Attention is currently required from: fixeria, neels, pespin.

Timur Davydov has posted comments on this change by Timur Davydov. ( https://gerrit.osmocom.org/c/libosmocore/+/41813?usp=email )

Change subject: Add Emscripten build support and JS callback logging backend ......................................................................

Patch Set 5:

(15 comments)

Commit Message:

https://gerrit.osmocom.org/c/libosmocore/+/41813/comment/c8fae8a5_871cf07b?u... : PS5, Line 7: Add Emscripten build support and JS callback logging backend

Why is this needed? How is this used? What's the context? I'm missing a good description here for a […]

Thanks for pointing this out.

The purpose of this patch is to make libosmocore usable in sandboxed, non-POSIX environments, specifically WebAssembly targets built with the Emscripten toolchain.

In such environments, a number of assumptions made by libosmocore (process model, availability of standard OS interfaces, and I/O mechanisms) do not hold, which currently prevents the library from being built and used at all. This patch introduces the minimal platform-specific adjustments required to enable successful builds and basic runtime integration in that context.

The intended use case is running Osmocom protocol and control components inside a WebAssembly runtime, where interaction with the outside world is mediated by the host environment rather than direct OS access.

All changes are limited to Emscripten-specific build paths and are inactive for native targets, leaving existing deployments and behavior unchanged.

I will update the commit message to better reflect this context.

File include/osmocom/core/logging.h:

https://gerrit.osmocom.org/c/libosmocore/+/41813/comment/61a177af_16eba50d?u... : PS5, Line 285: LOG_TGT_TYPE_WEB, /*!< Web logging */

AFAIU this is not "web generic" but "emscripten" specific, so let's please rename to something more […]

Thanks for the clarification. You are right — the current naming suggests a generic “web” abstraction, while in practice this mechanism is only enabled for Emscripten builds.

I will rename the API accordingly and update the commit message to better explain the intended scope and usage. I will send an updated patch shortly.

File src/core/netdev.c:

https://gerrit.osmocom.org/c/libosmocore/+/41813/comment/9b4fa443_a0750960?u... : PS5, Line 67: #if (!EMBEDDED) && !defined(__EMSCRIPTEN__)

I'd say better do this logic inside configure. […]

Fully agree, that's cleaner and more in line with the rest of the library. I went ahead and added a new configure option `ENABLE_NETDEV` in configure.ac. For EMBEDDED and EMSCRIPTEN builds it is explicitly set to `false`

Right now it fails to compile

File src/core/osmo_io_internal.h:

https://gerrit.osmocom.org/c/libosmocore/+/41813/comment/3e3463c3_a1fc752c?u... : PS5, Line 7: HAVE_LIBSCTP

Indeed config.h should be at the top.

Thanks for catching this! You're right — I completely missed that config.h is included after this block

File src/core/serial.c:

https://gerrit.osmocom.org/c/libosmocore/+/41813/comment/e06b29b5_62274514?u... : PS5, Line 27: #if !defined(__EMSCRIPTEN__)

Again, I'd say better add a HAVE_SERIAL or alike in configure.ac. […]

Thanks for the note!

After taking a closer look at configure.ac, I noticed the project already uses `ENABLE_SERIAL` to conditionally skip compiling the serial.c file.

The file compiles without issues, but in emscripten/web environments there's simply no need for serial support at all.

File src/core/socket.c:

https://gerrit.osmocom.org/c/libosmocore/+/41813/comment/db2155dc_fb4f7485?u... : PS5, Line 2050: #ifdef HAVE_LIBSCTP

missing explanation for this change. It may be a separate patch.

Explained in the comment for the function `osmo_sock_multiaddr_get_ip_and_port`

https://gerrit.osmocom.org/c/libosmocore/+/41813/comment/1dc18b36_a3750508?u... : PS5, Line 1928: #ifdef HAVE_LIBSCTP

Why this change? I'm lacking any rationale here. Also, looks like this may be a separate patch.

About functions `osmo_sock_multiaddr_get_ip_and_port`, `osmo_multiaddr_ip_and_port_snprintf` and `osmo_sock_multiaddr_get_name_buf`:

These functions were previously conditionally compiled only under `HAVE_LIBSCTP`. However, `libosmo-netif` uses some of them without the same `HAVE_LIBSCTP` guards, leading to link failures when libsctp is disabled.

Additionally, the implementations already included a fallback to legacy (single-address) behavior, so they do not strictly depend on SCTP/multi-homing features.

Refactoring done as follows: - `osmo_sock_multiaddr_get_ip_and_port` was slightly adjusted for better fallback handling. - `osmo_multiaddr_ip_and_port_snprintf` and `osmo_sock_multiaddr_get_name_buf` were moved outside the `HAVE_LIBSCTP` conditional block, as they do not depend on libsctp at all.

File src/vty/Makefile.am:

https://gerrit.osmocom.org/c/libosmocore/+/41813/comment/b1057c93_11cc95e4?u... : PS5, Line 36: libosmovty_la_SOURCES += telnet_interface_dummy.c

This for sure also needs an explanation.

In Emscripten there are no sockets so `telnet_interface.c` fails to compile.

However, completely excluding the file from the build leads to link errors (undefined references to the telnet functions).

To resolve this, I added a new file `telnet_interface_dummy.c` with stub/empty implementations of the required functions.

I considered adding many `#ifdef __EMSCRIPTEN__` blocks directly inside telnet_interface.c, but decided that separate dummy stubs are the cleaner approach.

File src/vty/logging_vty.c:

https://gerrit.osmocom.org/c/libosmocore/+/41813/comment/69d27735_c443c3fd?u... : PS5, Line 907: #if !defined(__EMSCRIPTEN__)

this also needs explanation.

The conditional compilation here was added by mistake — it's not needed in this place.

This spot was accidentally affected when I removed the file logging parts (Emscripten only emulates files via IndexedDB with no native operations, so file logging isn't usable anyway).

https://gerrit.osmocom.org/c/libosmocore/+/41813/comment/8d31850a_a69015c0?u... : PS5, Line 1038: DEFUN(cfg_log_web, cfg_log_web_cmd,

#if defined(__EMSCRIPTEN__)

Thanks — added conditional compilation `#if defined(__EMSCRIPTEN__)`

https://gerrit.osmocom.org/c/libosmocore/+/41813/comment/d7daf5db_e7550d3b?u... : PS5, Line 1047: vty_out(vty, "%% Unable to create WEB log for %s", VTY_NEWLINE);

this string end looks wrong

Thanks, fixed

https://gerrit.osmocom.org/c/libosmocore/+/41813/comment/35256656_e28f4c34?u... : PS5, Line 1067: vty_out(vty, "%% Unable to find WEB log target for %s", VTY_NEWLINE);

this string end looks wrong

Thanks, fixed

https://gerrit.osmocom.org/c/libosmocore/+/41813/comment/446e4480_df154fdc?u... : PS5, Line 1130:

No need for a space here.

Thanks, fixed

https://gerrit.osmocom.org/c/libosmocore/+/41813/comment/f04400fe_0e2783cd?u... : PS5, Line 1361: install_lib_element(CONFIG_NODE, &cfg_log_web_cmd); : install_lib_element(CONFIG_NODE, &cfg_no_log_web_cmd);

I assume these commands are only relevant for emscripten? […]

Thanks — added conditional compilation `#if defined(__EMSCRIPTEN__)`

File src/vty/telnet_interface_dummy.c:

https://gerrit.osmocom.org/c/libosmocore/+/41813/comment/6ab07699_cea1ec23?u... : PS5, Line 18: #include <sys/socket.h> : #include <netinet/in.h> : #include <errno.h> : #include <stdlib.h> : #include <stdio.h> : #include <string.h> : #i

(Not critical) Most if the includes are not needed here and can be removed?

Thanks for pointing that out!

You're right — many of the includes here are indeed not needed anymore (leftover from copy-paste)