Lumpiasty/llama.cpp
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

cli : move to HTTP-based implementation (#24948)

Browse Source 
* cli: move to HTTP-based implementation

* wip

* working

* remote server ok

* cli support router mode

Co-authored-by: Piotr Wilkin <ilintar@gmail.com>

* case: router with only one model

* Apply suggestions from code review

Co-authored-by: Piotr Wilkin (ilintar) <piotr.wilkin@syndatis.com>

* remove outdated comment

* use destructor instead

* add ftype

* cli-view --> cli-ui

* pimpl

* no more json in header

* nits fixes

* also show model aliases

---------

Co-authored-by: Piotr Wilkin <ilintar@gmail.com>
Co-authored-by: Piotr Wilkin (ilintar) <piotr.wilkin@syndatis.com>
This commit is contained in:
Xuan-Son Nguyen
2026-07-08 14:52:43 +02:00
committed by GitHub
co-authored by Piotr Wilkin Piotr Wilkin
parent 07e012afdc
commit c264f65ff9
14 changed files with 1336 additions and 724 deletions
Download Patch File Download Diff File Expand all files Collapse all files
tools/server/server-context.cpp
+1
View File
@@ -4521,6 +4521,7 @@ void server_routes::init_routes() {
{ "default_generation_settings", default_generation_settings_for_props },
{ "total_slots", params.n_parallel },
{ "model_alias", meta->model_name },
{ "model_ftype", meta->model_ftype },
{ "model_path", meta->model_path },
{ "modalities", json {
{"vision", meta->has_inp_image},
tools/server/server-models.cpp
+3 -69
View File
@@ -7,6 +7,7 @@
#include "build-info.h"
#include "preset.h"
#include "download.h"
#include "http.h"
#include <cpp-httplib/httplib.h> // TODO: remove this once we use HTTP client from download.h
#include <optional>
@@ -28,14 +29,7 @@
#include <sstream>
#include <cstring>
#ifdef _WIN32
#include <winsock2.h>
#include <windows.h>
#else
#include <sys/socket.h>
#include <netinet/in.h>
#include <arpa/inet.h>
#include <unistd.h>
#ifndef _WIN32
extern char **environ;
#endif
@@ -716,66 +710,6 @@ std::optional<server_model_meta> server_models::get_meta(const std::string & nam
return std::nullopt;
}
static int get_free_port() {
#ifdef _WIN32
WSADATA wsaData;
if (WSAStartup(MAKEWORD(2, 2), &wsaData) != 0) {
return -1;
}
typedef SOCKET native_socket_t;
#define INVALID_SOCKET_VAL INVALID_SOCKET
#define CLOSE_SOCKET(s) closesocket(s)
#else
typedef int native_socket_t;
#define INVALID_SOCKET_VAL -1
#define CLOSE_SOCKET(s) close(s)
#endif
native_socket_t sock = socket(AF_INET, SOCK_STREAM, 0);
if (sock == INVALID_SOCKET_VAL) {
#ifdef _WIN32
WSACleanup();
#endif
return -1;
}
struct sockaddr_in serv_addr;
std::memset(&serv_addr, 0, sizeof(serv_addr));
serv_addr.sin_family = AF_INET;
serv_addr.sin_addr.s_addr = htonl(INADDR_ANY);
serv_addr.sin_port = htons(0);
if (bind(sock, (struct sockaddr*)&serv_addr, sizeof(serv_addr)) != 0) {
CLOSE_SOCKET(sock);
#ifdef _WIN32
WSACleanup();
#endif
return -1;
}
#ifdef _WIN32
int namelen = sizeof(serv_addr);
#else
socklen_t namelen = sizeof(serv_addr);
#endif
if (getsockname(sock, (struct sockaddr*)&serv_addr, &namelen) != 0) {
CLOSE_SOCKET(sock);
#ifdef _WIN32
WSACleanup();
#endif
return -1;
}
int port = ntohs(serv_addr.sin_port);
CLOSE_SOCKET(sock);
#ifdef _WIN32
WSACleanup();
#endif
return port;
}
// helper to convert vector<string> to char **
// pointers are only valid as long as the original vector is valid
static std::vector<char *> to_char_ptr_array(const std::vector<std::string> & vec) {
@@ -879,7 +813,7 @@ void server_models::load(const std::string & name, const load_options & opts) {
// prepare new instance info
instance_t inst;
inst.meta = meta;
inst.meta.port = get_free_port();
inst.meta.port = common_http_get_free_port();
inst.meta.status = SERVER_MODEL_STATUS_LOADING;
inst.meta.loaded_info = json{};
inst.meta.last_used = ggml_time_ms();
tools/server/server.cpp
+46 -23
View File
@@ -36,6 +36,19 @@ static inline void signal_handler(int signal) {
shutdown_handler(signal);
}
// satisfies -Wmissing-declarations (used by llama command)
int llama_server(int argc, char ** argv);
// to be used via CLI (argc / argv are used by router mode only)
int llama_server(common_params & params, int argc, char ** argv);
void llama_server_terminate();
void llama_server_terminate() {
if (shutdown_handler) {
shutdown_handler(0);
}
}
// wrapper function that handles exceptions and logs errors
// this is to make sure handler_t never throws exceptions; instead, it returns an error response
static server_http_context::handler_t ex_wrapper(server_http_context::handler_t func) {
@@ -72,9 +85,6 @@ static server_http_context::handler_t ex_wrapper(server_http_context::handler_t
};
}
// satisfies -Wmissing-declarations
int llama_server(int argc, char ** argv);
int llama_server(int argc, char ** argv) {
std::setlocale(LC_NUMERIC, "C");
@@ -94,16 +104,26 @@ int llama_server(int argc, char ** argv) {
llama_backend_init();
llama_numa_init(params.numa);
return llama_server(params, argc, argv);
}
int llama_server(common_params & params, int argc, char ** argv) {
bool is_run_by_cli = (argv == nullptr);
common_models_handler models_handler;
try {
models_handler = common_models_handler_init(params, LLAMA_EXAMPLE_SERVER);
if (common_models_handler_is_preset_repo(models_handler)) {
// apply the preset and start the server in router mode
common_models_handler_apply(models_handler, params);
// note: router mode also accepts -hf remote-preset, so we need to check that first
if (!is_run_by_cli && !params.model.hf_repo.empty()) {
try {
models_handler = common_models_handler_init(params, LLAMA_EXAMPLE_SERVER);
if (common_models_handler_is_preset_repo(models_handler)) {
// apply the preset and start the server in router mode
common_models_handler_apply(models_handler, params);
}
} catch (const std::exception & e) {
SRV_ERR("failed to fetch model metadata: %s\n", e.what());
return 1;
}
} catch (const std::exception & e) {
SRV_ERR("failed to fetch model metadata: %s\n", e.what());
return 1;
}
// router server never loads a model and must not touch the GPU
@@ -321,8 +341,9 @@ int llama_server(int argc, char ** argv) {
if (child.is_child() && child.get_mode() == SERVER_CHILD_MODE_DOWNLOAD) {
return child.run_download(params);
} else if (!is_router_server) {
} else if (!is_router_server && !is_run_by_cli) {
// single-model mode (NOT spawned by router)
// if this is invoked by CLI, model downloading should be already handled
try {
common_models_handler_apply(models_handler, params);
} catch (const std::exception & e) {
@@ -411,20 +432,22 @@ int llama_server(int argc, char ** argv) {
};
}
// TODO: refactor in common/console
// register signal handler if not running by CLI
if (!is_run_by_cli) {
#if defined (__unix__) || (defined (__APPLE__) && defined (__MACH__))
struct sigaction sigint_action;
sigint_action.sa_handler = signal_handler;
sigemptyset (&sigint_action.sa_mask);
sigint_action.sa_flags = 0;
sigaction(SIGINT, &sigint_action, NULL);
sigaction(SIGTERM, &sigint_action, NULL);
struct sigaction sigint_action;
sigint_action.sa_handler = signal_handler;
sigemptyset (&sigint_action.sa_mask);
sigint_action.sa_flags = 0;
sigaction(SIGINT, &sigint_action, NULL);
sigaction(SIGTERM, &sigint_action, NULL);
#elif defined (_WIN32)
auto console_ctrl_handler = +[](DWORD ctrl_type) -> BOOL {
return (ctrl_type == CTRL_C_EVENT) ? (signal_handler(SIGINT), true) : false;
};
SetConsoleCtrlHandler(reinterpret_cast<PHANDLER_ROUTINE>(console_ctrl_handler), true);
auto console_ctrl_handler = +[](DWORD ctrl_type) -> BOOL {
return (ctrl_type == CTRL_C_EVENT) ? (signal_handler(SIGINT), true) : false;
};
SetConsoleCtrlHandler(reinterpret_cast<PHANDLER_ROUTINE>(console_ctrl_handler), true);
#endif
}
SRV_INF("listening on %s\n", ctx_http.listening_address.c_str());