RmBuffer *rm_buffer_pool_get(RmBufferPool *pool) {
RmBuffer *buffer = NULL;
g_mutex_lock(&pool->lock);
{
while(!buffer) {
if(pool->stack) {
buffer = pool->stack->data;
pool->stack = g_slist_delete_link(pool->stack, pool->stack);
} else if(pool->avail_buffers > 0) {
buffer = rm_buffer_new(pool);
} else {
if(!pool->mem_warned) {
rm_log_warning_line(
"read buffer limit reached - waiting for "
"processing to catch up");
pool->mem_warned = true;
}
g_cond_wait(&pool->change, &pool->lock);
}
}
pool->avail_buffers--;
if(pool->avail_buffers < pool->min_kept_buffers) {
pool->min_kept_buffers = pool->avail_buffers;
}
}
g_mutex_unlock(&pool->lock);
rm_assert_gentle(buffer);
return buffer;
}
RmBuffer *rm_buffer_get(RmBufferPool *pool) {
RmBuffer *buffer = NULL;
g_mutex_lock(&pool->lock);
{
while(!buffer) {
buffer = rm_util_slist_pop(&pool->stack, NULL);
if (!buffer && pool->avail_buffers > 0) {
buffer = rm_buffer_new(pool);
}
if (!buffer) {
if(!pool->mem_warned) {
rm_log_warning_line(
"read buffer limit reached - waiting for "
"processing to catch up");
pool->mem_warned = true;
}
g_cond_wait(&pool->change, &pool->lock);
}
}
pool->avail_buffers--;
}
g_mutex_unlock(&pool->lock);
rm_assert_gentle(buffer);
return buffer;
}
int rm_json_cache_read(RmTrie *file_trie, const char *json_path) {
#if !HAVE_JSON_GLIB
(void)file_trie;
(void)json_path;
rm_log_info_line(_("caching is not supported due to missing json-glib library."));
return EXIT_FAILURE;
#else
rm_assert_gentle(file_trie);
rm_assert_gentle(json_path);
int result = EXIT_FAILURE;
GError *error = NULL;
size_t keys_in_table = rm_trie_size(file_trie);
JsonParser *parser = json_parser_new();
rm_log_info_line(_("Loading json-cache `%s'"), json_path);
if(!json_parser_load_from_file(parser, json_path, &error)) {
rm_log_warning_line(_("FAILED: %s\n"), error->message);
g_error_free(error);
goto failure;
}
JsonNode *root = json_parser_get_root(parser);
if(JSON_NODE_TYPE(root) != JSON_NODE_ARRAY) {
rm_log_warning_line(_("No valid json cache (no array in /)"));
goto failure;
}
/* Iterate over all objects in it */
json_array_foreach_element(json_node_get_array(root),
(JsonArrayForeach)rm_json_cache_parse_entry,
file_trie);
/* check if some entries were added */
result = (keys_in_table >= rm_trie_size(file_trie));
failure:
if(parser) {
g_object_unref(parser);
}
return result;
#endif
}
bool rm_session_was_aborted() {
gint rc = g_atomic_int_get(&SESSION_ABORTED);
static GOnce print_once = G_ONCE_INIT;
switch(rc) {
case 1:
g_once (&print_once, rm_session_print_first_abort_warn, NULL);
break;
case 2:
rm_log_warning_line(_("Received second Interrupt, stopping hard."));
exit(EXIT_FAILURE);
break;
}
return rc;
}
int rm_hasher_main(int argc, const char **argv) {
RmHasherSession tag;
/* List of paths we got passed (or NULL) */
tag.paths = NULL;
/* Print hashes in the same order as files in command line args */
tag.print_in_order = TRUE;
/* Print a hash with builtin identifier */
tag.print_multihash = FALSE;
/* Digest type (user option, default SHA1) */
tag.digest_type = RM_DIGEST_SHA1;
gint threads = 8;
gint64 buffer_mbytes = 256;
////////////// Option Parsing ///////////////
/* clang-format off */
const GOptionEntry entries[] = {
{"digest-type" , 'd' , 0 , G_OPTION_ARG_CALLBACK , (GOptionArgFunc)rm_hasher_parse_type , _("Digest type [SHA1]") , "[TYPE]"} ,
{"num-threads" , 't' , 0 , G_OPTION_ARG_INT , &threads , _("Number of hashing threads [8]") , "N"} ,
{"multihash" , 'm' , 0 , G_OPTION_ARG_NONE , &tag.print_multihash , _("Print hash as self identifying multihash") , NULL} ,
{"buffer-mbytes" , 'b' , 0 , G_OPTION_ARG_INT64 , &buffer_mbytes , _("Megabytes read buffer [256 MB]") , "MB"} ,
{"ignore-order" , 'i' , G_OPTION_FLAG_REVERSE , G_OPTION_ARG_NONE , &tag.print_in_order , _("Print hashes in order completed, not in order entered (reduces memory usage)") , NULL} ,
{"" , 0 , 0 , G_OPTION_ARG_FILENAME_ARRAY , &tag.paths , _("Space-separated list of files") , "[FILE…]"} ,
{NULL , 0 , 0 , 0 , NULL , NULL , NULL}};
/* clang-format on */
GError *error = NULL;
GOptionContext *context = g_option_context_new(_("Hash a list of files"));
GOptionGroup *main_group =
g_option_group_new(argv[0], _("Hash a list of files"), "", &tag, NULL);
char summary[4096];
memset(summary, 0, sizeof(summary));
g_snprintf(summary, sizeof(summary),
_("Multi-threaded file digest (hash) calculator.\n"
"\n Available digest types:"
"\n %s\n"
"\n Versions with different bit numbers:"
"\n %s\n"
"\n Supported, but not useful:"
"\n %s\n"),
"spooky, city, xxhash, sha{1,256,512}, md5, murmur",
"spooky{32,64,128}, city{128,256,512}, murmur{512}",
"farmhash, cumulative, paranoid, ext, bastard");
g_option_group_add_entries(main_group, entries);
g_option_context_set_main_group(context, main_group);
g_option_context_set_summary(context, summary);
if(!g_option_context_parse(context, &argc, (char ***)&argv, &error)) {
/* print g_option error message */
rm_log_error_line("%s", error->message);
exit(EXIT_FAILURE);
}
if(tag.paths == NULL) {
/* read paths from stdin */
char path_buf[PATH_MAX];
char *tokbuf = NULL;
GPtrArray *paths = g_ptr_array_new();
while(fgets(path_buf, PATH_MAX, stdin)) {
char *abs_path = realpath(strtok_r(path_buf, "\n", &tokbuf), NULL);
g_ptr_array_add(paths, abs_path);
}
tag.paths = (char **)g_ptr_array_free(paths, FALSE);
}
if(tag.paths == NULL || tag.paths[0] == NULL) {
rm_log_error_line(_("No valid paths given."));
exit(EXIT_FAILURE);
}
g_option_context_free(context);
////////// Implementation //////
if(tag.print_in_order) {
/* allocate buffer to collect results */
tag.completed_digests_buffer =
g_slice_alloc0((g_strv_length(tag.paths) + 1) * sizeof(RmDigest *));
tag.path_index = 0;
}
/* initialise structures */
g_mutex_init(&tag.lock);
RmHasher *hasher = rm_hasher_new(tag.digest_type,
threads,
FALSE,
4096,
1024 * 1024 * buffer_mbytes,
0,
(RmHasherCallback)rm_hasher_callback,
&tag);
/* Iterate over paths, pushing to hasher threads */
for(int i = 0; tag.paths && tag.paths[i]; ++i) {
/* check it is a regular file */
RmStat stat_buf;
if(rm_sys_stat(tag.paths[i], &stat_buf) == -1) {
rm_log_warning_line(_("Can't open directory or file \"%s\": %s"),
tag.paths[i], strerror(errno));
} else if(S_ISDIR(stat_buf.st_mode)) {
rm_log_warning_line(_("Directories are not supported: %s"), tag.paths[i]);
} else if(S_ISREG(stat_buf.st_mode)) {
RmHasherTask *task = rm_hasher_task_new(hasher, NULL, GINT_TO_POINTER(i));
rm_hasher_task_hash(task, tag.paths[i], 0, 0, FALSE);
rm_hasher_task_finish(task);
continue;
} else {
rm_log_warning_line(_("%s: Unknown file type"), tag.paths[i]);
}
/* dummy callback for failed paths */
g_free(tag.paths[i]);
tag.paths[i] = NULL;
rm_hasher_callback(hasher, NULL, &tag, GINT_TO_POINTER(i));
}
/* wait for all hasher threads to finish... */
rm_hasher_free(hasher, TRUE);
/* tidy up */
if(tag.print_in_order) {
g_slice_free1((g_strv_length(tag.paths) + 1) * sizeof(RmDigest *),
tag.completed_digests_buffer);
}
g_strfreev(tag.paths);
return EXIT_SUCCESS;
}
static gpointer rm_session_print_first_abort_warn(_UNUSED gpointer data) {
rm_log_warning("\r");
rm_log_warning_line(_("Received Interrupt, stopping..."));
return NULL;
}
/* Method to test if a file is non stripped binary. Uses libelf*/
bool rm_util_is_nonstripped(_U const char *path, _U RmStat *statp) {
bool is_ns = false;
#if HAVE_LIBELF
g_return_val_if_fail(path, false);
if(statp && (statp->st_mode & (S_IXUSR | S_IXGRP | S_IXOTH)) == 0) {
return false;
}
/* inspired by "jschmier"'s answer at http://stackoverflow.com/a/5159890 */
int fd;
/* ELF handle */
Elf *elf;
/* section descriptor pointer */
Elf_Scn *scn;
/* section header */
GElf_Shdr shdr;
/* Open ELF file to obtain file descriptor */
if((fd = rm_sys_open(path, O_RDONLY)) == -1) {
rm_log_warning_line(_("cannot open file '%s' for nonstripped test: "), path);
rm_log_perror("");
return 0;
}
/* Protect program from using an older library */
if(elf_version(EV_CURRENT) == EV_NONE) {
rm_log_error_line(_("ELF Library is out of date!"));
return false;
}
/* Initialize elf pointer for examining contents of file */
elf = elf_begin(fd, ELF_C_READ, NULL);
/* Initialize section descriptor pointer so that elf_nextscn()
* returns a pointer to the section descriptor at index 1.
* */
scn = NULL;
/* Iterate through ELF sections */
while((scn = elf_nextscn(elf, scn)) != NULL) {
/* Retrieve section header */
gelf_getshdr(scn, &shdr);
/* If a section header holding a symbol table (.symtab)
* is found, this ELF file has not been stripped. */
if(shdr.sh_type == SHT_SYMTAB) {
is_ns = true;
break;
}
}
elf_end(elf);
rm_sys_close(fd);
#endif
return is_ns;
}