On Tue, Dec 7, 2010 at 4:00 AM, Len Brown <lenb@xxxxxxxxxx> wrote: > From: Len Brown <len.brown@xxxxxxxxx> > > turbostat is a tool to help verify proper operation > of processor power management states on modern x86. > > Signed-off-by: Len Brown <len.brown@xxxxxxxxx> > --- > v4: fix open file descriptor leak, reported by David C. Wong > > Âtools/power/x86/turbostat/Makefile  Â|  Â8 + > Âtools/power/x86/turbostat/turbostat.8 | Â172 ++++++ > Âtools/power/x86/turbostat/turbostat.c | 1048 +++++++++++++++++++++++++++++++++ > Â3 files changed, 1228 insertions(+), 0 deletions(-) > > diff --git a/tools/power/x86/turbostat/turbostat.c b/tools/power/x86/turbostat/turbostat.c > new file mode 100644 > index 0000000..4c6983d > --- /dev/null > +++ b/tools/power/x86/turbostat/turbostat.c > @@ -0,0 +1,1048 @@ > +/* > + * turbostat -- show CPU frequency and C-state residency > + * on modern Intel turbo-capable processors. > + * > + * Copyright (c) 2010, Intel Corporation. > + * Len Brown <len.brown@xxxxxxxxx> > + * > + * This program is free software; you can redistribute it and/or modify it > + * under the terms and conditions of the GNU General Public License, > + * version 2, as published by the Free Software Foundation. > + * > + * This program is distributed in the hope it will be useful, but WITHOUT > + * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or > + * FITNESS FOR A PARTICULAR PURPOSE. ÂSee the GNU General Public License for > + * more details. > + * > + * You should have received a copy of the GNU General Public License along with > + * this program; if not, write to the Free Software Foundation, Inc., > + * 51 Franklin St - Fifth Floor, Boston, MA 02110-1301 USA. > + */ > + > +#include <stdio.h> > +#include <unistd.h> > +#include <sys/types.h> > +#include <sys/wait.h> > +#include <sys/stat.h> > +#include <sys/resource.h> > +#include <fcntl.h> > +#include <signal.h> > +#include <sys/time.h> > +#include <stdlib.h> > +#include <dirent.h> > +#include <string.h> > +#include <ctype.h> > + > +#define MSR_TSC    Â0x10 > +#define MSR_NEHALEM_PLATFORM_INFO   Â0xCE > +#define MSR_NEHALEM_TURBO_RATIO_LIMIT Â0x1AD > +#define MSR_APERF   Â0xE8 > +#define MSR_MPERF   Â0xE7 > +#define MSR_PKG_C2_RESIDENCY  0x60D  /* SNB only */ > +#define MSR_PKG_C3_RESIDENCY  0x3F8 > +#define MSR_PKG_C6_RESIDENCY  0x3F9 > +#define MSR_PKG_C7_RESIDENCY  0x3FA  /* SNB only */ > +#define MSR_CORE_C3_RESIDENCY Â0x3FC > +#define MSR_CORE_C6_RESIDENCY Â0x3FD > +#define MSR_CORE_C7_RESIDENCY Â0x3FE  /* SNB only */ > + > +char *proc_stat = "/proc/stat"; > +unsigned int interval_sec = 5; /* set with -i interval_sec */ > +unsigned int verbose;     Â/* set with -v */ > +unsigned int skip_c0; > +unsigned int skip_c1; > +unsigned int do_nhm_cstates; > +unsigned int do_snb_cstates; > +unsigned int has_aperf; > +unsigned int units = 1000000000;    /* Ghz etc */ > +unsigned int genuine_intel; > +unsigned int has_invariant_tsc; > +unsigned int do_nehalem_platform_info; > +unsigned int do_nehalem_turbo_ratio_limit; > +unsigned int extra_msr_offset; > +double bclk; > +unsigned int show_pkg; > +unsigned int show_core; > +unsigned int show_cpu; > + > +int aperf_mperf_unstable; > +int backwards_count; > +char *progname; > +int need_reinitialize; > + > +int num_cpus; > + > +typedef struct per_cpu_counters { > +    unsigned long long tsc;     /* per thread */ > +    unsigned long long aperf;    /* per thread */ > +    unsigned long long mperf;    /* per thread */ > +    unsigned long long c1; Â/* per thread (calculated) */ > +    unsigned long long c3; Â/* per core */ > +    unsigned long long c6; Â/* per core */ > +    unsigned long long c7; Â/* per core */ > +    unsigned long long pc2; /* per package */ > +    unsigned long long pc3; /* per package */ > +    unsigned long long pc6; /* per package */ > +    unsigned long long pc7; /* per package */ > +    unsigned long long extra_msr;  /* per thread */ > +    int pkg; > +    int core; > +    int cpu; > +    struct per_cpu_counters *next; > +} PCC; > + Why the typedef? Isn't preferable to use struct per_cpu_counters all around (yeah I know it requires more typing, but is less obfuscated). > +PCC *pcc_even; > +PCC *pcc_odd; > +PCC *pcc_delta; > +PCC *pcc_average; > +struct timeval tv_even; > +struct timeval tv_odd; > +struct timeval tv_delta; > + > +unsigned long long get_msr(int cpu, off_t offset) > +{ > +    ssize_t retval; > +    unsigned long long msr; > +    char pathname[32]; > +    int fd; > + > +    sprintf(pathname, "/dev/cpu/%d/msr", cpu); > +    fd = open(pathname, O_RDONLY); > +    if (fd < 0) { > +        perror(pathname); > +        need_reinitialize = 1; > +        return 0; > +    } > + > +    retval = pread(fd, &msr, sizeof msr, offset); > +    if (retval != sizeof msr) { > +        fprintf(stderr, "cpu%d pread(..., 0x%zx) = %jd\n", > +            cpu, offset, retval); > +        exit(-2); > +    } > + > +    close(fd); > +    return msr; > +} > + > +void print_header() use void in the paramenter? Also make this static? > + > +#define SUBTRACT_COUNTER(after, before, delta) (delta = (after - before), (before > after)) > + > + Remove this extra blank line? > + > +void get_counters(PCC *pcc) > +{ > +    for ( ; pcc; pcc = pcc->next) { > +        pcc->tsc = get_msr(pcc->cpu, MSR_TSC); > +        if (do_nhm_cstates) > +            pcc->c3 = get_msr(pcc->cpu, MSR_CORE_C3_RESIDENCY); > +        if (do_nhm_cstates) > +            pcc->c6 = get_msr(pcc->cpu, MSR_CORE_C6_RESIDENCY); > +        if (do_snb_cstates) > +            pcc->c7 = get_msr(pcc->cpu, MSR_CORE_C7_RESIDENCY); > +        if (has_aperf) > +            pcc->aperf = get_msr(pcc->cpu, MSR_APERF); > +        if (has_aperf) > +            pcc->mperf = get_msr(pcc->cpu, MSR_MPERF); > +        if (do_snb_cstates) > +            pcc->pc2 = get_msr(pcc->cpu, MSR_PKG_C2_RESIDENCY); > +        if (do_nhm_cstates) > +            pcc->pc3 = get_msr(pcc->cpu, MSR_PKG_C3_RESIDENCY); > +        if (do_nhm_cstates) > +            pcc->pc6 = get_msr(pcc->cpu, MSR_PKG_C6_RESIDENCY); > +        if (do_snb_cstates) > +            pcc->pc7 = get_msr(pcc->cpu, MSR_PKG_C7_RESIDENCY); > +        if (extra_msr_offset) > +            pcc->extra_msr = get_msr(pcc->cpu, extra_msr_offset); > +    } > +} > + > + Remove this blank line too? > +void print_nehalem_info() > +{ Add void into the parameter? > +    unsigned long long msr; > +    unsigned int ratio; > + > +    if (!do_nehalem_platform_info) > +        return; > + > +    msr = get_msr(0, MSR_NEHALEM_PLATFORM_INFO); > + > +    ratio = (msr >> 40) & 0xFF; > +    fprintf(stderr, "%d * %.0f = %.0f MHz max efficiency\n", > +        ratio, bclk, ratio * bclk); > + > +    ratio = (msr >> 8) & 0xFF; > +    fprintf(stderr, "%d * %.0f = %.0f MHz TSC frequency\n", > +        ratio, bclk, ratio * bclk); > + > +    if (verbose > 1) > +        fprintf(stderr, "MSR_NEHALEM_PLATFORM_INFO: 0x%llx\n", msr); > + > +    if (!do_nehalem_turbo_ratio_limit) > +        return; > + > +    msr = get_msr(0, MSR_NEHALEM_TURBO_RATIO_LIMIT); > + > +    ratio = (msr >> 24) & 0xFF; > +    if (ratio) > +        fprintf(stderr, "%d * %.0f = %.0f MHz max turbo 4 active cores\n", > +            ratio, bclk, ratio * bclk); > + > +    ratio = (msr >> 16) & 0xFF; > +    if (ratio) > +        fprintf(stderr, "%d * %.0f = %.0f MHz max turbo 3 active cores\n", > +            ratio, bclk, ratio * bclk); > + > +    ratio = (msr >> 8) & 0xFF; > +    if (ratio) > +        fprintf(stderr, "%d * %.0f = %.0f MHz max turbo 2 active cores\n", > +            ratio, bclk, ratio * bclk); > + > +    ratio = (msr >> 0) & 0xFF; > +    if (ratio) > +        fprintf(stderr, "%d * %.0f = %.0f MHz max turbo 1 active cores\n", > +            ratio, bclk, ratio * bclk); > + > +} > + > +void free_counter_list(PCC *list) > +{ > +    PCC *p; > + > +    for (p = list; p; ) { > +        PCC *free_me; > + > +        free_me = p; > +        p = p->next; > +        free(free_me); > +    } > +    return; Why the return here? Isn't redundant? > + > +void insert_cpu_counters(PCC **list, PCC *new) > +{ > +    PCC *prev; > + > +    /* > +    Â* list was empty > +    Â*/ > +    if (*list == NULL) { > +        new->next = *list; > +        *list = new; > +        return; > +    } > + > +    show_cpu = 1;  /* there is more than one CPU */ > + > +    /* > +    Â* insert on front of list. > +    Â* It is sorted by ascending package#, core#, cpu# > +    Â*/ > +    if (((*list)->pkg > new->pkg) || > +      (((*list)->pkg == new->pkg) && ((*list)->core > new->core)) || > +      (((*list)->pkg == new->pkg) && ((*list)->core == new->core) && ((*list)->cpu > new->cpu))) { > +        new->next = *list; > +        *list = new; > +        return; > +    } > + > +    prev = *list; > + > +    while (prev->next && (prev->next->pkg < new->pkg)) { > +        prev = prev->next; > +        show_pkg = 1;  /* there is more than 1 package */ > +    } > + > +    while (prev->next && (prev->next->pkg == new->pkg) > +        && (prev->next->core < new->core)) { > +        prev = prev->next; > +        show_core = 1; Â/* there is more than 1 core */ > +    } > + > +    while (prev->next && (prev->next->pkg == new->pkg) > +        && (prev->next->core == new->core) > +        && (prev->next->cpu < new->cpu)) { > +        prev = prev->next; > +    } > + > +    /* > +    Â* insert after "prev" > +    Â*/ > +    new->next = prev->next; > +    prev->next = new; > + > +    return; Same thing here about the return. > + > +int get_physical_package_id(int cpu) > +{ > +    char path[64]; > +    FILE *filep; > +    int pkg; > + > +    sprintf(path, "/sys/devices/system/cpu/cpu%d/topology/physical_package_id", cpu); > +    filep = fopen(path, "r"); > +    if (filep == NULL) { > +        perror(path); > +        exit(1); > +    } > +    fscanf(filep, "%d", &pkg); > +    fclose(filep); > +    return pkg; > +} > + > +int get_core_id(int cpu) > +{ > +    char path[64]; > +    FILE *filep; > +    int core; > + > +    sprintf(path, "/sys/devices/system/cpu/cpu%d/topology/core_id", cpu); > +    filep = fopen(path, "r"); > +    if (filep == NULL) { > +        perror(path); > +        exit(1); > +    } > +    fscanf(filep, "%d", &core); > +    fclose(filep); > +    return core; > +} > + > +/* > + * run func(index, cpu) on every cpu in /proc/stat > + */ > + > +int for_all_cpus(void (func)(int, int, int)) > +{ > +    FILE *fp; > +    int cpu_count; > +    int retval; > + > +    fp = fopen(proc_stat, "r"); > +    if (fp == NULL) { > +        perror(proc_stat); > +        exit(1); > +    } > + > +    retval = fscanf(fp, "cpu %*d %*d %*d %*d %*d %*d %*d %*d %*d %*d\n"); > +    if (retval != 0) { > +        perror("/proc/stat format"); > +        exit(1); > +    } > + > +    for (cpu_count = 0; ; cpu_count++) { > +        int cpu; > + > +        retval = fscanf(fp, "cpu%u %*d %*d %*d %*d %*d %*d %*d %*d %*d %*d\n", &cpu); > +        if (retval != 1) > +            break; > + > +        func(get_physical_package_id(cpu), get_core_id(cpu), cpu); > +    } > +    fclose(fp); > +    return cpu_count; > +} > + > +void re_initialize(void) > +{ > +    printf("turbostat: topology changed, re-initializing.\n"); > +    free_all_counters(); > +    num_cpus = for_all_cpus(alloc_new_cpu_counters); > +    need_reinitialize = 0; > +    printf("num_cpus is now %d\n", num_cpus); > +} > + > +void dummy(int pkg, int core, int cpu) { return; } > +/* > + * check to see if a cpu came on-line > + */ > +void verify_num_cpus() Use void on parameter? > +{ > +    int new_num_cpus; > + > +    new_num_cpus = for_all_cpus(dummy); > + > +    if (new_num_cpus != num_cpus) { > +        if (verbose) > +            printf("num_cpus was %d, is now Â%d\n", > +                num_cpus, new_num_cpus); > +        need_reinitialize = 1; > +    } > + > +    return; > +} No need of this return, right? _______________________________________________ linux-pm mailing list linux-pm@xxxxxxxxxxxxxxxxxxxxxxxxxx https://lists.linux-foundation.org/mailman/listinfo/linux-pm