vbatts/llama.cpp
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

rpc : fix segfault with nkvo (#9389)

Browse source 
* rpc : fix nkvo

* rpc : buf_size must not be static

ref: #9337

---------

Co-authored-by: slaren <slarengh@gmail.com>
This commit is contained in:
Radoslav Gerganov 2024-09-09 18:40:10 +03:00 • committed by GitHub
parent 5fac4d5764
commit 293bebe077
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
3 changed files with 16 additions and 10 deletions
Download patch file Download diff file Expand all files Collapse all files

18
ggml/src/ggml-rpc.cpp
View file

@ -883,15 +883,17 @@ ggml_tensor * rpc_server::deserialize_tensor(struct ggml_context * ctx, const rp
}
result->buffer = reinterpret_cast<ggml_backend_buffer_t>(tensor->buffer);
if (result->buffer && buffers.find(result->buffer) == buffers.end()) {
return nullptr;
result->buffer = nullptr;
}
// require that the tensor data does not go beyond the buffer end
uint64_t tensor_size = (uint64_t) ggml_nbytes(result);
uint64_t buffer_start = (uint64_t) ggml_backend_buffer_get_base(result->buffer);
uint64_t buffer_size = (uint64_t) ggml_backend_buffer_get_size(result->buffer);
GGML_ASSERT(tensor->data + tensor_size >= tensor->data); // check for overflow
GGML_ASSERT(tensor->data >= buffer_start && tensor->data + tensor_size <= buffer_start + buffer_size);
if (result->buffer) {
// require that the tensor data does not go beyond the buffer end
uint64_t tensor_size = (uint64_t) ggml_nbytes(result);
uint64_t buffer_start = (uint64_t) ggml_backend_buffer_get_base(result->buffer);
uint64_t buffer_size = (uint64_t) ggml_backend_buffer_get_size(result->buffer);
GGML_ASSERT(tensor->data + tensor_size >= tensor->data); // check for overflow
GGML_ASSERT(tensor->data >= buffer_start && tensor->data + tensor_size <= buffer_start + buffer_size);
}
result->op = (ggml_op) tensor->op;
for (uint32_t i = 0; i < GGML_MAX_OP_PARAMS / sizeof(int32_t); i++) {
@ -1060,7 +1062,7 @@ bool rpc_server::graph_compute(const std::vector<uint8_t> & input, std::vector<u
const rpc_tensor * tensors = (const rpc_tensor *)(input.data() + sizeof(n_nodes) + n_nodes*sizeof(uint64_t) + sizeof(n_tensors));
GGML_PRINT_DEBUG("[%s] n_nodes: %u, n_tensors: %u\n", __func__, n_nodes, n_tensors);
static size_t buf_size = ggml_tensor_overhead()*(n_nodes + n_tensors) + ggml_graph_overhead_custom(n_nodes, false);
size_t buf_size = ggml_tensor_overhead()*(n_nodes + n_tensors) + ggml_graph_overhead_custom(n_nodes, false);
struct ggml_init_params params = {
/*.mem_size =*/ buf_size,
/*.mem_buffer =*/ NULL,