.gitignore

@@ -5,8 +5,5 @@ initrd/boot/moon
 env-local.sh
 initrd/bin/**
 base/usr/**
-base/etc/skel/LICENSE
 .fakeroot
 kernel/config.cmake
-ports/out/
-ports/temp/

CMakeLists.txt

@@ -5,7 +5,7 @@ set(CMAKE_CXX_COMPILER_WORKS 1)
 
 set(CMAKE_CROSSCOMPILING true)
 
-project(Luna LANGUAGES C CXX ASM ASM_NASM VERSION 0.5.0)
+project(Luna LANGUAGES C CXX ASM ASM_NASM VERSION 0.4.0)
 set(LUNA_RELEASE_NAME "Mercury") # Name for alpha releases
 
 set(LUNA_ROOT ${CMAKE_CURRENT_LIST_DIR})
@@ -33,7 +33,7 @@ set(COMMON_FLAGS -Wall -Wextra -Werror -Wvla
     -Wdisabled-optimization -Wformat=2 -Winit-self 
     -Wmissing-include-dirs -Wswitch-default -Wcast-qual 
     -Wundef -Wcast-align -Wwrite-strings -Wlogical-op 
-    -Wredundant-decls -Wshadow -Wconversion -Wbidi-chars=any
+    -Wredundant-decls -Wshadow -Wconversion
     -fno-asynchronous-unwind-tables -fno-omit-frame-pointer 
     -std=c++20 -fno-rtti -fno-exceptions)
 

README.md

@@ -1,5 +1,5 @@
 # Luna
-A simple POSIX-based operating system for personal computers, written in C++. [![Build Status](https://drone.cloudapio.eu/api/badges/apio/Luna/status.svg)](https://drone.cloudapio.eu/apio/Luna)
+A very basic POSIX-based operating system for personal computers, written in C++. [![Build Status](https://drone.cloudapio.eu/api/badges/apio/Luna/status.svg)](https://drone.cloudapio.eu/apio/Luna)
 
 ## Another UNIX clone?
 [Yes, another UNIX clone](https://wiki.osdev.org/User:Sortie/Yes_Another_Unix_Clone).
@@ -8,18 +8,15 @@ A simple POSIX-based operating system for personal computers, written in C++. [!
 - x86_64-compatible lightweight [kernel](kernel/).
 - Preemptive multitasking, with a round-robin [scheduler](kernel/src/thread/).
 - [Virtual file system](kernel/src/fs/) with a simple [tmpfs](kernel/src/fs/tmpfs/) and read-only [ext2](kernel/src/fs/ext2/) support.
-- Can [load ELF programs](kernel/src/binfmt/ELF.cpp), [shebang scripts](kernel/src/binfmt/Script.cpp) or [arbitrary binary formats](kernel/src/binfmt/BinaryFormat.h) (registered through kernel modules, which are not supported yet =D).
+- Can [load ELF programs](kernel/src/thread/ELF.cpp) from the file system as userspace tasks.
 - Boots from an [ext2](apps/preinit.cpp) root filesystem (a bit slow for now).
-- [System call](kernel/src/sys/) interface and [C Library](libc/), aiming to be almost POSIX-compatible.
-- Support for [several third-party programs](ports/), including the [GNU binutils](ports/binutils/PACKAGE) suite of utilities.
+- [System call](kernel/src/sys/) interface and [C Library](libc/), aiming to be mostly POSIX-compatible.
 - POSIX [signal](libc/src/signal.cpp) support.
 - Designed to be [portable](kernel/src/arch), no need to be restricted to x86_64.
-- Everything is [UTF-8](libluna/include/luna/Utf8.h).
-- [UNIX local domain sockets](kernel/src/net/UnixSocket.cpp), allowing for local IPC.
-- [POSIX shared memory](libc/include/sys/mman.h) support.
+- Designed around [UTF-8](libluna/include/luna/Utf8.h).
 - Environment-agnostic [utility library](libluna/), which can be used in both kernel and userspace.
-- Return-oriented [error propagation](libluna/include/luna/Result.h), inspired by Rust and SerenityOS. No exceptions here :).
-- An extensive set of [standard Unix utilities](apps/), from [ls](apps/ls.cpp) to [uname](apps/uname.cpp) to [base64](apps/base64.cpp). Written in modern C++ and very small amounts of code, using Luna's practical [OS library](libos/).
+- Return-oriented [error propagation](libluna/include/luna/Result.h), inspired by Rust and SerenityOS.
+- Build system uses [CMake](CMakeLists.txt).
 
 ## Setup
 
@@ -53,7 +50,7 @@ These images do reflect the latest changes on the `main` branch, but are obvious
 
 ## Is there third-party software I can use on Luna?
 
-Yes! A ports system is in place, and you can use the build scripts to add some ports to your image. More information in the [Ports](ports/README.md) page.
+There is no infrastructure for porting third-party software nor there are any patches in the repo for now, but some third-party programs can run on Luna, including [my own library](https://git.cloudapio.eu/apio/minitar).
 
 ## License
-Luna is open-source and free software under the [BSD-2-Clause License](LICENSE).
+Luna is open-source and free software under the [BSD-2 License](LICENSE).

apps/CMakeLists.txt

@@ -28,6 +28,8 @@ luna_app(stat.cpp stat)
 luna_app(uname.cpp uname)
 luna_app(base64.cpp base64)
 luna_app(login.cpp login)
+luna_app(ipc-test.cpp ipc-test)
+luna_app(signal-test.cpp signal-test)
 luna_app(mount.cpp mount)
 luna_app(umount.cpp umount)
 luna_app(ps.cpp ps)
@@ -35,8 +37,8 @@ luna_app(time.cpp time)
 luna_app(ln.cpp ln)
 luna_app(mktemp.cpp mktemp)
 luna_app(sysfuzz.cpp sysfuzz)
+luna_app(pivot_root.cpp pivot_root)
 luna_app(cp.cpp cp)
 luna_app(kill.cpp kill)
 luna_app(gol.cpp gol)
-luna_app(touch.cpp touch)
-luna_app(free.cpp free)
+luna_app(buffer-test.cpp buffer-test)

apps/buffer-test.cpp

@@ -0,0 +1,27 @@
+#include <assert.h>
+#include <stdio.h>
+#include <unistd.h>
+
+int main()
+{
+    fprintf(stderr, "Writing incomplete line to stdout (_IOLBF=%d)...\n", stdout->_buf.mode);
+    fputs("hi!", stdout);
+    sleep(3);
+    putchar('\n');
+    fprintf(stderr, "Incomplete line should have been written.\n");
+
+    FILE* f = fopen("/dev/console", "w+");
+    assert(f);
+    assert(setvbuf(f, NULL, _IOFBF, 0) == 0);
+
+    fprintf(stderr, "Writing long text to file (_IOFBF=%d)...\n", f->_buf.mode);
+
+    fputs("Hello world!\nHow are you doing!\nThis is a test for many lines of buffering.\n", f);
+
+    sleep(3);
+    fflush(f);
+
+    fprintf(stderr, "Long text should have been written.\n");
+
+    fclose(f);
+}

apps/edit.cpp

@@ -1,8 +1,6 @@
 #include <luna/String.h>
 #include <os/ArgumentParser.h>
 #include <os/File.h>
-#include <os/FileSystem.h>
-#include <os/Prompt.h>
 
 using os::File;
 
@@ -16,20 +14,10 @@ Result<int> luna_main(int argc, char** argv)
     parser.add_positional_argument(pathname, "path"_sv, true);
     parser.parse(argc, argv);
 
-    if (os::FileSystem::exists(pathname))
-    {
-        String prompt = TRY(String::format("File %s already exists. Overwrite?"_sv, pathname.chars()));
-        bool overwrite = os::conditional_prompt(prompt.chars(), os::DefaultNo);
-        if (!overwrite) return 0;
-    }
-
     auto file = TRY(File::open_or_create(pathname, File::WriteOnly));
 
     auto input = File::standard_input();
 
-    os::println("- Editing %s. Press Enter to start a new line, or Ctrl+D at the start of a line to save and exit. -",
-                pathname.chars());
-
     while (1)
     {
         String line = TRY(input->read_line());

apps/free.cpp

@@ -1,50 +0,0 @@
-#include <luna/Units.h>
-#include <os/ArgumentParser.h>
-#include <os/File.h>
-#include <sys/memstat.h>
-
-Result<int> luna_main(int argc, char** argv)
-{
-    bool reserved { false };
-    bool human_readable { false };
-
-    os::ArgumentParser parser;
-    parser.add_description("Query the amount of memory available and used in the system.");
-    parser.add_system_program_info("free"_sv);
-    parser.add_switch_argument(reserved, 'r', "show-reserved"_sv, "show reserved system memory"_sv);
-    parser.add_switch_argument(human_readable, 'h', "human-readable"_sv, "show output in human-readable units"_sv);
-    parser.parse(argc, argv);
-
-    struct membuf buf;
-    memstat(&buf);
-
-    if (human_readable)
-    {
-        auto total = TRY(to_dynamic_unit(buf.mem_total, 10, false, Unit::Binary, true));
-        auto free = TRY(to_dynamic_unit(buf.mem_free, 10, false, Unit::Binary, true));
-        auto used = TRY(to_dynamic_unit(buf.mem_used, 10, false, Unit::Binary, true));
-
-        os::println("%s total memory, out of which:", total.chars());
-        os::println("%s used", used.chars());
-        os::println("%s free", free.chars());
-    }
-    else
-    {
-        os::println("%lu total memory, out of which:", buf.mem_total);
-        os::println("%lu used", buf.mem_used);
-        os::println("%lu free", buf.mem_free);
-    }
-
-    if (reserved)
-    {
-        if (human_readable)
-        {
-            auto reserved_string = TRY(to_dynamic_unit(buf.mem_reserved, 10, false, Unit::Binary, true));
-
-            os::println("%s reserved", reserved_string.chars());
-        }
-        else { os::println("%lu reserved", buf.mem_reserved); }
-    }
-
-    return 0;
-}

apps/gol.cpp

@@ -6,7 +6,6 @@
 #include <stdio.h>
 #include <stdlib.h>
 #include <sys/ioctl.h>
-#include <sys/mman.h>
 #include <time.h>
 #include <unistd.h>
 
@@ -25,7 +24,6 @@ static int g_fb_height;
 static int g_fd;
 
 static Cell* g_cells;
-static char* g_fb;
 
 static Result<void> fill_cells()
 {
@@ -49,29 +47,27 @@ static Cell& find_cell(int row, int column)
 
 static constexpr int BYTES_PER_PIXEL = 4;
 
+static char* g_buf;
+
 static void draw_cells()
 {
+    lseek(g_fd, 0, SEEK_SET);
+
     const int CELL_WIDTH = g_fb_width / g_num_columns;
     const int CELL_HEIGHT = g_fb_height / g_num_rows;
 
     for (int i = 0; i < g_num_rows; i++)
     {
+        memset(g_buf, 0, g_fb_width * BYTES_PER_PIXEL);
 
         for (int j = 0; j < g_num_columns; j++)
         {
-            char* buf = g_fb + (i * g_fb_width * CELL_HEIGHT * BYTES_PER_PIXEL);
-
             auto& cell = find_cell(i, j);
-            u8 color = cell.state ? 0xff : 0x00;
-            for (int k = 0; k < CELL_HEIGHT; k++)
-            {
-                memset(buf + (j * CELL_WIDTH * BYTES_PER_PIXEL), color, CELL_WIDTH * BYTES_PER_PIXEL);
-                buf += g_fb_width * BYTES_PER_PIXEL;
-            }
+            if (cell.state) memset(g_buf + (j * CELL_WIDTH * BYTES_PER_PIXEL), 0xff, CELL_WIDTH * BYTES_PER_PIXEL);
         }
-    }
 
-    msync(g_fb, g_fb_height * g_fb_width * BYTES_PER_PIXEL, MS_SYNC);
+        for (int j = 0; j < CELL_HEIGHT; j++) { write(g_fd, g_buf, g_fb_width * BYTES_PER_PIXEL); }
+    }
 }
 
 static int find_neighbors(int row, int column)
@@ -141,7 +137,7 @@ Result<int> luna_main(int argc, char** argv)
     else
         srand((unsigned)time(NULL));
 
-    g_fd = open("/dev/fb0", O_RDWR);
+    g_fd = open("/dev/fb0", O_WRONLY);
     if (g_fd < 0)
     {
         perror("gol: cannot open framebuffer for writing");
@@ -153,13 +149,7 @@ Result<int> luna_main(int argc, char** argv)
 
     TRY(fill_cells());
 
-    g_fb =
-        (char*)mmap(nullptr, g_fb_height * g_fb_width * BYTES_PER_PIXEL, PROT_READ | PROT_WRITE, MAP_SHARED, g_fd, 0);
-    if (g_fb == MAP_FAILED)
-    {
-        perror("gol: cannot map framebuffer into memory");
-        return 1;
-    }
+    g_buf = (char*)TRY(calloc_impl(g_fb_width, BYTES_PER_PIXEL));
 
     draw_cells();
 
@@ -172,8 +162,5 @@ Result<int> luna_main(int argc, char** argv)
 
     usleep(delay_at_end * 1000);
 
-    munmap(g_fb, g_fb_height * g_fb_width * BYTES_PER_PIXEL);
-    free(g_cells);
-
     return 0;
 }

apps/init.cpp

@@ -3,14 +3,10 @@
 #include <luna/PathParser.h>
 #include <luna/Sort.h>
 #include <luna/String.h>
-#include <luna/Utf8.h>
 #include <luna/Vector.h>
-#include <os/ArgumentParser.h>
 #include <os/Directory.h>
 #include <os/File.h>
 #include <os/Process.h>
-#include <os/Security.h>
-#include <pwd.h>
 #include <signal.h>
 #include <stdio.h>
 #include <stdlib.h>
@@ -21,10 +17,20 @@
 #include <sys/sysmacros.h>
 #include <unistd.h>
 
-static bool g_is_system = false;
-
 FILE* g_init_log = nullptr;
 
+// Request a successful exit from the system (for tests)
+void sigterm_handler(int)
+{
+    _exit(0);
+}
+
+// Request a failure exit from the system (for tests)
+void sigquit_handler(int)
+{
+    _exit(1);
+}
+
 struct Service
 {
     String name;
@@ -34,8 +40,6 @@ struct Service
     String standard_output;
     String standard_error;
     String standard_input;
-    Option<uid_t> user {};
-    Option<gid_t> group {};
     bool wait { false };
     Option<pid_t> pid {};
 };
@@ -52,20 +56,6 @@ static void do_log(const char* format, ...)
     va_end(ap);
 }
 
-// Request a successful exit from the system (for tests)
-void sigterm_handler(int)
-{
-    do_log("[init] successful exit requested, complying\n");
-    exit(0);
-}
-
-// Request a failure exit from the system (for tests)
-void sigquit_handler(int)
-{
-    do_log("[init] failure exit requested, complying\n");
-    exit(1);
-}
-
 static Result<void> service_child(const Service& service, SharedPtr<os::File> output, SharedPtr<os::File> error,
                                   SharedPtr<os::File> input)
 {
@@ -75,12 +65,6 @@ static Result<void> service_child(const Service& service, SharedPtr<os::File> ou
     if (error) dup2(error->fd(), STDERR_FILENO);
     if (input) dup2(input->fd(), STDIN_FILENO);
 
-    if (service.user.has_value())
-    {
-        setgid(service.group.value());
-        setuid(service.user.value());
-    }
-
     if (service.environment.is_empty()) { TRY(os::Process::exec(args[0].view(), args.slice(), false)); }
     else
     {
@@ -191,10 +175,28 @@ static Result<void> load_service(const os::Path& path)
 
         if (parts[0].view() == "Command")
         {
+            if (!service.command.is_empty())
+            {
+                do_log("[init] 'Command' cannot be specified after 'Script' has already been set! (%s)\n",
+                       line.chars());
+                return {};
+            }
             service.command = move(parts[1]);
             continue;
         }
 
+        if (parts[0].view() == "Script")
+        {
+            if (!service.command.is_empty())
+            {
+                do_log("[init] 'Script' cannot be specified after 'Command' has already been set! (%s)\n",
+                       line.chars());
+                return {};
+            }
+            service.command = TRY(String::format("/bin/sh -- %s"_sv, parts[1].chars()));
+            continue;
+        }
+
         if (parts[0].view() == "Restart")
         {
             if (parts[1].view() == "true" || parts[1].view().to_uint().value_or(0) == 1)
@@ -230,15 +232,6 @@ static Result<void> load_service(const os::Path& path)
             continue;
         }
 
-        if (g_is_system && parts[0].view() == "User")
-        {
-            auto* pw = getpwnam(parts[1].chars());
-            if (!pw) continue;
-            service.user = pw->pw_uid;
-            service.group = pw->pw_gid;
-            continue;
-        }
-
         if (parts[0].view() == "Wait")
         {
             if (parts[1].view() == "true" || parts[1].view().to_uint().value_or(0) == 1)
@@ -261,7 +254,7 @@ static Result<void> load_service(const os::Path& path)
 
     if (service.command.is_empty())
     {
-        do_log("[init] service file is missing 'Command' entry, aborting!\n");
+        do_log("[init] service file is missing 'Command' or 'Script' entry, aborting!\n");
         return {};
     }
 
@@ -272,9 +265,9 @@ static Result<void> load_service(const os::Path& path)
     return {};
 }
 
-static Result<void> load_services(StringView path)
+static Result<void> load_services()
 {
-    auto dir = TRY(os::Directory::open(path));
+    auto dir = TRY(os::Directory::open("/etc/init"));
 
     auto services = TRY(dir->list_names(os::Directory::Filter::ParentAndBase));
     sort(services.begin(), services.end(), String::compare);
@@ -284,9 +277,9 @@ static Result<void> load_services(StringView path)
     return {};
 }
 
-static Result<void> start_services(StringView path)
+static Result<void> start_services()
 {
-    TRY(load_services(path));
+    TRY(load_services());
     for (auto& service : g_services)
     {
         do_log("[init] starting service %s\n", service.name.chars());
@@ -303,19 +296,6 @@ static Result<void> set_hostname()
     auto hostname = TRY(file->read_line());
     hostname.trim("\n");
 
-    if (hostname.is_empty())
-    {
-        do_log("[init] /etc/hostname is empty or invalid, keeping the default hostname\n");
-        return {};
-    }
-
-    Utf8StringDecoder decoder(hostname.chars());
-    if (decoder.code_points().has_error())
-    {
-        do_log("[init] /etc/hostname is not valid UTF-8, keeping the default hostname\n");
-        return {};
-    }
-
     if (sethostname(hostname.chars(), hostname.length()) < 0) return {};
 
     do_log("[init] successfully set system hostname to '%s'\n", hostname.chars());
@@ -330,16 +310,7 @@ static void mount_tmpfs()
     if (chmod("/tmp", 01777) < 0) exit(255);
 }
 
-static void mount_shmfs()
-{
-    if (mkdir("/dev/shm", 0755) < 0) exit(255);
-
-    if (mount("/dev/shm", "tmpfs", "tmpfs") < 0) exit(255);
-
-    if (chmod("/dev/shm", 01777) < 0) exit(255);
-}
-
-Result<int> sysinit(StringView path)
+int main()
 {
     if (getpid() != 1)
     {
@@ -347,35 +318,25 @@ Result<int> sysinit(StringView path)
         return 1;
     }
 
-    g_is_system = true;
-
     // Before this point, we don't even have an stdin, stdout and stderr. Set it up now so that child processes (and us)
     // can print stuff.
     stdin = fopen("/dev/console", "r");
     stdout = fopen("/dev/console", "w");
     stderr = fopen("/dev/console", "w");
 
-    TRY(os::Security::pledge("stdio rpath wpath cpath fattr host mount proc exec id", nullptr));
-
     mount_tmpfs();
-    mount_shmfs();
 
     umask(022);
 
-    g_init_log = fopen("/dev/uart0", "w");
-    check(g_init_log);
-    setlinebuf(g_init_log);
+    g_init_log = fopen("/dev/uart0", "w+");
     fcntl(fileno(g_init_log), F_SETFD, FD_CLOEXEC);
 
     set_hostname();
 
-    if (signal(SIGTERM, sigterm_handler) == SIG_ERR) do_log("[init] failed to register handler for SIGTERM\n");
-    if (signal(SIGQUIT, sigquit_handler) == SIG_ERR) do_log("[init] failed to register handler for SIGQUIT\n");
+    if (signal(SIGTERM, sigterm_handler) == SIG_ERR) do_log("[init] failed to register handler for SIGTERM");
+    if (signal(SIGQUIT, sigquit_handler) == SIG_ERR) do_log("[init] failed to register handler for SIGQUIT");
 
-    TRY(os::Security::pledge("stdio rpath wpath cpath proc exec id", nullptr));
-
-    if (path.is_empty()) path = "/etc/init";
-    start_services(path);
+    start_services();
 
     while (1)
     {
@@ -410,70 +371,3 @@ Result<int> sysinit(StringView path)
         }
     }
 }
-
-Result<int> user_init(StringView path)
-{
-    setpgid(0, 0);
-
-    g_init_log = fopen("/dev/uart0", "w");
-    check(g_init_log);
-    setlinebuf(g_init_log);
-    fcntl(fileno(g_init_log), F_SETFD, FD_CLOEXEC);
-
-    TRY(os::Security::pledge("stdio rpath wpath cpath proc exec", nullptr));
-
-    if (path.is_empty()) path = "/etc/user";
-    start_services(path);
-
-    TRY(os::Security::pledge("stdio rpath wpath proc exec", nullptr));
-
-    while (1)
-    {
-        int status;
-        auto rc = os::Process::wait(os::Process::ANY_CHILD, &status);
-        if (rc.has_error()) continue;
-
-        pid_t child = rc.release_value();
-
-        for (auto& service : g_services)
-        {
-            if (service.pid.has_value() && service.pid.value() == child)
-            {
-                if (WIFEXITED(status))
-                {
-                    do_log("[init] service %s exited with status %d\n", service.name.chars(), WEXITSTATUS(status));
-                }
-                else
-                {
-                    do_log("[init] service %s was terminated by signal %d\n", service.name.chars(), WTERMSIG(status));
-                }
-
-                if (service.restart)
-                {
-                    do_log("[init] restarting service %s\n", service.name.chars());
-
-                    start_service(service);
-                }
-
-                break;
-            }
-        }
-    }
-}
-
-Result<int> luna_main(int argc, char** argv)
-{
-    bool user;
-    StringView service_path;
-
-    os::ArgumentParser parser;
-    parser.add_description("The init system for Luna.");
-    parser.add_system_program_info("init"_sv);
-    parser.add_switch_argument(user, 'u', "user"_sv, "initialize a user session instead of the system");
-    parser.add_value_argument(service_path, 's', "service-path"_sv,
-                              "change the default service path (/etc/init or /etc/user)");
-    parser.parse(argc, argv);
-
-    if (user) return user_init(service_path);
-    return sysinit(service_path);
-}

apps/ipc-test.cpp

@@ -0,0 +1,45 @@
+#include <stdio.h>
+#include <string.h>
+#include <sys/wait.h>
+#include <unistd.h>
+
+int main()
+{
+    int pfds[2];
+
+    if (pipe(pfds) < 0)
+    {
+        perror("pipe");
+        return 1;
+    }
+
+    pid_t child = fork();
+    if (child == 0)
+    {
+        close(pfds[1]);
+
+        char buffer[4096];
+        size_t nread = read(pfds[0], buffer, sizeof(buffer) - 1);
+        buffer[nread] = 0;
+        close(pfds[0]);
+
+        puts(buffer);
+
+        return 0;
+    }
+    else if (child == -1)
+    {
+        perror("fork");
+        return 1;
+    }
+
+    close(pfds[0]);
+
+    const char* string = "Hello from a child process who just received this message from its parent!";
+    write(pfds[1], string, strlen(string));
+    close(pfds[1]);
+
+    wait(NULL);
+
+    return 0;
+}

apps/login.cpp

@@ -26,12 +26,11 @@ Result<int> luna_main(int argc, char** argv)
 
     String name;
 
-    setpgid(0, 0);
-    signal(SIGTTOU, SIG_IGN);
-    if (isatty(STDIN_FILENO)) tcsetpgrp(STDIN_FILENO, getpgid(0));
-
     if (username.is_empty())
     {
+        signal(SIGTTOU, SIG_IGN);
+
+        if (isatty(STDIN_FILENO)) tcsetpgrp(STDIN_FILENO, getpgid(0));
 
         auto input = os::File::standard_input();
 
@@ -45,7 +44,7 @@ Result<int> luna_main(int argc, char** argv)
         username = name.view();
     }
 
-    execl("/usr/bin/su", "login", "-lp", "--", username.chars(), nullptr);
+    execl("/bin/su", "login", "-lp", "--", username.chars(), nullptr);
 
     perror("su");
     return 1;

apps/ls.cpp

@@ -8,7 +8,6 @@
 #include <os/FileSystem.h>
 #include <os/Mode.h>
 #include <pwd.h>
-#include <unistd.h>
 
 void find_user_and_group(struct stat& st, StringView& owner, StringView& group)
 {
@@ -48,46 +47,18 @@ int sort_reverse(const os::Directory::Entry* a, const os::Directory::Entry* b)
     return 0;
 }
 
-#define RESET_COLORS "\x1b[m"
-#define SYMLINK_COLOR "\x1b[36m"
-#define FILE_COLOR "\x1b[1;32m"
-#define DIR_COLOR "\x1b[1;34m"
-#define SOCKET_COLOR "\x1b[33m"
-#define SPECIAL_COLOR "\x1b[35m"
-#define STICKY_COLOR "\x1b[30;1;42m"
-#define SETUID_COLOR "\x1b[30;1;41m"
-#define EXEC_COLOR "\x1b[1;31m"
-
-static const char* file_type_color(const os::Directory::Entry& entry)
-{
-    if (entry.mode & S_ISVTX) return STICKY_COLOR;
-    if (entry.mode & S_ISUID || entry.mode & S_ISGID) return SETUID_COLOR;
-
-    switch (entry.mode & S_IFMT)
-    {
-    case S_IFREG: return entry.mode & S_IXUSR ? EXEC_COLOR : FILE_COLOR;
-    case S_IFDIR: return DIR_COLOR;
-    case S_IFLNK: return SYMLINK_COLOR;
-    case S_IFSOCK: return SOCKET_COLOR;
-    default: return SPECIAL_COLOR;
-    }
-}
-
-static Result<String> entry_join(const Vector<os::Directory::Entry>& vec, StringView delim, bool colors)
+static Result<String> entry_join(const Vector<os::Directory::Entry>& vec, StringView delim)
 {
     if (vec.size() == 0) return String {};
+    if (vec.size() == 1) return vec[0].name.clone();
 
     StringBuilder sb;
-    if (colors) TRY(sb.add(StringView { file_type_color(vec[0]) }));
     TRY(sb.add(vec[0].name));
-    if (colors) TRY(sb.add(StringView { RESET_COLORS }));
 
     for (usize i = 1; i < vec.size(); i++)
     {
         TRY(sb.add(delim));
-        if (colors) TRY(sb.add(StringView { file_type_color(vec[i]) }));
         TRY(sb.add(vec[i].name));
-        if (colors) TRY(sb.add(StringView { RESET_COLORS }));
     }
 
     return sb.string();
@@ -104,7 +75,6 @@ Result<int> luna_main(int argc, char** argv)
     bool follow_symlink_args { false };
     bool one_per_line { false };
     bool list_directories { false };
-    bool no_colors { false };
 
     StringView sort_type { "name" };
 
@@ -128,8 +98,6 @@ Result<int> luna_main(int argc, char** argv)
     parser.add_switch_argument(list_directories, 'd', "directory"_sv, "list directories instead of their contents"_sv);
     parser.add_value_argument(sort_type, ' ', "sort"_sv, "sort by name, size or time"_sv);
     parser.add_switch_argument(reverse_sort, 'r', "reverse"_sv, "reverse order while sorting"_sv);
-    parser.add_switch_argument(no_colors, ' ', "no-colors"_sv,
-                               "disable coloring of output (defaults to true when not in a TTY)"_sv);
     parser.parse(argc, argv);
 
     Vector<os::Directory::Entry> files;
@@ -186,13 +154,11 @@ Result<int> luna_main(int argc, char** argv)
 
     if (!long_list)
     {
-        auto list = TRY(entry_join(files, one_per_line ? "\n"_sv : " "_sv, !no_colors && isatty(STDIN_FILENO)));
+        auto list = TRY(entry_join(files, one_per_line ? "\n"_sv : " "_sv));
         if (!list.is_empty()) os::println("%s", list.chars());
     }
     else
     {
-        bool colors = !no_colors && isatty(STDIN_FILENO);
-
         for (const auto& file : files)
         {
             struct stat st;
@@ -210,32 +176,14 @@ Result<int> luna_main(int argc, char** argv)
 
             if (!human_readable && !si)
             {
-                if (colors)
-                {
-                    os::println("%s %u %4s %4s %10lu %s%s" RESET_COLORS "%s" SYMLINK_COLOR "%s" RESET_COLORS,
-                                formatted_mode, st.st_nlink, owner.chars(), group.chars(), st.st_size,
-                                file_type_color(file), file.name.chars(), link.is_empty() ? "" : " -> ", link.chars());
-                }
-                else
-                {
-                    os::println("%s %u %4s %4s %10lu %s%s%s", formatted_mode, st.st_nlink, owner.chars(), group.chars(),
-                                st.st_size, file.name.chars(), link.is_empty() ? "" : " -> ", link.chars());
-                }
+                os::println("%s %u %4s %4s %10lu %s%s%s", formatted_mode, st.st_nlink, owner.chars(), group.chars(),
+                            st.st_size, file.name.chars(), link.is_empty() ? "" : " -> ", link.chars());
             }
             else
             {
                 auto size = TRY(to_dynamic_unit(st.st_size, 10, false, si ? Unit::SI : Unit::Binary, false));
-                if (colors)
-                {
-                    os::println("%s %u %4s %4s %6s %s%s" RESET_COLORS "%s" SYMLINK_COLOR "%s" RESET_COLORS,
-                                formatted_mode, st.st_nlink, owner.chars(), group.chars(), size.chars(),
-                                file_type_color(file), file.name.chars(), link.is_empty() ? "" : " -> ", link.chars());
-                }
-                else
-                {
-                    os::println("%s %u %4s %4s %6s %s%s%s", formatted_mode, st.st_nlink, owner.chars(), group.chars(),
-                                size.chars(), file.name.chars(), link.is_empty() ? "" : " -> ", link.chars());
-                }
+                os::println("%s %u %4s %4s %6s %s%s%s", formatted_mode, st.st_nlink, owner.chars(), group.chars(),
+                            size.chars(), file.name.chars(), link.is_empty() ? "" : " -> ", link.chars());
             }
         }
     }

apps/pivot_root.cpp

@@ -0,0 +1,19 @@
+#include <os/ArgumentParser.h>
+#include <sys/syscall.h>
+#include <unistd.h>
+
+Result<int> luna_main(int argc, char** argv)
+{
+    StringView new_root;
+    StringView put_old;
+
+    os::ArgumentParser parser;
+    parser.add_description("Move the current root directory to another directory and replace it with another mount.");
+    parser.add_system_program_info("pivot_root"_sv);
+    parser.add_positional_argument(new_root, "new_root", true);
+    parser.add_positional_argument(put_old, "put_old", true);
+    parser.parse(argc, argv);
+
+    long rc = syscall(SYS_pivot_root, new_root.chars(), put_old.chars());
+    return Result<int>::from_syscall(rc);
+}

apps/preinit.cpp

@@ -72,9 +72,13 @@ int main()
     // Now, mount the /dev file system on the new root.
     mount_devfs();
 
-    char* argv[] = { "/usr/bin/init", nullptr };
+    /*setenv("PATH", "/bin:/usr/bin", 1);
+    char* argv[] = { "init", nullptr };
     char* envp[] = { nullptr };
-    execve(argv[0], argv, envp);
+    execvpe(argv[0], argv, envp);*/
+
+    char* argv[] = { "/usr/bin/init", nullptr };
+    execv(argv[0], argv);
 
     fail_errno("Failed to execute init");
 

apps/rm.cpp

@@ -1,35 +1,10 @@
 #include <os/ArgumentParser.h>
-#include <os/Directory.h>
-#include <os/File.h>
 #include <os/FileSystem.h>
 
-Result<void> remove_wrapper(const os::Path& path, bool verbose)
-{
-    TRY(os::FileSystem::remove(path));
-    if (verbose) os::println("removed '%s'", path.name().chars());
-    return {};
-}
-
-Result<void> remove_tree(const os::Path& path, bool verbose)
-{
-    auto rc = remove_wrapper(path, verbose);
-    if (!rc.has_error()) return {};
-    if (rc.error() != ENOTEMPTY) return rc.release_error();
-
-    auto dir = TRY(os::Directory::open(path));
-
-    Vector<String> entries = TRY(dir->list_names(os::Directory::Filter::ParentAndBase));
-
-    for (const auto& entry : entries) { TRY(remove_tree({ dir->fd(), entry.view() }, verbose)); }
-
-    return remove_wrapper(path, verbose);
-}
-
 Result<int> luna_main(int argc, char** argv)
 {
     StringView path;
     bool recursive;
-    bool verbose;
 
     os::ArgumentParser parser;
     parser.add_description("Remove a path from the file system."_sv);
@@ -37,12 +12,11 @@ Result<int> luna_main(int argc, char** argv)
     parser.add_positional_argument(path, "path"_sv, true);
     parser.add_switch_argument(recursive, 'r', "recursive"_sv,
                                "remove a directory recursively (by default, rm removes only empty directories)"_sv);
-    parser.add_switch_argument(verbose, 'v', "verbose"_sv, "log every removed file and directory"_sv);
     parser.parse(argc, argv);
 
-    if (!recursive) TRY(remove_wrapper(path, verbose));
+    if (!recursive) TRY(os::FileSystem::remove(path));
     else
-        TRY(remove_tree(path, verbose));
+        TRY(os::FileSystem::remove_tree(path));
 
     return 0;
 }

apps/signal-test.cpp

@@ -0,0 +1,23 @@
+#include <signal.h>
+#include <stdio.h>
+#include <string.h>
+
+void handler(int)
+{
+    puts("I caught a segfault!");
+}
+
+int main()
+{
+    struct sigaction sa;
+    sa.sa_handler = handler;
+    sigemptyset(&sa.sa_mask);
+    sa.sa_flags = SA_RESETHAND;
+    sigaction(SIGSEGV, &sa, NULL);
+
+#pragma GCC diagnostic ignored "-Wnonnull"
+    char* str = nullptr;
+    memset(str, 0, 2);
+
+    return 0;
+}

apps/stat.cpp

@@ -2,7 +2,7 @@
 #include <os/FileSystem.h>
 #include <os/Mode.h>
 #include <stdio.h>
-#include <time.h>
+#include <sys/stat.h>
 
 static const char* file_type(mode_t mode)
 {
@@ -24,7 +24,7 @@ Result<int> luna_main(int argc, char** argv)
     bool follow_symlinks { false };
 
     os::ArgumentParser parser;
-    parser.add_description("Display file metadata.");
+    parser.add_description("Display file status.");
     parser.add_system_program_info("stat"_sv);
     parser.add_positional_argument(path, "path", true);
     parser.add_switch_argument(follow_symlinks, 'L', "dereference"_sv, "follow symlinks");
@@ -36,20 +36,10 @@ Result<int> luna_main(int argc, char** argv)
     char buf[11];
     os::format_mode(st.st_mode, buf);
 
-    char atime[256];
-    strftime(atime, sizeof(atime), "%Y-%m-%d %H:%M:%S", gmtime(&st.st_atim.tv_sec));
-    char mtime[256];
-    strftime(mtime, sizeof(mtime), "%Y-%m-%d %H:%M:%S", gmtime(&st.st_mtim.tv_sec));
-    char ctime[256];
-    strftime(ctime, sizeof(ctime), "%Y-%m-%d %H:%M:%S", gmtime(&st.st_ctim.tv_sec));
-
-    printf("  File: %s\n", path.chars());
-    printf("  Size: %zu (%s)\n", st.st_size, file_type(st.st_mode));
-    printf(" Inode: %lu Links: %lu\n", st.st_ino, st.st_nlink);
-    printf("  Mode: (%#o/%s) UID: %u GID: %u\n", st.st_mode & ~S_IFMT, buf, st.st_uid, st.st_gid);
-    printf("Access: %s.%.9ld\n", atime, st.st_atim.tv_nsec);
-    printf("Modify: %s.%.9ld\n", mtime, st.st_mtim.tv_nsec);
-    printf("Change: %s.%.9ld\n", ctime, st.st_ctim.tv_nsec);
+    printf(" File: %s\n", path.chars());
+    printf(" Size: %zu (%s)\n", st.st_size, file_type(st.st_mode));
+    printf("Inode: %lu Links: %lu\n", st.st_ino, st.st_nlink);
+    printf(" Mode: (%#o/%s) UID: %u GID: %u\n", st.st_mode & ~S_IFMT, buf, st.st_uid, st.st_gid);
 
     return 0;
 }

apps/su.cpp

@@ -127,7 +127,7 @@ Result<int> luna_main(int argc, char** argv)
     {
         chdir(entry->pw_dir);
         clearenv();
-        setenv("PATH", "/usr/bin:/usr/local/bin", 1);
+        setenv("PATH", "/bin:/sbin", 1);
         setpgid(0, 0);
     }
 

apps/sysfuzz.cpp

@@ -1,4 +1,4 @@
-#include <moon/Syscall.h>
+#include <luna/Syscall.h>
 #include <os/ArgumentParser.h>
 #include <stdio.h>
 #include <stdlib.h>

apps/touch.cpp

@@ -1,47 +0,0 @@
-#include <errno.h>
-#include <os/ArgumentParser.h>
-#include <os/File.h>
-#include <sys/stat.h>
-
-Result<int> luna_main(int argc, char** argv)
-{
-    Vector<StringView> files;
-
-    bool only_atime { false };
-    bool only_mtime { false };
-    bool no_create { false };
-    bool no_dereference { false };
-
-    os::ArgumentParser parser;
-    parser.add_description("Update the access and modification times of files."_sv);
-    parser.add_system_program_info("touch"_sv);
-    parser.set_vector_argument(files, "files", true);
-    parser.add_switch_argument(only_atime, 'a', ""_sv, "change only the access time"_sv);
-    parser.add_switch_argument(no_create, 'c', "no-create"_sv, "do not create new files"_sv);
-    parser.add_switch_argument(no_dereference, 'h', "no-dereference"_sv, "do not follow symbolic links"_sv);
-    parser.add_switch_argument(only_mtime, 'm', ""_sv, "change only the modification time"_sv);
-    TRY(parser.parse(argc, argv));
-
-    if (only_atime && only_mtime)
-    {
-        os::eprintln("%s: only one of -a and -m can be specified.", argv[0]);
-        parser.short_usage(argv[0]);
-    }
-
-    struct timespec times[2] = {
-        { .tv_sec = 0, .tv_nsec = only_mtime ? UTIME_OMIT : UTIME_NOW },
-        { .tv_sec = 0, .tv_nsec = only_atime ? UTIME_OMIT : UTIME_NOW },
-    };
-
-    for (auto& filename : files)
-    {
-        SharedPtr<os::File> file;
-        if (no_create) file = TRY(os::File::open(filename, os::File::ReadOnly));
-        else
-            file = TRY(os::File::open_or_create(filename, os::File::ReadOnly));
-
-        if (futimens(file->fd(), times, no_dereference ? AT_SYMLINK_NOFOLLOW : 0) < 0) return err(errno);
-    }
-
-    return 0;
-}

base/etc/init/00-home

@@ -1,4 +1,4 @@
 Name=mount-home
 Description=Mount the user's home directory on a writable filesystem.
-Command=/etc/startup/mount-home.sh
+Script=/etc/startup/mount-home.sh
 Wait=true

base/etc/passwd

@@ -1,2 +1,2 @@
-root:toor:0:0:Administrator:/:/usr/bin/sh
-selene:moon:1000:1000:User:/home/selene:/usr/bin/sh
+root:toor:0:0:Administrator:/:/bin/sh
+selene:moon:1000:1000:User:/home/selene:/bin/sh

base/etc/skel/welcome

@@ -1,14 +0,0 @@
-Welcome to the Luna operating system!
-You are running on the default user account, selene.
-
-If you are familiar with Unix-style operating systems (like Linux or *BSD), you should be able to use the Luna terminal without much problems.
-
-Following the traditional Unix filesystem structure, programs are installed in /usr/bin (/bin is a symlink to /usr/bin). The command `ls /bin` will show all commands available on your current Luna installation.
-
-Currently, because of driver limitations, the root file system is mounted read-only. Your home folder is writable, but volatile; it is created and populated on boot, and its contents will vanish after a reboot.
-
-The system is booted using the 'init' program. You can read its configuration files in the /etc/init directory to learn more about the boot process.
-
-Luna is free software, released under the BSD-2-Clause license. The license is included in the LICENSE file in your home directory.
-
-View the source code and read more about Luna at https://git.cloudapio.eu/apio/Luna.

base/etc/startup/mount-home.sh

@@ -1,6 +1,2 @@
-#!/bin/sh
-# Create and populate a volatile home directory.
 mount -t tmpfs tmpfs /home/selene
 chown selene:selene /home/selene
-cp /etc/skel/welcome /home/selene/
-cp /etc/skel/LICENSE /home/selene/

kernel/CMakeLists.txt

@@ -6,7 +6,6 @@ set(SOURCES
     ${HEADERS}
     src/main.cpp
     src/Log.cpp
-    src/Pledge.cpp
     src/cxxabi.cpp
     src/video/Framebuffer.cpp
     src/video/TextConsole.cpp
@@ -15,7 +14,6 @@ set(SOURCES
     src/memory/KernelVM.cpp
     src/memory/AddressSpace.cpp
     src/memory/MemoryMap.cpp
-    src/memory/SharedMemory.cpp
     src/boot/Init.cpp
     src/arch/Serial.cpp
     src/arch/Timer.cpp
@@ -43,36 +41,24 @@ set(SOURCES
     src/sys/mount.cpp
     src/sys/resource.cpp
     src/sys/signal.cpp
-    src/sys/socket.cpp
-    src/sys/poll.cpp
-    src/sys/alarm.cpp
-    src/sys/pledge.cpp
-    src/sys/memstat.cpp
     src/fs/VFS.cpp
     src/fs/Pipe.cpp
     src/fs/Mount.cpp
     src/fs/MBR.cpp
     src/fs/GPT.cpp
-    src/fs/StorageCache.cpp
-    src/net/UnixSocket.cpp
     src/fs/tmpfs/FileSystem.cpp
     src/fs/tmpfs/Inode.cpp
     src/fs/ext2/FileSystem.cpp
     src/fs/ext2/Inode.cpp
     src/fs/devices/DeviceRegistry.cpp
-    src/fs/devices/BlockDevice.cpp
     src/fs/devices/NullDevice.cpp
     src/fs/devices/ZeroDevice.cpp
     src/fs/devices/FullDevice.cpp
     src/fs/devices/ConsoleDevice.cpp
     src/fs/devices/FramebufferDevice.cpp
     src/fs/devices/UARTDevice.cpp
-    src/fs/devices/MouseDevice.cpp
-    src/fs/devices/KeyboardDevice.cpp
     src/fs/InitRD.cpp
-    src/binfmt/ELF.cpp
-    src/binfmt/BinaryFormat.cpp
-    src/binfmt/Script.cpp
+    src/thread/ELF.cpp
 )
 
 if("${LUNA_ARCH}" MATCHES "x86_64")
@@ -90,7 +76,6 @@ if("${LUNA_ARCH}" MATCHES "x86_64")
         src/arch/x86_64/init/GDT.cpp
         src/arch/x86_64/init/IDT.cpp
         src/arch/x86_64/init/PIC.cpp
-        src/arch/x86_64/PS2Mouse.cpp
     )
 endif()
 

kernel/config.cmake.template

@@ -15,7 +15,3 @@
 # control characters, leading/trailing spaces, problematic characters and invalid UTF-8). Keep in mind that this restriction
 # is only enforced when creating files; existing files with such illegal filenames are parsed correctly and fully usable.
 # target_compile_definitions(moon PRIVATE MOON_DISABLE_FILENAME_RESTRICTIONS)
-
-# Uncomment the line below to make the kernel also calculate stack traces for userspace addresses on program crashes. 
-# This can aid in debugging, but makes the kernel more unstable as stack tracing will access arbitrary userspace memory.
-# target_compile_definitions(moon PRIVATE MOON_ENABLE_USERSPACE_STACK_TRACES)

kernel/debug.cmake

@@ -11,5 +11,4 @@ target_compile_definitions(moon PRIVATE EXT2_DEBUG)
 target_compile_definitions(moon PRIVATE DEVICE_REGISTRY_DEBUG)
 target_compile_definitions(moon PRIVATE FORK_DEBUG)
 target_compile_definitions(moon PRIVATE MOUNT_DEBUG)
-target_compile_definitions(moon PRIVATE CACHE_DEBUG)
 target_compile_options(moon PRIVATE -fsanitize=undefined)

kernel/src/Log.cpp

@@ -27,8 +27,8 @@ static void log_serial(LogLevel level, const char* format, va_list origin)
     va_list ap;
     va_copy(ap, origin);
 
-    /*const SafeScopeLock lock { g_serial_lock };
-    if (!lock.did_succeed()) return;*/
+    const SafeScopeLock lock { g_serial_lock };
+    if (!lock.did_succeed()) return;
 
     Serial::printf("\x1b[%sm"
                    "%c"
@@ -61,7 +61,7 @@ static void log_text_console(LogLevel level, const char* format, va_list origin)
     va_list ap;
     va_copy(ap, origin);
 
-    /*const ScopeLock lock { g_console_lock };*/
+    const ScopeLock lock { g_console_lock };
 
     const u32 original_foreground = TextConsole::foreground();
     const u32 original_background = TextConsole::background();

kernel/src/Pledge.cpp

@@ -1,69 +0,0 @@
-#include "Pledge.h"
-#include "Log.h"
-#include "memory/MemoryManager.h"
-
-static const char* promise_names[] = {
-#define __enumerate(promise) #promise,
-    enumerate_promises(__enumerate)
-#undef __enumerate
-};
-
-Result<void> check_pledge(Thread* thread, Promise promise)
-{
-    // Thread has not called pledge().
-    if (thread->promises < 0) return {};
-    int mask = (1 << (int)promise);
-    if ((thread->promises & mask) != mask)
-    {
-        kerrorln("Pledge violation in thread %d! Has not pledged %s", thread->id, promise_names[(int)promise]);
-        if (thread->promises & (1 << (int)Promise::p_error)) return err(ENOSYS);
-
-        // Kill this thread with an uncatchable SIGABRT. For this, we reset the disposition of SIGABRT to the default
-        // (dump core). We could just kill the thread here and be done, but that discards anything on the current stack,
-        // which means that some destructors might not be called. Instead, leave the job to the next call of
-        // Thread::process_pending_signals().
-        thread->signal_handlers[SIGABRT - 1].sa_handler = SIG_DFL;
-
-        // Unblock SIGABRT.
-        thread->signal_mask.set(SIGABRT - 1, false);
-
-        // If there are any other pending signals, they might be processed before SIGABRT. Avoid that by resetting the
-        // thread's pending signals.
-        thread->pending_signals.clear();
-
-        thread->send_signal(SIGABRT);
-
-        // This should never arrive to userspace, unless we're init and have ignored SIGABRT.
-        return err(ENOSYS);
-    }
-
-    return {};
-}
-
-Result<int> parse_promises(u64 pledge)
-{
-    if (!pledge) return -1;
-
-    auto text = TRY(MemoryManager::strdup_from_user(pledge));
-    if (text.is_empty()) return 0;
-
-    auto promises = TRY(text.split(" "));
-
-    int result = 0;
-    for (const auto& promise : promises)
-    {
-        for (int i = 0; i < (int)Promise::num_promises; i++)
-        {
-            if (promise.view() == promise_names[i])
-            {
-                result |= (1 << i);
-                goto found;
-            }
-        }
-        return err(EINVAL);
-    found:
-        continue;
-    }
-
-    return result;
-}

kernel/src/Pledge.h

@@ -1,19 +0,0 @@
-#pragma once
-#include "thread/Thread.h"
-#include <luna/Result.h>
-
-#define enumerate_promises(_p)                                                                                         \
-    _p(stdio) _p(rpath) _p(wpath) _p(cpath) _p(fattr) _p(chown) _p(unix) _p(tty) _p(proc) _p(exec) _p(prot_exec)       \
-        _p(id) _p(mount) _p(host) _p(error)
-
-enum class Promise
-{
-#define __enumerate(promise) p_##promise,
-    enumerate_promises(__enumerate)
-#undef __enumerate
-        num_promises,
-};
-
-Result<void> check_pledge(Thread* thread, Promise promise);
-
-Result<int> parse_promises(u64 pledge);

kernel/src/api/Keyboard.h

@@ -1,131 +0,0 @@
-#pragma once
-#include <luna/Types.h>
-
-namespace moon
-{
-    enum KeyCode : u8
-    {
-        // Function keys
-        K_F1,
-        K_F2,
-        K_F3,
-        K_F4,
-        K_F5,
-        K_F6,
-        K_F7,
-        K_F8,
-        K_F9,
-        K_F10,
-        K_F11,
-        K_F12,
-        // System keys
-        K_Esc,
-        K_PrtScr,
-        K_Pause,
-        K_Super,
-        K_Menu,
-        // Modifier keys
-        K_LeftShift,
-        K_RightShift,
-        K_LeftAlt,
-        K_RightAlt, // or AltGr on some keyboards
-        K_LeftControl,
-        K_RightControl,
-        // Navigation keys
-        K_Tab,
-        K_Home,
-        K_End,
-        K_PageUp,
-        K_PageDown,
-        K_RightArrow,
-        K_LeftArrow,
-        K_UpArrow,
-        K_DownArrow,
-        // Editing keys
-        K_Backspace,
-        K_Enter,
-        K_Insert,
-        K_Delete,
-        K_KeypadEnter,
-        // Lock keys
-        K_ScrollLock,
-        K_CapsLock,
-        K_NumLock,
-        // Keypad keys
-        K_Keypad0,
-        K_Keypad1,
-        K_Keypad2,
-        K_Keypad3,
-        K_Keypad4,
-        K_Keypad5,
-        K_Keypad6,
-        K_Keypad7,
-        K_Keypad8,
-        K_Keypad9,
-        K_KeypadDot,
-        K_KeypadPlus,
-        K_KeypadMinus,
-        K_KeypadMul,
-        K_KeypadDiv,
-        // Character keys (depending on keyboard layout), examples in US QWERTY
-        K_CH00, // `
-        K_CH01, // 1
-        K_CH02, // 2
-        K_CH03, // 3
-        K_CH04, // 4
-        K_CH05, // 5
-        K_CH06, // 6
-        K_CH07, // 7
-        K_CH08, // 8
-        K_CH09, // 9
-        K_CH10, // 0
-        K_CH11, // -
-        K_CH12, // =
-        K_CH13, // Q
-        K_CH14, // W
-        K_CH15, // E
-        K_CH16, // R
-        K_CH17, // T
-        K_CH18, // Y
-        K_CH19, // U
-        K_CH20, // I
-        K_CH21, // O
-        K_CH22, // P
-        K_CH23, // [
-        K_CH24, // ]
-        K_CH25, // A
-        K_CH26, // S
-        K_CH27, // D
-        K_CH28, // F
-        K_CH29, // G
-        K_CH30, // H
-        K_CH31, // J
-        K_CH32, // K
-        K_CH33, // L
-        K_CH34, // ;
-        K_CH35, // '
-        K_CH36, // #
-        K_CH37, // Backslash
-        K_CH38, // Z
-        K_CH39, // X
-        K_CH40, // C
-        K_CH41, // V
-        K_CH42, // B
-        K_CH43, // N
-        K_CH44, // M
-        K_CH45, // ,
-        K_CH46, // .
-        K_CH47, // /
-        K_CH48, // Space
-        // Unknown key
-        K_Unknown,
-    };
-
-    struct [[gnu::packed]] KeyboardPacket
-    {
-        u8 key;
-        bool released;
-    };
-
-    static_assert(sizeof(KeyboardPacket) == 2);
-}

kernel/src/api/Mouse.h

@@ -1,22 +0,0 @@
-#pragma once
-#include <luna/Types.h>
-
-namespace moon
-{
-    enum MouseButton
-    {
-        Left = (1 << 0),
-        Middle = (1 << 1),
-        Right = (1 << 2),
-    };
-
-    struct [[gnu::packed]] MousePacket
-    {
-        i16 xdelta;
-        i16 ydelta;
-        u8 buttons;
-        u8 _padding[3];
-    };
-
-    static_assert(sizeof(MousePacket) == 8);
-}

kernel/src/arch/Keyboard.h

@@ -1,10 +1,9 @@
 #pragma once
-#include "api/Keyboard.h"
 #include <luna/Option.h>
 
 namespace Keyboard
 {
-    struct TTYKeyboardState
+    struct KeyboardState
     {
         bool ignore_next { false };
         bool left_shift { false };
@@ -13,12 +12,5 @@ namespace Keyboard
         bool capslock { false };
     };
 
-    struct KeyboardState
-    {
-        bool ignore_next { false };
-    };
-
-    Option<char> decode_scancode_tty(u8 scancode, TTYKeyboardState& state);
-
-    Option<moon::KeyboardPacket> decode_scancode(u8 scancode, KeyboardState& state);
+    Option<char> decode_scancode_tty(u8 scancode, KeyboardState& state);
 }

kernel/src/arch/MMU.h

@@ -29,15 +29,15 @@ namespace MMU
 
     u64 translate_physical_address(u64 phys);
 
+    Result<void> map(u64 virt, u64 phys, int flags, UseHugePages use_huge_pages);
+    Result<u64> unmap(u64 virt);
+    Result<u64> get_physical(u64 virt);
+    Result<int> get_flags(u64 virt);
+    Result<void> remap(u64 virt, int flags);
+
     void switch_page_directory(PageDirectory* dir);
     PageDirectory* get_page_directory();
 
-    Result<void> map(u64 virt, u64 phys, int flags, UseHugePages use_huge_pages, PageDirectory* directory = nullptr);
-    Result<u64> unmap(u64 virt, PageDirectory* directory = nullptr);
-    Result<u64> get_physical(u64 virt, PageDirectory* directory = nullptr);
-    Result<int> get_flags(u64 virt, PageDirectory* directory = nullptr);
-    Result<void> remap(u64 virt, int flags);
-
     void flush_all();
 
     Result<PageDirectory*> create_page_directory_for_userspace();

kernel/src/arch/x86_64/CPU.cpp

@@ -1,12 +1,10 @@
 #include "arch/CPU.h"
 #include "Log.h"
-#include "api/Mouse.h"
 #include "arch/Keyboard.h"
 #include "arch/Timer.h"
 #include "arch/x86_64/CPU.h"
 #include "arch/x86_64/IO.h"
 #include "fs/devices/ConsoleDevice.h"
-#include "fs/devices/MouseDevice.h"
 #include "memory/MemoryManager.h"
 #include "sys/Syscall.h"
 #include "thread/Scheduler.h"
@@ -29,9 +27,8 @@ extern void change_pic_masks(u8 pic1_mask, u8 pic2_mask);
 extern void pic_eoi(unsigned char irq);
 extern void pic_eoi(Registers* regs);
 extern void setup_idt();
-extern void init_mouse();
 
-Thread* g_io_thread;
+static Thread* g_io_thread;
 
 typedef void (*interrupt_handler_t)(Registers*, void*);
 
@@ -137,35 +134,27 @@ extern "C" void handle_x86_exception(Registers* regs)
     case 19: FIXME_UNHANDLED_INTERRUPT("SIMD floating-point exception");
     case 20: FIXME_UNHANDLED_INTERRUPT("Virtualization exception");
     case 21: FIXME_UNHANDLED_INTERRUPT("Control-protection exception");
-    default: fail("Caught invalid reserved exception or #DF/#MC, which shouldn't call handle_x86_exception");
+    default: FIXME_UNHANDLED_INTERRUPT("Reserved exception or #DF/#MC, which shouldn't call handle_x86_exception");
     }
 }
 
 CircularQueue<u8, 60> scancode_queue;
-CircularQueue<moon::MousePacket, 20> mouse_queue;
 
 void io_thread()
 {
     while (true)
     {
         u8 scancode;
-        moon::MousePacket packet;
+        while (!scancode_queue.try_pop(scancode)) kernel_wait_for_event();
 
-        if (scancode_queue.try_pop(scancode)) { ConsoleDevice::did_press_or_release_key(scancode); }
-        else if (mouse_queue.try_pop(packet)) { MouseDevice::add_mouse_event(packet); }
-        else
-            kernel_wait_for_event();
+        ConsoleDevice::did_press_or_release_key(scancode);
     }
 }
 
 static void timer_interrupt(Registers* regs, void*)
 {
     Timer::tick();
-    if (should_invoke_scheduler())
-    {
-        Scheduler::invoke(regs);
-        TextConsole::tick_cursor();
-    }
+    if (should_invoke_scheduler()) Scheduler::invoke(regs);
 }
 
 static void keyboard_interrupt(Registers*, void*)
@@ -272,8 +261,6 @@ namespace CPU
 
         remap_pic();
 
-        init_mouse();
-
         sync_interrupts();
     }
 
@@ -339,10 +326,7 @@ namespace CPU
     static void backtrace_impl(u64 base_pointer, void (*callback)(u64, void*), void* arg)
     {
         StackFrame* current_frame = (StackFrame*)base_pointer;
-        while (current_frame &&
-#ifndef MOON_ENABLE_USERSPACE_STACK_TRACES
-               (u64)current_frame >= 0xFFFF'FFFF'8000'0000 &&
-#endif
+        while (current_frame && (u64)current_frame >= 0xFFFF'FFFF'8000'0000 &&
                MemoryManager::validate_access(current_frame, sizeof(*current_frame), MemoryManager::DEFAULT_ACCESS) &&
                current_frame->instruction)
         {

kernel/src/arch/x86_64/Keyboard.cpp

@@ -19,20 +19,21 @@ constexpr u8 RIGHT_SHIFT = 0x36;
 constexpr u8 CAPS_LOCK = 0x3a;
 
 constexpr u8 LEFT_CONTROL = 0x1D;
+constexpr u8 TAB = 0x0F;
 constexpr u8 LEFT_ALT = 0x38;
 constexpr u8 F11 = 0x57;
 constexpr u8 F12 = 0x58;
 
 static bool should_ignore_key(u8 scancode)
 {
-    return (scancode > 0x3A && scancode < 0x47) || scancode == F11 || scancode == F12;
+    return (scancode > 0x3A && scancode < 0x47) || scancode == TAB || scancode == F11 || scancode == F12;
 }
 
 constexpr char key_table[] = {
     '\0',
     '\1', // escape
     '1',  '2', '3', '4', '5', '6', '7', '8', '9', '0', '-',  '=', '\b',
-    '\t', // tab
+    '\0', // tab
     'q',  'w', 'e', 'r', 't', 'y', 'u', 'i', 'o', 'p', '[',  ']', '\n',
     '\0', // left ctrl
     'a',  's', 'd', 'f', 'g', 'h', 'j', 'k', 'l', ';', '\'', '`',
@@ -74,7 +75,7 @@ constexpr char shifted_key_table[] = {
     '\0',
     '\1', // escape
     '!',  '@', '#', '$', '%', '^', '&', '*', '(', ')', '_', '+', '\b',
-    '\t', // tab
+    '\0', // tab
     'Q',  'W', 'E', 'R', 'T', 'Y', 'U', 'I', 'O', 'P', '{', '}', '\n',
     '\0', // left ctrl
     'A',  'S', 'D', 'F', 'G', 'H', 'J', 'K', 'L', ':', '"', '~',
@@ -112,71 +113,7 @@ constexpr char shifted_key_table[] = {
     '.',  // keypad .
 };
 
-using namespace moon;
-
-constexpr KeyCode keycode_table[] = {
-    K_Unknown,    K_Esc,
-    K_CH01, // 1
-    K_CH02, // 2
-    K_CH03, // 3
-    K_CH04, // 4
-    K_CH05, // 5
-    K_CH06, // 6
-    K_CH07, // 7
-    K_CH08, // 8
-    K_CH09, // 9
-    K_CH10, // 0
-    K_CH11, // -
-    K_CH12, // =
-    K_Backspace,  K_Tab,
-    K_CH13, // Q
-    K_CH14, // W
-    K_CH15, // E
-    K_CH16, // R
-    K_CH17, // T
-    K_CH18, // Y
-    K_CH19, // U
-    K_CH20, // I
-    K_CH21, // O
-    K_CH22, // P
-    K_CH23, // [
-    K_CH24, // ]
-    K_Enter,      K_LeftControl,
-    K_CH25, // A
-    K_CH26, // S
-    K_CH27, // D
-    K_CH28, // F
-    K_CH29, // G
-    K_CH30, // H
-    K_CH31, // J
-    K_CH32, // K
-    K_CH33, // L
-    K_CH34, // ;
-    K_CH35, // '
-    K_CH00, // `
-    K_LeftShift,
-    K_CH36, // #
-    K_CH38, // Z
-    K_CH39, // X
-    K_CH40, // C
-    K_CH41, // V
-    K_CH42, // B
-    K_CH43, // N
-    K_CH44, // M
-    K_CH45, // ,
-    K_CH46, // .
-    K_CH47, // /
-    K_RightShift, K_KeypadMul,   K_LeftAlt,
-    K_CH48, // Space
-    K_CapsLock,   K_F1,          K_F2,          K_F3,      K_F4,        K_F5,         K_F6,
-    K_F7,         K_F8,          K_F9,          K_F10,     K_NumLock,   K_ScrollLock, K_Keypad7,
-    K_Keypad8,    K_Keypad9,     K_KeypadMinus, K_Keypad4, K_Keypad5,   K_Keypad6,    K_KeypadPlus,
-    K_Keypad1,    K_Keypad2,     K_Keypad3,     K_Keypad0, K_KeypadDot, K_Unknown,    K_Unknown,
-    K_CH37, // Backslash
-    K_F11,        K_F12,
-};
-
-static bool is_shifted(const Keyboard::TTYKeyboardState& state)
+static bool is_shifted(const Keyboard::KeyboardState& state)
 {
     if (state.capslock) return !(state.left_shift || state.right_shift);
     return state.left_shift || state.right_shift;
@@ -184,7 +121,7 @@ static bool is_shifted(const Keyboard::TTYKeyboardState& state)
 
 namespace Keyboard
 {
-    Option<char> decode_scancode_tty(u8 scancode, TTYKeyboardState& state)
+    Option<char> decode_scancode_tty(u8 scancode, KeyboardState& state)
     {
         if (state.ignore_next)
         {
@@ -247,24 +184,4 @@ namespace Keyboard
         if (is_shifted(state)) return shifted_key_table[scancode];
         return key_table[scancode];
     }
-
-    Option<KeyboardPacket> decode_scancode(u8 scancode, KeyboardState& state)
-    {
-        if (state.ignore_next)
-        {
-            state.ignore_next = false;
-            return {};
-        }
-
-        // FIXME: Support extended scancodes.
-        if (scancode == EXTENDED_KEY_CODE)
-        {
-            state.ignore_next = true;
-            return {};
-        }
-
-        bool released = is_key_released(scancode);
-
-        return KeyboardPacket { keycode_table[scancode], released };
-    }
 }

kernel/src/arch/x86_64/MMU.cpp

@@ -110,11 +110,10 @@ namespace MMU
         return result;
     }
 
-    PageTableEntry& l4_entry(u64 virt, PageDirectory* directory = nullptr)
+    PageTableEntry& l4_entry(u64 virt)
     {
         auto index = l4_index(virt);
-        auto* vdir = directory ? translate_physical(directory) : get_virtual_page_directory();
-        return vdir->entries[index];
+        return get_virtual_page_directory()->entries[index];
     }
 
     PageDirectory& page_table(const PageTableEntry& entry)
@@ -140,9 +139,9 @@ namespace MMU
         return page_table(entry).entries[index];
     }
 
-    Result<PageTableEntry*> find_entry(u64 virt, PageDirectory* directory = nullptr)
+    Result<PageTableEntry*> find_entry(u64 virt)
     {
-        const auto& l4 = l4_entry(virt, directory);
+        const auto& l4 = l4_entry(virt);
         if (!l4.present) return err(EFAULT);
         auto& l3 = l3_entry(l4, virt);
         if (!l3.present) return err(EFAULT);
@@ -153,9 +152,9 @@ namespace MMU
         return &l1_entry(l2, virt);
     }
 
-    Result<PageTableEntry*> apply_cascading_flags(u64 virt, int flags, PageDirectory* directory = nullptr)
+    Result<PageTableEntry*> apply_cascading_flags(u64 virt, int flags)
     {
-        auto& l4 = l4_entry(virt, directory);
+        auto& l4 = l4_entry(virt);
         if (!l4.present) return err(EFAULT);
         if (flags & Flags::ReadWrite) l4.read_write = true;
         if (flags & Flags::User) l4.user = true;
@@ -184,9 +183,9 @@ namespace MMU
         entry.set_address(phys);
     }
 
-    Result<void> map(u64 virt, u64 phys, int flags, UseHugePages use_huge_pages, PageDirectory* directory)
+    Result<void> map(u64 virt, u64 phys, int flags, UseHugePages use_huge_pages)
     {
-        auto& l4 = l4_entry(virt, directory);
+        auto& l4 = l4_entry(virt);
         if (!l4.present)
         {
             const u64 addr = TRY(MemoryManager::alloc_frame());
@@ -253,9 +252,9 @@ namespace MMU
         return {};
     }
 
-    Result<u64> unmap(u64 virt, PageDirectory* directory)
+    Result<u64> unmap(u64 virt)
     {
-        auto& l1 = *TRY(find_entry(virt, directory));
+        auto& l1 = *TRY(find_entry(virt));
         if (!l1.present) return err(EFAULT);
         const u64 address = l1.get_address();
         l1.clear();
@@ -263,16 +262,16 @@ namespace MMU
         return address;
     }
 
-    Result<u64> get_physical(u64 virt, PageDirectory* directory)
+    Result<u64> get_physical(u64 virt)
     {
-        const auto& l1 = *TRY(find_entry(virt, directory));
+        const auto& l1 = *TRY(find_entry(virt));
         if (!l1.present) return err(EFAULT);
         return l1.get_address();
     }
 
-    Result<int> get_flags(u64 virt, PageDirectory* directory)
+    Result<int> get_flags(u64 virt)
     {
-        const auto& l1 = *TRY(find_entry(virt, directory));
+        const auto& l1 = *TRY(find_entry(virt));
         if (!l1.present) return err(EFAULT);
         return arch_flags_to_mmu(l1);
     }
@@ -441,8 +440,12 @@ namespace MMU
 
                     for (u64 l = 0; l < 512; l++)
                     {
+                        PageTableEntry& old_l1 = old_pt->entries[l];
+                        if (!old_l1.present) continue;
                         PageTableEntry& new_l1 = new_pt->entries[l];
-                        new_l1.present = false; // The entry must be filled in by the caller.
+                        new_l1.set_address(TRY(MemoryManager::alloc_frame()));
+
+                        memcpy(&page_table(new_l1), &page_table(old_l1), ARCH_PAGE_SIZE);
                     }
                 }
             }

kernel/src/arch/x86_64/PS2Mouse.cpp

@@ -1,118 +0,0 @@
-#include "Log.h"
-#include "api/Mouse.h"
-#include "arch/CPU.h"
-#include "arch/x86_64/IO.h"
-#include "fs/devices/MouseDevice.h"
-#include "thread/Thread.h"
-
-extern Thread* g_io_thread;
-
-static u8 g_mouse_packet[3];
-static int g_index = 0;
-
-#define PS2_MOUSE_Y_OVERFLOW 0x80
-#define PS2_MOUSE_X_OVERFLOW 0x40
-#define PS2_MOUSE_Y_SIGN 0x20
-#define PS2_MOUSE_X_SIGN 0x10
-#define PS2_MOUSE_ALWAYS_1 0x08
-#define PS2_MOUSE_MIDDLE_BTN 0x04
-#define PS2_MOUSE_RIGHT_BTN 0x02
-#define PS2_MOUSE_LEFT_BTN 0x01
-
-extern CircularQueue<moon::MousePacket, 20> mouse_queue;
-
-static void process_mouse_event(u8 data)
-{
-    if (g_index == 0)
-    {
-        // NOTE: https://wiki.osdev.org/Mouse_Input#PS2_Mouse says to discard the packet if X or Y overflow is 1, but
-        // https://wiki.osdev.org/PS2_Mouse uses it. Discard for now.
-        if (data & PS2_MOUSE_X_OVERFLOW) return;
-        if (data & PS2_MOUSE_Y_OVERFLOW) return;
-
-        if ((data & PS2_MOUSE_ALWAYS_1) == 0) return;
-    }
-
-    g_mouse_packet[g_index++] = data;
-    if (g_index < 3) return;
-    g_index = 0;
-
-    moon::MousePacket packet;
-    packet.xdelta = g_mouse_packet[1];
-    packet.ydelta = g_mouse_packet[2];
-    packet.buttons = 0;
-
-    u8 flags = g_mouse_packet[0];
-    if (flags & PS2_MOUSE_X_SIGN) packet.xdelta = -(256 - packet.xdelta);
-    if (flags & PS2_MOUSE_Y_SIGN) packet.ydelta = -(256 - packet.ydelta);
-
-    if (flags & PS2_MOUSE_MIDDLE_BTN) packet.buttons |= moon::MouseButton::Middle;
-    if (flags & PS2_MOUSE_RIGHT_BTN) packet.buttons |= moon::MouseButton::Right;
-    if (flags & PS2_MOUSE_LEFT_BTN) packet.buttons |= moon::MouseButton::Left;
-
-    MouseDevice::add_mouse_event(packet);
-}
-
-static void mouse_interrupt(Registers*, void*)
-{
-    const u8 data = IO::inb(0x60);
-    process_mouse_event(data);
-}
-
-#define PS2_MOUSE_TIMEOUT 100000
-
-static void mouse_wait()
-{
-    int timeout = PS2_MOUSE_TIMEOUT;
-    while (--timeout)
-    {
-        if ((IO::inb(0x64) & 0b10) == 0) return;
-        CPU::pause();
-    }
-}
-
-static void mouse_wait_for_input()
-{
-    int timeout = PS2_MOUSE_TIMEOUT;
-    while (--timeout)
-    {
-        if (IO::inb(0x64) & 0b1) return;
-        CPU::pause();
-    }
-}
-
-static void mouse_write_data(u8 byte)
-{
-    mouse_wait();
-    IO::outb(0x64, 0xD4);
-    mouse_wait();
-    IO::outb(0x60, byte);
-}
-
-void init_mouse()
-{
-    CPU::register_interrupt(12, mouse_interrupt, nullptr);
-
-    IO::outb(0x64, 0xA8); // Enable PS/2 auxiliary port
-
-    mouse_wait();
-    IO::outb(0x64, 0x20); // Get Compaq status byte
-    mouse_wait_for_input();
-    u8 status = IO::inb(0x60);
-    status |= 0x02;  // Enable IRQ12
-    status &= ~0x20; // Disable Mouse Clock
-    mouse_wait();
-    IO::outb(0x64, 0x60); // Set Compaq status byte
-    mouse_wait();
-    IO::outb(0x60, status);
-
-    mouse_write_data(0xF6); // Reset defaults
-    mouse_wait();
-    IO::inb(0x60);
-
-    mouse_write_data(0xF4); // Send automatic packets when the mouse moves or is clicked
-    mouse_wait();
-    IO::inb(0x60);
-
-    g_index = 0;
-}

kernel/src/arch/x86_64/Thread.cpp

@@ -106,7 +106,7 @@ bool Thread::deliver_signal(int signo, Registers* current_regs)
 
     memcpy(&regs, current_regs, sizeof(regs));
 
-    kinfoln("signal: delivering signal %d for thread %d, ip=%p, sp=%p, handler=%p, sigreturn=%p", signo, id,
+    kinfoln("signal: delivering signal %d for thread %ld, ip=%p, sp=%p, handler=%p, sigreturn=%p", signo, id,
             (void*)regs.rip, (void*)regs.rsp, (void*)handler.sa_handler, (void*)handler.__sa_sigreturn);
 
     regs.rsp -= 128; // Skip the red zone
@@ -124,7 +124,7 @@ bool Thread::deliver_signal(int signo, Registers* current_regs)
     if (push_mem_on_stack((u8*)&handler.__sa_sigreturn, sizeof(void*)).has_error()) return false;
 
     signal_mask = handler.sa_mask;
-    if ((handler.sa_flags & SA_NODEFER) == 0) signal_mask.set(signo - 1, true);
+    if ((handler.sa_flags & SA_NODEFER) == 0) signal_mask |= (1 << (signo - 1));
 
     rsp = regs.rsp;
 

kernel/src/arch/x86_64/disk/ATA.cpp

@@ -6,7 +6,6 @@
 #include "fs/MBR.h"
 #include "memory/MemoryManager.h"
 #include <luna/Alignment.h>
-#include <luna/Buffer.h>
 #include <luna/CType.h>
 #include <luna/SafeArithmetic.h>
 #include <luna/Vector.h>
@@ -162,7 +161,7 @@ namespace ATA
     {
         if (!(read_bm(BusmasterRegister::Status) & BMS_IRQPending)) return;
 
-        if (m_current_drive < 2 && m_drives[m_current_drive].has_value()) m_drives[m_current_drive]->irq_handler();
+        if (m_current_drive < 2 && m_drives[m_current_drive]) m_drives[m_current_drive]->irq_handler();
 
         m_irq_called = true;
 
@@ -308,7 +307,14 @@ namespace ATA
 
             kinfoln("ata: Channel %d has a drive on slot %d!", m_channel_index, drive);
 
-            m_drives[drive] = Drive { this, drive, {} };
+            auto rc = adopt_shared_if_nonnull(new (std::nothrow) Drive(this, drive, {}));
+            if (rc.has_error())
+            {
+                kinfoln("ata: Failed to create drive object: %s", rc.error_string());
+                return false;
+            }
+
+            m_drives[drive] = rc.release_value();
 
             if (!m_drives[drive]->initialize())
             {
@@ -321,7 +327,7 @@ namespace ATA
 
         for (u8 drive = 0; drive < 2; drive++)
         {
-            if (m_drives[drive].has_value())
+            if (m_drives[drive])
             {
                 if (!m_drives[drive]->post_initialize())
                 {
@@ -329,7 +335,7 @@ namespace ATA
                     return false;
                 }
 
-                auto rc = ATADevice::create(m_drives[drive].value_ptr());
+                auto rc = ATADevice::create(m_drives[drive]);
 
                 if (rc.has_error())
                 {
@@ -432,18 +438,6 @@ namespace ATA
             m_uses_dma = false;
         }
 
-        if (m_drive_index == 0 && !(status & BMS_MasterInit))
-        {
-            kwarnln("ata: Drive %d does not have DMA support", m_drive_index);
-            m_uses_dma = false;
-        }
-
-        if (m_drive_index == 1 && !(status & BMS_SlaveInit))
-        {
-            kwarnln("ata: Drive %d does not have DMA support", m_drive_index);
-            m_uses_dma = false;
-        }
-
         auto frame = MemoryManager::alloc_frame();
         if (frame.has_error() || frame.value() > 0xffffffff)
         {
@@ -729,7 +723,7 @@ namespace ATA
 
 static u32 next_minor = 0;
 
-Result<String> ATA::Drive::create_drive_name(ATA::Drive* drive)
+Result<String> ATA::Drive::create_drive_name(SharedPtr<ATA::Drive> drive)
 {
     static u32 cd_index = 0;
     static u32 sd_index = 0;
@@ -737,28 +731,67 @@ Result<String> ATA::Drive::create_drive_name(ATA::Drive* drive)
     return String::format("%s%d"_sv, drive->m_is_atapi ? "cd" : "sd", drive->m_is_atapi ? cd_index++ : sd_index++);
 }
 
-Result<SharedPtr<Device>> ATADevice::create(ATA::Drive* drive)
+Result<SharedPtr<Device>> ATADevice::create(SharedPtr<ATA::Drive> drive)
 {
-    auto device = TRY(adopt_shared_if_nonnull(new (std::nothrow) ATADevice(drive)));
+    auto device = TRY(adopt_shared_if_nonnull(new (std::nothrow) ATADevice()));
+    device->m_drive = drive;
     device->m_device_path = TRY(ATA::Drive::create_drive_name(drive));
     TRY(DeviceRegistry::register_special_device(DeviceRegistry::Disk, next_minor++, device, 0400));
     return (SharedPtr<Device>)device;
 }
 
-ATADevice::ATADevice(ATA::Drive* drive) : BlockDevice(drive->block_size(), drive->block_count()), m_drive(drive)
+Result<u64> ATADevice::read(u8* buf, usize offset, usize size) const
 {
-}
+    if (size == 0) return 0;
+
+    if (offset > m_drive->capacity()) return 0;
+    if (offset + size > m_drive->capacity()) size = m_drive->capacity() - offset;
+
+    usize length = size;
+
+    auto block_size = m_drive->block_size();
 
-Result<void> ATADevice::read_block(Buffer& buf, u64 block) const
-{
     ScopedKMutexLock<100> lock(m_drive->channel()->lock());
 
-    if (buf.size() != m_drive->block_size())
+    // FIXME: Don't always allocate this if we don't need it.
+    auto* temp = TRY(make_array<u8>(block_size));
+    auto guard = make_scope_guard([temp] { delete[] temp; });
+
+    if (offset % block_size)
     {
-        kwarnln("ata: error while reading block %lu: cache entry size mismatch (%lu), data=%p", block, buf.size(),
-                buf.data());
-        fail("Cache entry size mismatch");
+        // The size we need to read to round up to a block.
+        usize extra_size = block_size - (offset % block_size);
+        // Maybe we don't even want enough to get to the next block?
+        if (extra_size > size) extra_size = size;
+
+        TRY(m_drive->read_lba(offset / block_size, temp, 1));
+        memcpy(buf, temp + (offset % block_size), extra_size);
+        offset += extra_size;
+        size -= extra_size;
+        buf += extra_size;
     }
 
-    return m_drive->read_lba(block, buf.data(), 1);
+    while (size >= ARCH_PAGE_SIZE)
+    {
+        TRY(m_drive->read_lba(offset / block_size, buf, ARCH_PAGE_SIZE / block_size));
+        offset += ARCH_PAGE_SIZE;
+        size -= ARCH_PAGE_SIZE;
+        buf += ARCH_PAGE_SIZE;
+    }
+
+    while (size >= block_size)
+    {
+        TRY(m_drive->read_lba(offset / block_size, buf, 1));
+        offset += block_size;
+        size -= block_size;
+        buf += block_size;
+    }
+
+    if (size)
+    {
+        TRY(m_drive->read_lba(offset / block_size, temp, 1));
+        memcpy(buf, temp, size);
+    }
+
+    return length;
 }

kernel/src/arch/x86_64/disk/ATA.h

@@ -1,6 +1,5 @@
 #pragma once
 #include "arch/PCI.h"
-#include "fs/devices/BlockDevice.h"
 #include "fs/devices/DeviceRegistry.h"
 #include "lib/KMutex.h"
 #include <luna/Atomic.h>
@@ -165,7 +164,7 @@ namespace ATA
 
         Result<void> read_lba(u64 lba, void* out, usize nblocks);
 
-        static Result<String> create_drive_name(ATA::Drive* drive);
+        static Result<String> create_drive_name(SharedPtr<ATA::Drive> drive);
 
       private:
         bool identify_ata();
@@ -257,6 +256,7 @@ namespace ATA
         Controller* m_controller;
         u8 m_channel_index;
         bool m_is_pci_native_mode;
+
         u8 m_interrupt_line;
 
         KMutex<100> m_lock {};
@@ -271,10 +271,10 @@ namespace ATA
 
         u8 m_current_drive = (u8)-1;
 
-        Option<Drive> m_drives[2];
+        SharedPtr<Drive> m_drives[2];
     };
 
-    class Controller : public Shareable
+    class Controller
     {
       public:
         static Result<void> scan();
@@ -297,29 +297,48 @@ namespace ATA
 
 }
 
-class ATADevice : public BlockDevice
+class ATADevice : public Device
 {
   public:
     // Initializer for DeviceRegistry.
-    static Result<SharedPtr<Device>> create(ATA::Drive* drive);
+    static Result<SharedPtr<Device>> create(SharedPtr<ATA::Drive> drive);
 
-    Result<void> read_block(Buffer& buf, u64 block) const override;
+    Result<usize> read(u8*, usize, usize) const override;
 
-    bool will_block_if_read() const override
+    Result<usize> write(const u8*, usize, usize) override
+    {
+        return err(ENOTSUP);
+    }
+
+    bool blocking() const override
     {
         return false;
     }
 
+    bool is_block_device() const override
+    {
+        return true;
+    }
+
+    usize size() const override
+    {
+        return m_drive->capacity();
+    }
+
+    Result<usize> block_size() const override
+    {
+        return m_drive->block_size();
+    }
+
     StringView device_path() const override
     {
         return m_device_path.view();
     }
 
-    ATADevice(ATA::Drive* drive);
     virtual ~ATADevice() = default;
 
   private:
     ATADevice() = default;
-    ATA::Drive* m_drive;
+    SharedPtr<ATA::Drive> m_drive;
     String m_device_path;
 };

kernel/src/binfmt/BinaryFormat.cpp

@@ -1,42 +0,0 @@
-#include "binfmt/BinaryFormat.h"
-#include "binfmt/ELF.h"
-#include "binfmt/Script.h"
-
-struct BinaryFormatDescriptor
-{
-    binfmt_loader_creator_t creator;
-    void* arg;
-};
-
-Vector<BinaryFormatDescriptor> g_binary_formats;
-
-Result<void> BinaryFormat::init()
-{
-    TRY(register_binary_format(ELFLoader::create, nullptr));
-    TRY(register_binary_format(ScriptLoader::create, nullptr));
-
-    return {};
-}
-
-Result<void> BinaryFormat::register_binary_format(binfmt_loader_creator_t creator, void* arg)
-{
-    return g_binary_formats.try_append({ creator, arg });
-}
-
-Result<SharedPtr<BinaryFormatLoader>> BinaryFormat::create_loader(SharedPtr<VFS::Inode> inode, int recursion_level)
-{
-    if (recursion_level >= 8) return err(ELOOP);
-
-    for (const auto& format : g_binary_formats)
-    {
-        auto loader = TRY(format.creator(inode, format.arg, recursion_level));
-        if (TRY(loader->sniff())) return loader;
-    }
-
-    return err(ENOEXEC);
-}
-
-BinaryFormatLoader::BinaryFormatLoader(SharedPtr<VFS::Inode> inode, int recursion_level)
-    : m_inode(inode), m_recursion_level(recursion_level)
-{
-}

kernel/src/binfmt/BinaryFormat.h

@@ -1,114 +0,0 @@
-#pragma once
-#include "fs/VFS.h"
-#include "memory/AddressSpace.h"
-
-class BinaryFormatLoader : public Shareable
-{
-  public:
-    /**
-     * @brief Determine if the given executable file matches this binary format.
-     *
-     * @return Result<bool> An error, or whether the file matches the binary format.
-     */
-    virtual Result<bool> sniff() = 0;
-
-    /**
-     * @brief Load the given executable binary file into an address space.
-     *
-     * Depending on the binary format, this function may load an arbitrary interpreter instead, which will interpret the
-     * file on its own.
-     *
-     * @param space The address space to load the executable into.
-     * @return Result<u64> An error, or the entry point of the executable in memory.
-     */
-    virtual Result<u64> load(AddressSpace* space) = 0;
-
-    /**
-     * @brief Return the short name associated with this format.
-     *
-     * @return StringView The format's name.
-     */
-    virtual StringView format() const = 0;
-
-    /**
-     * @brief Transform an interpreted program's command line arguments.
-     *
-     * Example: A script 'foo.sh' with a shebang line '#!/bin/sh' is loaded using a BinaryFormatLoader.
-     *
-     * This function will then be called with path="/path/to/foo.sh" and args={"foo.sh", "--enable-bar"}
-     *
-     * The function should return {"/bin/sh", "/path/to/foo.sh", "--enable-bar"} (prepending the interpreter command
-     * line and substituting args[0] for the full path of the script).
-     *
-     * For native executable formats that do not require an interpreter (e.g. ELF), this function should just ignore
-     * path and return args unmodified.
-     *
-     * @param path The path (absolute or relative to the current process's working directory) of the current program
-     * file. This should be used instead of args[0] as arbitrary values can be passed there, leaving the interpreter
-     * unable to find the target program.
-     * @param args The original command line arguments passed to execve().
-     * @return Result<Vector<String>> An error, or the transformed command line arguments.
-     */
-    virtual Result<Vector<String>> cmdline(const String& path, Vector<String> args) = 0;
-
-    virtual ~BinaryFormatLoader() = default;
-
-    /**
-     * @brief Construct a new BinaryFormatLoader.
-     *
-     * This should not be directly used, instead each subclass of BinaryFormatLoader should implement a static create()
-     * method which implements the binfmt_loader_creator_t type and register it using
-     * BinaryFormat::register_binary_format().
-     *
-     * Then, anyone that needs an appropriate BinaryFormatLoader for an executable file should call
-     * BinaryFormat::create_loader(), which will find an appropriate loader out of all default/registered loaders.
-     *
-     * @param inode The executable program file to load into memory.
-     * @param recursion_level In normal cases, 0. If the BinaryFormatLoader is created inside another loader (for
-     * example, a script loader loading the script's interpreter), the caller shall pass its recursion_level + 1 to
-     * BinaryFormat::create_loader(), which will forward it to this constructor. This avoids infinite recursion.
-     */
-    BinaryFormatLoader(SharedPtr<VFS::Inode> inode, int recursion_level);
-
-  protected:
-    SharedPtr<VFS::Inode> m_inode;
-    int m_recursion_level;
-};
-
-/**
- * @brief The factory function signature for binary format loaders.
- */
-typedef Result<SharedPtr<BinaryFormatLoader>> (*binfmt_loader_creator_t)(SharedPtr<VFS::Inode>, void*, int);
-
-namespace BinaryFormat
-{
-    /**
-     * @brief Register the default binary format loaders.
-     *
-     * @return Result<void> Whether the operation succeeded.
-     */
-    Result<void> init();
-
-    /**
-     * @brief Register a new binary format loader type.
-     *
-     * @param creator A factory function to create said binary format loader. The function shall be passed the inode to
-     * load the program from, the arbitrary argument passed to this function, and a recursion index to avoid infinite
-     * recursion.
-     * @param arg An arbitrary argument that will be passed to the above factory function.
-     * @return Result<void> Whether the operation succeeded.
-     */
-    Result<void> register_binary_format(binfmt_loader_creator_t creator, void* arg);
-
-    /**
-     * @brief Create an appropriate loader object for an executable file.
-     *
-     * @param inode The executable file to create the loader for. If no appropriate loader could be found for this file
-     * type, this function shall return ENOEXEC.
-     * @param recursion_level In most cases, 0. If called inside another loader, the loader shall pass its
-     * own recursion_level variable + 1, to avoid infinite recursion. If recursion_level >= 8, this function shall
-     * immediately return ELOOP.
-     * @return Result<SharedPtr<BinaryFormatLoader>> An error, or the new loader object created for this file.
-     */
-    Result<SharedPtr<BinaryFormatLoader>> create_loader(SharedPtr<VFS::Inode> inode, int recursion_level = 0);
-}

kernel/src/binfmt/ELF.cpp

@@ -1,154 +0,0 @@
-#include "binfmt/ELF.h"
-#include "Log.h"
-#include "arch/CPU.h"
-#include "arch/MMU.h"
-#include "memory/MemoryManager.h"
-#include <bits/mmap-flags.h>
-#include <luna/Alignment.h>
-#include <luna/Alloc.h>
-#include <luna/CString.h>
-#include <luna/Common.h>
-#include <luna/ScopeGuard.h>
-
-static bool can_execute_segment(u32 flags)
-{
-    return flags & 1;
-}
-
-static bool can_write_segment(u32 flags)
-{
-    return flags & 2;
-}
-
-/*static bool can_write_and_execute_segment(u32 flags)
-{
-    return can_write_segment(flags) && can_execute_segment(flags);
-}*/
-
-Result<bool> ELFLoader::sniff()
-{
-    u8 buf[SELFMAG];
-    usize nread = TRY(m_inode->read(buf, 0, sizeof buf));
-    if (nread < SELFMAG) return false;
-
-    return !memcmp(buf, ELFMAG, SELFMAG);
-}
-
-Result<u64> ELFLoader::load(AddressSpace* space)
-{
-    Elf64_Ehdr elf_header;
-    usize nread = TRY(m_inode->read((u8*)&elf_header, 0, sizeof elf_header));
-
-    if (nread < sizeof elf_header)
-    {
-        kerrorln("Error while loading ELF: ELF header does not fit in file");
-        return err(ENOEXEC);
-    }
-
-    if (memcmp(elf_header.e_ident, ELFMAG, SELFMAG) != 0)
-    {
-        kerrorln("Error while loading ELF: ELF header has no valid magic");
-        return err(ENOEXEC);
-    }
-
-    if (elf_header.e_ident[EI_CLASS] != ELFCLASS64)
-    {
-        kerrorln("Error while loading ELF: ELF object is not 64-bit");
-        return err(ENOEXEC);
-    }
-
-    if (elf_header.e_ident[EI_DATA] != ELFDATA2LSB)
-    {
-        kerrorln("Error while loading ELF: ELF object is not 2's complement little-endian");
-        return err(ENOEXEC);
-    }
-
-    if (elf_header.e_type != ET_EXEC)
-    {
-        kerrorln("Error while loading ELF: ELF object is not an executable");
-        return err(ENOEXEC);
-    }
-
-    if (elf_header.e_machine != EM_MACH)
-    {
-        kerrorln("Error while loading ELF: ELF object's target architecture does not match the current one (%s)",
-                 CPU::platform_string().chars());
-        return err(ENOEXEC);
-    }
-
-    if (elf_header.e_phnum == 0)
-    {
-        kerrorln("Error while loading ELF: ELF object has no program headers");
-        return err(ENOEXEC);
-    }
-
-#ifdef ELF_DEBUG
-    kdbgln("ELF: Loading ELF with entry=%#.16lx", elf_header.e_entry);
-#endif
-
-    usize i;
-    Elf64_Phdr program_header;
-
-    for (TRY(m_inode->read((u8*)&program_header, elf_header.e_phoff, sizeof program_header)), i = 0;
-         i < elf_header.e_phnum;
-         i++, TRY(m_inode->read((u8*)&program_header, elf_header.e_phoff + (i * elf_header.e_phentsize),
-                                sizeof program_header)))
-    {
-        if (program_header.p_type == PT_LOAD)
-        {
-#ifdef ELF_DEBUG
-            kdbgln("ELF: Loading segment (offset=%zu, base=%#.16lx, filesize=%zu, memsize=%zu)",
-                   program_header.p_offset, program_header.p_vaddr, program_header.p_filesz, program_header.p_memsz);
-#endif
-
-            u64 base_vaddr = align_down<ARCH_PAGE_SIZE>(program_header.p_vaddr);
-            u64 vaddr_diff = program_header.p_vaddr - base_vaddr;
-            /*expect(!can_write_and_execute_segment(program_header.p_flags),
-                   "Segment is both writable and executable");*/
-
-            int flags = MMU::User | MMU::NoExecute;
-            if (can_write_segment(program_header.p_flags)) flags |= MMU::ReadWrite;
-            if (can_execute_segment(program_header.p_flags)) flags &= ~MMU::NoExecute;
-
-            int prot = PROT_READ;
-            if (can_write_segment(program_header.p_flags)) prot |= PROT_WRITE;
-            if (can_execute_segment(program_header.p_flags)) prot |= PROT_EXEC;
-
-            if (!TRY(space->test_and_alloc_region(base_vaddr,
-                                                  ceil_div(program_header.p_memsz + vaddr_diff, ARCH_PAGE_SIZE), prot,
-                                                  MAP_ANONYMOUS | MAP_PRIVATE, 0, true)))
-                return err(ENOMEM);
-
-            // Allocate physical memory for the segment
-            TRY(MemoryManager::alloc_at(base_vaddr, ceil_div(program_header.p_memsz + vaddr_diff, ARCH_PAGE_SIZE),
-                                        flags));
-
-            // Zero out unused memory (before the start of the segment)
-            memset((void*)base_vaddr, 0, vaddr_diff);
-
-            // Load the file section of the segment
-            m_inode->read((u8*)program_header.p_vaddr, program_header.p_offset, program_header.p_filesz);
-
-            // Fill out the rest of the segment with 0s
-            memset((void*)(program_header.p_vaddr + program_header.p_filesz), 0,
-                   program_header.p_memsz - program_header.p_filesz);
-        }
-        else { kwarnln("ELF: Encountered non-loadable program header, skipping"); }
-    }
-
-    return elf_header.e_entry;
-}
-
-Result<Vector<String>> ELFLoader::cmdline(const String&, Vector<String> args)
-{
-    return args;
-}
-
-ELFLoader::ELFLoader(SharedPtr<VFS::Inode> inode, int recursion_level) : BinaryFormatLoader(inode, recursion_level)
-{
-}
-
-Result<SharedPtr<BinaryFormatLoader>> ELFLoader::create(SharedPtr<VFS::Inode> inode, void*, int recursion_level)
-{
-    return (SharedPtr<BinaryFormatLoader>)TRY(make_shared<ELFLoader>(inode, recursion_level));
-}

kernel/src/binfmt/Script.cpp

@@ -1,68 +0,0 @@
-#include "binfmt/Script.h"
-#include "binfmt/ELF.h"
-#include "thread/Scheduler.h"
-
-#define SHEBANG "#!"
-
-Result<bool> ScriptLoader::sniff()
-{
-    u8 buf[2];
-    usize nread = TRY(m_inode->read(buf, 0, sizeof buf));
-    if (nread < 2) return false;
-
-    return !memcmp(buf, SHEBANG, 2);
-}
-
-Result<u64> ScriptLoader::load(AddressSpace* space)
-{
-    u8 buf[256];
-    usize nread = TRY(m_inode->read(buf, 2, 255));
-    if (!nread) return err(ENOEXEC);
-    for (usize i = 0; i < nread; i++)
-    {
-        if (buf[i] == '\n') buf[i] = '\0';
-        else if (buf[i] == '\r' && (i + 1) < nread && buf[i + 1] == '\n')
-            buf[i] = buf[i + 1] = '\0';
-        else
-            continue;
-        break;
-    }
-
-    auto view = StringView { (const char*)buf };
-    m_interpreter_cmdline = TRY(view.split(" "));
-    if (!m_interpreter_cmdline.size()) return err(ENOEXEC);
-
-    auto& interpreter_path = m_interpreter_cmdline[0];
-    auto* current = Scheduler::current();
-
-    auto interpreter =
-        TRY(VFS::resolve_path(interpreter_path.chars(), current->auth, current->current_directory, true));
-    if (!VFS::can_execute(interpreter, current->auth)) return err(EACCES);
-
-    auto loader = TRY(BinaryFormat::create_loader(interpreter, m_recursion_level + 1));
-    u64 entry = TRY(loader->load(space));
-
-    m_interpreter_cmdline = TRY(loader->cmdline(interpreter_path, move(m_interpreter_cmdline)));
-    return entry;
-}
-
-Result<Vector<String>> ScriptLoader::cmdline(const String& path, Vector<String> args)
-{
-    Vector<String> new_args;
-    TRY(new_args.try_reserve(m_interpreter_cmdline.size() + args.size() + 1));
-    for (auto& arg : m_interpreter_cmdline) { TRY(new_args.try_append(move(arg))); }
-    auto arg = TRY(path.clone());
-    TRY(new_args.try_append(move(arg)));
-    for (usize i = 1; i < args.size(); i++) { TRY(new_args.try_append(move(args[i]))); }
-    return new_args;
-}
-
-ScriptLoader::ScriptLoader(SharedPtr<VFS::Inode> inode, int recursion_level)
-    : BinaryFormatLoader(inode, recursion_level)
-{
-}
-
-Result<SharedPtr<BinaryFormatLoader>> ScriptLoader::create(SharedPtr<VFS::Inode> inode, void*, int recursion_level)
-{
-    return (SharedPtr<BinaryFormatLoader>)TRY(make_shared<ScriptLoader>(inode, recursion_level));
-}

kernel/src/binfmt/Script.h

@@ -1,25 +0,0 @@
-#pragma once
-#include "binfmt/BinaryFormat.h"
-#include "fs/VFS.h"
-#include "memory/AddressSpace.h"
-
-class ScriptLoader : public BinaryFormatLoader
-{
-  public:
-    Result<bool> sniff() override;
-    Result<u64> load(AddressSpace* space) override;
-
-    Result<Vector<String>> cmdline(const String& path, Vector<String> args) override;
-
-    StringView format() const override
-    {
-        return "script";
-    }
-
-    ScriptLoader(SharedPtr<VFS::Inode> inode, int recursion_level);
-
-    static Result<SharedPtr<BinaryFormatLoader>> create(SharedPtr<VFS::Inode> inode, void*, int);
-
-  private:
-    Vector<String> m_interpreter_cmdline;
-};

kernel/src/fs/GPT.cpp

@@ -41,8 +41,8 @@ namespace GPT
 
         u32 partition_index = 1;
 
-        auto* table = (PartitionEntry*)TRY(calloc_impl(header.num_partitions, sizeof(PartitionEntry)));
-        auto guard = make_scope_guard([table] { free_impl(table); });
+        auto* table = TRY(make_array<PartitionEntry>(header.num_partitions));
+        auto guard = make_scope_guard([table] { delete[] table; });
 
         nread = TRY(device->read((u8*)table, partition_table_start, sizeof(PartitionEntry) * header.num_partitions));
         check(nread == sizeof(PartitionEntry) * header.num_partitions);
@@ -66,9 +66,6 @@ namespace GPT
     {
         header.checksum = 0;
 
-        CRC32 crc;
-        crc.append((u8*)&header, 0x5c);
-
-        return crc.digest();
+        return CRC32::checksum((u8*)&header, 0x5c);
     }
 }

kernel/src/fs/InitRD.cpp

@@ -6,7 +6,7 @@
 #include "memory/MemoryManager.h"
 #include "thread/Thread.h"
 #include <bits/modes.h>
-#include <luna/Common.h>
+#include <luna/Alignment.h>
 
 TarStream g_initrd;
 extern const BOOTBOOT bootboot;
@@ -18,15 +18,17 @@ void InitRD::initialize()
     g_initrd.initialize((void*)virtual_initrd_address, bootboot.initrd_size);
 }
 
-static Result<void> vfs_create_dir_if_not_exists(const char* path, mode_t mode)
+static Result<void> vfs_create_dir_if_not_exists(const char* path, mode_t mode, uid_t uid, gid_t gid)
 {
-    auto rc = VFS::create_directory(path, mode & (mode_t)~S_IFMT, Credentials {});
+    auto rc = VFS::create_directory(path, Credentials {});
     if (rc.has_error())
     {
         if (rc.error() == EEXIST) return {};
         return rc.release_error();
     }
     auto dir = rc.value();
+    dir->chmod(mode & (mode_t)~S_IFMT);
+    dir->chown(uid, gid);
     return {};
 }
 
@@ -37,17 +39,19 @@ Result<void> InitRD::populate_vfs()
     {
         if (entry.type == TarStream::EntryType::RegularFile)
         {
-            auto file = TRY(VFS::create_file(entry.name.chars(), entry.mode & (mode_t)~S_IFMT, Credentials {}));
+            auto file = TRY(VFS::create_file(entry.name.chars(), Credentials {}));
             file->write(entry.data(), 0, entry.size);
+            file->chmod(entry.mode & (mode_t)~S_IFMT);
+            file->chown(entry.uid, entry.gid);
         }
         else if (entry.type == TarStream::EntryType::Directory)
         {
-            TRY(vfs_create_dir_if_not_exists(entry.name.chars(), entry.mode));
+            TRY(vfs_create_dir_if_not_exists(entry.name.chars(), entry.mode, entry.uid, entry.gid));
         }
     }
 
     // Now we don't need the original initrd anymore
-    MemoryManager::free_frames(bootboot.initrd_ptr, ceil_div(bootboot.initrd_size, ARCH_PAGE_SIZE));
+    MemoryManager::free_frames(bootboot.initrd_ptr, get_blocks_from_size(bootboot.initrd_size, ARCH_PAGE_SIZE));
 
     return {};
 }

kernel/src/fs/MBR.h

@@ -41,7 +41,7 @@ namespace MBR
 
         Result<usize> write(const u8* buf, usize offset, usize length) override;
 
-        bool will_block_if_read() const override
+        bool blocking() const override
         {
             return false;
         }

kernel/src/fs/Mount.h

@@ -37,19 +37,44 @@ class MountInode : public VFS::Inode, public LinkedListNode<MountInode>
         return err(EISDIR);
     }
 
-    bool will_block_if_read() const override
+    bool blocking() const override
     {
         return false;
     }
 
-    const VFS::InodeMetadata& metadata() const override
+    usize size() const override
     {
-        return m_mount_root_inode->metadata();
+        return 0;
     }
 
-    Result<void> set_metadata(const VFS::InodeMetadata& metadata) override
+    mode_t mode() const override
     {
-        return m_mount_root_inode->set_metadata(metadata);
+        return m_mount_root_inode->mode();
+    }
+
+    u32 uid() const override
+    {
+        return m_mount_root_inode->uid();
+    }
+
+    u32 gid() const override
+    {
+        return m_mount_root_inode->gid();
+    }
+
+    nlink_t nlinks() const override
+    {
+        return m_mount_root_inode->nlinks();
+    }
+
+    Result<void> chmod(mode_t mode) override
+    {
+        return m_mount_root_inode->chmod(mode);
+    }
+
+    Result<void> chown(u32 uid, u32 gid) override
+    {
+        return m_mount_root_inode->chown(uid, gid);
     }
 
     VFS::FileSystem* fs() const override
@@ -57,6 +82,11 @@ class MountInode : public VFS::Inode, public LinkedListNode<MountInode>
         return m_mountee.ptr();
     }
 
+    usize inode_number() const override
+    {
+        return m_mount_root_inode->inode_number();
+    }
+
     VFS::InodeType type() const override
     {
         return VFS::InodeType::Directory;
@@ -92,14 +122,14 @@ class MountInode : public VFS::Inode, public LinkedListNode<MountInode>
         return m_mount_root_inode->remove_entry(name);
     }
 
-    Result<SharedPtr<VFS::Inode>> create_file(const char* name, mode_t mode) override
+    Result<SharedPtr<VFS::Inode>> create_file(const char* name) override
     {
-        return m_mount_root_inode->create_file(name, mode);
+        return m_mount_root_inode->create_file(name);
     }
 
-    Result<SharedPtr<VFS::Inode>> create_subdirectory(const char* name, mode_t mode) override
+    Result<SharedPtr<VFS::Inode>> create_subdirectory(const char* name) override
     {
-        return m_mount_root_inode->create_subdirectory(name, mode);
+        return m_mount_root_inode->create_subdirectory(name);
     }
 
     Result<void> add_entry(SharedPtr<VFS::Inode> inode, const char* name) override

kernel/src/fs/Pipe.cpp

@@ -14,14 +14,8 @@ Result<void> Pipe::create(SharedPtr<VFS::Inode>& rpipe, SharedPtr<VFS::Inode>& w
     pipe->m_reader = reader.ptr();
 
     writer->m_pipe = reader->m_pipe = pipe;
-    writer->m_metadata.inum = 0;
-    writer->m_metadata.uid = auth.euid;
-    writer->m_metadata.gid = auth.egid;
-    writer->m_metadata.mode = 0200;
-    reader->m_metadata.inum = 0;
-    reader->m_metadata.uid = auth.euid;
-    reader->m_metadata.gid = auth.egid;
-    reader->m_metadata.mode = 0400;
+    writer->chown(auth.euid, auth.egid);
+    reader->chown(auth.euid, auth.egid);
 
     rpipe = reader;
     wpipe = writer;
@@ -44,11 +38,7 @@ Result<usize> Pipe::read(u8* buf, usize, usize length)
 
 Result<usize> Pipe::write(const u8* buf, usize, usize length)
 {
-    if (!m_reader)
-    {
-        Scheduler::current()->send_signal(SIGPIPE);
-        return err(EPIPE);
-    }
+    if (!m_reader) return length;
 
     u8* slice = TRY(m_data_buffer.slice_at_end(length));
     memcpy(slice, buf, length);
@@ -56,7 +46,7 @@ Result<usize> Pipe::write(const u8* buf, usize, usize length)
     return length;
 }
 
-bool Pipe::will_block_if_read() const
+bool Pipe::blocking() const
 {
     return !m_data_buffer.size() && m_writer;
 }

kernel/src/fs/Pipe.h

@@ -6,7 +6,7 @@ class PipeInodeBase;
 class PipeReader;
 class PipeWriter;
 
-class Pipe : public Shareable
+class Pipe
 {
   public:
     static Result<void> create(SharedPtr<VFS::Inode>& rpipe, SharedPtr<VFS::Inode>& wpipe);
@@ -15,7 +15,7 @@ class Pipe : public Shareable
 
     Result<usize> write(const u8* buf, usize, usize length);
 
-    bool will_block_if_read() const;
+    bool blocking() const;
 
   private:
     Buffer m_data_buffer;
@@ -40,15 +40,14 @@ class PipeInodeBase : public VFS::FileInode
         return err(ENOTSUP);
     }
 
-    bool will_block_if_read() const override
+    bool blocking() const override
     {
-        return m_pipe->will_block_if_read();
+        return m_pipe->blocking();
     }
 
-    const VFS::InodeMetadata& metadata() const override
+    usize size() const override
     {
-        m_metadata.size = m_pipe->m_data_buffer.size();
-        return m_metadata;
+        return m_pipe->m_data_buffer.size();
     }
 
     void did_link() override
@@ -59,6 +58,23 @@ class PipeInodeBase : public VFS::FileInode
     {
     }
 
+    Result<void> chmod(mode_t) override
+    {
+        return err(ENOTSUP);
+    }
+
+    Result<void> chown(u32 uid, u32 gid) override
+    {
+        m_uid = uid;
+        m_gid = gid;
+        return {};
+    }
+
+    usize inode_number() const override
+    {
+        return 0;
+    }
+
     VFS::FileSystem* fs() const override
     {
         return nullptr;
@@ -70,6 +86,8 @@ class PipeInodeBase : public VFS::FileInode
 
   protected:
     SharedPtr<Pipe> m_pipe;
+    u32 m_uid { 0 };
+    u32 m_gid { 0 };
 };
 
 class PipeWriter : public PipeInodeBase
@@ -86,6 +104,11 @@ class PipeWriter : public PipeInodeBase
         return m_pipe->write(buf, offset, length);
     }
 
+    mode_t mode() const override
+    {
+        return 0200;
+    }
+
     ~PipeWriter();
 
     friend class Pipe;
@@ -105,6 +128,11 @@ class PipeReader : public PipeInodeBase
         check(false);
     }
 
+    mode_t mode() const override
+    {
+        return 0400;
+    }
+
     ~PipeReader();
 
     friend class Pipe;

kernel/src/fs/StorageCache.cpp

@@ -1,49 +0,0 @@
-#include "fs/StorageCache.h"
-#include "Log.h"
-#include <luna/Heap.h>
-#include <luna/ScopeGuard.h>
-
-static LinkedList<StorageCache> g_storage_caches;
-
-Result<StorageCache::CacheEntry*> StorageCache::fetch_entry(u64 block)
-{
-    {
-        CacheEntry* entry = m_cache_entries.try_get_ref(block);
-        if (entry && !entry->buffer.is_empty()) return entry;
-    }
-
-    CacheEntry entry {};
-    TRY(m_cache_entries.try_set(block, move(entry)));
-
-#ifdef CACHE_DEBUG
-    kdbgln("cache: Created new cache entry for block %lu", block);
-#endif
-
-    return m_cache_entries.try_get_ref(block);
-}
-
-void StorageCache::clear()
-{
-    m_mutex.lock();
-
-    kdbgln("cache: clearing %lu entries, out of %lu buckets", m_cache_entries.size(), m_cache_entries.capacity());
-    m_cache_entries.clear();
-    kdbgln("cache: done");
-
-    m_mutex.unlock();
-}
-
-StorageCache::StorageCache()
-{
-    g_storage_caches.append(this);
-}
-
-StorageCache::~StorageCache()
-{
-    g_storage_caches.remove(this);
-}
-
-void StorageCache::clear_caches()
-{
-    for (auto* cache : g_storage_caches) { cache->clear(); }
-}

kernel/src/fs/StorageCache.h

@@ -1,37 +0,0 @@
-#pragma once
-#include "lib/KMutex.h"
-#include <luna/Buffer.h>
-#include <luna/HashMap.h>
-#include <luna/LinkedList.h>
-
-class StorageCache : public LinkedListNode<StorageCache>
-{
-  public:
-    struct CacheEntry
-    {
-        Buffer buffer {};
-    };
-
-    void lock()
-    {
-        return m_mutex.lock();
-    }
-
-    void unlock()
-    {
-        return m_mutex.unlock();
-    }
-
-    Result<CacheEntry*> fetch_entry(u64 block);
-
-    void clear();
-
-    static void clear_caches();
-
-    StorageCache();
-    ~StorageCache();
-
-  private:
-    HashMap<u64, CacheEntry> m_cache_entries;
-    KMutex<100> m_mutex;
-};

kernel/src/fs/VFS.cpp

@@ -69,8 +69,7 @@ namespace VFS
         return resolve_path_impl(path, auth, current_inode, follow_last_symlink, symlinks_followed);
     }
 
-    Result<SharedPtr<Inode>> create_directory(const char* path, mode_t mode, Credentials auth,
-                                              SharedPtr<Inode> working_directory)
+    Result<SharedPtr<Inode>> create_directory(const char* path, Credentials auth, SharedPtr<Inode> working_directory)
     {
         auto parent_path = TRY(PathParser::dirname(path));
 
@@ -82,11 +81,10 @@ namespace VFS
 
         TRY(validate_filename(child_name.view()));
 
-        return parent_inode->create_subdirectory(child_name.chars(), mode);
+        return parent_inode->create_subdirectory(child_name.chars());
     }
 
-    Result<SharedPtr<Inode>> create_file(const char* path, mode_t mode, Credentials auth,
-                                         SharedPtr<Inode> working_directory)
+    Result<SharedPtr<Inode>> create_file(const char* path, Credentials auth, SharedPtr<Inode> working_directory)
     {
         auto parent_path = TRY(PathParser::dirname(path));
 
@@ -98,7 +96,7 @@ namespace VFS
 
         TRY(validate_filename(child_name.view()));
 
-        return parent_inode->create_file(child_name.chars(), mode);
+        return parent_inode->create_file(child_name.chars());
     }
 
     Result<void> validate_filename(StringView name)
@@ -137,51 +135,45 @@ namespace VFS
     {
         if (auth.euid == 0) return true;
 
-        const auto& metadata = inode->metadata();
+        if (inode->uid() == auth.euid) { return inode->mode() & S_IXUSR; }
+        if (inode->gid() == auth.egid) { return inode->mode() & S_IXGRP; }
 
-        if (metadata.uid == auth.euid) { return metadata.mode & S_IXUSR; }
-        if (metadata.gid == auth.egid) { return metadata.mode & S_IXGRP; }
-
-        return metadata.mode & S_IXOTH;
+        return inode->mode() & S_IXOTH;
     }
 
     bool can_write(SharedPtr<Inode> inode, Credentials auth)
     {
         if (auth.euid == 0) return true;
 
-        const auto& metadata = inode->metadata();
+        if (inode->uid() == auth.euid) { return inode->mode() & S_IWUSR; }
+        if (inode->gid() == auth.egid) { return inode->mode() & S_IWGRP; }
 
-        if (metadata.uid == auth.euid) { return metadata.mode & S_IWUSR; }
-        if (metadata.gid == auth.egid) { return metadata.mode & S_IWGRP; }
-
-        return metadata.mode & S_IWOTH;
+        return inode->mode() & S_IWOTH;
     }
 
     bool can_read(SharedPtr<Inode> inode, Credentials auth)
     {
         if (auth.euid == 0) return true;
 
-        const auto& metadata = inode->metadata();
+        if (inode->uid() == auth.euid) { return inode->mode() & S_IRUSR; }
+        if (inode->gid() == auth.egid) { return inode->mode() & S_IRGRP; }
 
-        if (metadata.uid == auth.euid) { return metadata.mode & S_IRUSR; }
-        if (metadata.gid == auth.egid) { return metadata.mode & S_IRGRP; }
-
-        return metadata.mode & S_IROTH;
+        return inode->mode() & S_IROTH;
     }
 
     bool is_setuid(SharedPtr<Inode> inode)
     {
-        return inode->metadata().mode & S_ISUID;
+        return inode->mode() & S_ISUID;
     }
 
     bool is_setgid(SharedPtr<Inode> inode)
     {
-        return inode->metadata().mode & S_ISGID;
+        return inode->mode() & S_ISGID;
     }
 
     bool is_sticky(SharedPtr<Inode> inode)
     {
-        return inode->metadata().mode & S_ISVTX;
+        return inode->mode() & S_ISVTX;
     }
 
     bool is_seekable(SharedPtr<Inode> inode)

kernel/src/fs/VFS.h

@@ -1,7 +1,5 @@
 #pragma once
-#include "arch/Timer.h"
 #include <bits/makedev.h>
-#include <bits/timespec.h>
 #include <luna/SharedPtr.h>
 #include <luna/StaticString.h>
 #include <luna/StringView.h>
@@ -19,44 +17,23 @@ namespace VFS
         BlockDevice,
         Symlink,
         FIFO,
-        Socket,
-    };
-
-    struct InodeMetadata
-    {
-        ino_t inum;
-        size_t size { 0 };
-        mode_t mode;
-        nlink_t nlinks { 1 };
-        uid_t uid { 0 };
-        gid_t gid { 0 };
-        dev_t devid { 0 };
-
-        struct timespec atime;
-        struct timespec mtime;
-        struct timespec ctime;
     };
 
     class Inode;
 
-    class FileSystem : public Shareable
+    class FileSystem
     {
       public:
         virtual SharedPtr<Inode> root_inode() const = 0;
 
-        virtual Result<SharedPtr<Inode>> create_file_inode(mode_t mode) = 0;
+        virtual Result<SharedPtr<Inode>> create_file_inode() = 0;
 
-        virtual Result<SharedPtr<Inode>> create_dir_inode(SharedPtr<Inode> parent, mode_t mode) = 0;
+        virtual Result<SharedPtr<Inode>> create_dir_inode(SharedPtr<Inode> parent) = 0;
 
-        virtual Result<SharedPtr<Inode>> create_device_inode(u32 major, u32 minor, mode_t mode) = 0;
+        virtual Result<SharedPtr<Inode>> create_device_inode(u32 major, u32 minor) = 0;
 
         virtual Result<SharedPtr<Inode>> create_symlink_inode(StringView link) = 0;
 
-        virtual Result<u64> allocate_inode_number()
-        {
-            return err(ENOTSUP);
-        }
-
         virtual Result<void> set_mount_dir(SharedPtr<Inode> parent) = 0;
 
         virtual Result<void> reset_mount_dir() = 0;
@@ -96,7 +73,7 @@ namespace VFS
         StaticString<128> name;
     };
 
-    class Inode : public Shareable
+    class Inode
     {
       public:
         virtual Result<u64> ioctl(int, void*)
@@ -109,23 +86,14 @@ namespace VFS
             return err(ENOTTY);
         }
 
-        virtual void did_close()
-        {
-        }
-
-        virtual Result<u64> query_shared_memory(off_t, usize)
-        {
-            return err(EACCES);
-        }
-
         // Directory-specific methods
         virtual Result<SharedPtr<Inode>> find(const char* name) const = 0;
 
         virtual Option<DirectoryEntry> get(usize index) const = 0;
 
-        virtual Result<SharedPtr<Inode>> create_file(const char* name, mode_t mode) = 0;
+        virtual Result<SharedPtr<Inode>> create_file(const char* name) = 0;
 
-        virtual Result<SharedPtr<Inode>> create_subdirectory(const char* name, mode_t mode) = 0;
+        virtual Result<SharedPtr<Inode>> create_subdirectory(const char* name) = 0;
 
         virtual Result<void> add_entry(SharedPtr<Inode> inode, const char* name) = 0;
 
@@ -142,7 +110,7 @@ namespace VFS
 
         virtual Result<void> truncate(usize size) = 0;
 
-        virtual bool will_block_if_read() const = 0;
+        virtual bool blocking() const = 0;
 
         // Symlink-specific methods
         virtual Result<StringView> readlink()
@@ -150,16 +118,29 @@ namespace VFS
             return StringView {};
         }
 
-        virtual const InodeMetadata& metadata() const
+        virtual dev_t device_id() const
         {
-            return m_metadata;
+            return luna_dev_makedev(0, 0);
         }
 
-        virtual Result<void> set_metadata(const InodeMetadata& metadata)
+        // Metadata accessors
+        virtual usize size() const = 0;
+
+        virtual mode_t mode() const = 0;
+
+        virtual nlink_t nlinks() const
         {
-            m_metadata = metadata;
-            m_metadata.ctime = *Timer::realtime_clock();
-            return {};
+            return 1;
+        }
+
+        virtual u32 uid() const
+        {
+            return 0;
+        }
+
+        virtual u32 gid() const
+        {
+            return 0;
         }
 
         virtual bool is_mountpoint() const
@@ -170,6 +151,11 @@ namespace VFS
         virtual void did_link() = 0;
         virtual void did_unlink() = 0;
 
+        // Metadata changers
+        virtual Result<void> chmod(mode_t mode) = 0;
+
+        virtual Result<void> chown(u32 uid, u32 gid) = 0;
+
         // Generic VFS-related methods
         virtual FileSystem* fs() const = 0;
 
@@ -177,6 +163,8 @@ namespace VFS
 
         virtual InodeType type() const = 0;
 
+        virtual usize inode_number() const = 0;
+
         virtual void add_handle()
         {
             auto* f = fs();
@@ -188,13 +176,9 @@ namespace VFS
             auto* f = fs();
             if (f) f->remove_handle();
         }
-
-      protected:
-        mutable InodeMetadata m_metadata;
-        Option<u64> m_shmid {};
     };
 
-    class FileInode : public Inode
+    class FileInode : Inode
     {
       public:
         Result<SharedPtr<Inode>> find(const char*) const override
@@ -207,12 +191,12 @@ namespace VFS
             return {};
         }
 
-        Result<SharedPtr<Inode>> create_file(const char*, mode_t) override
+        Result<SharedPtr<Inode>> create_file(const char*) override
         {
             return err(ENOTDIR);
         }
 
-        Result<SharedPtr<Inode>> create_subdirectory(const char*, mode_t) override
+        Result<SharedPtr<Inode>> create_subdirectory(const char*) override
         {
             return err(ENOTDIR);
         }
@@ -237,7 +221,7 @@ namespace VFS
             return 0;
         }
 
-        bool will_block_if_read() const override
+        bool blocking() const override
         {
             return false;
         }
@@ -250,7 +234,7 @@ namespace VFS
         virtual ~FileInode() = default;
     };
 
-    class DeviceInode : public Inode
+    class DeviceInode : Inode
     {
       public:
         Result<SharedPtr<Inode>> find(const char*) const override
@@ -263,12 +247,12 @@ namespace VFS
             return {};
         }
 
-        Result<SharedPtr<Inode>> create_file(const char*, mode_t) override
+        Result<SharedPtr<Inode>> create_file(const char*) override
         {
             return err(ENOTDIR);
         }
 
-        Result<SharedPtr<Inode>> create_subdirectory(const char*, mode_t) override
+        Result<SharedPtr<Inode>> create_subdirectory(const char*) override
         {
             return err(ENOTDIR);
         }
@@ -305,10 +289,10 @@ namespace VFS
                                           SharedPtr<VFS::Inode> working_directory = {},
                                           bool follow_last_symlink = true);
 
-    Result<SharedPtr<Inode>> create_directory(const char* path, mode_t mode, Credentials auth,
+    Result<SharedPtr<Inode>> create_directory(const char* path, Credentials auth,
                                               SharedPtr<VFS::Inode> working_directory = {});
 
-    Result<SharedPtr<Inode>> create_file(const char* path, mode_t mode, Credentials auth,
+    Result<SharedPtr<Inode>> create_file(const char* path, Credentials auth,
                                          SharedPtr<VFS::Inode> working_directory = {});
 
     Result<void> validate_filename(StringView name);

kernel/src/fs/devices/BlockDevice.cpp

@@ -1,51 +0,0 @@
-#include "fs/devices/BlockDevice.h"
-#include "arch/MMU.h"
-#include <luna/Common.h>
-
-BlockDevice::BlockDevice(u64 block_size, u64 num_blocks) : m_cache(), m_block_size(block_size), m_num_blocks(num_blocks)
-{
-}
-
-Result<usize> BlockDevice::read(u8* buf, usize offset, usize length) const
-{
-    if (length == 0) return 0;
-
-    if (offset > size()) return 0;
-    if (offset + length > size()) length = size() - offset;
-
-    const usize saved = length;
-
-    auto read_data = [&](u64 off, u64 len) -> Result<void> {
-        const u64 block = offset / m_block_size;
-
-        m_cache.lock();
-        auto guard = make_scope_guard([&] { m_cache.unlock(); });
-
-        auto* entry = TRY(m_cache.fetch_entry(block));
-        if (entry->buffer.is_empty())
-        {
-            // TODO: What if the cache needs clearing in the middle of try_resize()? That could lead to some weird
-            // state...
-            TRY(entry->buffer.try_resize(m_block_size));
-            TRY(read_block(entry->buffer, block));
-        }
-        memcpy(buf, entry->buffer.data() + off, len);
-        offset += len;
-        length -= len;
-        buf += len;
-        return {};
-    };
-
-    if (offset % m_block_size)
-    {
-        const u64 off = offset % m_block_size;
-        const u64 len = min(m_block_size - off, length);
-        TRY(read_data(off, len));
-    }
-
-    while (length >= m_block_size) { TRY(read_data(0, m_block_size)); }
-
-    if (length) { TRY(read_data(0, length)); }
-
-    return saved;
-}

kernel/src/fs/devices/BlockDevice.h

@@ -1,40 +0,0 @@
-#pragma once
-#include "fs/StorageCache.h"
-#include "fs/devices/Device.h"
-
-class BlockDevice : public Device
-{
-  public:
-    BlockDevice(u64 block_size, u64 num_blocks);
-
-    Result<usize> read(u8* buf, usize offset, usize length) const override;
-
-    virtual Result<void> read_block(Buffer& buf, u64 block) const = 0;
-
-    Result<usize> write(const u8*, usize, usize) override
-    {
-        return err(ENOTSUP);
-    }
-
-    usize size() const override
-    {
-        return m_block_size * m_num_blocks;
-    }
-
-    bool is_block_device() const override
-    {
-        return true;
-    }
-
-    Result<usize> block_size() const override
-    {
-        return m_block_size;
-    }
-
-    virtual ~BlockDevice() = default;
-
-  protected:
-    mutable StorageCache m_cache;
-    u64 m_block_size;
-    u64 m_num_blocks;
-};

kernel/src/fs/devices/ConsoleDevice.cpp

@@ -1,8 +1,6 @@
 #include "fs/devices/ConsoleDevice.h"
 #include "Log.h"
-#include "Pledge.h"
 #include "fs/devices/DeviceRegistry.h"
-#include "fs/devices/KeyboardDevice.h"
 #include "memory/MemoryManager.h"
 #include "thread/Scheduler.h"
 #include "video/TextConsole.h"
@@ -13,7 +11,6 @@
 #include <luna/Vector.h>
 
 Vector<SharedPtr<ConsoleDevice>> ConsoleDevice::m_console_devices;
-bool ConsoleDevice::s_is_in_graphical_mode { false };
 
 Result<void> ConsoleDevice::create()
 {
@@ -34,9 +31,9 @@ Result<void> ConsoleDevice::handle_background_process_group(bool can_succeed, in
     auto foreground_pgrp = m_foreground_process_group.value();
 
     auto* current = Scheduler::current();
-    if (current->pgid == foreground_pgrp) return {};
+    if ((pid_t)current->pgid == foreground_pgrp) return {};
 
-    if ((current->signal_mask.get(signo - 1)) || (current->signal_handlers[signo - 1].sa_handler == SIG_IGN))
+    if ((current->signal_mask & (1 << (signo - 1))) || (current->signal_handlers[signo - 1].sa_handler == SIG_IGN))
     {
         if (can_succeed) return {};
         return err(EIO);
@@ -52,7 +49,13 @@ Result<usize> ConsoleDevice::read(u8* buf, usize, usize length) const
 {
     TRY(handle_background_process_group(false, SIGTTIN));
 
-    length = m_input_buffer.dequeue_data(buf, length);
+    if (length > m_input_buffer.size()) length = m_input_buffer.size();
+
+    memcpy(buf, m_input_buffer.data(), length);
+
+    memmove(m_input_buffer.data(), m_input_buffer.data() + length, m_input_buffer.size() - length);
+
+    m_input_buffer.try_resize(m_input_buffer.size() - length).release_value();
 
     if (!length && m_may_read_without_blocking) m_may_read_without_blocking = false;
 
@@ -63,27 +66,18 @@ Result<usize> ConsoleDevice::write(const u8* buf, usize, usize length)
 {
     if (m_settings.c_lflag & TOSTOP) TRY(handle_background_process_group(true, SIGTTOU));
 
-    if (s_is_in_graphical_mode) return length;
-
     TextConsole::write((const char*)buf, length);
     return length;
 }
 
-bool ConsoleDevice::will_block_if_read() const
+bool ConsoleDevice::blocking() const
 {
     return m_may_read_without_blocking ? false : m_input_buffer.size() == 0;
 }
 
 void ConsoleDevice::did_press_or_release_key(u8 scancode)
 {
-    if (!s_is_in_graphical_mode)
-        for (const auto& device : m_console_devices) { device->process_key_event(scancode); }
-    else
-    {
-        static Keyboard::KeyboardState state = {};
-        auto packet = Keyboard::decode_scancode(scancode, state);
-        if (packet.has_value()) KeyboardDevice::add_keyboard_event(packet.release_value());
-    }
+    for (const auto& device : m_console_devices) { device->process_key_event(scancode); }
 }
 
 void ConsoleDevice::process_key_event(u8 scancode)
@@ -106,7 +100,6 @@ void ConsoleDevice::process_key_event(u8 scancode)
                 {
                     TextConsole::putwchar(L'\b');
                     if (_iscntrl(maybe_char.value())) TextConsole::putwchar(L'\b');
-                    if (maybe_char.value() == '\t') TextConsole::wprint(L"\b\b");
                 }
                 return;
             }
@@ -180,7 +173,7 @@ void ConsoleDevice::process_key_event(u8 scancode)
         if (m_settings.c_lflag & ECHOCTL)
         {
             bool should_echo = true;
-            if (key == '\n' || key == '\t' || key == '\0' || key == m_settings.c_cc[VEOF]) should_echo = false;
+            if (key == '\n' || key == '\0' || key == m_settings.c_cc[VEOF]) should_echo = false;
 
             if (should_echo)
             {
@@ -201,9 +194,6 @@ void ConsoleDevice::process_key_event(u8 scancode)
 
 Result<u64> ConsoleDevice::ioctl(int request, void* arg)
 {
-    auto* current = Scheduler::current();
-    TRY(check_pledge(current, Promise::p_tty));
-
     switch (request)
     {
     case TCGETS: {
@@ -245,13 +235,6 @@ Result<u64> ConsoleDevice::ioctl(int request, void* arg)
         if (!MemoryManager::copy_to_user_typed((struct winsize*)arg, &window)) return err(EFAULT);
         return 0;
     }
-    case TTYSETGFX: {
-        s_is_in_graphical_mode = (bool)arg;
-        if (!s_is_in_graphical_mode) TextConsole::enable_cursor();
-        else
-            TextConsole::disable_cursor();
-        return 0;
-    }
     default: return err(EINVAL);
     }
 }

kernel/src/fs/devices/ConsoleDevice.h

@@ -16,7 +16,7 @@ class ConsoleDevice : public Device
 
     static void did_press_or_release_key(u8 scancode);
 
-    bool will_block_if_read() const override;
+    bool blocking() const override;
 
     Result<u64> ioctl(int request, void* arg) override;
 
@@ -37,10 +37,9 @@ class ConsoleDevice : public Device
     mutable Buffer m_input_buffer;
     Option<pid_t> m_foreground_process_group {};
     Vector<u8> m_line_buffer;
-    mutable Keyboard::TTYKeyboardState m_kb_state;
+    mutable Keyboard::KeyboardState m_kb_state;
 
     static Vector<SharedPtr<ConsoleDevice>> m_console_devices;
-    static bool s_is_in_graphical_mode;
 
     void process_key_event(u8 scancode);
 

kernel/src/fs/devices/Device.h

@@ -1,22 +1,14 @@
 #pragma once
 #include "Log.h"
 #include <luna/Result.h>
-#include <luna/SharedPtr.h>
-#include <luna/StringView.h>
-#include <sys/types.h>
 
-class Device : public Shareable
+class Device
 {
   public:
     virtual Result<usize> read(u8* buf, usize offset, usize length) const = 0;
 
     virtual Result<usize> write(const u8* buf, usize offset, usize length) = 0;
 
-    virtual Result<u64> query_shared_memory(off_t, usize)
-    {
-        return err(EACCES);
-    }
-
     virtual Result<u64> ioctl(int, void*)
     {
         return err(ENOTTY);
@@ -48,10 +40,7 @@ class Device : public Shareable
     // Path in devfs.
     virtual StringView device_path() const = 0;
 
-    virtual bool will_block_if_read() const = 0;
+    virtual bool blocking() const = 0;
 
     virtual ~Device() = default;
-
-  protected:
-    Option<u64> m_shmid;
 };

kernel/src/fs/devices/DeviceRegistry.cpp

@@ -4,8 +4,6 @@
 #include "fs/devices/ConsoleDevice.h"
 #include "fs/devices/FramebufferDevice.h"
 #include "fs/devices/FullDevice.h"
-#include "fs/devices/KeyboardDevice.h"
-#include "fs/devices/MouseDevice.h"
 #include "fs/devices/NullDevice.h"
 #include "fs/devices/UARTDevice.h"
 #include "fs/devices/ZeroDevice.h"
@@ -38,7 +36,8 @@ namespace DeviceRegistry
 
     Result<void> create_special_device_inode(SharedPtr<VFS::FileSystem> fs, const DeviceDescriptor& descriptor)
     {
-        auto inode = TRY(fs->create_device_inode(descriptor.major, descriptor.minor, descriptor.mode));
+        auto inode = TRY(fs->create_device_inode(descriptor.major, descriptor.minor));
+        inode->chmod(descriptor.mode);
         TRY(fs->root_inode()->add_entry(inode, descriptor.name));
 
         return {};
@@ -72,8 +71,6 @@ namespace DeviceRegistry
         ConsoleDevice::create();
         FramebufferDevice::create();
         UARTDevice::create();
-        MouseDevice::create();
-        KeyboardDevice::create();
 
         return {};
     }

kernel/src/fs/devices/DeviceRegistry.h

@@ -16,7 +16,6 @@ namespace DeviceRegistry
         Disk = 4,
         DiskPartition = 5,
         Serial = 6,
-        Input = 7,
     };
 
     Result<SharedPtr<Device>> fetch_special_device(u32 major, u32 minor);

kernel/src/fs/devices/FramebufferDevice.cpp

@@ -1,6 +1,4 @@
 #include "fs/devices/FramebufferDevice.h"
-#include "arch/MMU.h"
-#include "memory/SharedMemory.h"
 #include "video/Framebuffer.h"
 #include <bits/ioctl-defs.h>
 #include <luna/CString.h>
@@ -25,43 +23,12 @@ Result<usize> FramebufferDevice::write(const u8* buf, usize offset, usize length
     return length;
 }
 
-Result<u64> FramebufferDevice::query_shared_memory(off_t offset, usize count)
-{
-    if (offset + (count * ARCH_PAGE_SIZE) > Framebuffer::size()) return err(EINVAL);
-
-    if (!m_shmid.has_value())
-    {
-        u64 shmid = TRY(SharedMemory::create(Framebuffer::ptr() + offset, offset, count));
-        m_shmid = shmid;
-        auto* shm = g_shared_memory_map.try_get_ref(shmid);
-        shm->device = SharedPtr<Device> { this };
-        return shmid;
-    }
-
-    auto* shm = g_shared_memory_map.try_get_ref(*m_shmid);
-    if (!shm)
-    {
-        m_shmid = {};
-        return query_shared_memory(offset, count);
-    }
-    if (shm->offset > offset)
-    {
-        TRY(shm->grow_backward(Framebuffer::ptr() + offset, (shm->offset - offset) / ARCH_PAGE_SIZE));
-    }
-    if (shm->frames.size() < count)
-    {
-        TRY(shm->grow_forward(Framebuffer::ptr() + offset + (shm->frames.size() * ARCH_PAGE_SIZE),
-                              count - shm->frames.size()));
-    }
-    return *m_shmid;
-}
-
 usize FramebufferDevice::size() const
 {
     return Framebuffer::size();
 }
 
-bool FramebufferDevice::will_block_if_read() const
+bool FramebufferDevice::blocking() const
 {
     return false;
 }

kernel/src/fs/devices/FramebufferDevice.h

@@ -11,9 +11,7 @@ class FramebufferDevice : public Device
 
     Result<usize> write(const u8*, usize, usize) override;
 
-    Result<u64> query_shared_memory(off_t offset, usize count) override;
-
-    bool will_block_if_read() const override;
+    bool blocking() const override;
 
     bool is_block_device() const override
     {

kernel/src/fs/devices/FullDevice.h

@@ -19,7 +19,7 @@ class FullDevice : public Device
         return err(ENOSPC);
     }
 
-    bool will_block_if_read() const override
+    bool blocking() const override
     {
         return false;
     }

kernel/src/fs/devices/KeyboardDevice.cpp

@@ -1,34 +0,0 @@
-#include "fs/devices/KeyboardDevice.h"
-
-SharedPtr<KeyboardDevice> KeyboardDevice::s_keyboard_device = {};
-
-Result<void> KeyboardDevice::create()
-{
-    auto device = TRY(make_shared<KeyboardDevice>());
-    s_keyboard_device = device;
-    return DeviceRegistry::register_special_device(DeviceRegistry::Input, 1, device, 0600);
-}
-
-Result<usize> KeyboardDevice::read(u8* buf, usize, usize length) const
-{
-    length = m_packet_buffer.dequeue_data(buf, length);
-
-    return length;
-}
-
-Result<usize> KeyboardDevice::write(const u8* buf, usize, usize length)
-{
-    TRY(m_packet_buffer.append_data(buf, length));
-
-    return length;
-}
-
-void KeyboardDevice::add_keyboard_event(const moon::KeyboardPacket& packet)
-{
-    if (s_keyboard_device) s_keyboard_device->write((const u8*)&packet, 0, sizeof(packet));
-}
-
-bool KeyboardDevice::will_block_if_read() const
-{
-    return !m_packet_buffer.size();
-}

kernel/src/fs/devices/KeyboardDevice.h

@@ -1,31 +0,0 @@
-#pragma once
-#include "api/Keyboard.h"
-#include "fs/devices/DeviceRegistry.h"
-#include <luna/Buffer.h>
-
-class KeyboardDevice : public Device
-{
-  public:
-    // Initializer for DeviceRegistry.
-    static Result<void> create();
-
-    Result<usize> read(u8*, usize, usize) const override;
-
-    Result<usize> write(const u8*, usize, usize) override;
-
-    static void add_keyboard_event(const moon::KeyboardPacket& packet);
-
-    bool will_block_if_read() const override;
-
-    StringView device_path() const override
-    {
-        return "kbd";
-    }
-
-    virtual ~KeyboardDevice() = default;
-
-  private:
-    mutable Buffer m_packet_buffer;
-
-    static SharedPtr<KeyboardDevice> s_keyboard_device;
-};

kernel/src/fs/devices/MouseDevice.cpp

@@ -1,50 +0,0 @@
-#include "fs/devices/MouseDevice.h"
-
-SharedPtr<MouseDevice> MouseDevice::s_mouse_device = {};
-
-Result<void> MouseDevice::create()
-{
-    auto device = TRY(make_shared<MouseDevice>());
-    s_mouse_device = device;
-    return DeviceRegistry::register_special_device(DeviceRegistry::Input, 0, device, 0600);
-}
-
-Result<usize> MouseDevice::read(u8* buf, usize, usize length) const
-{
-    usize nread = 0;
-    while (length >= sizeof(moon::MousePacket))
-    {
-        if (!m_packet_queue.try_pop(*(moon::MousePacket*)buf)) break;
-
-        buf += sizeof(moon::MousePacket);
-        length -= sizeof(moon::MousePacket);
-        nread += sizeof(moon::MousePacket);
-    }
-
-    return nread;
-}
-
-Result<usize> MouseDevice::write(const u8* buf, usize, usize length)
-{
-    usize nwritten = 0;
-    while (length >= sizeof(moon::MousePacket))
-    {
-        if (!m_packet_queue.try_push(*(const moon::MousePacket*)buf)) break;
-
-        buf += sizeof(moon::MousePacket);
-        length -= sizeof(moon::MousePacket);
-        nwritten += sizeof(moon::MousePacket);
-    }
-
-    return nwritten;
-}
-
-void MouseDevice::add_mouse_event(const moon::MousePacket& packet)
-{
-    if (s_mouse_device) s_mouse_device->m_packet_queue.try_push(packet);
-}
-
-bool MouseDevice::will_block_if_read() const
-{
-    return m_packet_queue.is_empty();
-}

kernel/src/fs/devices/MouseDevice.h

@@ -1,31 +0,0 @@
-#pragma once
-#include "api/Mouse.h"
-#include "fs/devices/DeviceRegistry.h"
-#include <luna/CircularQueue.h>
-
-class MouseDevice : public Device
-{
-  public:
-    // Initializer for DeviceRegistry.
-    static Result<void> create();
-
-    Result<usize> read(u8*, usize, usize) const override;
-
-    Result<usize> write(const u8*, usize, usize) override;
-
-    static void add_mouse_event(const moon::MousePacket& packet);
-
-    bool will_block_if_read() const override;
-
-    StringView device_path() const override
-    {
-        return "mouse";
-    }
-
-    virtual ~MouseDevice() = default;
-
-  private:
-    mutable CircularQueue<moon::MousePacket, 200> m_packet_queue;
-
-    static SharedPtr<MouseDevice> s_mouse_device;
-};

kernel/src/fs/devices/NullDevice.h

@@ -17,7 +17,7 @@ class NullDevice : public Device
         return 0;
     }
 
-    bool will_block_if_read() const override
+    bool blocking() const override
     {
         return false;
     }

kernel/src/fs/devices/UARTDevice.h

@@ -14,7 +14,7 @@ class UARTDevice : public Device
 
     Result<usize> write(const u8*, usize, usize) override;
 
-    bool will_block_if_read() const override
+    bool blocking() const override
     {
         return false;
     }

kernel/src/fs/devices/ZeroDevice.h

@@ -19,7 +19,7 @@ class ZeroDevice : public Device
         return 0;
     }
 
-    bool will_block_if_read() const override
+    bool blocking() const override
     {
         return false;
     }

kernel/src/fs/ext2/FileSystem.cpp

@@ -1,6 +1,6 @@
 #include "fs/ext2/FileSystem.h"
 #include "fs/ext2/Inode.h"
-#include <luna/Common.h>
+#include <luna/Alignment.h>
 
 static VFS::InodeType vfs_type_from_ext2_type(mode_t mode)
 {
@@ -49,21 +49,11 @@ namespace Ext2
         auto inode = TRY(adopt_shared_if_nonnull(new (std::nothrow) Ext2::Inode({}, this)));
         TRY(m_host_device->read((u8*)&inode->m_raw_inode, inode_address, INODE_SIZE));
         inode->m_type = vfs_type_from_ext2_type(inode->m_raw_inode.mode);
-        inode->m_metadata.inum = inum;
-        inode->m_metadata.size = (m_uses_extended_size && (inode->m_type == VFS::InodeType::RegularFile))
-                                     ? ((u64)inode->m_raw_inode.size_high << 32) | (u64)inode->m_raw_inode.size_low
-                                     : inode->m_raw_inode.size_low;
-        inode->m_metadata.mode = inode->m_raw_inode.mode & 07777;
-        inode->m_metadata.nlinks = inode->m_raw_inode.nlinks;
-        inode->m_metadata.uid = inode->m_raw_inode.uid;
-        inode->m_metadata.gid = inode->m_raw_inode.gid;
-        inode->m_metadata.atime = { .tv_sec = inode->m_raw_inode.atime, .tv_nsec = 0 };
-        inode->m_metadata.mtime = { .tv_sec = inode->m_raw_inode.mtime, .tv_nsec = 0 };
-        inode->m_metadata.ctime = { .tv_sec = inode->m_raw_inode.create_time, .tv_nsec = 0 };
+        inode->m_inum = inum;
 
 #ifdef EXT2_DEBUG
         kdbgln("ext2: Read inode %lu with mode %#x (%#x + %#o), size %lu", inum, inode->m_raw_inode.mode,
-               inode->m_raw_inode.mode & 0xf000, inode->metadata().mode, inode->metadata().size);
+               inode->m_raw_inode.mode & 0xf000, inode->mode(), inode->size());
 #endif
 
         m_inode_cache.try_set(inum, inode);
@@ -117,7 +107,7 @@ namespace Ext2
         fs->m_host_device = host_device;
 
         fs->m_block_size = 1024 << fs->m_superblock.log_block_size;
-        fs->m_block_groups = ceil_div(fs->m_superblock.nr_blocks, fs->m_superblock.blocks_per_block_group);
+        fs->m_block_groups = get_blocks_from_size(fs->m_superblock.nr_blocks, fs->m_superblock.blocks_per_block_group);
 
 #ifdef EXT2_DEBUG
         kdbgln("ext2: Mounting new Ext2 file system, block size=%lu, blocks=%u, inodes=%u, block group=(%u blocks, %u "

kernel/src/fs/ext2/FileSystem.h

@@ -127,17 +127,17 @@ namespace Ext2
             return m_root_inode;
         }
 
-        Result<SharedPtr<VFS::Inode>> create_file_inode(mode_t) override
+        Result<SharedPtr<VFS::Inode>> create_file_inode() override
         {
             return err(EROFS);
         }
 
-        Result<SharedPtr<VFS::Inode>> create_dir_inode(SharedPtr<VFS::Inode>, mode_t) override
+        Result<SharedPtr<VFS::Inode>> create_dir_inode(SharedPtr<VFS::Inode>) override
         {
             return err(EROFS);
         }
 
-        Result<SharedPtr<VFS::Inode>> create_device_inode(u32, u32, mode_t) override
+        Result<SharedPtr<VFS::Inode>> create_device_inode(u32, u32) override
         {
             return err(EROFS);
         }

kernel/src/fs/ext2/Inode.cpp

@@ -7,12 +7,19 @@ namespace Ext2
     {
     }
 
+    usize Inode::size() const
+    {
+        return (m_fs->m_uses_extended_size && (m_type == VFS::InodeType::RegularFile))
+                   ? ((u64)m_raw_inode.size_high << 32) | (u64)m_raw_inode.size_low
+                   : m_raw_inode.size_low;
+    }
+
     Result<usize> Inode::read(u8* buf, usize offset, usize length) const
     {
         if (length == 0) return 0;
 
-        if (offset > m_metadata.size) return 0;
-        if (offset + length > m_metadata.size) length = m_metadata.size - offset;
+        if (offset > size()) return 0;
+        if (offset + length > size()) length = size() - offset;
 
         const usize block_size = m_fs->m_block_size;
 
@@ -88,21 +95,22 @@ namespace Ext2
     {
         check(m_type == VFS::InodeType::Directory);
 
-        const usize inode_size = m_metadata.size;
+        const usize inode_size = size();
         const usize block_size = m_fs->m_block_size;
 
-        auto buf = TRY(Buffer::create_sized(block_size));
+        u8* const buf = TRY(make_array<u8>(block_size));
+        auto guard = make_scope_guard([buf] { delete[] buf; });
 
         m_entries.clear();
 
         for (usize offset = 0; offset < inode_size; offset += block_size)
         {
-            TRY(read(buf.data(), offset, block_size));
+            TRY(read(buf, offset, block_size));
 
             usize dir_offset = 0;
             while (dir_offset < block_size)
             {
-                auto& entry = *(Ext2::RawDirectoryEntry*)&buf.data()[dir_offset];
+                auto& entry = *(Ext2::RawDirectoryEntry*)&buf[dir_offset];
 
                 if (entry.inum != 0)
                 {
@@ -176,7 +184,7 @@ namespace Ext2
 
         if (!m_link.is_empty()) return m_link.view();
 
-        const usize length = m_metadata.size;
+        const usize length = size();
 
         if (length < 60)
         {

kernel/src/fs/ext2/Inode.h

@@ -21,6 +21,33 @@ namespace Ext2
             return m_type;
         }
 
+        usize size() const override;
+
+        mode_t mode() const override
+        {
+            return m_raw_inode.mode & 07777;
+        }
+
+        nlink_t nlinks() const override
+        {
+            return m_raw_inode.nlinks;
+        }
+
+        u32 uid() const override
+        {
+            return m_raw_inode.uid;
+        }
+
+        u32 gid() const override
+        {
+            return m_raw_inode.gid;
+        }
+
+        usize inode_number() const override
+        {
+            return m_inum;
+        }
+
         VFS::FileSystem* fs() const override
         {
             return m_fs;
@@ -34,12 +61,12 @@ namespace Ext2
         {
         }
 
-        Result<u64> query_shared_memory(off_t, usize) override
+        Result<void> chmod(mode_t) override
         {
-            return err(ENOTSUP);
+            return err(EROFS);
         }
 
-        Result<void> set_metadata(const VFS::InodeMetadata&) override
+        Result<void> chown(u32, u32) override
         {
             return err(EROFS);
         }
@@ -60,14 +87,14 @@ namespace Ext2
 
         Option<VFS::DirectoryEntry> get(usize) const override;
 
-        Result<SharedPtr<VFS::Inode>> create_file(const char*, mode_t) override
+        Result<SharedPtr<VFS::Inode>> create_file(const char*) override
         {
             if (m_type != VFS::InodeType::Directory) return err(ENOTDIR);
 
             return err(EROFS);
         }
 
-        Result<SharedPtr<VFS::Inode>> create_subdirectory(const char*, mode_t) override
+        Result<SharedPtr<VFS::Inode>> create_subdirectory(const char*) override
         {
             if (m_type != VFS::InodeType::Directory) return err(ENOTDIR);
 
@@ -95,7 +122,7 @@ namespace Ext2
             return m_entries.size();
         }
 
-        bool will_block_if_read() const override
+        bool blocking() const override
         {
             return false;
         }

kernel/src/fs/tmpfs/FileSystem.cpp

@@ -1,5 +1,4 @@
 #include "fs/tmpfs/FileSystem.h"
-#include "arch/Timer.h"
 #include "fs/devices/DeviceRegistry.h"
 #include "fs/tmpfs/Inode.h"
 #include <luna/Alloc.h>
@@ -11,23 +10,17 @@ namespace TmpFS
     Result<SharedPtr<VFS::FileSystem>> FileSystem::create()
     {
         SharedPtr<FileSystem> fs = TRY(adopt_shared_if_nonnull(new (std::nothrow) FileSystem()));
-        SharedPtr<VFS::Inode> root = TRY(fs->create_dir_inode({}, 0755));
+        SharedPtr<VFS::Inode> root = TRY(fs->create_dir_inode({}));
+        root->chmod(0755);
         fs->set_root(root);
         return (SharedPtr<VFS::FileSystem>)fs;
     }
 
-    Result<u64> FileSystem::allocate_inode_number()
-    {
-        return m_next_inode_number++;
-    }
-
-    Result<SharedPtr<VFS::Inode>> FileSystem::create_file_inode(mode_t mode)
+    Result<SharedPtr<VFS::Inode>> FileSystem::create_file_inode()
     {
         SharedPtr<FileInode> inode = TRY(make_shared<FileInode>());
         inode->set_fs(*this, {});
         inode->set_inode_number(m_next_inode_number++, {});
-        inode->m_metadata.mode = mode;
-        inode->m_metadata.atime = inode->m_metadata.ctime = inode->m_metadata.mtime = *Timer::realtime_clock();
         return (SharedPtr<VFS::Inode>)inode;
     }
 
@@ -37,11 +30,10 @@ namespace TmpFS
         inode->set_fs(*this, {});
         TRY(inode->set_link(link, {}));
         inode->set_inode_number(m_next_inode_number++, {});
-        inode->m_metadata.atime = inode->m_metadata.ctime = inode->m_metadata.mtime = *Timer::realtime_clock();
         return (SharedPtr<VFS::Inode>)inode;
     }
 
-    Result<SharedPtr<VFS::Inode>> FileSystem::create_dir_inode(SharedPtr<VFS::Inode> parent, mode_t mode)
+    Result<SharedPtr<VFS::Inode>> FileSystem::create_dir_inode(SharedPtr<VFS::Inode> parent)
     {
         SharedPtr<DirInode> inode = TRY(make_shared<DirInode>());
 
@@ -51,13 +43,11 @@ namespace TmpFS
         inode->set_self(inode, {});
         inode->set_fs(*this, {});
         inode->set_inode_number(m_next_inode_number++, {});
-        inode->m_metadata.mode = mode;
-        inode->m_metadata.atime = inode->m_metadata.ctime = inode->m_metadata.mtime = *Timer::realtime_clock();
 
         return (SharedPtr<VFS::Inode>)inode;
     }
 
-    Result<SharedPtr<VFS::Inode>> FileSystem::create_device_inode(u32 major, u32 minor, mode_t mode)
+    Result<SharedPtr<VFS::Inode>> FileSystem::create_device_inode(u32 major, u32 minor)
     {
         SharedPtr<Device> device = TRY(DeviceRegistry::fetch_special_device(major, minor));
 
@@ -68,9 +58,6 @@ namespace TmpFS
         // FIXME: This should be queried from Device directly, but Device doesn't have an API to store and retrieve its
         // device ID atm.
         inode->set_device_id(luna_dev_makedev(major, minor), {});
-        inode->m_metadata.mode = mode;
-        inode->m_metadata.atime = inode->m_metadata.ctime = inode->m_metadata.mtime = *Timer::realtime_clock();
-        inode->m_metadata.size = device->size();
 
         return (SharedPtr<VFS::Inode>)inode;
     }

kernel/src/fs/tmpfs/FileSystem.h

@@ -13,13 +13,11 @@ namespace TmpFS
             return m_root_inode;
         }
 
-        Result<SharedPtr<VFS::Inode>> create_file_inode(mode_t mode) override;
-        Result<SharedPtr<VFS::Inode>> create_dir_inode(SharedPtr<VFS::Inode> parent, mode_t mode) override;
-        Result<SharedPtr<VFS::Inode>> create_device_inode(u32 major, u32 minor, mode_t mode) override;
+        Result<SharedPtr<VFS::Inode>> create_file_inode() override;
+        Result<SharedPtr<VFS::Inode>> create_dir_inode(SharedPtr<VFS::Inode> parent) override;
+        Result<SharedPtr<VFS::Inode>> create_device_inode(u32 major, u32 minor) override;
         Result<SharedPtr<VFS::Inode>> create_symlink_inode(StringView link) override;
 
-        Result<u64> allocate_inode_number() override;
-
         Result<void> set_mount_dir(SharedPtr<VFS::Inode> parent) override;
 
         Result<void> reset_mount_dir() override;

kernel/src/fs/tmpfs/Inode.cpp

@@ -1,6 +1,4 @@
 #include "fs/tmpfs/Inode.h"
-#include "arch/MMU.h"
-#include "memory/SharedMemory.h"
 
 namespace TmpFS
 {
@@ -45,8 +43,6 @@ namespace TmpFS
 
         inode->did_link();
 
-        m_metadata.mtime = *Timer::realtime_clock();
-
         return {};
     }
 
@@ -63,23 +59,21 @@ namespace TmpFS
 
         inode->did_unlink();
 
-        m_metadata.mtime = *Timer::realtime_clock();
-
         return {};
     }
 
-    Result<SharedPtr<VFS::Inode>> DirInode::create_file(const char* name, mode_t mode)
+    Result<SharedPtr<VFS::Inode>> DirInode::create_file(const char* name)
     {
-        auto inode = TRY(m_fs->create_file_inode(mode));
+        auto inode = TRY(m_fs->create_file_inode());
 
         TRY(add_entry(inode, name));
 
         return inode;
     }
 
-    Result<SharedPtr<VFS::Inode>> DirInode::create_subdirectory(const char* name, mode_t mode)
+    Result<SharedPtr<VFS::Inode>> DirInode::create_subdirectory(const char* name)
     {
-        auto inode = TRY(m_fs->create_dir_inode(m_self, mode));
+        auto inode = TRY(m_fs->create_dir_inode(m_self));
 
         TRY(add_entry(inode, name));
 
@@ -116,10 +110,6 @@ namespace TmpFS
         u8* slice = TRY(m_data_buffer.slice(offset, length));
         memcpy(slice, buf, length);
 
-        m_metadata.size = m_data_buffer.size();
-
-        m_metadata.mtime = *Timer::realtime_clock();
-
         return length;
     }
 
@@ -131,42 +121,11 @@ namespace TmpFS
 
         if (size > old_size) memset(m_data_buffer.data() + old_size, 0, size - old_size);
 
-        m_metadata.size = m_data_buffer.size();
-
-        m_metadata.mtime = *Timer::realtime_clock();
-
         return {};
     }
 
-    Result<u64> FileInode::query_shared_memory(off_t offset, usize count)
+    usize FileInode::size() const
     {
-        if (offset + (count * ARCH_PAGE_SIZE) > m_data_buffer.size()) return err(EINVAL);
-
-        if (!m_shmid.has_value())
-        {
-            u64 shmid = TRY(SharedMemory::create(m_data_buffer.data() + offset, offset, count));
-            m_shmid = shmid;
-            auto* shm = g_shared_memory_map.try_get_ref(shmid);
-            shm->inode = SharedPtr<VFS::Inode> { (VFS::Inode*)this };
-            return shmid;
-        }
-
-        auto* shm = g_shared_memory_map.try_get_ref(*m_shmid);
-        if (!shm)
-        {
-            m_shmid = {};
-            return query_shared_memory(offset, count);
-        }
-
-        if (shm->offset > offset)
-        {
-            TRY(shm->grow_backward(m_data_buffer.data() + offset, (shm->offset - offset) / ARCH_PAGE_SIZE));
-        }
-        if (shm->frames.size() < count)
-        {
-            TRY(shm->grow_forward(m_data_buffer.data() + offset + (shm->frames.size() * ARCH_PAGE_SIZE),
-                                  count - shm->frames.size()));
-        }
-        return *m_shmid;
+        return m_data_buffer.size();
     }
 }

kernel/src/fs/tmpfs/Inode.h

@@ -20,29 +20,68 @@ namespace TmpFS
 
         void set_inode_number(usize inum, Badge<FileSystem>)
         {
-            m_metadata.inum = inum;
+            m_inode_number = inum;
         }
 
         VFS::FileSystem* fs() const override
         {
             return m_fs;
         }
+
+        usize inode_number() const override
+        {
+            return m_inode_number;
+        }
+
         Result<usize> read(u8*, usize, usize) const override;
 
         Result<usize> write(const u8*, usize, usize) override;
 
         Result<void> truncate(usize size) override;
 
-        Result<u64> query_shared_memory(off_t offset, usize count) override;
+        usize size() const override;
+
+        mode_t mode() const override
+        {
+            return m_mode;
+        }
+
+        u32 uid() const override
+        {
+            return m_uid;
+        }
+
+        u32 gid() const override
+        {
+            return m_gid;
+        }
+
+        nlink_t nlinks() const override
+        {
+            return (nlink_t)m_nlinks;
+        }
+
+        Result<void> chmod(mode_t mode) override
+        {
+            m_mode = mode;
+            return {};
+        }
+
+        Result<void> chown(u32 uid, u32 gid) override
+        {
+            m_uid = uid;
+            m_gid = gid;
+            return {};
+        }
 
         void did_link() override
         {
-            m_metadata.nlinks++;
+            m_nlinks++;
         }
 
         void did_unlink() override
         {
-            m_metadata.nlinks--;
+            m_nlinks--;
         }
 
         virtual ~FileInode() = default;
@@ -50,8 +89,11 @@ namespace TmpFS
       private:
         VFS::FileSystem* m_fs;
         Buffer m_data_buffer;
-
-        friend class FileSystem;
+        usize m_inode_number;
+        mode_t m_mode;
+        u32 m_uid { 0 };
+        u32 m_gid { 0 };
+        u32 m_nlinks { 0 };
     };
 
     class SymlinkInode : public VFS::FileInode
@@ -71,7 +113,7 @@ namespace TmpFS
 
         void set_inode_number(usize inum, Badge<FileSystem>)
         {
-            m_metadata.inum = inum;
+            m_inode_number = inum;
         }
 
         Result<void> set_link(StringView link, Badge<FileSystem>)
@@ -85,6 +127,11 @@ namespace TmpFS
             return m_fs;
         }
 
+        usize inode_number() const override
+        {
+            return m_inode_number;
+        }
+
         Result<usize> read(u8*, usize, usize) const override
         {
             return err(ENOTSUP);
@@ -100,14 +147,51 @@ namespace TmpFS
             return err(ENOTSUP);
         }
 
+        usize size() const override
+        {
+            return m_link.length();
+        }
+
+        mode_t mode() const override
+        {
+            return 0777;
+        }
+
+        u32 uid() const override
+        {
+            return m_uid;
+        }
+
+        u32 gid() const override
+        {
+            return m_gid;
+        }
+
+        nlink_t nlinks() const override
+        {
+            return (nlink_t)m_nlinks;
+        }
+
+        Result<void> chmod(mode_t) override
+        {
+            return {};
+        }
+
+        Result<void> chown(u32 uid, u32 gid) override
+        {
+            m_uid = uid;
+            m_gid = gid;
+            return {};
+        }
+
         void did_link() override
         {
-            m_metadata.nlinks++;
+            m_nlinks++;
         }
 
         void did_unlink() override
         {
-            m_metadata.nlinks--;
+            m_nlinks--;
         }
 
         Result<StringView> readlink() override
@@ -120,8 +204,10 @@ namespace TmpFS
       private:
         VFS::FileSystem* m_fs;
         String m_link;
-
-        friend class FileSystem;
+        usize m_inode_number;
+        u32 m_uid { 0 };
+        u32 m_gid { 0 };
+        u32 m_nlinks { 0 };
     };
 
     class DeviceInode : public VFS::DeviceInode
@@ -141,7 +227,7 @@ namespace TmpFS
 
         void set_inode_number(usize inum, Badge<FileSystem>)
         {
-            m_metadata.inum = inum;
+            m_inode_number = inum;
         }
 
         void set_device(SharedPtr<Device> device, Badge<FileSystem>)
@@ -151,12 +237,7 @@ namespace TmpFS
 
         void set_device_id(dev_t id, Badge<FileSystem>)
         {
-            m_metadata.devid = id;
-        }
-
-        Result<u64> query_shared_memory(off_t offset, usize count) override
-        {
-            return m_device->query_shared_memory(offset, count);
+            m_device_id = id;
         }
 
         VFS::FileSystem* fs() const override
@@ -164,6 +245,16 @@ namespace TmpFS
             return m_fs;
         }
 
+        dev_t device_id() const override
+        {
+            return m_device_id;
+        }
+
+        usize inode_number() const override
+        {
+            return m_inode_number;
+        }
+
         Result<usize> read(u8* buf, usize offset, usize length) const override
         {
             return m_device->read(buf, offset, length);
@@ -190,19 +281,57 @@ namespace TmpFS
             return m_device->isatty();
         }
 
-        bool will_block_if_read() const override
+        bool blocking() const override
         {
-            return m_device->will_block_if_read();
+            return m_device->blocking();
+        }
+
+        usize size() const override
+        {
+            return m_device->size();
+        }
+
+        mode_t mode() const override
+        {
+            return m_mode;
+        }
+
+        u32 uid() const override
+        {
+            return m_uid;
+        }
+
+        u32 gid() const override
+        {
+            return m_gid;
+        }
+
+        nlink_t nlinks() const override
+        {
+            return (nlink_t)m_nlinks;
+        }
+
+        Result<void> chmod(mode_t mode) override
+        {
+            m_mode = mode;
+            return {};
+        }
+
+        Result<void> chown(u32 uid, u32 gid) override
+        {
+            m_uid = uid;
+            m_gid = gid;
+            return {};
         }
 
         void did_link() override
         {
-            m_metadata.nlinks++;
+            m_nlinks++;
         }
 
         void did_unlink() override
         {
-            m_metadata.nlinks--;
+            m_nlinks--;
         }
 
         virtual ~DeviceInode() = default;
@@ -210,8 +339,12 @@ namespace TmpFS
       private:
         VFS::FileSystem* m_fs;
         SharedPtr<Device> m_device;
-
-        friend class FileSystem;
+        usize m_inode_number;
+        mode_t m_mode;
+        u32 m_uid { 0 };
+        u32 m_gid { 0 };
+        u32 m_nlinks { 0 };
+        dev_t m_device_id { 0 };
     };
 
     class DirInode : public VFS::Inode
@@ -226,7 +359,7 @@ namespace TmpFS
 
         void set_inode_number(usize inum, Badge<FileSystem>)
         {
-            m_metadata.inum = inum;
+            m_inode_number = inum;
         }
 
         void set_self(SharedPtr<VFS::Inode> self, Badge<FileSystem>)
@@ -252,16 +385,54 @@ namespace TmpFS
             return err(EISDIR);
         }
 
-        bool will_block_if_read() const override
+        bool blocking() const override
         {
             return false;
         }
 
+        usize size() const override
+        {
+            return 0;
+        }
+
+        mode_t mode() const override
+        {
+            return m_mode;
+        }
+
+        u32 uid() const override
+        {
+            return m_uid;
+        }
+
+        u32 gid() const override
+        {
+            return m_gid;
+        }
+
+        Result<void> chmod(mode_t mode) override
+        {
+            m_mode = mode;
+            return {};
+        }
+
+        Result<void> chown(u32 uid, u32 gid) override
+        {
+            m_uid = uid;
+            m_gid = gid;
+            return {};
+        }
+
         VFS::FileSystem* fs() const override
         {
             return m_fs;
         }
 
+        usize inode_number() const override
+        {
+            return m_inode_number;
+        }
+
         VFS::InodeType type() const override
         {
             return VFS::InodeType::Directory;
@@ -284,8 +455,8 @@ namespace TmpFS
 
         Result<void> remove_entry(const char* name) override;
 
-        Result<SharedPtr<VFS::Inode>> create_file(const char* name, mode_t mode) override;
-        Result<SharedPtr<VFS::Inode>> create_subdirectory(const char* name, mode_t mode) override;
+        Result<SharedPtr<VFS::Inode>> create_file(const char* name) override;
+        Result<SharedPtr<VFS::Inode>> create_subdirectory(const char* name) override;
 
         Result<void> add_entry(SharedPtr<VFS::Inode> inode, const char* name);
         Result<void> replace_entry(SharedPtr<VFS::Inode> inode, const char* name);
@@ -294,9 +465,13 @@ namespace TmpFS
 
       private:
         VFS::FileSystem* m_fs;
-        SharedPtr<VFS::Inode> m_self;
-        Vector<VFS::DirectoryEntry> m_entries;
+        usize m_inode_number;
+        mode_t m_mode;
+        u32 m_uid { 0 };
+        u32 m_gid { 0 };
 
-        friend class FileSystem;
+        SharedPtr<VFS::Inode> m_self;
+
+        Vector<VFS::DirectoryEntry> m_entries;
     };
 }

kernel/src/main.cpp

@@ -1,7 +1,6 @@
 #include "Log.h"
 #include "arch/CPU.h"
 #include "arch/Timer.h"
-#include "binfmt/BinaryFormat.h"
 #include "boot/Init.h"
 #include "config.h"
 #include "fs/InitRD.h"
@@ -31,56 +30,38 @@ void reap_thread()
     }
 }
 
-void oom_thread()
-{
-    while (true)
-    {
-        kernel_wait_for_event();
-        // OOM! Do everything we can to recover memory.
-        StorageCache::clear_caches();
-    }
-}
-
 [[noreturn]] void init()
 {
-    {
-        kinfoln("Starting Moon %s %s", MOON_VERSION, MOON_RELEASE);
+    kinfoln("Starting Moon %s, built on %s at %s", MOON_VERSION, __DATE__, __TIME__);
 
-        // Default hostname if nobody from userspace changes it
-        set_host_name("moon"_sv);
+    // Default hostname if nobody from userspace changes it
+    set_host_name("moon"_sv);
 
-        kinfoln("Current platform: %s", CPU::platform_string().chars());
-        kinfoln("Current processor: %s", CPU::identify().value_or("(unknown)"_sv).chars());
+    kinfoln("Current platform: %s", CPU::platform_string().chars());
+    kinfoln("Current processor: %s", CPU::identify().value_or("(unknown)"_sv).chars());
 
-        auto root = mark_critical(TmpFS::FileSystem::create(), "Failed to create initial ramfs");
-        mark_critical(VFS::mount_root(root), "Failed to mount the initial ramfs as the root filesystem");
-        mark_critical(InitRD::populate_vfs(), "Failed to load files from the initial ramdisk");
-        mark_critical(DeviceRegistry::init(), "Failed to register initial devices");
+    auto root = mark_critical(TmpFS::FileSystem::create(), "Failed to create initial ramfs");
+    mark_critical(VFS::mount_root(root), "Failed to mount the initial ramfs as the root filesystem");
+    mark_critical(InitRD::populate_vfs(), "Failed to load files from the initial ramdisk");
+    mark_critical(DeviceRegistry::init(), "Failed to register initial devices");
 
-        mark_critical(BinaryFormat::init(), "Failed to register initial binary formats");
+    auto init =
+        mark_critical(VFS::resolve_path("/bin/preinit", Credentials {}), "Can't find init in the initial ramfs!");
+    auto init_thread =
+        mark_critical(Scheduler::new_userspace_thread(init, "/bin/preinit"), "Failed to create PID 1 process for init");
 
-        auto init =
-            mark_critical(VFS::resolve_path("/bin/preinit", Credentials {}), "Can't find init in the initial ramfs!");
-        auto init_thread = mark_critical(Scheduler::new_userspace_thread(init, "/bin/preinit"),
-                                         "Failed to create PID 1 process for init");
-
-        auto reap = mark_critical(Scheduler::new_kernel_thread(reap_thread, "[reap]"),
-                                  "Failed to create the process reaper kernel thread");
-        Scheduler::set_reap_thread(reap);
-
-        auto oom = mark_critical(Scheduler::new_kernel_thread(oom_thread, "[oom]"),
-                                 "Failed to create the out-of-memory kernel thread");
-        Scheduler::set_oom_thread(oom);
+    auto reap = mark_critical(Scheduler::new_kernel_thread(reap_thread, "[reap]"),
+                              "Failed to create the process reaper kernel thread");
+    Scheduler::set_reap_thread(reap);
 
 #ifdef ARCH_X86_64
-        ATA::Controller::scan();
+    ATA::Controller::scan();
 #endif
 
-        // Disable console logging before transferring control to userspace.
-        setup_log(log_debug_enabled(), log_serial_enabled(), false);
+    // Disable console logging before transferring control to userspace.
+    setup_log(log_debug_enabled(), log_serial_enabled(), false);
 
-        init_thread->wake_up();
-    }
+    init_thread->wake_up();
 
     kernel_exit();
 }

kernel/src/memory/AddressSpace.cpp

@@ -1,9 +1,7 @@
 #include "memory/AddressSpace.h"
+#include "Log.h"
 #include "arch/MMU.h"
 #include "memory/Heap.h"
-#include "memory/MemoryManager.h"
-#include "memory/SharedMemory.h"
-#include <bits/mmap-flags.h>
 #include <luna/CString.h>
 #include <luna/ScopeGuard.h>
 
@@ -51,26 +49,15 @@ Result<OwnedPtr<AddressSpace>> AddressSpace::clone()
 {
     OwnedPtr<AddressSpace> ptr = TRY(make_owned<AddressSpace>());
 
-    ptr->m_directory = TRY(MMU::clone_userspace_page_directory(m_directory));
-
     for (const auto* region : m_regions)
     {
         auto* new_region = TRY(make<VMRegion>());
         memcpy(new_region, region, sizeof(*region));
         ptr->m_regions.append(new_region);
-
-        if (new_region->used && new_region->prot != 0 && new_region->flags & MAP_SHARED)
-        {
-            TRY(MemoryManager::copy_region(new_region->start, new_region->count, m_directory, ptr->m_directory));
-            auto* shm = g_shared_memory_map.try_get_ref(new_region->shmid);
-            if (shm) shm->refs++;
-        }
-        else if (new_region->used && new_region->prot != 0)
-        {
-            TRY(MemoryManager::copy_region_data(new_region->start, new_region->count, m_directory, ptr->m_directory));
-        }
     }
 
+    ptr->m_directory = TRY(MMU::clone_userspace_page_directory(m_directory));
+
     return move(ptr);
 }
 
@@ -94,17 +81,8 @@ AddressSpace& AddressSpace::operator=(AddressSpace&& other)
     return *this;
 }
 
-Result<u64> AddressSpace::alloc_region(usize count, int prot, int flags, off_t offset, u64 shmid, bool persistent)
+Result<u64> AddressSpace::alloc_region(usize count, bool persistent)
 {
-    auto update_region = [=](VMRegion* region) {
-        region->used = true;
-        region->persistent = persistent;
-        region->prot = prot;
-        region->flags = flags;
-        region->offset = offset;
-        region->shmid = shmid;
-    };
-
     for (auto* region = m_regions.expect_last(); region; region = m_regions.previous(region).value_or(nullptr))
     {
         if (!region->used)
@@ -112,7 +90,8 @@ Result<u64> AddressSpace::alloc_region(usize count, int prot, int flags, off_t o
             if (region->count < count) continue;
             if (region->count == count)
             {
-                update_region(region);
+                region->used = true;
+                region->persistent = persistent;
                 u64 address = region->start;
                 try_merge_region_with_neighbors(region);
                 return address;
@@ -121,7 +100,8 @@ Result<u64> AddressSpace::alloc_region(usize count, int prot, int flags, off_t o
             u64 boundary = region->end - (count * ARCH_PAGE_SIZE);
 
             auto* new_region = TRY(split_region(region, boundary));
-            update_region(new_region);
+            new_region->used = true;
+            new_region->persistent = persistent;
             try_merge_region_with_neighbors(new_region);
 
             return boundary;
@@ -131,23 +111,12 @@ Result<u64> AddressSpace::alloc_region(usize count, int prot, int flags, off_t o
     return err(ENOMEM);
 }
 
-Result<bool> AddressSpace::set_region(u64 address, usize count, bool used, int prot, int flags, off_t offset, u64 shmid,
-                                      bool persistent)
+Result<bool> AddressSpace::set_region(u64 address, usize count, bool used, bool persistent)
 {
     if (address >= VM_END) return err(EINVAL);
 
     u64 end = address + (count * ARCH_PAGE_SIZE);
 
-    auto update_region = [=](VMRegion* region) {
-        if (!used) region->cleanup_shared();
-        region->used = used;
-        region->persistent = persistent;
-        region->prot = prot;
-        region->flags = flags;
-        region->offset = offset;
-        region->shmid = shmid;
-    };
-
     for (auto* region : m_regions)
     {
         if (region->end < address) continue;
@@ -162,13 +131,13 @@ Result<bool> AddressSpace::set_region(u64 address, usize count, bool used, int p
 
         if (region->start >= address && region->end <= end)
         {
-            update_region(region);
+            region->used = used;
+            region->persistent = persistent;
             if (region->start == address && region->end == end)
             {
                 try_merge_region_with_neighbors(region);
                 return true;
             }
-            try_merge_region_with_neighbors(region);
             continue;
         }
 
@@ -176,7 +145,8 @@ Result<bool> AddressSpace::set_region(u64 address, usize count, bool used, int p
         {
             auto* middle_region = TRY(split_region(region, address));
             TRY(split_region(middle_region, end));
-            update_region(middle_region);
+            middle_region->used = used;
+            middle_region->persistent = persistent;
             return true;
         }
 
@@ -184,7 +154,8 @@ Result<bool> AddressSpace::set_region(u64 address, usize count, bool used, int p
         {
             bool finished = region->end == end;
             auto* split = TRY(split_region(region, address));
-            update_region(split);
+            split->used = used;
+            split->persistent = persistent;
             try_merge_region_with_neighbors(split);
             if (!finished) continue;
             return true;
@@ -193,7 +164,8 @@ Result<bool> AddressSpace::set_region(u64 address, usize count, bool used, int p
         if (region->end > end)
         {
             TRY(split_region(region, end));
-            update_region(region);
+            region->used = used;
+            region->persistent = persistent;
             try_merge_region_with_neighbors(region);
             return true;
         }
@@ -202,23 +174,6 @@ Result<bool> AddressSpace::set_region(u64 address, usize count, bool used, int p
     return true;
 }
 
-Result<void> AddressSpace::sync_regions(u64 address, usize count)
-{
-    if (address >= VM_END) return err(EINVAL);
-
-    u64 end = address + (count * ARCH_PAGE_SIZE);
-
-    for (auto* region : m_regions)
-    {
-        if (region->end < address) continue;
-        if (region->start > end) return {};
-
-        region->sync_shared();
-    }
-
-    return {};
-}
-
 void AddressSpace::merge_contiguous_regions(VMRegion* a, VMRegion* b)
 {
     a->end = b->end;
@@ -229,24 +184,18 @@ void AddressSpace::merge_contiguous_regions(VMRegion* a, VMRegion* b)
 
 void AddressSpace::try_merge_region_with_neighbors(VMRegion* region)
 {
-    auto equals = [](VMRegion* a, VMRegion* b) {
-        if (a->used != b->used) return false;
-        if (a->persistent != b->persistent) return false;
-        if (a->prot != b->prot) return false;
-        if (a->flags != b->flags) return false;
-        if (a->shmid != b->shmid) return false;
-        return true;
-    };
-
     auto prev = m_regions.previous(region);
-    if (prev.has_value() && equals(*prev, region))
+    if (prev.has_value() && (*prev)->used == region->used && (*prev)->persistent == region->persistent)
     {
         merge_contiguous_regions(*prev, region);
         region = *prev;
     }
 
     auto next = m_regions.next(region);
-    if (next.has_value() && equals(*next, region)) { merge_contiguous_regions(region, *next); }
+    if (next.has_value() && (*next)->used == region->used && (*next)->persistent == region->persistent)
+    {
+        merge_contiguous_regions(region, *next);
+    }
 }
 
 Result<VMRegion*> AddressSpace::split_region(VMRegion* parent, u64 boundary)
@@ -258,9 +207,6 @@ Result<VMRegion*> AddressSpace::split_region(VMRegion* parent, u64 boundary)
     region->count = (region->end - region->start) / ARCH_PAGE_SIZE;
     region->used = parent->used;
     region->persistent = parent->persistent;
-    region->prot = parent->prot;
-    region->flags = parent->flags;
-    region->shmid = parent->shmid;
     m_regions.add_after(parent, region);
 
     parent->end = boundary;
@@ -271,43 +217,6 @@ Result<VMRegion*> AddressSpace::split_region(VMRegion* parent, u64 boundary)
 
 AddressSpace::~AddressSpace()
 {
-    auto* directory = MMU::get_page_directory();
-    MMU::switch_page_directory(this->m_directory);
-    m_regions.consume([this](VMRegion* region) {
-        region->cleanup_shared();
-        delete region;
-    });
-    MMU::switch_page_directory(directory);
-
+    m_regions.consume([](VMRegion* region) { delete region; });
     if (m_directory) MMU::delete_userspace_page_directory(m_directory);
 }
-
-void VMRegion::cleanup_shared()
-{
-    if (used && (flags & MAP_SHARED))
-    {
-        SharedMemory* shmem = g_shared_memory_map.try_get_ref(shmid);
-        if (shmem)
-        {
-            for (u64 addr = start; addr < end; addr += ARCH_PAGE_SIZE) { MMU::unmap(addr); }
-            if (--shmem->refs == 0)
-            {
-                shmem->free();
-                g_shared_memory_map.try_remove(shmid);
-            }
-        }
-    }
-}
-
-void VMRegion::sync_shared()
-{
-    if (used && (flags & MAP_SHARED) && (prot & PROT_WRITE))
-    {
-        SharedMemory* shmem = g_shared_memory_map.try_get_ref(shmid);
-        if (shmem)
-        {
-            if (shmem->inode) shmem->inode->write((const u8*)start, offset, count * ARCH_PAGE_SIZE);
-            if (shmem->device) shmem->device->write((const u8*)start, offset, count * ARCH_PAGE_SIZE);
-        }
-    }
-}

kernel/src/memory/AddressSpace.h

@@ -3,7 +3,6 @@
 #include <luna/LinkedList.h>
 #include <luna/OwnedPtr.h>
 #include <luna/Result.h>
-#include <sys/types.h>
 
 class VMRegion : LinkedListNode<VMRegion>
 {
@@ -13,13 +12,6 @@ class VMRegion : LinkedListNode<VMRegion>
     usize count;
     bool used { true };
     bool persistent { false };
-    int flags { 0 };
-    int prot { 0 };
-    u64 shmid;
-    off_t offset { 0 };
-
-    void cleanup_shared();
-    void sync_shared();
 };
 
 class AddressSpace
@@ -30,21 +22,18 @@ class AddressSpace
 
     AddressSpace& operator=(AddressSpace&& other);
 
-    Result<u64> alloc_region(usize count, int prot, int flags, off_t offset, u64 shmid = 0, bool persistent = false);
+    Result<u64> alloc_region(usize count, bool persistent = false);
 
-    Result<bool> test_and_alloc_region(u64 address, usize count, int prot, int flags, off_t offset, u64 shmid = 0,
-                                       bool persistent = false)
+    Result<bool> test_and_alloc_region(u64 address, usize count, bool persistent = false)
     {
-        return set_region(address, count, true, prot, flags, offset, shmid, persistent);
+        return set_region(address, count, true, persistent);
     }
 
     Result<bool> free_region(u64 address, usize count)
     {
-        return set_region(address, count, false, 0, 0, 0, 0, false);
+        return set_region(address, count, false, false);
     }
 
-    Result<void> sync_regions(u64 address, usize count);
-
     static Result<OwnedPtr<AddressSpace>> try_create();
 
     Result<OwnedPtr<AddressSpace>> clone();
@@ -55,8 +44,7 @@ class AddressSpace
     }
 
   private:
-    Result<bool> set_region(u64 address, usize count, bool used, int prot, int flags, off_t offset, u64 shmid,
-                            bool persistent);
+    Result<bool> set_region(u64 address, usize count, bool used, bool persistent);
     Result<void> create_default_region();
     Result<void> create_null_region();
     void try_merge_region_with_neighbors(VMRegion* region);

kernel/src/memory/MemoryManager.cpp

@@ -1,12 +1,9 @@
 #include "memory/MemoryManager.h"
-#include "Log.h"
 #include "arch/MMU.h"
-#include "fs/StorageCache.h"
 #include "memory/KernelVM.h"
 #include "memory/MemoryMap.h"
 #include <luna/Alignment.h>
 #include <luna/Bitmap.h>
-#include <luna/Common.h>
 #include <luna/ScopeGuard.h>
 #include <luna/Spinlock.h>
 #include <luna/SystemError.h>
@@ -42,11 +39,11 @@ namespace MemoryManager
     Result<void> protect_kernel_sections()
     {
         const usize rodata_size = (usize)(end_of_kernel_rodata - start_of_kernel_rodata);
-        const usize rodata_pages = ceil_div(rodata_size, ARCH_PAGE_SIZE);
+        const usize rodata_pages = get_blocks_from_size(rodata_size, ARCH_PAGE_SIZE);
         TRY(remap((u64)start_of_kernel_rodata, rodata_pages, MMU::NoExecute));
 
         const usize data_size = (usize)(end_of_kernel_data - start_of_kernel_data);
-        const usize data_pages = ceil_div(data_size, ARCH_PAGE_SIZE);
+        const usize data_pages = get_blocks_from_size(data_size, ARCH_PAGE_SIZE);
         TRY(remap((u64)start_of_kernel_data, data_pages, MMU::NoExecute | MMU::ReadWrite));
 
         return {};
@@ -69,7 +66,7 @@ namespace MemoryManager
         // We store our frame bitmap at the beginning of the largest free memory block.
         char* const frame_bitmap_addr = (char*)largest_free_entry.ptr();
 
-        const usize frame_bitmap_size = ceil_div(physical_address_space_size / ARCH_PAGE_SIZE, 8UL);
+        const usize frame_bitmap_size = get_blocks_from_size(physical_address_space_size / ARCH_PAGE_SIZE, 8UL);
 
         // This should never happen, unless memory is very fragmented. Usually there is always a very big block of
         // usable memory and then some tiny blocks around it.
@@ -98,7 +95,7 @@ namespace MemoryManager
         }
 
         // Make sure that the physical frames used by the bitmap aren't handed out to anyone else.
-        lock_frames(largest_free_entry.address(), ceil_div(frame_bitmap_size, ARCH_PAGE_SIZE));
+        lock_frames(largest_free_entry.address(), get_blocks_from_size(frame_bitmap_size, ARCH_PAGE_SIZE));
     }
 
     void init()
@@ -144,12 +141,7 @@ namespace MemoryManager
 
         usize index;
         bool ok = frame_bitmap->find_and_toggle(false, start_index).try_set_value(index);
-        if (!ok)
-        {
-            kwarnln("OOM alert! Scheduling the OOM thread...");
-            Scheduler::signal_oom_thread();
-            return err(ENOMEM);
-        }
+        if (!ok) return err(ENOMEM);
 
         start_index = index + 1;
 
@@ -234,57 +226,6 @@ namespace MemoryManager
         return {};
     }
 
-    Result<void> copy_region(u64 virt, usize count, PageDirectory* oldpd, PageDirectory* newpd)
-    {
-        CHECK_PAGE_ALIGNED(virt);
-
-        usize pages_mapped = 0;
-
-        // Let's clean up after ourselves if we fail.
-        auto guard = make_scope_guard(
-            [=, &pages_mapped] { kwarnln("copy_region failed, sorry! cannot reclaim already copied pages"); });
-
-        while (pages_mapped < count)
-        {
-            u64 phys = TRY(MMU::get_physical(virt, oldpd));
-            int flags = TRY(MMU::get_flags(virt, oldpd));
-            TRY(MMU::map(virt, phys, flags, MMU::UseHugePages::No, newpd));
-            virt += ARCH_PAGE_SIZE;
-            pages_mapped++;
-        }
-
-        guard.deactivate();
-
-        return {};
-    }
-
-    Result<void> copy_region_data(u64 virt, usize count, PageDirectory* oldpd, PageDirectory* newpd)
-    {
-        CHECK_PAGE_ALIGNED(virt);
-
-        usize pages_mapped = 0;
-
-        // Let's clean up after ourselves if we fail.
-        auto guard = make_scope_guard(
-            [=, &pages_mapped] { kwarnln("copy_region_data failed, sorry! cannot reclaim already copied pages"); });
-
-        while (pages_mapped < count)
-        {
-            u64 frame = TRY(alloc_frame());
-            u64 phys = TRY(MMU::get_physical(virt, oldpd));
-            int flags = TRY(MMU::get_flags(virt, oldpd));
-            memcpy((void*)MMU::translate_physical_address(frame), (void*)MMU::translate_physical_address(phys),
-                   ARCH_PAGE_SIZE);
-            TRY(MMU::map(virt, frame, flags, MMU::UseHugePages::No, newpd));
-            virt += ARCH_PAGE_SIZE;
-            pages_mapped++;
-        }
-
-        guard.deactivate();
-
-        return {};
-    }
-
     Result<void> map_huge_frames_at(u64 virt, u64 phys, usize count, int flags)
     {
         CHECK_PAGE_ALIGNED(virt);
@@ -417,20 +358,6 @@ namespace MemoryManager
         return {};
     }
 
-    Result<void> unmap_owned_if_possible(u64 virt, usize count)
-    {
-        CHECK_PAGE_ALIGNED(virt);
-
-        while (count--)
-        {
-            const auto frame = MMU::unmap(virt);
-            if (frame.has_value()) TRY(free_frame(frame.value()));
-            virt += ARCH_PAGE_SIZE;
-        }
-
-        return {};
-    }
-
     Result<void> unmap_owned_and_free_vm(u64 virt, usize count)
     {
         CHECK_PAGE_ALIGNED(virt);
@@ -515,7 +442,7 @@ namespace MemoryManager
         {
             TRY(result.try_append(*(char*)address));
             address++;
-            if ((address % ARCH_PAGE_SIZE) == 0)
+            if (address % ARCH_PAGE_SIZE)
             {
                 if (!validate_page_access(address, MMU::User)) return err(EFAULT);
             }
@@ -533,7 +460,7 @@ namespace MemoryManager
 
         uintptr_t diff = address - page;
 
-        usize pages = ceil_div(size + diff, ARCH_PAGE_SIZE);
+        usize pages = get_blocks_from_size(size + diff, ARCH_PAGE_SIZE);
 
         while (pages--)
         {
@@ -569,7 +496,7 @@ namespace MemoryManager
         while (size--)
         {
             // Crossed a page boundary, gotta check the page tables again before touching any memory!!
-            if ((user_ptr % ARCH_PAGE_SIZE) == 0)
+            if (user_ptr % ARCH_PAGE_SIZE)
             {
                 if (!validate_page_access(user_ptr, MMU::ReadWrite | MMU::User)) return false;
             }
@@ -597,7 +524,7 @@ namespace MemoryManager
         while (size--)
         {
             // Crossed a page boundary, gotta check the page tables again before touching any memory!!
-            if ((user_ptr % ARCH_PAGE_SIZE) == 0)
+            if (user_ptr % ARCH_PAGE_SIZE)
             {
                 if (!validate_page_access(user_ptr, MMU::User)) return false;
             }

kernel/src/memory/MemoryManager.h

@@ -67,9 +67,6 @@ namespace MemoryManager
     Result<void> map_frames_at(u64 virt, u64 phys, usize count, int flags);
     Result<void> map_huge_frames_at(u64 virt, u64 phys, usize count, int flags);
 
-    Result<void> copy_region(u64 virt, usize count, PageDirectory* oldpd, PageDirectory* newpd);
-    Result<void> copy_region_data(u64 virt, usize count, PageDirectory* oldpd, PageDirectory* newpd);
-
     Result<u64> alloc_at(u64 virt, usize count, int flags);
     Result<u64> alloc_at_zeroed(u64, usize count, int flags);
     Result<u64> alloc_for_kernel(usize count, int flags);
@@ -78,7 +75,6 @@ namespace MemoryManager
 
     Result<void> unmap_owned(u64 virt, usize count);
     Result<void> unmap_owned_and_free_vm(u64 virt, usize count);
-    Result<void> unmap_owned_if_possible(u64 virt, usize count);
     Result<void> unmap_weak(u64 virt, usize count);
     Result<void> unmap_weak_and_free_vm(u64 virt, usize count);
 

kernel/src/memory/SharedMemory.cpp

@@ -1,124 +0,0 @@
-#include "memory/SharedMemory.h"
-#include "memory/MemoryManager.h"
-#include <bits/mmap-flags.h>
-#include <luna/ScopeGuard.h>
-
-HashMap<u64, SharedMemory> g_shared_memory_map;
-Atomic<u64> g_next_shmem_id;
-
-Result<u64> SharedMemory::create(u8* mem, off_t offset, usize count)
-{
-    SharedMemory shmem;
-    shmem.offset = offset;
-
-    auto guard = make_scope_guard([&shmem] {
-        for (u64 frame : shmem.frames) { MemoryManager::free_frame(frame); }
-    });
-
-    while (count--)
-    {
-        u64 frame = TRY(MemoryManager::alloc_frame());
-        TRY(shmem.frames.try_append(frame));
-        if (mem)
-        {
-            memcpy((void*)MMU::translate_physical_address(frame), mem, ARCH_PAGE_SIZE);
-            mem += ARCH_PAGE_SIZE;
-        }
-        else { memset((void*)MMU::translate_physical_address(frame), 0, ARCH_PAGE_SIZE); }
-    }
-
-    const u64 id = g_next_shmem_id++;
-
-    check(TRY(g_shared_memory_map.try_set(id, move(shmem))));
-
-    guard.deactivate();
-
-    return id;
-}
-
-Result<void> SharedMemory::grow_forward(u8* mem, usize count)
-{
-    u64 old_count = frames.size();
-
-    auto guard = make_scope_guard([old_count, this] {
-        while (old_count < this->frames.size()) { MemoryManager::free_frame(*this->frames.try_pop()); }
-    });
-
-    while (count--)
-    {
-        u64 frame = TRY(MemoryManager::alloc_frame());
-        TRY(frames.try_append(frame));
-        if (mem)
-        {
-            memcpy((void*)MMU::translate_physical_address(frame), mem, ARCH_PAGE_SIZE);
-            mem += ARCH_PAGE_SIZE;
-        }
-    }
-
-    guard.deactivate();
-
-    return {};
-}
-
-Result<void> SharedMemory::grow_backward(u8* mem, usize count)
-{
-    Vector<u64> new_frames;
-
-    const usize bytes = count * ARCH_PAGE_SIZE;
-
-    auto guard = make_scope_guard([&new_frames, count] {
-        for (u64 i = 0; i < count && i < new_frames.size(); i++) { MemoryManager::free_frame(new_frames[i]); }
-    });
-
-    while (count--)
-    {
-        u64 frame = TRY(MemoryManager::alloc_frame());
-        TRY(new_frames.try_append(frame));
-        if (mem)
-        {
-            memcpy((void*)MMU::translate_physical_address(frame), mem, ARCH_PAGE_SIZE);
-            mem += ARCH_PAGE_SIZE;
-        }
-    }
-
-    for (u64 frame : frames) { TRY(new_frames.try_append(frame)); }
-
-    frames = move(new_frames);
-    offset -= bytes;
-
-    guard.deactivate();
-
-    return {};
-}
-
-Result<void> SharedMemory::map(u64 virt, int flags, off_t _offset, usize count)
-{
-    usize off = _offset / ARCH_PAGE_SIZE;
-    if (off + count > frames.size()) return err(EINVAL);
-
-    for (usize i = off; i < count; i++)
-    {
-        TRY(MMU::map(virt, frames[i], flags, MMU::UseHugePages::No));
-        virt += ARCH_PAGE_SIZE;
-    }
-
-    return {};
-}
-
-void SharedMemory::free()
-{
-    if ((inode || device) && (prot & PROT_WRITE))
-    {
-        for (u64 i = 0; i < frames.size(); i++)
-        {
-            if (inode)
-                inode->write((u8*)MMU::translate_physical_address(frames[i]), offset + (i * ARCH_PAGE_SIZE),
-                             ARCH_PAGE_SIZE);
-            if (device)
-                device->write((u8*)MMU::translate_physical_address(frames[i]), offset + (i * ARCH_PAGE_SIZE),
-                              ARCH_PAGE_SIZE);
-        }
-    }
-
-    for (u64 frame : frames) { MemoryManager::free_frame(frame); }
-}

kernel/src/memory/SharedMemory.h

@@ -1,32 +0,0 @@
-#pragma once
-#include "fs/VFS.h"
-#include "fs/devices/Device.h"
-#include <luna/HashMap.h>
-#include <luna/Vector.h>
-#include <sys/types.h>
-
-struct SharedMemory
-{
-    Vector<u64> frames;
-    off_t offset;
-    int prot;
-    SharedPtr<VFS::Inode> inode {};
-    SharedPtr<Device> device {};
-    int refs { 0 };
-
-    SharedMemory() = default;
-    SharedMemory(SharedMemory&&) = default;
-
-    static Result<u64> create(u8* mem, off_t offset, usize count);
-
-    Result<void> grow_forward(u8* mem, usize count);
-    Result<void> grow_backward(u8* mem, usize count);
-
-    Result<void> map(u64 virt, int flags, off_t offset, usize count);
-
-    void free();
-
-    ~SharedMemory() = default;
-};
-
-extern HashMap<u64, SharedMemory> g_shared_memory_map;

kernel/src/net/Socket.h

@@ -1,72 +0,0 @@
-#pragma once
-#include "arch/CPU.h"
-#include "fs/VFS.h"
-#include "thread/Thread.h"
-#include <bits/socket.h>
-
-class Socket : public VFS::FileInode
-{
-  public:
-    Socket() = default;
-
-    VFS::InodeType type() const override
-    {
-        return VFS::InodeType::Socket;
-    }
-
-    void set_fs(VFS::FileSystem* fs)
-    {
-        m_fs = fs;
-    }
-
-    void set_inode_number(usize inum)
-    {
-        m_metadata.inum = inum;
-    }
-
-    VFS::FileSystem* fs() const override
-    {
-        return m_fs;
-    }
-
-    Result<usize> read(u8* buf, usize, usize length) const override
-    {
-        return recv(buf, length, 0);
-    }
-
-    Result<usize> write(const u8* buf, usize, usize length) override
-    {
-        return send(buf, length, 0);
-    }
-
-    virtual Result<usize> send(const u8*, usize, int) = 0;
-
-    virtual Result<usize> recv(u8*, usize, int) const = 0;
-
-    virtual Result<void> bind(struct sockaddr*, socklen_t) = 0;
-    virtual Result<void> connect(Registers*, int, struct sockaddr*, socklen_t) = 0;
-
-    virtual Result<SharedPtr<OpenFileDescription>> accept(Registers*, int, struct sockaddr**, socklen_t*) = 0;
-
-    virtual Result<void> listen(int backlog) = 0;
-
-    Result<void> truncate(usize) override
-    {
-        return err(EINVAL);
-    }
-
-    void did_link() override
-    {
-        m_metadata.nlinks++;
-    }
-
-    void did_unlink() override
-    {
-        m_metadata.nlinks--;
-    }
-
-    virtual ~Socket() = default;
-
-  protected:
-    VFS::FileSystem* m_fs { nullptr };
-};

kernel/src/net/UnixSocket.cpp

@@ -1,211 +0,0 @@
-#include "net/UnixSocket.h"
-#include <bits/open-flags.h>
-#include <luna/PathParser.h>
-#include <thread/Scheduler.h>
-
-UnixSocket::UnixSocket()
-{
-}
-
-UnixSocket::UnixSocket(UnixSocket* peer) : m_state(State::Connected), m_peer(peer)
-{
-}
-
-UnixSocket::~UnixSocket()
-{
-    did_close();
-}
-
-void UnixSocket::did_close()
-{
-    if (m_peer)
-    {
-        m_peer->m_peer = nullptr;
-        m_peer->m_state = State::Reset;
-    }
-    m_state = State::Inactive;
-}
-
-void UnixSocket::connect_to_peer(UnixSocket* peer)
-{
-    m_peer = peer;
-    m_state = State::Connected;
-}
-
-Result<usize> UnixSocket::send(const u8* buf, usize length, int)
-{
-    if (m_state == State::Reset) return err(ECONNRESET);
-    if (m_state != State::Connected) return err(ENOTCONN);
-
-    check(m_peer);
-
-    TRY(m_peer->m_data.append_data(buf, length));
-
-    return length;
-}
-
-Result<usize> UnixSocket::recv(u8* buf, usize length, int) const
-{
-    if (m_state == State::Reset && !m_data.size()) return err(ECONNRESET);
-    if (m_state != State::Connected && m_state != State::Reset) return err(ENOTCONN);
-
-    return m_data.dequeue_data(buf, length);
-}
-
-static Result<void> bind_socket_to_fs(const char* path, Credentials auth, SharedPtr<VFS::Inode> working_directory,
-                                      SharedPtr<UnixSocket> socket)
-{
-    auto parent_path = TRY(PathParser::dirname(path));
-
-    auto parent_inode = TRY(VFS::resolve_path(parent_path.chars(), auth, working_directory));
-
-    if (!VFS::can_write(parent_inode, auth)) return err(EACCES);
-
-    auto child_name = TRY(PathParser::basename(path));
-
-    TRY(VFS::validate_filename(child_name.view()));
-
-    socket->set_inode_number(TRY(parent_inode->fs()->allocate_inode_number()));
-    socket->set_fs(parent_inode->fs());
-
-    return parent_inode->add_entry(socket, child_name.chars());
-}
-
-Result<void> UnixSocket::bind(struct sockaddr* addr, socklen_t addrlen)
-{
-    if (!addr) return err(EDESTADDRREQ);
-    if (addr->sa_family != AF_UNIX) return err(EAFNOSUPPORT);
-    if ((usize)addrlen > sizeof(sockaddr_un)) return err(EINVAL);
-
-    if (m_state == State::Connected) return err(EISCONN);
-    if (m_state != State::Inactive) return err(EINVAL);
-
-    struct sockaddr_un* un_address = (struct sockaddr_un*)addr;
-
-    String path = TRY(String::from_string_view(
-        StringView::from_fixed_size_cstring(un_address->sun_path, addrlen - sizeof(sa_family_t))));
-
-    auto* current = Scheduler::current();
-
-    m_metadata.mode = 0777 & ~current->umask;
-    m_metadata.uid = current->auth.euid;
-    m_metadata.gid = current->auth.egid;
-
-    auto rc = bind_socket_to_fs(path.chars(), current->auth, current->current_directory, SharedPtr<Socket> { this });
-    if (rc.has_error())
-    {
-        if (rc.error() == EEXIST) return err(EADDRINUSE);
-        return rc.release_error();
-    }
-
-    memcpy(&m_addr, un_address, addrlen);
-    m_addrlen = addrlen;
-
-    m_state = State::Bound;
-
-    return {};
-}
-
-Result<void> UnixSocket::connect(Registers* regs, int flags, struct sockaddr* addr, socklen_t addrlen)
-{
-    if (!addr) return err(EINVAL);
-    if (addr->sa_family != AF_UNIX) return err(EAFNOSUPPORT);
-    if ((usize)addrlen > sizeof(sockaddr_un)) return err(EINVAL);
-
-    if (m_state == State::Connected) return err(EISCONN);
-    if (m_state == State::Connecting) return err(EALREADY);
-    if (m_state != State::Inactive) return err(EINVAL);
-
-    struct sockaddr_un* un_address = (struct sockaddr_un*)addr;
-
-    String path = TRY(String::from_string_view(
-        StringView::from_fixed_size_cstring(un_address->sun_path, addrlen - sizeof(sa_family_t))));
-
-    auto* current = Scheduler::current();
-
-    auto inode = TRY(VFS::resolve_path(path.chars(), current->auth, current->current_directory));
-    if (inode->type() != VFS::InodeType::Socket)
-        return err(ENOTSOCK); // FIXME: POSIX doesn't say what error to return here?
-    if (!VFS::can_write(inode, current->auth)) return err(EACCES);
-
-    auto socket = (SharedPtr<UnixSocket>)inode;
-    if (socket->m_state != State::Listening) return err(ECONNREFUSED);
-    if (!socket->m_listen_queue.try_push(this)) return err(ECONNREFUSED);
-    if (socket->m_blocked_thread) socket->m_blocked_thread->wake_up();
-
-    m_state = Connecting;
-    if (flags & O_NONBLOCK) return err(EINPROGRESS);
-
-    while (1)
-    {
-        m_blocked_thread = current;
-        kernel_wait_for_event();
-        m_blocked_thread = nullptr;
-        if (current->interrupted)
-        {
-            if (current->will_ignore_pending_signal())
-            {
-                current->process_pending_signals(regs);
-                continue;
-            }
-            return err(EINTR);
-        }
-        break;
-    }
-
-    check(m_state == Connected);
-    check(m_peer);
-
-    return {};
-}
-
-Result<void> UnixSocket::listen(int backlog)
-{
-    if (backlog < 0) backlog = 0;
-    if (m_state == State::Listening || m_state == State::Connected) return err(EINVAL);
-    if (m_state != State::Bound) return err(EDESTADDRREQ);
-    TRY(m_listen_queue.set_size(backlog));
-    m_state = State::Listening;
-    return {};
-}
-
-Result<SharedPtr<OpenFileDescription>> UnixSocket::accept(Registers* regs, int flags, struct sockaddr** addr,
-                                                          socklen_t* addrlen)
-{
-    if (m_state != State::Listening) return err(EINVAL);
-
-    auto* current = Scheduler::current();
-
-    UnixSocket* peer = nullptr;
-    while (!m_listen_queue.try_pop(peer))
-    {
-        if (flags & O_NONBLOCK) return err(EAGAIN);
-        m_blocked_thread = current;
-        kernel_wait_for_event();
-        m_blocked_thread = nullptr;
-        if (current->interrupted)
-        {
-            if (current->will_ignore_pending_signal())
-            {
-                current->process_pending_signals(regs);
-                continue;
-            }
-            return err(EINTR);
-        }
-    }
-
-    check(peer);
-
-    auto socket = TRY(make_shared<UnixSocket>(peer));
-    auto description = TRY(make_shared<OpenFileDescription>(socket, O_RDWR));
-
-    peer->m_peer = socket.ptr();
-    peer->m_state = State::Connected;
-
-    if (peer->m_blocked_thread) peer->m_blocked_thread->wake_up();
-
-    *addr = (struct sockaddr*)&peer->m_addr;
-    *addrlen = peer->m_addrlen;
-
-    return description;
-}

kernel/src/net/UnixSocket.h

@@ -1,57 +0,0 @@
-#pragma once
-#include "net/Socket.h"
-#include "thread/Thread.h"
-#include <luna/Buffer.h>
-#include <luna/CircularQueue.h>
-#include <luna/String.h>
-#include <sys/un.h>
-
-class UnixSocket : public Socket
-{
-  public:
-    UnixSocket();
-    UnixSocket(UnixSocket* peer);
-
-    bool will_block_if_read() const override
-    {
-        return (m_state == Connected || m_state == Reset) && !m_data.size();
-    }
-
-    Result<usize> send(const u8*, usize, int) override;
-    Result<usize> recv(u8*, usize, int) const override;
-
-    Result<void> bind(struct sockaddr*, socklen_t) override;
-    Result<void> connect(Registers*, int, struct sockaddr*, socklen_t) override;
-    Result<SharedPtr<OpenFileDescription>> accept(Registers*, int, struct sockaddr**, socklen_t*) override;
-
-    Result<void> listen(int backlog) override;
-
-    void did_close() override;
-
-    void connect_to_peer(UnixSocket* peer);
-
-    virtual ~UnixSocket();
-
-  private:
-    enum State
-    {
-        Inactive,
-        Bound,
-        Listening,
-        Connecting,
-        Connected,
-        Reset,
-    };
-
-    State m_state = State::Inactive;
-    UnixSocket* m_peer = nullptr;
-
-    mutable Buffer m_data;
-
-    Thread* m_blocked_thread { nullptr };
-
-    DynamicCircularQueue<UnixSocket*> m_listen_queue;
-
-    struct sockaddr_un m_addr = { .sun_family = AF_UNIX, .sun_path = {} };
-    socklen_t m_addrlen = sizeof(sa_family_t);
-};

kernel/src/sys/Syscall.h

@@ -1,7 +1,7 @@
 #pragma once
-#include "api/Syscall.h"
 #include "arch/CPU.h"
 #include <luna/Result.h>
+#include <luna/Syscall.h>
 
 typedef u64 SyscallArgs[6];
 

kernel/src/sys/alarm.cpp

@@ -1,18 +0,0 @@
-#include "Pledge.h"
-#include "sys/Syscall.h"
-#include "thread/Scheduler.h"
-
-Result<u64> sys_alarm(Registers*, SyscallArgs args)
-{
-    unsigned int seconds = (unsigned int)args[0];
-
-    auto* current = Scheduler::current();
-
-    TRY(check_pledge(current, Promise::p_stdio));
-
-    u64 time_left = current->alarm_ticks_left;
-
-    current->alarm_ticks_left = seconds * 1000;
-
-    return time_left * 1000;
-}

kernel/src/sys/chdir.cpp

@@ -1,4 +1,3 @@
-#include "Pledge.h"
 #include "memory/MemoryManager.h"
 #include "sys/Syscall.h"
 #include "thread/Scheduler.h"
@@ -10,8 +9,6 @@ Result<u64> sys_chdir(Registers*, SyscallArgs args)
 
     Thread* current = Scheduler::current();
 
-    TRY(check_pledge(current, Promise::p_rpath));
-
     if (PathParser::is_absolute(path.view()))
     {
         SharedPtr<VFS::Inode> inode = TRY(VFS::resolve_path(path.chars(), current->auth));

kernel/src/sys/clock_gettime.cpp

@@ -1,8 +1,6 @@
-#include "Pledge.h"
 #include "arch/Timer.h"
 #include "memory/MemoryManager.h"
 #include "sys/Syscall.h"
-#include "thread/Scheduler.h"
 #include <bits/clockid.h>
 #include <bits/timespec.h>
 
@@ -11,10 +9,6 @@ Result<u64> sys_clock_gettime(Registers*, SyscallArgs args)
     clockid_t id = (clockid_t)args[0];
     struct timespec* ts = (struct timespec*)args[1];
 
-    auto* current = Scheduler::current();
-
-    TRY(check_pledge(current, Promise::p_stdio));
-
     switch (id)
     {
     case CLOCK_MONOTONIC: {

kernel/src/sys/exec.cpp

@@ -1,9 +1,8 @@
 #include "Log.h"
-#include "Pledge.h"
-#include "binfmt/BinaryFormat.h"
 #include "fs/VFS.h"
 #include "memory/MemoryManager.h"
 #include "sys/Syscall.h"
+#include "thread/ELF.h"
 #include "thread/Scheduler.h"
 #include "thread/ThreadImage.h"
 #include <bits/modes.h>
@@ -56,18 +55,13 @@ Result<u64> sys_execve(Registers* regs, SyscallArgs args)
 {
     auto path = TRY(MemoryManager::strdup_from_user(args[0]));
     auto argv = TRY(copy_string_vector_from_userspace(args[1]));
-    Vector<String> envp;
-    if (args[2]) envp = TRY(copy_string_vector_from_userspace(args[2]));
-
-    String cmdline = TRY(String::join(argv, " "));
+    auto envp = TRY(copy_string_vector_from_userspace(args[2]));
 
     if ((calculate_userspace_stack_size(argv) + calculate_userspace_stack_size(envp)) > MAX_ARGV_STACK_SIZE)
         return err(E2BIG);
 
     auto current = Scheduler::current();
 
-    TRY(check_pledge(current, Promise::p_exec));
-
     auto inode = TRY(VFS::resolve_path(path.chars(), current->auth, current->current_directory));
 
     if (!VFS::can_execute(inode, current->auth)) return err(EACCES);
@@ -76,22 +70,9 @@ Result<u64> sys_execve(Registers* regs, SyscallArgs args)
     kdbgln("exec: attempting to replace current image with %s", path.chars());
 #endif
 
-    bool is_setuid = VFS::is_setuid(inode);
-    bool is_setgid = VFS::is_setgid(inode);
-    bool is_secure_environment = is_setgid || is_setuid;
-    if (is_secure_environment && current->execpromises >= 0) return err(EACCES);
-
-    auto loader = TRY(BinaryFormat::create_loader(inode));
-
-#ifdef EXEC_DEBUG
-    kdbgln("exec: created loader for binary format %s", loader->format().chars());
-#endif
-
     auto guard = make_scope_guard([current] { MMU::switch_page_directory(current->self_directory()); });
 
-    auto image = TRY(ThreadImage::try_load_from_binary(loader));
-
-    argv = TRY(loader->cmdline(path, move(argv)));
+    auto image = TRY(ThreadImage::try_load_from_elf(inode));
 
     u64 user_argv = TRY(image->push_string_vector_on_stack(argv));
     usize user_argc = argv.size();
@@ -111,13 +92,17 @@ Result<u64> sys_execve(Registers* regs, SyscallArgs args)
     {
         auto& descriptor = current->fd_table[i];
         if (!descriptor.has_value()) continue;
-        if (descriptor->flags & O_CLOEXEC) { descriptor = {}; }
+        if (descriptor->flags & O_CLOEXEC)
+        {
+            descriptor->inode->remove_handle();
+            descriptor = {};
+        }
     }
 
-    if (is_setuid) current->auth.euid = current->auth.suid = inode->metadata().uid;
-    if (is_setgid) current->auth.egid = current->auth.sgid = inode->metadata().gid;
+    if (VFS::is_setuid(inode)) current->auth.euid = current->auth.suid = inode->uid();
+    if (VFS::is_setgid(inode)) current->auth.egid = current->auth.sgid = inode->gid();
 
-    current->cmdline = cmdline.chars();
+    current->name = path.chars();
 
     image->apply(current);
 
@@ -125,9 +110,6 @@ Result<u64> sys_execve(Registers* regs, SyscallArgs args)
 
     current->set_arguments(user_argc, user_argv, user_envc, user_envp);
 
-    current->promises = current->execpromises;
-    current->execpromises = -1;
-
     memcpy(regs, &current->regs, sizeof(*regs));
 
     for (int i = 0; i < NSIG; i++)
@@ -137,7 +119,7 @@ Result<u64> sys_execve(Registers* regs, SyscallArgs args)
 
     current->has_called_exec = true;
 
-    kinfoln("exec: thread %d was replaced with %s", current->id, path.chars());
+    kinfoln("exec: thread %lu was replaced with %s", current->id, path.chars());
 
     return 0;
 }
@@ -146,8 +128,6 @@ Result<u64> sys_fork(Registers* regs, SyscallArgs)
 {
     auto current = Scheduler::current();
 
-    TRY(check_pledge(current, Promise::p_proc));
-
     auto guard = make_scope_guard([current] { MMU::switch_page_directory(current->self_directory()); });
 
     memcpy(&current->regs, regs, sizeof(*regs));
@@ -161,22 +141,28 @@ Result<u64> sys_fork(Registers* regs, SyscallArgs)
     thread->state = ThreadState::Runnable;
     thread->is_kernel = false;
     thread->fp_data.save();
-    thread->cmdline = current->cmdline;
+    thread->name = current->name;
     thread->auth = current->auth;
     thread->current_directory = current->current_directory;
     thread->current_directory_path = move(current_directory_path);
     thread->umask = current->umask;
     thread->parent = current;
-    thread->promises = current->promises;
-    thread->execpromises = current->execpromises;
 
-    for (int i = 0; i < FD_MAX; i++) { thread->fd_table[i] = current->fd_table[i]; }
+    for (int i = 0; i < FD_MAX; i++)
+    {
+        thread->fd_table[i] = current->fd_table[i];
+        if (current->fd_table[i].has_value()) current->fd_table[i]->inode->add_handle();
+    }
 
     image->apply(thread);
 
     memcpy(&thread->regs, regs, sizeof(*regs));
 
-    for (int i = 0; i < NSIG; i++) thread->signal_handlers[i] = current->signal_handlers[i];
+    for (int i = 0; i < NSIG; i++)
+    {
+        auto sighandler = current->signal_handlers[i].sa_handler;
+        thread->signal_handlers[i] = { .sa_handler = sighandler, .sa_mask = 0, .sa_flags = 0 };
+    }
     thread->signal_mask = current->signal_mask;
 
     thread->set_return(0);
@@ -184,7 +170,7 @@ Result<u64> sys_fork(Registers* regs, SyscallArgs)
     Scheduler::add_thread(thread);
 
 #ifdef FORK_DEBUG
-    kdbgln("fork: thread %d forked into child %d", current->id, thread->id);
+    kdbgln("fork: thread %lu forked into child %lu", current->id, thread->id);
 #endif
 
     return thread->id;

kernel/src/sys/file.cpp

@@ -1,15 +1,12 @@
 #include "Log.h"
-#include "Pledge.h"
 #include "fs/Pipe.h"
 #include "fs/VFS.h"
 #include "memory/MemoryManager.h"
 #include "sys/Syscall.h"
 #include "thread/Scheduler.h"
-#include <bits/atfile.h>
 #include <bits/fcntl.h>
 #include <bits/open-flags.h>
 #include <bits/seek.h>
-#include <bits/utime.h>
 #include <luna/SafeArithmetic.h>
 #include <sys/types.h>
 
@@ -25,13 +22,11 @@ Result<u64> sys_read(Registers* regs, SyscallArgs args)
 
     Thread* current = Scheduler::current();
 
-    TRY(check_pledge(current, Promise::p_stdio));
-
     auto& descriptor = *TRY(current->resolve_fd(fd));
 
     if (!descriptor.is_readable()) return err(EBADF);
 
-    while (descriptor.inode()->will_block_if_read())
+    while (descriptor.inode->blocking())
     {
         if (descriptor.should_block()) kernel_sleep(10);
         else
@@ -40,18 +35,14 @@ Result<u64> sys_read(Registers* regs, SyscallArgs args)
         if (current->interrupted)
         {
             kdbgln("signal: read interrupted by signal");
-            if (current->will_ignore_pending_signal())
-            {
-                current->process_pending_signals(regs);
-                continue;
-            }
-            return err(EINTR);
+            if (current->will_invoke_signal_handler()) return err(EINTR);
+            current->process_pending_signals(regs);
         }
     }
 
-    usize nread = TRY(descriptor.inode()->read(buf, descriptor.offset, size));
+    usize nread = TRY(descriptor.inode->read(buf, descriptor.offset, size));
 
-    if (VFS::is_seekable(descriptor.inode())) descriptor.offset += nread;
+    if (VFS::is_seekable(descriptor.inode)) descriptor.offset += nread;
 
     return nread;
 }
@@ -68,18 +59,15 @@ Result<u64> sys_write(Registers*, SyscallArgs args)
 
     Thread* current = Scheduler::current();
 
-    TRY(check_pledge(current, Promise::p_stdio));
-
     auto& descriptor = *TRY(current->resolve_fd(fd));
 
     if (!descriptor.is_writable()) return err(EBADF);
 
-    if (descriptor.should_append() && VFS::is_seekable(descriptor.inode()))
-        descriptor.offset = descriptor.inode()->metadata().size;
+    if (descriptor.should_append() && VFS::is_seekable(descriptor.inode)) descriptor.offset = descriptor.inode->size();
 
-    usize nwritten = TRY(descriptor.inode()->write(buf, descriptor.offset, size));
+    usize nwritten = TRY(descriptor.inode->write(buf, descriptor.offset, size));
 
-    if (VFS::is_seekable(descriptor.inode())) descriptor.offset += nwritten;
+    if (VFS::is_seekable(descriptor.inode)) descriptor.offset += nwritten;
 
     return nwritten;
 }
@@ -92,13 +80,11 @@ Result<u64> sys_lseek(Registers*, SyscallArgs args)
 
     Thread* current = Scheduler::current();
 
-    TRY(check_pledge(current, Promise::p_stdio));
-
     auto& descriptor = *TRY(current->resolve_fd(fd));
 
-    if (descriptor.inode()->type() == VFS::InodeType::FIFO) return err(ESPIPE);
+    if (descriptor.inode->type() == VFS::InodeType::FIFO) return err(ESPIPE);
 
-    if (!VFS::is_seekable(descriptor.inode())) return descriptor.offset;
+    if (!VFS::is_seekable(descriptor.inode)) return descriptor.offset;
 
     off_t new_offset;
 
@@ -106,7 +92,7 @@ Result<u64> sys_lseek(Registers*, SyscallArgs args)
     {
     case SEEK_SET: new_offset = offset; break;
     case SEEK_CUR: new_offset = TRY(safe_add((long)descriptor.offset, offset)); break;
-    case SEEK_END: new_offset = TRY(safe_add((long)descriptor.inode()->metadata().size, offset)); break;
+    case SEEK_END: new_offset = TRY(safe_add((long)descriptor.inode->size(), offset)); break;
     default: return err(EINVAL);
     }
 
@@ -124,8 +110,6 @@ Result<u64> sys_fcntl(Registers*, SyscallArgs args)
 
     Thread* current = Scheduler::current();
 
-    TRY(check_pledge(current, Promise::p_stdio));
-
     auto& descriptor = *TRY(current->resolve_fd(fd));
 
     bool is_cloexec = true;
@@ -153,14 +137,14 @@ Result<u64> sys_fcntl(Registers*, SyscallArgs args)
             descriptor.flags &= ~O_CLOEXEC;
         return 0;
     }
-    case F_GETFL: return (u64)descriptor.status_flags();
+    case F_GETFL: return (u64)(descriptor.flags & ~O_CLOEXEC);
     case F_SETFL: {
         int arg = (int)args[2];
 
-        descriptor.status_flags() &= ~(O_APPEND | O_NONBLOCK);
+        descriptor.flags &= ~(O_APPEND | O_NONBLOCK);
         arg &= (O_APPEND | O_NONBLOCK);
 
-        descriptor.status_flags() |= arg;
+        descriptor.flags |= arg;
 
         return 0;
     }
@@ -177,7 +161,7 @@ Result<u64> sys_ioctl(Registers*, SyscallArgs args)
     Thread* current = Scheduler::current();
     auto& descriptor = *TRY(current->resolve_fd(fd));
 
-    return descriptor.inode()->ioctl(request, arg);
+    return descriptor.inode->ioctl(request, arg);
 }
 
 Result<u64> sys_isatty(Registers*, SyscallArgs args)
@@ -185,10 +169,9 @@ Result<u64> sys_isatty(Registers*, SyscallArgs args)
     int fd = (int)args[0];
 
     Thread* current = Scheduler::current();
-    TRY(check_pledge(current, Promise::p_stdio));
     auto& descriptor = *TRY(current->resolve_fd(fd));
 
-    return descriptor.inode()->isatty();
+    return descriptor.inode->isatty();
 }
 
 Result<u64> sys_dup2(Registers*, SyscallArgs args)
@@ -198,8 +181,6 @@ Result<u64> sys_dup2(Registers*, SyscallArgs args)
 
     Thread* current = Scheduler::current();
 
-    TRY(check_pledge(current, Promise::p_stdio));
-
     if (newfd < 0 || newfd >= FD_MAX) return err(EBADF);
 
     auto descriptor = *TRY(current->resolve_fd(oldfd));
@@ -218,8 +199,6 @@ Result<u64> sys_pipe(Registers*, SyscallArgs args)
 
     Thread* current = Scheduler::current();
 
-    TRY(check_pledge(current, Promise::p_stdio));
-
     int rfd = TRY(current->allocate_fd(0));
     int wfd = TRY(current->allocate_fd(rfd + 1));
 
@@ -231,8 +210,8 @@ Result<u64> sys_pipe(Registers*, SyscallArgs args)
 
     TRY(Pipe::create(rpipe, wpipe));
 
-    current->fd_table[rfd] = FileDescriptor { TRY(make_shared<OpenFileDescription>(rpipe, O_RDONLY)), 0 };
-    current->fd_table[wfd] = FileDescriptor { TRY(make_shared<OpenFileDescription>(wpipe, O_WRONLY)), 0 };
+    current->fd_table[rfd] = FileDescriptor { rpipe, 0, O_RDONLY };
+    current->fd_table[wfd] = FileDescriptor { wpipe, 0, O_WRONLY };
 
     return 0;
 }
@@ -243,98 +222,9 @@ Result<u64> sys_umask(Registers*, SyscallArgs args)
 
     auto* current = Scheduler::current();
 
-    TRY(check_pledge(current, Promise::p_stdio));
-
     mode_t old_umask = current->umask;
 
     current->umask = new_umask & 0777;
 
     return old_umask;
 }
-
-Result<u64> sys_truncate(Registers*, SyscallArgs args)
-{
-    auto path = TRY(MemoryManager::strdup_from_user(args[0]));
-    size_t length = (size_t)args[1];
-
-    auto* current = Scheduler::current();
-    TRY(check_pledge(current, Promise::p_wpath));
-    auto inode = TRY(VFS::resolve_path(path.chars(), current->auth, current->current_directory));
-
-    if (!VFS::can_write(inode, current->auth)) return err(EACCES);
-
-    TRY(inode->truncate(length));
-
-    return 0;
-}
-
-Result<u64> sys_ftruncate(Registers*, SyscallArgs args)
-{
-    int fd = (int)args[0];
-    size_t length = (size_t)args[1];
-
-    auto* current = Scheduler::current();
-    TRY(check_pledge(current, Promise::p_stdio));
-    auto description = TRY(current->resolve_fd(fd))->description;
-    if (!(description->flags & O_WRONLY)) return err(EBADF);
-
-    TRY(description->inode->truncate(length));
-
-    return 0;
-}
-
-Result<u64> sys_utimensat(Registers*, SyscallArgs args)
-{
-    int dirfd = (int)args[0];
-    auto path = TRY(MemoryManager::strdup_from_user(args[1]));
-    const auto* times = (const struct timespec*)args[2];
-    int flags = (int)args[3];
-
-    auto* current = Scheduler::current();
-    TRY(check_pledge(current, Promise::p_fattr));
-    auto inode = TRY(current->resolve_atfile(dirfd, path, flags & AT_EMPTY_PATH, !(flags & AT_SYMLINK_NOFOLLOW)));
-
-    struct timespec ktimes[2];
-    ktimes[0].tv_sec = ktimes[1].tv_sec = 0;
-    ktimes[0].tv_nsec = ktimes[1].tv_nsec = UTIME_NOW;
-
-    if (times && !MemoryManager::copy_from_user(times, ktimes, sizeof(ktimes))) return err(EFAULT);
-
-    // No permission checks are performed, since no actual modification is done, but the above checks are still
-    // performed.
-    if (ktimes[0].tv_nsec == UTIME_OMIT && ktimes[1].tv_nsec == UTIME_OMIT) return 0;
-
-    bool allow_write_access = ktimes[0].tv_nsec == UTIME_NOW && ktimes[1].tv_nsec == UTIME_NOW;
-
-    if (allow_write_access)
-    {
-        if (!VFS::can_write(inode, current->auth) && current->auth.euid != inode->metadata().uid &&
-            current->auth.euid != 0)
-            return err(EACCES);
-    }
-    else if (current->auth.euid != inode->metadata().uid && current->auth.euid != 0)
-        return err(EPERM);
-
-    auto metadata = inode->metadata();
-    if (ktimes[0].tv_nsec != UTIME_OMIT)
-    {
-        if (ktimes[0].tv_nsec == UTIME_NOW) metadata.atime = *Timer::realtime_clock();
-        else
-        {
-            if (ktimes[0].tv_nsec < 0 || ktimes[0].tv_nsec > 999'999'999) return err(EINVAL);
-            metadata.atime = ktimes[0];
-        }
-    }
-    if (ktimes[1].tv_nsec != UTIME_OMIT)
-    {
-        if (ktimes[1].tv_nsec == UTIME_NOW) metadata.mtime = *Timer::realtime_clock();
-        else
-        {
-            if (ktimes[1].tv_nsec < 0 || ktimes[1].tv_nsec > 999'999'999) return err(EINVAL);
-            metadata.mtime = ktimes[1];
-        }
-    }
-    TRY(inode->set_metadata(metadata));
-
-    return 0;
-}

kernel/src/sys/getdents.cpp

@@ -1,4 +1,3 @@
-#include "Pledge.h"
 #include "fs/VFS.h"
 #include "memory/MemoryManager.h"
 #include "sys/Syscall.h"
@@ -12,23 +11,22 @@ Result<u64> sys_getdents(Registers*, SyscallArgs args)
     usize count = (usize)args[2];
 
     Thread* current = Scheduler::current();
-    TRY(check_pledge(current, Promise::p_stdio));
 
     auto& descriptor = *TRY(current->resolve_fd(fd));
 
-    if (descriptor.inode()->type() != VFS::InodeType::Directory) return err(ENOTDIR);
+    if (descriptor.inode->type() != VFS::InodeType::Directory) return err(ENOTDIR);
 
     usize nwrite = 0;
     while (nwrite < count)
     {
         VFS::DirectoryEntry entry;
-        bool ok = descriptor.inode()->get(descriptor.offset).try_set_value(entry);
+        bool ok = descriptor.inode->get(descriptor.offset).try_set_value(entry);
         if (!ok) break;
 
         descriptor.offset++;
 
         luna_dirent kent;
-        kent.inode = entry.inode->metadata().inum;
+        kent.inode = entry.inode->inode_number();
         strlcpy(kent.name, entry.name.chars(), entry.name.length() + 1);
 
         if (!MemoryManager::copy_to_user_typed(ent, &kent)) return err(EFAULT);

kernel/src/sys/id.cpp

@@ -1,4 +1,3 @@
-#include "Pledge.h"
 #include "memory/MemoryManager.h"
 #include "sys/Syscall.h"
 #include "thread/Scheduler.h"
@@ -6,54 +5,41 @@
 
 Result<u64> sys_getpid(Registers*, SyscallArgs)
 {
-    auto* current = Scheduler::current();
-    TRY(check_pledge(current, Promise::p_stdio));
-    return current->id;
+    return Scheduler::current()->id;
 }
 
 Result<u64> sys_getppid(Registers*, SyscallArgs)
 {
-    auto* current = Scheduler::current();
-    TRY(check_pledge(current, Promise::p_stdio));
-    auto* parent = current->parent;
+    auto* parent = Scheduler::current()->parent;
+
     return parent ? parent->id : 0;
 }
 
 Result<u64> sys_getuid(Registers*, SyscallArgs)
 {
-    auto* current = Scheduler::current();
-    TRY(check_pledge(current, Promise::p_stdio));
-    return current->auth.uid;
+    return Scheduler::current()->auth.uid;
 }
 
 Result<u64> sys_geteuid(Registers*, SyscallArgs)
 {
-    auto* current = Scheduler::current();
-    TRY(check_pledge(current, Promise::p_stdio));
-    return current->auth.euid;
+    return Scheduler::current()->auth.euid;
 }
 
 Result<u64> sys_getgid(Registers*, SyscallArgs)
 {
-    auto* current = Scheduler::current();
-    TRY(check_pledge(current, Promise::p_stdio));
-    return current->auth.gid;
+    return Scheduler::current()->auth.gid;
 }
 
 Result<u64> sys_getegid(Registers*, SyscallArgs)
 {
-    auto* current = Scheduler::current();
-    TRY(check_pledge(current, Promise::p_stdio));
-    return current->auth.egid;
+    return Scheduler::current()->auth.egid;
 }
 
 Result<u64> sys_setuid(Registers*, SyscallArgs args)
 {
     u32 uid = (u32)args[0];
 
-    auto* current = Scheduler::current();
-    TRY(check_pledge(current, Promise::p_id));
-    Credentials& auth = current->auth;
+    Credentials& auth = Scheduler::current()->auth;
 
     if (auth.euid == 0)
     {
@@ -71,9 +57,7 @@ Result<u64> sys_seteuid(Registers*, SyscallArgs args)
 {
     u32 uid = (u32)args[0];
 
-    auto* current = Scheduler::current();
-    TRY(check_pledge(current, Promise::p_id));
-    Credentials& auth = current->auth;
+    Credentials& auth = Scheduler::current()->auth;
 
     if (auth.euid != 0 && uid != auth.uid && uid != auth.suid) return err(EPERM);
     auth.euid = uid;
@@ -85,9 +69,7 @@ Result<u64> sys_setgid(Registers*, SyscallArgs args)
 {
     u32 gid = (u32)args[0];
 
-    auto* current = Scheduler::current();
-    TRY(check_pledge(current, Promise::p_id));
-    Credentials& auth = current->auth;
+    Credentials& auth = Scheduler::current()->auth;
 
     if (auth.euid == 0)
     {
@@ -105,9 +87,7 @@ Result<u64> sys_setegid(Registers*, SyscallArgs args)
 {
     u32 gid = (u32)args[0];
 
-    auto* current = Scheduler::current();
-    TRY(check_pledge(current, Promise::p_id));
-    Credentials& auth = current->auth;
+    Credentials& auth = Scheduler::current()->auth;
 
     if (auth.euid != 0 && gid != auth.gid && gid != auth.sgid) return err(EPERM);
     auth.egid = gid;
@@ -121,9 +101,8 @@ Result<u64> sys_setpgid(Registers*, SyscallArgs args)
     pid_t pgid = (pid_t)args[1];
 
     auto* current = Scheduler::current();
-    TRY(check_pledge(current, Promise::p_proc));
-    if (pid == 0) pid = current->id;
-    if (pgid == 0) pgid = current->id;
+    if (pid == 0) pid = (pid_t)current->id;
+    if (pgid == 0) pgid = (pid_t)current->id;
 
     if (pgid < 0) return err(EINVAL);
 
@@ -134,7 +113,7 @@ Result<u64> sys_setpgid(Registers*, SyscallArgs args)
 
     if (thread->has_called_exec) return err(EPERM);
 
-    if (pgid != current->id)
+    if (pgid != (pid_t)current->id)
     {
         bool pgid_exists = false;
         Scheduler::for_each_in_process_group(pgid, [&pgid_exists](Thread*) {
@@ -154,8 +133,7 @@ Result<u64> sys_getpgid(Registers*, SyscallArgs args)
     pid_t pid = (pid_t)args[0];
 
     auto* current = Scheduler::current();
-    TRY(check_pledge(current, Promise::p_stdio));
-    if (pid == 0) pid = current->id;
+    if (pid == 0) pid = (pid_t)current->id;
 
     if (pid < 0) return err(EINVAL);
 
@@ -172,15 +150,12 @@ Result<u64> sys_fchmodat(Registers*, SyscallArgs args)
     int flags = (int)args[3];
 
     auto* current = Scheduler::current();
-    TRY(check_pledge(current, Promise::p_wpath));
 
     auto inode = TRY(current->resolve_atfile(dirfd, path, flags & AT_EMPTY_PATH, !(flags & AT_SYMLINK_NOFOLLOW)));
 
-    if (current->auth.euid != 0 && current->auth.euid != inode->metadata().uid) return err(EPERM);
+    if (current->auth.euid != 0 && current->auth.euid != inode->uid()) return err(EPERM);
 
-    auto metadata = inode->metadata();
-    metadata.mode = mode;
-    TRY(inode->set_metadata(metadata));
+    TRY(inode->chmod(mode));
 
     return 0;
 }
@@ -189,21 +164,17 @@ Result<u64> sys_fchownat(Registers*, SyscallArgs args)
 {
     int dirfd = (int)args[0];
     auto path = TRY(MemoryManager::strdup_from_user(args[1]));
-    uid_t uid = (u32)args[2];
-    gid_t gid = (u32)args[3];
+    u32 uid = (u32)args[2];
+    u32 gid = (u32)args[3];
     int flags = (int)args[4];
 
     auto* current = Scheduler::current();
-    TRY(check_pledge(current, Promise::p_chown));
 
     auto inode = TRY(current->resolve_atfile(dirfd, path, flags & AT_EMPTY_PATH, !(flags & AT_SYMLINK_NOFOLLOW)));
 
     if (current->auth.euid != 0) return err(EPERM);
 
-    auto metadata = inode->metadata();
-    if (uid != (uid_t)-1) metadata.uid = uid;
-    if (gid != (gid_t)-1) metadata.gid = gid;
-    TRY(inode->set_metadata(metadata));
+    TRY(inode->chown(uid == (u32)-1 ? inode->uid() : uid, gid == (u32)-1 ? inode->gid() : gid));
 
     return 0;
 }

kernel/src/sys/link.cpp

@@ -1,5 +1,4 @@
 #include "Log.h"
-#include "Pledge.h"
 #include "memory/MemoryManager.h"
 #include "sys/Syscall.h"
 #include "thread/Scheduler.h"
@@ -13,7 +12,6 @@ Result<u64> sys_unlinkat(Registers*, SyscallArgs args)
     int flags = (int)args[2];
 
     Thread* current = Scheduler::current();
-    TRY(check_pledge(current, Promise::p_cpath));
 
     auto dirname = TRY(PathParser::dirname(path.view()));
     auto basename = TRY(PathParser::basename(path.view()));
@@ -28,8 +26,8 @@ Result<u64> sys_unlinkat(Registers*, SyscallArgs args)
     auto child = TRY(inode->find(basename.chars()));
     if (flags == AT_REMOVEDIR && child->type() != VFS::InodeType::Directory) return err(ENOTDIR);
 
-    if (current->auth.euid != 0 && VFS::is_sticky(inode) && current->auth.euid != inode->metadata().uid &&
-        current->auth.euid != child->metadata().uid)
+    if (current->auth.euid != 0 && VFS::is_sticky(inode) && current->auth.euid != inode->uid() &&
+        current->auth.euid != child->uid())
         return err(EACCES);
 
     TRY(inode->remove_entry(basename.chars()));
@@ -46,7 +44,6 @@ Result<u64> sys_symlinkat(Registers*, SyscallArgs args)
     if (target.is_empty()) return err(ENOENT);
 
     auto* current = Scheduler::current();
-    TRY(check_pledge(current, Promise::p_cpath));
 
     auto parent = TRY(PathParser::dirname(linkpath.view()));
 
@@ -59,10 +56,7 @@ Result<u64> sys_symlinkat(Registers*, SyscallArgs args)
     TRY(VFS::validate_filename(child_name.view()));
 
     auto inode = TRY(parent_inode->fs()->create_symlink_inode(target.view()));
-    auto metadata = inode->metadata();
-    metadata.uid = current->auth.euid;
-    metadata.gid = current->auth.egid;
-    TRY(inode->set_metadata(metadata));
+    TRY(inode->chown(current->auth.euid, current->auth.egid));
     TRY(parent_inode->add_entry(inode, child_name.chars()));
 
     return 0;
@@ -76,7 +70,6 @@ Result<u64> sys_readlinkat(Registers*, SyscallArgs args)
     usize bufsiz = (usize)args[3];
 
     auto* current = Scheduler::current();
-    TRY(check_pledge(current, Promise::p_rpath));
 
     auto symlink = TRY(current->resolve_atfile(dirfd, path, true, false));
 
@@ -88,6 +81,8 @@ Result<u64> sys_readlinkat(Registers*, SyscallArgs args)
     usize nread = linkpath.length();
     if (nread > bufsiz) nread = bufsiz;
 
+    kdbgln("readlink: reading %zu bytes from symlink (%s)", nread, linkpath.chars());
+
     if (!MemoryManager::copy_to_user(buf, linkpath.chars(), nread)) return err(EFAULT);
 
     return nread;
@@ -102,7 +97,6 @@ Result<u64> sys_linkat(Registers*, SyscallArgs args)
     int flags = (int)args[4];
 
     auto* current = Scheduler::current();
-    TRY(check_pledge(current, Promise::p_cpath));
 
     auto parent = TRY(PathParser::dirname(newpath.view()));