1
0
mirror of https://git.FreeBSD.org/src.git synced 2024-12-29 12:03:03 +00:00
freebsd/sys/i386/linux/linux_misc.c

944 lines
22 KiB
C
Raw Normal View History

/*-
* Copyright (c) 1994-1995 S<EFBFBD>ren Schmidt
* All rights reserved.
*
* 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
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
* notice, this list of conditions and the following disclaimer
* in this position and unchanged.
* 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.
* 3. The name of the author may not be used to endorse or promote products
* derived from this software withough specific prior written permission
*
* THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND 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 BE LIABLE 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: linux_misc.c,v 1.13 1996/03/02 19:37:57 peter Exp $
*/
#include <sys/param.h>
#include <sys/systm.h>
#include <sys/sysproto.h>
#include <sys/kernel.h>
#include <sys/exec.h>
#include <sys/mman.h>
#include <sys/proc.h>
#include <sys/dirent.h>
#include <sys/file.h>
#include <sys/filedesc.h>
#include <sys/ioctl.h>
#include <sys/imgact_aout.h>
#include <sys/mount.h>
#include <sys/namei.h>
#include <sys/resource.h>
#include <sys/resourcevar.h>
#include <sys/stat.h>
#include <sys/sysctl.h>
#include <sys/times.h>
#include <sys/utsname.h>
#include <sys/vnode.h>
#include <sys/wait.h>
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
#include <sys/time.h>
#include <vm/vm.h>
#include <vm/vm_param.h>
#include <vm/pmap.h>
#include <vm/lock.h>
#include <vm/vm_kern.h>
#include <vm/vm_prot.h>
#include <vm/vm_map.h>
1995-12-14 22:35:45 +00:00
#include <vm/vm_extern.h>
#include <machine/cpu.h>
#include <machine/psl.h>
#include <i386/linux/linux.h>
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
#include <i386/linux/linux_proto.h>
#include <i386/linux/linux_util.h>
int
linux_alarm(struct proc *p, struct linux_alarm_args *args, int *retval)
{
struct itimerval it, old_it;
struct timeval tv;
int s;
#ifdef DEBUG
printf("Linux-emul(%d): alarm(%d)\n", p->p_pid, args->secs);
#endif
it.it_value.tv_sec = (long)args->secs;
it.it_value.tv_usec = 0;
it.it_interval.tv_sec = 0;
it.it_interval.tv_usec = 0;
s = splclock();
old_it = p->p_realtimer;
tv = time;
if (timerisset(&old_it.it_value))
if (timercmp(&old_it.it_value, &tv, <))
timerclear(&old_it.it_value);
else
timevalsub(&old_it.it_value, &tv);
splx(s);
if (itimerfix(&it.it_value) || itimerfix(&it.it_interval))
return EINVAL;
s = splclock();
untimeout(realitexpire, (caddr_t)p);
tv = time;
if (timerisset(&it.it_value)) {
timevaladd(&it.it_value, &tv);
timeout(realitexpire, (caddr_t)p, hzto(&it.it_value));
}
p->p_realtimer = it;
splx(s);
if (old_it.it_value.tv_usec)
old_it.it_value.tv_sec++;
*retval = old_it.it_value.tv_sec;
return 0;
}
int
linux_brk(struct proc *p, struct linux_brk_args *args, int *retval)
{
#if 0
struct vmspace *vm = p->p_vmspace;
vm_offset_t new, old;
int error;
if ((vm_offset_t)args->dsend < (vm_offset_t)vm->vm_daddr)
return EINVAL;
if (((caddr_t)args->dsend - (caddr_t)vm->vm_daddr)
> p->p_rlimit[RLIMIT_DATA].rlim_cur)
return ENOMEM;
old = round_page((vm_offset_t)vm->vm_daddr) + ctob(vm->vm_dsize);
new = round_page((vm_offset_t)args->dsend);
*retval = old;
if ((new-old) > 0) {
if (swap_pager_full)
return ENOMEM;
1996-01-19 22:59:24 +00:00
error = vm_map_find(&vm->vm_map, NULL, 0, &old, (new-old), FALSE,
VM_PROT_ALL, VM_PROT_ALL, 0);
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
if (error)
return error;
vm->vm_dsize += btoc((new-old));
*retval = (int)(vm->vm_daddr + ctob(vm->vm_dsize));
}
return 0;
#else
struct vmspace *vm = p->p_vmspace;
vm_offset_t new, old;
struct obreak_args /* {
char * nsize;
} */ tmp;
#ifdef DEBUG
printf("Linux-emul(%d): brk(%08x)\n", p->p_pid, args->dsend);
#endif
old = (vm_offset_t)vm->vm_daddr + ctob(vm->vm_dsize);
new = (vm_offset_t)args->dsend;
tmp.nsize = (char *) new;
if (((caddr_t)new > vm->vm_daddr) && !obreak(p, &tmp, retval))
retval[0] = (int)new;
else
retval[0] = (int)old;
return 0;
#endif
}
int
linux_uselib(struct proc *p, struct linux_uselib_args *args, int *retval)
{
struct nameidata ni;
struct vnode *vp;
struct exec *a_out;
struct vattr attr;
unsigned long vmaddr, file_offset;
unsigned long buffer, bss_size;
char *ptr;
int error;
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
caddr_t sg;
int locked;
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
sg = stackgap_init();
CHECKALTEXIST(p, &sg, args->library);
#ifdef DEBUG
printf("Linux-emul(%d): uselib(%s)\n", p->p_pid, args->library);
#endif
a_out = NULL;
locked = 0;
vp = NULL;
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
NDINIT(&ni, LOOKUP, FOLLOW | LOCKLEAF, UIO_SYSSPACE, args->library, p);
if (error = namei(&ni))
goto cleanup;
vp = ni.ni_vp;
if (vp == NULL) {
error = ENOEXEC; /* ?? */
goto cleanup;
}
/*
* From here on down, we have a locked vnode that must be unlocked.
*/
locked++;
/*
* Writable?
*/
if (vp->v_writecount) {
error = ETXTBSY;
goto cleanup;
}
/*
* Executable?
*/
if (error = VOP_GETATTR(vp, &attr, p->p_ucred, p))
goto cleanup;
if ((vp->v_mount->mnt_flag & MNT_NOEXEC) ||
((attr.va_mode & 0111) == 0) ||
(attr.va_type != VREG)) {
error = ENOEXEC;
goto cleanup;
}
/*
* Sensible size?
*/
if (attr.va_size == 0) {
error = ENOEXEC;
goto cleanup;
}
/*
* Can we access it?
*/
if (error = VOP_ACCESS(vp, VEXEC, p->p_ucred, p))
goto cleanup;
if (error = VOP_OPEN(vp, FREAD, p->p_ucred, p))
goto cleanup;
/*
* Lock no longer needed
*/
VOP_UNLOCK(vp);
locked = 0;
/*
* Pull in executable header into kernel_map
*/
error = vm_mmap(kernel_map, (vm_offset_t *)&a_out, PAGE_SIZE,
VM_PROT_READ, VM_PROT_READ, 0, (caddr_t)vp, 0);
if (error)
goto cleanup;
/*
* Is it a Linux binary ?
*/
if (((a_out->a_magic >> 16) & 0xff) != 0x64) {
error = ENOEXEC;
goto cleanup;
}
/* While we are here, we should REALLY do some more checks */
/*
* Set file/virtual offset based on a.out variant.
*/
switch ((int)(a_out->a_magic & 0xffff)) {
case 0413: /* ZMAGIC */
file_offset = 1024;
break;
case 0314: /* QMAGIC */
file_offset = 0;
break;
default:
error = ENOEXEC;
goto cleanup;
}
bss_size = round_page(a_out->a_bss);
/*
* Check various fields in header for validity/bounds.
*/
if (a_out->a_text % NBPG || a_out->a_data % NBPG) {
error = ENOEXEC;
goto cleanup;
}
/* text + data can't exceed file size */
if (a_out->a_data + a_out->a_text > attr.va_size) {
error = EFAULT;
goto cleanup;
}
/*
* text/data/bss must not exceed limits
* XXX: this is not complete. it should check current usage PLUS
* the resources needed by this library.
*/
if (a_out->a_text > MAXTSIZ || a_out->a_data + bss_size > MAXDSIZ ||
a_out->a_data+bss_size > p->p_rlimit[RLIMIT_DATA].rlim_cur) {
error = ENOMEM;
goto cleanup;
}
/*
* prevent more writers
*/
vp->v_flag |= VTEXT;
/*
* Check if file_offset page aligned,.
* Currently we cannot handle misalinged file offsets,
* and so we read in the entire image (what a waste).
*/
if (file_offset & PGOFSET) {
#ifdef DEBUG
printf("uselib: Non page aligned binary %d\n", file_offset);
#endif
/*
* Map text+data read/write/execute
*/
/* a_entry is the load address and is page aligned */
vmaddr = trunc_page(a_out->a_entry);
/* get anon user mapping, read+write+execute */
error = vm_map_find(&p->p_vmspace->vm_map, NULL, 0, &vmaddr,
a_out->a_text + a_out->a_data, FALSE,
VM_PROT_ALL, VM_PROT_ALL, 0);
if (error)
goto cleanup;
/* map file into kernel_map */
error = vm_mmap(kernel_map, &buffer,
round_page(a_out->a_text + a_out->a_data + file_offset),
VM_PROT_READ, VM_PROT_READ, MAP_FILE,
(caddr_t)vp, trunc_page(file_offset));
if (error)
goto cleanup;
/* copy from kernel VM space to user space */
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
error = copyout((caddr_t)(buffer + file_offset), (caddr_t)vmaddr,
a_out->a_text + a_out->a_data);
/* release temporary kernel space */
vm_map_remove(kernel_map, buffer,
round_page(a_out->a_text + a_out->a_data + file_offset));
if (error)
goto cleanup;
}
else {
#ifdef DEBUG
printf("uselib: Page aligned binary %d\n", file_offset);
#endif
/*
* for QMAGIC, a_entry is 20 bytes beyond the load address
* to skip the executable header
*/
vmaddr = trunc_page(a_out->a_entry);
/*
* Map it all into the process's space as a single copy-on-write
* "data" segment.
*/
error = vm_mmap(&p->p_vmspace->vm_map, &vmaddr,
a_out->a_text + a_out->a_data,
VM_PROT_ALL, VM_PROT_ALL, MAP_PRIVATE | MAP_FIXED,
(caddr_t)vp, file_offset);
if (error)
goto cleanup;
}
#ifdef DEBUG
printf("mem=%08x = %08x %08x\n", vmaddr, ((int*)vmaddr)[0], ((int*)vmaddr)[1]);
#endif
if (bss_size != 0) {
/*
* Calculate BSS start address
*/
vmaddr = trunc_page(a_out->a_entry) + a_out->a_text + a_out->a_data;
/*
* allocate some 'anon' space
*/
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
error = vm_map_find(&p->p_vmspace->vm_map, NULL, 0, &vmaddr,
1996-01-19 22:59:24 +00:00
bss_size, FALSE,
VM_PROT_ALL, VM_PROT_ALL, 0);
if (error)
goto cleanup;
}
cleanup:
/*
* Unlock vnode if needed
*/
if (locked)
VOP_UNLOCK(vp);
/*
* Release the kernel mapping.
*/
if (a_out)
vm_map_remove(kernel_map, (vm_offset_t)a_out, PAGE_SIZE);
return error;
}
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
/* XXX move */
struct linux_select_argv {
int nfds;
fd_set *readfds;
fd_set *writefds;
fd_set *exceptfds;
struct timeval *timeout;
};
int
linux_select(struct proc *p, struct linux_select_args *args, int *retval)
{
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
struct linux_select_argv linux_args;
struct linux_newselect_args newsel;
int error;
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
#ifdef SELECT_DEBUG
printf("Linux-emul(%d): select(%x)\n",
p->p_pid, args->ptr);
#endif
if ((error = copyin((caddr_t)args->ptr, (caddr_t)&linux_args,
sizeof(linux_args))))
return error;
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
newsel.nfds = linux_args.nfds;
newsel.readfds = linux_args.readfds;
newsel.writefds = linux_args.writefds;
newsel.exceptfds = linux_args.exceptfds;
newsel.timeout = linux_args.timeout;
return linux_newselect(p, &newsel, retval);
}
int
linux_newselect(struct proc *p, struct linux_newselect_args *args, int *retval)
{
struct select_args bsa;
struct timeval tv0, tv1, utv, *tvp;
caddr_t sg;
int error;
#ifdef DEBUG
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
printf("Linux-emul(%d): newselect(%d, %x, %x, %x, %x)\n",
p->p_pid, args->nfds, args->readfds, args->writefds,
args->exceptfds, args->timeout);
#endif
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
error = 0;
bsa.nd = args->nfds;
bsa.in = args->readfds;
bsa.ou = args->writefds;
bsa.ex = args->exceptfds;
bsa.tv = args->timeout;
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
/*
* Store current time for computation of the amount of
* time left.
*/
if (args->timeout) {
if ((error = copyin(args->timeout, &utv, sizeof(utv))))
goto select_out;
#ifdef DEBUG
printf("Linux-emul(%d): incoming timeout (%d/%d)\n",
p->p_pid, utv.tv_sec, utv.tv_usec);
#endif
if (itimerfix(&utv)) {
/*
* The timeval was invalid. Convert it to something
* valid that will act as it does under Linux.
*/
sg = stackgap_init();
tvp = stackgap_alloc(&sg, sizeof(utv));
utv.tv_sec += utv.tv_usec / 1000000;
utv.tv_usec %= 1000000;
if (utv.tv_usec < 0) {
utv.tv_sec -= 1;
utv.tv_usec += 1000000;
}
if (utv.tv_sec < 0)
timerclear(&utv);
if ((error = copyout(&utv, tvp, sizeof(utv))))
goto select_out;
bsa.tv = tvp;
}
microtime(&tv0);
}
error = select(p, &bsa, retval);
#ifdef DEBUG
printf("Linux-emul(%d): real select returns %d\n",
p->p_pid, error);
#endif
if (error) {
/*
* See fs/select.c in the Linux kernel. Without this,
* Maelstrom doesn't work.
*/
if (error == ERESTART)
error = EINTR;
goto select_out;
}
if (args->timeout) {
if (*retval) {
/*
* Compute how much time was left of the timeout,
* by subtracting the current time and the time
* before we started the call, and subtracting
* that result from the user-supplied value.
*/
microtime(&tv1);
timevalsub(&tv1, &tv0);
timevalsub(&utv, &tv1);
if (utv.tv_sec < 0)
timerclear(&utv);
} else
timerclear(&utv);
#ifdef DEBUG
printf("Linux-emul(%d): outgoing timeout (%d/%d)\n",
p->p_pid, utv.tv_sec, utv.tv_usec);
#endif
if ((error = copyout(&utv, args->timeout, sizeof(utv))))
goto select_out;
}
select_out:
#ifdef DEBUG
printf("Linux-emul(%d): newselect_out -> %d\n",
p->p_pid, error);
#endif
return error;
}
int
linux_getpgid(struct proc *p, struct linux_getpgid_args *args, int *retval)
{
struct proc *curproc;
#ifdef DEBUG
printf("Linux-emul(%d): getpgid(%d)\n", p->p_pid, args->pid);
#endif
if (args->pid != p->p_pid) {
if (!(curproc = pfind(args->pid)))
return ESRCH;
}
else
curproc = p;
*retval = curproc->p_pgid;
return 0;
}
int
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
linux_fork(struct proc *p, struct linux_fork_args *args, int *retval)
{
int error;
#ifdef DEBUG
printf("Linux-emul(%d): fork()\n", p->p_pid);
#endif
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
if (error = fork(p, (struct fork_args *)args, retval))
return error;
if (retval[1] == 1)
retval[0] = 0;
return 0;
}
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
/* XXX move */
struct linux_mmap_argv {
linux_caddr_t addr;
int len;
int prot;
int flags;
int fd;
int pos;
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
linux_mmap(struct proc *p, struct linux_mmap_args *args, int *retval)
{
struct mmap_args /* {
caddr_t addr;
size_t len;
int prot;
int flags;
int fd;
long pad;
off_t pos;
} */ bsd_args;
int error;
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
struct linux_mmap_argv linux_args;
if ((error = copyin((caddr_t)args->ptr, (caddr_t)&linux_args,
sizeof(linux_args))))
return error;
#ifdef DEBUG
printf("Linux-emul(%d): mmap(%08x, %d, %d, %08x, %d, %d)\n",
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
p->p_pid, linux_args.addr, linux_args.len, linux_args.prot,
linux_args.flags, linux_args.fd, linux_args.pos);
#endif
bsd_args.flags = 0;
if (linux_args.flags & LINUX_MAP_SHARED)
bsd_args.flags |= MAP_SHARED;
if (linux_args.flags & LINUX_MAP_PRIVATE)
bsd_args.flags |= MAP_PRIVATE;
if (linux_args.flags & LINUX_MAP_FIXED)
bsd_args.flags |= MAP_FIXED;
if (linux_args.flags & LINUX_MAP_ANON)
bsd_args.flags |= MAP_ANON;
bsd_args.addr = linux_args.addr;
bsd_args.len = linux_args.len;
bsd_args.prot = linux_args.prot;
bsd_args.fd = linux_args.fd;
bsd_args.pos = linux_args.pos;
bsd_args.pad = 0;
return mmap(p, &bsd_args, retval);
}
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
linux_msync(struct proc *p, struct linux_msync_args *args, int *retval)
{
struct msync_args bsd_args;
bsd_args.addr = args->addr;
bsd_args.len = args->len;
bsd_args.flags = 0; /* XXX ignore */
return msync(p, &bsd_args, retval);
}
int
linux_pipe(struct proc *p, struct linux_pipe_args *args, int *retval)
{
int error;
#ifdef DEBUG
printf("Linux-emul(%d): pipe(*)\n", p->p_pid);
#endif
if (error = pipe(p, 0, retval))
return error;
if (error = copyout(retval, args->pipefds, 2*sizeof(int)))
return error;
*retval = 0;
return 0;
}
int
linux_time(struct proc *p, struct linux_time_args *args, int *retval)
{
struct timeval tv;
linux_time_t tm;
int error;
#ifdef DEBUG
printf("Linux-emul(%d): time(*)\n", p->p_pid);
#endif
microtime(&tv);
tm = tv.tv_sec;
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
if (args->tm && (error = copyout(&tm, args->tm, sizeof(linux_time_t))))
return error;
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
*retval = tm;
return 0;
}
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
struct linux_times_argv {
long tms_utime;
long tms_stime;
long tms_cutime;
long tms_cstime;
};
#define CLK_TCK 100 /* Linux uses 100 */
#define CONVTCK(r) (r.tv_sec * CLK_TCK + r.tv_usec / (1000000 / CLK_TCK))
int
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
linux_times(struct proc *p, struct linux_times_args *args, int *retval)
{
struct timeval tv;
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
struct linux_times_argv tms;
struct rusage ru;
int error, s;
#ifdef DEBUG
printf("Linux-emul(%d): times(*)\n", p->p_pid);
#endif
calcru(p, &ru.ru_utime, &ru.ru_stime, NULL);
tms.tms_utime = CONVTCK(ru.ru_utime);
tms.tms_stime = CONVTCK(ru.ru_stime);
tms.tms_cutime = CONVTCK(p->p_stats->p_cru.ru_utime);
tms.tms_cstime = CONVTCK(p->p_stats->p_cru.ru_stime);
if ((error = copyout((caddr_t)&tms, (caddr_t)args->buf,
sizeof(struct linux_times_argv))))
return error;
microtime(&tv);
timevalsub(&tv, &boottime);
*retval = (int)CONVTCK(tv);
return 0;
}
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
/* XXX move */
struct linux_newuname_t {
char sysname[65];
char nodename[65];
char release[65];
char version[65];
char machine[65];
char domainname[65];
};
int
linux_newuname(struct proc *p, struct linux_newuname_args *args, int *retval)
{
struct linux_newuname_t linux_newuname;
#ifdef DEBUG
printf("Linux-emul(%d): newuname(*)\n", p->p_pid);
#endif
bzero(&linux_newuname, sizeof(struct linux_newuname_args));
strncpy(linux_newuname.sysname, ostype, 64);
strncpy(linux_newuname.nodename, hostname, 64);
strncpy(linux_newuname.release, osrelease, 64);
strncpy(linux_newuname.version, version, 64);
strncpy(linux_newuname.machine, machine, 64);
strncpy(linux_newuname.domainname, domainname, 64);
return (copyout((caddr_t)&linux_newuname, (caddr_t)args->buf,
sizeof(struct linux_newuname_t)));
}
struct linux_utimbuf {
linux_time_t l_actime;
linux_time_t l_modtime;
};
int
linux_utime(struct proc *p, struct linux_utime_args *args, int *retval)
{
struct utimes_args /* {
char *path;
struct timeval *tptr;
} */ bsdutimes;
struct timeval tv[2], *tvp;
struct linux_utimbuf lut;
int error;
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
caddr_t sg;
sg = stackgap_init();
CHECKALTEXIST(p, &sg, args->fname);
#ifdef DEBUG
printf("Linux-emul(%d): utime(%s, *)\n", p->p_pid, args->fname);
#endif
if (args->times) {
if ((error = copyin(args->times, &lut, sizeof lut)))
return error;
tv[0].tv_sec = lut.l_actime;
tv[0].tv_usec = 0;
tv[1].tv_sec = lut.l_modtime;
tv[1].tv_usec = 0;
/* so that utimes can copyin */
tvp = (struct timeval *)stackgap_alloc(&sg, sizeof(tv));
if ((error = copyout(tv, tvp, sizeof(tv))))
return error;
bsdutimes.tptr = tvp;
} else
bsdutimes.tptr = NULL;
bsdutimes.path = args->fname;
return utimes(p, &bsdutimes, retval);
}
int
linux_waitpid(struct proc *p, struct linux_waitpid_args *args, int *retval)
{
struct wait_args /* {
int pid;
int *status;
int options;
struct rusage *rusage;
} */ tmp;
int error, tmpstat;
#ifdef DEBUG
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
printf("Linux-emul(%d): waitpid(%d, 0x%x, %d)\n",
p->p_pid, args->pid, args->status, args->options);
#endif
tmp.pid = args->pid;
tmp.status = args->status;
tmp.options = args->options;
tmp.rusage = NULL;
if (error = wait4(p, &tmp, retval))
return error;
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
if (args->status) {
if (error = copyin(args->status, &tmpstat, sizeof(int)))
return error;
if (WIFSIGNALED(tmpstat))
tmpstat = (tmpstat & 0xffffff80) |
bsd_to_linux_signal[WTERMSIG(tmpstat)];
else if (WIFSTOPPED(tmpstat))
tmpstat = (tmpstat & 0xffff00ff) |
(bsd_to_linux_signal[WSTOPSIG(tmpstat)]<<8);
return copyout(&tmpstat, args->status, sizeof(int));
} else
return 0;
}
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
linux_wait4(struct proc *p, struct linux_wait4_args *args, int *retval)
{
struct wait_args /* {
int pid;
int *status;
int options;
struct rusage *rusage;
} */ tmp;
int error, tmpstat;
#ifdef DEBUG
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
printf("Linux-emul(%d): wait4(%d, 0x%x, %d, 0x%x)\n",
p->p_pid, args->pid, args->status, args->options, args->rusage);
#endif
tmp.pid = args->pid;
tmp.status = args->status;
tmp.options = args->options;
tmp.rusage = args->rusage;
if (error = wait4(p, &tmp, retval))
return error;
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
p->p_siglist &= ~sigmask(SIGCHLD);
if (args->status) {
if (error = copyin(args->status, &tmpstat, sizeof(int)))
return error;
if (WIFSIGNALED(tmpstat))
tmpstat = (tmpstat & 0xffffff80) |
bsd_to_linux_signal[WTERMSIG(tmpstat)];
else if (WIFSTOPPED(tmpstat))
tmpstat = (tmpstat & 0xffff00ff) |
(bsd_to_linux_signal[WSTOPSIG(tmpstat)]<<8);
return copyout(&tmpstat, args->status, sizeof(int));
} else
return 0;
}
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
linux_mknod(struct proc *p, struct linux_mknod_args *args, int *retval)
{
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
caddr_t sg;
struct mknod_args bsd_mknod;
struct mkfifo_args bsd_mkfifo;
sg = stackgap_init();
CHECKALTCREAT(p, &sg, args->path);
#ifdef DEBUG
printf("Linux-emul(%d): mknod(%s, %d, %d)\n",
p->p_pid, args->path, args->mode, args->dev);
#endif
if (args->mode & S_IFIFO) {
bsd_mkfifo.path = args->path;
bsd_mkfifo.mode = args->mode;
return mkfifo(p, &bsd_mkfifo, retval);
} else {
bsd_mknod.path = args->path;
bsd_mknod.mode = args->mode;
bsd_mknod.dev = args->dev;
return mknod(p, &bsd_mknod, retval);
}
}
/*
* UGH! This is just about the dumbest idea I've ever heard!!
*/
int
linux_personality(struct proc *p, struct linux_personality_args *args,
int *retval)
{
#ifdef DEBUG
printf("Linux-emul(%d): personality(%d)\n",
p->p_pid, args->per);
#endif
if (args->per != 0)
return EINVAL;
/* Yes Jim, it's still a Linux... */
retval[0] = 0;
return 0;
}
/*
* Wrappers for get/setitimer for debugging..
*/
int
linux_setitimer(struct proc *p, struct linux_setitimer_args *args, int *retval)
{
struct setitimer_args bsa;
struct itimerval foo;
int error;
#ifdef DEBUG
printf("Linux-emul(%d): setitimer(%08x, %08x)\n",
p->p_pid, args->itv, args->oitv);
#endif
bsa.which = args->which;
bsa.itv = args->itv;
bsa.oitv = args->oitv;
if (args->itv) {
if ((error = copyin((caddr_t)args->itv, (caddr_t)&foo,
sizeof(foo))))
return error;
#ifdef DEBUG
printf("setitimer: value: sec: %d, usec: %d\n", foo.it_value.tv_sec, foo.it_value.tv_usec);
printf("setitimer: interval: sec: %d, usec: %d\n", foo.it_interval.tv_sec, foo.it_interval.tv_usec);
#endif
}
return setitimer(p, &bsa, retval);
}
int
linux_getitimer(struct proc *p, struct linux_getitimer_args *args, int *retval)
{
struct getitimer_args bsa;
#ifdef DEBUG
printf("Linux-emul(%d): getitimer(%08x)\n",
p->p_pid, args->itv);
#endif
bsa.which = args->which;
bsa.itv = args->itv;
return getitimer(p, &bsa, retval);
}