darshan-posix.c 58.2 KB
Newer Older
1
/*
Shane Snyder's avatar
Shane Snyder committed
2 3 4
 * Copyright (C) 2015 University of Chicago.
 * See COPYRIGHT notice in top-level directory.
 *
5 6
 */

7 8 9
#define _XOPEN_SOURCE 500
#define _GNU_SOURCE

10
#include "darshan-runtime-config.h"
11 12 13 14 15 16 17 18 19 20 21 22 23 24
#include <stdio.h>
#include <unistd.h>
#include <sys/types.h>
#include <sys/stat.h>
#include <fcntl.h>
#include <stdarg.h>
#include <string.h>
#include <time.h>
#include <stdlib.h>
#include <errno.h>
#include <sys/uio.h>
#include <sys/mman.h>
#include <search.h>
#include <assert.h>
25
#include <libgen.h>
26
#include <aio.h>
27
#include <pthread.h>
28

29
#include "uthash.h"
30
#include "darshan.h"
31
#include "darshan-posix-log-format.h"
32

33
#ifndef HAVE_OFF64_T
34 35
typedef int64_t off64_t;
#endif
36 37 38
#ifndef HAVE_AIOCB64
#define aiocb64 aiocb
#endif
39

40
/* TODO: this probably shouldn't be here long term -- MPI symbols mess up LD_PRELOAD */
Shane Snyder's avatar
Shane Snyder committed
41 42 43 44
#ifdef DARSHAN_PRELOAD
extern double (*__real_PMPI_Comm_rank)(MPI_Comm comm, int *rank);
#endif

45 46
/* TODO: more libc, fgetc, etc etc etc. */

47 48
DARSHAN_FORWARD_DECL(open, int, (const char *path, int flags, ...));
DARSHAN_FORWARD_DECL(open64, int, (const char *path, int flags, ...));
49 50
DARSHAN_FORWARD_DECL(creat, int, (const char* path, mode_t mode));
DARSHAN_FORWARD_DECL(creat64, int, (const char* path, mode_t mode));
51 52
DARSHAN_FORWARD_DECL(fopen, FILE*, (const char *path, const char *mode));
DARSHAN_FORWARD_DECL(fopen64, FILE*, (const char *path, const char *mode));
53 54 55 56
DARSHAN_FORWARD_DECL(mkstemp, int, (char *template));
DARSHAN_FORWARD_DECL(mkostemp, int, (char *template, int flags));
DARSHAN_FORWARD_DECL(mkstemps, int, (char *template, int suffixlen));
DARSHAN_FORWARD_DECL(mkostemps, int, (char *template, int suffixlen, int flags));
57 58 59 60 61 62 63 64
DARSHAN_FORWARD_DECL(read, ssize_t, (int fd, void *buf, size_t count));
DARSHAN_FORWARD_DECL(write, ssize_t, (int fd, const void *buf, size_t count));
DARSHAN_FORWARD_DECL(pread, ssize_t, (int fd, void *buf, size_t count, off_t offset));
DARSHAN_FORWARD_DECL(pwrite, ssize_t, (int fd, const void *buf, size_t count, off_t offset));
DARSHAN_FORWARD_DECL(pread64, ssize_t, (int fd, void *buf, size_t count, off64_t offset));
DARSHAN_FORWARD_DECL(pwrite64, ssize_t, (int fd, const void *buf, size_t count, off64_t offset));
DARSHAN_FORWARD_DECL(readv, ssize_t, (int fd, const struct iovec *iov, int iovcnt));
DARSHAN_FORWARD_DECL(writev, ssize_t, (int fd, const struct iovec *iov, int iovcnt));
65 66 67 68 69
DARSHAN_FORWARD_DECL(fread, size_t, (void *ptr, size_t size, size_t nmemb, FILE *stream));
DARSHAN_FORWARD_DECL(fwrite, size_t, (const void *ptr, size_t size, size_t nmemb, FILE *stream));
DARSHAN_FORWARD_DECL(lseek, off_t, (int fd, off_t offset, int whence));
DARSHAN_FORWARD_DECL(lseek64, off64_t, (int fd, off64_t offset, int whence));
DARSHAN_FORWARD_DECL(fseek, int, (FILE *stream, long offset, int whence));
70 71 72 73 74 75
DARSHAN_FORWARD_DECL(__xstat, int, (int vers, const char* path, struct stat *buf));
DARSHAN_FORWARD_DECL(__xstat64, int, (int vers, const char* path, struct stat64 *buf));
DARSHAN_FORWARD_DECL(__lxstat, int, (int vers, const char* path, struct stat *buf));
DARSHAN_FORWARD_DECL(__lxstat64, int, (int vers, const char* path, struct stat64 *buf));
DARSHAN_FORWARD_DECL(__fxstat, int, (int vers, int fd, struct stat *buf));
DARSHAN_FORWARD_DECL(__fxstat64, int, (int vers, int fd, struct stat64 *buf));
Shane Snyder's avatar
Shane Snyder committed
76 77
DARSHAN_FORWARD_DECL(mmap, void*, (void *addr, size_t length, int prot, int flags, int fd, off_t offset));
DARSHAN_FORWARD_DECL(mmap64, void*, (void *addr, size_t length, int prot, int flags, int fd, off64_t offset));
78 79
DARSHAN_FORWARD_DECL(fsync, int, (int fd));
DARSHAN_FORWARD_DECL(fdatasync, int, (int fd));
80
DARSHAN_FORWARD_DECL(close, int, (int fd));
81
DARSHAN_FORWARD_DECL(fclose, int, (FILE *fp));
82 83
/* TODO aio */
/* TODO listio */
84

85 86 87 88 89
/* maximum number of access sizes and stride sizes that darshan will track
 * per file at runtime; at log time they will be reduced into the 4 most
 * frequently occurring ones 
 */
#define POSIX_MAX_ACCESS_COUNT_RUNTIME 32
90

91 92 93 94 95 96
enum posix_io_type
{
    POSIX_READ = 1,
    POSIX_WRITE = 2,
};

97 98 99 100 101 102 103 104 105 106 107 108
enum posix_counter_type
{
    POSIX_COUNTER_ACCESS,
    POSIX_COUNTER_STRIDE
};

struct posix_access_counter
{
    int64_t size;
    int freq;
};

109 110 111 112 113 114
/* The posix_file_runtime structure maintains necessary runtime metadata
 * for the POSIX file record (darshan_posix_file structure, defined in
 * darshan-posix-log-format.h) pointed to by 'file_record'. This metadata
 * assists with the instrumenting of specific statistics in the file record.
 * 'hlink' is a hash table link structure used to add/remove this record
 * from the hash table of POSIX file records for this process. 
115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130
 *
 * RATIONALE: the POSIX module needs to track some stateful, volatile 
 * information about each open file (like the current file offset, most recent 
 * access time, etc.) to aid in instrumentation, but this information can't be
 * stored in the darshan_posix_file struct because we don't want it to appear in
 * the final darshan log file.  We therefore associate a posix_file_runtime
 * struct with each darshan_posix_file struct in order to track this information.
  *
 * NOTE: There is a one-to-one mapping of posix_file_runtime structs to
 * darshan_posix_file structs.
 *
 * NOTE: The posix_file_runtime struct contains a pointer to a darshan_posix_file
 * struct (see the *file_record member) rather than simply embedding an entire
 * darshan_posix_file struct.  This is done so that all of the darshan_posix_file
 * structs can be kept contiguous in memory as a single array to simplify
 * reduction, compression, and storage.
131
 */
132
struct posix_file_runtime
133
{
134
    struct darshan_posix_file* file_record;
135 136 137 138
    int64_t offset;
    int64_t last_byte_read;
    int64_t last_byte_written;
    enum posix_io_type last_io_type;
139 140 141
    double last_meta_end;
    double last_read_end;
    double last_write_end;
142 143 144 145
    void* access_root;
    int access_count;
    void* stride_root;
    int stride_count;
146
    UT_hash_handle hlink;
147
};
148

149 150 151 152 153 154 155 156 157 158 159
/* The posix_file_runtime_ref structure is used to associate a POSIX
 * file descriptor with an already existing POSIX file record. This is
 * necessary as many POSIX I/O functions take only an input file descriptor,
 * but POSIX file records are indexed by their full file paths (i.e., darshan
 * record identifiers for POSIX files are created by hashing the file path).
 * In other words, this structure is necessary as it allows us to look up a
 * file record either by a pathname (posix_file_runtime) or by POSIX file
 * descriptor (posix_file_runtime_ref), depending on which parameters are
 * available. This structure includes another hash table link, since separate
 * hashes are maintained for posix_file_runtime structures and posix_file_runtime_ref
 * structures.
160 161 162 163 164 165 166 167 168 169 170
 *
 * RATIONALE: In theory the fd information could be included in the
 * posix_file_runtime struct rather than in a separate structure here.  The
 * reason we don't do that is because the same file could be opened multiple
 * times by a given process with different file descriptors and thus
 * simulataneously referenced using different file descriptors.  This practice is
 * not common, but we must support it.
 *
 * NOTE: there are potentially multiple posix_file_runtime_ref structures
 * referring to a single posix_file_runtime structure.  Most of the time there is
 * only one, however.
171
 */
172
struct posix_file_runtime_ref
173
{
174
    struct posix_file_runtime* file;
175 176 177 178
    int fd;
    UT_hash_handle hlink;
};

179 180 181 182
/* The posix_runtime structure maintains necessary state for storing
 * POSIX file records and for coordinating with darshan-core at 
 * shutdown time.
 */
183 184
struct posix_runtime
{
185
    struct posix_file_runtime* file_runtime_array;
186
    struct darshan_posix_file* file_record_array;
187
    int file_array_size;
188
    int file_array_ndx;
189 190
    struct posix_file_runtime* file_hash;
    struct posix_file_runtime_ref* fd_hash;
191 192
    void *red_buf;
    int shared_rec_count;
193 194
};

195
static struct posix_runtime *posix_runtime = NULL;
196
static pthread_mutex_t posix_runtime_mutex = PTHREAD_RECURSIVE_MUTEX_INITIALIZER_NP;
Shane Snyder's avatar
Shane Snyder committed
197
static int instrumentation_disabled = 0;
198
static int my_rank = -1;
199
static int darshan_mem_alignment = 1;
200

201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225
/* global variables for determining 4 most common accesses/strides */
static struct posix_file_runtime* walker_file = NULL;
static int walker_validx;
static int walker_cntidx;

static void posix_runtime_initialize(void);
static struct posix_file_runtime* posix_file_by_name(const char *name);
static struct posix_file_runtime* posix_file_by_name_setfd(const char* name, int fd);
static struct posix_file_runtime* posix_file_by_fd(int fd);
static void posix_file_close_fd(int fd);
static void posix_access_counter(struct posix_file_runtime* file, ssize_t size,
    enum posix_counter_type type);
static void posix_access_walker(const void* nodep, const VISIT which, const int depth);
static void posix_walk_file_accesses(void);
static int posix_access_compare(const void* a_p, const void* b_p);
static int posix_file_compare(const void* a, const void* b);

static void posix_disable_instrumentation(void);
static void posix_prepare_for_reduction(darshan_record_id *shared_recs,
    int *shared_rec_count, void **send_buf, void **recv_buf, int *rec_size);
static void posix_record_reduction_op(void* infile_v, void* inoutfile_v,
    int *len, MPI_Datatype *datatype);
static void posix_get_output_data(void **buffer, int *size);
static void posix_shutdown(void);

226 227
#define POSIX_LOCK() pthread_mutex_lock(&posix_runtime_mutex)
#define POSIX_UNLOCK() pthread_mutex_unlock(&posix_runtime_mutex)
228

229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253
#define POSIX_COMMON_COUNTER_INC(__rec_p, __value, __count, __validx, __cntidx) do {\
    int i; \
    int set = 0; \
    int64_t min = DARSHAN_COUNTER_VALUE(__rec_p, __cntidx); \
    int min_index = 0; \
    if(__value == 0) break; \
    for(i=0; i<4; i++) { \
        /* increment bucket if already exists */ \
        if(DARSHAN_COUNTER_VALUE(__rec_p, __validx + i) == __value) { \
            DARSHAN_COUNTER_INC(__rec_p, __cntidx + i, __count); \
            set = 1; \
            break; \
        } \
        /* otherwise find the least frequently used bucket */ \
        else if(DARSHAN_COUNTER_VALUE(__rec_p, __cntidx + i) < min) { \
            min = DARSHAN_COUNTER_VALUE(__rec_p, __cntidx + i); \
            min_index = i; \
        } \
    } \
    if(!set && (__count > min)) { \
        DARSHAN_COUNTER_SET(__rec_p, __cntidx+min_index, __count); \
        DARSHAN_COUNTER_SET(__rec_p, __validx+min_index, __value); \
    } \
} while(0)

254
#define POSIX_RECORD_OPEN(__ret, __path, __mode, __stream_flag, __tm1, __tm2) do { \
255
    struct posix_file_runtime* file; \
256 257 258
    char* exclude; \
    int tmp_index = 0; \
    if(__ret < 0) break; \
259
    while((exclude = darshan_path_exclusions[tmp_index])) { \
260 261 262 263 264
        if(!(strncmp(exclude, __path, strlen(exclude)))) \
            break; \
        tmp_index++; \
    } \
    if(exclude) break; \
265
    file = posix_file_by_name_setfd(__path, __ret); \
266
    if(!file) break; \
267
    file->file_record->rank = my_rank; \
268
    if(__mode) \
269 270 271 272
        DARSHAN_COUNTER_SET(file->file_record, POSIX_MODE, __mode); \
    file->offset = 0; \
    file->last_byte_written = 0; \
    file->last_byte_read = 0; \
273
    if(__stream_flag)\
274
        DARSHAN_COUNTER_INC(file->file_record, POSIX_FOPENS, 1); \
275
    else \
276 277 278 279
        DARSHAN_COUNTER_INC(file->file_record, POSIX_OPENS, 1); \
    if(DARSHAN_COUNTER_F_VALUE(file->file_record, POSIX_F_OPEN_TIMESTAMP) == 0) \
        DARSHAN_COUNTER_F_SET(file->file_record, POSIX_F_OPEN_TIMESTAMP, __tm1); \
    DARSHAN_COUNTER_F_INC_NO_OVERLAP(file->file_record, __tm1, __tm2, file->last_meta_end, POSIX_F_META_TIME); \
280 281 282
} while(0)

#define POSIX_RECORD_READ(__ret, __fd, __pread_flag, __pread_offset, __aligned, __stream_flag, __tm1, __tm2) do{ \
283
    size_t stride; \
284
    int64_t this_offset; \
285
    struct posix_file_runtime* file; \
286
    int64_t file_alignment; \
287 288 289 290
    double __elapsed = __tm2-__tm1; \
    if(__ret < 0) break; \
    file = posix_file_by_fd(__fd); \
    if(!file) break; \
291
    if(__pread_flag) \
292 293 294 295
        this_offset = __pread_offset; \
    else \
        this_offset = file->offset; \
    if(this_offset > file->last_byte_read) \
296
        DARSHAN_COUNTER_INC(file->file_record, POSIX_SEQ_READS, 1); \
297
    if(this_offset == (file->last_byte_read + 1)) \
298
        DARSHAN_COUNTER_INC(file->file_record, POSIX_CONSEC_READS, 1); \
299
    if(this_offset > 0 && this_offset > file->last_byte_read \
300 301 302
        && file->last_byte_read != 0) \
        stride = this_offset - file->last_byte_read - 1; \
    else \
303
        stride = 0; \
304 305
    file->last_byte_read = this_offset + __ret - 1; \
    file->offset = this_offset + __ret; \
306 307
    DARSHAN_COUNTER_MAX(file->file_record, POSIX_MAX_BYTE_READ, (this_offset + __ret - 1)); \
    DARSHAN_COUNTER_INC(file->file_record, POSIX_BYTES_READ, __ret); \
308
    if(__stream_flag)\
309
        DARSHAN_COUNTER_INC(file->file_record, POSIX_FREADS, 1); \
310
    else\
311
        DARSHAN_COUNTER_INC(file->file_record, POSIX_READS, 1); \
312 313 314
    DARSHAN_BUCKET_INC(file->file_record, POSIX_SIZE_READ_0_100, __ret); \
    posix_access_counter(file, __ret, POSIX_COUNTER_ACCESS); \
    posix_access_counter(file, stride, POSIX_COUNTER_STRIDE); \
315
    if(!__aligned) \
316 317
        DARSHAN_COUNTER_INC(file->file_record, POSIX_MEM_NOT_ALIGNED, 1); \
    file_alignment = DARSHAN_COUNTER_VALUE(file->file_record, POSIX_FILE_ALIGNMENT); \
318
    if(file_alignment > 0 && (this_offset % file_alignment) != 0) \
319
        DARSHAN_COUNTER_INC(file->file_record, POSIX_FILE_NOT_ALIGNED, 1); \
320 321 322 323 324 325 326 327 328 329
    if(file->last_io_type == POSIX_WRITE) \
        DARSHAN_COUNTER_INC(file->file_record, POSIX_RW_SWITCHES, 1); \
    file->last_io_type = POSIX_READ; \
    DARSHAN_COUNTER_F_INC_NO_OVERLAP(file->file_record, __tm1, __tm2, file->last_read_end, POSIX_F_READ_TIME); \
    if(DARSHAN_COUNTER_F_VALUE(file->file_record, POSIX_F_READ_START_TIMESTAMP) == 0) \
        DARSHAN_COUNTER_F_SET(file->file_record, POSIX_F_READ_START_TIMESTAMP, __tm1); \
    DARSHAN_COUNTER_F_SET(file->file_record, POSIX_F_READ_END_TIMESTAMP, __tm2); \
    if(DARSHAN_COUNTER_F_VALUE(file->file_record, POSIX_F_MAX_READ_TIME) < __elapsed){ \
        DARSHAN_COUNTER_F_SET(file->file_record, POSIX_F_MAX_READ_TIME, __elapsed); \
        DARSHAN_COUNTER_SET(file->file_record, POSIX_MAX_READ_TIME_SIZE, __ret); } \
330 331 332
} while(0)

#define POSIX_RECORD_WRITE(__ret, __fd, __pwrite_flag, __pwrite_offset, __aligned, __stream_flag, __tm1, __tm2) do{ \
333
    size_t stride; \
334
    int64_t this_offset; \
335
    struct posix_file_runtime* file; \
336
    int64_t file_alignment; \
337 338 339 340
    double __elapsed = __tm2-__tm1; \
    if(__ret < 0) break; \
    file = posix_file_by_fd(__fd); \
    if(!file) break; \
341
    if(__pwrite_flag) \
342 343 344 345
        this_offset = __pwrite_offset; \
    else \
        this_offset = file->offset; \
    if(this_offset > file->last_byte_written) \
346
        DARSHAN_COUNTER_INC(file->file_record, POSIX_SEQ_WRITES, 1); \
347
    if(this_offset == (file->last_byte_written + 1)) \
348
        DARSHAN_COUNTER_INC(file->file_record, POSIX_CONSEC_WRITES, 1); \
349
    if(this_offset > 0 && this_offset > file->last_byte_written \
350 351
        && file->last_byte_written != 0) \
        stride = this_offset - file->last_byte_written - 1; \
352
    else \
353
        stride = 0; \
354 355
    file->last_byte_written = this_offset + __ret - 1; \
    file->offset = this_offset + __ret; \
356 357
    DARSHAN_COUNTER_MAX(file->file_record, POSIX_MAX_BYTE_WRITTEN, (this_offset + __ret - 1)); \
    DARSHAN_COUNTER_INC(file->file_record, POSIX_BYTES_WRITTEN, __ret); \
358
    if(__stream_flag) \
359
        DARSHAN_COUNTER_INC(file->file_record, POSIX_FWRITES, 1); \
360
    else \
361
        DARSHAN_COUNTER_INC(file->file_record, POSIX_WRITES, 1); \
362 363 364
    DARSHAN_BUCKET_INC(file->file_record, POSIX_SIZE_WRITE_0_100, __ret); \
    posix_access_counter(file, __ret, POSIX_COUNTER_ACCESS); \
    posix_access_counter(file, stride, POSIX_COUNTER_STRIDE);  \
365
    if(!__aligned) \
366 367
        DARSHAN_COUNTER_INC(file->file_record, POSIX_MEM_NOT_ALIGNED, 1); \
    file_alignment = DARSHAN_COUNTER_VALUE(file->file_record, POSIX_FILE_ALIGNMENT); \
368
    if(file_alignment > 0 && (this_offset % file_alignment) != 0) \
369
        DARSHAN_COUNTER_INC(file->file_record, POSIX_FILE_NOT_ALIGNED, 1); \
370 371 372 373 374 375 376 377 378 379
    if(file->last_io_type == POSIX_READ) \
        DARSHAN_COUNTER_INC(file->file_record, POSIX_RW_SWITCHES, 1); \
    file->last_io_type = POSIX_WRITE; \
    DARSHAN_COUNTER_F_INC_NO_OVERLAP(file->file_record, __tm1, __tm2, file->last_write_end, POSIX_F_WRITE_TIME); \
    if(DARSHAN_COUNTER_F_VALUE(file->file_record, POSIX_F_WRITE_START_TIMESTAMP) == 0) \
        DARSHAN_COUNTER_F_SET(file->file_record, POSIX_F_WRITE_START_TIMESTAMP, __tm1); \
    DARSHAN_COUNTER_F_SET(file->file_record, POSIX_F_WRITE_END_TIMESTAMP, __tm2); \
    if(DARSHAN_COUNTER_F_VALUE(file->file_record, POSIX_F_MAX_WRITE_TIME) < __elapsed){ \
        DARSHAN_COUNTER_F_SET(file->file_record, POSIX_F_MAX_WRITE_TIME, __elapsed); \
        DARSHAN_COUNTER_SET(file->file_record, POSIX_MAX_WRITE_TIME_SIZE, __ret); } \
380
} while(0)
381

382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404
#define POSIX_LOOKUP_RECORD_STAT(__path, __statbuf, __tm1, __tm2) do { \
    char* exclude; \
    int tmp_index = 0; \
    struct posix_file_runtime* file; \
    while((exclude = darshan_path_exclusions[tmp_index])) { \
        if(!(strncmp(exclude, __path, strlen(exclude)))) \
            break; \
        tmp_index++; \
    } \
    if(exclude) break; \
    file = posix_file_by_name(__path); \
    if(file) \
    { \
        POSIX_RECORD_STAT(file, __statbuf, __tm1, __tm2); \
    } \
} while(0)

#define POSIX_RECORD_STAT(__file, __statbuf, __tm1, __tm2) do { \
    (__file)->file_record->rank = my_rank; \
    DARSHAN_COUNTER_F_INC_NO_OVERLAP((__file)->file_record, __tm1, __tm2, (__file)->last_meta_end, POSIX_F_META_TIME); \
    DARSHAN_COUNTER_INC((__file)->file_record, POSIX_STATS, 1); \
} while(0)

405 406 407 408
/**********************************************************
 *      Wrappers for POSIX I/O functions of interest      * 
 **********************************************************/

409
int DARSHAN_DECL(open)(const char *path, int flags, ...)
410 411 412 413 414
{
    int mode = 0;
    int ret;
    double tm1, tm2;

415 416
    MAP_OR_FAIL(open);

417
    if(flags & O_CREAT) 
418 419 420 421 422 423
    {
        va_list arg;
        va_start(arg, flags);
        mode = va_arg(arg, int);
        va_end(arg);

424
        tm1 = darshan_core_wtime();
425
        ret = __real_open(path, flags, mode);
426
        tm2 = darshan_core_wtime();
427 428 429
    }
    else
    {
430
        tm1 = darshan_core_wtime();
431
        ret = __real_open(path, flags);
432
        tm2 = darshan_core_wtime();
433 434
    }

435 436
    POSIX_LOCK();
    posix_runtime_initialize();
437
    POSIX_RECORD_OPEN(ret, path, mode, 0, tm1, tm2);
438
    POSIX_UNLOCK();
439 440 441 442

    return(ret);
}

Shane Snyder's avatar
Shane Snyder committed
443 444 445 446 447 448
int DARSHAN_DECL(open64)(const char *path, int flags, ...)
{
    int mode = 0;
    int ret;
    double tm1, tm2;

449
    MAP_OR_FAIL(open64);
Shane Snyder's avatar
Shane Snyder committed
450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470

    if(flags & O_CREAT)
    {
        va_list arg;
        va_start(arg, flags);
        mode = va_arg(arg, int);
        va_end(arg);

        tm1 = darshan_core_wtime();
        ret = __real_open64(path, flags, mode);
        tm2 = darshan_core_wtime();
    }
    else
    {
        tm1 = darshan_core_wtime();
        ret = __real_open64(path, flags);
        tm2 = darshan_core_wtime();
    }

    POSIX_LOCK();
    posix_runtime_initialize();
471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501
    POSIX_RECORD_OPEN(ret, path, mode, 0, tm1, tm2);
    POSIX_UNLOCK();

    return(ret);
}

int DARSHAN_DECL(creat)(const char* path, mode_t mode)
{
    int ret;
    double tm1, tm2;

    MAP_OR_FAIL(creat);

    tm1 = darshan_core_wtime();
    ret = __real_creat(path, mode);
    tm2 = darshan_core_wtime();

    POSIX_LOCK();
    posix_runtime_initialize();
    POSIX_RECORD_OPEN(ret, path, mode, 0, tm1, tm2);
    POSIX_UNLOCK();

    return(ret);
}

int DARSHAN_DECL(creat64)(const char* path, mode_t mode)
{
    int ret;
    double tm1, tm2;

    MAP_OR_FAIL(creat64);
Shane Snyder's avatar
Shane Snyder committed
502

503 504 505 506 507 508
    tm1 = darshan_core_wtime();
    ret = __real_creat64(path, mode);
    tm2 = darshan_core_wtime();

    POSIX_LOCK();
    posix_runtime_initialize();
Shane Snyder's avatar
Shane Snyder committed
509 510 511 512 513 514
    POSIX_RECORD_OPEN(ret, path, mode, 0, tm1, tm2);
    POSIX_UNLOCK();

    return(ret);
}

515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564
FILE* DARSHAN_DECL(fopen)(const char *path, const char *mode)
{
    FILE* ret;
    int fd;
    double tm1, tm2;

    MAP_OR_FAIL(fopen);

    tm1 = darshan_core_wtime();
    ret = __real_fopen(path, mode);
    tm2 = darshan_core_wtime();

    if(ret == NULL)
        fd = -1;
    else
        fd = fileno(ret);

    POSIX_LOCK();
    posix_runtime_initialize();
    POSIX_RECORD_OPEN(fd, path, 0, 1, tm1, tm2);
    POSIX_UNLOCK();

    return(ret);
}

FILE* DARSHAN_DECL(fopen64)(const char *path, const char *mode)
{
    FILE* ret;
    int fd;
    double tm1, tm2;

    MAP_OR_FAIL(fopen64);

    tm1 = darshan_core_wtime();
    ret = __real_fopen64(path, mode);
    tm2 = darshan_core_wtime();

    if(ret == NULL)
        fd = -1;
    else
        fd = fileno(ret);

    POSIX_LOCK();
    posix_runtime_initialize();
    POSIX_RECORD_OPEN(fd, path, 0, 1, tm1, tm2);
    POSIX_UNLOCK();

    return(ret);
}

565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640
int DARSHAN_DECL(mkstemp)(char* template)
{
    int ret;
    double tm1, tm2;

    MAP_OR_FAIL(mkstemp);

    tm1 = darshan_core_wtime();
    ret = __real_mkstemp(template);
    tm2 = darshan_core_wtime();

    POSIX_LOCK();
    posix_runtime_initialize();
    POSIX_RECORD_OPEN(ret, template, 0, 0, tm1, tm2);
    POSIX_UNLOCK();

    return(ret);
}

int DARSHAN_DECL(mkostemp)(char* template, int flags)
{
    int ret;
    double tm1, tm2;

    MAP_OR_FAIL(mkostemp);

    tm1 = darshan_core_wtime();
    ret = __real_mkostemp(template, flags);
    tm2 = darshan_core_wtime();

    POSIX_LOCK();
    posix_runtime_initialize();
    POSIX_RECORD_OPEN(ret, template, 0, 0, tm1, tm2);
    POSIX_UNLOCK();

    return(ret);
}

int DARSHAN_DECL(mkstemps)(char* template, int suffixlen)
{
    int ret;
    double tm1, tm2;

    MAP_OR_FAIL(mkstemps);

    tm1 = darshan_core_wtime();
    ret = __real_mkstemps(template, suffixlen);
    tm2 = darshan_core_wtime();

    POSIX_LOCK();
    posix_runtime_initialize();
    POSIX_RECORD_OPEN(ret, template, 0, 0, tm1, tm2);
    POSIX_UNLOCK();

    return(ret);
}

int DARSHAN_DECL(mkostemps)(char* template, int suffixlen, int flags)
{
    int ret;
    double tm1, tm2;

    MAP_OR_FAIL(mkostemps);

    tm1 = darshan_core_wtime();
    ret = __real_mkostemps(template, suffixlen, flags);
    tm2 = darshan_core_wtime();

    POSIX_LOCK();
    posix_runtime_initialize();
    POSIX_RECORD_OPEN(ret, template, 0, 0, tm1, tm2);
    POSIX_UNLOCK();

    return(ret);
}

641 642 643 644 645 646
ssize_t DARSHAN_DECL(read)(int fd, void *buf, size_t count)
{
    ssize_t ret;
    int aligned_flag = 0;
    double tm1, tm2;

Shane Snyder's avatar
Shane Snyder committed
647
    MAP_OR_FAIL(read);
648

649
    if((unsigned long)buf % darshan_mem_alignment == 0) aligned_flag = 1;
650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670

    tm1 = darshan_core_wtime();
    ret = __real_read(fd, buf, count);
    tm2 = darshan_core_wtime();

    POSIX_LOCK();
    posix_runtime_initialize();
    POSIX_RECORD_READ(ret, fd, 0, 0, aligned_flag, 0, tm1, tm2);
    POSIX_UNLOCK();

    return(ret);
}

ssize_t DARSHAN_DECL(write)(int fd, const void *buf, size_t count)
{
    ssize_t ret;
    int aligned_flag = 0;
    double tm1, tm2;

    MAP_OR_FAIL(write);

671
    if((unsigned long)buf % darshan_mem_alignment == 0) aligned_flag = 1;
672 673 674 675 676 677 678 679 680 681 682 683 684 685 686 687 688 689 690 691 692

    tm1 = darshan_core_wtime();
    ret = __real_write(fd, buf, count);
    tm2 = darshan_core_wtime();

    POSIX_LOCK();
    posix_runtime_initialize();
    POSIX_RECORD_WRITE(ret, fd, 0, 0, aligned_flag, 0, tm1, tm2);
    POSIX_UNLOCK();

    return(ret);
}

ssize_t DARSHAN_DECL(pread)(int fd, void *buf, size_t count, off_t offset)
{
    ssize_t ret;
    int aligned_flag = 0;
    double tm1, tm2;

    MAP_OR_FAIL(pread);

693
    if((unsigned long)buf % darshan_mem_alignment == 0) aligned_flag = 1;
694 695 696 697 698 699 700 701 702 703 704 705 706 707 708 709 710 711 712 713 714

    tm1 = darshan_core_wtime();
    ret = __real_pread(fd, buf, count, offset);
    tm2 = darshan_core_wtime();

    POSIX_LOCK();
    posix_runtime_initialize();
    POSIX_RECORD_READ(ret, fd, 1, offset, aligned_flag, 0, tm1, tm2);
    POSIX_UNLOCK();

    return(ret);
}

ssize_t DARSHAN_DECL(pwrite)(int fd, const void *buf, size_t count, off_t offset)
{
    ssize_t ret;
    int aligned_flag = 0;
    double tm1, tm2;

    MAP_OR_FAIL(pwrite);

715
    if((unsigned long)buf % darshan_mem_alignment == 0) aligned_flag = 1;
716 717 718 719 720 721 722 723 724 725 726 727 728 729 730 731 732 733 734 735 736

    tm1 = darshan_core_wtime();
    ret = __real_pwrite(fd, buf, count, offset);
    tm2 = darshan_core_wtime();

    POSIX_LOCK();
    posix_runtime_initialize();
    POSIX_RECORD_WRITE(ret, fd, 1, offset, aligned_flag, 0, tm1, tm2);
    POSIX_UNLOCK();

    return(ret);
}

ssize_t DARSHAN_DECL(pread64)(int fd, void *buf, size_t count, off64_t offset)
{
    ssize_t ret;
    int aligned_flag = 0;
    double tm1, tm2;

    MAP_OR_FAIL(pread64);

737
    if((unsigned long)buf % darshan_mem_alignment == 0) aligned_flag = 1;
738 739 740 741 742 743 744 745 746 747 748 749 750 751 752 753 754 755 756 757 758

    tm1 = darshan_core_wtime();
    ret = __real_pread64(fd, buf, count, offset);
    tm2 = darshan_core_wtime();

    POSIX_LOCK();
    posix_runtime_initialize();
    POSIX_RECORD_READ(ret, fd, 1, offset, aligned_flag, 0, tm1, tm2);
    POSIX_UNLOCK();

    return(ret);
}

ssize_t DARSHAN_DECL(pwrite64)(int fd, const void *buf, size_t count, off64_t offset)
{
    ssize_t ret;
    int aligned_flag = 0;
    double tm1, tm2;

    MAP_OR_FAIL(pwrite64);

759
    if((unsigned long)buf % darshan_mem_alignment == 0) aligned_flag = 1;
760 761 762 763 764 765 766 767 768 769 770 771 772 773 774 775

    tm1 = darshan_core_wtime();
    ret = __real_pwrite64(fd, buf, count, offset);
    tm2 = darshan_core_wtime();

    POSIX_LOCK();
    posix_runtime_initialize();
    POSIX_RECORD_WRITE(ret, fd, 1, offset, aligned_flag, 0, tm1, tm2);
    POSIX_UNLOCK();

    return(ret);
}

ssize_t DARSHAN_DECL(readv)(int fd, const struct iovec *iov, int iovcnt)
{
    ssize_t ret;
776
    int aligned_flag = 1;
777 778 779 780
    int i;
    double tm1, tm2;

    MAP_OR_FAIL(readv);
781

782 783 784 785
    for(i=0; i<iovcnt; i++)
    {
        if(((unsigned long)iov[i].iov_base % darshan_mem_alignment) != 0)
            aligned_flag = 0;
786
    }
787 788 789 790 791 792

    tm1 = darshan_core_wtime();
    ret = __real_readv(fd, iov, iovcnt);
    tm2 = darshan_core_wtime();

    POSIX_LOCK();
793
    posix_runtime_initialize();
794 795 796 797 798 799 800 801 802
    POSIX_RECORD_READ(ret, fd, 0, 0, aligned_flag, 0, tm1, tm2);
    POSIX_UNLOCK();

    return(ret);
}

ssize_t DARSHAN_DECL(writev)(int fd, const struct iovec *iov, int iovcnt)
{
    ssize_t ret;
803
    int aligned_flag = 1;
804 805 806 807
    int i;
    double tm1, tm2;

    MAP_OR_FAIL(writev);
808

809 810 811 812
    for(i=0; i<iovcnt; i++)
    {
        if(((unsigned long)iov[i].iov_base % darshan_mem_alignment) != 0)
            aligned_flag = 0;
813
    }
814 815 816 817 818 819

    tm1 = darshan_core_wtime();
    ret = __real_writev(fd, iov, iovcnt);
    tm2 = darshan_core_wtime();

    POSIX_LOCK();
820
    posix_runtime_initialize();
821 822 823 824 825 826
    POSIX_RECORD_WRITE(ret, fd, 0, 0, aligned_flag, 0, tm1, tm2);
    POSIX_UNLOCK();

    return(ret);
}

827 828 829 830 831 832 833 834
size_t DARSHAN_DECL(fread)(void *ptr, size_t size, size_t nmemb, FILE *stream)
{
    size_t ret;
    int aligned_flag = 0;
    double tm1, tm2;

    MAP_OR_FAIL(fread);

835
    if((unsigned long)ptr % darshan_mem_alignment == 0) aligned_flag = 1;
836 837 838 839 840 841 842 843 844 845 846 847 848 849 850 851 852 853 854 855 856 857 858 859 860 861 862 863 864 865

    tm1 = darshan_core_wtime();
    ret = __real_fread(ptr, size, nmemb, stream);
    tm2 = darshan_core_wtime();

    POSIX_LOCK();
    posix_runtime_initialize();
    if(ret > 0)
    {
        POSIX_RECORD_READ(size*ret, fileno(stream), 0, 0,
            aligned_flag, 1, tm1, tm2);
    }
    else
    {
        POSIX_RECORD_READ(ret, fileno(stream), 0, 0,
            aligned_flag, 1, tm1, tm2);
    }
    POSIX_UNLOCK();

    return(ret);
}

size_t DARSHAN_DECL(fwrite)(const void *ptr, size_t size, size_t nmemb, FILE *stream)
{
    size_t ret;
    int aligned_flag = 0;
    double tm1, tm2;

    MAP_OR_FAIL(fwrite);

866
    if((unsigned long)ptr % darshan_mem_alignment == 0) aligned_flag = 1;
867 868 869 870 871 872 873 874 875 876 877 878 879 880 881 882 883 884 885 886 887 888 889 890 891 892 893 894 895 896 897 898 899 900 901 902 903 904 905 906 907 908 909 910 911 912 913 914 915 916 917 918 919 920 921 922 923 924 925 926 927 928 929 930 931 932 933 934 935 936 937 938 939 940 941 942 943 944 945 946 947 948 949 950 951 952 953 954 955 956 957 958 959 960 961 962 963 964 965 966 967

    tm1 = darshan_core_wtime();
    ret = __real_fwrite(ptr, size, nmemb, stream);
    tm2 = darshan_core_wtime();

    POSIX_LOCK();
    posix_runtime_initialize();
    if(ret > 0)
    {
        POSIX_RECORD_WRITE(size*ret, fileno(stream), 0, 0,
            aligned_flag, 1, tm1, tm2);
    }
    else
    {
        POSIX_RECORD_WRITE(ret, fileno(stream), 0, 0,
            aligned_flag, 1, tm1, tm2);
    }
    POSIX_UNLOCK();

    return(ret);
}

off_t DARSHAN_DECL(lseek)(int fd, off_t offset, int whence)
{
    off_t ret;
    struct posix_file_runtime* file;
    double tm1, tm2;

    MAP_OR_FAIL(lseek);

    tm1 = darshan_core_wtime();
    ret = __real_lseek(fd, offset, whence);
    tm2 = darshan_core_wtime();

    if(ret >= 0)
    {
        POSIX_LOCK();
        posix_runtime_initialize();
        file = posix_file_by_fd(fd);
        if(file)
        {
            file->offset = ret;
            DARSHAN_COUNTER_F_INC_NO_OVERLAP(file->file_record,
                tm1, tm2, file->last_meta_end, POSIX_F_META_TIME);
            DARSHAN_COUNTER_INC(file->file_record, POSIX_SEEKS, 1);
        }
        POSIX_UNLOCK();
    }

    return(ret);
}

off_t DARSHAN_DECL(lseek64)(int fd, off_t offset, int whence)
{
    off_t ret;
    struct posix_file_runtime* file;
    double tm1, tm2;

    MAP_OR_FAIL(lseek64);

    tm1 = darshan_core_wtime();
    ret = __real_lseek64(fd, offset, whence);
    tm2 = darshan_core_wtime();

    if(ret >= 0)
    {
        POSIX_LOCK();
        posix_runtime_initialize();
        file = posix_file_by_fd(fd);
        if(file)
        {
            file->offset = ret;
            DARSHAN_COUNTER_F_INC_NO_OVERLAP(file->file_record,
                tm1, tm2, file->last_meta_end, POSIX_F_META_TIME);
            DARSHAN_COUNTER_INC(file->file_record, POSIX_SEEKS, 1);
        }
        POSIX_UNLOCK();
    }

    return(ret);
}

int DARSHAN_DECL(fseek)(FILE *stream, long offset, int whence)
{
    int ret;
    struct posix_file_runtime* file;
    double tm1, tm2;

    MAP_OR_FAIL(fseek);

    tm1 = darshan_core_wtime();
    ret = __real_fseek(stream, offset, whence);
    tm2 = darshan_core_wtime();

    if(ret >= 0)
    {
        POSIX_LOCK();
        posix_runtime_initialize();
        file = posix_file_by_fd(fileno(stream));
        if(file)
        {
968
            file->offset = ftell(stream);
969 970 971 972 973 974 975 976 977 978
            DARSHAN_COUNTER_F_INC_NO_OVERLAP(file->file_record,
                tm1, tm2, file->last_meta_end, POSIX_F_META_TIME);
            DARSHAN_COUNTER_INC(file->file_record, POSIX_FSEEKS, 1);
        }
        POSIX_UNLOCK();
    }

    return(ret);
}

979 980 981 982 983 984 985 986 987 988 989 990 991 992 993 994 995 996 997 998 999 1000 1001 1002 1003 1004 1005 1006 1007 1008 1009 1010 1011 1012 1013 1014 1015 1016 1017 1018 1019 1020 1021 1022 1023 1024 1025 1026 1027 1028 1029 1030 1031 1032 1033 1034 1035 1036 1037 1038 1039 1040 1041 1042 1043 1044 1045 1046 1047 1048 1049 1050 1051 1052 1053 1054 1055 1056 1057 1058 1059 1060 1061 1062 1063 1064 1065 1066 1067 1068 1069 1070 1071 1072 1073 1074 1075 1076 1077 1078 1079 1080 1081 1082 1083 1084 1085 1086 1087 1088 1089 1090 1091 1092 1093 1094 1095 1096 1097 1098 1099 1100 1101 1102 1103 1104 1105 1106 1107 1108 1109 1110 1111 1112 1113 1114 1115 1116 1117 1118 1119 1120
int DARSHAN_DECL(__xstat)(int vers, const char *path, struct stat *buf)
{
    int ret;
    double tm1, tm2;

    MAP_OR_FAIL(__xstat);

    tm1 = darshan_core_wtime();
    ret = __real___xstat(vers, path, buf);
    tm2 = darshan_core_wtime();

    if(ret < 0 || !S_ISREG(buf->st_mode))
        return(ret);

    POSIX_LOCK();
    posix_runtime_initialize();
    POSIX_LOOKUP_RECORD_STAT(path, buf, tm1, tm2);
    POSIX_UNLOCK();

    return(ret);
}

int DARSHAN_DECL(__xstat64)(int vers, const char *path, struct stat64 *buf)
{
    int ret;
    double tm1, tm2;

    MAP_OR_FAIL(__xstat64);

    tm1 = darshan_core_wtime();
    ret = __real___xstat64(vers, path, buf);
    tm2 = darshan_core_wtime();

    if(ret < 0 || !S_ISREG(buf->st_mode))
        return(ret);

    POSIX_LOCK();
    posix_runtime_initialize();
    POSIX_LOOKUP_RECORD_STAT(path, buf, tm1, tm2);
    POSIX_UNLOCK();

    return(ret);
}

int DARSHAN_DECL(__lxstat)(int vers, const char *path, struct stat *buf)
{
    int ret;
    double tm1, tm2;

    MAP_OR_FAIL(__lxstat);

    tm1 = darshan_core_wtime();
    ret = __real___lxstat(vers, path, buf);
    tm2 = darshan_core_wtime();

    if(ret < 0 || !S_ISREG(buf->st_mode))
        return(ret);

    POSIX_LOCK();
    posix_runtime_initialize();
    POSIX_LOOKUP_RECORD_STAT(path, buf, tm1, tm2);
    POSIX_UNLOCK();

    return(ret);
}

int DARSHAN_DECL(__lxstat64)(int vers, const char *path, struct stat64 *buf)
{
    int ret;
    double tm1, tm2;

    MAP_OR_FAIL(__lxstat64);

    tm1 = darshan_core_wtime();
    ret = __real___lxstat64(vers, path, buf);
    tm2 = darshan_core_wtime();

    if(ret < 0 || !S_ISREG(buf->st_mode))
        return(ret);

    POSIX_LOCK();
    posix_runtime_initialize();
    POSIX_LOOKUP_RECORD_STAT(path, buf, tm1, tm2);
    POSIX_UNLOCK();

    return(ret);
}

int DARSHAN_DECL(__fxstat)(int vers, int fd, struct stat *buf)
{
    int ret;
    struct posix_file_runtime* file;
    double tm1, tm2;

    MAP_OR_FAIL(__fxstat);

    tm1 = darshan_core_wtime();
    ret = __real___fxstat(vers, fd, buf);
    tm2 = darshan_core_wtime();

    if(ret < 0 || !S_ISREG(buf->st_mode))
        return(ret);

    POSIX_LOCK();
    posix_runtime_initialize();
    file = posix_file_by_fd(fd);
    if(file)
    {
        POSIX_RECORD_STAT(file, buf, tm1, tm2);
    }
    POSIX_UNLOCK();

    return(ret);
}

int DARSHAN_DECL(__fxstat64)(int vers, int fd, struct stat64 *buf)
{
    int ret;
    struct posix_file_runtime* file;
    double tm1, tm2;

    MAP_OR_FAIL(__fxstat64);

    tm1 = darshan_core_wtime();
    ret = __real___fxstat64(vers, fd, buf);
    tm2 = darshan_core_wtime();

    if(ret < 0 || !S_ISREG(buf->st_mode))
        return(ret);

    POSIX_LOCK();
    posix_runtime_initialize();
    file = posix_file_by_fd(fd);
    if(file)
    {
        POSIX_RECORD_STAT(file, buf, tm1, tm2);
    }
    POSIX_UNLOCK();

    return(ret);
}

Shane Snyder's avatar
Shane Snyder committed
1121 1122 1123 1124 1125 1126 1127 1128 1129 1130 1131 1132 1133 1134 1135 1136 1137 1138 1139 1140 1141 1142 1143 1144 1145 1146 1147 1148 1149 1150 1151 1152 1153 1154 1155 1156 1157 1158 1159 1160 1161 1162 1163 1164 1165 1166 1167 1168
void* DARSHAN_DECL(mmap)(void *addr, size_t length, int prot, int flags,
    int fd, off_t offset)
{
    void* ret;
    struct posix_file_runtime* file;

    MAP_OR_FAIL(mmap);

    ret = __real_mmap(addr, length, prot, flags, fd, offset);
    if(ret == MAP_FAILED)
        return(ret);

    POSIX_LOCK();
    posix_runtime_initialize();
    file = posix_file_by_fd(fd);
    if(file)
    {
        DARSHAN_COUNTER_INC(file->file_record, POSIX_MMAPS, 1);
    }
    POSIX_UNLOCK();

    return(ret);
}

void* DARSHAN_DECL(mmap64)(void *addr, size_t length, int prot, int flags,
    int fd, off64_t offset)
{
    void* ret;
    struct posix_file_runtime* file;

    MAP_OR_FAIL(mmap64);

    ret = __real_mmap64(addr, length, prot, flags, fd, offset);
    if(ret == MAP_FAILED)
        return(ret);

    POSIX_LOCK();
    posix_runtime_initialize();
    file = posix_file_by_fd(fd);
    if(file)
    {
        DARSHAN_COUNTER_INC(file->file_record, POSIX_MMAPS, 1);
    }
    POSIX_UNLOCK();

    return(ret);
}

1169 1170 1171 1172 1173 1174 1175 1176 1177 1178 1179 1180 1181 1182 1183 1184 1185 1186 1187 1188 1189 1190 1191 1192 1193 1194 1195 1196 1197 1198 1199 1200 1201 1202 1203 1204 1205 1206 1207 1208 1209 1210 1211 1212 1213 1214 1215 1216 1217 1218 1219 1220 1221 1222 1223 1224 1225 1226
int DARSHAN_DECL(fsync)(int fd)
{
    int ret;
    struct posix_file_runtime* file;
    double tm1, tm2;

    MAP_OR_FAIL(fsync);

    tm1 = darshan_core_wtime();
    ret = __real_fsync(fd);
    tm2 = darshan_core_wtime();

    if(ret < 0)
        return(ret);

    POSIX_LOCK();
    posix_runtime_initialize();
    file = posix_file_by_fd(fd);
    if(file)
    {
        DARSHAN_COUNTER_F_INC_NO_OVERLAP(file->file_record,
            tm1, tm2, file->last_write_end, POSIX_F_WRITE_TIME);
        DARSHAN_COUNTER_INC(file->file_record, POSIX_FSYNCS, 1);
    }
    POSIX_UNLOCK();

    return(ret);
}

int DARSHAN_DECL(fdatasync)(int fd)
{
    int ret;
    struct posix_file_runtime* file;
    double tm1, tm2;

    MAP_OR_FAIL(fdatasync);

    tm1 = darshan_core_wtime();
    ret = __real_fdatasync(fd);
    tm2 = darshan_core_wtime();

    if(ret < 0)
        return(ret);

    POSIX_LOCK();
    posix_runtime_initialize();
    file = posix_file_by_fd(fd);
    if(file)
    {
        DARSHAN_COUNTER_F_INC_NO_OVERLAP(file->file_record,
            tm1, tm2, file->last_write_end, POSIX_F_WRITE_TIME); 
        DARSHAN_COUNTER_INC(file->file_record, POSIX_FDSYNCS, 1);
    }
    POSIX_UNLOCK();

    return(ret);
}

1227 1228
int DARSHAN_DECL(close)(int fd)
{
1229
    struct posix_file_runtime* file;
1230 1231 1232 1233 1234 1235 1236 1237 1238 1239 1240
    double tm1, tm2;
    int ret;

    MAP_OR_FAIL(close);

    tm1 = darshan_core_wtime();
    ret = __real_close(fd);
    tm2 = darshan_core_wtime();

    POSIX_LOCK();
    posix_runtime_initialize();
Shane Snyder's avatar
Shane Snyder committed
1241 1242 1243
    file = posix_file_by_fd(fd);
    if(file)
    {
1244 1245 1246 1247 1248 1249
        file->last_byte_written = 0;
        file->last_byte_read = 0;
        DARSHAN_COUNTER_F_SET(file->file_record,
            POSIX_F_CLOSE_TIMESTAMP, darshan_core_wtime());
        DARSHAN_COUNTER_F_INC_NO_OVERLAP(file->file_record,
             tm1, tm2, file->last_meta_end, POSIX_F_META_TIME);
Shane Snyder's avatar
Shane Snyder committed
1250 1251
        posix_file_close_fd(fd);
    }
1252 1253 1254 1255 1256
    POSIX_UNLOCK();    

    return(ret);
}

1257 1258 1259 1260 1261 1262 1263 1264 1265 1266 1267 1268 1269 1270 1271 1272 1273 1274 1275 1276 1277 1278 1279 1280 1281 1282 1283 1284 1285 1286 1287
int DARSHAN_DECL(fclose)(FILE *fp)
{
    struct posix_file_runtime* file;
    int fd = fileno(fp);
    double tm1, tm2;
    int ret;

    MAP_OR_FAIL(fclose);

    tm1 = darshan_core_wtime();
    ret = __real_fclose(fp);
    tm2 = darshan_core_wtime();

    POSIX_LOCK();
    posix_runtime_initialize();
    file = posix_file_by_fd(fd);
    if(file)
    {
        file->last_byte_written = 0;
        file->last_byte_read = 0;
        DARSHAN_COUNTER_F_SET(file->file_record,
            POSIX_F_CLOSE_TIMESTAMP, darshan_core_wtime());
        DARSHAN_COUNTER_F_INC_NO_OVERLAP(file->file_record,
            tm1, tm2, file->last_meta_end, POSIX_F_META_TIME);
        posix_file_close_fd(fd);
    }
    POSIX_UNLOCK();

    return(ret);
}

1288 1289 1290
/**********************************************************
 * Internal functions for manipulating POSIX module state *
 **********************************************************/
1291

1292
/* initialize internal POSIX module data structures and register with darshan-core */
1293
static void posix_runtime_initialize()
1294
{
1295 1296 1297
    int mem_limit;
    struct darshan_module_funcs posix_mod_fns =
    {
Shane Snyder's avatar
Shane Snyder committed
1298
        .disable_instrumentation = &posix_disable_instrumentation,
1299
        .prepare_for_reduction = &posix_prepare_for_reduction,
1300
        .record_reduction_op = &posix_record_reduction_op,
1301
        .get_output_data = &posix_get_output_data,
1302
        .shutdown = &posix_shutdown
1303
    };
1304

Shane Snyder's avatar
Shane Snyder committed
1305 1306
    /* don't do anything if already initialized or instrumenation is disabled */
    if(posix_runtime || instrumentation_disabled)
1307
        return;
1308

1309 1310
    /* register the posix module with darshan core */
    darshan_core_register_module(
1311
        DARSHAN_POSIX_MOD,
1312
        &posix_mod_fns,
1313 1314
        &mem_limit,
        &darshan_mem_alignment);
1315 1316 1317 1318 1319 1320 1321 1322 1323 1324 1325 1326

    /* return if no memory assigned by darshan core */
    if(mem_limit == 0)
        return;

    posix_runtime = malloc(sizeof(*posix_runtime));
    if(!posix_runtime)
        return;
    memset(posix_runtime, 0, sizeof(*posix_runtime));

    /* set maximum number of file records according to max memory limit */
    /* NOTE: maximum number of records is based on the size of a posix file record */
1327
    /* TODO: should we base memory usage off file record or total runtime structure sizes? */
1328
    posix_runtime->file_array_size = mem_limit / sizeof(struct darshan_posix_file);
1329
    posix_runtime->file_array_ndx = 0;
1330 1331

    /* allocate array of runtime file records */
1332
    posix_runtime->file_runtime_array = malloc(posix_runtime->file_array_size *
1333
                                               sizeof(struct posix_file_runtime));
1334 1335 1336
    posix_runtime->file_record_array = malloc(posix_runtime->file_array_size *
                                              sizeof(struct darshan_posix_file));
    if(!posix_runtime->file_runtime_array || !posix_runtime->file_record_array)
1337 1338 1339 1340
    {
        posix_runtime->file_array_size = 0;
        return;
    }
1341
    memset(posix_runtime->file_runtime_array, 0, posix_runtime->file_array_size *
1342
           sizeof(struct posix_file_runtime));
1343 1344
    memset(posix_runtime->file_record_array, 0, posix_runtime->file_array_size *
           sizeof(struct darshan_posix_file));
1345

Shane Snyder's avatar
Shane Snyder committed
1346 1347
    /* TODO: can we move this out of here? perhaps register_module returns rank? */
    DARSHAN_MPI_CALL(PMPI_Comm_rank)(MPI_COMM_WORLD, &my_rank);
1348

1349 1350
    return;
}
1351

1352
/* get a POSIX file record for the given file path */
1353
static struct posix_file_runtime* posix_file_by_name(const char *name)
1354
{
1355
    struct posix_file_runtime *file = NULL;
1356
    char *newname = NULL;
1357
    darshan_record_id file_id;
1358
    int file_alignment;
1359

Shane Snyder's avatar
Shane Snyder committed
1360
    if(!posix_runtime || instrumentation_disabled)
1361
        return(NULL);
1362

1363
    newname = darshan_clean_file_path(name);
1364
    if(!newname)
1365
        newname = (char*)name;
1366

1367
    /* get a unique id for this file from darshan core */
Shane Snyder's avatar
Shane Snyder committed
1368
    darshan_core_register_record(
1369 1370 1371
        (void*)newname,
        strlen(newname),
        1,
1372
        DARSHAN_POSIX_MOD,
1373 1374
        &file_id,
        &file_alignment);
1375

1376
    /* search the hash table for this file record, and return if found */
1377
    HASH_FIND(hlink, posix_runtime->file_hash, &file_id, sizeof(darshan_record_id), file);
1378
    if(file)
1379
    {
1380
        if(newname != name)
1381
            free(newname);
1382
        return(file);
1383 1384
    }

1385 1386 1387 1388 1389 1390
    if(posix_runtime->file_array_ndx < posix_runtime->file_array_size);
    {
        /* no existing record, assign a new file record from the global array */
        file = &(posix_runtime->file_runtime_array[posix_runtime->file_array_ndx]);
        file->file_record = &(posix_runtime->file_record_array[posix_runtime->file_array_ndx]);
        file->file_record->f_id = file_id;
1391 1392
        DARSHAN_COUNTER_SET(file->file_record, POSIX_MEM_ALIGNMENT, darshan_mem_alignment);
        DARSHAN_COUNTER_SET(file->file_record, POSIX_FILE_ALIGNMENT, file_alignment);
1393

1394
        /* add new record to file hash table */
1395
        HASH_ADD(hlink, posix_runtime->file_hash, file_record->f_id, sizeof(darshan_record_id), file);
1396

1397 1398
        posix_runtime->file_array_ndx++;
    }
1399 1400

    if(newname != name)
1401
        free(newname);
1402 1403 1404
    return(file);
}

1405 1406 1407
/* get a POSIX file record for the given file path, and also create a
 * reference structure using the returned file descriptor
 */
1408
static struct posix_file_runtime* posix_file_by_name_setfd(const char* name, int fd)
1409
{
1410 1411
    struct posix_file_runtime* file;
    struct posix_file_runtime_ref* ref;
1412

Shane Snyder's avatar
Shane Snyder committed
1413
    if(!posix_runtime || instrumentation_disabled)
1414 1415 1416 1417 1418 1419 1420 1421 1422 1423 1424 1425 1426 1427 1428 1429 1430 1431 1432 1433 1434 1435 1436 1437 1438 1439
        return(NULL);

    /* find file record by name first */
    file = posix_file_by_name(name);

    if(!file)
        return(NULL);

    /* search hash table for existing file ref for this fd */
    HASH_FIND(hlink, posix_runtime->fd_hash, &fd, sizeof(int), ref);
    if(ref)
    {
        /* we have a reference.  Make sure it points to the correct file
         * and return it
         */
        ref->file = file;
        return(file);
    }

    /* if we hit this point, then we don't have a reference for this fd
     * in the table yet.  Add it.
     */
    ref = malloc(sizeof(*ref));
    if(!ref)
        return(NULL);
    memset(ref, 0, sizeof(*ref));
Shane Snyder's avatar
Shane Snyder committed
1440

1441 1442 1443 1444 1445 1446 1447
    ref->file = file;
    ref->fd = fd;    
    HASH_ADD(hlink, posix_runtime->fd_hash, fd, sizeof(int), ref);

    return(file);
}

1448
/* get a POSIX file record for the given file descriptor */
1449
static struct posix_file_runtime* posix_file_by_fd(int fd)
Shane Snyder's avatar
Shane Snyder committed
1450
{
1451
    struct posix_file_runtime_ref* ref;
Shane Snyder's avatar
Shane Snyder committed
1452 1453 1454 1455 1456 1457 1458 1459 1460 1461 1462 1463

    if(!posix_runtime || instrumentation_disabled)
        return(NULL);

    /* search hash table for existing file ref for this fd */
    HASH_FIND(hlink, posix_runtime->fd_hash, &fd, sizeof(int), ref);
    if(ref)
        return(ref->file);

    return(NULL);
}

1464
/* free up reference data structures for the given file descriptor */
1465 1466
static void posix_file_close_fd(int fd)
{
1467
    struct posix_file_runtime_ref* ref;
1468

Shane Snyder's avatar
Shane Snyder committed
1469
    if(!posix_runtime || instrumentation_disabled)
1470 1471 1472 1473
        return;

    /* search hash table for this fd */
    HASH_FIND(hlink, posix_runtime->fd_hash, &fd, sizeof(int), ref);
1474
    if(ref)
1475 1476 1477 1478 1479 1480 1481
    {
        /* we have a reference, delete it */
        HASH_DELETE(hlink, posix_runtime->fd_hash, ref);
        free(ref);
    }

    return;
1482 1483
}

1484 1485 1486 1487 1488 1489 1490 1491 1492 1493 1494 1495 1496 1497 1498 1499 1500 1501 1502 1503 1504 1505 1506 1507 1508 1509 1510 1511 1512 1513 1514 1515 1516 1517 1518 1519 1520 1521 1522 1523 1524 1525 1526 1527 1528 1529 1530 1531 1532 1533 1534 1535 1536 1537 1538 1539 1540 1541 1542 1543 1544 1545 1546 1547 1548 1549 1550 1551 1552 1553 1554 1555 1556 1557 1558 1559 1560 1561 1562 1563 1564 1565 1566 1567 1568 1569 1570 1571 1572 1573 1574 1575 1576 1577 1578 1579 1580 1581 1582 1583 1584 1585 1586 1587 1588 1589 1590 1591 1592 1593 1594 1595 1596 1597 1598 1599 1600 1601 1602 1603 1604 1605 1606 1607 1608 1609 1610 1611
static void posix_access_counter(struct posix_file_runtime* file, ssize_t size,
    enum posix_counter_type type)
{
    struct posix_access_counter* counter;
    struct posix_access_counter* found;
    void* tmp;
    void** root;
    int* count;
    struct posix_access_counter tmp_counter;

    /* don't count sizes or strides of 0 */
    if(size == 0)
        return;

    switch(type)
    {
        case POSIX_COUNTER_ACCESS:
            root = &file->access_root;
            count = &file->access_count;
            break;
        case POSIX_COUNTER_STRIDE:
            root = &file->stride_root;
            count = &file->stride_count;
            break;
        default:
            return;
    }

    /* check to see if this size is already recorded */
    tmp_counter.size = size;
    tmp_counter.freq = 1;
    tmp = tfind(&tmp_counter, root, posix_access_compare);
    if(tmp)
    {
        found = *(struct posix_access_counter**)tmp;
        found->freq++;
        return;
    }

    /* we can add a new one as long as we haven't hit the limit */
    if(*count < POSIX_MAX_ACCESS_COUNT_RUNTIME)
    {
        counter = malloc(sizeof(*counter));
        if(!counter)
        {
            return;
        }

        counter->size = size;
        counter->freq = 1;

        tmp = tsearch(counter, root, posix_access_compare);
        found = *(struct posix_access_counter**)tmp;
        /* if we get a new answer out here we are in trouble; this was
         * already checked with the tfind()
         */
        assert(found == counter);

        (*count)++;
    }

    return;
}

static void posix_access_walker(const void* nodep, const VISIT which, const int depth)
{
    struct posix_access_counter* counter;

    switch (which)
    {
        case postorder:
        case leaf:
            counter = *(struct posix_access_counter**)nodep;
            POSIX_COMMON_COUNTER_INC(walker_file->file_record, counter->size, counter->freq,
                walker_validx, walker_cntidx);
        default:
            break;
    }

    return;
};

/* posix_walk_file_accesses()
 *
 * goes through runtime collections of accesses sizes and chooses the 4 most
 * common for logging
 */
static void posix_walk_file_accesses()
{
    int i;

    for(i=0; i<posix_runtime->file_array_ndx; i++)
    {
        /* walk trees for both access sizes and stride sizes to pick 4 most
         * common of each
         */

        /* NOTE: setting global variables here for cp_access_walker() */
        walker_file = &posix_runtime->file_runtime_array[i];
        walker_validx = POSIX_ACCESS1_ACCESS;
        walker_cntidx = POSIX_ACCESS1_COUNT;
        twalk(walker_file->access_root,
            posix_access_walker);
        tdestroy(walker_file->access_root, free);

        walker_validx = POSIX_STRIDE1_STRIDE;
        walker_cntidx = POSIX_STRIDE1_COUNT;
        twalk(walker_file->stride_root,
            posix_access_walker);
        tdestroy(walker_file->stride_root, free);
    }

    return;
}

static int posix_access_compare(const void* a_p, const void* b_p)
{
    const struct posix_access_counter* a = a_p;
    const struct posix_access_counter* b = b_p;

    if(a->size < b->size)
        return(-1);
    if(a->size > b->size)
        return(1);
    return(0);
}


1612
/* compare function for sorting file records by descending rank */
1613
static int posix_file_compare(const void* a_p, const void* b_p)
1614
{
1615 1616
    const struct darshan_posix_file* a = a_p;
    const struct darshan_posix_file* b = b_p;
1617

1618
    if(a->rank < b->rank)
1619
        return 1;
1620
    if(a->rank > b->rank)
1621 1622 1623 1624 1625
        return -1;

    return 0;
}

1626 1627 1628
/************************************************************************
 * Functions exported by this module for coordinating with darshan-core *
 ************************************************************************/
1629

Shane Snyder's avatar
Shane Snyder committed
1630
static void posix_disable_instrumentation()
1631
{
1632 1633
    assert(posix_runtime);

Shane Snyder's avatar
Shane Snyder committed
1634
    POSIX_LOCK();
1635 1636 1637

    posix_walk_file_accesses();

Shane Snyder's avatar
Shane Snyder committed
1638 1639
    instrumentation_disabled = 1;
    POSIX_UNLOCK();
1640

Shane Snyder's avatar
Shane Snyder committed
1641 1642 1643
    return;
}

1644 1645 1646 1647 1648 1649 1650
static void posix_prepare_for_reduction(
    darshan_record_id *shared_recs,
    int *shared_rec_count,
    void **send_buf,
    void **recv_buf,
    int *rec_size)
{
1651
    struct posix_file_runtime *file;
1652
    int i;
1653
    double posix_time;
1654 1655 1656 1657 1658 1659 1660 1661 1662 1663

    assert(posix_runtime);

    /* necessary initialization of shared records (e.g., change rank to -1) */
    for(i = 0; i < *shared_rec_count; i++)
    {
        HASH_FIND(hlink, posix_runtime->file_hash, &shared_recs[i],
            sizeof(darshan_record_id), file);
        assert(file);

1664 1665 1666 1667 1668 1669 1670 1671 1672 1673 1674 1675 1676 1677 1678 1679 1680 1681 1682 1683 1684 1685 1686 1687 1688
        posix_time =
            file->file_record->fcounters[POSIX_F_READ_TIME] +
            file->file_record->fcounters[POSIX_F_WRITE_TIME] +
            file->file_record->fcounters[POSIX_F_META_TIME];

        /* initialize fastest/slowest info prior to the reduction */
        file->file_record->counters[POSIX_FASTEST_RANK] =
            file->file_record->rank;
        file->file_record->counters[POSIX_FASTEST_RANK_BYTES] = 
            file->file_record->counters[POSIX_BYTES_READ] +
            file->file_record->counters[POSIX_BYTES_WRITTEN];
        file->file_record->fcounters[POSIX_F_FASTEST_RANK_TIME] = 
            posix_time;

        /* until reduction occurs, we assume that this rank is both
         * the fastest and slowest. It is up to the reduction operator
         * to find the true min and max.
         */
        file->file_record->counters[POSIX_SLOWEST_RANK] =
            file->file_record->counters[POSIX_FASTEST_RANK];
        file->file_record->counters[POSIX_SLOWEST_RANK_BYTES] =
            file->file_record->counters[POSIX_FASTEST_RANK_BYTES];
        file->file_record->fcounters[POSIX_F_SLOWEST_RANK_TIME] =
            file->file_record->fcounters[POSIX_F_FASTEST_RANK_TIME];