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