remove support for extra args

common/arg.cpp

@@ -2488,13 +2488,6 @@ common_params_context common_params_parser_init(common_params & params, llama_ex
             params.models_max = value;
         }
     ).set_examples({LLAMA_EXAMPLE_SERVER}).set_env("LLAMA_ARG_MODELS_MAX"));
-    add_opt(common_arg(
-        {"--models-allow-extra-args"},
-        string_format("for router server, allow extra arguments for models; important: some arguments can allow users to access local file system, use with caution (default: %s)", params.models_allow_extra_args ? "enabled" : "disabled"),
-        [](common_params & params) {
-            params.models_allow_extra_args = true;
-        }
-    ).set_examples({LLAMA_EXAMPLE_SERVER}).set_env("LLAMA_ARG_MODELS_ALLOW_EXTRA_ARGS"));
     add_opt(common_arg(
         {"--no-models-autoload"},
         "disables automatic loading of models (default: enabled)",

common/common.h

@@ -462,7 +462,6 @@ struct common_params {
     std::string models_dir = ""; // directory containing models for the router server
     int models_max = 4;          // maximum number of models to load simultaneously
     bool models_autoload = true; // automatically load models when requested via the router server
-    bool models_allow_extra_args = false; // allow passing extra arguments when loading models via the router server
 
     bool log_json = false;
 

tools/server/server-models.cpp

@@ -324,7 +324,7 @@ void server_models::unload_lru() {
     }
 }
 
-void server_models::load(const std::string & name, const std::vector<std::string> & extra_args, bool auto_load) {
+void server_models::load(const std::string & name, bool auto_load) {
     if (!has_model(name)) {
         throw std::runtime_error("model name=" + name + " is not found");
     }
@@ -381,13 +381,6 @@ void server_models::load(const std::string & name, const std::vector<std::string
             child_args.push_back(inst.meta.name);
             child_args.push_back("--port");
             child_args.push_back(std::to_string(inst.meta.port));
-
-            // append extra args
-            if (base_params.models_allow_extra_args) {
-                for (const auto & arg : extra_args) {
-                    child_args.push_back(arg);
-                }
-            }
         }
 
         std::vector<std::string> child_env = base_env; // copy
@@ -536,7 +529,7 @@ bool server_models::ensure_model_loaded(const std::string & name) {
     }
     if (meta->status == SERVER_MODEL_STATUS_UNLOADED) {
         SRV_INF("model name=%s is not loaded, loading...\n", name.c_str());
-        load(name, {}, true);
+        load(name, true);
     }
 
     SRV_INF("waiting until model name=%s is fully loaded...\n", name.c_str());

tools/server/server-models.h

@@ -103,7 +103,7 @@ public:
     std::vector<server_model_meta> get_all_meta();
 
     // if auto_load is true, load the model with previous args if any
-    void load(const std::string & name, const std::vector<std::string> & extra_args, bool auto_load);
+    void load(const std::string & name, bool auto_load);
     void unload(const std::string & name);
     void unload_all();
 

tools/server/server.cpp

@@ -3311,11 +3311,6 @@ public:
         auto res = std::make_unique<server_res_generator>(ctx_server);
         json body = json::parse(req.body);
         std::string name = json_value(body, "model", std::string());
-        std::vector<std::string> extra_args = json_value(body, "extra_args", std::vector<std::string>());
-        if (!params.models_allow_extra_args && !extra_args.empty()) {
-            res->error(format_error_response("extra_args is not allowed", ERROR_TYPE_INVALID_REQUEST));
-            return res;
-        }
         auto model = models->get_meta(name);
         if (!model.has_value()) {
             res->error(format_error_response("model is not found", ERROR_TYPE_NOT_FOUND));
@@ -3325,7 +3320,7 @@ public:
             res->error(format_error_response("model is already loaded", ERROR_TYPE_INVALID_REQUEST));
             return res;
         }
-        models->load(name, extra_args, false);
+        models->load(name, false);
         res->ok({{"success", true}});
         return res;
     };
@@ -4014,9 +4009,6 @@ int main(int argc, char ** argv, char ** envp) {
         LOG_INF("%s: router server is listening on %s\n", __func__, ctx_http.listening_address.c_str());
         LOG_INF("%s: NOTE: router mode is experimental\n", __func__);
         LOG_INF("%s:       it is not recommended to use this mode in untrusted environments\n", __func__);
-        if (params.models_allow_extra_args) {
-            LOG_WRN("%s: extra_args is enabled; this may lead to security issues if the server is exposed to untrusted clients\n", __func__);
-        }
         if (ctx_http.thread.joinable()) {
             ctx_http.thread.join(); // keep the main thread alive
         }