freebsd-nq/sys/kern/kern_exec.c

617 lines
15 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.
*
* $Id: kern_exec.c,v 1.45 1996/07/27 03:23:41 dyson Exp $
1994-05-24 10:09:53 +00:00
*/
#include <sys/param.h>
#include <sys/systm.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/malloc.h>
#include <sys/namei.h>
#include <sys/sysent.h>
#include <sys/syslog.h>
#include <sys/shm.h>
#include <sys/sysctl.h>
#include <sys/vnode.h>
#include <vm/vm.h>
#include <vm/vm_param.h>
#include <vm/vm_prot.h>
#include <vm/lock.h>
#include <vm/pmap.h>
#include <vm/vm_map.h>
#include <vm/vm_kern.h>
#include <vm/vm_extern.h>
#include <machine/reg.h>
static int *exec_copyout_strings __P((struct image_params *));
static int exec_check_permissions(struct image_params *);
/*
* XXX trouble here if sizeof(caddr_t) != sizeof(int), other parts
* of the sysctl code also assumes this, and sizeof(int) == sizeof(long).
*/
1996-06-03 04:09:36 +00:00
static struct ps_strings *ps_strings = PS_STRINGS;
SYSCTL_INT(_kern, KERN_PS_STRINGS, ps_strings, 0, &ps_strings, 0, "");
static caddr_t usrstack = (caddr_t)USRSTACK;
SYSCTL_INT(_kern, KERN_USRSTACK, usrstack, 0, &usrstack, 0, "");
/*
* execsw_set is constructed for us by the linker. Each of the items
* is a pointer to a `const struct execsw', hence the double pointer here.
*/
static const struct execsw **execsw =
(const struct execsw **)&execsw_set.ls_items[0];
#ifndef _SYS_SYSPROTO_H_
struct execve_args {
char *fname;
char **argv;
char **envv;
};
#endif
/*
* execve() system call.
*/
int
execve(p, uap, retval)
struct proc *p;
register struct execve_args *uap;
int *retval;
{
struct nameidata nd, *ndp;
int *stack_base;
int error, len, i;
struct image_params image_params, *imgp;
struct vattr attr;
imgp = &image_params;
/*
* Initialize part of the common data
*/
imgp->proc = p;
imgp->uap = uap;
imgp->attr = &attr;
imgp->image_header = NULL;
imgp->argc = imgp->envc = 0;
imgp->entry_addr = 0;
imgp->vmspace_destroyed = 0;
imgp->interpreted = 0;
imgp->interpreter_name[0] = '\0';
imgp->auxargs = NULL;
/*
* Allocate temporary demand zeroed space for argument and
* environment strings
*/
imgp->stringbase = (char *)kmem_alloc_wait(exec_map, ARG_MAX);
if (imgp->stringbase == NULL) {
error = ENOMEM;
goto exec_fail;
}
imgp->stringp = imgp->stringbase;
imgp->stringspace = 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);
1995-05-30 08:16:23 +00:00
goto exec_fail;
}
imgp->vp = ndp->ni_vp;
if (imgp->vp == NULL) {
error = ENOEXEC;
goto exec_fail_dealloc;
}
/*
* Check file permissions (also 'opens' file)
*/
error = exec_check_permissions(imgp);
/*
* Lose the lock on the vnode. It's no longer needed, and must not
* exist for the pagefault paging to work below.
*/
VOP_UNLOCK(imgp->vp);
if (error)
goto exec_fail_dealloc;
/*
* Map the image header (first page) of the file into
* kernel address space
*/
This set of commits to the VM system does the following, and contain contributions or ideas from Stephen McKay <syssgm@devetir.qld.gov.au>, Alan Cox <alc@cs.rice.edu>, David Greenman <davidg@freebsd.org> and me: More usage of the TAILQ macros. Additional minor fix to queue.h. Performance enhancements to the pageout daemon. Addition of a wait in the case that the pageout daemon has to run immediately. Slightly modify the pageout algorithm. Significant revamp of the pmap/fork code: 1) PTE's and UPAGES's are NO LONGER in the process's map. 2) PTE's and UPAGES's reside in their own objects. 3) TOTAL elimination of recursive page table pagefaults. 4) The page directory now resides in the PTE object. 5) Implemented pmap_copy, thereby speeding up fork time. 6) Changed the pv entries so that the head is a pointer and not an entire entry. 7) Significant cleanup of pmap_protect, and pmap_remove. 8) Removed significant amounts of machine dependent fork code from vm_glue. Pushed much of that code into the machine dependent pmap module. 9) Support more completely the reuse of already zeroed pages (Page table pages and page directories) as being already zeroed. Performance and code cleanups in vm_map: 1) Improved and simplified allocation of map entries. 2) Improved vm_map_copy code. 3) Corrected some minor problems in the simplify code. Implemented splvm (combo of splbio and splimp.) The VM code now seldom uses splhigh. Improved the speed of and simplified kmem_malloc. Minor mod to vm_fault to avoid using pre-zeroed pages in the case of objects with backing objects along with the already existant condition of having a vnode. (If there is a backing object, there will likely be a COW... With a COW, it isn't necessary to start with a pre-zeroed page.) Minor reorg of source to perhaps improve locality of ref.
1996-05-18 03:38:05 +00:00
error = vm_mmap(exech_map, /* map */
(vm_offset_t *)&imgp->image_header, /* address */
PAGE_SIZE, /* size */
VM_PROT_READ, /* protection */
VM_PROT_READ, /* max protection */
0, /* flags */
(caddr_t)imgp->vp, /* vnode */
0); /* offset */
if (error) {
uprintf("mmap failed: %d\n",error);
goto exec_fail_dealloc;
}
/*
* Loop through list of image activators, calling each one.
* If there is no match, the activator returns -1. If there
* is a match, but there was an error during the activation,
* the error is returned. Otherwise 0 means success. If the
* image is interpreted, loop back up and try activating
* the interpreter.
*/
for (i = 0; execsw[i]; ++i) {
if (execsw[i]->ex_imgact)
error = (*execsw[i]->ex_imgact)(imgp);
else
continue;
if (error == -1)
continue;
if (error)
goto exec_fail_dealloc;
if (imgp->interpreted) {
/* free old vnode and name buffer */
vrele(ndp->ni_vp);
FREE(ndp->ni_cnd.cn_pnbuf, M_NAMEI);
This set of commits to the VM system does the following, and contain contributions or ideas from Stephen McKay <syssgm@devetir.qld.gov.au>, Alan Cox <alc@cs.rice.edu>, David Greenman <davidg@freebsd.org> and me: More usage of the TAILQ macros. Additional minor fix to queue.h. Performance enhancements to the pageout daemon. Addition of a wait in the case that the pageout daemon has to run immediately. Slightly modify the pageout algorithm. Significant revamp of the pmap/fork code: 1) PTE's and UPAGES's are NO LONGER in the process's map. 2) PTE's and UPAGES's reside in their own objects. 3) TOTAL elimination of recursive page table pagefaults. 4) The page directory now resides in the PTE object. 5) Implemented pmap_copy, thereby speeding up fork time. 6) Changed the pv entries so that the head is a pointer and not an entire entry. 7) Significant cleanup of pmap_protect, and pmap_remove. 8) Removed significant amounts of machine dependent fork code from vm_glue. Pushed much of that code into the machine dependent pmap module. 9) Support more completely the reuse of already zeroed pages (Page table pages and page directories) as being already zeroed. Performance and code cleanups in vm_map: 1) Improved and simplified allocation of map entries. 2) Improved vm_map_copy code. 3) Corrected some minor problems in the simplify code. Implemented splvm (combo of splbio and splimp.) The VM code now seldom uses splhigh. Improved the speed of and simplified kmem_malloc. Minor mod to vm_fault to avoid using pre-zeroed pages in the case of objects with backing objects along with the already existant condition of having a vnode. (If there is a backing object, there will likely be a COW... With a COW, it isn't necessary to start with a pre-zeroed page.) Minor reorg of source to perhaps improve locality of ref.
1996-05-18 03:38:05 +00:00
if (vm_map_remove(exech_map, (vm_offset_t)imgp->image_header,
(vm_offset_t)imgp->image_header + PAGE_SIZE))
panic("execve: header dealloc failed (1)");
/* set new name to that of the interpreter */
NDINIT(ndp, LOOKUP, LOCKLEAF | FOLLOW | SAVENAME,
UIO_SYSSPACE, imgp->interpreter_name, p);
goto interpret;
}
break;
}
/* If we made it through all the activators and none matched, exit. */
if (error == -1) {
error = ENOEXEC;
goto exec_fail_dealloc;
}
/*
* 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);
/* 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 it's own resources back
*/
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. Disallow if the process is
* being traced.
*/
if ((attr.va_mode & (VSUID | VSGID)) &&
(p->p_flag & P_TRACED) == 0) {
/*
* Turn off syscall tracing for set-id programs, except for
* root.
*/
if (p->p_tracep && suser(p->p_ucred, &p->p_acflag)) {
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)
p->p_ucred->cr_uid = attr.va_uid;
if (attr.va_mode & VSGID)
p->p_ucred->cr_groups[0] = attr.va_gid;
p->p_flag |= P_SUGID;
} else {
p->p_flag &= ~P_SUGID;
}
/*
* 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;
/*
* If tracing the process, trap to debugger so breakpoints
* can be set before the program executes.
*/
if (p->p_flag & P_TRACED)
psignal(p, SIGTRAP);
/* clear "fork but no exec" flag, as we _are_ execing */
p->p_acflag &= ~AFORK;
/* Set entry address */
setregs(p, imgp->entry_addr, (u_long)stack_base);
/*
* free various allocated resources
*/
kmem_free_wakeup(exec_map, (vm_offset_t)imgp->stringbase, ARG_MAX);
This set of commits to the VM system does the following, and contain contributions or ideas from Stephen McKay <syssgm@devetir.qld.gov.au>, Alan Cox <alc@cs.rice.edu>, David Greenman <davidg@freebsd.org> and me: More usage of the TAILQ macros. Additional minor fix to queue.h. Performance enhancements to the pageout daemon. Addition of a wait in the case that the pageout daemon has to run immediately. Slightly modify the pageout algorithm. Significant revamp of the pmap/fork code: 1) PTE's and UPAGES's are NO LONGER in the process's map. 2) PTE's and UPAGES's reside in their own objects. 3) TOTAL elimination of recursive page table pagefaults. 4) The page directory now resides in the PTE object. 5) Implemented pmap_copy, thereby speeding up fork time. 6) Changed the pv entries so that the head is a pointer and not an entire entry. 7) Significant cleanup of pmap_protect, and pmap_remove. 8) Removed significant amounts of machine dependent fork code from vm_glue. Pushed much of that code into the machine dependent pmap module. 9) Support more completely the reuse of already zeroed pages (Page table pages and page directories) as being already zeroed. Performance and code cleanups in vm_map: 1) Improved and simplified allocation of map entries. 2) Improved vm_map_copy code. 3) Corrected some minor problems in the simplify code. Implemented splvm (combo of splbio and splimp.) The VM code now seldom uses splhigh. Improved the speed of and simplified kmem_malloc. Minor mod to vm_fault to avoid using pre-zeroed pages in the case of objects with backing objects along with the already existant condition of having a vnode. (If there is a backing object, there will likely be a COW... With a COW, it isn't necessary to start with a pre-zeroed page.) Minor reorg of source to perhaps improve locality of ref.
1996-05-18 03:38:05 +00:00
if (vm_map_remove(exech_map, (vm_offset_t)imgp->image_header,
(vm_offset_t)imgp->image_header + PAGE_SIZE))
panic("execve: header dealloc failed (2)");
vrele(ndp->ni_vp);
FREE(ndp->ni_cnd.cn_pnbuf, M_NAMEI);
return (0);
exec_fail_dealloc:
if (imgp->stringbase != NULL)
kmem_free_wakeup(exec_map, (vm_offset_t)imgp->stringbase, ARG_MAX);
if (imgp->image_header && imgp->image_header != (char *)-1)
This set of commits to the VM system does the following, and contain contributions or ideas from Stephen McKay <syssgm@devetir.qld.gov.au>, Alan Cox <alc@cs.rice.edu>, David Greenman <davidg@freebsd.org> and me: More usage of the TAILQ macros. Additional minor fix to queue.h. Performance enhancements to the pageout daemon. Addition of a wait in the case that the pageout daemon has to run immediately. Slightly modify the pageout algorithm. Significant revamp of the pmap/fork code: 1) PTE's and UPAGES's are NO LONGER in the process's map. 2) PTE's and UPAGES's reside in their own objects. 3) TOTAL elimination of recursive page table pagefaults. 4) The page directory now resides in the PTE object. 5) Implemented pmap_copy, thereby speeding up fork time. 6) Changed the pv entries so that the head is a pointer and not an entire entry. 7) Significant cleanup of pmap_protect, and pmap_remove. 8) Removed significant amounts of machine dependent fork code from vm_glue. Pushed much of that code into the machine dependent pmap module. 9) Support more completely the reuse of already zeroed pages (Page table pages and page directories) as being already zeroed. Performance and code cleanups in vm_map: 1) Improved and simplified allocation of map entries. 2) Improved vm_map_copy code. 3) Corrected some minor problems in the simplify code. Implemented splvm (combo of splbio and splimp.) The VM code now seldom uses splhigh. Improved the speed of and simplified kmem_malloc. Minor mod to vm_fault to avoid using pre-zeroed pages in the case of objects with backing objects along with the already existant condition of having a vnode. (If there is a backing object, there will likely be a COW... With a COW, it isn't necessary to start with a pre-zeroed page.) Minor reorg of source to perhaps improve locality of ref.
1996-05-18 03:38:05 +00:00
if (vm_map_remove(exech_map, (vm_offset_t)imgp->image_header,
(vm_offset_t)imgp->image_header + PAGE_SIZE))
panic("execve: header dealloc failed (3)");
if (ndp->ni_vp)
vrele(ndp->ni_vp);
FREE(ndp->ni_cnd.cn_pnbuf, M_NAMEI);
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);
}
}
/*
* 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 - SGROWSIZ);
imgp->vmspace_destroyed = 1;
/* Blow away entire process VM */
if (vmspace->vm_shm)
shmexit(imgp->proc);
vm_map_remove(&vmspace->vm_map, 0, USRSTACK);
/* Allocate a new stack */
error = vm_map_find(&vmspace->vm_map, NULL, 0, (vm_offset_t *)&stack_addr,
SGROWSIZ, FALSE, VM_PROT_ALL, VM_PROT_ALL, 0);
if (error)
return(error);
vmspace->vm_ssize = SGROWSIZ >> PAGE_SHIFT;
/* Initialize maximum stack address */
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, length;
/*
* extract arguments first
*/
argv = imgp->uap->argv;
if (argv) {
while ((argp = (caddr_t) fuword(argv++))) {
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++;
}
}
/*
* extract environment strings
*/
envv = imgp->uap->envv;
if (envv) {
while ((envp = (caddr_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.
*/
int *
exec_copyout_strings(imgp)
struct image_params *imgp;
{
int argc, envc;
char **vectp;
char *stringp, *destp;
int *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 = 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)
/*
* The '+ 2' is for the null pointers at the end of each of the
* arg and env vector sets, and 'AT_COUNT*2' is room for the
* ELF Auxargs data.
*/
vectp = (char **)(destp - (imgp->argc + imgp->envc + 2 +
AT_COUNT*2) * 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 = (int *)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, (int)vectp);
suword(&arginfo->ps_nargvstr, argc);
/*
* Fill in argument portion of vector table.
*/
for (; argc > 0; --argc) {
suword(vectp++, (int)destp);
while (*stringp++ != 0)
destp++;
destp++;
}
/* a null vector table pointer seperates the argp's from the envp's */
suword(vectp++, 0);
suword(&arginfo->ps_envstr, (int)vectp);
suword(&arginfo->ps_nenvstr, envc);
/*
* Fill in environment portion of vector table.
*/
for (; envc > 0; --envc) {
suword(vectp++, (int)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.
* Return 0 for success or error code on failure.
1994-05-24 10:09:53 +00:00
*/
static 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;
/*
* Check number of open-for-writes on the file and deny execution
* if there are any.
*/
if (vp->v_writecount) {
return (ETXTBSY);
}
/* 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);
/*
* Disable setuid/setgid if the filesystem prohibits it or if
* the process is being traced.
*/
if ((vp->v_mount->mnt_flag & MNT_NOSUID) || (p->p_flag & P_TRACED))
attr->va_mode &= ~(VSUID | VSGID);
/*
* Check for execute permission to file based on current credentials.
* Then call filesystem specific open routine (which does nothing
* in the general case).
*/
error = VOP_ACCESS(vp, VEXEC, p->p_ucred, p);
if (error)
return (error);
error = VOP_OPEN(vp, FREAD, p->p_ucred, p);
if (error)
return (error);
return (0);
1994-05-24 10:09:53 +00:00
}