freebsd-dev/sys/i386/linux/linux_sysvec.c

1181 lines
34 KiB
C
Raw Normal View History

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
/*-
2012-01-15 13:23:18 +00:00
* Copyright (c) 1994-1996 Søren Schmidt
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
* 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
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
* 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 without specific prior written permission
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
*
* 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.
*/
2003-06-02 16:56:40 +00:00
#include <sys/cdefs.h>
__FBSDID("$FreeBSD$");
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/param.h>
#include <sys/systm.h>
#include <sys/exec.h>
#include <sys/fcntl.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/imgact.h>
#include <sys/imgact_aout.h>
#include <sys/imgact_elf.h>
#include <sys/kernel.h>
2001-10-11 17:52:20 +00:00
#include <sys/lock.h>
#include <sys/malloc.h>
#include <sys/module.h>
#include <sys/mutex.h>
#include <sys/proc.h>
#include <sys/signalvar.h>
#include <sys/syscallsubr.h>
#include <sys/sysent.h>
#include <sys/sysproto.h>
#include <sys/vnode.h>
#include <sys/eventhandler.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 <vm/vm.h>
#include <vm/pmap.h>
#include <vm/vm_extern.h>
#include <vm/vm_map.h>
#include <vm/vm_object.h>
#include <vm/vm_page.h>
#include <vm/vm_param.h>
#include <machine/cpu.h>
#include <machine/cputypes.h>
#include <machine/md_var.h>
#include <machine/pcb.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.h>
#include <i386/linux/linux_proto.h>
#include <compat/linux/linux_emul.h>
#include <compat/linux/linux_futex.h>
#include <compat/linux/linux_ioctl.h>
#include <compat/linux/linux_mib.h>
#include <compat/linux/linux_misc.h>
#include <compat/linux/linux_signal.h>
#include <compat/linux/linux_util.h>
MODULE_VERSION(linux, 1);
MALLOC_DEFINE(M_LINUX, "linux", "Linux mode structures");
#if BYTE_ORDER == LITTLE_ENDIAN
#define SHELLMAGIC 0x2123 /* #! */
#else
#define SHELLMAGIC 0x2321
#endif
/*
* Allow the sendsig functions to use the ldebug() facility
* even though they are not syscalls themselves. Map them
* to syscall 0. This is slightly less bogus than using
* ldebug(sigreturn).
*/
#define LINUX_SYS_linux_rt_sendsig 0
#define LINUX_SYS_linux_sendsig 0
#define LINUX_PS_STRINGS (LINUX_USRSTACK - sizeof(struct ps_strings))
extern char linux_sigcode[];
extern int linux_szsigcode;
extern struct sysent linux_sysent[LINUX_SYS_MAXSYSCALL];
SET_DECLARE(linux_ioctl_handler_set, struct linux_ioctl_handler);
SET_DECLARE(linux_device_handler_set, struct linux_device_handler);
2002-03-20 07:51:46 +00:00
static int linux_fixup(register_t **stack_base,
struct image_params *iparams);
2002-03-20 07:51:46 +00:00
static int elf_linux_fixup(register_t **stack_base,
struct image_params *iparams);
1. Change prototype of trapsignal and sendsig to use ksiginfo_t *, most changes in MD code are trivial, before this change, trapsignal and sendsig use discrete parameters, now they uses member fields of ksiginfo_t structure. For sendsig, this change allows us to pass POSIX realtime signal value to user code. 2. Remove cpu_thread_siginfo, it is no longer needed because we now always generate ksiginfo_t data and feed it to libpthread. 3. Add p_sigqueue to proc structure to hold shared signals which were blocked by all threads in the proc. 4. Add td_sigqueue to thread structure to hold all signals delivered to thread. 5. i386 and amd64 now return POSIX standard si_code, other arches will be fixed. 6. In this sigqueue implementation, pending signal set is kept as before, an extra siginfo list holds additional siginfo_t data for signals. kernel code uses psignal() still behavior as before, it won't be failed even under memory pressure, only exception is when deleting a signal, we should call sigqueue_delete to remove signal from sigqueue but not SIGDELSET. Current there is no kernel code will deliver a signal with additional data, so kernel should be as stable as before, a ksiginfo can carry more information, for example, allow signal to be delivered but throw away siginfo data if memory is not enough. SIGKILL and SIGSTOP have fast path in sigqueue_add, because they can not be caught or masked. The sigqueue() syscall allows user code to queue a signal to target process, if resource is unavailable, EAGAIN will be returned as specification said. Just before thread exits, signal queue memory will be freed by sigqueue_flush. Current, all signals are allowed to be queued, not only realtime signals. Earlier patch reviewed by: jhb, deischen Tested on: i386, amd64
2005-10-14 12:43:47 +00:00
static void linux_sendsig(sig_t catcher, ksiginfo_t *ksi, sigset_t *mask);
static void exec_linux_setregs(struct thread *td,
struct image_params *imgp, u_long stack);
static register_t *linux_copyout_strings(struct image_params *imgp);
static boolean_t linux_trans_osrel(const Elf_Note *note, int32_t *osrel);
static int linux_szplatform;
const char *linux_platform;
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
static eventhandler_tag linux_exit_tag;
static eventhandler_tag linux_exec_tag;
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 syscalls return negative errno's, we do positive and map them
* Reference:
* FreeBSD: src/sys/sys/errno.h
* Linux: linux-2.6.17.8/include/asm-generic/errno-base.h
* linux-2.6.17.8/include/asm-generic/errno.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
*/
static int bsd_to_linux_errno[ELAST + 1] = {
-0, -1, -2, -3, -4, -5, -6, -7, -8, -9,
-10, -35, -12, -13, -14, -15, -16, -17, -18, -19,
-20, -21, -22, -23, -24, -25, -26, -27, -28, -29,
-30, -31, -32, -33, -34, -11,-115,-114, -88, -89,
-90, -91, -92, -93, -94, -95, -96, -97, -98, -99,
-100,-101,-102,-103,-104,-105,-106,-107,-108,-109,
-110,-111, -40, -36,-112,-113, -39, -11, -87,-122,
-116, -66, -6, -6, -6, -6, -6, -37, -38, -9,
-6, -6, -43, -42, -75,-125, -84, -95, -16, -74,
-72, -67, -71
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 bsd_to_linux_signal[LINUX_SIGTBLSZ] = {
LINUX_SIGHUP, LINUX_SIGINT, LINUX_SIGQUIT, LINUX_SIGILL,
LINUX_SIGTRAP, LINUX_SIGABRT, 0, LINUX_SIGFPE,
LINUX_SIGKILL, LINUX_SIGBUS, LINUX_SIGSEGV, LINUX_SIGSYS,
LINUX_SIGPIPE, LINUX_SIGALRM, LINUX_SIGTERM, LINUX_SIGURG,
LINUX_SIGSTOP, LINUX_SIGTSTP, LINUX_SIGCONT, LINUX_SIGCHLD,
LINUX_SIGTTIN, LINUX_SIGTTOU, LINUX_SIGIO, LINUX_SIGXCPU,
LINUX_SIGXFSZ, LINUX_SIGVTALRM, LINUX_SIGPROF, LINUX_SIGWINCH,
0, LINUX_SIGUSR1, LINUX_SIGUSR2
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_to_bsd_signal[LINUX_SIGTBLSZ] = {
SIGHUP, SIGINT, SIGQUIT, SIGILL,
SIGTRAP, SIGABRT, SIGBUS, SIGFPE,
SIGKILL, SIGUSR1, SIGSEGV, SIGUSR2,
SIGPIPE, SIGALRM, SIGTERM, SIGBUS,
SIGCHLD, SIGCONT, SIGSTOP, SIGTSTP,
SIGTTIN, SIGTTOU, SIGURG, SIGXCPU,
SIGXFSZ, SIGVTALRM, SIGPROF, SIGWINCH,
SIGIO, SIGURG, SIGSYS
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
};
#define LINUX_T_UNKNOWN 255
static int _bsd_to_linux_trapcode[] = {
LINUX_T_UNKNOWN, /* 0 */
6, /* 1 T_PRIVINFLT */
LINUX_T_UNKNOWN, /* 2 */
3, /* 3 T_BPTFLT */
LINUX_T_UNKNOWN, /* 4 */
LINUX_T_UNKNOWN, /* 5 */
16, /* 6 T_ARITHTRAP */
254, /* 7 T_ASTFLT */
LINUX_T_UNKNOWN, /* 8 */
13, /* 9 T_PROTFLT */
1, /* 10 T_TRCTRAP */
LINUX_T_UNKNOWN, /* 11 */
14, /* 12 T_PAGEFLT */
LINUX_T_UNKNOWN, /* 13 */
17, /* 14 T_ALIGNFLT */
LINUX_T_UNKNOWN, /* 15 */
LINUX_T_UNKNOWN, /* 16 */
LINUX_T_UNKNOWN, /* 17 */
0, /* 18 T_DIVIDE */
2, /* 19 T_NMI */
4, /* 20 T_OFLOW */
5, /* 21 T_BOUND */
7, /* 22 T_DNA */
8, /* 23 T_DOUBLEFLT */
9, /* 24 T_FPOPFLT */
10, /* 25 T_TSSFLT */
11, /* 26 T_SEGNPFLT */
12, /* 27 T_STKFLT */
18, /* 28 T_MCHK */
19, /* 29 T_XMMFLT */
15 /* 30 T_RESERVED */
};
#define bsd_to_linux_trapcode(code) \
((code)<sizeof(_bsd_to_linux_trapcode)/sizeof(*_bsd_to_linux_trapcode)? \
_bsd_to_linux_trapcode[(code)]: \
LINUX_T_UNKNOWN)
/*
* If FreeBSD & Linux have a difference of opinion about what a trap
* means, deal with it here.
*
* MPSAFE
*/
static int
translate_traps(int signal, int trap_code)
{
if (signal != SIGBUS)
return signal;
switch (trap_code) {
case T_PROTFLT:
case T_TSSFLT:
case T_DOUBLEFLT:
case T_PAGEFLT:
return SIGSEGV;
default:
return signal;
}
}
1998-02-09 06:11:36 +00:00
static int
linux_fixup(register_t **stack_base, struct image_params *imgp)
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
{
register_t *argv, *envp;
argv = *stack_base;
envp = *stack_base + (imgp->args->argc + 1);
(*stack_base)--;
suword(*stack_base, (intptr_t)(void *)envp);
(*stack_base)--;
suword(*stack_base, (intptr_t)(void *)argv);
(*stack_base)--;
suword(*stack_base, imgp->args->argc);
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
}
1998-02-09 06:11:36 +00:00
static int
elf_linux_fixup(register_t **stack_base, struct image_params *imgp)
{
struct proc *p;
Elf32_Auxargs *args;
Elf32_Addr *uplatform;
struct ps_strings *arginfo;
register_t *pos;
KASSERT(curthread->td_proc == imgp->proc,
("unsafe elf_linux_fixup(), should be curproc"));
p = imgp->proc;
arginfo = (struct ps_strings *)p->p_sysent->sv_psstrings;
uplatform = (Elf32_Addr *)((caddr_t)arginfo - linux_szplatform);
args = (Elf32_Auxargs *)imgp->auxargs;
pos = *stack_base + (imgp->args->argc + imgp->args->envc + 2);
AUXARGS_ENTRY(pos, LINUX_AT_HWCAP, cpu_feature);
/*
* Do not export AT_CLKTCK when emulating Linux kernel prior to 2.4.0,
* as it has appeared in the 2.4.0-rc7 first time.
* Being exported, AT_CLKTCK is returned by sysconf(_SC_CLK_TCK),
* glibc falls back to the hard-coded CLK_TCK value when aux entry
* is not present.
* Also see linux_times() implementation.
*/
if (linux_kernver(curthread) >= LINUX_KERNVER_2004000)
AUXARGS_ENTRY(pos, LINUX_AT_CLKTCK, stclohz);
AUXARGS_ENTRY(pos, AT_PHDR, args->phdr);
AUXARGS_ENTRY(pos, AT_PHENT, args->phent);
AUXARGS_ENTRY(pos, AT_PHNUM, args->phnum);
AUXARGS_ENTRY(pos, AT_PAGESZ, args->pagesz);
AUXARGS_ENTRY(pos, AT_FLAGS, args->flags);
AUXARGS_ENTRY(pos, AT_ENTRY, args->entry);
AUXARGS_ENTRY(pos, AT_BASE, args->base);
AUXARGS_ENTRY(pos, LINUX_AT_SECURE, 0);
o Merge contents of struct pcred into struct ucred. Specifically, add the real uid, saved uid, real gid, and saved gid to ucred, as well as the pcred->pc_uidinfo, which was associated with the real uid, only rename it to cr_ruidinfo so as not to conflict with cr_uidinfo, which corresponds to the effective uid. o Remove p_cred from struct proc; add p_ucred to struct proc, replacing original macro that pointed. p->p_ucred to p->p_cred->pc_ucred. o Universally update code so that it makes use of ucred instead of pcred, p->p_ucred instead of p->p_pcred, cr_ruidinfo instead of p_uidinfo, cr_{r,sv}{u,g}id instead of p_*, etc. o Remove pcred0 and its initialization from init_main.c; initialize cr_ruidinfo there. o Restruction many credential modification chunks to always crdup while we figure out locking and optimizations; generally speaking, this means moving to a structure like this: newcred = crdup(oldcred); ... p->p_ucred = newcred; crfree(oldcred); It's not race-free, but better than nothing. There are also races in sys_process.c, all inter-process authorization, fork, exec, and exit. o Remove sigio->sio_ruid since sigio->sio_ucred now contains the ruid; remove comments indicating that the old arrangement was a problem. o Restructure exec1() a little to use newcred/oldcred arrangement, and use improved uid management primitives. o Clean up exit1() so as to do less work in credential cleanup due to pcred removal. o Clean up fork1() so as to do less work in credential cleanup and allocation. o Clean up ktrcanset() to take into account changes, and move to using suser_xxx() instead of performing a direct uid==0 comparision. o Improve commenting in various kern_prot.c credential modification calls to better document current behavior. In a couple of places, current behavior is a little questionable and we need to check POSIX.1 to make sure it's "right". More commenting work still remains to be done. o Update credential management calls, such as crfree(), to take into account new ruidinfo reference. o Modify or add the following uid and gid helper routines: change_euid() change_egid() change_ruid() change_rgid() change_svuid() change_svgid() In each case, the call now acts on a credential not a process, and as such no longer requires more complicated process locking/etc. They now assume the caller will do any necessary allocation of an exclusive credential reference. Each is commented to document its reference requirements. o CANSIGIO() is simplified to require only credentials, not processes and pcreds. o Remove lots of (p_pcred==NULL) checks. o Add an XXX to authorization code in nfs_lock.c, since it's questionable, and needs to be considered carefully. o Simplify posix4 authorization code to require only credentials, not processes and pcreds. Note that this authorization, as well as CANSIGIO(), needs to be updated to use the p_cansignal() and p_cansched() centralized authorization routines, as they currently do not take into account some desirable restrictions that are handled by the centralized routines, as well as being inconsistent with other similar authorization instances. o Update libkvm to take these changes into account. Obtained from: TrustedBSD Project Reviewed by: green, bde, jhb, freebsd-arch, freebsd-audit
2001-05-25 16:59:11 +00:00
AUXARGS_ENTRY(pos, AT_UID, imgp->proc->p_ucred->cr_ruid);
AUXARGS_ENTRY(pos, AT_EUID, imgp->proc->p_ucred->cr_svuid);
AUXARGS_ENTRY(pos, AT_GID, imgp->proc->p_ucred->cr_rgid);
AUXARGS_ENTRY(pos, AT_EGID, imgp->proc->p_ucred->cr_svgid);
AUXARGS_ENTRY(pos, LINUX_AT_PLATFORM, PTROUT(uplatform));
if (args->execfd != -1)
AUXARGS_ENTRY(pos, AT_EXECFD, args->execfd);
AUXARGS_ENTRY(pos, AT_NULL, 0);
free(imgp->auxargs, M_TEMP);
imgp->auxargs = 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
(*stack_base)--;
suword(*stack_base, (register_t)imgp->args->argc);
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
/*
* Copied from kern/kern_exec.c
*/
static register_t *
linux_copyout_strings(struct image_params *imgp)
{
int argc, envc;
char **vectp;
char *stringp, *destp;
register_t *stack_base;
struct ps_strings *arginfo;
struct proc *p;
/*
* Calculate string base and vector table pointers.
* Also deal with signal trampoline code for this exec type.
*/
p = imgp->proc;
arginfo = (struct ps_strings *)p->p_sysent->sv_psstrings;
destp = (caddr_t)arginfo - SPARE_USRSPACE - linux_szplatform -
roundup((ARG_MAX - imgp->args->stringspace), sizeof(char *));
/*
* install LINUX_PLATFORM
*/
copyout(linux_platform, ((caddr_t)arginfo - linux_szplatform),
linux_szplatform);
/*
* If we have a valid auxargs ptr, prepare some room
* on the stack.
*/
if (imgp->auxargs) {
/*
* 'AT_COUNT*2' is size for the ELF Auxargs data. This is for
* lower compatibility.
*/
imgp->auxarg_size = (imgp->auxarg_size) ? imgp->auxarg_size :
(LINUX_AT_COUNT * 2);
/*
* The '+ 2' is for the null pointers at the end of each of
* the arg and env vector sets,and imgp->auxarg_size is room
* for argument of Runtime loader.
*/
vectp = (char **)(destp - (imgp->args->argc +
imgp->args->envc + 2 + imgp->auxarg_size) * sizeof(char *));
} else {
/*
* The '+ 2' is for the null pointers at the end of each of
* the arg and env vector sets
*/
vectp = (char **)(destp - (imgp->args->argc + imgp->args->envc + 2) *
sizeof(char *));
}
/*
* vectp also becomes our initial stack base
*/
stack_base = (register_t *)vectp;
stringp = imgp->args->begin_argv;
argc = imgp->args->argc;
envc = imgp->args->envc;
/*
* Copy out strings - arguments and environment.
*/
copyout(stringp, destp, ARG_MAX - imgp->args->stringspace);
/*
* Fill in "ps_strings" struct for ps, w, etc.
*/
suword(&arginfo->ps_argvstr, (long)(intptr_t)vectp);
suword(&arginfo->ps_nargvstr, argc);
/*
* Fill in argument portion of vector table.
*/
for (; argc > 0; --argc) {
suword(vectp++, (long)(intptr_t)destp);
while (*stringp++ != 0)
destp++;
destp++;
}
/* a null vector table pointer separates the argp's from the envp's */
suword(vectp++, 0);
suword(&arginfo->ps_envstr, (long)(intptr_t)vectp);
suword(&arginfo->ps_nenvstr, envc);
/*
* Fill in environment portion of vector table.
*/
for (; envc > 0; --envc) {
suword(vectp++, (long)(intptr_t)destp);
while (*stringp++ != 0)
destp++;
destp++;
}
/* end of vector table is a null pointer */
suword(vectp, 0);
return (stack_base);
}
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
extern int _ucodesel, _udatasel;
extern unsigned long linux_sznonrtsigcode;
static void
1. Change prototype of trapsignal and sendsig to use ksiginfo_t *, most changes in MD code are trivial, before this change, trapsignal and sendsig use discrete parameters, now they uses member fields of ksiginfo_t structure. For sendsig, this change allows us to pass POSIX realtime signal value to user code. 2. Remove cpu_thread_siginfo, it is no longer needed because we now always generate ksiginfo_t data and feed it to libpthread. 3. Add p_sigqueue to proc structure to hold shared signals which were blocked by all threads in the proc. 4. Add td_sigqueue to thread structure to hold all signals delivered to thread. 5. i386 and amd64 now return POSIX standard si_code, other arches will be fixed. 6. In this sigqueue implementation, pending signal set is kept as before, an extra siginfo list holds additional siginfo_t data for signals. kernel code uses psignal() still behavior as before, it won't be failed even under memory pressure, only exception is when deleting a signal, we should call sigqueue_delete to remove signal from sigqueue but not SIGDELSET. Current there is no kernel code will deliver a signal with additional data, so kernel should be as stable as before, a ksiginfo can carry more information, for example, allow signal to be delivered but throw away siginfo data if memory is not enough. SIGKILL and SIGSTOP have fast path in sigqueue_add, because they can not be caught or masked. The sigqueue() syscall allows user code to queue a signal to target process, if resource is unavailable, EAGAIN will be returned as specification said. Just before thread exits, signal queue memory will be freed by sigqueue_flush. Current, all signals are allowed to be queued, not only realtime signals. Earlier patch reviewed by: jhb, deischen Tested on: i386, amd64
2005-10-14 12:43:47 +00:00
linux_rt_sendsig(sig_t catcher, ksiginfo_t *ksi, sigset_t *mask)
{
struct thread *td = curthread;
struct proc *p = td->td_proc;
struct sigacts *psp;
struct trapframe *regs;
Round of cleanups and enhancements. These include (in random order): o Introduce private types for use in linux syscalls for two reasons: 1. establish type independence for ease in porting and, 2. provide a visual queue as to which syscalls have proper prototypes to further cleanup the i386/alpha split. Linuxulator types are prefixed by 'l_'. void and char have not been "virtualized". o Provide dummy functions for all syscalls and remove dummy functions or implementations of truely obsolete syscalls. o Sanitize the shm*, sem* and msg* syscalls. o Make a first attempt to implement the linux_sysctl syscall. At this time it only returns one MIB (KERN_VERSION), but most importantly, it tells us when we need to add additional sysctls :-) o Bump the kenel version up to 2.4.2 (this is not the same as the KERN_VERSION MIB, BTW). o Implement new syscalls, of which most are specific to i386. Our syscall table is now up to date with Linux 2.4.2. Some highlights: - Implement the 32-bit uid_t and gid_t bases syscalls. - Implement a couple of 64-bit file size/offset bases syscalls. o Fix or improve numerous syscalls and prototypes. o Reduce style(9) violations while I'm here. Especially indentation inconsistencies within the same file are addressed. Re-indenting did not obfuscate actual changes to the extend that it could not be combined. NOTE: I spend some time testing these changes and found that if there were regressions, they were not caused by these changes AFAICT. It was observed that installing a RH 7.1 runtime environment did make matters worse. Hangs and/or reboots have been observed with and without these changes, so when it failed to make life better in cases it doesn't look like it made it worse.
2001-09-08 19:07:04 +00:00
struct l_rt_sigframe *fp, frame;
1. Change prototype of trapsignal and sendsig to use ksiginfo_t *, most changes in MD code are trivial, before this change, trapsignal and sendsig use discrete parameters, now they uses member fields of ksiginfo_t structure. For sendsig, this change allows us to pass POSIX realtime signal value to user code. 2. Remove cpu_thread_siginfo, it is no longer needed because we now always generate ksiginfo_t data and feed it to libpthread. 3. Add p_sigqueue to proc structure to hold shared signals which were blocked by all threads in the proc. 4. Add td_sigqueue to thread structure to hold all signals delivered to thread. 5. i386 and amd64 now return POSIX standard si_code, other arches will be fixed. 6. In this sigqueue implementation, pending signal set is kept as before, an extra siginfo list holds additional siginfo_t data for signals. kernel code uses psignal() still behavior as before, it won't be failed even under memory pressure, only exception is when deleting a signal, we should call sigqueue_delete to remove signal from sigqueue but not SIGDELSET. Current there is no kernel code will deliver a signal with additional data, so kernel should be as stable as before, a ksiginfo can carry more information, for example, allow signal to be delivered but throw away siginfo data if memory is not enough. SIGKILL and SIGSTOP have fast path in sigqueue_add, because they can not be caught or masked. The sigqueue() syscall allows user code to queue a signal to target process, if resource is unavailable, EAGAIN will be returned as specification said. Just before thread exits, signal queue memory will be freed by sigqueue_flush. Current, all signals are allowed to be queued, not only realtime signals. Earlier patch reviewed by: jhb, deischen Tested on: i386, amd64
2005-10-14 12:43:47 +00:00
int sig, code;
int oonstack;
1. Change prototype of trapsignal and sendsig to use ksiginfo_t *, most changes in MD code are trivial, before this change, trapsignal and sendsig use discrete parameters, now they uses member fields of ksiginfo_t structure. For sendsig, this change allows us to pass POSIX realtime signal value to user code. 2. Remove cpu_thread_siginfo, it is no longer needed because we now always generate ksiginfo_t data and feed it to libpthread. 3. Add p_sigqueue to proc structure to hold shared signals which were blocked by all threads in the proc. 4. Add td_sigqueue to thread structure to hold all signals delivered to thread. 5. i386 and amd64 now return POSIX standard si_code, other arches will be fixed. 6. In this sigqueue implementation, pending signal set is kept as before, an extra siginfo list holds additional siginfo_t data for signals. kernel code uses psignal() still behavior as before, it won't be failed even under memory pressure, only exception is when deleting a signal, we should call sigqueue_delete to remove signal from sigqueue but not SIGDELSET. Current there is no kernel code will deliver a signal with additional data, so kernel should be as stable as before, a ksiginfo can carry more information, for example, allow signal to be delivered but throw away siginfo data if memory is not enough. SIGKILL and SIGSTOP have fast path in sigqueue_add, because they can not be caught or masked. The sigqueue() syscall allows user code to queue a signal to target process, if resource is unavailable, EAGAIN will be returned as specification said. Just before thread exits, signal queue memory will be freed by sigqueue_flush. Current, all signals are allowed to be queued, not only realtime signals. Earlier patch reviewed by: jhb, deischen Tested on: i386, amd64
2005-10-14 12:43:47 +00:00
sig = ksi->ksi_signo;
code = ksi->ksi_code;
PROC_LOCK_ASSERT(p, MA_OWNED);
psp = p->p_sigacts;
mtx_assert(&psp->ps_mtx, MA_OWNED);
regs = td->td_frame;
oonstack = sigonstack(regs->tf_esp);
#ifdef DEBUG
Round of cleanups and enhancements. These include (in random order): o Introduce private types for use in linux syscalls for two reasons: 1. establish type independence for ease in porting and, 2. provide a visual queue as to which syscalls have proper prototypes to further cleanup the i386/alpha split. Linuxulator types are prefixed by 'l_'. void and char have not been "virtualized". o Provide dummy functions for all syscalls and remove dummy functions or implementations of truely obsolete syscalls. o Sanitize the shm*, sem* and msg* syscalls. o Make a first attempt to implement the linux_sysctl syscall. At this time it only returns one MIB (KERN_VERSION), but most importantly, it tells us when we need to add additional sysctls :-) o Bump the kenel version up to 2.4.2 (this is not the same as the KERN_VERSION MIB, BTW). o Implement new syscalls, of which most are specific to i386. Our syscall table is now up to date with Linux 2.4.2. Some highlights: - Implement the 32-bit uid_t and gid_t bases syscalls. - Implement a couple of 64-bit file size/offset bases syscalls. o Fix or improve numerous syscalls and prototypes. o Reduce style(9) violations while I'm here. Especially indentation inconsistencies within the same file are addressed. Re-indenting did not obfuscate actual changes to the extend that it could not be combined. NOTE: I spend some time testing these changes and found that if there were regressions, they were not caused by these changes AFAICT. It was observed that installing a RH 7.1 runtime environment did make matters worse. Hangs and/or reboots have been observed with and without these changes, so when it failed to make life better in cases it doesn't look like it made it worse.
2001-09-08 19:07:04 +00:00
if (ldebug(rt_sendsig))
printf(ARGS(rt_sendsig, "%p, %d, %p, %u"),
catcher, sig, (void*)mask, code);
#endif
/*
* Allocate space for the signal handler context.
*/
if ((td->td_pflags & TDP_ALTSTACK) && !oonstack &&
SIGISMEMBER(psp->ps_sigonstack, sig)) {
fp = (struct l_rt_sigframe *)(td->td_sigstk.ss_sp +
td->td_sigstk.ss_size - sizeof(struct l_rt_sigframe));
} else
Round of cleanups and enhancements. These include (in random order): o Introduce private types for use in linux syscalls for two reasons: 1. establish type independence for ease in porting and, 2. provide a visual queue as to which syscalls have proper prototypes to further cleanup the i386/alpha split. Linuxulator types are prefixed by 'l_'. void and char have not been "virtualized". o Provide dummy functions for all syscalls and remove dummy functions or implementations of truely obsolete syscalls. o Sanitize the shm*, sem* and msg* syscalls. o Make a first attempt to implement the linux_sysctl syscall. At this time it only returns one MIB (KERN_VERSION), but most importantly, it tells us when we need to add additional sysctls :-) o Bump the kenel version up to 2.4.2 (this is not the same as the KERN_VERSION MIB, BTW). o Implement new syscalls, of which most are specific to i386. Our syscall table is now up to date with Linux 2.4.2. Some highlights: - Implement the 32-bit uid_t and gid_t bases syscalls. - Implement a couple of 64-bit file size/offset bases syscalls. o Fix or improve numerous syscalls and prototypes. o Reduce style(9) violations while I'm here. Especially indentation inconsistencies within the same file are addressed. Re-indenting did not obfuscate actual changes to the extend that it could not be combined. NOTE: I spend some time testing these changes and found that if there were regressions, they were not caused by these changes AFAICT. It was observed that installing a RH 7.1 runtime environment did make matters worse. Hangs and/or reboots have been observed with and without these changes, so when it failed to make life better in cases it doesn't look like it made it worse.
2001-09-08 19:07:04 +00:00
fp = (struct l_rt_sigframe *)regs->tf_esp - 1;
mtx_unlock(&psp->ps_mtx);
/*
* Build the argument list for the signal handler.
*/
if (p->p_sysent->sv_sigtbl)
if (sig <= p->p_sysent->sv_sigsize)
sig = p->p_sysent->sv_sigtbl[_SIG_IDX(sig)];
bzero(&frame, sizeof(frame));
frame.sf_handler = catcher;
frame.sf_sig = sig;
frame.sf_siginfo = &fp->sf_si;
frame.sf_ucontext = &fp->sf_sc;
/* Fill in POSIX parts */
ksiginfo_to_lsiginfo(ksi, &frame.sf_si, sig);
/*
* Build the signal context to be used by sigreturn.
*/
frame.sf_sc.uc_flags = 0; /* XXX ??? */
frame.sf_sc.uc_link = NULL; /* XXX ??? */
frame.sf_sc.uc_stack.ss_sp = td->td_sigstk.ss_sp;
frame.sf_sc.uc_stack.ss_size = td->td_sigstk.ss_size;
frame.sf_sc.uc_stack.ss_flags = (td->td_pflags & TDP_ALTSTACK)
? ((oonstack) ? LINUX_SS_ONSTACK : 0) : LINUX_SS_DISABLE;
2001-01-24 00:27:28 +00:00
PROC_UNLOCK(p);
bsd_to_linux_sigset(mask, &frame.sf_sc.uc_sigmask);
frame.sf_sc.uc_mcontext.sc_mask = frame.sf_sc.uc_sigmask.__bits[0];
frame.sf_sc.uc_mcontext.sc_gs = rgs();
frame.sf_sc.uc_mcontext.sc_fs = regs->tf_fs;
frame.sf_sc.uc_mcontext.sc_es = regs->tf_es;
frame.sf_sc.uc_mcontext.sc_ds = regs->tf_ds;
frame.sf_sc.uc_mcontext.sc_edi = regs->tf_edi;
frame.sf_sc.uc_mcontext.sc_esi = regs->tf_esi;
frame.sf_sc.uc_mcontext.sc_ebp = regs->tf_ebp;
frame.sf_sc.uc_mcontext.sc_ebx = regs->tf_ebx;
frame.sf_sc.uc_mcontext.sc_edx = regs->tf_edx;
frame.sf_sc.uc_mcontext.sc_ecx = regs->tf_ecx;
frame.sf_sc.uc_mcontext.sc_eax = regs->tf_eax;
frame.sf_sc.uc_mcontext.sc_eip = regs->tf_eip;
frame.sf_sc.uc_mcontext.sc_cs = regs->tf_cs;
frame.sf_sc.uc_mcontext.sc_eflags = regs->tf_eflags;
frame.sf_sc.uc_mcontext.sc_esp_at_signal = regs->tf_esp;
frame.sf_sc.uc_mcontext.sc_ss = regs->tf_ss;
frame.sf_sc.uc_mcontext.sc_err = regs->tf_err;
frame.sf_sc.uc_mcontext.sc_cr2 = (register_t)ksi->ksi_addr;
frame.sf_sc.uc_mcontext.sc_trapno = bsd_to_linux_trapcode(code);
#ifdef DEBUG
Round of cleanups and enhancements. These include (in random order): o Introduce private types for use in linux syscalls for two reasons: 1. establish type independence for ease in porting and, 2. provide a visual queue as to which syscalls have proper prototypes to further cleanup the i386/alpha split. Linuxulator types are prefixed by 'l_'. void and char have not been "virtualized". o Provide dummy functions for all syscalls and remove dummy functions or implementations of truely obsolete syscalls. o Sanitize the shm*, sem* and msg* syscalls. o Make a first attempt to implement the linux_sysctl syscall. At this time it only returns one MIB (KERN_VERSION), but most importantly, it tells us when we need to add additional sysctls :-) o Bump the kenel version up to 2.4.2 (this is not the same as the KERN_VERSION MIB, BTW). o Implement new syscalls, of which most are specific to i386. Our syscall table is now up to date with Linux 2.4.2. Some highlights: - Implement the 32-bit uid_t and gid_t bases syscalls. - Implement a couple of 64-bit file size/offset bases syscalls. o Fix or improve numerous syscalls and prototypes. o Reduce style(9) violations while I'm here. Especially indentation inconsistencies within the same file are addressed. Re-indenting did not obfuscate actual changes to the extend that it could not be combined. NOTE: I spend some time testing these changes and found that if there were regressions, they were not caused by these changes AFAICT. It was observed that installing a RH 7.1 runtime environment did make matters worse. Hangs and/or reboots have been observed with and without these changes, so when it failed to make life better in cases it doesn't look like it made it worse.
2001-09-08 19:07:04 +00:00
if (ldebug(rt_sendsig))
printf(LMSG("rt_sendsig flags: 0x%x, sp: %p, ss: 0x%x, mask: 0x%x"),
frame.sf_sc.uc_stack.ss_flags, td->td_sigstk.ss_sp,
td->td_sigstk.ss_size, frame.sf_sc.uc_mcontext.sc_mask);
#endif
if (copyout(&frame, fp, sizeof(frame)) != 0) {
/*
* Process has trashed its stack; give it an illegal
* instruction to halt it in its tracks.
*/
#ifdef DEBUG
if (ldebug(rt_sendsig))
printf(LMSG("rt_sendsig: bad stack %p, oonstack=%x"),
fp, oonstack);
#endif
PROC_LOCK(p);
sigexit(td, SIGILL);
}
/*
* Build context to run handler in.
*/
regs->tf_esp = (int)fp;
regs->tf_eip = p->p_sysent->sv_sigcode_base + linux_sznonrtsigcode;
regs->tf_eflags &= ~(PSL_T | PSL_VM | PSL_D);
regs->tf_cs = _ucodesel;
regs->tf_ds = _udatasel;
regs->tf_es = _udatasel;
regs->tf_fs = _udatasel;
regs->tf_ss = _udatasel;
PROC_LOCK(p);
mtx_lock(&psp->ps_mtx);
}
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
/*
* Send an interrupt to process.
*
* Stack is set up to allow sigcode stored
* in u. to call routine, followed by kcall
* to sigreturn routine below. After sigreturn
* resets the signal mask, the stack, and the
* frame pointer, it returns to the user
* specified pc, psl.
*/
1998-02-09 06:11:36 +00:00
static void
1. Change prototype of trapsignal and sendsig to use ksiginfo_t *, most changes in MD code are trivial, before this change, trapsignal and sendsig use discrete parameters, now they uses member fields of ksiginfo_t structure. For sendsig, this change allows us to pass POSIX realtime signal value to user code. 2. Remove cpu_thread_siginfo, it is no longer needed because we now always generate ksiginfo_t data and feed it to libpthread. 3. Add p_sigqueue to proc structure to hold shared signals which were blocked by all threads in the proc. 4. Add td_sigqueue to thread structure to hold all signals delivered to thread. 5. i386 and amd64 now return POSIX standard si_code, other arches will be fixed. 6. In this sigqueue implementation, pending signal set is kept as before, an extra siginfo list holds additional siginfo_t data for signals. kernel code uses psignal() still behavior as before, it won't be failed even under memory pressure, only exception is when deleting a signal, we should call sigqueue_delete to remove signal from sigqueue but not SIGDELSET. Current there is no kernel code will deliver a signal with additional data, so kernel should be as stable as before, a ksiginfo can carry more information, for example, allow signal to be delivered but throw away siginfo data if memory is not enough. SIGKILL and SIGSTOP have fast path in sigqueue_add, because they can not be caught or masked. The sigqueue() syscall allows user code to queue a signal to target process, if resource is unavailable, EAGAIN will be returned as specification said. Just before thread exits, signal queue memory will be freed by sigqueue_flush. Current, all signals are allowed to be queued, not only realtime signals. Earlier patch reviewed by: jhb, deischen Tested on: i386, amd64
2005-10-14 12:43:47 +00:00
linux_sendsig(sig_t catcher, ksiginfo_t *ksi, sigset_t *mask)
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 thread *td = curthread;
struct proc *p = td->td_proc;
struct sigacts *psp;
struct trapframe *regs;
Round of cleanups and enhancements. These include (in random order): o Introduce private types for use in linux syscalls for two reasons: 1. establish type independence for ease in porting and, 2. provide a visual queue as to which syscalls have proper prototypes to further cleanup the i386/alpha split. Linuxulator types are prefixed by 'l_'. void and char have not been "virtualized". o Provide dummy functions for all syscalls and remove dummy functions or implementations of truely obsolete syscalls. o Sanitize the shm*, sem* and msg* syscalls. o Make a first attempt to implement the linux_sysctl syscall. At this time it only returns one MIB (KERN_VERSION), but most importantly, it tells us when we need to add additional sysctls :-) o Bump the kenel version up to 2.4.2 (this is not the same as the KERN_VERSION MIB, BTW). o Implement new syscalls, of which most are specific to i386. Our syscall table is now up to date with Linux 2.4.2. Some highlights: - Implement the 32-bit uid_t and gid_t bases syscalls. - Implement a couple of 64-bit file size/offset bases syscalls. o Fix or improve numerous syscalls and prototypes. o Reduce style(9) violations while I'm here. Especially indentation inconsistencies within the same file are addressed. Re-indenting did not obfuscate actual changes to the extend that it could not be combined. NOTE: I spend some time testing these changes and found that if there were regressions, they were not caused by these changes AFAICT. It was observed that installing a RH 7.1 runtime environment did make matters worse. Hangs and/or reboots have been observed with and without these changes, so when it failed to make life better in cases it doesn't look like it made it worse.
2001-09-08 19:07:04 +00:00
struct l_sigframe *fp, frame;
l_sigset_t lmask;
1. Change prototype of trapsignal and sendsig to use ksiginfo_t *, most changes in MD code are trivial, before this change, trapsignal and sendsig use discrete parameters, now they uses member fields of ksiginfo_t structure. For sendsig, this change allows us to pass POSIX realtime signal value to user code. 2. Remove cpu_thread_siginfo, it is no longer needed because we now always generate ksiginfo_t data and feed it to libpthread. 3. Add p_sigqueue to proc structure to hold shared signals which were blocked by all threads in the proc. 4. Add td_sigqueue to thread structure to hold all signals delivered to thread. 5. i386 and amd64 now return POSIX standard si_code, other arches will be fixed. 6. In this sigqueue implementation, pending signal set is kept as before, an extra siginfo list holds additional siginfo_t data for signals. kernel code uses psignal() still behavior as before, it won't be failed even under memory pressure, only exception is when deleting a signal, we should call sigqueue_delete to remove signal from sigqueue but not SIGDELSET. Current there is no kernel code will deliver a signal with additional data, so kernel should be as stable as before, a ksiginfo can carry more information, for example, allow signal to be delivered but throw away siginfo data if memory is not enough. SIGKILL and SIGSTOP have fast path in sigqueue_add, because they can not be caught or masked. The sigqueue() syscall allows user code to queue a signal to target process, if resource is unavailable, EAGAIN will be returned as specification said. Just before thread exits, signal queue memory will be freed by sigqueue_flush. Current, all signals are allowed to be queued, not only realtime signals. Earlier patch reviewed by: jhb, deischen Tested on: i386, amd64
2005-10-14 12:43:47 +00:00
int sig, code;
int oonstack, i;
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
PROC_LOCK_ASSERT(p, MA_OWNED);
psp = p->p_sigacts;
1. Change prototype of trapsignal and sendsig to use ksiginfo_t *, most changes in MD code are trivial, before this change, trapsignal and sendsig use discrete parameters, now they uses member fields of ksiginfo_t structure. For sendsig, this change allows us to pass POSIX realtime signal value to user code. 2. Remove cpu_thread_siginfo, it is no longer needed because we now always generate ksiginfo_t data and feed it to libpthread. 3. Add p_sigqueue to proc structure to hold shared signals which were blocked by all threads in the proc. 4. Add td_sigqueue to thread structure to hold all signals delivered to thread. 5. i386 and amd64 now return POSIX standard si_code, other arches will be fixed. 6. In this sigqueue implementation, pending signal set is kept as before, an extra siginfo list holds additional siginfo_t data for signals. kernel code uses psignal() still behavior as before, it won't be failed even under memory pressure, only exception is when deleting a signal, we should call sigqueue_delete to remove signal from sigqueue but not SIGDELSET. Current there is no kernel code will deliver a signal with additional data, so kernel should be as stable as before, a ksiginfo can carry more information, for example, allow signal to be delivered but throw away siginfo data if memory is not enough. SIGKILL and SIGSTOP have fast path in sigqueue_add, because they can not be caught or masked. The sigqueue() syscall allows user code to queue a signal to target process, if resource is unavailable, EAGAIN will be returned as specification said. Just before thread exits, signal queue memory will be freed by sigqueue_flush. Current, all signals are allowed to be queued, not only realtime signals. Earlier patch reviewed by: jhb, deischen Tested on: i386, amd64
2005-10-14 12:43:47 +00:00
sig = ksi->ksi_signo;
code = ksi->ksi_code;
mtx_assert(&psp->ps_mtx, MA_OWNED);
if (SIGISMEMBER(psp->ps_siginfo, sig)) {
/* Signal handler installed with SA_SIGINFO. */
1. Change prototype of trapsignal and sendsig to use ksiginfo_t *, most changes in MD code are trivial, before this change, trapsignal and sendsig use discrete parameters, now they uses member fields of ksiginfo_t structure. For sendsig, this change allows us to pass POSIX realtime signal value to user code. 2. Remove cpu_thread_siginfo, it is no longer needed because we now always generate ksiginfo_t data and feed it to libpthread. 3. Add p_sigqueue to proc structure to hold shared signals which were blocked by all threads in the proc. 4. Add td_sigqueue to thread structure to hold all signals delivered to thread. 5. i386 and amd64 now return POSIX standard si_code, other arches will be fixed. 6. In this sigqueue implementation, pending signal set is kept as before, an extra siginfo list holds additional siginfo_t data for signals. kernel code uses psignal() still behavior as before, it won't be failed even under memory pressure, only exception is when deleting a signal, we should call sigqueue_delete to remove signal from sigqueue but not SIGDELSET. Current there is no kernel code will deliver a signal with additional data, so kernel should be as stable as before, a ksiginfo can carry more information, for example, allow signal to be delivered but throw away siginfo data if memory is not enough. SIGKILL and SIGSTOP have fast path in sigqueue_add, because they can not be caught or masked. The sigqueue() syscall allows user code to queue a signal to target process, if resource is unavailable, EAGAIN will be returned as specification said. Just before thread exits, signal queue memory will be freed by sigqueue_flush. Current, all signals are allowed to be queued, not only realtime signals. Earlier patch reviewed by: jhb, deischen Tested on: i386, amd64
2005-10-14 12:43:47 +00:00
linux_rt_sendsig(catcher, ksi, mask);
return;
}
regs = td->td_frame;
oonstack = sigonstack(regs->tf_esp);
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
Round of cleanups and enhancements. These include (in random order): o Introduce private types for use in linux syscalls for two reasons: 1. establish type independence for ease in porting and, 2. provide a visual queue as to which syscalls have proper prototypes to further cleanup the i386/alpha split. Linuxulator types are prefixed by 'l_'. void and char have not been "virtualized". o Provide dummy functions for all syscalls and remove dummy functions or implementations of truely obsolete syscalls. o Sanitize the shm*, sem* and msg* syscalls. o Make a first attempt to implement the linux_sysctl syscall. At this time it only returns one MIB (KERN_VERSION), but most importantly, it tells us when we need to add additional sysctls :-) o Bump the kenel version up to 2.4.2 (this is not the same as the KERN_VERSION MIB, BTW). o Implement new syscalls, of which most are specific to i386. Our syscall table is now up to date with Linux 2.4.2. Some highlights: - Implement the 32-bit uid_t and gid_t bases syscalls. - Implement a couple of 64-bit file size/offset bases syscalls. o Fix or improve numerous syscalls and prototypes. o Reduce style(9) violations while I'm here. Especially indentation inconsistencies within the same file are addressed. Re-indenting did not obfuscate actual changes to the extend that it could not be combined. NOTE: I spend some time testing these changes and found that if there were regressions, they were not caused by these changes AFAICT. It was observed that installing a RH 7.1 runtime environment did make matters worse. Hangs and/or reboots have been observed with and without these changes, so when it failed to make life better in cases it doesn't look like it made it worse.
2001-09-08 19:07:04 +00:00
if (ldebug(sendsig))
printf(ARGS(sendsig, "%p, %d, %p, %u"),
catcher, sig, (void*)mask, code);
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
#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
/*
* Allocate space for the signal handler context.
*/
if ((td->td_pflags & TDP_ALTSTACK) && !oonstack &&
SIGISMEMBER(psp->ps_sigonstack, sig)) {
fp = (struct l_sigframe *)(td->td_sigstk.ss_sp +
td->td_sigstk.ss_size - sizeof(struct l_sigframe));
} else
Round of cleanups and enhancements. These include (in random order): o Introduce private types for use in linux syscalls for two reasons: 1. establish type independence for ease in porting and, 2. provide a visual queue as to which syscalls have proper prototypes to further cleanup the i386/alpha split. Linuxulator types are prefixed by 'l_'. void and char have not been "virtualized". o Provide dummy functions for all syscalls and remove dummy functions or implementations of truely obsolete syscalls. o Sanitize the shm*, sem* and msg* syscalls. o Make a first attempt to implement the linux_sysctl syscall. At this time it only returns one MIB (KERN_VERSION), but most importantly, it tells us when we need to add additional sysctls :-) o Bump the kenel version up to 2.4.2 (this is not the same as the KERN_VERSION MIB, BTW). o Implement new syscalls, of which most are specific to i386. Our syscall table is now up to date with Linux 2.4.2. Some highlights: - Implement the 32-bit uid_t and gid_t bases syscalls. - Implement a couple of 64-bit file size/offset bases syscalls. o Fix or improve numerous syscalls and prototypes. o Reduce style(9) violations while I'm here. Especially indentation inconsistencies within the same file are addressed. Re-indenting did not obfuscate actual changes to the extend that it could not be combined. NOTE: I spend some time testing these changes and found that if there were regressions, they were not caused by these changes AFAICT. It was observed that installing a RH 7.1 runtime environment did make matters worse. Hangs and/or reboots have been observed with and without these changes, so when it failed to make life better in cases it doesn't look like it made it worse.
2001-09-08 19:07:04 +00:00
fp = (struct l_sigframe *)regs->tf_esp - 1;
mtx_unlock(&psp->ps_mtx);
2001-01-24 00:27:28 +00:00
PROC_UNLOCK(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
/*
* Build the argument list for the signal handler.
*/
if (p->p_sysent->sv_sigtbl)
if (sig <= p->p_sysent->sv_sigsize)
sig = p->p_sysent->sv_sigtbl[_SIG_IDX(sig)];
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
bzero(&frame, sizeof(frame));
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
frame.sf_handler = catcher;
frame.sf_sig = sig;
bsd_to_linux_sigset(mask, &lmask);
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
/*
* Build the signal context to be used by sigreturn.
*/
frame.sf_sc.sc_mask = lmask.__bits[0];
frame.sf_sc.sc_gs = rgs();
frame.sf_sc.sc_fs = regs->tf_fs;
frame.sf_sc.sc_es = regs->tf_es;
frame.sf_sc.sc_ds = regs->tf_ds;
frame.sf_sc.sc_edi = regs->tf_edi;
frame.sf_sc.sc_esi = regs->tf_esi;
frame.sf_sc.sc_ebp = regs->tf_ebp;
frame.sf_sc.sc_ebx = regs->tf_ebx;
frame.sf_sc.sc_edx = regs->tf_edx;
frame.sf_sc.sc_ecx = regs->tf_ecx;
frame.sf_sc.sc_eax = regs->tf_eax;
frame.sf_sc.sc_eip = regs->tf_eip;
frame.sf_sc.sc_cs = regs->tf_cs;
frame.sf_sc.sc_eflags = regs->tf_eflags;
frame.sf_sc.sc_esp_at_signal = regs->tf_esp;
frame.sf_sc.sc_ss = regs->tf_ss;
frame.sf_sc.sc_err = regs->tf_err;
frame.sf_sc.sc_cr2 = (register_t)ksi->ksi_addr;
1. Change prototype of trapsignal and sendsig to use ksiginfo_t *, most changes in MD code are trivial, before this change, trapsignal and sendsig use discrete parameters, now they uses member fields of ksiginfo_t structure. For sendsig, this change allows us to pass POSIX realtime signal value to user code. 2. Remove cpu_thread_siginfo, it is no longer needed because we now always generate ksiginfo_t data and feed it to libpthread. 3. Add p_sigqueue to proc structure to hold shared signals which were blocked by all threads in the proc. 4. Add td_sigqueue to thread structure to hold all signals delivered to thread. 5. i386 and amd64 now return POSIX standard si_code, other arches will be fixed. 6. In this sigqueue implementation, pending signal set is kept as before, an extra siginfo list holds additional siginfo_t data for signals. kernel code uses psignal() still behavior as before, it won't be failed even under memory pressure, only exception is when deleting a signal, we should call sigqueue_delete to remove signal from sigqueue but not SIGDELSET. Current there is no kernel code will deliver a signal with additional data, so kernel should be as stable as before, a ksiginfo can carry more information, for example, allow signal to be delivered but throw away siginfo data if memory is not enough. SIGKILL and SIGSTOP have fast path in sigqueue_add, because they can not be caught or masked. The sigqueue() syscall allows user code to queue a signal to target process, if resource is unavailable, EAGAIN will be returned as specification said. Just before thread exits, signal queue memory will be freed by sigqueue_flush. Current, all signals are allowed to be queued, not only realtime signals. Earlier patch reviewed by: jhb, deischen Tested on: i386, amd64
2005-10-14 12:43:47 +00:00
frame.sf_sc.sc_trapno = bsd_to_linux_trapcode(ksi->ksi_trapno);
for (i = 0; i < (LINUX_NSIG_WORDS-1); i++)
frame.sf_extramask[i] = lmask.__bits[i+1];
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 (copyout(&frame, fp, sizeof(frame)) != 0) {
/*
* Process has trashed its stack; give it an illegal
* instruction to halt it in its tracks.
*/
PROC_LOCK(p);
sigexit(td, SIGILL);
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
}
/*
* Build context to run handler in.
*/
regs->tf_esp = (int)fp;
regs->tf_eip = p->p_sysent->sv_sigcode_base;
regs->tf_eflags &= ~(PSL_T | PSL_VM | PSL_D);
regs->tf_cs = _ucodesel;
regs->tf_ds = _udatasel;
regs->tf_es = _udatasel;
regs->tf_fs = _udatasel;
regs->tf_ss = _udatasel;
Round of cleanups and enhancements. These include (in random order): o Introduce private types for use in linux syscalls for two reasons: 1. establish type independence for ease in porting and, 2. provide a visual queue as to which syscalls have proper prototypes to further cleanup the i386/alpha split. Linuxulator types are prefixed by 'l_'. void and char have not been "virtualized". o Provide dummy functions for all syscalls and remove dummy functions or implementations of truely obsolete syscalls. o Sanitize the shm*, sem* and msg* syscalls. o Make a first attempt to implement the linux_sysctl syscall. At this time it only returns one MIB (KERN_VERSION), but most importantly, it tells us when we need to add additional sysctls :-) o Bump the kenel version up to 2.4.2 (this is not the same as the KERN_VERSION MIB, BTW). o Implement new syscalls, of which most are specific to i386. Our syscall table is now up to date with Linux 2.4.2. Some highlights: - Implement the 32-bit uid_t and gid_t bases syscalls. - Implement a couple of 64-bit file size/offset bases syscalls. o Fix or improve numerous syscalls and prototypes. o Reduce style(9) violations while I'm here. Especially indentation inconsistencies within the same file are addressed. Re-indenting did not obfuscate actual changes to the extend that it could not be combined. NOTE: I spend some time testing these changes and found that if there were regressions, they were not caused by these changes AFAICT. It was observed that installing a RH 7.1 runtime environment did make matters worse. Hangs and/or reboots have been observed with and without these changes, so when it failed to make life better in cases it doesn't look like it made it worse.
2001-09-08 19:07:04 +00:00
PROC_LOCK(p);
mtx_lock(&psp->ps_mtx);
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
}
/*
* System call to cleanup state after a signal
* has been taken. Reset signal mask and
* stack state from context left by sendsig (above).
* Return to previous pc and psl as specified by
* context left by sendsig. Check carefully to
* make sure that the user has not modified the
* psl to gain improper privileges or to cause
* a machine fault.
*/
int
linux_sigreturn(struct thread *td, struct linux_sigreturn_args *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
{
Round of cleanups and enhancements. These include (in random order): o Introduce private types for use in linux syscalls for two reasons: 1. establish type independence for ease in porting and, 2. provide a visual queue as to which syscalls have proper prototypes to further cleanup the i386/alpha split. Linuxulator types are prefixed by 'l_'. void and char have not been "virtualized". o Provide dummy functions for all syscalls and remove dummy functions or implementations of truely obsolete syscalls. o Sanitize the shm*, sem* and msg* syscalls. o Make a first attempt to implement the linux_sysctl syscall. At this time it only returns one MIB (KERN_VERSION), but most importantly, it tells us when we need to add additional sysctls :-) o Bump the kenel version up to 2.4.2 (this is not the same as the KERN_VERSION MIB, BTW). o Implement new syscalls, of which most are specific to i386. Our syscall table is now up to date with Linux 2.4.2. Some highlights: - Implement the 32-bit uid_t and gid_t bases syscalls. - Implement a couple of 64-bit file size/offset bases syscalls. o Fix or improve numerous syscalls and prototypes. o Reduce style(9) violations while I'm here. Especially indentation inconsistencies within the same file are addressed. Re-indenting did not obfuscate actual changes to the extend that it could not be combined. NOTE: I spend some time testing these changes and found that if there were regressions, they were not caused by these changes AFAICT. It was observed that installing a RH 7.1 runtime environment did make matters worse. Hangs and/or reboots have been observed with and without these changes, so when it failed to make life better in cases it doesn't look like it made it worse.
2001-09-08 19:07:04 +00:00
struct l_sigframe frame;
struct trapframe *regs;
Round of cleanups and enhancements. These include (in random order): o Introduce private types for use in linux syscalls for two reasons: 1. establish type independence for ease in porting and, 2. provide a visual queue as to which syscalls have proper prototypes to further cleanup the i386/alpha split. Linuxulator types are prefixed by 'l_'. void and char have not been "virtualized". o Provide dummy functions for all syscalls and remove dummy functions or implementations of truely obsolete syscalls. o Sanitize the shm*, sem* and msg* syscalls. o Make a first attempt to implement the linux_sysctl syscall. At this time it only returns one MIB (KERN_VERSION), but most importantly, it tells us when we need to add additional sysctls :-) o Bump the kenel version up to 2.4.2 (this is not the same as the KERN_VERSION MIB, BTW). o Implement new syscalls, of which most are specific to i386. Our syscall table is now up to date with Linux 2.4.2. Some highlights: - Implement the 32-bit uid_t and gid_t bases syscalls. - Implement a couple of 64-bit file size/offset bases syscalls. o Fix or improve numerous syscalls and prototypes. o Reduce style(9) violations while I'm here. Especially indentation inconsistencies within the same file are addressed. Re-indenting did not obfuscate actual changes to the extend that it could not be combined. NOTE: I spend some time testing these changes and found that if there were regressions, they were not caused by these changes AFAICT. It was observed that installing a RH 7.1 runtime environment did make matters worse. Hangs and/or reboots have been observed with and without these changes, so when it failed to make life better in cases it doesn't look like it made it worse.
2001-09-08 19:07:04 +00:00
l_sigset_t lmask;
sigset_t bmask;
int eflags, i;
1. Change prototype of trapsignal and sendsig to use ksiginfo_t *, most changes in MD code are trivial, before this change, trapsignal and sendsig use discrete parameters, now they uses member fields of ksiginfo_t structure. For sendsig, this change allows us to pass POSIX realtime signal value to user code. 2. Remove cpu_thread_siginfo, it is no longer needed because we now always generate ksiginfo_t data and feed it to libpthread. 3. Add p_sigqueue to proc structure to hold shared signals which were blocked by all threads in the proc. 4. Add td_sigqueue to thread structure to hold all signals delivered to thread. 5. i386 and amd64 now return POSIX standard si_code, other arches will be fixed. 6. In this sigqueue implementation, pending signal set is kept as before, an extra siginfo list holds additional siginfo_t data for signals. kernel code uses psignal() still behavior as before, it won't be failed even under memory pressure, only exception is when deleting a signal, we should call sigqueue_delete to remove signal from sigqueue but not SIGDELSET. Current there is no kernel code will deliver a signal with additional data, so kernel should be as stable as before, a ksiginfo can carry more information, for example, allow signal to be delivered but throw away siginfo data if memory is not enough. SIGKILL and SIGSTOP have fast path in sigqueue_add, because they can not be caught or masked. The sigqueue() syscall allows user code to queue a signal to target process, if resource is unavailable, EAGAIN will be returned as specification said. Just before thread exits, signal queue memory will be freed by sigqueue_flush. Current, all signals are allowed to be queued, not only realtime signals. Earlier patch reviewed by: jhb, deischen Tested on: i386, amd64
2005-10-14 12:43:47 +00:00
ksiginfo_t ksi;
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
regs = td->td_frame;
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
if (ldebug(sigreturn))
printf(ARGS(sigreturn, "%p"), (void *)args->sfp);
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
#endif
/*
* The trampoline code hands us the sigframe.
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
* It is unsafe to keep track of it ourselves, in the event that a
* program jumps out of a signal handler.
*/
if (copyin(args->sfp, &frame, sizeof(frame)) != 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
return (EFAULT);
/*
* Check for security violations.
*/
#define EFLAGS_SECURE(ef, oef) ((((ef) ^ (oef)) & ~PSL_USERCHANGE) == 0)
eflags = frame.sf_sc.sc_eflags;
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 do allow users to change the privileged flag PSL_RF. The
* cpu sets PSL_RF in tf_eflags for faults. Debuggers should
* sometimes set it there too. tf_eflags is kept in the signal
* context during signal handling and there is no other place
* to remember it, so the PSL_RF bit may be corrupted by the
* signal handler without us knowing. Corruption of the PSL_RF
* bit at worst causes one more or one less debugger trap, so
* allowing it is fairly harmless.
*/
if (!EFLAGS_SECURE(eflags & ~PSL_RF, regs->tf_eflags & ~PSL_RF))
return(EINVAL);
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
/*
* Don't allow users to load a valid privileged %cs. Let the
* hardware check for invalid selectors, excess privilege in
* other selectors, invalid %eip's and invalid %esp's.
*/
#define CS_SECURE(cs) (ISPL(cs) == SEL_UPL)
if (!CS_SECURE(frame.sf_sc.sc_cs)) {
1. Change prototype of trapsignal and sendsig to use ksiginfo_t *, most changes in MD code are trivial, before this change, trapsignal and sendsig use discrete parameters, now they uses member fields of ksiginfo_t structure. For sendsig, this change allows us to pass POSIX realtime signal value to user code. 2. Remove cpu_thread_siginfo, it is no longer needed because we now always generate ksiginfo_t data and feed it to libpthread. 3. Add p_sigqueue to proc structure to hold shared signals which were blocked by all threads in the proc. 4. Add td_sigqueue to thread structure to hold all signals delivered to thread. 5. i386 and amd64 now return POSIX standard si_code, other arches will be fixed. 6. In this sigqueue implementation, pending signal set is kept as before, an extra siginfo list holds additional siginfo_t data for signals. kernel code uses psignal() still behavior as before, it won't be failed even under memory pressure, only exception is when deleting a signal, we should call sigqueue_delete to remove signal from sigqueue but not SIGDELSET. Current there is no kernel code will deliver a signal with additional data, so kernel should be as stable as before, a ksiginfo can carry more information, for example, allow signal to be delivered but throw away siginfo data if memory is not enough. SIGKILL and SIGSTOP have fast path in sigqueue_add, because they can not be caught or masked. The sigqueue() syscall allows user code to queue a signal to target process, if resource is unavailable, EAGAIN will be returned as specification said. Just before thread exits, signal queue memory will be freed by sigqueue_flush. Current, all signals are allowed to be queued, not only realtime signals. Earlier patch reviewed by: jhb, deischen Tested on: i386, amd64
2005-10-14 12:43:47 +00:00
ksiginfo_init_trap(&ksi);
ksi.ksi_signo = SIGBUS;
ksi.ksi_code = BUS_OBJERR;
ksi.ksi_trapno = T_PROTFLT;
ksi.ksi_addr = (void *)regs->tf_eip;
trapsignal(td, &ksi);
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
return(EINVAL);
}
lmask.__bits[0] = frame.sf_sc.sc_mask;
for (i = 0; i < (LINUX_NSIG_WORDS-1); i++)
lmask.__bits[i+1] = frame.sf_extramask[i];
linux_to_bsd_sigset(&lmask, &bmask);
kern_sigprocmask(td, SIG_SETMASK, &bmask, NULL, 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
/*
* Restore signal context.
*/
/* %gs was restored by the trampoline. */
regs->tf_fs = frame.sf_sc.sc_fs;
regs->tf_es = frame.sf_sc.sc_es;
regs->tf_ds = frame.sf_sc.sc_ds;
regs->tf_edi = frame.sf_sc.sc_edi;
regs->tf_esi = frame.sf_sc.sc_esi;
regs->tf_ebp = frame.sf_sc.sc_ebp;
regs->tf_ebx = frame.sf_sc.sc_ebx;
regs->tf_edx = frame.sf_sc.sc_edx;
regs->tf_ecx = frame.sf_sc.sc_ecx;
regs->tf_eax = frame.sf_sc.sc_eax;
regs->tf_eip = frame.sf_sc.sc_eip;
regs->tf_cs = frame.sf_sc.sc_cs;
regs->tf_eflags = eflags;
regs->tf_esp = frame.sf_sc.sc_esp_at_signal;
regs->tf_ss = frame.sf_sc.sc_ss;
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
return (EJUSTRETURN);
}
/*
* System call to cleanup state after a signal
* has been taken. Reset signal mask and
* stack state from context left by rt_sendsig (above).
* Return to previous pc and psl as specified by
* context left by sendsig. Check carefully to
* make sure that the user has not modified the
* psl to gain improper privileges or to cause
* a machine fault.
*/
int
linux_rt_sigreturn(struct thread *td, struct linux_rt_sigreturn_args *args)
{
Round of cleanups and enhancements. These include (in random order): o Introduce private types for use in linux syscalls for two reasons: 1. establish type independence for ease in porting and, 2. provide a visual queue as to which syscalls have proper prototypes to further cleanup the i386/alpha split. Linuxulator types are prefixed by 'l_'. void and char have not been "virtualized". o Provide dummy functions for all syscalls and remove dummy functions or implementations of truely obsolete syscalls. o Sanitize the shm*, sem* and msg* syscalls. o Make a first attempt to implement the linux_sysctl syscall. At this time it only returns one MIB (KERN_VERSION), but most importantly, it tells us when we need to add additional sysctls :-) o Bump the kenel version up to 2.4.2 (this is not the same as the KERN_VERSION MIB, BTW). o Implement new syscalls, of which most are specific to i386. Our syscall table is now up to date with Linux 2.4.2. Some highlights: - Implement the 32-bit uid_t and gid_t bases syscalls. - Implement a couple of 64-bit file size/offset bases syscalls. o Fix or improve numerous syscalls and prototypes. o Reduce style(9) violations while I'm here. Especially indentation inconsistencies within the same file are addressed. Re-indenting did not obfuscate actual changes to the extend that it could not be combined. NOTE: I spend some time testing these changes and found that if there were regressions, they were not caused by these changes AFAICT. It was observed that installing a RH 7.1 runtime environment did make matters worse. Hangs and/or reboots have been observed with and without these changes, so when it failed to make life better in cases it doesn't look like it made it worse.
2001-09-08 19:07:04 +00:00
struct l_ucontext uc;
struct l_sigcontext *context;
sigset_t bmask;
Round of cleanups and enhancements. These include (in random order): o Introduce private types for use in linux syscalls for two reasons: 1. establish type independence for ease in porting and, 2. provide a visual queue as to which syscalls have proper prototypes to further cleanup the i386/alpha split. Linuxulator types are prefixed by 'l_'. void and char have not been "virtualized". o Provide dummy functions for all syscalls and remove dummy functions or implementations of truely obsolete syscalls. o Sanitize the shm*, sem* and msg* syscalls. o Make a first attempt to implement the linux_sysctl syscall. At this time it only returns one MIB (KERN_VERSION), but most importantly, it tells us when we need to add additional sysctls :-) o Bump the kenel version up to 2.4.2 (this is not the same as the KERN_VERSION MIB, BTW). o Implement new syscalls, of which most are specific to i386. Our syscall table is now up to date with Linux 2.4.2. Some highlights: - Implement the 32-bit uid_t and gid_t bases syscalls. - Implement a couple of 64-bit file size/offset bases syscalls. o Fix or improve numerous syscalls and prototypes. o Reduce style(9) violations while I'm here. Especially indentation inconsistencies within the same file are addressed. Re-indenting did not obfuscate actual changes to the extend that it could not be combined. NOTE: I spend some time testing these changes and found that if there were regressions, they were not caused by these changes AFAICT. It was observed that installing a RH 7.1 runtime environment did make matters worse. Hangs and/or reboots have been observed with and without these changes, so when it failed to make life better in cases it doesn't look like it made it worse.
2001-09-08 19:07:04 +00:00
l_stack_t *lss;
stack_t ss;
struct trapframe *regs;
int eflags;
1. Change prototype of trapsignal and sendsig to use ksiginfo_t *, most changes in MD code are trivial, before this change, trapsignal and sendsig use discrete parameters, now they uses member fields of ksiginfo_t structure. For sendsig, this change allows us to pass POSIX realtime signal value to user code. 2. Remove cpu_thread_siginfo, it is no longer needed because we now always generate ksiginfo_t data and feed it to libpthread. 3. Add p_sigqueue to proc structure to hold shared signals which were blocked by all threads in the proc. 4. Add td_sigqueue to thread structure to hold all signals delivered to thread. 5. i386 and amd64 now return POSIX standard si_code, other arches will be fixed. 6. In this sigqueue implementation, pending signal set is kept as before, an extra siginfo list holds additional siginfo_t data for signals. kernel code uses psignal() still behavior as before, it won't be failed even under memory pressure, only exception is when deleting a signal, we should call sigqueue_delete to remove signal from sigqueue but not SIGDELSET. Current there is no kernel code will deliver a signal with additional data, so kernel should be as stable as before, a ksiginfo can carry more information, for example, allow signal to be delivered but throw away siginfo data if memory is not enough. SIGKILL and SIGSTOP have fast path in sigqueue_add, because they can not be caught or masked. The sigqueue() syscall allows user code to queue a signal to target process, if resource is unavailable, EAGAIN will be returned as specification said. Just before thread exits, signal queue memory will be freed by sigqueue_flush. Current, all signals are allowed to be queued, not only realtime signals. Earlier patch reviewed by: jhb, deischen Tested on: i386, amd64
2005-10-14 12:43:47 +00:00
ksiginfo_t ksi;
regs = td->td_frame;
#ifdef DEBUG
if (ldebug(rt_sigreturn))
printf(ARGS(rt_sigreturn, "%p"), (void *)args->ucp);
#endif
/*
* The trampoline code hands us the ucontext.
* It is unsafe to keep track of it ourselves, in the event that a
* program jumps out of a signal handler.
*/
if (copyin(args->ucp, &uc, sizeof(uc)) != 0)
return (EFAULT);
context = &uc.uc_mcontext;
/*
* Check for security violations.
*/
#define EFLAGS_SECURE(ef, oef) ((((ef) ^ (oef)) & ~PSL_USERCHANGE) == 0)
eflags = context->sc_eflags;
/*
* XXX do allow users to change the privileged flag PSL_RF. The
* cpu sets PSL_RF in tf_eflags for faults. Debuggers should
* sometimes set it there too. tf_eflags is kept in the signal
* context during signal handling and there is no other place
* to remember it, so the PSL_RF bit may be corrupted by the
* signal handler without us knowing. Corruption of the PSL_RF
* bit at worst causes one more or one less debugger trap, so
* allowing it is fairly harmless.
*/
if (!EFLAGS_SECURE(eflags & ~PSL_RF, regs->tf_eflags & ~PSL_RF))
return(EINVAL);
/*
* Don't allow users to load a valid privileged %cs. Let the
* hardware check for invalid selectors, excess privilege in
* other selectors, invalid %eip's and invalid %esp's.
*/
#define CS_SECURE(cs) (ISPL(cs) == SEL_UPL)
if (!CS_SECURE(context->sc_cs)) {
1. Change prototype of trapsignal and sendsig to use ksiginfo_t *, most changes in MD code are trivial, before this change, trapsignal and sendsig use discrete parameters, now they uses member fields of ksiginfo_t structure. For sendsig, this change allows us to pass POSIX realtime signal value to user code. 2. Remove cpu_thread_siginfo, it is no longer needed because we now always generate ksiginfo_t data and feed it to libpthread. 3. Add p_sigqueue to proc structure to hold shared signals which were blocked by all threads in the proc. 4. Add td_sigqueue to thread structure to hold all signals delivered to thread. 5. i386 and amd64 now return POSIX standard si_code, other arches will be fixed. 6. In this sigqueue implementation, pending signal set is kept as before, an extra siginfo list holds additional siginfo_t data for signals. kernel code uses psignal() still behavior as before, it won't be failed even under memory pressure, only exception is when deleting a signal, we should call sigqueue_delete to remove signal from sigqueue but not SIGDELSET. Current there is no kernel code will deliver a signal with additional data, so kernel should be as stable as before, a ksiginfo can carry more information, for example, allow signal to be delivered but throw away siginfo data if memory is not enough. SIGKILL and SIGSTOP have fast path in sigqueue_add, because they can not be caught or masked. The sigqueue() syscall allows user code to queue a signal to target process, if resource is unavailable, EAGAIN will be returned as specification said. Just before thread exits, signal queue memory will be freed by sigqueue_flush. Current, all signals are allowed to be queued, not only realtime signals. Earlier patch reviewed by: jhb, deischen Tested on: i386, amd64
2005-10-14 12:43:47 +00:00
ksiginfo_init_trap(&ksi);
ksi.ksi_signo = SIGBUS;
ksi.ksi_code = BUS_OBJERR;
ksi.ksi_trapno = T_PROTFLT;
ksi.ksi_addr = (void *)regs->tf_eip;
trapsignal(td, &ksi);
return(EINVAL);
}
linux_to_bsd_sigset(&uc.uc_sigmask, &bmask);
kern_sigprocmask(td, SIG_SETMASK, &bmask, NULL, 0);
/*
* Restore signal context
*/
/* %gs was restored by the trampoline. */
regs->tf_fs = context->sc_fs;
regs->tf_es = context->sc_es;
regs->tf_ds = context->sc_ds;
regs->tf_edi = context->sc_edi;
regs->tf_esi = context->sc_esi;
regs->tf_ebp = context->sc_ebp;
regs->tf_ebx = context->sc_ebx;
regs->tf_edx = context->sc_edx;
regs->tf_ecx = context->sc_ecx;
regs->tf_eax = context->sc_eax;
regs->tf_eip = context->sc_eip;
regs->tf_cs = context->sc_cs;
regs->tf_eflags = eflags;
regs->tf_esp = context->sc_esp_at_signal;
regs->tf_ss = context->sc_ss;
/*
* call sigaltstack & ignore results..
*/
lss = &uc.uc_stack;
ss.ss_sp = lss->ss_sp;
ss.ss_size = lss->ss_size;
ss.ss_flags = linux_to_bsd_sigaltstack(lss->ss_flags);
#ifdef DEBUG
if (ldebug(rt_sigreturn))
printf(LMSG("rt_sigret flags: 0x%x, sp: %p, ss: 0x%x, mask: 0x%x"),
ss.ss_flags, ss.ss_sp, ss.ss_size, context->sc_mask);
#endif
(void)kern_sigaltstack(td, &ss, NULL);
return (EJUSTRETURN);
}
Reorganize syscall entry and leave handling. Extend struct sysvec with three new elements: sv_fetch_syscall_args - the method to fetch syscall arguments from usermode into struct syscall_args. The structure is machine-depended (this might be reconsidered after all architectures are converted). sv_set_syscall_retval - the method to set a return value for usermode from the syscall. It is a generalization of cpu_set_syscall_retval(9) to allow ABIs to override the way to set a return value. sv_syscallnames - the table of syscall names. Use sv_set_syscall_retval in kern_sigsuspend() instead of hardcoding the call to cpu_set_syscall_retval(). The new functions syscallenter(9) and syscallret(9) are provided that use sv_*syscall* pointers and contain the common repeated code from the syscall() implementations for the architecture-specific syscall trap handlers. Syscallenter() fetches arguments, calls syscall implementation from ABI sysent table, and set up return frame. The end of syscall bookkeeping is done by syscallret(). Take advantage of single place for MI syscall handling code and implement ptrace_lwpinfo pl_flags PL_FLAG_SCE, PL_FLAG_SCX and PL_FLAG_EXEC. The SCE and SCX flags notify the debugger that the thread is stopped at syscall entry or return point respectively. The EXEC flag augments SCX and notifies debugger that the process address space was changed by one of exec(2)-family syscalls. The i386, amd64, sparc64, sun4v, powerpc and ia64 syscall()s are changed to use syscallenter()/syscallret(). MIPS and arm are not converted and use the mostly unchanged syscall() implementation. Reviewed by: jhb, marcel, marius, nwhitehorn, stas Tested by: marcel (ia64), marius (sparc64), nwhitehorn (powerpc), stas (mips) MFC after: 1 month
2010-05-23 18:32:02 +00:00
static int
linux_fetch_syscall_args(struct thread *td, struct syscall_args *sa)
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
{
Reorganize syscall entry and leave handling. Extend struct sysvec with three new elements: sv_fetch_syscall_args - the method to fetch syscall arguments from usermode into struct syscall_args. The structure is machine-depended (this might be reconsidered after all architectures are converted). sv_set_syscall_retval - the method to set a return value for usermode from the syscall. It is a generalization of cpu_set_syscall_retval(9) to allow ABIs to override the way to set a return value. sv_syscallnames - the table of syscall names. Use sv_set_syscall_retval in kern_sigsuspend() instead of hardcoding the call to cpu_set_syscall_retval(). The new functions syscallenter(9) and syscallret(9) are provided that use sv_*syscall* pointers and contain the common repeated code from the syscall() implementations for the architecture-specific syscall trap handlers. Syscallenter() fetches arguments, calls syscall implementation from ABI sysent table, and set up return frame. The end of syscall bookkeeping is done by syscallret(). Take advantage of single place for MI syscall handling code and implement ptrace_lwpinfo pl_flags PL_FLAG_SCE, PL_FLAG_SCX and PL_FLAG_EXEC. The SCE and SCX flags notify the debugger that the thread is stopped at syscall entry or return point respectively. The EXEC flag augments SCX and notifies debugger that the process address space was changed by one of exec(2)-family syscalls. The i386, amd64, sparc64, sun4v, powerpc and ia64 syscall()s are changed to use syscallenter()/syscallret(). MIPS and arm are not converted and use the mostly unchanged syscall() implementation. Reviewed by: jhb, marcel, marius, nwhitehorn, stas Tested by: marcel (ia64), marius (sparc64), nwhitehorn (powerpc), stas (mips) MFC after: 1 month
2010-05-23 18:32:02 +00:00
struct proc *p;
struct trapframe *frame;
p = td->td_proc;
frame = td->td_frame;
sa->code = frame->tf_eax;
sa->args[0] = frame->tf_ebx;
sa->args[1] = frame->tf_ecx;
sa->args[2] = frame->tf_edx;
sa->args[3] = frame->tf_esi;
sa->args[4] = frame->tf_edi;
sa->args[5] = frame->tf_ebp; /* Unconfirmed */
if (sa->code >= p->p_sysent->sv_size)
sa->callp = &p->p_sysent->sv_table[0];
else
sa->callp = &p->p_sysent->sv_table[sa->code];
sa->narg = sa->callp->sy_narg;
td->td_retval[0] = 0;
td->td_retval[1] = frame->tf_edx;
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
}
/*
* If a linux binary is exec'ing something, try this image activator
* first. We override standard shell script execution in order to
* be able to modify the interpreter path. We only do this if a linux
* binary is doing the exec, so we do not create an EXEC module for it.
*/
2002-03-20 07:51:46 +00:00
static int exec_linux_imgact_try(struct image_params *iparams);
static int
exec_linux_imgact_try(struct image_params *imgp)
{
const char *head = (const char *)imgp->image_header;
char *rpath;
int error = -1;
/*
* The interpreter for shell scripts run from a linux binary needs
* to be located in /compat/linux if possible in order to recursively
* maintain linux path emulation.
*/
if (((const short *)head)[0] == SHELLMAGIC) {
/*
* Run our normal shell image activator. If it succeeds attempt
* to use the alternate path for the interpreter. If an alternate
* path is found, use our stringspace to store it.
*/
if ((error = exec_shell_imgact(imgp)) == 0) {
linux_emul_convpath(FIRST_THREAD_IN_PROC(imgp->proc),
imgp->interpreter_name, UIO_SYSSPACE, &rpath, 0, AT_FDCWD);
if (rpath != NULL)
imgp->args->fname_buf =
imgp->interpreter_name = rpath;
}
}
return (error);
}
/*
* exec_setregs may initialize some registers differently than Linux
* does, thus potentially confusing Linux binaries. If necessary, we
* override the exec_setregs default(s) here.
*/
static void
exec_linux_setregs(struct thread *td, struct image_params *imgp, u_long stack)
{
struct pcb *pcb = td->td_pcb;
exec_setregs(td, imgp, stack);
/* Linux sets %gs to 0, we default to _udatasel */
pcb->pcb_gs = 0;
load_gs(0);
pcb->pcb_initial_npxcw = __LINUX_NPXCW__;
}
static void
linux_get_machine(const char **dst)
{
switch (cpu_class) {
case CPUCLASS_686:
*dst = "i686";
break;
case CPUCLASS_586:
*dst = "i586";
break;
case CPUCLASS_486:
*dst = "i486";
break;
default:
*dst = "i386";
}
}
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 sysentvec linux_sysvec = {
.sv_size = LINUX_SYS_MAXSYSCALL,
.sv_table = linux_sysent,
.sv_mask = 0,
.sv_sigsize = LINUX_SIGTBLSZ,
.sv_sigtbl = bsd_to_linux_signal,
.sv_errsize = ELAST + 1,
.sv_errtbl = bsd_to_linux_errno,
.sv_transtrap = translate_traps,
.sv_fixup = linux_fixup,
.sv_sendsig = linux_sendsig,
.sv_sigcode = linux_sigcode,
.sv_szsigcode = &linux_szsigcode,
Reorganize syscall entry and leave handling. Extend struct sysvec with three new elements: sv_fetch_syscall_args - the method to fetch syscall arguments from usermode into struct syscall_args. The structure is machine-depended (this might be reconsidered after all architectures are converted). sv_set_syscall_retval - the method to set a return value for usermode from the syscall. It is a generalization of cpu_set_syscall_retval(9) to allow ABIs to override the way to set a return value. sv_syscallnames - the table of syscall names. Use sv_set_syscall_retval in kern_sigsuspend() instead of hardcoding the call to cpu_set_syscall_retval(). The new functions syscallenter(9) and syscallret(9) are provided that use sv_*syscall* pointers and contain the common repeated code from the syscall() implementations for the architecture-specific syscall trap handlers. Syscallenter() fetches arguments, calls syscall implementation from ABI sysent table, and set up return frame. The end of syscall bookkeeping is done by syscallret(). Take advantage of single place for MI syscall handling code and implement ptrace_lwpinfo pl_flags PL_FLAG_SCE, PL_FLAG_SCX and PL_FLAG_EXEC. The SCE and SCX flags notify the debugger that the thread is stopped at syscall entry or return point respectively. The EXEC flag augments SCX and notifies debugger that the process address space was changed by one of exec(2)-family syscalls. The i386, amd64, sparc64, sun4v, powerpc and ia64 syscall()s are changed to use syscallenter()/syscallret(). MIPS and arm are not converted and use the mostly unchanged syscall() implementation. Reviewed by: jhb, marcel, marius, nwhitehorn, stas Tested by: marcel (ia64), marius (sparc64), nwhitehorn (powerpc), stas (mips) MFC after: 1 month
2010-05-23 18:32:02 +00:00
.sv_prepsyscall = NULL,
.sv_name = "Linux a.out",
.sv_coredump = NULL,
.sv_imgact_try = exec_linux_imgact_try,
.sv_minsigstksz = LINUX_MINSIGSTKSZ,
.sv_pagesize = PAGE_SIZE,
.sv_minuser = VM_MIN_ADDRESS,
.sv_maxuser = VM_MAXUSER_ADDRESS,
.sv_usrstack = LINUX_USRSTACK,
.sv_psstrings = PS_STRINGS,
.sv_stackprot = VM_PROT_ALL,
.sv_copyout_strings = exec_copyout_strings,
.sv_setregs = exec_linux_setregs,
.sv_fixlimit = NULL,
.sv_maxssiz = NULL,
Reorganize syscall entry and leave handling. Extend struct sysvec with three new elements: sv_fetch_syscall_args - the method to fetch syscall arguments from usermode into struct syscall_args. The structure is machine-depended (this might be reconsidered after all architectures are converted). sv_set_syscall_retval - the method to set a return value for usermode from the syscall. It is a generalization of cpu_set_syscall_retval(9) to allow ABIs to override the way to set a return value. sv_syscallnames - the table of syscall names. Use sv_set_syscall_retval in kern_sigsuspend() instead of hardcoding the call to cpu_set_syscall_retval(). The new functions syscallenter(9) and syscallret(9) are provided that use sv_*syscall* pointers and contain the common repeated code from the syscall() implementations for the architecture-specific syscall trap handlers. Syscallenter() fetches arguments, calls syscall implementation from ABI sysent table, and set up return frame. The end of syscall bookkeeping is done by syscallret(). Take advantage of single place for MI syscall handling code and implement ptrace_lwpinfo pl_flags PL_FLAG_SCE, PL_FLAG_SCX and PL_FLAG_EXEC. The SCE and SCX flags notify the debugger that the thread is stopped at syscall entry or return point respectively. The EXEC flag augments SCX and notifies debugger that the process address space was changed by one of exec(2)-family syscalls. The i386, amd64, sparc64, sun4v, powerpc and ia64 syscall()s are changed to use syscallenter()/syscallret(). MIPS and arm are not converted and use the mostly unchanged syscall() implementation. Reviewed by: jhb, marcel, marius, nwhitehorn, stas Tested by: marcel (ia64), marius (sparc64), nwhitehorn (powerpc), stas (mips) MFC after: 1 month
2010-05-23 18:32:02 +00:00
.sv_flags = SV_ABI_LINUX | SV_AOUT | SV_IA32 | SV_ILP32,
.sv_set_syscall_retval = cpu_set_syscall_retval,
.sv_fetch_syscall_args = linux_fetch_syscall_args,
.sv_syscallnames = NULL,
.sv_shared_page_base = LINUX_SHAREDPAGE,
.sv_shared_page_len = PAGE_SIZE,
.sv_schedtail = linux_schedtail,
};
INIT_SYSENTVEC(aout_sysvec, &linux_sysvec);
struct sysentvec elf_linux_sysvec = {
.sv_size = LINUX_SYS_MAXSYSCALL,
.sv_table = linux_sysent,
.sv_mask = 0,
.sv_sigsize = LINUX_SIGTBLSZ,
.sv_sigtbl = bsd_to_linux_signal,
.sv_errsize = ELAST + 1,
.sv_errtbl = bsd_to_linux_errno,
.sv_transtrap = translate_traps,
.sv_fixup = elf_linux_fixup,
.sv_sendsig = linux_sendsig,
.sv_sigcode = linux_sigcode,
.sv_szsigcode = &linux_szsigcode,
Reorganize syscall entry and leave handling. Extend struct sysvec with three new elements: sv_fetch_syscall_args - the method to fetch syscall arguments from usermode into struct syscall_args. The structure is machine-depended (this might be reconsidered after all architectures are converted). sv_set_syscall_retval - the method to set a return value for usermode from the syscall. It is a generalization of cpu_set_syscall_retval(9) to allow ABIs to override the way to set a return value. sv_syscallnames - the table of syscall names. Use sv_set_syscall_retval in kern_sigsuspend() instead of hardcoding the call to cpu_set_syscall_retval(). The new functions syscallenter(9) and syscallret(9) are provided that use sv_*syscall* pointers and contain the common repeated code from the syscall() implementations for the architecture-specific syscall trap handlers. Syscallenter() fetches arguments, calls syscall implementation from ABI sysent table, and set up return frame. The end of syscall bookkeeping is done by syscallret(). Take advantage of single place for MI syscall handling code and implement ptrace_lwpinfo pl_flags PL_FLAG_SCE, PL_FLAG_SCX and PL_FLAG_EXEC. The SCE and SCX flags notify the debugger that the thread is stopped at syscall entry or return point respectively. The EXEC flag augments SCX and notifies debugger that the process address space was changed by one of exec(2)-family syscalls. The i386, amd64, sparc64, sun4v, powerpc and ia64 syscall()s are changed to use syscallenter()/syscallret(). MIPS and arm are not converted and use the mostly unchanged syscall() implementation. Reviewed by: jhb, marcel, marius, nwhitehorn, stas Tested by: marcel (ia64), marius (sparc64), nwhitehorn (powerpc), stas (mips) MFC after: 1 month
2010-05-23 18:32:02 +00:00
.sv_prepsyscall = NULL,
.sv_name = "Linux ELF",
.sv_coredump = elf32_coredump,
.sv_imgact_try = exec_linux_imgact_try,
.sv_minsigstksz = LINUX_MINSIGSTKSZ,
.sv_pagesize = PAGE_SIZE,
.sv_minuser = VM_MIN_ADDRESS,
.sv_maxuser = VM_MAXUSER_ADDRESS,
.sv_usrstack = LINUX_USRSTACK,
.sv_psstrings = LINUX_PS_STRINGS,
.sv_stackprot = VM_PROT_ALL,
.sv_copyout_strings = linux_copyout_strings,
.sv_setregs = exec_linux_setregs,
.sv_fixlimit = NULL,
.sv_maxssiz = NULL,
.sv_flags = SV_ABI_LINUX | SV_IA32 | SV_ILP32 | SV_SHP,
Reorganize syscall entry and leave handling. Extend struct sysvec with three new elements: sv_fetch_syscall_args - the method to fetch syscall arguments from usermode into struct syscall_args. The structure is machine-depended (this might be reconsidered after all architectures are converted). sv_set_syscall_retval - the method to set a return value for usermode from the syscall. It is a generalization of cpu_set_syscall_retval(9) to allow ABIs to override the way to set a return value. sv_syscallnames - the table of syscall names. Use sv_set_syscall_retval in kern_sigsuspend() instead of hardcoding the call to cpu_set_syscall_retval(). The new functions syscallenter(9) and syscallret(9) are provided that use sv_*syscall* pointers and contain the common repeated code from the syscall() implementations for the architecture-specific syscall trap handlers. Syscallenter() fetches arguments, calls syscall implementation from ABI sysent table, and set up return frame. The end of syscall bookkeeping is done by syscallret(). Take advantage of single place for MI syscall handling code and implement ptrace_lwpinfo pl_flags PL_FLAG_SCE, PL_FLAG_SCX and PL_FLAG_EXEC. The SCE and SCX flags notify the debugger that the thread is stopped at syscall entry or return point respectively. The EXEC flag augments SCX and notifies debugger that the process address space was changed by one of exec(2)-family syscalls. The i386, amd64, sparc64, sun4v, powerpc and ia64 syscall()s are changed to use syscallenter()/syscallret(). MIPS and arm are not converted and use the mostly unchanged syscall() implementation. Reviewed by: jhb, marcel, marius, nwhitehorn, stas Tested by: marcel (ia64), marius (sparc64), nwhitehorn (powerpc), stas (mips) MFC after: 1 month
2010-05-23 18:32:02 +00:00
.sv_set_syscall_retval = cpu_set_syscall_retval,
.sv_fetch_syscall_args = linux_fetch_syscall_args,
.sv_syscallnames = NULL,
.sv_shared_page_base = LINUX_SHAREDPAGE,
.sv_shared_page_len = PAGE_SIZE,
.sv_schedtail = linux_schedtail,
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
};
INIT_SYSENTVEC(elf_sysvec, &elf_linux_sysvec);
static char GNU_ABI_VENDOR[] = "GNU";
static int GNULINUX_ABI_DESC = 0;
static boolean_t
linux_trans_osrel(const Elf_Note *note, int32_t *osrel)
{
const Elf32_Word *desc;
uintptr_t p;
p = (uintptr_t)(note + 1);
p += roundup2(note->n_namesz, sizeof(Elf32_Addr));
desc = (const Elf32_Word *)p;
if (desc[0] != GNULINUX_ABI_DESC)
return (FALSE);
/*
* For linux we encode osrel as follows (see linux_mib.c):
* VVVMMMIII (version, major, minor), see linux_mib.c.
*/
*osrel = desc[1] * 1000000 + desc[2] * 1000 + desc[3];
return (TRUE);
}
static Elf_Brandnote linux_brandnote = {
.hdr.n_namesz = sizeof(GNU_ABI_VENDOR),
.hdr.n_descsz = 16, /* XXX at least 16 */
.hdr.n_type = 1,
.vendor = GNU_ABI_VENDOR,
.flags = BN_TRANSLATE_OSREL,
.trans_osrel = linux_trans_osrel
};
static Elf32_Brandinfo linux_brand = {
.brand = ELFOSABI_LINUX,
.machine = EM_386,
.compat_3_brand = "Linux",
.emul_path = "/compat/linux",
.interp_path = "/lib/ld-linux.so.1",
.sysvec = &elf_linux_sysvec,
.interp_newpath = NULL,
.brand_note = &linux_brandnote,
.flags = BI_CAN_EXEC_DYN | BI_BRAND_NOTE
};
static Elf32_Brandinfo linux_glibc2brand = {
.brand = ELFOSABI_LINUX,
.machine = EM_386,
.compat_3_brand = "Linux",
.emul_path = "/compat/linux",
.interp_path = "/lib/ld-linux.so.2",
.sysvec = &elf_linux_sysvec,
.interp_newpath = NULL,
.brand_note = &linux_brandnote,
.flags = BI_CAN_EXEC_DYN | BI_BRAND_NOTE
};
Elf32_Brandinfo *linux_brandlist[] = {
&linux_brand,
&linux_glibc2brand,
NULL
};
static int
linux_elf_modevent(module_t mod, int type, void *data)
{
Elf32_Brandinfo **brandinfo;
int error;
struct linux_ioctl_handler **lihp;
struct linux_device_handler **ldhp;
error = 0;
switch(type) {
case MOD_LOAD:
for (brandinfo = &linux_brandlist[0]; *brandinfo != NULL;
++brandinfo)
if (elf32_insert_brand_entry(*brandinfo) < 0)
error = EINVAL;
if (error == 0) {
SET_FOREACH(lihp, linux_ioctl_handler_set)
linux_ioctl_register_handler(*lihp);
SET_FOREACH(ldhp, linux_device_handler_set)
linux_device_register_handler(*ldhp);
mtx_init(&emul_lock, "emuldata lock", NULL, MTX_DEF);
sx_init(&emul_shared_lock, "emuldata->shared lock");
LIST_INIT(&futex_list);
mtx_init(&futex_mtx, "ftllk", NULL, MTX_DEF);
linux_exit_tag = EVENTHANDLER_REGISTER(process_exit, linux_proc_exit,
NULL, 1000);
linux_exec_tag = EVENTHANDLER_REGISTER(process_exec, linux_proc_exec,
NULL, 1000);
linux_get_machine(&linux_platform);
linux_szplatform = roundup(strlen(linux_platform) + 1,
sizeof(char *));
linux_osd_jail_register();
stclohz = (stathz ? stathz : hz);
if (bootverbose)
printf("Linux ELF exec handler installed\n");
} else
printf("cannot insert Linux ELF brand handler\n");
break;
case MOD_UNLOAD:
for (brandinfo = &linux_brandlist[0]; *brandinfo != NULL;
++brandinfo)
if (elf32_brand_inuse(*brandinfo))
error = EBUSY;
if (error == 0) {
for (brandinfo = &linux_brandlist[0];
*brandinfo != NULL; ++brandinfo)
if (elf32_remove_brand_entry(*brandinfo) < 0)
error = EINVAL;
}
if (error == 0) {
SET_FOREACH(lihp, linux_ioctl_handler_set)
linux_ioctl_unregister_handler(*lihp);
SET_FOREACH(ldhp, linux_device_handler_set)
linux_device_unregister_handler(*ldhp);
mtx_destroy(&emul_lock);
sx_destroy(&emul_shared_lock);
mtx_destroy(&futex_mtx);
EVENTHANDLER_DEREGISTER(process_exit, linux_exit_tag);
EVENTHANDLER_DEREGISTER(process_exec, linux_exec_tag);
linux_osd_jail_deregister();
if (bootverbose)
printf("Linux ELF exec handler removed\n");
} else
printf("Could not deinstall ELF interpreter entry\n");
break;
default:
return EOPNOTSUPP;
}
return error;
}
static moduledata_t linux_elf_mod = {
"linuxelf",
linux_elf_modevent,
0
};
DECLARE_MODULE_TIED(linuxelf, linux_elf_mod, SI_SUB_EXEC, SI_ORDER_ANY);