root/llama.cpp
1
0
Fork 0
mirror of https://github.com/ggerganov/llama.cpp.git synced 2024-12-25 02:44:36 +00:00
Code Issues Actions 9 Packages Projects Releases Wiki Activity

ggml : check tensor name lengths in gguf files (#10100)

Browse Source
This commit is contained in:
Diego Devesa 2024-10-31 11:40:59 +01:00 committed by GitHub
parent 1329c0a75e
commit dea5e86051
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
2 changed files with 41 additions and 11 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

45
ggml/src/ggml.c
View File

@ -22102,18 +22102,46 @@ static size_t gguf_type_size(enum gguf_type type) {
return GGUF_TYPE_SIZE[type];
}
static void gguf_tensor_info_sanitize(struct gguf_tensor_info * info) {
GGML_ASSERT(info->n_dims <= GGML_MAX_DIMS);
GGML_ASSERT(0 <= info->type && info->type < GGML_TYPE_COUNT);
static bool gguf_tensor_info_sanitize(struct gguf_tensor_info * info) {
if (info->n_dims > GGML_MAX_DIMS) {
fprintf(stderr, "%s: invalid number of dimensions (%" PRIu32 ")\n", __func__, info->n_dims);
return false;
}
if (info->type < 0 || info->type >= GGML_TYPE_COUNT) {
fprintf(stderr, "%s: invalid type (%d)\n", __func__, info->type);
return false;
}
if (strlen(info->name.data) >= GGML_MAX_NAME) {
fprintf(stderr, "%s: tensor '%s' name is too long\n", __func__, info->name.data);
return false;
}
for (uint32_t i = 0; i < info->n_dims; ++i) {
GGML_ASSERT(info->ne[i] > 0);
if (info->ne[i] <= 0) {
fprintf(stderr, "%s: invalid number of elements (%" PRIu64 ")\n", __func__, info->ne[i]);
return false;
}
}
// prevent overflow for total number of elements
GGML_ASSERT(INT64_MAX/info->ne[1] > info->ne[0]);
GGML_ASSERT(INT64_MAX/info->ne[2] > info->ne[0]*info->ne[1]);
GGML_ASSERT(INT64_MAX/info->ne[3] > info->ne[0]*info->ne[1]*info->ne[2]);
if (INT64_MAX/info->ne[1] <= info->ne[0]) {
fprintf(stderr, "%s: invalid number of elements (%" PRIu64 ")\n", __func__, info->ne[1]);
return false;
}
if (INT64_MAX/info->ne[2] <= info->ne[0]*info->ne[1]) {
fprintf(stderr, "%s: invalid number of elements (%" PRIu64 ")\n", __func__, info->ne[2]);
return false;
}
if (INT64_MAX/info->ne[3] <= info->ne[0]*info->ne[1]*info->ne[2]) {
fprintf(stderr, "%s: invalid number of elements (%" PRIu64 ")\n", __func__, info->ne[3]);
return false;
}
return true;
}
static bool gguf_fread_el(FILE * file, void * dst, size_t size, size_t * offset) {
@ -22414,8 +22442,7 @@ struct gguf_context * gguf_init_from_file(const char * fname, struct gguf_init_p
ok = ok && gguf_fread_el (file, &info->type, sizeof(info->type), &offset);
ok = ok && gguf_fread_el (file, &info->offset, sizeof(info->offset), &offset);
// TODO: return an error instead of crashing with GGML_ASSERT
gguf_tensor_info_sanitize(info);
ok = ok && gguf_tensor_info_sanitize(info);
// make sure there is no duplicated tensor names
for (uint64_t j = 0; j < i && ok; ++j) {

7
src/llama.cpp
View File

@ -4273,8 +4273,11 @@ struct llama_model_loader {
llama_tensor_weight(const llama_file * file, uint16_t idx, const char * name, const struct gguf_context * gguf_ctx, ggml_tensor * tensor) : idx(idx), tensor(tensor) {
const int tensor_idx = gguf_find_tensor(gguf_ctx, name);
offs = gguf_get_data_offset(gguf_ctx) + gguf_get_tensor_offset(gguf_ctx, tensor_idx);
if (tensor_idx < 0) {
throw std::runtime_error(format("tensor '%s' not found in the model", name));
}
offs = gguf_get_data_offset(gguf_ctx) + gguf_get_tensor_offset(gguf_ctx, tensor_idx);
if (offs + ggml_nbytes(tensor) < offs || offs + ggml_nbytes(tensor) > file->size) {
throw std::runtime_error(format("tensor '%s' data is not within the file bounds, model is corrupted or incomplete", name));
}
@ -7426,7 +7429,7 @@ static bool llm_load_tensors(
if (flags & llama_model_loader::TENSOR_NOT_REQUIRED) {
return nullptr;
}
throw std::runtime_error(format("missing tensor %s", tn.str().c_str()));
throw std::runtime_error(format("missing tensor '%s'", tn.str().c_str()));
}
// some models use the token embedding tensor as the output, but since these are used in different layers and with different ops