darshan-convert.c 10.4 KB
Newer Older
1
/*
2 3 4
 * Copyright (C) 2015 University of Chicago.
 * See COPYRIGHT notice in top-level directory.
 *
5 6 7 8 9 10 11 12 13 14 15 16 17
 */

#include <stdio.h>
#include <sys/types.h>
#include <sys/stat.h>
#include <fcntl.h>
#include <unistd.h>
#include <string.h>
#include <zlib.h>
#include <time.h>
#include <stdlib.h>
#include <getopt.h>
#include <assert.h>
18
#include <errno.h>
19 20 21

#include "darshan-logutils.h"

22
extern uint32_t darshan_hashlittle(const void *key, size_t length, uint32_t initval);
23 24 25 26 27 28

int usage (char *exename)
{
    fprintf(stderr, "Usage: %s [options] <infile> <outfile>\n", exename);
    fprintf(stderr, "       Converts darshan log from infile to outfile.\n");
    fprintf(stderr, "       rewrites the log file into the newest format.\n");
29
    fprintf(stderr, "       --bzip2 Use bzip2 compression instead of zlib.\n");
30
    fprintf(stderr, "       --obfuscate Obfuscate items in the log.\n");
31 32
    fprintf(stderr, "       --key <key> Key to use when obfuscating.\n");
    fprintf(stderr, "       --annotate <string> Additional metadata to add.\n");
33
    fprintf(stderr, "       --file <hash> Limit output to specified (hashed) file only.\n");
34
    fprintf(stderr, "       --reset-md Reset old metadata during conversion.\n");
35 36 37 38

    exit(1);
}

39
void parse_args (int argc, char **argv, char **infile, char **outfile,
40 41
                 int *bzip2, int *obfuscate, int *reset_md, int *key,
                 char **annotate, uint64_t* hash)
42 43
{
    int index;
44 45
    int ret;

46 47
    static struct option long_opts[] =
    {
48
        {"bzip2", 0, NULL, 'b'},
49
        {"annotate", 1, NULL, 'a'},
50
        {"obfuscate", 0, NULL, 'o'},
51
        {"reset-md", 0, NULL, 'r'},
52
        {"key", 1, NULL, 'k'},
53
        {"file", 1, NULL, 'f'},
54 55
        {"help",  0, NULL, 0},
        { 0, 0, 0, 0 }
56 57
    };

58
    *bzip2 = 0;
59
    *obfuscate = 0;
60
    *reset_md = 0;
61
    *key = 0;
62 63
    *hash = 0;

64 65 66 67 68 69 70 71
    while(1)
    {
        int c = getopt_long(argc, argv, "", long_opts, &index);

        if (c == -1) break;

        switch(c)
        {
72 73 74
            case 'b':
                *bzip2 = 1;
                break;
75 76 77
            case 'a':
                *annotate = optarg;
                break;
78 79 80
            case 'o':
                *obfuscate = 1;
                break;
81 82 83
            case 'r':
                *reset_md = 1;
                break;
84 85 86
            case 'k':
                *key = atoi(optarg);
                break;
87 88 89 90 91
            case 'f':
                ret = sscanf(optarg, "%" PRIu64, hash);
                if(ret != 1)
                    usage(argv[0]);
                break;
92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112
            case 0:
            case '?':
            default:
                usage(argv[0]);
                break;
        }
    }

    if (optind + 2 == argc)
    {
        *infile = argv[optind];
        *outfile = argv[optind+1];
    }
    else
    {
        usage(argv[0]);
    }

    return;
}

113 114 115 116 117 118
static void reset_md_job(struct darshan_job *job)
{
    job->metadata[0] = '\0';
    return;
}

119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140
void obfuscate_job(int key, struct darshan_job *job)
{
    job->uid   = (int64_t) darshan_hashlittle(&job->uid, sizeof(job->uid), key);
    if (job->jobid != 0)
    {
        job->jobid = (int64_t) darshan_hashlittle(&job->jobid, sizeof(job->jobid), key);
    }

    return;
}

void obfuscate_exe(int key, char *exe)
{
    uint32_t hashed;

    hashed = darshan_hashlittle(exe, strlen(exe), key);
    memset(exe, 0, strlen(exe));
    sprintf(exe, "%u", hashed);

    return;
}

141
void obfuscate_filenames(int key, struct darshan_record_ref *rec_hash)
142
{
143
    struct darshan_record_ref *ref, *tmp;
144
    uint32_t hashed;
145
    char tmp_string[128] = {0};
146

147 148 149 150 151
    HASH_ITER(hlink, rec_hash, ref, tmp)
    {
        hashed = darshan_hashlittle(ref->rec.name, strlen(ref->rec.name), key);
        sprintf(tmp_string, "%u", hashed);
        free(ref->rec.name);
152
        ref->rec.name = malloc(strlen(tmp_string) + 1);
153 154
        assert(ref->rec.name);
        memcpy(ref->rec.name, tmp_string, strlen(tmp_string));
155
        ref->rec.name[strlen(tmp_string)] = '\0';
156
    }
157 158 159 160

    return;
}

161 162 163 164 165
void add_annotation (char *annotation,
                     struct darshan_job *job)
{
    char *token;
    char *save;
166
    int len;
167
    
168 169 170 171 172 173 174 175 176 177 178
    /* check for newline in existing metadata, insert if needed */
    len = strlen(job->metadata);
    if(len > 0 && len < sizeof(job->metadata))
    {
        if(job->metadata[len-1] != '\n')
        {
            job->metadata[len] = '\n';
            job->metadata[len+1] = '\0';
        }
    }

179 180
    /* determine remaining space in metadata string */
    int remaining = sizeof(job->metadata) - strlen(job->metadata);
181 182 183 184 185

    for(token=strtok_r(annotation, "\t", &save);
        token != NULL;
        token=strtok_r(NULL, "\t", &save))
    {
186 187 188 189 190 191 192 193 194 195
        if ((strlen(token)+1) < remaining)
        {
            strcat(job->metadata, token);
            strcat(job->metadata, "\n");
            remaining -= (strlen(token)+1);
        }
        else
        {
            fprintf(stderr,
                    "not enough space left in metadata for: current=%s token=%s (remain=%d:need=%d)\n",
196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213
                    job->metadata, token, remaining-1, (int)strlen(token)+1);
        }
    }

    return;
}

static void remove_hash_recs(struct darshan_record_ref **rec_hash, darshan_record_id hash)
{
    struct darshan_record_ref *ref, *tmp;

    HASH_ITER(hlink, *rec_hash, ref, tmp)
    {
        if(ref->rec.id != hash)
        {
            HASH_DELETE(hlink, *rec_hash, ref);
            free(ref->rec.name);
            free(ref);
214
        }
215 216 217 218 219
    }

    return;
}

220 221 222 223 224 225
int main(int argc, char **argv)
{
    int ret;
    char *infile_name;
    char *outfile_name;
    struct darshan_job job;
226
    char tmp_string[4096] = {0};
227 228 229 230
    darshan_fd infile;
    darshan_fd outfile;
    int i;
    int mount_count;
231
    struct darshan_mnt_info *mnt_data_array;
232 233
    struct darshan_record_ref *rec_hash = NULL;
    struct darshan_record_ref *ref, *tmp;
234
    char *mod_buf;
235
    enum darshan_comp_type comp_type;
236 237 238
    int bzip2;
    int obfuscate;
    int key;
239
    char *annotation = NULL;
240 241
    darshan_record_id hash;
    int reset_md;
242

243 244
    parse_args(argc, argv, &infile_name, &outfile_name, &bzip2, &obfuscate,
               &reset_md, &key, &annotation, &hash);
245

246
    infile = darshan_log_open(infile_name);
247 248 249
    if(!infile)
        return(-1);
 
250
    comp_type = bzip2 ? comp_type = DARSHAN_BZIP2_COMP : DARSHAN_ZLIB_COMP;
251
    outfile = darshan_log_create(outfile_name, comp_type, infile->partial_flag);
252 253
    if(!outfile)
    {
254 255 256 257
        darshan_log_close(infile);
        return(-1);
    }

258 259 260 261 262
    /* read job info */
    ret = darshan_log_getjob(infile, &job);
    if(ret < 0)
    {
        darshan_log_close(infile);
263
        darshan_log_close(outfile);
264
        unlink(outfile_name);
265 266 267
        return(-1);
    }

268
    if (reset_md) reset_md_job(&job);
269
    if (obfuscate) obfuscate_job(key, &job);
270
    if (annotation) add_annotation(annotation, &job);
271

272 273 274
    ret = darshan_log_putjob(outfile, &job);
    if (ret < 0)
    {
275
        darshan_log_close(infile);
276 277 278 279
        darshan_log_close(outfile);
        return(-1);
    }

280
    ret = darshan_log_getexe(infile, tmp_string);
281 282 283
    if(ret < 0)
    {
        darshan_log_close(infile);
284
        darshan_log_close(outfile);
285
        unlink(outfile_name);
286 287
        return(-1);
    }
288

289 290
    if (obfuscate) obfuscate_exe(key, tmp_string);

291 292 293
    ret = darshan_log_putexe(outfile, tmp_string);
    if(ret < 0)
    {
294
        darshan_log_close(infile);
295 296 297
        darshan_log_close(outfile);
        return(-1);
    }
298

299
    ret = darshan_log_getmounts(infile, &mnt_data_array, &mount_count);
300 301 302
    if(ret < 0)
    {
        darshan_log_close(infile);
303
        darshan_log_close(outfile);
304
        unlink(outfile_name);
305 306 307
        return(-1);
    }

308
    ret = darshan_log_putmounts(outfile, mnt_data_array, mount_count);
309 310
    if(ret < 0)
    {
311
        darshan_log_close(infile);
312 313 314 315
        darshan_log_close(outfile);
        return(-1);
    }

316
    ret = darshan_log_gethash(infile, &rec_hash);
317 318
    if(ret < 0)
    {
319 320
        darshan_log_close(infile);
        darshan_log_close(outfile);
321
        unlink(outfile_name);
322
        return(-1);
323
    }
324

325 326 327 328 329 330
    /* NOTE: obfuscating filepaths breaks the ability to map files
     * to the corresponding FS & mount info maintained by darshan
     */
    if(obfuscate) obfuscate_filenames(key, rec_hash);
    if(hash) remove_hash_recs(&rec_hash, hash);

331 332
    ret = darshan_log_puthash(outfile, rec_hash);
    if(ret < 0)
333
    {
334 335 336
        darshan_log_close(infile);
        darshan_log_close(outfile);
        return(-1);
337 338
    }

339 340 341 342 343 344 345 346
    mod_buf = malloc(DEF_MOD_BUF_SIZE);
    if (!mod_buf)
    {
        darshan_log_close(infile);
        darshan_log_close(outfile);
        return(-1);
    }

347
    /* loop over each module and convert it's data to the new format */
348
    for(i=0; i<DARSHAN_MAX_MODS; i++)
349
    {
350 351
        darshan_record_id rec_id;

352
        /* check each module for any data */
353
        if(infile->mod_map[i].len == 0)
354
            continue;
355
        else if(!mod_logutils[i])
356 357
        {
            fprintf(stderr, "Warning: no log utility handlers defined "
358
                "for module %s, SKIPPING.\n", darshan_module_names[i]);
359 360 361
            continue;
        }

362
        /* we have module data to convert */
363
        memset(mod_buf, 0, DEF_MOD_BUF_SIZE);
364

365 366
        ret = mod_logutils[i]->log_get_record(infile, mod_buf, &rec_id);
        if(ret != 1)
367
        {
368
            fprintf(stderr, "Error: failed to parse the first %s module record.\n",
369 370 371
                darshan_module_names[i]);
            darshan_log_close(infile);
            darshan_log_close(outfile);
372
            unlink(outfile_name);
373
            free(mod_buf);
374 375
            return(-1);
        }
376

377 378 379 380 381
        /* loop over each of the module's records and convert */
        do
        {
            if(!hash || hash == rec_id)
            {
382
                ret = mod_logutils[i]->log_put_record(outfile, mod_buf, infile->mod_ver[i]);
383 384 385 386 387 388 389 390 391 392
                if(ret < 0)
                {
                    darshan_log_close(infile);
                    darshan_log_close(outfile);
                    return(-1);
                }

                memset(mod_buf, 0, DEF_MOD_BUF_SIZE);
            }
        } while((ret = mod_logutils[i]->log_get_record(infile, mod_buf, &rec_id)) == 1);
393
    }
394
    free(mod_buf);
395 396 397 398

    darshan_log_close(infile);
    darshan_log_close(outfile);

399
    if(mount_count > 0)
400
        free(mnt_data_array);
401 402 403 404 405 406 407 408

    HASH_ITER(hlink, rec_hash, ref, tmp)
    {
        HASH_DELETE(hlink, rec_hash, ref);
        free(ref->rec.name);
        free(ref);
    }

409
    return(ret);
410 411
}

412 413 414 415 416 417 418 419
/*
 * Local variables:
 *  c-indent-level: 4
 *  c-basic-offset: 4
 * End:
 *
 * vim: ts=8 sts=4 sw=4 expandtab
 */