d/freebsd-nq
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity
b49b12158f
freebsd-nq/sys/i386/linux/linux_file.c

771 lines
18 KiB
C
Raw Normal View History

First incarnation of our Linux emulator or rather compatibility code. This first shot only incorporaties so much functionality that DOOM can run (the X version), signal handling is VERY weak, so is many other things. But it meets my milestone number one (you guessed it - running DOOM). Uses /compat/linux as prefix for loading shared libs, so it won't conflict with our own libs. Kernel must be compiled with "options COMPAT_LINUX" for this to work.
1995-06-25 17:32:43 +00:00
/*-
* 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
* 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.
*
Make the long-awaited change from $Id$ to $FreeBSD$ This will make a number of things easier in the future, as well as (finally!) avoiding the Id-smashing problem which has plagued developers for so long. Boy, I'm glad we're not using sup anymore. This update would have been insane otherwise.
1997-01-14 07:20:47 +00:00
* $FreeBSD$
First incarnation of our Linux emulator or rather compatibility code. This first shot only incorporaties so much functionality that DOOM can run (the X version), signal handling is VERY weak, so is many other things. But it meets my milestone number one (you guessed it - running DOOM). Uses /compat/linux as prefix for loading shared libs, so it won't conflict with our own libs. Kernel must be compiled with "options COMPAT_LINUX" for this to work.
1995-06-25 17:32:43 +00:00
*/
#include <sys/param.h>
#include <sys/systm.h>
Completed function declarations and added prototypes. Removed some unnecessary #includes. Fixed warnings about nested externs.
1995-11-22 07:43:53 +00:00
#include <sys/sysproto.h>
First incarnation of our Linux emulator or rather compatibility code. This first shot only incorporaties so much functionality that DOOM can run (the X version), signal handling is VERY weak, so is many other things. But it meets my milestone number one (you guessed it - running DOOM). Uses /compat/linux as prefix for loading shared libs, so it won't conflict with our own libs. Kernel must be compiled with "options COMPAT_LINUX" for this to work.
1995-06-25 17:32:43 +00:00
#include <sys/fcntl.h>
#include <sys/file.h>
#include <sys/filedesc.h>
#include <sys/proc.h>
#include <sys/ioctl.h>
#include <sys/stat.h>
#include <sys/vnode.h>
#include <sys/malloc.h>
#include <sys/exec.h>
#include <sys/dirent.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/sysproto.h>
#include <sys/conf.h>
#include <sys/tty.h>
Completed function declarations and added prototypes. Removed some unnecessary #includes. Fixed warnings about nested externs.
1995-11-22 07:43:53 +00:00
First incarnation of our Linux emulator or rather compatibility code. This first shot only incorporaties so much functionality that DOOM can run (the X version), signal handling is VERY weak, so is many other things. But it meets my milestone number one (you guessed it - running DOOM). Uses /compat/linux as prefix for loading shared libs, so it won't conflict with our own libs. Kernel must be compiled with "options COMPAT_LINUX" for this to work.
1995-06-25 17:32:43 +00:00
#include <ufs/ufs/dir.h>
Completed function declarations and added prototypes. Removed some unnecessary #includes. Fixed warnings about nested externs.
1995-11-22 07:43:53 +00:00
#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>
First incarnation of our Linux emulator or rather compatibility code. This first shot only incorporaties so much functionality that DOOM can run (the X version), signal handling is VERY weak, so is many other things. But it meets my milestone number one (you guessed it - running DOOM). Uses /compat/linux as prefix for loading shared libs, so it won't conflict with our own libs. Kernel must be compiled with "options COMPAT_LINUX" for this to work.
1995-06-25 17:32:43 +00:00
int
linux_creat(struct proc *p, struct linux_creat_args *args, int *retval)
{
Clean up some warnings by using the generated structures in <sys/sysproto.h> for passing to the bsd system calls, rather than inveninting our own equivalent structures.
1995-12-15 03:06:57 +00:00
struct open_args /* {
First incarnation of our Linux emulator or rather compatibility code. This first shot only incorporaties so much functionality that DOOM can run (the X version), signal handling is VERY weak, so is many other things. But it meets my milestone number one (you guessed it - running DOOM). Uses /compat/linux as prefix for loading shared libs, so it won't conflict with our own libs. Kernel must be compiled with "options COMPAT_LINUX" for this to work.
1995-06-25 17:32:43 +00:00
char *path;
int flags;
int mode;
Clean up some warnings by using the generated structures in <sys/sysproto.h> for passing to the bsd system calls, rather than inveninting our own equivalent structures.
1995-12-15 03:06:57 +00:00
} */ bsd_open_args;
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();
CHECKALTCREAT(p, &sg, args->path);
First incarnation of our Linux emulator or rather compatibility code. This first shot only incorporaties so much functionality that DOOM can run (the X version), signal handling is VERY weak, so is many other things. But it meets my milestone number one (you guessed it - running DOOM). Uses /compat/linux as prefix for loading shared libs, so it won't conflict with our own libs. Kernel must be compiled with "options COMPAT_LINUX" for this to work.
1995-06-25 17:32:43 +00:00
#ifdef DEBUG
printf("Linux-emul(%d): creat(%s, %d)\n",
p->p_pid, args->path, args->mode);
#endif
bsd_open_args.path = args->path;
bsd_open_args.mode = args->mode;
bsd_open_args.flags = O_WRONLY | O_CREAT | O_TRUNC;
return open(p, &bsd_open_args, retval);
}
int
linux_open(struct proc *p, struct linux_open_args *args, int *retval)
{
Clean up some warnings by using the generated structures in <sys/sysproto.h> for passing to the bsd system calls, rather than inveninting our own equivalent structures.
1995-12-15 03:06:57 +00:00
struct open_args /* {
First incarnation of our Linux emulator or rather compatibility code. This first shot only incorporaties so much functionality that DOOM can run (the X version), signal handling is VERY weak, so is many other things. But it meets my milestone number one (you guessed it - running DOOM). Uses /compat/linux as prefix for loading shared libs, so it won't conflict with our own libs. Kernel must be compiled with "options COMPAT_LINUX" for this to work.
1995-06-25 17:32:43 +00:00
char *path;
int flags;
int mode;
Clean up some warnings by using the generated structures in <sys/sysproto.h> for passing to the bsd system calls, rather than inveninting our own equivalent structures.
1995-12-15 03:06:57 +00:00
} */ bsd_open_args;
First incarnation of our Linux emulator or rather compatibility code. This first shot only incorporaties so much functionality that DOOM can run (the X version), signal handling is VERY weak, so is many other things. But it meets my milestone number one (you guessed it - running DOOM). Uses /compat/linux as prefix for loading shared libs, so it won't conflict with our own libs. Kernel must be compiled with "options COMPAT_LINUX" for this to work.
1995-06-25 17:32:43 +00:00
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();
First incarnation of our Linux emulator or rather compatibility code. This first shot only incorporaties so much functionality that DOOM can run (the X version), signal handling is VERY weak, so is many other things. But it meets my milestone number one (you guessed it - running DOOM). Uses /compat/linux as prefix for loading shared libs, so it won't conflict with our own libs. Kernel must be compiled with "options COMPAT_LINUX" for this to work.
1995-06-25 17:32:43 +00:00
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->flags & LINUX_O_CREAT)
CHECKALTCREAT(p, &sg, args->path);
else
CHECKALTEXIST(p, &sg, args->path);
First incarnation of our Linux emulator or rather compatibility code. This first shot only incorporaties so much functionality that DOOM can run (the X version), signal handling is VERY weak, so is many other things. But it meets my milestone number one (you guessed it - running DOOM). Uses /compat/linux as prefix for loading shared libs, so it won't conflict with our own libs. Kernel must be compiled with "options COMPAT_LINUX" for this to work.
1995-06-25 17:32:43 +00:00
#ifdef DEBUG
printf("Linux-emul(%d): open(%s, 0x%x, 0x%x)\n",
p->p_pid, args->path, args->flags, args->mode);
#endif
bsd_open_args.flags = 0;
if (args->flags & LINUX_O_RDONLY)
bsd_open_args.flags |= O_RDONLY;
if (args->flags & LINUX_O_WRONLY)
bsd_open_args.flags |= O_WRONLY;
if (args->flags & LINUX_O_RDWR)
bsd_open_args.flags |= O_RDWR;
if (args->flags & LINUX_O_NDELAY)
bsd_open_args.flags |= O_NONBLOCK;
if (args->flags & LINUX_O_APPEND)
bsd_open_args.flags |= O_APPEND;
if (args->flags & LINUX_O_SYNC)
bsd_open_args.flags |= O_FSYNC;
if (args->flags & LINUX_O_NONBLOCK)
bsd_open_args.flags |= O_NONBLOCK;
if (args->flags & LINUX_FASYNC)
bsd_open_args.flags |= O_ASYNC;
if (args->flags & LINUX_O_CREAT)
bsd_open_args.flags |= O_CREAT;
if (args->flags & LINUX_O_TRUNC)
bsd_open_args.flags |= O_TRUNC;
if (args->flags & LINUX_O_EXCL)
bsd_open_args.flags |= O_EXCL;
if (args->flags & LINUX_O_NOCTTY)
bsd_open_args.flags |= O_NOCTTY;
bsd_open_args.path = args->path;
bsd_open_args.mode = args->mode;
error = open(p, &bsd_open_args, retval);
if (!error && !(bsd_open_args.flags & O_NOCTTY) &&
SESS_LEADER(p) && !(p->p_flag & P_CONTROLT)) {
struct filedesc *fdp = p->p_fd;
struct file *fp = fdp->fd_ofiles[*retval];
if (fp->f_type == DTYPE_VNODE)
(fp->f_ops->fo_ioctl)(fp, TIOCSCTTY, (caddr_t) 0, p);
}
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 DEBUG
printf("Linux-emul(%d): open returns error %d\n",
p->p_pid, error);
#endif
First incarnation of our Linux emulator or rather compatibility code. This first shot only incorporaties so much functionality that DOOM can run (the X version), signal handling is VERY weak, so is many other things. But it meets my milestone number one (you guessed it - running DOOM). Uses /compat/linux as prefix for loading shared libs, so it won't conflict with our own libs. Kernel must be compiled with "options COMPAT_LINUX" for this to work.
1995-06-25 17:32:43 +00:00
return error;
}
struct linux_flock {
short l_type;
short l_whence;
linux_off_t l_start;
linux_off_t l_len;
linux_pid_t l_pid;
};
static void
linux_to_bsd_flock(struct linux_flock *linux_flock, struct flock *bsd_flock)
{
switch (linux_flock->l_type) {
case LINUX_F_RDLCK:
bsd_flock->l_type = F_RDLCK;
break;
case LINUX_F_WRLCK:
bsd_flock->l_type = F_WRLCK;
break;
case LINUX_F_UNLCK:
bsd_flock->l_type = F_UNLCK;
break;
}
bsd_flock->l_whence = linux_flock->l_whence;
bsd_flock->l_start = (off_t)linux_flock->l_start;
bsd_flock->l_len = (off_t)linux_flock->l_len;
bsd_flock->l_pid = (pid_t)linux_flock->l_pid;
}
static void
bsd_to_linux_flock(struct flock *bsd_flock, struct linux_flock *linux_flock)
{
switch (bsd_flock->l_type) {
case F_RDLCK:
linux_flock->l_type = LINUX_F_RDLCK;
break;
case F_WRLCK:
linux_flock->l_type = LINUX_F_WRLCK;
break;
case F_UNLCK:
linux_flock->l_type = LINUX_F_UNLCK;
break;
}
linux_flock->l_whence = bsd_flock->l_whence;
linux_flock->l_start = (linux_off_t)bsd_flock->l_start;
linux_flock->l_len = (linux_off_t)bsd_flock->l_len;
linux_flock->l_pid = (linux_pid_t)bsd_flock->l_pid;
}
int
linux_fcntl(struct proc *p, struct linux_fcntl_args *args, int *retval)
{
int error, result;
Clean up some warnings by using the generated structures in <sys/sysproto.h> for passing to the bsd system calls, rather than inveninting our own equivalent structures.
1995-12-15 03:06:57 +00:00
struct fcntl_args /* {
First incarnation of our Linux emulator or rather compatibility code. This first shot only incorporaties so much functionality that DOOM can run (the X version), signal handling is VERY weak, so is many other things. But it meets my milestone number one (you guessed it - running DOOM). Uses /compat/linux as prefix for loading shared libs, so it won't conflict with our own libs. Kernel must be compiled with "options COMPAT_LINUX" for this to work.
1995-06-25 17:32:43 +00:00
int fd;
int cmd;
int arg;
Clean up some warnings by using the generated structures in <sys/sysproto.h> for passing to the bsd system calls, rather than inveninting our own equivalent structures.
1995-12-15 03:06:57 +00:00
} */ fcntl_args;
First incarnation of our Linux emulator or rather compatibility code. This first shot only incorporaties so much functionality that DOOM can run (the X version), signal handling is VERY weak, so is many other things. But it meets my milestone number one (you guessed it - running DOOM). Uses /compat/linux as prefix for loading shared libs, so it won't conflict with our own libs. Kernel must be compiled with "options COMPAT_LINUX" for this to work.
1995-06-25 17:32:43 +00:00
struct linux_flock linux_flock;
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 flock *bsd_flock;
struct filedesc *fdp;
struct file *fp;
struct vnode *vp;
struct vattr va;
long pgid;
struct pgrp *pgrp;
struct tty *tp, *(*d_tty) __P((dev_t));
caddr_t sg;
sg = stackgap_init();
bsd_flock = (struct flock *)stackgap_alloc(&sg, sizeof(struct flock));
d_tty = NULL;
First incarnation of our Linux emulator or rather compatibility code. This first shot only incorporaties so much functionality that DOOM can run (the X version), signal handling is VERY weak, so is many other things. But it meets my milestone number one (you guessed it - running DOOM). Uses /compat/linux as prefix for loading shared libs, so it won't conflict with our own libs. Kernel must be compiled with "options COMPAT_LINUX" for this to work.
1995-06-25 17:32:43 +00:00
#ifdef DEBUG
printf("Linux-emul(%d): fcntl(%d, %08x, *)\n",
p->p_pid, args->fd, args->cmd);
#endif
fcntl_args.fd = args->fd;
fcntl_args.arg = 0;
switch (args->cmd) {
case LINUX_F_DUPFD:
fcntl_args.cmd = F_DUPFD;
return fcntl(p, &fcntl_args, retval);
case LINUX_F_GETFD:
fcntl_args.cmd = F_GETFD;
return fcntl(p, &fcntl_args, retval);
case LINUX_F_SETFD:
fcntl_args.cmd = F_SETFD;
return fcntl(p, &fcntl_args, retval);
case LINUX_F_GETFL:
fcntl_args.cmd = F_GETFL;
error = fcntl(p, &fcntl_args, &result);
*retval = 0;
if (result & O_RDONLY) *retval |= LINUX_O_RDONLY;
if (result & O_WRONLY) *retval |= LINUX_O_WRONLY;
if (result & O_RDWR) *retval |= LINUX_O_RDWR;
if (result & O_NDELAY) *retval |= LINUX_O_NONBLOCK;
if (result & O_APPEND) *retval |= LINUX_O_APPEND;
if (result & O_FSYNC) *retval |= LINUX_O_SYNC;
return error;
case LINUX_F_SETFL:
if (args->arg & LINUX_O_NDELAY) fcntl_args.arg |= O_NONBLOCK;
if (args->arg & LINUX_O_APPEND) fcntl_args.arg |= O_APPEND;
if (args->arg & LINUX_O_SYNC) fcntl_args.arg |= O_FSYNC;
fcntl_args.cmd = F_SETFL;
return fcntl(p, &fcntl_args, retval);
case LINUX_F_GETLK:
if ((error = copyin((caddr_t)args->arg, (caddr_t)&linux_flock,
sizeof(struct linux_flock))))
return error;
linux_to_bsd_flock(&linux_flock, bsd_flock);
fcntl_args.cmd = F_GETLK;
fcntl_args.arg = (int)bsd_flock;
if (error = fcntl(p, &fcntl_args, retval))
return error;
bsd_to_linux_flock(bsd_flock, &linux_flock);
return copyout((caddr_t)&linux_flock, (caddr_t)args->arg,
sizeof(struct linux_flock));
case LINUX_F_SETLK:
if ((error = copyin((caddr_t)args->arg, (caddr_t)&linux_flock,
sizeof(struct linux_flock))))
return error;
linux_to_bsd_flock(&linux_flock, bsd_flock);
fcntl_args.cmd = F_SETLK;
fcntl_args.arg = (int)bsd_flock;
return fcntl(p, &fcntl_args, retval);
case LINUX_F_SETLKW:
if ((error = copyin((caddr_t)args->arg, (caddr_t)&linux_flock,
sizeof(struct linux_flock))))
return error;
linux_to_bsd_flock(&linux_flock, bsd_flock);
fcntl_args.cmd = F_SETLKW;
fcntl_args.arg = (int)bsd_flock;
return fcntl(p, &fcntl_args, retval);
case LINUX_F_SETOWN:
case LINUX_F_GETOWN:
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
/*
* We need to route around the normal fcntl() for these calls,
* since it uses TIOC{G,S}PGRP, which is too restrictive for
* Linux F_{G,S}ETOWN semantics. For sockets, this problem
* does not exist.
*/
fdp = p->p_fd;
if ((u_int)args->fd >= fdp->fd_nfiles ||
(fp = fdp->fd_ofiles[args->fd]) == NULL)
return EBADF;
if (fp->f_type == DTYPE_SOCKET) {
fcntl_args.cmd = args->cmd == LINUX_F_SETOWN ? F_SETOWN : F_GETOWN;
return fcntl(p, &fcntl_args, retval);
}
vp = (struct vnode *)fp->f_data;
if (vp->v_type != VCHR)
return EINVAL;
if ((error = VOP_GETATTR(vp, &va, p->p_ucred, p)))
return error;
d_tty = cdevsw[major(va.va_rdev)]->d_devtotty;
if (!d_tty || (!(tp = (*d_tty)(va.va_rdev))))
return EINVAL;
if (args->cmd == LINUX_F_GETOWN) {
retval[0] = tp->t_pgrp ? tp->t_pgrp->pg_id : NO_PID;
return 0;
}
if ((long)args->arg <= 0) {
pgid = -(long)args->arg;
} else {
struct proc *p1 = pfind((long)args->arg);
if (p1 == 0)
return (ESRCH);
pgid = (long)p1->p_pgrp->pg_id;
}
pgrp = pgfind(pgid);
if (pgrp == NULL || pgrp->pg_session != p->p_session)
return EPERM;
tp->t_pgrp = pgrp;
return 0;
First incarnation of our Linux emulator or rather compatibility code. This first shot only incorporaties so much functionality that DOOM can run (the X version), signal handling is VERY weak, so is many other things. But it meets my milestone number one (you guessed it - running DOOM). Uses /compat/linux as prefix for loading shared libs, so it won't conflict with our own libs. Kernel must be compiled with "options COMPAT_LINUX" for this to work.
1995-06-25 17:32:43 +00:00
}
return EINVAL;
}
int
linux_lseek(struct proc *p, struct linux_lseek_args *args, int *retval)
{
Clean up some warnings by using the generated structures in <sys/sysproto.h> for passing to the bsd system calls, rather than inveninting our own equivalent structures.
1995-12-15 03:06:57 +00:00
struct lseek_args /* {
int fd;
First incarnation of our Linux emulator or rather compatibility code. This first shot only incorporaties so much functionality that DOOM can run (the X version), signal handling is VERY weak, so is many other things. But it meets my milestone number one (you guessed it - running DOOM). Uses /compat/linux as prefix for loading shared libs, so it won't conflict with our own libs. Kernel must be compiled with "options COMPAT_LINUX" for this to work.
1995-06-25 17:32:43 +00:00
int pad;
Clean up some warnings by using the generated structures in <sys/sysproto.h> for passing to the bsd system calls, rather than inveninting our own equivalent structures.
1995-12-15 03:06:57 +00:00
off_t offset;
First incarnation of our Linux emulator or rather compatibility code. This first shot only incorporaties so much functionality that DOOM can run (the X version), signal handling is VERY weak, so is many other things. But it meets my milestone number one (you guessed it - running DOOM). Uses /compat/linux as prefix for loading shared libs, so it won't conflict with our own libs. Kernel must be compiled with "options COMPAT_LINUX" for this to work.
1995-06-25 17:32:43 +00:00
int whence;
Clean up some warnings by using the generated structures in <sys/sysproto.h> for passing to the bsd system calls, rather than inveninting our own equivalent structures.
1995-12-15 03:06:57 +00:00
} */ tmp_args;
First incarnation of our Linux emulator or rather compatibility code. This first shot only incorporaties so much functionality that DOOM can run (the X version), signal handling is VERY weak, so is many other things. But it meets my milestone number one (you guessed it - running DOOM). Uses /compat/linux as prefix for loading shared libs, so it won't conflict with our own libs. Kernel must be compiled with "options COMPAT_LINUX" for this to work.
1995-06-25 17:32:43 +00:00
int error;
#ifdef DEBUG
printf("Linux-emul(%d): lseek(%d, %d, %d)\n",
p->p_pid, args->fdes, args->off, args->whence);
#endif
Clean up some warnings by using the generated structures in <sys/sysproto.h> for passing to the bsd system calls, rather than inveninting our own equivalent structures.
1995-12-15 03:06:57 +00:00
tmp_args.fd = args->fdes;
tmp_args.offset = (off_t)args->off;
First incarnation of our Linux emulator or rather compatibility code. This first shot only incorporaties so much functionality that DOOM can run (the X version), signal handling is VERY weak, so is many other things. But it meets my milestone number one (you guessed it - running DOOM). Uses /compat/linux as prefix for loading shared libs, so it won't conflict with our own libs. Kernel must be compiled with "options COMPAT_LINUX" for this to work.
1995-06-25 17:32:43 +00:00
tmp_args.whence = args->whence;
Clean up some warnings by using the generated structures in <sys/sysproto.h> for passing to the bsd system calls, rather than inveninting our own equivalent structures.
1995-12-15 03:06:57 +00:00
error = lseek(p, &tmp_args, retval);
First incarnation of our Linux emulator or rather compatibility code. This first shot only incorporaties so much functionality that DOOM can run (the X version), signal handling is VERY weak, so is many other things. But it meets my milestone number one (you guessed it - running DOOM). Uses /compat/linux as prefix for loading shared libs, so it won't conflict with our own libs. Kernel must be compiled with "options COMPAT_LINUX" for this to work.
1995-06-25 17:32:43 +00:00
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
int
linux_llseek(struct proc *p, struct linux_llseek_args *args, int *retval)
{
struct lseek_args bsd_args;
int error;
off_t off;
#ifdef DEBUG
printf("Linux-emul(%d): llseek(%d, %d:%d, %d)\n",
p->p_pid, args->fd, args->ohigh, args->olow, args->whence);
#endif
off = (args->olow) | (((off_t) args->ohigh) << 32);
bsd_args.fd = args->fd;
bsd_args.offset = off;
bsd_args.whence = args->whence;
if ((error = lseek(p, &bsd_args, retval)))
return error;
if ((error = copyout(retval, (caddr_t)args->res, sizeof (off_t))))
return error;
retval[0] = 0;
return 0;
}
First incarnation of our Linux emulator or rather compatibility code. This first shot only incorporaties so much functionality that DOOM can run (the X version), signal handling is VERY weak, so is many other things. But it meets my milestone number one (you guessed it - running DOOM). Uses /compat/linux as prefix for loading shared libs, so it won't conflict with our own libs. Kernel must be compiled with "options COMPAT_LINUX" for this to work.
1995-06-25 17:32:43 +00:00
struct linux_dirent {
long dino;
linux_off_t doff;
unsigned short dreclen;
char dname[LINUX_NAME_MAX + 1];
};
#define LINUX_RECLEN(de,namlen) \
ALIGN((((char *)&(de)->dname - (char *)de) + (namlen) + 1))
int
linux_readdir(struct proc *p, struct linux_readdir_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_getdents_args lda;
lda.fd = args->fd;
lda.dent = args->dent;
lda.count = 1;
return linux_getdents(p, &lda, retval);
}
int
linux_getdents(struct proc *p, struct linux_getdents_args *args, int *retval)
First incarnation of our Linux emulator or rather compatibility code. This first shot only incorporaties so much functionality that DOOM can run (the X version), signal handling is VERY weak, so is many other things. But it meets my milestone number one (you guessed it - running DOOM). Uses /compat/linux as prefix for loading shared libs, so it won't conflict with our own libs. Kernel must be compiled with "options COMPAT_LINUX" for this to work.
1995-06-25 17:32:43 +00:00
{
register struct dirent *bdp;
struct vnode *vp;
caddr_t inp, buf; /* BSD-format */
int len, reclen; /* BSD-format */
caddr_t outp; /* Linux-format */
int resid, linuxreclen=0; /* Linux-format */
struct file *fp;
struct uio auio;
struct iovec aiov;
struct vattr va;
off_t off;
struct linux_dirent linux_dirent;
Modified linux_readdir() function to properly handle Linux readdir() calls with a byte size of 1. This special case was not correctly emulated. Now programs such as a simple 'ls' to a commercial Macintosh emulator called 'executor' will work correctly.
1995-08-28 00:50:08 +00:00
int buflen, error, eofflag, nbytes, justone, blockoff;
First incarnation of our Linux emulator or rather compatibility code. This first shot only incorporaties so much functionality that DOOM can run (the X version), signal handling is VERY weak, so is many other things. But it meets my milestone number one (you guessed it - running DOOM). Uses /compat/linux as prefix for loading shared libs, so it won't conflict with our own libs. Kernel must be compiled with "options COMPAT_LINUX" for this to work.
1995-06-25 17:32:43 +00:00
#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): getdents(%d, *, %d)\n",
First incarnation of our Linux emulator or rather compatibility code. This first shot only incorporaties so much functionality that DOOM can run (the X version), signal handling is VERY weak, so is many other things. But it meets my milestone number one (you guessed it - running DOOM). Uses /compat/linux as prefix for loading shared libs, so it won't conflict with our own libs. Kernel must be compiled with "options COMPAT_LINUX" for this to work.
1995-06-25 17:32:43 +00:00
p->p_pid, args->fd, args->count);
#endif
Modified linux_readdir() function to properly handle Linux readdir() calls with a byte size of 1. This special case was not correctly emulated. Now programs such as a simple 'ls' to a commercial Macintosh emulator called 'executor' will work correctly.
1995-08-28 00:50:08 +00:00
if ((error = getvnode(p->p_fd, args->fd, &fp)) != 0) {
First incarnation of our Linux emulator or rather compatibility code. This first shot only incorporaties so much functionality that DOOM can run (the X version), signal handling is VERY weak, so is many other things. But it meets my milestone number one (you guessed it - running DOOM). Uses /compat/linux as prefix for loading shared libs, so it won't conflict with our own libs. Kernel must be compiled with "options COMPAT_LINUX" for this to work.
1995-06-25 17:32:43 +00:00
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
}
First incarnation of our Linux emulator or rather compatibility code. This first shot only incorporaties so much functionality that DOOM can run (the X version), signal handling is VERY weak, so is many other things. But it meets my milestone number one (you guessed it - running DOOM). Uses /compat/linux as prefix for loading shared libs, so it won't conflict with our own libs. Kernel must be compiled with "options COMPAT_LINUX" for this to work.
1995-06-25 17:32:43 +00:00
if ((fp->f_flag & FREAD) == 0)
return (EBADF);
vp = (struct vnode *) fp->f_data;
if (vp->v_type != VDIR)
return (EINVAL);
Modified linux_readdir() function to properly handle Linux readdir() calls with a byte size of 1. This special case was not correctly emulated. Now programs such as a simple 'ls' to a commercial Macintosh emulator called 'executor' will work correctly.
1995-08-28 00:50:08 +00:00
if ((error = VOP_GETATTR(vp, &va, p->p_ucred, p))) {
First incarnation of our Linux emulator or rather compatibility code. This first shot only incorporaties so much functionality that DOOM can run (the X version), signal handling is VERY weak, so is many other things. But it meets my milestone number one (you guessed it - running DOOM). Uses /compat/linux as prefix for loading shared libs, so it won't conflict with our own libs. Kernel must be compiled with "options COMPAT_LINUX" for this to work.
1995-06-25 17:32:43 +00:00
return error;
Fix the getdirentries of ibcs2 to handle uneven DIRBLKSIZ offsets. Slight modification from previous fix. Also, fix problem where an entry would be skipped next call if not enough room in buffer current call.
1995-10-10 23:13:27 +00:00
}
First incarnation of our Linux emulator or rather compatibility code. This first shot only incorporaties so much functionality that DOOM can run (the X version), signal handling is VERY weak, so is many other things. But it meets my milestone number one (you guessed it - running DOOM). Uses /compat/linux as prefix for loading shared libs, so it won't conflict with our own libs. Kernel must be compiled with "options COMPAT_LINUX" for this to work.
1995-06-25 17:32:43 +00:00
nbytes = args->count;
if (nbytes == 1) {
nbytes = sizeof (struct linux_dirent);
justone = 1;
}
else
justone = 0;
Modified linux_readdir() function to properly handle Linux readdir() calls with a byte size of 1. This special case was not correctly emulated. Now programs such as a simple 'ls' to a commercial Macintosh emulator called 'executor' will work correctly.
1995-08-28 00:50:08 +00:00
off = fp->f_offset;
Fix the getdirentries of ibcs2 to handle uneven DIRBLKSIZ offsets. Slight modification from previous fix. Also, fix problem where an entry would be skipped next call if not enough room in buffer current call.
1995-10-10 23:13:27 +00:00
blockoff = off % DIRBLKSIZ;
buflen = max(DIRBLKSIZ, nbytes + blockoff);
buflen = min(buflen, MAXBSIZE);
First incarnation of our Linux emulator or rather compatibility code. This first shot only incorporaties so much functionality that DOOM can run (the X version), signal handling is VERY weak, so is many other things. But it meets my milestone number one (you guessed it - running DOOM). Uses /compat/linux as prefix for loading shared libs, so it won't conflict with our own libs. Kernel must be compiled with "options COMPAT_LINUX" for this to work.
1995-06-25 17:32:43 +00:00
buf = malloc(buflen, M_TEMP, M_WAITOK);
This is the kernel Lite/2 commit. There are some requisite userland changes, so don't expect to be able to run the kernel as-is (very well) without the appropriate Lite/2 userland changes. The system boots and can mount UFS filesystems. Untested: ext2fs, msdosfs, NFS Known problems: Incorrect Berkeley ID strings in some files. Mount_std mounts will not work until the getfsent library routine is changed. Reviewed by: various people Submitted by: Jeffery Hsu <hsu@freebsd.org>
1997-02-10 02:22:35 +00:00
vn_lock(vp, LK_EXCLUSIVE | LK_RETRY, p);
First incarnation of our Linux emulator or rather compatibility code. This first shot only incorporaties so much functionality that DOOM can run (the X version), signal handling is VERY weak, so is many other things. But it meets my milestone number one (you guessed it - running DOOM). Uses /compat/linux as prefix for loading shared libs, so it won't conflict with our own libs. Kernel must be compiled with "options COMPAT_LINUX" for this to work.
1995-06-25 17:32:43 +00:00
again:
aiov.iov_base = buf;
aiov.iov_len = buflen;
auio.uio_iov = &aiov;
auio.uio_iovcnt = 1;
auio.uio_rw = UIO_READ;
auio.uio_segflg = UIO_SYSSPACE;
auio.uio_procp = p;
auio.uio_resid = buflen;
Modified linux_readdir() function to properly handle Linux readdir() calls with a byte size of 1. This special case was not correctly emulated. Now programs such as a simple 'ls' to a commercial Macintosh emulator called 'executor' will work correctly.
1995-08-28 00:50:08 +00:00
auio.uio_offset = off - (off_t)blockoff;
First incarnation of our Linux emulator or rather compatibility code. This first shot only incorporaties so much functionality that DOOM can run (the X version), signal handling is VERY weak, so is many other things. But it meets my milestone number one (you guessed it - running DOOM). Uses /compat/linux as prefix for loading shared libs, so it won't conflict with our own libs. Kernel must be compiled with "options COMPAT_LINUX" for this to work.
1995-06-25 17:32:43 +00:00
Make this compile again after the Lite2 merge. VOP_UNLOCK was being called with the wrong mumber of arguments. Also silenced a -Wall warning.
1997-02-10 16:34:16 +00:00
error = VOP_READDIR(vp, &auio, fp->f_cred, &eofflag, NULL, NULL);
Modified linux_readdir() function to properly handle Linux readdir() calls with a byte size of 1. This special case was not correctly emulated. Now programs such as a simple 'ls' to a commercial Macintosh emulator called 'executor' will work correctly.
1995-08-28 00:50:08 +00:00
if (error) {
First incarnation of our Linux emulator or rather compatibility code. This first shot only incorporaties so much functionality that DOOM can run (the X version), signal handling is VERY weak, so is many other things. But it meets my milestone number one (you guessed it - running DOOM). Uses /compat/linux as prefix for loading shared libs, so it won't conflict with our own libs. Kernel must be compiled with "options COMPAT_LINUX" for this to work.
1995-06-25 17:32:43 +00:00
goto out;
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
}
First incarnation of our Linux emulator or rather compatibility code. This first shot only incorporaties so much functionality that DOOM can run (the X version), signal handling is VERY weak, so is many other things. But it meets my milestone number one (you guessed it - running DOOM). Uses /compat/linux as prefix for loading shared libs, so it won't conflict with our own libs. Kernel must be compiled with "options COMPAT_LINUX" for this to work.
1995-06-25 17:32:43 +00:00
inp = buf;
Modified linux_readdir() function to properly handle Linux readdir() calls with a byte size of 1. This special case was not correctly emulated. Now programs such as a simple 'ls' to a commercial Macintosh emulator called 'executor' will work correctly.
1995-08-28 00:50:08 +00:00
inp += blockoff;
First incarnation of our Linux emulator or rather compatibility code. This first shot only incorporaties so much functionality that DOOM can run (the X version), signal handling is VERY weak, so is many other things. But it meets my milestone number one (you guessed it - running DOOM). Uses /compat/linux as prefix for loading shared libs, so it won't conflict with our own libs. Kernel must be compiled with "options COMPAT_LINUX" for this to work.
1995-06-25 17:32:43 +00:00
outp = (caddr_t) args->dent;
resid = nbytes;
Modified linux_readdir() function to properly handle Linux readdir() calls with a byte size of 1. This special case was not correctly emulated. Now programs such as a simple 'ls' to a commercial Macintosh emulator called 'executor' will work correctly.
1995-08-28 00:50:08 +00:00
if ((len = buflen - auio.uio_resid - blockoff) == 0) {
First incarnation of our Linux emulator or rather compatibility code. This first shot only incorporaties so much functionality that DOOM can run (the X version), signal handling is VERY weak, so is many other things. But it meets my milestone number one (you guessed it - running DOOM). Uses /compat/linux as prefix for loading shared libs, so it won't conflict with our own libs. Kernel must be compiled with "options COMPAT_LINUX" for this to work.
1995-06-25 17:32:43 +00:00
goto eof;
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
}
First incarnation of our Linux emulator or rather compatibility code. This first shot only incorporaties so much functionality that DOOM can run (the X version), signal handling is VERY weak, so is many other things. But it meets my milestone number one (you guessed it - running DOOM). Uses /compat/linux as prefix for loading shared libs, so it won't conflict with our own libs. Kernel must be compiled with "options COMPAT_LINUX" for this to work.
1995-06-25 17:32:43 +00:00
while (len > 0) {
bdp = (struct dirent *) inp;
Modified linux_readdir() function to properly handle Linux readdir() calls with a byte size of 1. This special case was not correctly emulated. Now programs such as a simple 'ls' to a commercial Macintosh emulator called 'executor' will work correctly.
1995-08-28 00:50:08 +00:00
reclen = bdp->d_reclen;
if (reclen & 3) {
printf("linux_readdir: reclen=%d\n", reclen);
error = EFAULT;
goto out;
}
First incarnation of our Linux emulator or rather compatibility code. This first shot only incorporaties so much functionality that DOOM can run (the X version), signal handling is VERY weak, so is many other things. But it meets my milestone number one (you guessed it - running DOOM). Uses /compat/linux as prefix for loading shared libs, so it won't conflict with our own libs. Kernel must be compiled with "options COMPAT_LINUX" for this to work.
1995-06-25 17:32:43 +00:00
if (bdp->d_fileno == 0) {
inp += reclen;
Fix the getdirentries of ibcs2 to handle uneven DIRBLKSIZ offsets. Slight modification from previous fix. Also, fix problem where an entry would be skipped next call if not enough room in buffer current call.
1995-10-10 23:13:27 +00:00
off += reclen;
Modified linux_readdir() function to properly handle Linux readdir() calls with a byte size of 1. This special case was not correctly emulated. Now programs such as a simple 'ls' to a commercial Macintosh emulator called 'executor' will work correctly.
1995-08-28 00:50:08 +00:00
len -= reclen;
First incarnation of our Linux emulator or rather compatibility code. This first shot only incorporaties so much functionality that DOOM can run (the X version), signal handling is VERY weak, so is many other things. But it meets my milestone number one (you guessed it - running DOOM). Uses /compat/linux as prefix for loading shared libs, so it won't conflict with our own libs. Kernel must be compiled with "options COMPAT_LINUX" for this to work.
1995-06-25 17:32:43 +00:00
continue;
}
linuxreclen = LINUX_RECLEN(&linux_dirent, bdp->d_namlen);
if (reclen > len || resid < linuxreclen) {
outp++;
break;
}
linux_dirent.dino = (long) bdp->d_fileno;
Fix the getdents() emulation, the Linux ELF libraries use this, and this code was not quite right (linux has a readdir and getdents syscall, with the same args. readdir only returns one entry and uses a mutant dirent structure. This code was also returning the mutant form for getdents as well. My fault for missing this before.)
1996-03-10 22:27:51 +00:00
if (justone) {
/*
* old linux-style readdir usage.
*/
linux_dirent.doff = (linux_off_t) linuxreclen;
linux_dirent.dreclen = (u_short) bdp->d_namlen;
} else {
linux_dirent.doff = (linux_off_t) off;
linux_dirent.dreclen = (u_short) linuxreclen;
}
First incarnation of our Linux emulator or rather compatibility code. This first shot only incorporaties so much functionality that DOOM can run (the X version), signal handling is VERY weak, so is many other things. But it meets my milestone number one (you guessed it - running DOOM). Uses /compat/linux as prefix for loading shared libs, so it won't conflict with our own libs. Kernel must be compiled with "options COMPAT_LINUX" for this to work.
1995-06-25 17:32:43 +00:00
strcpy(linux_dirent.dname, bdp->d_name);
Modified linux_readdir() function to properly handle Linux readdir() calls with a byte size of 1. This special case was not correctly emulated. Now programs such as a simple 'ls' to a commercial Macintosh emulator called 'executor' will work correctly.
1995-08-28 00:50:08 +00:00
if ((error = copyout((caddr_t)&linux_dirent, outp, linuxreclen))) {
First incarnation of our Linux emulator or rather compatibility code. This first shot only incorporaties so much functionality that DOOM can run (the X version), signal handling is VERY weak, so is many other things. But it meets my milestone number one (you guessed it - running DOOM). Uses /compat/linux as prefix for loading shared libs, so it won't conflict with our own libs. Kernel must be compiled with "options COMPAT_LINUX" for this to work.
1995-06-25 17:32:43 +00:00
goto out;
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
}
First incarnation of our Linux emulator or rather compatibility code. This first shot only incorporaties so much functionality that DOOM can run (the X version), signal handling is VERY weak, so is many other things. But it meets my milestone number one (you guessed it - running DOOM). Uses /compat/linux as prefix for loading shared libs, so it won't conflict with our own libs. Kernel must be compiled with "options COMPAT_LINUX" for this to work.
1995-06-25 17:32:43 +00:00
inp += reclen;
Fix the getdirentries of ibcs2 to handle uneven DIRBLKSIZ offsets. Slight modification from previous fix. Also, fix problem where an entry would be skipped next call if not enough room in buffer current call.
1995-10-10 23:13:27 +00:00
off += reclen;
First incarnation of our Linux emulator or rather compatibility code. This first shot only incorporaties so much functionality that DOOM can run (the X version), signal handling is VERY weak, so is many other things. But it meets my milestone number one (you guessed it - running DOOM). Uses /compat/linux as prefix for loading shared libs, so it won't conflict with our own libs. Kernel must be compiled with "options COMPAT_LINUX" for this to work.
1995-06-25 17:32:43 +00:00
outp += linuxreclen;
resid -= linuxreclen;
len -= reclen;
if (justone)
break;
}
if (outp == (caddr_t) args->dent)
goto again;
fp->f_offset = off;
if (justone)
nbytes = resid + linuxreclen;
Modified linux_readdir() function to properly handle Linux readdir() calls with a byte size of 1. This special case was not correctly emulated. Now programs such as a simple 'ls' to a commercial Macintosh emulator called 'executor' will work correctly.
1995-08-28 00:50:08 +00:00
First incarnation of our Linux emulator or rather compatibility code. This first shot only incorporaties so much functionality that DOOM can run (the X version), signal handling is VERY weak, so is many other things. But it meets my milestone number one (you guessed it - running DOOM). Uses /compat/linux as prefix for loading shared libs, so it won't conflict with our own libs. Kernel must be compiled with "options COMPAT_LINUX" for this to work.
1995-06-25 17:32:43 +00:00
eof:
*retval = nbytes - resid;
out:
Make this compile again after the Lite2 merge. VOP_UNLOCK was being called with the wrong mumber of arguments. Also silenced a -Wall warning.
1997-02-10 16:34:16 +00:00
VOP_UNLOCK(vp, 0, p);
First incarnation of our Linux emulator or rather compatibility code. This first shot only incorporaties so much functionality that DOOM can run (the X version), signal handling is VERY weak, so is many other things. But it meets my milestone number one (you guessed it - running DOOM). Uses /compat/linux as prefix for loading shared libs, so it won't conflict with our own libs. Kernel must be compiled with "options COMPAT_LINUX" for this to work.
1995-06-25 17:32:43 +00:00
free(buf, M_TEMP);
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
/*
* These exist mainly for hooks for doing /compat/linux translation.
*/
int
linux_access(struct proc *p, struct linux_access_args *args, int *retval)
{
struct access_args bsd;
caddr_t sg;
sg = stackgap_init();
CHECKALTEXIST(p, &sg, args->path);
#ifdef DEBUG
printf("Linux-emul(%d): access(%s, %d)\n",
p->p_pid, args->path, args->flags);
#endif
bsd.path = args->path;
bsd.flags = args->flags;
return access(p, &bsd, retval);
}
int
linux_unlink(struct proc *p, struct linux_unlink_args *args, int *retval)
{
struct unlink_args bsd;
caddr_t sg;
sg = stackgap_init();
CHECKALTEXIST(p, &sg, args->path);
#ifdef DEBUG
printf("Linux-emul(%d): unlink(%s)\n",
p->p_pid, args->path);
#endif
bsd.path = args->path;
return unlink(p, &bsd, retval);
}
int
linux_chdir(struct proc *p, struct linux_chdir_args *args, int *retval)
{
struct chdir_args bsd;
caddr_t sg;
sg = stackgap_init();
CHECKALTEXIST(p, &sg, args->path);
#ifdef DEBUG
printf("Linux-emul(%d): chdir(%s)\n",
p->p_pid, args->path);
#endif
bsd.path = args->path;
return chdir(p, &bsd, retval);
}
int
linux_chmod(struct proc *p, struct linux_chmod_args *args, int *retval)
{
struct chmod_args bsd;
caddr_t sg;
sg = stackgap_init();
CHECKALTEXIST(p, &sg, args->path);
#ifdef DEBUG
printf("Linux-emul(%d): chmod(%s, %d)\n",
p->p_pid, args->path, args->mode);
#endif
bsd.path = args->path;
bsd.mode = args->mode;
return chmod(p, &bsd, retval);
}
int
linux_chown(struct proc *p, struct linux_chown_args *args, int *retval)
{
struct chown_args bsd;
caddr_t sg;
sg = stackgap_init();
CHECKALTEXIST(p, &sg, args->path);
#ifdef DEBUG
printf("Linux-emul(%d): chown(%s, %d, %d)\n",
p->p_pid, args->path, args->uid, args->gid);
#endif
bsd.path = args->path;
/* XXX size casts here */
bsd.uid = args->uid;
bsd.gid = args->gid;
return chown(p, &bsd, retval);
}
int
linux_mkdir(struct proc *p, struct linux_mkdir_args *args, int *retval)
{
struct mkdir_args bsd;
caddr_t sg;
sg = stackgap_init();
CHECKALTCREAT(p, &sg, args->path);
#ifdef DEBUG
printf("Linux-emul(%d): mkdir(%s, %d)\n",
p->p_pid, args->path, args->mode);
#endif
bsd.path = args->path;
bsd.mode = args->mode;
return mkdir(p, &bsd, retval);
}
int
linux_rmdir(struct proc *p, struct linux_rmdir_args *args, int *retval)
{
struct rmdir_args bsd;
caddr_t sg;
sg = stackgap_init();
CHECKALTEXIST(p, &sg, args->path);
#ifdef DEBUG
printf("Linux-emul(%d): rmdir(%s)\n",
p->p_pid, args->path);
#endif
bsd.path = args->path;
return rmdir(p, &bsd, retval);
}
int
linux_rename(struct proc *p, struct linux_rename_args *args, int *retval)
{
struct rename_args bsd;
caddr_t sg;
sg = stackgap_init();
CHECKALTEXIST(p, &sg, args->from);
CHECKALTCREAT(p, &sg, args->to);
#ifdef DEBUG
printf("Linux-emul(%d): rename(%s, %s)\n",
p->p_pid, args->from, args->to);
#endif
bsd.from = args->from;
bsd.to = args->to;
return rename(p, &bsd, retval);
}
int
linux_symlink(struct proc *p, struct linux_symlink_args *args, int *retval)
{
struct symlink_args bsd;
caddr_t sg;
sg = stackgap_init();
CHECKALTEXIST(p, &sg, args->path);
CHECKALTCREAT(p, &sg, args->to);
#ifdef DEBUG
printf("Linux-emul(%d): symlink(%s, %s)\n",
p->p_pid, args->path, args->to);
#endif
bsd.path = args->path;
bsd.link = args->to;
return symlink(p, &bsd, retval);
}
int
linux_execve(struct proc *p, struct linux_execve_args *args, int *retval)
{
struct execve_args bsd;
caddr_t sg;
sg = stackgap_init();
CHECKALTEXIST(p, &sg, args->path);
#ifdef DEBUG
printf("Linux-emul(%d): execve(%s)\n",
p->p_pid, args->path);
#endif
bsd.fname = args->path;
bsd.argv = args->argp;
bsd.envv = args->envp;
return execve(p, &bsd, retval);
}
int
linux_readlink(struct proc *p, struct linux_readlink_args *args, int *retval)
{
struct readlink_args bsd;
caddr_t sg;
sg = stackgap_init();
CHECKALTEXIST(p, &sg, args->name);
#ifdef DEBUG
printf("Linux-emul(%d): readlink(%s, 0x%x, %d)\n",
p->p_pid, args->name, args->buf, args->count);
#endif
bsd.path = args->name;
bsd.buf = args->buf;
bsd.count = args->count;
return readlink(p, &bsd, retval);
}
int
linux_truncate(struct proc *p, struct linux_truncate_args *args, int *retval)
{
struct otruncate_args bsd;
caddr_t sg;
sg = stackgap_init();
CHECKALTEXIST(p, &sg, args->path);
#ifdef DEBUG
printf("Linux-emul(%d): truncate(%s)\n",
p->p_pid, args->path);
#endif
bsd.path = args->path;
return otruncate(p, &bsd, retval);
}
Reference in New Issue Copy Permalink