MAX_BRAND_STRING_LEN   (CPUID_LEAF_BRAND_NUM*4*sizeof(uint32_t))
        static const char * const supported_brands[] = {
                "E5-2658 v3",
                "E5-2648L v3", "E5-2628L v3",
                "E5-2618L v3", "E5-2608L v3",
                "E5-2658A v3", "E3-1258L v4",
                "E3-1278L v4"
        };
        struct cpuid_out res;
        int ret = PQOS_RETVAL_OK,
                match_found = 0;
        uint32_t brand[MAX_BRAND_STRING_LEN / 4 + 1];
        char *brand_str = (char *)brand;
        uint32_t *brand_u32 = brand;
        unsigned i = 0;

        /**
         * Assume \a cap is not NULL at this stage.
         * Adequate check has to be done in the caller.
         */
        ASSERT(cap != NULL);

        lcpuid(0x80000000, 0, &res);
        if (res.eax < CPUID_LEAF_BRAND_END) {
                LOG_ERROR("Brand string CPU-ID extended functions "
                          "not supported\n");
                return PQOS_RETVAL_ERROR;
        }

        memset(brand, 0, sizeof(brand));

        for (i = 0; i < CPUID_LEAF_BRAND_NUM; i++) {
                lcpuid((unsigned)CPUID_LEAF_BRAND_START + i, 0, &res);
                *brand_u32++ = res.eax;
                *brand_u32++ = res.ebx;
                *brand_u32++ = res.ecx;
                *brand_u32++ = res.edx;
        }

        LOG_DEBUG("CPU brand string '%s'\n", brand_str);

        /**
         * match brand against supported ones
         */
        for (i = 0; i < DIM(supported_brands); i++)
                if (strstr(brand_str, supported_brands[i]) != NULL) {
                        LOG_INFO("Cache allocation detected for model name "
                                 "'%s'\n", brand_str);
                        match_found = 1;
                        break;
                }

        if (!match_found) {
		LOG_WARN("Cache allocation not supported on model name '%s'!\n",
                         brand_str);
                return PQOS_RETVAL_RESOURCE;
        }

        /**
         * Figure out number of ways and CBM (1:1)
         * using CPUID.0x4.0x3
         */
        cap->num_classes = 4;

        return ret;
}

/**
 * @brief Detects presence of L3 CAT based on CPUID
 *
 * @param cap CAT structure to be initialized
 * @param cpu CPU topology structure
 *
 * @return Operation status
 * @retval PQOS_RETVAL_OK success
 * @retval PQOS_RETVAL_RESOURCE technology not supported
 */
static int
discover_alloc_l3_cpuid(struct pqos_cap_l3ca *cap,
                        const struct pqos_cpuinfo *cpu)
{
        struct cpuid_out res;
        int ret = PQOS_RETVAL_OK;

        /**
         * We can go to CPUID.0x10.0 to explore
         * allocation capabilities
         */
        lcpuid(0x10, 0x0, &res);
        if (!(res.ebx & (1 << PQOS_RES_ID_L3_ALLOCATION))) {
                LOG_INFO("CPUID.0x10.0: L3 CAT not detected.\n");
                return PQOS_RETVAL_RESOURCE;
        }

        /**
         * L3 CAT detected
         * - get more info about it
         */
        lcpuid(0x10, PQOS_RES_ID_L3_ALLOCATION, &res);
        cap->num_classes = res.edx + 1;
        cap->num_ways = res.eax + 1;
        cap->cdp = (res.ecx >> PQOS_CPUID_CAT_CDP_BIT) & 1;
        cap->cdp_on = 0;
	cap->os_cdp = 0;
        cap->way_contention = (uint64_t) res.ebx;

        if (cap->cdp) {
                /**
                 * CDP is supported but is it on?
                 */
                int cdp_on = 0;

                ret = l3cdp_is_enabled(cpu, &cdp_on);
                if (ret != PQOS_RETVAL_OK) {
                        LOG_ERROR("L3 CDP detection error!\n");
                        return ret;
                }
                cap->cdp_on = cdp_on;
                if (cdp_on)
                        cap->num_classes = cap->num_classes / 2;
        }

        return ret;
}

/**
 * @brief Discovers L3 CAT
 *
 * First it tries to detects CAT through CPUID.0x7.0
 * if this fails then falls into brand string check.
 *
 * Function allocates memory for CAT capabilities
 * and returns it to the caller through \a r_cap.
 *
 * \a cpu is only needed to detect CDP status.
 *
 * @param r_cap place to store CAT capabilities structure
 * @param cpu detected cpu topology
 *
 * @return Operation status
 * @retval PQOS_RETVAL_OK success
 */
static int
discover_alloc_l3(struct pqos_cap_l3ca **r_cap,
                  const struct pqos_cpuinfo *cpu)
{
        struct cpuid_out res;
        struct pqos_cap_l3ca *cap = NULL;
        const unsigned sz = sizeof(*cap);
        unsigned l3_size = 0;
        int ret = PQOS_RETVAL_OK;

        cap = (struct pqos_cap_l3ca *)malloc(sz);
        if (cap == NULL)
                return PQOS_RETVAL_RESOURCE;

        ASSERT(cap != NULL);

        memset(cap, 0, sz);
        cap->mem_size = sz;

        /**
         * Run CPUID.0x7.0 to check
         * for allocation capability (bit 15 of ebx)
         */
        lcpuid(0x7, 0x0, &res);

        if (res.ebx & (1 << 15)) {
                /**
                 * Use CPUID method
                 */
                LOG_INFO("CPUID.0x7.0: L3 CAT supported\n");
                ret = discover_alloc_l3_cpuid(cap, cpu);
                if (ret == PQOS_RETVAL_OK)
                        ret = get_cache_info(&cpu->l3, NULL, &l3_size);
        } else {
                /**
                 * Use brand string matching method 1st.
                 * If it fails then try register probing.
                 */
                LOG_INFO("CPUID.0x7.0: L3 CAT not detected. "
			 "Checking brand string...\n");
                ret = discover_alloc_l3_brandstr(cap);
                if (ret != PQOS_RETVAL_OK)
                        ret = discover_alloc_l3_probe(cap, cpu);
                if (ret == PQOS_RETVAL_OK)
                        ret = get_cache_info(&cpu->l3, &cap->num_ways,
                                             &l3_size);
        }

        if (ret == PQOS_RETVAL_OK) {
                if (cap->num_ways > 0)
                        cap->way_size = l3_size / cap->num_ways;
                LOG_INFO("L3 CAT details: CDP support=%d, CDP on=%d, "
                         "#COS=%u, #ways=%u, ways contention bit-mask 0x%x\n",
                         cap->cdp, cap->cdp_on, cap->num_classes,
                         cap->num_ways, cap->way_contention);
                LOG_INFO("L3 CAT details: cache size %u bytes, "
                         "way size %u bytes\n", l3_size, cap->way_size);
        }

        if (ret == PQOS_RETVAL_OK)
                (*r_cap) = cap;
        else
                free(cap);

        return ret;
}

/**
 * @brief Discovers L2 CAT
 *
 * @param r_cap place to store L2 CAT capabilities structure
 * @param cpu CPU topology structure
 *
 * @return Operation status
 * @retval PQOS_RETVAL_OK success
 */
static int
discover_alloc_l2(struct pqos_cap_l2ca **r_cap,
                  const struct pqos_cpuinfo *cpu)
{
        struct cpuid_out res;
        struct pqos_cap_l2ca *cap = NULL;
        const unsigned sz = sizeof(*cap);
	unsigned l2_size = 0;
        int ret = PQOS_RETVAL_OK;

        ASSERT(cpu != NULL);

        cap = (struct pqos_cap_l2ca *)malloc(sz);
        if (cap == NULL)
                return PQOS_RETVAL_RESOURCE;

        ASSERT(cap != NULL);

        memset(cap, 0, sz);
        cap->mem_size = sz;

        /**
         * Run CPUID.0x7.0 to check
         * for allocation capability (bit 15 of ebx)
         */
        lcpuid(0x7, 0x0, &res);
        if (!(res.ebx & (1 << 15))) {
                LOG_INFO("CPUID.0x7.0: L2 CAT not supported\n");
                free(cap);
                return PQOS_RETVAL_RESOURCE;
	}

        /**
         * We can go to CPUID.0x10.0 to obtain more info
         */
        lcpuid(0x10, 0x0, &res);
        if (!(res.ebx & (1 << PQOS_RES_ID_L2_ALLOCATION))) {
		LOG_INFO("CPUID 0x10.0: L2 CAT not supported!\n");
                free(cap);
                return PQOS_RETVAL_RESOURCE;
	}

	lcpuid(0x10, PQOS_RES_ID_L2_ALLOCATION, &res);

	cap->num_classes = res.edx+1;
	cap->num_ways = res.eax+1;
        cap->cdp = (res.ecx >> PQOS_CPUID_CAT_CDP_BIT) & 1;
        cap->cdp_on = 0;
	cap->os_cdp = 0;
	cap->way_contention = (uint64_t) res.ebx;

        if (cap->cdp) {
                int cdp_on = 0;

                /*
                 * Check if L2 CDP is enabled
                 */
                ret = l2cdp_is_enabled(cpu, &cdp_on);
                if (ret != PQOS_RETVAL_OK) {
                        LOG_ERROR("L2 CDP detection error!\n");
                        free(cap);
                        return ret;
                }
                cap->cdp_on = cdp_on;
                if (cdp_on)
                        cap->num_classes = cap->num_classes / 2;
        }

	ret = get_cache_info(&cpu->l2, NULL, &l2_size);
	if (ret != PQOS_RETVAL_OK) {
		LOG_ERROR("Error reading L2 info!\n");
                free(cap);
		return PQOS_RETVAL_ERROR;
	}
	if (cap->num_ways > 0)
		cap->way_size = l2_size / cap->num_ways;

	LOG_INFO("L2 CAT details: CDP support=%d, CDP on=%d, "
		 "#COS=%u, #ways=%u, ways contention bit-mask 0x%x\n",
		 cap->cdp, cap->cdp_on, cap->num_classes, cap->num_ways,
                 cap->way_contention);
	LOG_INFO("L2 CAT details: cache size %u bytes, way size %u bytes\n",
		 l2_size, cap->way_size);

	(*r_cap) = cap;
        return ret;
}

/**
 * @brief Discovers MBA
 *
 * @param r_cap place to store MBA capabilities structure
 *
 * @return Operation status
 * @retval PQOS_RETVAL_OK success
 */
static int
discover_alloc_mba(struct pqos_cap_mba **r_cap)
{
        struct cpuid_out res;
        struct pqos_cap_mba *cap = NULL;
        const unsigned sz = sizeof(*cap);
        int ret = PQOS_RETVAL_OK;

        cap = (struct pqos_cap_mba *)malloc(sz);
        if (cap == NULL)
                return PQOS_RETVAL_RESOURCE;

        ASSERT(cap != NULL);

        memset(cap, 0, sz);
        cap->mem_size = sz;

        /**
         * Run CPUID.0x7.0 to check
         * for allocation capability (bit 15 of ebx)
         */
        lcpuid(0x7, 0x0, &res);
        if (!(res.ebx & (1 << 15))) {
                LOG_INFO("CPUID.0x7.0: MBA not supported\n");
                free(cap);
                return PQOS_RETVAL_RESOURCE;
	}

        /**
         * We can go to CPUID.0x10.0 to obtain more info
         */
        lcpuid(0x10, 0x0, &res);
        if (!(res.ebx & (1 << PQOS_RES_ID_MB_ALLOCATION))) {
		LOG_INFO("CPUID 0x10.0: MBA not supported!\n");
                free(cap);
                return PQOS_RETVAL_RESOURCE;
	}

	lcpuid(0x10, PQOS_RES_ID_MB_ALLOCATION, &res);

	cap->num_classes = (res.edx & 0xffff) + 1;
        cap->throttle_max = (res.eax & 0xfff) + 1;
        cap->is_linear = (res.ecx >> 2) & 1;
        if (cap->is_linear)
                cap->throttle_step = 100 - cap->throttle_max;
        else {
                LOG_WARN("MBA non-linear mode not supported yet!\n");
                free(cap);
                return PQOS_RETVAL_RESOURCE;
        }

	LOG_INFO("MBA details: "
		 "#COS=%u, %slinear, max=%u, step=%u\n",
		 cap->num_classes,
                 cap->is_linear ? "" : "non-",
                 cap->throttle_max, cap->throttle_step);

	(*r_cap) = cap;
        return ret;
}

/**
 * @brief Runs detection of platform monitoring and allocation capabilities
 *
 * @param p_cap place to store allocated capabilities structure
 * @param cpu detected cpu topology
 *
 * @return Operation status
 * @retval PQOS_RETVAL_OK success
 */
static int
discover_capabilities(struct pqos_cap **p_cap,
                      const struct pqos_cpuinfo *cpu)
{
        struct pqos_cap_mon *det_mon = NULL;
        struct pqos_cap_l3ca *det_l3ca = NULL;
        struct pqos_cap_l2ca *det_l2ca = NULL;
        struct pqos_cap_mba *det_mba = NULL;
        struct pqos_cap *_cap = NULL;
        struct pqos_capability *item = NULL;
        unsigned sz = 0;
        int ret = PQOS_RETVAL_OK;

        /**
         * Monitoring init
         */
        ret = discover_monitoring(&det_mon, cpu);
        switch (ret) {
        case PQOS_RETVAL_OK:
                LOG_INFO("Monitoring capability detected\n");
                sz += sizeof(struct pqos_capability);
                break;
        case PQOS_RETVAL_RESOURCE:
                LOG_INFO("Monitoring capability not detected\n");
                break;
        default:
                LOG_ERROR("Error encounter in monitoring discovery!\n");
                ret = PQOS_RETVAL_ERROR;
                goto error_exit;
        }

        /**
         * L3 Cache allocation init
         */
        ret = discover_alloc_l3(&det_l3ca, cpu);
        switch (ret) {
        case PQOS_RETVAL_OK:
                LOG_INFO("L3CA capability detected\n");
                sz += sizeof(struct pqos_capability);
                break;
        case PQOS_RETVAL_RESOURCE:
                LOG_INFO("L3CA capability not detected\n");
                break;
        default:
                LOG_ERROR("Fatal error encounter in L3 CAT discovery!\n");
                ret = PQOS_RETVAL_ERROR;
                goto error_exit;
        }

        /**
         * L2 Cache allocation init
         */
        ret = discover_alloc_l2(&det_l2ca, cpu);
        switch (ret) {
        case PQOS_RETVAL_OK:
                LOG_INFO("L2CA capability detected\n");
                sz += sizeof(struct pqos_capability);
                break;
        case PQOS_RETVAL_RESOURCE:
                LOG_INFO("L2CA capability not detected\n");
                break;
        default:
                LOG_ERROR("Fatal error encounter in L2 CAT discovery!\n");
                ret = PQOS_RETVAL_ERROR;
                goto error_exit;
        }

        /**
         * Memory bandwidth allocation init
         */
        ret = discover_alloc_mba(&det_mba);
        switch (ret) {
        case PQOS_RETVAL_OK:
                LOG_INFO("MBA capability detected\n");
                sz += sizeof(struct pqos_capability);
                break;
        case PQOS_RETVAL_RESOURCE:
                LOG_INFO("MBA capability not detected\n");
                break;
        default:
                LOG_ERROR("Fatal error encounter in MBA discovery!\n");
                ret = PQOS_RETVAL_ERROR;
                goto error_exit;
        }

        if (sz == 0) {
                LOG_ERROR("No Platform QoS capability discovered\n");
                ret = PQOS_RETVAL_ERROR;
                goto error_exit;
        }

        sz += sizeof(struct pqos_cap);
        _cap = (struct pqos_cap *)malloc(sz);
        if (_cap == NULL) {
                LOG_ERROR("Allocation error in %s()\n", __func__);
                ret = PQOS_RETVAL_ERROR;
                goto error_exit;
        }

        memset(_cap, 0, sz);
        _cap->mem_size = sz;
        _cap->version = PQOS_VERSION;
        item = &_cap->capabilities[0];

        if (det_mon != NULL) {
                _cap->num_cap++;
                item->type = PQOS_CAP_TYPE_MON;
                item->u.mon = det_mon;
                item++;
                ret = PQOS_RETVAL_OK;
        }

        if (det_l3ca != NULL) {
                _cap->num_cap++;
                item->type = PQOS_CAP_TYPE_L3CA;
                item->u.l3ca = det_l3ca;
                item++;
                ret = PQOS_RETVAL_OK;
        }

        if (det_l2ca != NULL) {
                _cap->num_cap++;
                item->type = PQOS_CAP_TYPE_L2CA;
                item->u.l2ca = det_l2ca;
                item++;
                ret = PQOS_RETVAL_OK;
        }

        if (det_mba != NULL) {
                _cap->num_cap++;
                item->type = PQOS_CAP_TYPE_MBA;
                item->u.mba = det_mba;
                item++;
                ret = PQOS_RETVAL_OK;
        }

        (*p_cap) = _cap;

 error_exit:
        if (ret != PQOS_RETVAL_OK) {
                if (det_mon != NULL)
                        free(det_mon);
                if (det_l3ca != NULL)
                        free(det_l3ca);
                if (det_l2ca != NULL)
                        free(det_l2ca);
                if (det_mba != NULL)
                        free(det_mba);
        }

        return ret;
}

#ifdef __linux__
/**
 * @brief Checks file fname to detect str and set a flag
 *
 * @param fname name of the file to be searched
 * @param str string being searched for
 * @param supported pointer to os_supported flag
 *
 * @return Operation status
 * @retval PQOS_RETVAL_OK success
 */
static int
detect_os_support(const char *fname, const char *str, int *supported)
{
        FILE *fd;
        char temp[1024];

        if (fname == NULL || str == NULL || supported == NULL)
                return PQOS_RETVAL_PARAM;

        fd = fopen(fname, "r");
        if (fd == NULL) {
                LOG_DEBUG("%s not found.\n", fname);
                *supported = 0;
                return PQOS_RETVAL_OK;
        }

        while (fgets(temp, sizeof(temp), fd) != NULL) {
                if (strstr(temp, str) != NULL) {
                        *supported = 1;
                        fclose(fd);
                        return PQOS_RETVAL_OK;
                }
        }

        fclose(fd);
        return PQOS_RETVAL_OK;
}

/**
 * @brief Check if event is supported by resctrl monitoring
 *
 * @param [in] event monitoring event type
 * @param [out] supported set to 1 if resctrl support is present
 *
 * @return Operation status
 * @retval PQOS_RETVAL_OK success
 */
static int
detect_resctrl_support(const enum pqos_mon_event event, int *supported) {
        char buf[128];
        struct stat st;
        const char *event_name = NULL;

        ASSERT(supported != NULL);

        *supported = 0;

        switch (event) {
        case PQOS_MON_EVENT_L3_OCCUP:
                event_name = "llc_occupancy";
                break;
        case PQOS_MON_EVENT_LMEM_BW:
                event_name = "mbm_total_bytes";
                break;
        case PQOS_MON_EVENT_TMEM_BW:
                event_name = "mbm_local_bytes";
                break;
        default:
                return PQOS_RETVAL_OK;
                break;
        }

        /** Check if resctrl is mounted */
        memset(buf, 0, sizeof(buf));
        if (snprintf(buf, sizeof(buf) - 1, "%s/info", RESCTRL_PATH) < 0)
                return PQOS_RETVAL_ERROR;

        /**
         * If resctrl is not mounted perform negative check for perf support
         */
        if (stat(buf, &st) != 0) {
                /* perf is not present, assume that resctrl is supported*/
                if (stat(PERF_MON_PATH, &st) != 0)
                        *supported = 1;
                return PQOS_RETVAL_OK;
        }

        /**
         * Resctrl is mounted check if L3 monitoring is supported in resctrl
         * info dir
         */
        memset(buf, 0, sizeof(buf));
        if (snprintf(buf, sizeof(buf) - 1, "%s/info/L3_MON/mon_features",
                RESCTRL_PATH) < 0)
                return PQOS_RETVAL_ERROR;

        if (stat(buf, &st) == 0)
                return detect_os_support(buf, event_name, supported);

        return PQOS_RETVAL_OK;
}

/**
 * @brief Get event name string to search in cpuinfo
 *
 * @param [in] event monitoring event type to look for
 *
 * @return cpuinfo flag representing monitoring event
 */
static const char *
get_os_event_name(int event)
{
        switch (event) {
        case PQOS_MON_EVENT_L3_OCCUP:
                return "cqm_occup_llc";
        case PQOS_MON_EVENT_LMEM_BW:
                return "cqm_mbm_local";
        case PQOS_MON_EVENT_TMEM_BW:
                return "cqm_mbm_total";
        default:
                return NULL;
        }
}

/**
 * @brief Get event disable flag to seartch in kernel cmdline
 *
 * @param [in] event monitoring event type to look for
 *
 * @return cmdline flag
 */
static const char *
get_os_event_disabled(int event)
{
        switch (event) {
        case PQOS_MON_EVENT_L3_OCCUP:
                return "!cmt";
        case PQOS_MON_EVENT_LMEM_BW:
                return "!mbmlocal";
        case PQOS_MON_EVENT_TMEM_BW:
                return "!mbmtotal";
        default:
                return NULL;
        }
}

/**
 * @brief Runs detection of OS monitoring events
 *
 * @param mon Monitoring capability structure
 *
 * @return Operation status
 * @retval PQOS_RETVAL_OK success
 */
static int
discover_os_monitoring(struct pqos_cap_mon *mon,
                       enum pqos_interface interface)
{
	int ret = PQOS_RETVAL_OK;
	unsigned i;
        enum pqos_os_mon lmem_support = PQOS_OS_MON_UNSUPPORTED;
	enum pqos_os_mon tmem_support = PQOS_OS_MON_UNSUPPORTED;
        int resctrl_support = 0;

        ASSERT(mon != NULL);

        for (i = 0; i < mon->num_events; i++) {
                struct pqos_monitor *event = &(mon->events[i]);
                const char *str = NULL;
                int os_support;
                int os_disabled = 0;
                int os_resctrl;

		/**
		 * Assume support of perf events
		 */
		if (event->type == PQOS_PERF_EVENT_LLC_MISS ||
		    event->type == PQOS_PERF_EVENT_IPC) {
			event->os_support = PQOS_OS_MON_PERF;
			continue;
		}

		str = get_os_event_name(event->type);
		if (str == NULL)
			continue;

                /**
                 * Check if event is supported by kernel
                 */
		ret = detect_os_support(PROC_CPUINFO, str, &os_support);
		if (ret != PQOS_RETVAL_OK) {
                        LOG_ERROR("Fatal error encountered in OS monitoring"
                                  " event detection!\n");
                        return ret;
                }

                /** Event not supported - continue*/
                if (!os_support) {
                        event->os_support = PQOS_OS_MON_UNSUPPORTED;
                        continue;
                }

                /**
                 * Check if feature is not disabled
                 */
                str = get_os_event_disabled(event->type);
                if (str != NULL) {
                        ret = detect_os_support("/proc/cmdline", str,
                                &os_disabled);
                        if (ret != PQOS_RETVAL_OK) {
                                LOG_ERROR("Fatal error encountered while "
                                          "checking for disabled events\n");
                                return ret;
                        }
                }

                /** Event disabled - continue */
                if (os_disabled) {
                        event->os_support = PQOS_OS_MON_UNSUPPORTED;
                        continue;
                }

                /**
                 * Check for resctrl support
                 */
                ret = detect_resctrl_support(event->type, &os_resctrl);
                if (ret != PQOS_RETVAL_OK) {
                        LOG_ERROR("Fatal error encountered while checking for "
                                  "resctrl monitoring support\n");
                        return ret;
                }

                if (os_resctrl) {
                        event->os_support = PQOS_OS_MON_RESCTRL;
                        resctrl_support = 1;
                } else
                        event->os_support = PQOS_OS_MON_PERF;

		if (event->type == PQOS_MON_EVENT_TMEM_BW)
			tmem_support = event->os_support;
		if (event->type == PQOS_MON_EVENT_LMEM_BW)
			lmem_support = event->os_support;
	}

	/**
	* RMEM is supported when both LMEM and TMEM are supported
	*/
	for (i = 0; i < mon->num_events; i++) {
		struct pqos_monitor *event = &(mon->events[i]);

		if (event->type == PQOS_MON_EVENT_RMEM_BW) {
			if (lmem_support == tmem_support)
				event->os_support = lmem_support;
			else
				event->os_support = PQOS_OS_MON_UNSUPPORTED;
			break;
		}
	}

        if (interface == PQOS_INTER_OS_RESCTRL_MON && !resctrl_support) {
                LOG_ERROR("Resctrl monitoring selected but not supported\n");
                return PQOS_RETVAL_INTER;
        }

	return ret;
}

/**
 * @brief Runs detection of OS monitoring and allocation capabilities
 *
 * @param p_cap place to store allocated capabilities structure
 * @param interface flag, stores the selected interface
 *
 * @return Operation status
 * @retval PQOS_RETVAL_OK success
 */
static int
discover_os_capabilities(struct pqos_cap *p_cap, enum pqos_interface interface)
{
        int ret = PQOS_RETVAL_OK;
        int res_flag = 0;
        unsigned i = 0;
        /**
         * This table is used to assist the discovery of OS capabilities
         */
        static struct os_caps {
                const char *fname;
                const char *str;
                const char *desc;
        } tab[PQOS_CAP_TYPE_NUMOF] = {
                { PROC_CPUINFO, "cqm", "CMT"},
                { PROC_CPUINFO, "cat_l3", "L3 CAT"},
                { PROC_CPUINFO, "cat_l2", "L2 CAT"},
                { PROC_CPUINFO, "mba", "MBA"},
        };

        /**
         * resctrl detection
         */
        ret = detect_os_support("/proc/filesystems", "resctrl", &res_flag);
        if (ret != PQOS_RETVAL_OK) {
                LOG_ERROR("Fatal error encountered in resctrl detection!\n");
                return ret;
        }
        LOG_INFO("%s\n", res_flag ?
                 "resctrl detected" :
                 "resctrl not detected. "
                 "Kernel version 4.10 or higher required");

        if ((interface == PQOS_INTER_OS ||
                interface == PQOS_INTER_OS_RESCTRL_MON) && res_flag == 0) {
                LOG_ERROR("OS interface selected but not supported\n");
                if (interface == PQOS_INTER_OS_RESCTRL_MON)
                        return PQOS_RETVAL_INTER;
                return PQOS_RETVAL_ERROR;
        }
        /**
         * Detect OS support for all HW capabilities
         */
        for (i = 0; i < p_cap->num_cap; i++) {
                struct pqos_capability *capability = &(p_cap->capabilities[i]);
                int type = capability->type;

                ASSERT(type < PQOS_CAP_TYPE_NUMOF);

                int *os_ptr = &(capability->os_support);

                ret = detect_os_support(tab[type].fname, tab[type].str, os_ptr);
                if (ret != PQOS_RETVAL_OK) {
                        LOG_ERROR("Fatal error encountered in"
                                  " OS detection!\n");
                        return ret;
                }

		/**
		 * Discover available monitoring events
		 */
		if (type == PQOS_CAP_TYPE_MON && res_flag) {
			ret = discover_os_monitoring(capability->u.mon,
                                                     interface);
			if (ret != PQOS_RETVAL_OK)
				return ret;
		}

                /**
                 * If resctrl is supported and L3 CAT is detected in HW then
                 * resctrl supports L3 CAT
                 */
                if (type == PQOS_CAP_TYPE_L3CA && *os_ptr == 0 && res_flag)
                        *os_ptr = 1;

		/**
		 * Discover L3 CDP support
		 */
		if (type == PQOS_CAP_TYPE_L3CA && *os_ptr) {
			ret = detect_os_support(PROC_CPUINFO, "cdp_l3",
			                        &capability->u.l3ca->os_cdp);
			if (ret != PQOS_RETVAL_OK) {
				LOG_ERROR("Fatal error encountered in L3 CDP "
				          "detection!\n");
				return ret;
			}
		}

                /**
                 * Discover L2 CDP support
                 */
                if (type == PQOS_CAP_TYPE_L2CA && *os_ptr) {
                        ret = detect_os_support(PROC_CPUINFO, "cdp_l2",
                                                &capability->u.l2ca->os_cdp);
                        if (ret != PQOS_RETVAL_OK) {
                                LOG_ERROR("Fatal error encountered in L2 CDP "
                                          "detection!\n");
                                return ret;
                        }
                }

                LOG_INFO("OS support for %s %s\n", tab[type].desc, *os_ptr ?
                         "detected" : "not detected");
        }
        /**
         * Check if resctrl is mounted
         */
        if (access(RESCTRL_PATH"/cpus", F_OK) != 0) {
                LOG_INFO("resctrl not mounted\n");
                return PQOS_RETVAL_RESOURCE;
        } else if (interface == PQOS_INTER_MSR)
                LOG_WARN("resctl filesystem mounted! Using MSR "
                         "interface may corrupt resctrl filesystem "
                         "and cause unexpected behaviour\n");

        return PQOS_RETVAL_OK;
}

/**
 * @brief Print information about capabilities that are not supported by the OS
 *
 * @param p_cap place to store allocated capabilities structure
 *
 * @return Operational status
 */
static int
log_hw_caps(struct pqos_cap *p_cap)
{
        unsigned i;

        /**
         * Log capabilities not supported by the OS
         */
        for (i = 0; i < p_cap->num_cap; i++)
                if (p_cap->capabilities[i].os_support == 0) {
                        if (p_cap->capabilities[i].type == PQOS_CAP_TYPE_MON)
                                LOG_INFO("Monitoring available in HW but not"
                                         " supported by OS. Disabling this"
                                         " capability.\n");
                        if (p_cap->capabilities[i].type == PQOS_CAP_TYPE_L3CA)
                                LOG_INFO("L3 CAT available in HW but not"
                                         " supported by OS. Disabling this"
                                         " capability.\n");
                        if (p_cap->capabilities[i].type == PQOS_CAP_TYPE_L2CA)
                                LOG_INFO("L2 CAT available in HW but not"
                                         " supported by OS. Disabling this"
                                         " capability.\n");
                        if (p_cap->capabilities[i].type == PQOS_CAP_TYPE_MBA)
                                LOG_INFO("MBA available in HW but not"
                                         " supported by OS. Disabling this"
                                         " capability.\n");
                }

        return PQOS_RETVAL_OK;
}
#endif /* __linux__ */

/*
 * =======================================
 * =======================================
 *
 * initialize and shutdown
 *
 * =======================================
 * =======================================
 */
int
pqos_init(const struct pqos_config *config)
{
        int ret = PQOS_RETVAL_OK;
        unsigned i = 0, max_core = 0;
        int cat_init = 0, mon_init = 0;
        char *environment = NULL;

        if (config == NULL)
                return PQOS_RETVAL_PARAM;

        environment = getenv("RDT_IFACE");
        if (environment != NULL) {
                if (strncasecmp(environment, "OS", 2) == 0) {
                        if (config->interface != PQOS_INTER_OS) {
                                fprintf(stderr, "Interface initialization "
                                        "error!\nYour system has been "
                                        "restricted to use the OS interface "
                                        "only!\n");
                                return PQOS_RETVAL_ERROR;
                        }
                } else if (strncasecmp(environment, "MSR", 3) == 0) {
                        if (config->interface != PQOS_INTER_MSR) {
                                fprintf(stderr, "Interface initialization "
                                        "error!\nYour system has been "
                                        "restricted to use the MSR interface "
                                        "only!\n");
                                return PQOS_RETVAL_ERROR;
                        }
                } else {
                        fprintf(stderr, "Interface initialization error!\n"
                                "Invalid interface enforcement selection.\n");
                        return PQOS_RETVAL_ERROR;
                }
        }

        if (_pqos_api_init() != 0) {
                fprintf(stderr, "API lock initialization error!\n");
                return PQOS_RETVAL_ERROR;
        }

        _pqos_api_lock();

        ret = _pqos_check_init(0);
        if (ret != PQOS_RETVAL_OK) {
                _pqos_api_unlock();
                return ret;
        }

        ret = log_init(config->fd_log,
		config->callback_log,
		config->context_log,
		config->verbose);
        if (ret != LOG_RETVAL_OK) {
                fprintf(stderr, "log_init() error\n");
                goto init_error;
        }

        /**
         * Topology not provided through config.
         * CPU discovery done through internal mechanism.
         */
        ret = cpuinfo_init(&m_cpu);
        if (ret != 0 || m_cpu == NULL) {
                LOG_ERROR("cpuinfo_init() error %d\n", ret);
                ret = PQOS_RETVAL_ERROR;
                goto log_init_error;
        }

        /**
         * Find max core id in the topology
         */
        for (i = 0; i < m_cpu->num_cores; i++)
                if (m_cpu->cores[i].lcore > max_core)
                        max_core = m_cpu->cores[i].lcore;

        ret = machine_init(max_core);
        if (ret != PQOS_RETVAL_OK) {
                LOG_ERROR("machine_init() error %d\n", ret);
                goto cpuinfo_init_error;
        }

        ret = discover_capabilities(&m_cap, m_cpu);
        if (ret != PQOS_RETVAL_OK) {
                LOG_ERROR("discover_capabilities() error %d\n", ret);
                goto machine_init_error;
        }
        ASSERT(m_cap != NULL);
#ifdef __linux__
        ret = discover_os_capabilities(m_cap, config->interface);
        if (ret == PQOS_RETVAL_ERROR || ret == PQOS_RETVAL_INTER) {
                LOG_ERROR("discover_os_capabilities() error %d\n", ret);
                goto machine_init_error;
        }
#endif

        if (config->interface == PQOS_INTER_OS ||
                config->interface == PQOS_INTER_OS_RESCTRL_MON) {
                const struct pqos_capability *l2_cap = NULL;

                ret = pqos_cap_get_type(m_cap, PQOS_CAP_TYPE_L2CA, &l2_cap);
                if (ret != PQOS_RETVAL_OK && ret != PQOS_RETVAL_RESOURCE)
                        goto machine_init_error;

                /* L2 CDP enabled but not supported by OS interface */
                if (l2_cap != NULL && l2_cap->u.l2ca->cdp &&
                        !l2_cap->u.l2ca->os_cdp) {
                        LOG_ERROR("Detected L2 CDP feature enabled but not "
                                  "supported by the current OS version!\n"
                                  "Please disable L2 CDP through the HW "
                                  "interface and perform a CAT reset.\n");
                        ret = PQOS_RETVAL_ERROR;
                        goto machine_init_error;
                }

#ifdef __linux__
                ret = log_hw_caps(m_cap);
#else
                LOG_ERROR("OS interface not supported!\n");
                ret = PQOS_RETVAL_RESOURCE;
                goto machine_init_error;
#endif
        }
        if (ret == PQOS_RETVAL_ERROR) {
                LOG_ERROR("log_hw_caps() error %d\n", ret);
                goto machine_init_error;
        }

        ret = _pqos_utils_init(config->interface);
        if (ret != PQOS_RETVAL_OK) {
                fprintf(stderr, "Utils initialization error!\n");
                goto machine_init_error;
        }

        ret = api_init(config->interface);
        if (ret != PQOS_RETVAL_OK) {
                LOG_ERROR("_pqos_api_init() error %d\n", ret);
                goto machine_init_error;
        }
#ifdef __linux__
        m_interface = config->interface;
#endif
        ret = pqos_alloc_init(m_cpu, m_cap, config);
        switch (ret) {
        case PQOS_RETVAL_BUSY:
                LOG_ERROR("OS allocation init error!\n");
                goto machine_init_error;
        case PQOS_RETVAL_OK:
                LOG_DEBUG("allocation init OK\n");
                cat_init = 1;
                break;
        default:
                LOG_ERROR("allocation init error %d\n", ret);
                break;
        }

        /**
         * If monitoring capability has been discovered
         * then get max RMID supported by a CPU socket
         * and allocate memory for RMID table
         */
        ret = pqos_mon_init(m_cpu, m_cap, config);
        switch (ret) {
        case PQOS_RETVAL_RESOURCE:
                LOG_DEBUG("monitoring init aborted: feature not present\n");
                ret = PQOS_RETVAL_OK;
                break;
        case PQOS_RETVAL_OK:
                LOG_DEBUG("monitoring init OK\n");
                mon_init = 1;
                break;
        case PQOS_RETVAL_ERROR:
        default:
                LOG_ERROR("monitoring init error %d\n", ret);
                break;
        }

        if (cat_init == 0 && mon_init == 0) {
                LOG_ERROR("None of detected capabilities could be "
                          "initialized!\n");
                ret = PQOS_RETVAL_ERROR;
        }

 machine_init_error:
        if (ret != PQOS_RETVAL_OK)
                (void) machine_fini();
 cpuinfo_init_error:
        if (ret != PQOS_RETVAL_OK)
                (void) cpuinfo_fini();
 log_init_error:
        if (ret != PQOS_RETVAL_OK)
                (void) log_fini();
 init_error:
        if (ret != PQOS_RETVAL_OK) {
                if (m_cap != NULL) {
                        for (i = 0; i < m_cap->num_cap; i++)
                                free(m_cap->capabilities[i].u.generic_ptr);
                        free(m_cap);
                }
                m_cpu = NULL;
                m_cap = NULL;
        }

        if (ret == PQOS_RETVAL_OK)
                m_init_done = 1;

        _pqos_api_unlock();

        if (ret != PQOS_RETVAL_OK)
                _pqos_api_exit();

        return ret;
}

int
pqos_fini(void)
{
        int ret = PQOS_RETVAL_OK;
        int retval = PQOS_RETVAL_OK;
        unsigned i = 0;

        _pqos_api_lock();

        ret = _pqos_check_init(1);
        if (ret != PQOS_RETVAL_OK) {
                _pqos_api_unlock();
                _pqos_api_exit();
                return ret;
        }

        pqos_mon_fini();
        pqos_alloc_fini();

        ret = cpuinfo_fini();
        if (ret != 0) {
                retval = PQOS_RETVAL_ERROR;
                LOG_ERROR("cpuinfo_fini() error %d\n", ret);
        }

        ret = machine_fini();
        if (ret != PQOS_RETVAL_OK) {
                retval = ret;
                LOG_ERROR("machine_fini() error %d\n", ret);
        }

        ret = log_fini();
        if (ret != PQOS_RETVAL_OK)
                retval = ret;

        m_cpu = NULL;

        for (i = 0; i < m_cap->num_cap; i++)
                free(m_cap->capabilities[i].u.generic_ptr);
        free((void *)m_cap);
        m_cap = NULL;

        m_init_done = 0;

        _pqos_api_unlock();

        if (_pqos_api_exit() != 0)
                retval = PQOS_RETVAL_ERROR;

        return retval;
}

/**
 * =======================================
 * =======================================
 *
 * capabilities
 *
 * =======================================
 * =======================================
 */

int
pqos_cap_get(const struct pqos_cap **cap,
             const struct pqos_cpuinfo **cpu)
{
        int ret = PQOS_RETVAL_OK;

        if (cap == NULL && cpu == NULL)
                return PQOS_RETVAL_PARAM;

        _pqos_api_lock();

        ret = _pqos_check_init(1);
        if (ret != PQOS_RETVAL_OK) {
                _pqos_api_unlock();
                return ret;
        }

        if (cap != NULL) {
                ASSERT(m_cap != NULL);
                *cap = m_cap;
        }

        if (cpu != NULL) {
                ASSERT(m_cpu != NULL);
                *cpu = m_cpu;
        }

        _pqos_api_unlock();
        return PQOS_RETVAL_OK;
}

void
_pqos_cap_l3cdp_change(const enum pqos_cdp_config cdp)
{
        struct pqos_cap_l3ca *l3_cap = NULL;
        unsigned i;

        ASSERT(cdp == PQOS_REQUIRE_CDP_ON || cdp == PQOS_REQUIRE_CDP_OFF);
        ASSERT(m_cap != NULL);

        if (m_cap == NULL)
                return;

        for (i = 0; i < m_cap->num_cap && l3_cap == NULL; i++)
                if (m_cap->capabilities[i].type == PQOS_CAP_TYPE_L3CA)
                        l3_cap = m_cap->capabilities[i].u.l3ca;

        if (l3_cap == NULL)
                return;

        if (cdp == PQOS_REQUIRE_CDP_ON && !l3_cap->cdp_on) {
                /* turn on */
                l3_cap->cdp_on = 1;
                l3_cap->num_classes = l3_cap->num_classes / 2;
        }

        if (cdp == PQOS_REQUIRE_CDP_OFF && l3_cap->cdp_on) {
                /* turn off */
                l3_cap->cdp_on = 0;
                l3_cap->num_classes = l3_cap->num_classes * 2;
        }
}

void
_pqos_cap_l2cdp_change(const enum pqos_cdp_config cdp)
{
        struct pqos_cap_l2ca *l2_cap = NULL;
        unsigned i;

        ASSERT(cdp == PQOS_REQUIRE_CDP_ON || cdp == PQOS_REQUIRE_CDP_OFF);
        ASSERT(m_cap != NULL);

        if (m_cap == NULL)
                return;

        for (i = 0; i < m_cap->num_cap && l2_cap == NULL; i++)
                if (m_cap->capabilities[i].type == PQOS_CAP_TYPE_L2CA)
                        l2_cap = m_cap->capabilities[i].u.l2ca;

        if (l2_cap == NULL)
                return;

        if (cdp == PQOS_REQUIRE_CDP_ON && !l2_cap->cdp_on) {
                /* turn on */
                l2_cap->cdp_on = 1;
                l2_cap->num_classes = l2_cap->num_classes / 2;
        }

        if (cdp == PQOS_REQUIRE_CDP_OFF && l2_cap->cdp_on) {
                /* turn off */
                l2_cap->cdp_on = 0;
                l2_cap->num_classes = l2_cap->num_classes * 2;
        }
}