/*
* linux/fs/exec.c
*
* Copyright (C) 1991, 1992 Linus Torvalds
*/
/*
* #!-checking implemented by tytso.
*/
/*
* Demand-loading implemented 01.12.91 - no need to read anything but
* the header into memory. The inode of the executable is put into
* "current->executable", and page faults do the actual loading. Clean.
*
* Once more I can proudly say that linux stood up to being changed: it
* was less than 2 hours work to get demand-loading completely implemented.
*
* Demand loading changed July 1993 by Eric Youngdale. Use mmap instead,
* current->executable is only used by the procfs. This allows a dispatch
* table to check for several different types of binary formats. We keep
* trying until we recognize the file or we run out of supported binary
* formats.
*/
#include <linux/fs.h>
#include <linux/sched.h>
#include <linux/kernel.h>
#include <linux/mm.h>
#include <linux/slab.h>
#include <linux/file.h>
#include <linux/mman.h>
#include <linux/a.out.h>
#include <linux/errno.h>
#include <linux/signal.h>
#include <linux/string.h>
#include <linux/stat.h>
#include <linux/fcntl.h>
#include <linux/ptrace.h>
#include <linux/user.h>
#include <linux/binfmts.h>
#include <linux/personality.h>
#include <linux/smp.h>
#include <linux/smp_lock.h>
#include <linux/init.h>
#include <asm/system.h>
#include <asm/uaccess.h>
#include <asm/pgtable.h>
#include <asm/mmu_context.h>
#include <linux/config.h>
#ifdef CONFIG_KERNELD
#include <linux/kerneld.h>
#endif
asmlinkage int sys_exit(int exit_code);
asmlinkage int sys_brk(unsigned long);
/*
* Here are the actual binaries that will be accepted:
* add more with "register_binfmt()" if using modules...
*
* These are defined again for the 'real' modules if you are using a
* module definition for these routines.
*/
static struct linux_binfmt *formats = (struct linux_binfmt *) NULL;
__initfunc(void binfmt_setup(void))
{
#ifdef CONFIG_BINFMT_MISC
init_misc_binfmt();
#endif
#ifdef CONFIG_BINFMT_ELF
init_elf_binfmt();
#endif
#ifdef CONFIG_BINFMT_ELF32
init_elf32_binfmt();
#endif
#ifdef CONFIG_BINFMT_AOUT
init_aout_binfmt();
#endif
#ifdef CONFIG_BINFMT_AOUT32
init_aout32_binfmt();
#endif
#ifdef CONFIG_BINFMT_JAVA
init_java_binfmt();
#endif
#ifdef CONFIG_BINFMT_EM86
init_em86_binfmt();
#endif
/* This cannot be configured out of the kernel */
init_script_binfmt();
}
int register_binfmt(struct linux_binfmt * fmt)
{
struct linux_binfmt ** tmp = &formats;
if (!fmt)
return -EINVAL;
if (fmt->next)
return -EBUSY;
while (*tmp) {
if (fmt == *tmp)
return -EBUSY;
tmp = &(*tmp)->next;
}
fmt->next = formats;
formats = fmt;
return 0;
}
#ifdef CONFIG_MODULES
int unregister_binfmt(struct linux_binfmt * fmt)
{
struct linux_binfmt ** tmp = &formats;
while (*tmp) {
if (fmt == *tmp) {
*tmp = fmt->next;
return 0;
}
tmp = &(*tmp)->next;
}
return -EINVAL;
}
#endif /* CONFIG_MODULES */
int open_dentry(struct dentry * dentry, int mode)
{
int fd;
struct inode * inode = dentry->d_inode;
if (!inode->i_op || !inode->i_op->default_file_ops)
return -EINVAL;
fd = get_unused_fd();
if (fd >= 0) {
struct file * f = get_empty_filp();
if (!f) {
put_unused_fd(fd);
return -ENFILE;
}
f->f_flags = mode;
f->f_mode = (mode+1) & O_ACCMODE;
f->f_dentry = dentry;
f->f_pos = 0;
f->f_reada = 0;
f->f_op = inode->i_op->default_file_ops;
if (f->f_op->open) {
int error = f->f_op->open(inode,f);
if (error) {
put_filp(f);
put_unused_fd(fd);
return error;
}
}
current->files->fd[fd] = f;
dget(dentry);
}
return fd;
}
/*
* Note that a shared library must be both readable and executable due to
* security reasons.
*
* Also note that we take the address to load from from the file itself.
*/
asmlinkage int sys_uselib(const char * library)
{
int fd, retval;
struct file * file;
struct linux_binfmt * fmt;
lock_kernel();
fd = sys_open(library, 0, 0);
retval = fd;
if (fd < 0)
goto out;
file = current->files->fd[fd];
retval = -ENOEXEC;
if (file && file->f_dentry && file->f_op && file->f_op->read) {
for (fmt = formats ; fmt ; fmt = fmt->next) {
int (*fn)(int) = fmt->load_shlib;
if (!fn)
continue;
retval = fn(fd);
if (retval != -ENOEXEC)
break;
}
}
sys_close(fd);
out:
unlock_kernel();
return retval;
}
/*
* count() counts the number of arguments/envelopes
*/
static int count(char ** argv)
{
int i = 0;
if (argv != NULL) {
for (;;) {
char * p;
int error;
error = get_user(p,argv);
if (error)
return error;
if (!p)
break;
argv++;
i++;
}
}
return i;
}
/*
* 'copy_string()' copies argument/envelope strings from user
* memory to free pages in kernel mem. These are in a format ready
* to be put directly into the top of new user memory.
*
* Modified by TYT, 11/24/91 to add the from_kmem argument, which specifies
* whether the string and the string array are from user or kernel segments:
*
* from_kmem argv * argv **
* 0 user space user space
* 1 kernel space user space
* 2 kernel space kernel space
*
* We do this by playing games with the fs segment register. Since it
* is expensive to load a segment register, we try to avoid calling
* set_fs() unless we absolutely have to.
*/
unsigned long copy_strings(int argc,char ** argv,unsigned long *page,
unsigned long p, int from_kmem)
{
char *str;
unsigned long old_fs;
if (!p)
return 0; /* bullet-proofing */
old_fs = get_fs();
if (from_kmem==2)
set_fs(KERNEL_DS);
while (argc-- > 0) {
int len;
unsigned long pos;
if (from_kmem == 1)
set_fs(KERNEL_DS);
get_user(str, argv+argc);
if (!str)
panic("VFS: argc is wrong");
if (from_kmem == 1)
set_fs(old_fs);
len = strlen_user(str); /* includes the '\0' */
if (p < len) { /* this shouldn't happen - 128kB */
set_fs(old_fs);
return 0;
}
p -= len;
pos = p;
while (len) {
char *pag;
int offset, bytes_to_copy;
offset = pos % PAGE_SIZE;
if (!(pag = (char *) page[pos/PAGE_SIZE]) &&
!(pag = (char *) page[pos/PAGE_SIZE] =
(unsigned long *) get_free_page(GFP_USER))) {
if (from_kmem==2)
set_fs(old_fs);
return 0;
}
bytes_to_copy = PAGE_SIZE - offset;
if (bytes_to_copy > len)
bytes_to_copy = len;
copy_from_user(pag + offset, str, bytes_to_copy);
pos += bytes_to_copy;
str += bytes_to_copy;
len -= bytes_to_copy;
}
}
if (from_kmem==2)
set_fs(old_fs);
return p;
}
unsigned long setup_arg_pages(unsigned long p, struct linux_binprm * bprm)
{
unsigned long stack_base;
struct vm_area_struct *mpnt;
int i;
stack_base = STACK_TOP - MAX_ARG_PAGES*PAGE_SIZE;
p += stack_base;
if (bprm->loader)
bprm->loader += stack_base;
bprm->exec += stack_base;
mpnt = kmem_cache_alloc(vm_area_cachep, SLAB_KERNEL);
if (mpnt) {
mpnt->vm_mm = current->mm;
mpnt->vm_start = PAGE_MASK & (unsigned long) p;
mpnt->vm_end = STACK_TOP;
mpnt->vm_page_prot = PAGE_COPY;
mpnt->vm_flags = VM_STACK_FLAGS;
mpnt->vm_ops = NULL;
mpnt->vm_offset = 0;
mpnt->vm_dentry = NULL;
mpnt->vm_pte = 0;
insert_vm_struct(current->mm, mpnt);
current->mm->total_vm = (mpnt->vm_end - mpnt->vm_start) >> PAGE_SHIFT;
}
for (i = 0 ; i < MAX_ARG_PAGES ; i++) {
if (bprm->page[i]) {
current->mm->rss++;
put_dirty_page(current,bprm->page[i],stack_base);
}
stack_base += PAGE_SIZE;
}
return p;
}
/*
* Read in the complete executable. This is used for "-N" files
* that aren't on a block boundary, and for files on filesystems
* without bmap support.
*/
int read_exec(struct dentry *dentry, unsigned long offset,
char * addr, unsigned long count, int to_kmem)
{
struct file file;
struct inode * inode = dentry->d_inode;
int result = -ENOEXEC;
if (!inode->i_op || !inode->i_op->default_file_ops)
goto end_readexec;
if (init_private_file(&file, dentry, 1))
goto end_readexec;
if (!file.f_op->read)
goto close_readexec;
if (file.f_op->llseek) {
if (file.f_op->llseek(inode,&file,offset,0) != offset)
goto close_readexec;
} else
file.f_pos = offset;
if (to_kmem) {
unsigned long old_fs = get_fs();
set_fs(get_ds());
result = file.f_op->read(inode, &file, addr, count);
set_fs(old_fs);
} else {
result = verify_area(VERIFY_WRITE, addr, count);
if (result)
goto close_readexec;
result = file.f_op->read(inode, &file, addr, count);
}
close_readexec:
if (file.f_op->release)
file.f_op->release(inode,&file);
end_readexec:
return result;
}
static void exec_mmap(void)
{
/*
* The clear_page_tables done later on exec does the right thing
* to the page directory when shared, except for graceful abort
* (the oom is wrong there, too, IMHO)
*/
if (current->mm->count > 1) {
struct mm_struct *mm = kmem_cache_alloc(mm_cachep, SLAB_KERNEL);
if (!mm) {
/* this is wrong, I think. */
oom(current);
return;
}
*mm = *current->mm;
init_new_context(mm);
mm->def_flags = 0; /* should future lockings be kept? */
mm->cpu_vm_mask = (1 << smp_processor_id());
mm->count = 1;
mm->mmap = mm->mmap_cache = NULL;
mm->total_vm = 0;
mm->rss = 0;
current->mm->count--;
current->mm = mm;
new_page_tables(current);
return;
}
flush_cache_mm(current->mm);
exit_mmap(current->mm);
clear_page_tables(current);
flush_tlb_mm(current->mm);
}
/*
* These functions flushes out all traces of the currently running executable
* so that a new one can be started
*/
static inline void flush_old_signals(struct signal_struct *sig)
{
int i;
struct sigaction * sa = sig->action;
for (i=32 ; i != 0 ; i--) {
sa->sa_mask = 0;
sa->sa_flags = 0;
if (sa->sa_handler != SIG_IGN)
sa->sa_handler = NULL;
sa++;
}
}
static inline void flush_old_files(struct files_struct * files)
{
unsigned long j;
j = 0;
for (;;) {
unsigned long set, i;
i = j * __NFDBITS;
if (i >= NR_OPEN)
break;
set = files->close_on_exec.fds_bits[j];
files->close_on_exec.fds_bits[j] = 0;
j++;
for ( ; set ; i++,set >>= 1) {
if (set & 1)
sys_close(i);
}
}
}
void flush_old_exec(struct linux_binprm * bprm)
{
int i;
int ch;
char * name;
if (current->euid == current->uid && current->egid == current->gid)
current->dumpable = 1;
name = bprm->filename;
for (i=0; (ch = *(name++)) != '\0';) {
if (ch == '/')
i = 0;
else
if (i < 15)
current->comm[i++] = ch;
}
current->comm[i] = '\0';
/* Release all of the old mmap stuff. */
exec_mmap();
flush_thread();
if (bprm->e_uid != current->euid || bprm->e_gid != current->egid ||
permission(bprm->dentry->d_inode,MAY_READ))
current->dumpable = 0;
flush_old_signals(current->sig);
flush_old_files(current->files);
}
/*
* Fill the binprm structure from the inode.
* Check permissions, then read the first 512 bytes
*/
int prepare_binprm(struct linux_binprm *bprm)
{
int mode;
int retval,id_change;
struct inode * inode = bprm->dentry->d_inode;
mode = inode->i_mode;
if (!S_ISREG(mode)) /* must be regular file */
return -EACCES;
if (!(mode & 0111)) /* with at least _one_ execute bit set */
return -EACCES;
if (IS_NOEXEC(inode)) /* FS mustn't be mounted noexec */
return -EACCES;
if (!inode->i_sb)
return -EACCES;
if ((retval = permission(inode, MAY_EXEC)) != 0)
return retval;
/* better not execute files which are being written to */
if (inode->i_writecount > 0)
return -ETXTBSY;
bprm->e_uid = current->euid;
bprm->e_gid = current->egid;
id_change = 0;
/* Set-uid? */
if (mode & S_ISUID) {
bprm->e_uid = inode->i_uid;
if (bprm->e_uid != current->euid)
id_change = 1;
}
/* Set-gid? */
/*
* If setgid is set but no group execute bit then this
* is a candidate for mandatory locking, not a setgid
* executable.
*/
if ((mode & (S_ISGID | S_IXGRP)) == (S_ISGID | S_IXGRP)) {
bprm->e_gid = inode->i_gid;
if (!in_group_p(bprm->e_gid))
id_change = 1;
}
if (id_change) {
/* We can't suid-execute if we're sharing parts of the executable */
/* or if we're being traced (or if suid execs are not allowed) */
/* (current->mm->count > 1 is ok, as we'll get a new mm anyway) */
if (IS_NOSUID(inode)
|| (current->flags & PF_PTRACED)
|| (current->fs->count > 1)
|| (atomic_read(¤t->sig->count) > 1)
|| (current->files->count > 1)) {
if (!suser())
return -EPERM;
}
}
memset(bprm->buf,0,sizeof(bprm->buf));
return read_exec(bprm->dentry,0,bprm->buf,128,1);
}
void remove_arg_zero(struct linux_binprm *bprm)
{
if (bprm->argc) {
unsigned long offset;
char * page;
offset = bprm->p % PAGE_SIZE;
page = (char*)bprm->page[bprm->p/PAGE_SIZE];
while(bprm->p++,*(page+offset++))
if(offset==PAGE_SIZE){
offset=0;
page = (char*)bprm->page[bprm->p/PAGE_SIZE];
}
bprm->argc--;
}
}
/*
* cycle the list of binary formats handler, until one recognizes the image
*/
int search_binary_handler(struct linux_binprm *bprm,struct pt_regs *regs)
{
int try,retval=0;
struct linux_binfmt *fmt;
#ifdef __alpha__
/* handle /sbin/loader.. */
{
struct exec * eh = (struct exec *) bprm->buf;
if (!bprm->loader && eh->fh.f_magic == 0x183 &&
(eh->fh.f_flags & 0x3000) == 0x3000)
{
char * dynloader[] = { "/sbin/loader" };
struct dentry * dentry;
dput(bprm->dentry);
bprm->dentry = NULL;
remove_arg_zero(bprm);
bprm->p = copy_strings(1, dynloader, bprm->page, bprm->p, 2);
bprm->argc++;
bprm->loader = bprm->p;
dentry = open_namei(dynloader[0], 0, 0);
retval = PTR_ERR(dentry);
if (IS_ERR(dentry))
return retval;
bprm->dentry = dentry;
retval = prepare_binprm(bprm);
if (retval<0)
return retval;
/* should call search_binary_handler recursively here,
but it does not matter */
}
}
#endif
for (try=0; try<2; try++) {
for (fmt = formats ; fmt ; fmt = fmt->next) {
int (*fn)(struct linux_binprm *, struct pt_regs *) = fmt->load_binary;
if (!fn)
continue;
retval = fn(bprm, regs);
if (retval >= 0) {
if (bprm->dentry)
dput(bprm->dentry);
bprm->dentry = NULL;
current->did_exec = 1;
return retval;
}
if (retval != -ENOEXEC)
break;
if (!bprm->dentry) /* We don't have the dentry anymore */
return retval;
}
if (retval != -ENOEXEC) {
break;
#ifdef CONFIG_KERNELD
}else{
#define printable(c) (((c)=='\t') || ((c)=='\n') || (0x20<=(c) && (c)<=0x7e))
char modname[20];
if (printable(bprm->buf[0]) &&
printable(bprm->buf[1]) &&
printable(bprm->buf[2]) &&
printable(bprm->buf[3]))
break; /* -ENOEXEC */
sprintf(modname, "binfmt-%04x", *(unsigned short *)(&bprm->buf[2]));
request_module(modname);
#endif
}
}
return retval;
}
/*
* sys_execve() executes a new program.
*/
int do_execve(char * filename, char ** argv, char ** envp, struct pt_regs * regs)
{
struct linux_binprm bprm;
struct dentry * dentry;
int retval;
int i;
bprm.p = PAGE_SIZE*MAX_ARG_PAGES-sizeof(void *);
for (i=0 ; i<MAX_ARG_PAGES ; i++) /* clear page-table */
bprm.page[i] = 0;
dentry = open_namei(filename, 0, 0);
retval = PTR_ERR(dentry);
if (IS_ERR(dentry))
return retval;
bprm.dentry = dentry;
bprm.filename = filename;
bprm.sh_bang = 0;
bprm.java = 0;
bprm.loader = 0;
bprm.exec = 0;
if ((bprm.argc = count(argv)) < 0) {
dput(dentry);
return bprm.argc;
}
if ((bprm.envc = count(envp)) < 0) {
dput(dentry);
return bprm.envc;
}
retval = prepare_binprm(&bprm);
if (retval >= 0) {
bprm.p = copy_strings(1, &bprm.filename, bprm.page, bprm.p, 2);
bprm.exec = bprm.p;
bprm.p = copy_strings(bprm.envc,envp,bprm.page,bprm.p,0);
bprm.p = copy_strings(bprm.argc,argv,bprm.page,bprm.p,0);
if (!bprm.p)
retval = -E2BIG;
}
if (retval >= 0)
retval = search_binary_handler(&bprm,regs);
if (retval >= 0)
/* execve success */
return retval;
/* Something went wrong, return the inode and free the argument pages*/
if (bprm.dentry)
dput(bprm.dentry);
for (i=0 ; i<MAX_ARG_PAGES ; i++)
free_page(bprm.page[i]);
return retval;
}