src/runtime/coreparse.c

   1 /*
   2  * A saved SBCL system is a .core file; the code here helps us accept
   3  * such a file as input.
   4  */
   5
   6 /*
   7  * This software is part of the SBCL system. See the README file for
   8  * more information.
   9  *
  10  * This software is derived from the CMU CL system, which was
  11  * written at Carnegie Mellon University and released into the
  12  * public domain. The software is in the public domain and is
  13  * provided with absolutely no warranty. See the COPYING and CREDITS
  14  * files for more information.
  15  */
  16
  17 #include <stdio.h>
  18 #include <stdlib.h>
  19 #include <string.h>
  20 #include <sys/file.h>
  21 #include <sys/types.h>
  22 #include <sys/stat.h>
  23 #include <fcntl.h>
  24 #include <unistd.h>
  25
  26 #include "sbcl.h"
  27 #include "os.h"
  28 #include "runtime.h"
  29 #include "globals.h"
  30 #include "core.h"
  31 #include "arch.h"
  32 #include "interr.h"
  33 #include "thread.h"
  34
  35 #include "validate.h"
  36 #include "gc-internal.h"
  37
  38 /* lutex stuff */
  39 #if defined(LISP_FEATURE_SB_THREAD) && defined(LISP_FEATURE_SB_LUTEX)
  40 #include "genesis/sap.h"
  41 #include "pthread-lutex.h"
  42 #endif
  43
  44
  45 unsigned char build_id[] =
  46 #include "../../output/build-id.tmp"
  47 ;
  48
  49 int
  50 open_binary(char *filename, int mode)
  51 {
  52 #ifdef LISP_FEATURE_WIN32
  53     mode |= O_BINARY;
  54 #endif
  55
  56     return open(filename, mode);
  57 }
  58
  59
  60 static struct runtime_options *
  61 read_runtime_options(int fd)
  62 {
  63     size_t optarray[RUNTIME_OPTIONS_WORDS];
  64     struct runtime_options *options = NULL;
  65
  66     if (read(fd, optarray, RUNTIME_OPTIONS_WORDS * sizeof(size_t)) !=
  67         RUNTIME_OPTIONS_WORDS * sizeof(size_t)) {
  68         return NULL;
  69     }
  70
  71     if ((RUNTIME_OPTIONS_MAGIC != optarray[0]) || (0 == optarray[1])) {
  72         return NULL;
  73     }
  74
  75     options = successful_malloc(sizeof(struct runtime_options));
  76
  77     options->dynamic_space_size = optarray[2];
  78     options->thread_control_stack_size = optarray[3];
  79
  80     return options;
  81 }
  82
  83 void
  84 maybe_initialize_runtime_options(int fd)
  85 {
  86     off_t end_offset = sizeof(lispobj) +
  87         sizeof(os_vm_offset_t) +
  88         (RUNTIME_OPTIONS_WORDS * sizeof(size_t));
  89
  90     lseek(fd, -end_offset, SEEK_END);
  91     runtime_options = read_runtime_options(fd);
  92 }
  93
  94 /* Search 'filename' for an embedded core.  An SBCL core has, at the
  95  * end of the file, a trailer containing optional saved runtime
  96  * options, the start of the core (an os_vm_offset_t), and a final
  97  * signature word (the lispobj CORE_MAGIC).  If this trailer is found
  98  * at the end of the file, the start of the core can be determined
  99  * from the core size.
 100  *
 101  * If an embedded core is present, this returns the offset into the
 102  * file to load the core from, or -1 if no core is present. */
 103 os_vm_offset_t
 104 search_for_embedded_core(char *filename)
 105 {
 106     lispobj header;
 107     os_vm_offset_t lispobj_size = sizeof(lispobj);
 108     os_vm_offset_t trailer_size = lispobj_size + sizeof(os_vm_offset_t);
 109     os_vm_offset_t core_start, pos;
 110     int fd = -1;
 111
 112     if ((fd = open_binary(filename, O_RDONLY)) < 0)
 113         goto lose;
 114     if (lseek(fd, -lispobj_size, SEEK_END) < 0)
 115         goto lose;
 116     if (read(fd, &header, (size_t)lispobj_size) < lispobj_size)
 117         goto lose;
 118
 119     if (header == CORE_MAGIC) {
 120         if (lseek(fd, -trailer_size, SEEK_END) < 0)
 121             goto lose;
 122         if (read(fd, &core_start, sizeof(os_vm_offset_t)) < 0)
 123             goto lose;
 124
 125         if (lseek(fd, core_start, SEEK_SET) < 0)
 126             goto lose;
 127         pos = lseek(fd, 0, SEEK_CUR);
 128
 129         if (read(fd, &header, (size_t)lispobj_size) < lispobj_size)
 130             goto lose;
 131
 132         if (header != CORE_MAGIC)
 133             goto lose;
 134
 135         maybe_initialize_runtime_options(fd);
 136
 137         close(fd);
 138         return pos;
 139     }
 140
 141 lose:
 142     if (fd != -1)
 143         close(fd);
 144
 145     return -1;
 146 }
 147
 148 static void
 149 process_directory(int fd, lispobj *ptr, int count, os_vm_offset_t file_offset)
 150 {
 151     struct ndir_entry *entry;
 152
 153     FSHOW((stderr, "/process_directory(..), count=%d\n", count));
 154
 155     for (entry = (struct ndir_entry *) ptr; --count>= 0; ++entry) {
 156
 157         long id = entry->identifier;
 158         long offset = os_vm_page_size * (1 + entry->data_page);
 159         os_vm_address_t addr =
 160             (os_vm_address_t) (os_vm_page_size * entry->address);
 161         lispobj *free_pointer = (lispobj *) addr + entry->nwords;
 162         unsigned long len = os_vm_page_size * entry->page_count;
 163
 164         if (len != 0) {
 165             os_vm_address_t real_addr;
 166             FSHOW((stderr, "/mapping %ld(0x%lx) bytes at 0x%lx\n",
 167                    (long)len, (long)len, (unsigned long)addr));
 168             real_addr = os_map(fd, offset + file_offset, addr, len);
 169             if (real_addr != addr) {
 170                 lose("file mapped in wrong place! "
 171                      "(0x%08x != 0x%08lx)\n",
 172                      real_addr,
 173                      addr);
 174             }
 175         }
 176
 177         FSHOW((stderr, "/space id = %ld, free pointer = 0x%lx\n",
 178                id, (unsigned long)free_pointer));
 179
 180         switch (id) {
 181         case DYNAMIC_CORE_SPACE_ID:
 182             if (len > dynamic_space_size) {
 183                 fprintf(stderr,
 184                         "dynamic space too small for core: %ldKiB required, %ldKiB available.\n",
 185                         len >> 10,
 186                         (long)dynamic_space_size >> 10);
 187                 exit(1);
 188             }
 189 #ifdef LISP_FEATURE_GENCGC
 190             if (addr != (os_vm_address_t)DYNAMIC_SPACE_START) {
 191                 fprintf(stderr, "in core: 0x%lx; in runtime: 0x%lx \n",
 192                         (long)addr, (long)DYNAMIC_SPACE_START);
 193                 lose("core/runtime address mismatch: DYNAMIC_SPACE_START\n");
 194             }
 195 #else
 196             if ((addr != (os_vm_address_t)DYNAMIC_0_SPACE_START) &&
 197                 (addr != (os_vm_address_t)DYNAMIC_1_SPACE_START)) {
 198                 fprintf(stderr, "in core: 0x%lx; in runtime: 0x%lx or 0x%lx\n",
 199                         (long)addr,
 200                         (long)DYNAMIC_0_SPACE_START,
 201                         (long)DYNAMIC_1_SPACE_START);
 202                 lose("warning: core/runtime address mismatch: DYNAMIC_SPACE_START\n");
 203             }
 204 #endif
 205 #if defined(ALLOCATION_POINTER)
 206             SetSymbolValue(ALLOCATION_POINTER, (lispobj)free_pointer,0);
 207 #else
 208             dynamic_space_free_pointer = free_pointer;
 209 #endif
 210             /* For stop-and-copy GC, this will be whatever the GC was
 211              * using at the time. With GENCGC, this will always be
 212              * space 0. (We checked above that for GENCGC,
 213              * addr==DYNAMIC_SPACE_START.) */
 214             current_dynamic_space = (lispobj *)addr;
 215             break;
 216         case STATIC_CORE_SPACE_ID:
 217             if (addr != (os_vm_address_t)STATIC_SPACE_START) {
 218                 fprintf(stderr, "in core: 0x%lx - in runtime: 0x%lx\n",
 219                         (long)addr, (long)STATIC_SPACE_START);
 220                 lose("core/runtime address mismatch: STATIC_SPACE_START\n");
 221             }
 222             break;
 223         case READ_ONLY_CORE_SPACE_ID:
 224             if (addr != (os_vm_address_t)READ_ONLY_SPACE_START) {
 225                 fprintf(stderr, "in core: 0x%lx - in runtime: 0x%lx\n",
 226                         (long)addr, (long)READ_ONLY_SPACE_START);
 227                 lose("core/runtime address mismatch: READ_ONLY_SPACE_START\n");
 228             }
 229             break;
 230         default:
 231             lose("unknown space ID %ld addr 0x%lx\n", id, (long)addr);
 232         }
 233     }
 234 }
 235
 236 lispobj
 237 load_core_file(char *file, os_vm_offset_t file_offset)
 238 {
 239     lispobj *header, val, len, *ptr, remaining_len;
 240     int fd = open_binary(file, O_RDONLY);
 241     unsigned int count;
 242
 243     lispobj initial_function = NIL;
 244     FSHOW((stderr, "/entering load_core_file(%s)\n", file));
 245     if (fd < 0) {
 246         fprintf(stderr, "could not open file \"%s\"\n", file);
 247         perror("open");
 248         exit(1);
 249     }
 250
 251     lseek(fd, file_offset, SEEK_SET);
 252     header = calloc(os_vm_page_size / sizeof(u32), sizeof(u32));
 253
 254     count = read(fd, header, os_vm_page_size);
 255     if (count < os_vm_page_size) {
 256         lose("premature end of core file\n");
 257     }
 258     SHOW("successfully read first page of core");
 259
 260     ptr = header;
 261     val = *ptr++;
 262
 263     if (val != CORE_MAGIC) {
 264         lose("invalid magic number in core: 0x%lx should have been 0x%x.\n",
 265              val,
 266              CORE_MAGIC);
 267     }
 268     SHOW("found CORE_MAGIC");
 269
 270     while (val != END_CORE_ENTRY_TYPE_CODE) {
 271         val = *ptr++;
 272         len = *ptr++;
 273         remaining_len = len - 2; /* (-2 to cancel the two ++ operations) */
 274         FSHOW((stderr, "/val=0x%ld, remaining_len=0x%ld\n",
 275                (long)val, (long)remaining_len));
 276
 277         switch (val) {
 278
 279         case END_CORE_ENTRY_TYPE_CODE:
 280             SHOW("END_CORE_ENTRY_TYPE_CODE case");
 281             break;
 282
 283         case VERSION_CORE_ENTRY_TYPE_CODE:
 284             SHOW("VERSION_CORE_ENTRY_TYPE_CODE case");
 285             if (*ptr != SBCL_CORE_VERSION_INTEGER) {
 286                 lose("core file version (%d) != runtime library version (%d)\n",
 287                      *ptr,
 288                      SBCL_CORE_VERSION_INTEGER);
 289             }
 290             break;
 291
 292         case BUILD_ID_CORE_ENTRY_TYPE_CODE:
 293             SHOW("BUILD_ID_CORE_ENTRY_TYPE_CODE case");
 294             {
 295                 unsigned int i;
 296
 297                 FSHOW((stderr, "build_id[]=\"%s\"\n", build_id));
 298                 FSHOW((stderr, "remaining_len = %d\n", remaining_len));
 299                 if (remaining_len != strlen((const char *)build_id))
 300                     goto losing_build_id;
 301                 for (i = 0; i < remaining_len; ++i) {
 302                     FSHOW((stderr, "ptr[%d] = char = %d, expected=%d\n",
 303                            i, ptr[i], build_id[i]));
 304                     if (ptr[i] != build_id[i])
 305                         goto losing_build_id;
 306                 }
 307                 break;
 308             losing_build_id:
 309                 /* .core files are not binary-compatible between
 310                  * builds because we can't easily detect whether the
 311                  * sources were patched between the time the
 312                  * dumping-the-.core runtime was built and the time
 313                  * that the loading-the-.core runtime was built.
 314                  *
 315                  * (We could easily detect whether version.lisp-expr
 316                  * was changed, but people experimenting with patches
 317                  * don't necessarily update version.lisp-expr.) */
 318
 319                 lose("can't load .core for different runtime, sorry\n");
 320             }
 321
 322         case NEW_DIRECTORY_CORE_ENTRY_TYPE_CODE:
 323             SHOW("NEW_DIRECTORY_CORE_ENTRY_TYPE_CODE case");
 324             process_directory(fd,
 325                               ptr,
 326 #ifndef LISP_FEATURE_ALPHA
 327                               remaining_len / (sizeof(struct ndir_entry) /
 328                                                sizeof(long)),
 329 #else
 330                               remaining_len / (sizeof(struct ndir_entry) /
 331                                                sizeof(u32)),
 332 #endif
 333                               file_offset);
 334             break;
 335
 336         case INITIAL_FUN_CORE_ENTRY_TYPE_CODE:
 337             SHOW("INITIAL_FUN_CORE_ENTRY_TYPE_CODE case");
 338             initial_function = (lispobj)*ptr;
 339             break;
 340
 341 #if defined(LISP_FEATURE_SB_THREAD) && defined(LISP_FEATURE_SB_LUTEX)
 342         case LUTEX_TABLE_CORE_ENTRY_TYPE_CODE:
 343             SHOW("LUTEX_TABLE_CORE_ENTRY_TYPE_CODE case");
 344             {
 345                 size_t n_lutexes = *ptr;
 346                 size_t fdoffset = (*(ptr + 1) + 1) * (os_vm_page_size);
 347                 size_t data_length = n_lutexes * sizeof(struct sap *);
 348                 struct lutex **lutexes_to_resurrect = malloc(data_length);
 349                 long bytes_read;
 350
 351                 lseek(fd, fdoffset + file_offset, SEEK_SET);
 352
 353                 FSHOW((stderr, "attempting to read %ld lutexes from core\n", n_lutexes));
 354                 bytes_read = read(fd, lutexes_to_resurrect, data_length);
 355
 356                 /* XXX */
 357                 if (bytes_read != data_length) {
 358                     lose("Could not read the lutex table");
 359                 }
 360                 else {
 361                     int i;
 362
 363                     for (i=0; i<n_lutexes; ++i) {
 364                         struct lutex *lutex = lutexes_to_resurrect[i];
 365
 366                         FSHOW((stderr, "re-init'ing lutex @ %p\n", lutex));
 367                         lutex_init((tagged_lutex_t) lutex);
 368                     }
 369
 370                     free(lutexes_to_resurrect);
 371                 }
 372                 break;
 373             }
 374 #endif
 375
 376 #ifdef LISP_FEATURE_GENCGC
 377         case PAGE_TABLE_CORE_ENTRY_TYPE_CODE:
 378         {
 379             size_t size = *ptr;
 380             size_t fdoffset = (*(ptr+1) + 1) * (os_vm_page_size);
 381             size_t offset = 0;
 382             long bytes_read;
 383             unsigned long data[4096];
 384             lseek(fd, fdoffset + file_offset, SEEK_SET);
 385             while ((bytes_read = read(fd, data, (size < 4096 ? size : 4096 )))
 386                     > 0)
 387             {
 388                 int i = 0;
 389                 size -= bytes_read;
 390                 while (bytes_read) {
 391                     bytes_read -= sizeof(long);
 392                     /* Ignore all zeroes. The size of the page table
 393                      * core entry was rounded up to os_vm_page_size
 394                      * during the save, and might now have more
 395                      * elements than the page table.
 396                      */
 397                     if (data[i]) {
 398                         page_table[offset].region_start_offset = data[i];
 399                     }
 400                     i++;
 401                     offset++;
 402                 }
 403             }
 404
 405             gencgc_partial_pickup = 1;
 406             break;
 407         }
 408 #endif
 409         default:
 410             lose("unknown core file entry: %ld\n", (long)val);
 411         }
 412
 413         ptr += remaining_len;
 414         FSHOW((stderr, "/new ptr=%lx\n", (unsigned long)ptr));
 415     }
 416     SHOW("about to free(header)");
 417     free(header);
 418     SHOW("returning from load_core_file(..)");
 419     return initial_function;
 420 }
 421