freebsd-skq/sys/kern/kern_exec.c

845 lines
20 KiB
C
Raw Normal View History

/*
* Copyright (c) 1993, David Greenman
* All rights reserved.
1994-05-24 10:09:53 +00:00
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* 1. Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* 2. Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
*
* THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
1994-05-24 10:09:53 +00:00
* ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
* ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
1994-05-24 10:09:53 +00:00
* FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
* DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
* OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
* HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
* LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
* OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
* SUCH DAMAGE.
*
1999-08-28 01:08:13 +00:00
* $FreeBSD$
1994-05-24 10:09:53 +00:00
*/
#include <sys/param.h>
#include <sys/systm.h>
#include <sys/lock.h>
#include <sys/mutex.h>
#include <sys/sysproto.h>
#include <sys/signalvar.h>
#include <sys/kernel.h>
#include <sys/mount.h>
#include <sys/filedesc.h>
1995-10-21 08:38:13 +00:00
#include <sys/fcntl.h>
#include <sys/acct.h>
#include <sys/exec.h>
#include <sys/imgact.h>
#include <sys/imgact_elf.h>
#include <sys/wait.h>
#include <sys/proc.h>
#include <sys/pioctl.h>
#include <sys/malloc.h>
#include <sys/namei.h>
#include <sys/sysent.h>
#include <sys/shm.h>
#include <sys/sysctl.h>
#include <sys/vnode.h>
#include <vm/vm.h>
#include <vm/vm_param.h>
#include <vm/pmap.h>
#include <vm/vm_page.h>
#include <vm/vm_map.h>
#include <vm/vm_kern.h>
#include <vm/vm_extern.h>
#include <vm/vm_object.h>
#include <vm/vm_pager.h>
#include <machine/reg.h>
MALLOC_DEFINE(M_PARGS, "proc-args", "Process arguments");
static register_t *exec_copyout_strings __P((struct image_params *));
/* XXX This should be vm_size_t. */
static u_long ps_strings = PS_STRINGS;
SYSCTL_ULONG(_kern, KERN_PS_STRINGS, ps_strings, CTLFLAG_RD, &ps_strings, 0, "");
/* XXX This should be vm_size_t. */
static u_long usrstack = USRSTACK;
SYSCTL_ULONG(_kern, KERN_USRSTACK, usrstack, CTLFLAG_RD, &usrstack, 0, "");
u_long ps_arg_cache_limit = PAGE_SIZE / 16;
SYSCTL_LONG(_kern, OID_AUTO, ps_arg_cache_limit, CTLFLAG_RW,
&ps_arg_cache_limit, 0, "");
int ps_argsopen = 1;
SYSCTL_INT(_kern, OID_AUTO, ps_argsopen, CTLFLAG_RW, &ps_argsopen, 0, "");
/*
* Each of the items is a pointer to a `const struct execsw', hence the
* double pointer here.
*/
static const struct execsw **execsw;
#ifndef _SYS_SYSPROTO_H_
struct execve_args {
char *fname;
char **argv;
char **envv;
};
#endif
/*
* execve() system call.
*/
int
execve(p, uap)
struct proc *p;
register struct execve_args *uap;
{
struct nameidata nd, *ndp;
register_t *stack_base;
int error, len, i;
struct image_params image_params, *imgp;
struct vattr attr;
int (*img_first) __P((struct image_params *));
imgp = &image_params;
/*
* Initialize part of the common data
*/
imgp->proc = p;
imgp->uap = uap;
imgp->attr = &attr;
imgp->argc = imgp->envc = 0;
imgp->argv0 = NULL;
imgp->entry_addr = 0;
imgp->vmspace_destroyed = 0;
imgp->interpreted = 0;
imgp->interpreter_name[0] = '\0';
imgp->auxargs = NULL;
imgp->vp = NULL;
imgp->firstpage = NULL;
imgp->ps_strings = 0;
imgp->auxarg_size = 0;
/*
* Allocate temporary demand zeroed space for argument and
* environment strings
*/
imgp->stringbase = (char *)kmem_alloc_wait(exec_map, ARG_MAX + PAGE_SIZE);
if (imgp->stringbase == NULL) {
error = ENOMEM;
goto exec_fail;
}
imgp->stringp = imgp->stringbase;
imgp->stringspace = ARG_MAX;
imgp->image_header = imgp->stringbase + ARG_MAX;
/*
* Translate the file name. namei() returns a vnode pointer
* in ni_vp amoung other things.
*/
ndp = &nd;
NDINIT(ndp, LOOKUP, LOCKLEAF | FOLLOW | SAVENAME,
UIO_USERSPACE, uap->fname, p);
interpret:
error = namei(ndp);
if (error) {
kmem_free_wakeup(exec_map, (vm_offset_t)imgp->stringbase,
ARG_MAX + PAGE_SIZE);
1995-05-30 08:16:23 +00:00
goto exec_fail;
}
imgp->vp = ndp->ni_vp;
imgp->fname = uap->fname;
/*
* Check file permissions (also 'opens' file)
*/
error = exec_check_permissions(imgp);
if (error) {
VOP_UNLOCK(imgp->vp, 0, p);
goto exec_fail_dealloc;
}
error = exec_map_first_page(imgp);
VOP_UNLOCK(imgp->vp, 0, p);
if (error)
goto exec_fail_dealloc;
/*
* If the current process has a special image activator it
* wants to try first, call it. For example, emulating shell
* scripts differently.
*/
error = -1;
if ((img_first = imgp->proc->p_sysent->sv_imgact_try) != NULL)
error = img_first(imgp);
/*
* Loop through the list of image activators, calling each one.
* An activator returns -1 if there is no match, 0 on success,
* and an error otherwise.
*/
for (i = 0; error == -1 && execsw[i]; ++i) {
if (execsw[i]->ex_imgact == NULL ||
execsw[i]->ex_imgact == img_first) {
continue;
}
error = (*execsw[i]->ex_imgact)(imgp);
}
if (error) {
if (error == -1)
error = ENOEXEC;
goto exec_fail_dealloc;
}
/*
* Special interpreter operation, cleanup and loop up to try to
* activate the interpreter.
*/
if (imgp->interpreted) {
exec_unmap_first_page(imgp);
/* free name buffer and old vnode */
NDFREE(ndp, NDF_ONLY_PNBUF);
vrele(ndp->ni_vp);
/* set new name to that of the interpreter */
NDINIT(ndp, LOOKUP, LOCKLEAF | FOLLOW | SAVENAME,
UIO_SYSSPACE, imgp->interpreter_name, p);
goto interpret;
}
/*
* Copy out strings (args and env) and initialize stack base
*/
stack_base = exec_copyout_strings(imgp);
p->p_vmspace->vm_minsaddr = (char *)stack_base;
/*
* If custom stack fixup routine present for this process
* let it do the stack setup.
* Else stuff argument count as first item on stack
*/
if (p->p_sysent->sv_fixup)
(*p->p_sysent->sv_fixup)(&stack_base, imgp);
else
suword(--stack_base, imgp->argc);
/*
* For security and other reasons, the file descriptor table cannot
* be shared after an exec.
*/
if (p->p_fd->fd_refcnt > 1) {
struct filedesc *tmp;
tmp = fdcopy(p);
fdfree(p);
p->p_fd = tmp;
}
/* Stop profiling */
stopprofclock(p);
/* close files on exec */
fdcloseexec(p);
/* reset caught signals */
execsigs(p);
/* name this process - nameiexec(p, ndp) */
len = min(ndp->ni_cnd.cn_namelen,MAXCOMLEN);
bcopy(ndp->ni_cnd.cn_nameptr, p->p_comm, len);
p->p_comm[len] = 0;
1995-05-30 08:16:23 +00:00
/*
* mark as execed, wakeup the process that vforked (if any) and tell
* it that it now has its own resources back
*/
2001-03-07 03:27:32 +00:00
PROC_LOCK(p);
p->p_flag |= P_EXEC;
if (p->p_pptr && (p->p_flag & P_PPWAIT)) {
p->p_flag &= ~P_PPWAIT;
wakeup((caddr_t)p->p_pptr);
}
1995-05-30 08:16:23 +00:00
/*
* Implement image setuid/setgid.
*
* Don't honor setuid/setgid if the filesystem prohibits it or if
* the process is being traced.
*/
if ((((attr.va_mode & VSUID) && p->p_ucred->cr_uid != attr.va_uid) ||
((attr.va_mode & VSGID) && p->p_ucred->cr_gid != attr.va_gid)) &&
(imgp->vp->v_mount->mnt_flag & MNT_NOSUID) == 0 &&
(p->p_flag & P_TRACED) == 0) {
2001-03-07 03:27:32 +00:00
PROC_UNLOCK(p);
/*
* Turn off syscall tracing for set-id programs, except for
* root.
*/
if (p->p_tracep && suser(p)) {
p->p_traceflag = 0;
vrele(p->p_tracep);
p->p_tracep = NULL;
}
/*
* Set the new credentials.
*/
p->p_ucred = crcopy(p->p_ucred);
if (attr.va_mode & VSUID)
change_euid(p, attr.va_uid);
if (attr.va_mode & VSGID)
p->p_ucred->cr_gid = attr.va_gid;
setsugid(p);
setugidsafety(p);
} else {
if (p->p_ucred->cr_uid == p->p_cred->p_ruid &&
p->p_ucred->cr_gid == p->p_cred->p_rgid)
p->p_flag &= ~P_SUGID;
2001-03-07 03:27:32 +00:00
PROC_UNLOCK(p);
}
/*
* Implement correct POSIX saved-id behavior.
*/
p->p_cred->p_svuid = p->p_ucred->cr_uid;
p->p_cred->p_svgid = p->p_ucred->cr_gid;
/*
* Store the vp for use in procfs
*/
if (p->p_textvp) /* release old reference */
vrele(p->p_textvp);
VREF(ndp->ni_vp);
p->p_textvp = ndp->ni_vp;
/*
* notify others that we exec'd
*/
2001-03-07 03:27:32 +00:00
PROC_LOCK(p);
KNOTE(&p->p_klist, NOTE_EXEC);
/*
* If tracing the process, trap to debugger so breakpoints
* can be set before the program executes.
*/
2001-03-07 03:27:32 +00:00
_STOPEVENT(p, S_EXEC, 0);
if (p->p_flag & P_TRACED)
psignal(p, SIGTRAP);
/* clear "fork but no exec" flag, as we _are_ execing */
p->p_acflag &= ~AFORK;
/* Set values passed into the program in registers. */
setregs(p, imgp->entry_addr, (u_long)(uintptr_t)stack_base,
imgp->ps_strings);
/* Free any previous argument cache */
if (p->p_args && --p->p_args->ar_ref == 0)
FREE(p->p_args, M_PARGS);
p->p_args = NULL;
/* Cache arguments if they fit inside our allowance */
i = imgp->endargs - imgp->stringbase;
if (ps_arg_cache_limit >= i + sizeof(struct pargs)) {
2001-03-07 03:27:32 +00:00
PROC_UNLOCK(p);
MALLOC(p->p_args, struct pargs *, sizeof(struct pargs) + i,
M_PARGS, M_WAITOK);
2001-03-07 03:27:32 +00:00
KASSERT(p->p_args != NULL, ("malloc of p_args failed"));
PROC_LOCK(p);
p->p_args->ar_ref = 1;
p->p_args->ar_length = i;
bcopy(imgp->stringbase, p->p_args->ar_args, i);
}
2001-03-07 03:27:32 +00:00
PROC_UNLOCK(p);
exec_fail_dealloc:
/*
* free various allocated resources
*/
if (imgp->firstpage)
exec_unmap_first_page(imgp);
if (imgp->stringbase != NULL)
kmem_free_wakeup(exec_map, (vm_offset_t)imgp->stringbase,
ARG_MAX + PAGE_SIZE);
if (imgp->vp) {
NDFREE(ndp, NDF_ONLY_PNBUF);
vrele(imgp->vp);
}
if (error == 0)
return (0);
exec_fail:
if (imgp->vmspace_destroyed) {
/* sorry, no more process anymore. exit gracefully */
exit1(p, W_EXITCODE(0, SIGABRT));
/* NOT REACHED */
return(0);
} else {
return(error);
}
}
int
exec_map_first_page(imgp)
struct image_params *imgp;
{
2001-05-21 18:30:50 +00:00
int rv, i;
int initial_pagein;
vm_page_t ma[VM_INITIAL_PAGEIN];
vm_object_t object;
if (imgp->firstpage) {
exec_unmap_first_page(imgp);
}
VOP_GETVOBJECT(imgp->vp, &object);
mtx_lock(&vm_mtx);
ma[0] = vm_page_grab(object, 0, VM_ALLOC_NORMAL | VM_ALLOC_RETRY);
if ((ma[0]->valid & VM_PAGE_BITS_ALL) != VM_PAGE_BITS_ALL) {
initial_pagein = VM_INITIAL_PAGEIN;
if (initial_pagein > object->size)
initial_pagein = object->size;
for (i = 1; i < initial_pagein; i++) {
if ((ma[i] = vm_page_lookup(object, i)) != NULL) {
if ((ma[i]->flags & PG_BUSY) || ma[i]->busy)
break;
if (ma[i]->valid)
break;
vm_page_busy(ma[i]);
} else {
ma[i] = vm_page_alloc(object, i, VM_ALLOC_NORMAL);
if (ma[i] == NULL)
break;
}
}
initial_pagein = i;
rv = vm_pager_get_pages(object, ma, initial_pagein, 0);
ma[0] = vm_page_lookup(object, 0);
if ((rv != VM_PAGER_OK) || (ma[0] == NULL) || (ma[0]->valid == 0)) {
if (ma[0]) {
vm_page_protect(ma[0], VM_PROT_NONE);
vm_page_free(ma[0]);
}
mtx_unlock(&vm_mtx);
return EIO;
}
}
vm_page_wire(ma[0]);
vm_page_wakeup(ma[0]);
pmap_kenter((vm_offset_t) imgp->image_header, VM_PAGE_TO_PHYS(ma[0]));
imgp->firstpage = ma[0];
mtx_unlock(&vm_mtx);
return 0;
}
void
exec_unmap_first_page(imgp)
struct image_params *imgp;
{
if (imgp->firstpage) {
mtx_lock(&vm_mtx);
pmap_kremove((vm_offset_t) imgp->image_header);
vm_page_unwire(imgp->firstpage, 1);
mtx_unlock(&vm_mtx);
imgp->firstpage = NULL;
}
}
/*
* Destroy old address space, and allocate a new stack
* The new stack is only SGROWSIZ large because it is grown
* automatically in trap.c.
*/
int
exec_new_vmspace(imgp)
struct image_params *imgp;
{
int error;
struct vmspace *vmspace = imgp->proc->p_vmspace;
caddr_t stack_addr = (caddr_t) (USRSTACK - MAXSSIZ);
vm_map_t map = &vmspace->vm_map;
mtx_assert(&vm_mtx, MA_OWNED);
imgp->vmspace_destroyed = 1;
/*
* Blow away entire process VM, if address space not shared,
* otherwise, create a new VM space so that other threads are
* not disrupted
*/
if (vmspace->vm_refcnt == 1) {
if (vmspace->vm_shm)
shmexit(imgp->proc);
pmap_remove_pages(vmspace_pmap(vmspace), 0, VM_MAXUSER_ADDRESS);
vm_map_remove(map, 0, VM_MAXUSER_ADDRESS);
} else {
vmspace_exec(imgp->proc);
vmspace = imgp->proc->p_vmspace;
map = &vmspace->vm_map;
}
/* Allocate a new stack */
error = vm_map_stack (&vmspace->vm_map, (vm_offset_t)stack_addr,
(vm_size_t)MAXSSIZ, VM_PROT_ALL, VM_PROT_ALL, 0);
if (error)
return (error);
#ifdef __ia64__
{
/*
* Allocate backing store. We really need something
* similar to vm_map_stack which can allow the backing
* store to grow upwards. This will do for now.
*/
vm_offset_t bsaddr;
bsaddr = USRSTACK - 2*MAXSSIZ;
error = vm_map_find(&vmspace->vm_map, 0, 0, &bsaddr,
4*PAGE_SIZE, 0,
VM_PROT_ALL, VM_PROT_ALL, 0);
imgp->proc->p_md.md_bspstore = bsaddr;
}
#endif
/* vm_ssize and vm_maxsaddr are somewhat antiquated concepts in the
* VM_STACK case, but they are still used to monitor the size of the
* process stack so we can check the stack rlimit.
*/
vmspace->vm_ssize = SGROWSIZ >> PAGE_SHIFT;
vmspace->vm_maxsaddr = (char *)USRSTACK - MAXSSIZ;
return(0);
}
/*
* Copy out argument and environment strings from the old process
* address space into the temporary string buffer.
*/
int
exec_extract_strings(imgp)
struct image_params *imgp;
{
char **argv, **envv;
char *argp, *envp;
int error;
size_t length;
/*
* extract arguments first
*/
argv = imgp->uap->argv;
if (argv) {
argp = (caddr_t) (intptr_t) fuword(argv);
if (argp == (caddr_t) -1)
return (EFAULT);
if (argp)
argv++;
if (imgp->argv0)
argp = imgp->argv0;
if (argp) {
do {
if (argp == (caddr_t) -1)
return (EFAULT);
if ((error = copyinstr(argp, imgp->stringp,
imgp->stringspace, &length))) {
if (error == ENAMETOOLONG)
return(E2BIG);
return (error);
}
imgp->stringspace -= length;
imgp->stringp += length;
imgp->argc++;
} while ((argp = (caddr_t) (intptr_t) fuword(argv++)));
}
}
imgp->endargs = imgp->stringp;
/*
* extract environment strings
*/
envv = imgp->uap->envv;
if (envv) {
while ((envp = (caddr_t) (intptr_t) fuword(envv++))) {
if (envp == (caddr_t) -1)
return (EFAULT);
if ((error = copyinstr(envp, imgp->stringp,
imgp->stringspace, &length))) {
if (error == ENAMETOOLONG)
return(E2BIG);
return (error);
}
imgp->stringspace -= length;
imgp->stringp += length;
imgp->envc++;
}
}
return (0);
}
/*
* Copy strings out to the new process address space, constructing
* new arg and env vector tables. Return a pointer to the base
* so that it can be used as the initial stack pointer.
*/
register_t *
exec_copyout_strings(imgp)
struct image_params *imgp;
{
int argc, envc;
char **vectp;
char *stringp, *destp;
register_t *stack_base;
struct ps_strings *arginfo;
Mega-commit for Linux emulator update.. This has been stress tested under netscape-2.0 for Linux running all the Java stuff. The scrollbars are now working, at least on my machine. (whew! :-) I'm uncomfortable with the size of this commit, but it's too inter-dependant to easily seperate out. The main changes: COMPAT_LINUX is *GONE*. Most of the code has been moved out of the i386 machine dependent section into the linux emulator itself. The int 0x80 syscall code was almost identical to the lcall 7,0 code and a minor tweak allows them to both be used with the same C code. All kernels can now just modload the lkm and it'll DTRT without having to rebuild the kernel first. Like IBCS2, you can statically compile it in with "options LINUX". A pile of new syscalls implemented, including getdents(), llseek(), readv(), writev(), msync(), personality(). The Linux-ELF libraries want to use some of these. linux_select() now obeys Linux semantics, ie: returns the time remaining of the timeout value rather than leaving it the original value. Quite a few bugs removed, including incorrect arguments being used in syscalls.. eg: mixups between passing the sigset as an int, vs passing it as a pointer and doing a copyin(), missing return values, unhandled cases, SIOC* ioctls, etc. The build for the code has changed. i386/conf/files now knows how to build linux_genassym and generate linux_assym.h on the fly. Supporting changes elsewhere in the kernel: The user-mode signal trampoline has moved from the U area to immediately below the top of the stack (below PS_STRINGS). This allows the different binary emulations to have their own signal trampoline code (which gets rid of the hardwired syscall 103 (sigreturn on BSD, syslog on Linux)) and so that the emulator can provide the exact "struct sigcontext *" argument to the program's signal handlers. The sigstack's "ss_flags" now uses SS_DISABLE and SS_ONSTACK flags, which have the same values as the re-used SA_DISABLE and SA_ONSTACK which are intended for sigaction only. This enables the support of a SA_RESETHAND flag to sigaction to implement the gross SYSV and Linux SA_ONESHOT signal semantics where the signal handler is reset when it's triggered. makesyscalls.sh no longer appends the struct sysentvec on the end of the generated init_sysent.c code. It's a lot saner to have it in a seperate file rather than trying to update the structure inside the awk script. :-) At exec time, the dozen bytes or so of signal trampoline code are copied to the top of the user's stack, rather than obtaining the trampoline code the old way by getting a clone of the parent's user area. This allows Linux and native binaries to freely exec each other without getting trampolines mixed up.
1996-03-02 19:38:20 +00:00
int szsigcode;
/*
* Calculate string base and vector table pointers.
Mega-commit for Linux emulator update.. This has been stress tested under netscape-2.0 for Linux running all the Java stuff. The scrollbars are now working, at least on my machine. (whew! :-) I'm uncomfortable with the size of this commit, but it's too inter-dependant to easily seperate out. The main changes: COMPAT_LINUX is *GONE*. Most of the code has been moved out of the i386 machine dependent section into the linux emulator itself. The int 0x80 syscall code was almost identical to the lcall 7,0 code and a minor tweak allows them to both be used with the same C code. All kernels can now just modload the lkm and it'll DTRT without having to rebuild the kernel first. Like IBCS2, you can statically compile it in with "options LINUX". A pile of new syscalls implemented, including getdents(), llseek(), readv(), writev(), msync(), personality(). The Linux-ELF libraries want to use some of these. linux_select() now obeys Linux semantics, ie: returns the time remaining of the timeout value rather than leaving it the original value. Quite a few bugs removed, including incorrect arguments being used in syscalls.. eg: mixups between passing the sigset as an int, vs passing it as a pointer and doing a copyin(), missing return values, unhandled cases, SIOC* ioctls, etc. The build for the code has changed. i386/conf/files now knows how to build linux_genassym and generate linux_assym.h on the fly. Supporting changes elsewhere in the kernel: The user-mode signal trampoline has moved from the U area to immediately below the top of the stack (below PS_STRINGS). This allows the different binary emulations to have their own signal trampoline code (which gets rid of the hardwired syscall 103 (sigreturn on BSD, syslog on Linux)) and so that the emulator can provide the exact "struct sigcontext *" argument to the program's signal handlers. The sigstack's "ss_flags" now uses SS_DISABLE and SS_ONSTACK flags, which have the same values as the re-used SA_DISABLE and SA_ONSTACK which are intended for sigaction only. This enables the support of a SA_RESETHAND flag to sigaction to implement the gross SYSV and Linux SA_ONESHOT signal semantics where the signal handler is reset when it's triggered. makesyscalls.sh no longer appends the struct sysentvec on the end of the generated init_sysent.c code. It's a lot saner to have it in a seperate file rather than trying to update the structure inside the awk script. :-) At exec time, the dozen bytes or so of signal trampoline code are copied to the top of the user's stack, rather than obtaining the trampoline code the old way by getting a clone of the parent's user area. This allows Linux and native binaries to freely exec each other without getting trampolines mixed up.
1996-03-02 19:38:20 +00:00
* Also deal with signal trampoline code for this exec type.
*/
arginfo = (struct ps_strings *)PS_STRINGS;
Mega-commit for Linux emulator update.. This has been stress tested under netscape-2.0 for Linux running all the Java stuff. The scrollbars are now working, at least on my machine. (whew! :-) I'm uncomfortable with the size of this commit, but it's too inter-dependant to easily seperate out. The main changes: COMPAT_LINUX is *GONE*. Most of the code has been moved out of the i386 machine dependent section into the linux emulator itself. The int 0x80 syscall code was almost identical to the lcall 7,0 code and a minor tweak allows them to both be used with the same C code. All kernels can now just modload the lkm and it'll DTRT without having to rebuild the kernel first. Like IBCS2, you can statically compile it in with "options LINUX". A pile of new syscalls implemented, including getdents(), llseek(), readv(), writev(), msync(), personality(). The Linux-ELF libraries want to use some of these. linux_select() now obeys Linux semantics, ie: returns the time remaining of the timeout value rather than leaving it the original value. Quite a few bugs removed, including incorrect arguments being used in syscalls.. eg: mixups between passing the sigset as an int, vs passing it as a pointer and doing a copyin(), missing return values, unhandled cases, SIOC* ioctls, etc. The build for the code has changed. i386/conf/files now knows how to build linux_genassym and generate linux_assym.h on the fly. Supporting changes elsewhere in the kernel: The user-mode signal trampoline has moved from the U area to immediately below the top of the stack (below PS_STRINGS). This allows the different binary emulations to have their own signal trampoline code (which gets rid of the hardwired syscall 103 (sigreturn on BSD, syslog on Linux)) and so that the emulator can provide the exact "struct sigcontext *" argument to the program's signal handlers. The sigstack's "ss_flags" now uses SS_DISABLE and SS_ONSTACK flags, which have the same values as the re-used SA_DISABLE and SA_ONSTACK which are intended for sigaction only. This enables the support of a SA_RESETHAND flag to sigaction to implement the gross SYSV and Linux SA_ONESHOT signal semantics where the signal handler is reset when it's triggered. makesyscalls.sh no longer appends the struct sysentvec on the end of the generated init_sysent.c code. It's a lot saner to have it in a seperate file rather than trying to update the structure inside the awk script. :-) At exec time, the dozen bytes or so of signal trampoline code are copied to the top of the user's stack, rather than obtaining the trampoline code the old way by getting a clone of the parent's user area. This allows Linux and native binaries to freely exec each other without getting trampolines mixed up.
1996-03-02 19:38:20 +00:00
szsigcode = *(imgp->proc->p_sysent->sv_szsigcode);
destp = (caddr_t)arginfo - szsigcode - SPARE_USRSPACE -
roundup((ARG_MAX - imgp->stringspace), sizeof(char *));
Mega-commit for Linux emulator update.. This has been stress tested under netscape-2.0 for Linux running all the Java stuff. The scrollbars are now working, at least on my machine. (whew! :-) I'm uncomfortable with the size of this commit, but it's too inter-dependant to easily seperate out. The main changes: COMPAT_LINUX is *GONE*. Most of the code has been moved out of the i386 machine dependent section into the linux emulator itself. The int 0x80 syscall code was almost identical to the lcall 7,0 code and a minor tweak allows them to both be used with the same C code. All kernels can now just modload the lkm and it'll DTRT without having to rebuild the kernel first. Like IBCS2, you can statically compile it in with "options LINUX". A pile of new syscalls implemented, including getdents(), llseek(), readv(), writev(), msync(), personality(). The Linux-ELF libraries want to use some of these. linux_select() now obeys Linux semantics, ie: returns the time remaining of the timeout value rather than leaving it the original value. Quite a few bugs removed, including incorrect arguments being used in syscalls.. eg: mixups between passing the sigset as an int, vs passing it as a pointer and doing a copyin(), missing return values, unhandled cases, SIOC* ioctls, etc. The build for the code has changed. i386/conf/files now knows how to build linux_genassym and generate linux_assym.h on the fly. Supporting changes elsewhere in the kernel: The user-mode signal trampoline has moved from the U area to immediately below the top of the stack (below PS_STRINGS). This allows the different binary emulations to have their own signal trampoline code (which gets rid of the hardwired syscall 103 (sigreturn on BSD, syslog on Linux)) and so that the emulator can provide the exact "struct sigcontext *" argument to the program's signal handlers. The sigstack's "ss_flags" now uses SS_DISABLE and SS_ONSTACK flags, which have the same values as the re-used SA_DISABLE and SA_ONSTACK which are intended for sigaction only. This enables the support of a SA_RESETHAND flag to sigaction to implement the gross SYSV and Linux SA_ONESHOT signal semantics where the signal handler is reset when it's triggered. makesyscalls.sh no longer appends the struct sysentvec on the end of the generated init_sysent.c code. It's a lot saner to have it in a seperate file rather than trying to update the structure inside the awk script. :-) At exec time, the dozen bytes or so of signal trampoline code are copied to the top of the user's stack, rather than obtaining the trampoline code the old way by getting a clone of the parent's user area. This allows Linux and native binaries to freely exec each other without getting trampolines mixed up.
1996-03-02 19:38:20 +00:00
/*
* install sigcode
*/
if (szsigcode)
copyout(imgp->proc->p_sysent->sv_sigcode,
((caddr_t)arginfo - szsigcode), szsigcode);
/*
* If we have a valid auxargs ptr, prepare some room
* on the stack.
*/
if (imgp->auxargs) {
/*
* 'AT_COUNT*2' is size for the ELF Auxargs data. This is for
* lower compatibility.
*/
imgp->auxarg_size = (imgp->auxarg_size) ? imgp->auxarg_size
: (AT_COUNT * 2);
/*
* The '+ 2' is for the null pointers at the end of each of
* the arg and env vector sets,and imgp->auxarg_size is room
* for argument of Runtime loader.
*/
vectp = (char **) (destp - (imgp->argc + imgp->envc + 2 +
imgp->auxarg_size) * sizeof(char *));
} else
/*
* The '+ 2' is for the null pointers at the end of each of
* the arg and env vector sets
*/
vectp = (char **)
(destp - (imgp->argc + imgp->envc + 2) * sizeof(char *));
/*
* vectp also becomes our initial stack base
*/
stack_base = (register_t *)vectp;
stringp = imgp->stringbase;
argc = imgp->argc;
envc = imgp->envc;
/*
* Copy out strings - arguments and environment.
*/
copyout(stringp, destp, ARG_MAX - imgp->stringspace);
/*
* Fill in "ps_strings" struct for ps, w, etc.
*/
suword(&arginfo->ps_argvstr, (long)(intptr_t)vectp);
suword(&arginfo->ps_nargvstr, argc);
/*
* Fill in argument portion of vector table.
*/
for (; argc > 0; --argc) {
suword(vectp++, (long)(intptr_t)destp);
while (*stringp++ != 0)
destp++;
destp++;
}
/* a null vector table pointer separates the argp's from the envp's */
suword(vectp++, 0);
suword(&arginfo->ps_envstr, (long)(intptr_t)vectp);
suword(&arginfo->ps_nenvstr, envc);
/*
* Fill in environment portion of vector table.
*/
for (; envc > 0; --envc) {
suword(vectp++, (long)(intptr_t)destp);
while (*stringp++ != 0)
destp++;
destp++;
}
/* end of vector table is a null pointer */
suword(vectp, 0);
return (stack_base);
}
1994-05-24 10:09:53 +00:00
/*
* Check permissions of file to execute.
* Called with imgp->vp locked.
* Return 0 for success or error code on failure.
1994-05-24 10:09:53 +00:00
*/
int
exec_check_permissions(imgp)
struct image_params *imgp;
1994-05-24 10:09:53 +00:00
{
struct proc *p = imgp->proc;
struct vnode *vp = imgp->vp;
struct vattr *attr = imgp->attr;
int error;
/* Get file attributes */
error = VOP_GETATTR(vp, attr, p->p_ucred, p);
if (error)
return (error);
/*
* 1) Check if file execution is disabled for the filesystem that this
* file resides on.
* 2) Insure that at least one execute bit is on - otherwise root
* will always succeed, and we don't want to happen unless the
* file really is executable.
* 3) Insure that the file is a regular file.
*/
if ((vp->v_mount->mnt_flag & MNT_NOEXEC) ||
((attr->va_mode & 0111) == 0) ||
(attr->va_type != VREG)) {
return (EACCES);
}
1994-05-24 10:09:53 +00:00
/*
* Zero length files can't be exec'd
1994-05-24 10:09:53 +00:00
*/
if (attr->va_size == 0)
return (ENOEXEC);
/*
* Check for execute permission to file based on current credentials.
*/
error = VOP_ACCESS(vp, VEXEC, p->p_ucred, p);
if (error)
return (error);
/*
* Check number of open-for-writes on the file and deny execution
* if there are any.
*/
if (vp->v_writecount)
return (ETXTBSY);
/*
* Call filesystem specific open routine (which does nothing in the
* general case).
*/
error = VOP_OPEN(vp, FREAD, p->p_ucred, p);
if (error)
return (error);
return (0);
1994-05-24 10:09:53 +00:00
}
/*
* Exec handler registration
*/
int
exec_register(execsw_arg)
const struct execsw *execsw_arg;
{
const struct execsw **es, **xs, **newexecsw;
int count = 2; /* New slot and trailing NULL */
if (execsw)
for (es = execsw; *es; es++)
count++;
newexecsw = malloc(count * sizeof(*es), M_TEMP, M_WAITOK);
if (newexecsw == NULL)
return ENOMEM;
xs = newexecsw;
if (execsw)
for (es = execsw; *es; es++)
*xs++ = *es;
*xs++ = execsw_arg;
*xs = NULL;
if (execsw)
free(execsw, M_TEMP);
execsw = newexecsw;
return 0;
}
int
exec_unregister(execsw_arg)
const struct execsw *execsw_arg;
{
const struct execsw **es, **xs, **newexecsw;
int count = 1;
if (execsw == NULL)
panic("unregister with no handlers left?\n");
for (es = execsw; *es; es++) {
if (*es == execsw_arg)
break;
}
if (*es == NULL)
return ENOENT;
for (es = execsw; *es; es++)
if (*es != execsw_arg)
count++;
newexecsw = malloc(count * sizeof(*es), M_TEMP, M_WAITOK);
if (newexecsw == NULL)
return ENOMEM;
xs = newexecsw;
for (es = execsw; *es; es++)
if (*es != execsw_arg)
*xs++ = *es;
*xs = NULL;
if (execsw)
free(execsw, M_TEMP);
execsw = newexecsw;
return 0;
}