Multiuser, Capabilities and Access Controls (#54)

[lunaix-os.git] / lunaix-os / kernel / process / process.c

diff --git a/lunaix-os/kernel/process/process.c b/lunaix-os/kernel/process/process.c
index 1647e10dd8bc24765a4d9bc50283c3bc66416f50..2a24f50bbb6c9791144cb5376c3471178ba3de2d 100644 (file)
--- a/lunaix-os/kernel/process/process.c
+++ b/lunaix-os/kernel/process/process.c
@@ -1,97 +1,95 @@
 #include <klibc/string.h>
 #include <lunaix/clock.h>
 #include <klibc/string.h>
 #include <lunaix/clock.h>

-#include <lunaix/common.h>
-#include <lunaix/mm/pmm.h>
+#include <lunaix/mm/mmap.h>
+#include <lunaix/mm/vmm.h>

 #include <lunaix/mm/region.h>
 #include <lunaix/mm/valloc.h>
 #include <lunaix/mm/region.h>
 #include <lunaix/mm/valloc.h>

-#include <lunaix/mm/vmm.h>

 #include <lunaix/process.h>
 #include <lunaix/spike.h>
 #include <lunaix/status.h>
 #include <lunaix/syscall.h>
 #include <lunaix/syslog.h>
 #include <lunaix/process.h>
 #include <lunaix/spike.h>
 #include <lunaix/status.h>
 #include <lunaix/syscall.h>
 #include <lunaix/syslog.h>

+#include <lunaix/exec.h>
+#include <lunaix/fs.h>
+
+#include <asm/abi.h>
+#include <asm/mm_defs.h>

 
 LOG_MODULE("PROC")
 
 
 LOG_MODULE("PROC")
 

-void*
-__dup_pagetable(pid_t pid, uintptr_t mount_point)
+int
+spawn_process(struct thread** created, ptr_t entry, bool with_ustack) 

 {
 {

-    void* ptd_pp = pmm_alloc_page(pid, PP_FGPERSIST);
-    vmm_set_mapping(VMS_SELF, PG_MOUNT_1, ptd_pp, PG_PREM_RW, VMAP_NULL);
+    struct proc_info* kproc = alloc_process();
+    struct proc_mm* mm = vmspace(kproc);

 
 

-    x86_page_table* ptd = PG_MOUNT_1;
-    x86_page_table* pptd = (x86_page_table*)(mount_point | (0x3FF << 12));
+    procvm_initvms_mount(mm);

 
 

-    size_t kspace_l1inx = L1_INDEX(KERNEL_MM_BASE);
+    struct thread* kthread = create_thread(kproc, with_ustack);

 
 

-    for (size_t i = 0; i < PG_MAX_ENTRIES - 1; i++) {
+    if (!kthread) {
+        procvm_unmount(mm);
+        delete_process(kproc);
+        return -1;
+    }

 
 

-        x86_pte_t ptde = pptd->entry[i];
-        // 空或者是未在内存中的L1页表项直接照搬过去。
-        // 内核地址空间直接共享过去。
-        if (!ptde || i >= kspace_l1inx || !(ptde & PG_PRESENT)) {
-            ptd->entry[i] = ptde;
-            continue;
-        }
+    commit_process(kproc);
+    start_thread(kthread, entry);

 
 

-        // 复制L2页表
-        void* pt_pp = pmm_alloc_page(pid, PP_FGPERSIST);
-        vmm_set_mapping(VMS_SELF, PG_MOUNT_2, pt_pp, PG_PREM_RW, VMAP_NULL);
+    procvm_unmount(mm);

 
 

-        x86_page_table* ppt = (x86_page_table*)(mount_point | (i << 12));
-        x86_page_table* pt = PG_MOUNT_2;
+    if (created) {
+        *created = kthread;
+    }

 
 

-        for (size_t j = 0; j < PG_MAX_ENTRIES; j++) {
-            x86_pte_t pte = ppt->entry[j];
-            pmm_ref_page(pid, PG_ENTRY_ADDR(pte));
-            pt->entry[j] = pte;
-        }
+    return 0;
+}

 
 

-        ptd->entry[i] = (uintptr_t)pt_pp | PG_ENTRY_FLAGS(ptde);
+int
+spawn_process_usr(struct thread** created, char* path, 
+                    const char** argv, const char** envp)
+{
+    // FIXME remote injection of user stack not yet implemented
+
+    struct proc_info* proc   = alloc_process();
+    struct proc_mm*   mm     = vmspace(proc);
+    
+    assert(!kernel_process(proc));
+
+    procvm_initvms_mount(mm);
+
+    int errno = 0;
+    struct thread* main_thread;
+    if (!(main_thread = create_thread(proc, true))) {
+        errno = ENOMEM;
+        goto fail;

     }
 
     }
 

-    ptd->entry[PG_MAX_ENTRIES - 1] = NEW_L1_ENTRY(T_SELF_REF_PERM, ptd_pp);
+    struct exec_host container;
+    exec_init_container(&container, main_thread, VMS_MOUNT_1, argv, envp);
+    if ((errno = exec_load_byname(&container, path))) {
+        goto fail;
+    }

 
 

-    return ptd_pp;
-}
+    commit_process(proc);
+    start_thread(main_thread, container.exe.entry);

 
 

-void
-__del_pagetable(pid_t pid, uintptr_t mount_point)
-{
-    x86_page_table* pptd = (x86_page_table*)(mount_point | (0x3FF << 12));
-
-    // only remove user address space
-    for (size_t i = 0; i < L1_INDEX(KERNEL_MM_BASE); i++) {
-        x86_pte_t ptde = pptd->entry[i];
-        if (!ptde || !(ptde & PG_PRESENT)) {
-            continue;
-        }
-
-        x86_page_table* ppt = (x86_page_table*)(mount_point | (i << 12));
-
-        for (size_t j = 0; j < PG_MAX_ENTRIES; j++) {
-            x86_pte_t pte = ppt->entry[j];
-            // free the 4KB data page
-            if ((pte & PG_PRESENT)) {
-                pmm_free_page(pid, PG_ENTRY_ADDR(pte));
-            }
-        }
-        // free the L2 page table
-        pmm_free_page(pid, PG_ENTRY_ADDR(ptde));
+    if (created) {
+        *created = main_thread;

     }
     }

-    // free the L1 directory
-    pmm_free_page(pid, PG_ENTRY_ADDR(pptd->entry[PG_MAX_ENTRIES - 1]));
-}

 
 

-void*
-vmm_dup_vmspace(pid_t pid)
-{
-    return __dup_pagetable(pid, VMS_SELF);
+    procvm_unmount(mm);
+    return 0;
+
+fail:
+    procvm_unmount(mm);
+    delete_process(proc);
+    return errno;

 }
 
 }
 

-__DEFINE_LXSYSCALL(pid_t, fork)
-{
-    return dup_proc();
+
+ptr_t proc_vmroot() {
+    return __current->mm->vmroot;

 }
 
 __DEFINE_LXSYSCALL(pid_t, getpid)
 }
 
 __DEFINE_LXSYSCALL(pid_t, getpid)


@@ -114,7 +112,7 @@ __DEFINE_LXSYSCALL2(int, setpgid, pid_t, pid, pid_t, pgid)
     struct proc_info* proc = pid ? get_process(pid) : __current;
 
     if (!proc) {
     struct proc_info* proc = pid ? get_process(pid) : __current;
 
     if (!proc) {

-        __current->k_status = EINVAL;
+        syscall_result(EINVAL);

         return -1;
     }
 
         return -1;
     }
 


@@ -123,7 +121,7 @@ __DEFINE_LXSYSCALL2(int, setpgid, pid_t, pid, pid_t, pgid)
     struct proc_info* gruppenfuhrer = get_process(pgid);
 
     if (!gruppenfuhrer || proc->pgid == gruppenfuhrer->pid) {
     struct proc_info* gruppenfuhrer = get_process(pgid);
 
     if (!gruppenfuhrer || proc->pgid == gruppenfuhrer->pid) {

-        __current->k_status = EINVAL;
+        syscall_result(EINVAL);

         return -1;
     }
 
         return -1;
     }
 


@@ -134,141 +132,121 @@ __DEFINE_LXSYSCALL2(int, setpgid, pid_t, pid, pid_t, pgid)
     return 0;
 }
 
     return 0;
 }
 

-void
-init_proc_user_space(struct proc_info* pcb)
-{
-    vmm_mount_pd(VMS_MOUNT_1, pcb->page_table);
+static inline bool
+__can_change_real_id(const struct user_scope* procu, caps_t id_cap) {
+    if (uscope_with_capability(procu, id_cap)) {
+        return true;
+    }

 
 

-    /*---  分配用户栈  ---*/
+    if (check_current_acl(0, 0) != ACL_NO_MATCH) {
+        return true;
+    }
+
+    return false;
+}

 
 

-    struct mm_region* stack_vm;
+__DEFINE_LXSYSCALL1(int, setuid, uid_t, uid)
+{
+    struct user_scope* procu;

 
 

-    stack_vm = region_create_range(
-      USTACK_END, USTACK_SIZE, REGION_RW | REGION_RSHARED | REGION_ANON);
-    // 注册用户栈区域
-    region_add(&pcb->mm.regions, stack_vm);
+    procu = current_user_scope();

 
 

-    // 预留地址空间，具体物理页将由Page Fault Handler按需分配。
-    for (uintptr_t i = PG_ALIGN(USTACK_END); i < USTACK_TOP; i += PG_SIZE) {
-        vmm_set_mapping(VMS_MOUNT_1, i, 0, PG_ALLOW_USER | PG_WRITE, VMAP_NULL);
+    if (__can_change_real_id(procu, CAP_SETUID)) 
+    {
+        procu->ruid = uid;

     }
 
     }
 

-    // TODO other uspace initialization stuff
+    __current->suid = uid;
+    __current->euid = uid;

 
 

-    vmm_unmount_pd(VMS_MOUNT_1);
+    return 0;

 }
 
 }
 

-void
-__mark_region(uintptr_t start_vpn, uintptr_t end_vpn, int attr)
+__DEFINE_LXSYSCALL1(int, setgid, gid_t, gid)

 {
 {

-    for (size_t i = start_vpn; i <= end_vpn; i++) {
-        x86_pte_t* curproc = &PTE_MOUNTED(VMS_SELF, i);
-        x86_pte_t* newproc = &PTE_MOUNTED(VMS_MOUNT_1, i);
-        cpu_invplg(newproc);
-
-        if ((attr & REGION_MODE_MASK) == REGION_RSHARED) {
-            // 如果读共享，则将两者的都标注为只读，那么任何写入都将会应用COW策略。
-            cpu_invplg(curproc);
-            cpu_invplg(i << 12);
-            *curproc = *curproc & ~PG_WRITE;
-            *newproc = *newproc & ~PG_WRITE;
-        } else {
-            // 如果是私有页，则将该页从新进程中移除。
-            *newproc = 0;
-        }
+    struct user_scope* procu;
+
+    procu = current_user_scope();
+
+    if (__can_change_real_id(procu, CAP_SETGID)) 
+    {
+        procu->rgid = gid;

     }
     }

+
+    __current->sgid = gid;
+    __current->egid = gid;
+
+    return 0;

 }
 
 }
 

-void
-__copy_fdtable(struct proc_info* pcb)
+__DEFINE_LXSYSCALL1(int, seteuid, uid_t, euid)

 {
 {

-    for (size_t i = 0; i < VFS_MAX_FD; i++) {
-        struct v_fd* fd = __current->fdtable->fds[i];
-        if (!fd)
-            continue;
-        vfs_dup_fd(fd, &pcb->fdtable->fds[i]);
-    }
+    __current->euid = euid;
+
+    return 0;

 }
 
 }
 

-pid_t
-dup_proc()
+__DEFINE_LXSYSCALL1(int, setegid, gid_t, egid)

 {
 {

-    struct proc_info* pcb = alloc_process();
-    pcb->mm.u_heap = __current->mm.u_heap;
-    pcb->intr_ctx = __current->intr_ctx;
-    pcb->parent = __current;
+    __current->egid = egid;

 
 

-    memcpy(pcb->fxstate, __current->fxstate, 512);
-
-    if (__current->cwd) {
-        pcb->cwd = __current->cwd;
-        vfs_ref_dnode(pcb->cwd);
-    }
+    return 0;
+}

 
 

-    __copy_fdtable(pcb);
-    region_copy(&__current->mm.regions, &pcb->mm.regions);
+__DEFINE_LXSYSCALL2(int, setgroups, const gid_t*, gids, unsigned int, len)
+{
+    struct user_scope* procu;

 
 

-    setup_proc_mem(pcb, VMS_SELF);
+    procu = current_user_scope();

 
 

-    // 根据 mm_region 进一步配置页表
+    if (check_current_acl(0, 0) == ACL_NO_MATCH) {
+        return EPERM;
+    }

 
 

-    struct mm_region *pos, *n;
-    llist_for_each(pos, n, &pcb->mm.regions, head)
-    {
-        // 如果写共享，则不作处理。
-        if ((pos->attr & REGION_WSHARED)) {
-            continue;
-        }
-
-        uintptr_t start_vpn = pos->start >> 12;
-        uintptr_t end_vpn = pos->end >> 12;
-        __mark_region(start_vpn, end_vpn, pos->attr);
+    if (uscope_with_capability(procu, CAP_SETGID)) {
+        return EPERM;

     }
 
     }
 

-    vmm_unmount_pd(VMS_MOUNT_1);
+    return uscope_setgroups(procu, gids, len);
+}

 
 

-    // 正如同fork，返回两次。
-    pcb->intr_ctx.registers.eax = 0;

 
 

-    commit_process(pcb);
+__DEFINE_LXSYSCALL(uid_t, getuid)
+{
+    return current_user_scope()->ruid;
+}

 
 

-    return pcb->pid;
+__DEFINE_LXSYSCALL(gid_t, getgid)
+{
+    return current_user_scope()->rgid;

 }
 
 }
 

-extern void __kernel_end;
+__DEFINE_LXSYSCALL(uid_t, geteuid)
+{
+    return __current->euid;
+}

 
 

-void
-setup_proc_mem(struct proc_info* proc, uintptr_t usedMnt)
+__DEFINE_LXSYSCALL(gid_t, getegid)

 {
 {

-    // copy the entire kernel page table
-    pid_t pid = proc->pid;
-    void* pt_copy = __dup_pagetable(pid, usedMnt);
-
-    vmm_mount_pd(VMS_MOUNT_1, pt_copy); // 将新进程的页表挂载到挂载点#2
-
-    // copy the kernel stack
-    for (size_t i = KSTACK_START >> 12; i <= KSTACK_TOP >> 12; i++) {
-        volatile x86_pte_t* ppte = &PTE_MOUNTED(VMS_MOUNT_1, i);
-
-        /*
-            This is a fucking nightmare, the TLB caching keep the rewrite to PTE
-           from updating. Even the Nightmare Moon the Evil is far less nasty
-           than this. It took me hours of debugging to figure this out.
-
-            In the name of Celestia our glorious goddess, I will fucking HATE
-           the TLB for the rest of my LIFE!
-        */
-        cpu_invplg(ppte);
-
-        x86_pte_t p = *ppte;
-        void* ppa = vmm_dup_page(pid, PG_ENTRY_ADDR(p));
-        pmm_free_page(pid, PG_ENTRY_ADDR(p));
-        *ppte = (p & 0xfff) | (uintptr_t)ppa;
-    }
+    return __current->egid;
+}
+
+__DEFINE_LXSYSCALL2(int, getgroups, gid_t*, out_buf, unsigned int, len)
+{
+    struct user_scope* procu;
+    struct ugroup_obj* gobj;
+
+    procu = current_user_scope();
+    gobj  = user_groups(procu);

 
 

-    // 我们不需要分配内核的区域，因为所有的内核代码和数据段只能通过系统调用来访问，任何非法的访问
-    // 都会导致eip落在区域外面，从而segmentation fault.
+    assert(gobj);
+    len = MIN(gobj->maxcap, len);

 
 

-    // 至于其他的区域我们暂时没有办法知道，因为那需要知道用户程序的信息。我们留到之后在处理。
-    proc->page_table = pt_copy;
-}
\ No newline at end of file
+    unsigned i = 0;
+    for (; i < len && gobj->list[i] != grp_list_end; i++)
+    {
+        out_buf[i] = gobj->list[i];
+    }
+    
+    return i + 1;
+}