qemu/qemu-io-cmds.c

/*
 * Command line utility to exercise the QEMU I/O path.
 *
 * Copyright (C) 2009-2016 Red Hat, Inc.
 * Copyright (c) 2003-2005 Silicon Graphics, Inc.
 *
 * This work is licensed under the terms of the GNU GPL, version 2 or later.
 * See the COPYING file in the top-level directory.
 */

#include "qemu/osdep.h"
#include "qapi/error.h"
#include "qemu-io.h"
#include "sysemu/block-backend.h"
#include "block/block.h"
#include "block/block_int.h" /* for info_f() */
#include "block/qapi.h"
#include "qemu/error-report.h"
#include "qemu/main-loop.h"
#include "qemu/timer.h"
#include "sysemu/block-backend.h"
#include "qemu/cutils.h"

#define CMD_NOFILE_OK   0x01

bool qemuio_misalign;

static cmdinfo_t *cmdtab;
static int ncmds;

static int compare_cmdname(const void *a, const void *b)
{
    return strcmp(((const cmdinfo_t *)a)->name,
                  ((const cmdinfo_t *)b)->name);
}

void qemuio_add_command(const cmdinfo_t *ci)
{
    cmdtab = g_renew(cmdinfo_t, cmdtab, ++ncmds);
    cmdtab[ncmds - 1] = *ci;
    qsort(cmdtab, ncmds, sizeof(*cmdtab), compare_cmdname);
}

int qemuio_command_usage(const cmdinfo_t *ci)
{
    printf("%s %s -- %s\n", ci->name, ci->args, ci->oneline);
    return 0;
}

static int init_check_command(BlockBackend *blk, const cmdinfo_t *ct)
{
    if (ct->flags & CMD_FLAG_GLOBAL) {
        return 1;
    }
    if (!(ct->flags & CMD_NOFILE_OK) && !blk) {
        fprintf(stderr, "no file open, try 'help open'\n");
        return 0;
    }
    return 1;
}

static int command(BlockBackend *blk, const cmdinfo_t *ct, int argc,
                   char **argv)
{
    char *cmd = argv[0];

    if (!init_check_command(blk, ct)) {
        return 0;
    }

    if (argc - 1 < ct->argmin || (ct->argmax != -1 && argc - 1 > ct->argmax)) {
        if (ct->argmax == -1) {
            fprintf(stderr,
                    "bad argument count %d to %s, expected at least %d arguments\n",
                    argc-1, cmd, ct->argmin);
        } else if (ct->argmin == ct->argmax) {
            fprintf(stderr,
                    "bad argument count %d to %s, expected %d arguments\n",
                    argc-1, cmd, ct->argmin);
        } else {
            fprintf(stderr,
                    "bad argument count %d to %s, expected between %d and %d arguments\n",
                    argc-1, cmd, ct->argmin, ct->argmax);
        }
        return 0;
    }
    optind = 0;
    return ct->cfunc(blk, argc, argv);
}

static const cmdinfo_t *find_command(const char *cmd)
{
    cmdinfo_t *ct;

    for (ct = cmdtab; ct < &cmdtab[ncmds]; ct++) {
        if (strcmp(ct->name, cmd) == 0 ||
            (ct->altname && strcmp(ct->altname, cmd) == 0))
        {
            return (const cmdinfo_t *)ct;
        }
    }
    return NULL;
}

/* Invoke fn() for commands with a matching prefix */
void qemuio_complete_command(const char *input,
                             void (*fn)(const char *cmd, void *opaque),
                             void *opaque)
{
    cmdinfo_t *ct;
    size_t input_len = strlen(input);

    for (ct = cmdtab; ct < &cmdtab[ncmds]; ct++) {
        if (strncmp(input, ct->name, input_len) == 0) {
            fn(ct->name, opaque);
        }
    }
}

static char **breakline(char *input, int *count)
{
    int c = 0;
    char *p;
    char **rval = g_new0(char *, 1);

    while (rval && (p = qemu_strsep(&input, " ")) != NULL) {
        if (!*p) {
            continue;
        }
        c++;
        rval = g_renew(char *, rval, (c + 1));
        rval[c - 1] = p;
        rval[c] = NULL;
    }
    *count = c;
    return rval;
}

static int64_t cvtnum(const char *s)
{
    char *end;
    int64_t ret;

    ret = qemu_strtosz_suffix(s, &end, QEMU_STRTOSZ_DEFSUFFIX_B);
    if (*end != '\0') {
        /* Detritus at the end of the string */
        return -EINVAL;
    }
    return ret;
}

static void print_cvtnum_err(int64_t rc, const char *arg)
{
    switch (rc) {
    case -EINVAL:
        printf("Parsing error: non-numeric argument,"
               " or extraneous/unrecognized suffix -- %s\n", arg);
        break;
    case -ERANGE:
        printf("Parsing error: argument too large -- %s\n", arg);
        break;
    default:
        printf("Parsing error: %s\n", arg);
    }
}

#define EXABYTES(x)     ((long long)(x) << 60)
#define PETABYTES(x)    ((long long)(x) << 50)
#define TERABYTES(x)    ((long long)(x) << 40)
#define GIGABYTES(x)    ((long long)(x) << 30)
#define MEGABYTES(x)    ((long long)(x) << 20)
#define KILOBYTES(x)    ((long long)(x) << 10)

#define TO_EXABYTES(x)  ((x) / EXABYTES(1))
#define TO_PETABYTES(x) ((x) / PETABYTES(1))
#define TO_TERABYTES(x) ((x) / TERABYTES(1))
#define TO_GIGABYTES(x) ((x) / GIGABYTES(1))
#define TO_MEGABYTES(x) ((x) / MEGABYTES(1))
#define TO_KILOBYTES(x) ((x) / KILOBYTES(1))

static void cvtstr(double value, char *str, size_t size)
{
    char *trim;
    const char *suffix;

    if (value >= EXABYTES(1)) {
        suffix = " EiB";
        snprintf(str, size - 4, "%.3f", TO_EXABYTES(value));
    } else if (value >= PETABYTES(1)) {
        suffix = " PiB";
        snprintf(str, size - 4, "%.3f", TO_PETABYTES(value));
    } else if (value >= TERABYTES(1)) {
        suffix = " TiB";
        snprintf(str, size - 4, "%.3f", TO_TERABYTES(value));
    } else if (value >= GIGABYTES(1)) {
        suffix = " GiB";
        snprintf(str, size - 4, "%.3f", TO_GIGABYTES(value));
    } else if (value >= MEGABYTES(1)) {
        suffix = " MiB";
        snprintf(str, size - 4, "%.3f", TO_MEGABYTES(value));
    } else if (value >= KILOBYTES(1)) {
        suffix = " KiB";
        snprintf(str, size - 4, "%.3f", TO_KILOBYTES(value));
    } else {
        suffix = " bytes";
        snprintf(str, size - 6, "%f", value);
    }

    trim = strstr(str, ".000");
    if (trim) {
        strcpy(trim, suffix);
    } else {
        strcat(str, suffix);
    }
}



static struct timeval tsub(struct timeval t1, struct timeval t2)
{
    t1.tv_usec -= t2.tv_usec;
    if (t1.tv_usec < 0) {
        t1.tv_usec += 1000000;
        t1.tv_sec--;
    }
    t1.tv_sec -= t2.tv_sec;
    return t1;
}

static double tdiv(double value, struct timeval tv)
{
    return value / ((double)tv.tv_sec + ((double)tv.tv_usec / 1000000.0));
}

#define HOURS(sec)      ((sec) / (60 * 60))
#define MINUTES(sec)    (((sec) % (60 * 60)) / 60)
#define SECONDS(sec)    ((sec) % 60)

enum {
    DEFAULT_TIME        = 0x0,
    TERSE_FIXED_TIME    = 0x1,
    VERBOSE_FIXED_TIME  = 0x2,
};

static void timestr(struct timeval *tv, char *ts, size_t size, int format)
{
    double usec = (double)tv->tv_usec / 1000000.0;

    if (format & TERSE_FIXED_TIME) {
        if (!HOURS(tv->tv_sec)) {
            snprintf(ts, size, "%u:%02u.%02u",
                    (unsigned int) MINUTES(tv->tv_sec),
                    (unsigned int) SECONDS(tv->tv_sec),
                    (unsigned int) (usec * 100));
            return;
        }
        format |= VERBOSE_FIXED_TIME; /* fallback if hours needed */
    }

    if ((format & VERBOSE_FIXED_TIME) || tv->tv_sec) {
        snprintf(ts, size, "%u:%02u:%02u.%02u",
                (unsigned int) HOURS(tv->tv_sec),
                (unsigned int) MINUTES(tv->tv_sec),
                (unsigned int) SECONDS(tv->tv_sec),
                (unsigned int) (usec * 100));
    } else {
        snprintf(ts, size, "0.%04u sec", (unsigned int) (usec * 10000));
    }
}

/*
 * Parse the pattern argument to various sub-commands.
 *
 * Because the pattern is used as an argument to memset it must evaluate
 * to an unsigned integer that fits into a single byte.
 */
static int parse_pattern(const char *arg)
{
    char *endptr = NULL;
    long pattern;

    pattern = strtol(arg, &endptr, 0);
    if (pattern < 0 || pattern > UCHAR_MAX || *endptr != '\0') {
        printf("%s is not a valid pattern byte\n", arg);
        return -1;
    }

    return pattern;
}

/*
 * Memory allocation helpers.
 *
 * Make sure memory is aligned by default, or purposefully misaligned if
 * that is specified on the command line.
 */

#define MISALIGN_OFFSET     16
static void *qemu_io_alloc(BlockBackend *blk, size_t len, int pattern)
{
    void *buf;

    if (qemuio_misalign) {
        len += MISALIGN_OFFSET;
    }
    buf = blk_blockalign(blk, len);
    memset(buf, pattern, len);
    if (qemuio_misalign) {
        buf += MISALIGN_OFFSET;
    }
    return buf;
}

static void qemu_io_free(void *p)
{
    if (qemuio_misalign) {
        p -= MISALIGN_OFFSET;
    }
    qemu_vfree(p);
}

static void dump_buffer(const void *buffer, int64_t offset, int64_t len)
{
    uint64_t i;
    int j;
    const uint8_t *p;

    for (i = 0, p = buffer; i < len; i += 16) {
        const uint8_t *s = p;

        printf("%08" PRIx64 ":  ", offset + i);
        for (j = 0; j < 16 && i + j < len; j++, p++) {
            printf("%02x ", *p);
        }
        printf(" ");
        for (j = 0; j < 16 && i + j < len; j++, s++) {
            if (isalnum(*s)) {
                printf("%c", *s);
            } else {
                printf(".");
            }
        }
        printf("\n");
    }
}

static void print_report(const char *op, struct timeval *t, int64_t offset,
                         int64_t count, int64_t total, int cnt, bool Cflag)
{
    char s1[64], s2[64], ts[64];

    timestr(t, ts, sizeof(ts), Cflag ? VERBOSE_FIXED_TIME : 0);
    if (!Cflag) {
        cvtstr((double)total, s1, sizeof(s1));
        cvtstr(tdiv((double)total, *t), s2, sizeof(s2));
        printf("%s %"PRId64"/%"PRId64" bytes at offset %" PRId64 "\n",
               op, total, count, offset);
        printf("%s, %d ops; %s (%s/sec and %.4f ops/sec)\n",
               s1, cnt, ts, s2, tdiv((double)cnt, *t));
    } else {/* bytes,ops,time,bytes/sec,ops/sec */
        printf("%"PRId64",%d,%s,%.3f,%.3f\n",
            total, cnt, ts,
            tdiv((double)total, *t),
            tdiv((double)cnt, *t));
    }
}

/*
 * Parse multiple length statements for vectored I/O, and construct an I/O
 * vector matching it.
 */
static void *
create_iovec(BlockBackend *blk, QEMUIOVector *qiov, char **argv, int nr_iov,
             int pattern)
{
    size_t *sizes = g_new0(size_t, nr_iov);
    size_t count = 0;
    void *buf = NULL;
    void *p;
    int i;

    for (i = 0; i < nr_iov; i++) {
        char *arg = argv[i];
        int64_t len;

        len = cvtnum(arg);
        if (len < 0) {
            print_cvtnum_err(len, arg);
            goto fail;
        }

        /* should be SIZE_T_MAX, but that doesn't exist */
        if (len > INT_MAX) {
            printf("Argument '%s' exceeds maximum size %d\n", arg, INT_MAX);
            goto fail;
        }

        sizes[i] = len;
        count += len;
    }

    qemu_iovec_init(qiov, nr_iov);

    buf = p = qemu_io_alloc(blk, count, pattern);

    for (i = 0; i < nr_iov; i++) {
        qemu_iovec_add(qiov, p, sizes[i]);
        p += sizes[i];
    }

fail:
    g_free(sizes);
    return buf;
}

static int do_pread(BlockBackend *blk, char *buf, int64_t offset,
                    int64_t count, int64_t *total)
{
    if (count > INT_MAX) {
        return -ERANGE;
    }

    *total = blk_pread(blk, offset, (uint8_t *)buf, count);
    if (*total < 0) {
        return *total;
    }
    return 1;
}

static int do_pwrite(BlockBackend *blk, char *buf, int64_t offset,
                     int64_t count, int flags, int64_t *total)
{
    if (count > INT_MAX) {
        return -ERANGE;
    }

    *total = blk_pwrite(blk, offset, (uint8_t *)buf, count, flags);
    if (*total < 0) {
        return *total;
    }
    return 1;
}

typedef struct {
    BlockBackend *blk;
    int64_t offset;
    int64_t count;
    int64_t *total;
    int flags;
    int ret;
    bool done;
} CoWriteZeroes;

static void coroutine_fn co_write_zeroes_entry(void *opaque)
{
    CoWriteZeroes *data = opaque;

    data->ret = blk_co_write_zeroes(data->blk, data->offset, data->count,
                                    data->flags);
    data->done = true;
    if (data->ret < 0) {
        *data->total = data->ret;
        return;
    }

    *data->total = data->count;
}

static int do_co_write_zeroes(BlockBackend *blk, int64_t offset, int64_t count,
                              int flags, int64_t *total)
{
    Coroutine *co;
    CoWriteZeroes data = {
        .blk    = blk,
        .offset = offset,
        .count  = count,
        .total  = total,
        .flags  = flags,
        .done   = false,
    };

    if (count >> BDRV_SECTOR_BITS > INT_MAX) {
        return -ERANGE;
    }

    co = qemu_coroutine_create(co_write_zeroes_entry);
    qemu_coroutine_enter(co, &data);
    while (!data.done) {
        aio_poll(blk_get_aio_context(blk), true);
    }
    if (data.ret < 0) {
        return data.ret;
    } else {
        return 1;
    }
}

static int do_write_compressed(BlockBackend *blk, char *buf, int64_t offset,
                               int64_t count, int64_t *total)
{
    int ret;

    if (count >> 9 > INT_MAX) {
        return -ERANGE;
    }

    ret = blk_write_compressed(blk, offset >> 9, (uint8_t *)buf, count >> 9);
    if (ret < 0) {
        return ret;
    }
    *total = count;
    return 1;
}

static int do_load_vmstate(BlockBackend *blk, char *buf, int64_t offset,
                           int64_t count, int64_t *total)
{
    if (count > INT_MAX) {
        return -ERANGE;
    }

    *total = blk_load_vmstate(blk, (uint8_t *)buf, offset, count);
    if (*total < 0) {
        return *total;
    }
    return 1;
}

static int do_save_vmstate(BlockBackend *blk, char *buf, int64_t offset,
                           int64_t count, int64_t *total)
{
    if (count > INT_MAX) {
        return -ERANGE;
    }

    *total = blk_save_vmstate(blk, (uint8_t *)buf, offset, count);
    if (*total < 0) {
        return *total;
    }
    return 1;
}

#define NOT_DONE 0x7fffffff
static void aio_rw_done(void *opaque, int ret)
{
    *(int *)opaque = ret;
}

static int do_aio_readv(BlockBackend *blk, QEMUIOVector *qiov,
                        int64_t offset, int *total)
{
    int async_ret = NOT_DONE;

    blk_aio_preadv(blk, offset, qiov, 0, aio_rw_done, &async_ret);
    while (async_ret == NOT_DONE) {
        main_loop_wait(false);
    }

    *total = qiov->size;
    return async_ret < 0 ? async_ret : 1;
}

static int do_aio_writev(BlockBackend *blk, QEMUIOVector *qiov,
                         int64_t offset, int flags, int *total)
{
    int async_ret = NOT_DONE;

    blk_aio_pwritev(blk, offset, qiov, flags, aio_rw_done, &async_ret);
    while (async_ret == NOT_DONE) {
        main_loop_wait(false);
    }

    *total = qiov->size;
    return async_ret < 0 ? async_ret : 1;
}

static void read_help(void)
{
    printf(
"\n"
" reads a range of bytes from the given offset\n"
"\n"
" Example:\n"
" 'read -v 512 1k' - dumps 1 kilobyte read from 512 bytes into the file\n"
"\n"
" Reads a segment of the currently open file, optionally dumping it to the\n"
" standard output stream (with -v option) for subsequent inspection.\n"
" -b, -- read from the VM state rather than the virtual disk\n"
" -C, -- report statistics in a machine parsable format\n"
" -l, -- length for pattern verification (only with -P)\n"
" -p, -- ignored for backwards compatibility\n"
" -P, -- use a pattern to verify read data\n"
" -q, -- quiet mode, do not show I/O statistics\n"
" -s, -- start offset for pattern verification (only with -P)\n"
" -v, -- dump buffer to standard output\n"
"\n");
}

static int read_f(BlockBackend *blk, int argc, char **argv);

static const cmdinfo_t read_cmd = {
    .name       = "read",
    .altname    = "r",
    .cfunc      = read_f,
    .argmin     = 2,
    .argmax     = -1,
    .args       = "[-abCqv] [-P pattern [-s off] [-l len]] off len",
    .oneline    = "reads a number of bytes at a specified offset",
    .help       = read_help,
};

static int read_f(BlockBackend *blk, int argc, char **argv)
{
    struct timeval t1, t2;
    bool Cflag = false, qflag = false, vflag = false;
    bool Pflag = false, sflag = false, lflag = false, bflag = false;
    int c, cnt;
    char *buf;
    int64_t offset;
    int64_t count;
    /* Some compilers get confused and warn if this is not initialized.  */
    int64_t total = 0;
    int pattern = 0;
    int64_t pattern_offset = 0, pattern_count = 0;

    while ((c = getopt(argc, argv, "bCl:pP:qs:v")) != -1) {
        switch (c) {
        case 'b':
            bflag = true;
            break;
        case 'C':
            Cflag = true;
            break;
        case 'l':
            lflag = true;
            pattern_count = cvtnum(optarg);
            if (pattern_count < 0) {
                print_cvtnum_err(pattern_count, optarg);
                return 0;
            }
            break;
        case 'p':
            /* Ignored for backwards compatibility */
            break;
        case 'P':
            Pflag = true;
            pattern = parse_pattern(optarg);
            if (pattern < 0) {
                return 0;
            }
            break;
        case 'q':
            qflag = true;
            break;
        case 's':
            sflag = true;
            pattern_offset = cvtnum(optarg);
            if (pattern_offset < 0) {
                print_cvtnum_err(pattern_offset, optarg);
                return 0;
            }
            break;
        case 'v':
            vflag = true;
            break;
        default:
            return qemuio_command_usage(&read_cmd);
        }
    }

    if (optind != argc - 2) {
        return qemuio_command_usage(&read_cmd);
    }

    offset = cvtnum(argv[optind]);
    if (offset < 0) {
        print_cvtnum_err(offset, argv[optind]);
        return 0;
    }

    optind++;
    count = cvtnum(argv[optind]);
    if (count < 0) {
        print_cvtnum_err(count, argv[optind]);
        return 0;
    } else if (count > SIZE_MAX) {
        printf("length cannot exceed %" PRIu64 ", given %s\n",
               (uint64_t) SIZE_MAX, argv[optind]);
        return 0;
    }

    if (!Pflag && (lflag || sflag)) {
        return qemuio_command_usage(&read_cmd);
    }

    if (!lflag) {
        pattern_count = count - pattern_offset;
    }

    if ((pattern_count < 0) || (pattern_count + pattern_offset > count))  {
        printf("pattern verification range exceeds end of read data\n");
        return 0;
    }

    if (bflag) {
        if (offset & 0x1ff) {
            printf("offset %" PRId64 " is not sector aligned\n",
                   offset);
            return 0;
        }
        if (count & 0x1ff) {
            printf("count %"PRId64" is not sector aligned\n",
                   count);
            return 0;
        }
    }

    buf = qemu_io_alloc(blk, count, 0xab);

    gettimeofday(&t1, NULL);
    if (bflag) {
        cnt = do_load_vmstate(blk, buf, offset, count, &total);
    } else {
        cnt = do_pread(blk, buf, offset, count, &total);