Files
ollama/mlxrunner/xgrammar/dynamic.c
Jesse Gross 2e036e7cdf mlx, mlxrunner: move the MLX engine out of x/
The MLX runner is the only Go inference runner left and is no longer
experimental, so its packages leave x/. The bindings become a top-level
mlx package beside the carried patches in mlx/compat, mirroring how
llama/ holds the llama.cpp integration, and the runner becomes mlxrunner
with the architectures nested under the package they implement.
Subpackages move with their parent unless listed.

  x/mlxrunner/mlx            mlx
  x/internal/mlxthread       mlx/mlxthread
  x/internal/mlxthreadtest   mlx/mlxthread/mlxthreadtest
  x/internal/mlxtest         mlx/mlxtest
  x/quant                    mlx/quant
  mlx/compat/*.patch         mlx/compat/mlx-c   (MLX patches go in mlx/compat/mlx)
  x/mlxrunner                mlxrunner
  x/models/nn                mlxrunner/nn
  x/models/<arch>            mlxrunner/model/<arch>
  x/mlxrunner/imports.go     mlxrunner/model/architectures   (new package)
  x/create                   create
  x/safetensors              fs/safetensors
  x/tokenizer                mlxrunner/tokenizer

Every package keeps its name, so the Go changes are the import path
rewrites the moves force, and the CMake, Dockerfile, CI cache keys, drift
check and Darwin payload script follow the new paths. Four edits are not
paths: the runner's blank architecture imports become the package
mlxrunner/model/architectures, so the list to extend for a new model sits
beside the architecture directories; a depguard rule keeps the two test
harnesses out of non-test code, as the x/internal placement used to; the
CI change filter's two entries for the long-deleted x/imagegen/mlx now
name the bindings' CMake project and the carried patches, so a change to
either builds the payload; and the tokenizer parity test reads its
fixtures from its own testdata instead of walking out of x/.

x/server and x/imagegen/manifest stay for the next two commits.
2026-09-16 14:06:08 -07:00

159 lines
5.9 KiB
C

#include "dynamic.h"
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#ifdef _WIN32
#include <windows.h>
#ifndef LOAD_LIBRARY_SEARCH_DLL_LOAD_DIR
#define LOAD_LIBRARY_SEARCH_DLL_LOAD_DIR 0x00000100
#endif
#ifndef LOAD_LIBRARY_SEARCH_DEFAULT_DIRS
#define LOAD_LIBRARY_SEARCH_DEFAULT_DIRS 0x00001000
#endif
static void* open_library(const char* path) {
int length = MultiByteToWideChar(CP_UTF8, MB_ERR_INVALID_CHARS, path, -1, NULL, 0);
if (length == 0) {
return NULL;
}
wchar_t* wide_path = malloc((size_t)length * sizeof(wchar_t));
if (wide_path == NULL) {
return NULL;
}
if (MultiByteToWideChar(CP_UTF8, MB_ERR_INVALID_CHARS, path, -1, wide_path, length) == 0) {
free(wide_path);
return NULL;
}
HMODULE module = LoadLibraryExW(
wide_path,
NULL,
LOAD_LIBRARY_SEARCH_DLL_LOAD_DIR | LOAD_LIBRARY_SEARCH_DEFAULT_DIRS);
free(wide_path);
return (void*)module;
}
#define OPEN(path) open_library(path)
#define CLOSE(handle) FreeLibrary((HMODULE)(handle))
#define SYMBOL(handle, name) ((void*)GetProcAddress((HMODULE)(handle), name))
#else
#include <dlfcn.h>
#define OPEN(path) dlopen(path, RTLD_NOW | RTLD_LOCAL)
#define CLOSE(handle) dlclose(handle)
#define SYMBOL(handle, name) dlsym(handle, name)
#endif
static const char* (*version_fn)(void);
static const char* (*last_error_fn)(void);
static int (*compiler_new_fn)(const char*, size_t, const uint64_t*, size_t, int32_t, const int32_t*, size_t, int32_t, int64_t, ollama_xgrammar_compiler**);
static void (*compiler_free_fn)(ollama_xgrammar_compiler*);
static int (*matcher_new_fn)(ollama_xgrammar_compiler*, const char*, size_t, ollama_xgrammar_matcher**);
static void (*matcher_free_fn)(ollama_xgrammar_matcher*);
static int (*matcher_fill_fn)(ollama_xgrammar_matcher*, int32_t*, size_t, int*);
static int (*matcher_accept_fn)(ollama_xgrammar_matcher*, int32_t, int*);
static int (*matcher_rollback_fn)(ollama_xgrammar_matcher*, int32_t);
static int (*matcher_is_terminated_fn)(ollama_xgrammar_matcher*, int*);
static const char* load_error;
static void clear_symbols(void) {
version_fn = NULL;
last_error_fn = NULL;
compiler_new_fn = NULL;
compiler_free_fn = NULL;
matcher_new_fn = NULL;
matcher_free_fn = NULL;
matcher_fill_fn = NULL;
matcher_accept_fn = NULL;
matcher_rollback_fn = NULL;
matcher_is_terminated_fn = NULL;
}
#define LOAD(handle, field, name) do { \
*(void**)(&field) = SYMBOL((handle)->ctx, name); \
if ((field) == NULL) { load_error = "xgrammar library is missing symbol " name; goto fail; } \
} while (0)
int ollama_xgrammar_dynamic_load(ollama_xgrammar_dynamic_handle* handle, const char* path) {
clear_symbols();
if (handle == NULL || path == NULL) {
load_error = "invalid xgrammar library path";
return 1;
}
handle->ctx = OPEN(path);
if (handle->ctx == NULL) {
load_error = "unable to open xgrammar library";
return 1;
}
LOAD(handle, version_fn, "ollama_xgrammar_version");
LOAD(handle, last_error_fn, "ollama_xgrammar_last_error");
LOAD(handle, compiler_new_fn, "ollama_xgrammar_compiler_new");
LOAD(handle, compiler_free_fn, "ollama_xgrammar_compiler_free");
LOAD(handle, matcher_new_fn, "ollama_xgrammar_matcher_new");
LOAD(handle, matcher_free_fn, "ollama_xgrammar_matcher_free");
LOAD(handle, matcher_fill_fn, "ollama_xgrammar_matcher_fill");
LOAD(handle, matcher_accept_fn, "ollama_xgrammar_matcher_accept");
LOAD(handle, matcher_rollback_fn, "ollama_xgrammar_matcher_rollback");
LOAD(handle, matcher_is_terminated_fn, "ollama_xgrammar_matcher_is_terminated");
load_error = NULL;
return 0;
fail:
CLOSE(handle->ctx);
handle->ctx = NULL;
clear_symbols();
return 1;
}
const char* ollama_xgrammar_dynamic_version(void) {
return version_fn == NULL ? "" : version_fn();
}
const char* ollama_xgrammar_dynamic_error(void) {
if (last_error_fn != NULL) {
const char* message = last_error_fn();
if (message != NULL && message[0] != '\0') {
return message;
}
}
return load_error == NULL ? "xgrammar library error" : load_error;
}
static int capture_error(int result, char** error) {
if (error != NULL) {
*error = NULL;
}
if (result == 0 || error == NULL) {
return result;
}
const char* message = ollama_xgrammar_dynamic_error();
size_t size = strlen(message) + 1;
*error = malloc(size);
if (*error != NULL) {
memcpy(*error, message, size);
}
return result;
}
int ollama_xgrammar_dynamic_compiler_new(const char* d, size_t ds, const uint64_t* o, size_t n, int32_t v, const int32_t* s, size_t ns, int32_t mt, int64_t cb, ollama_xgrammar_compiler** m, char** error) {
return capture_error(compiler_new_fn(d, ds, o, n, v, s, ns, mt, cb, m), error);
}
void ollama_xgrammar_dynamic_compiler_free(ollama_xgrammar_compiler* m) { compiler_free_fn(m); }
int ollama_xgrammar_dynamic_matcher_new(ollama_xgrammar_compiler* m, const char* s, size_t n, ollama_xgrammar_matcher** out, char** error) {
return capture_error(matcher_new_fn(m, s, n, out), error);
}
void ollama_xgrammar_dynamic_matcher_free(ollama_xgrammar_matcher* m) { matcher_free_fn(m); }
int ollama_xgrammar_dynamic_matcher_fill(ollama_xgrammar_matcher* m, int32_t* b, size_t n, int* a, char** error) {
return capture_error(matcher_fill_fn(m, b, n, a), error);
}
int ollama_xgrammar_dynamic_matcher_accept(ollama_xgrammar_matcher* m, int32_t t, int* a, char** error) {
return capture_error(matcher_accept_fn(m, t, a), error);
}
int ollama_xgrammar_dynamic_matcher_rollback(ollama_xgrammar_matcher* m, int32_t n, char** error) {
return capture_error(matcher_rollback_fn(m, n), error);
}
int ollama_xgrammar_dynamic_matcher_is_terminated(ollama_xgrammar_matcher* m, int* t, char** error) {
return capture_error(matcher_is_terminated_fn(m, t), error);
}