/*
- * Copyright (c) 2000-2005 Apple Computer, Inc. All rights reserved.
+ * Copyright (c) 2000-2007 Apple Inc. All rights reserved.
*
- * @APPLE_LICENSE_OSREFERENCE_HEADER_START@
+ * @APPLE_OSREFERENCE_LICENSE_HEADER_START@
*
- * This file contains Original Code and/or Modifications of Original Code
- * as defined in and that are subject to the Apple Public Source License
- * Version 2.0 (the 'License'). You may not use this file except in
- * compliance with the License. The rights granted to you under the
- * License may not be used to create, or enable the creation or
- * redistribution of, unlawful or unlicensed copies of an Apple operating
- * system, or to circumvent, violate, or enable the circumvention or
- * violation of, any terms of an Apple operating system software license
- * agreement.
- *
- * Please obtain a copy of the License at
- * http://www.opensource.apple.com/apsl/ and read it before using this
- * file.
- *
- * The Original Code and all software distributed under the License are
- * distributed on an 'AS IS' basis, WITHOUT WARRANTY OF ANY KIND, EITHER
- * EXPRESS OR IMPLIED, AND APPLE HEREBY DISCLAIMS ALL SUCH WARRANTIES,
- * INCLUDING WITHOUT LIMITATION, ANY WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE, QUIET ENJOYMENT OR NON-INFRINGEMENT.
- * Please see the License for the specific language governing rights and
+ * This file contains Original Code and/or Modifications of Original Code
+ * as defined in and that are subject to the Apple Public Source License
+ * Version 2.0 (the 'License'). You may not use this file except in
+ * compliance with the License. The rights granted to you under the License
+ * may not be used to create, or enable the creation or redistribution of,
+ * unlawful or unlicensed copies of an Apple operating system, or to
+ * circumvent, violate, or enable the circumvention or violation of, any
+ * terms of an Apple operating system software license agreement.
+ *
+ * Please obtain a copy of the License at
+ * http://www.opensource.apple.com/apsl/ and read it before using this file.
+ *
+ * The Original Code and all software distributed under the License are
+ * distributed on an 'AS IS' basis, WITHOUT WARRANTY OF ANY KIND, EITHER
+ * EXPRESS OR IMPLIED, AND APPLE HEREBY DISCLAIMS ALL SUCH WARRANTIES,
+ * INCLUDING WITHOUT LIMITATION, ANY WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE, QUIET ENJOYMENT OR NON-INFRINGEMENT.
+ * Please see the License for the specific language governing rights and
* limitations under the License.
- *
- * @APPLE_LICENSE_OSREFERENCE_HEADER_END@
+ *
+ * @APPLE_OSREFERENCE_LICENSE_HEADER_END@
*/
#ifdef MACH_BSD
#include <mach_rt.h>
#include <i386/tss.h>
#include <i386/user_ldt.h>
#include <i386/fpu.h>
-#include <i386/iopb_entries.h>
#include <i386/machdep_call.h>
#include <i386/misc_protos.h>
#include <i386/cpu_data.h>
#include <i386/cpu_number.h>
#include <i386/mp_desc.h>
#include <i386/vmparam.h>
+#include <i386/trap.h>
+#include <mach/i386/syscall_sw.h>
#include <sys/syscall.h>
#include <sys/kdebug.h>
-#include <sys/ktrace.h>
+#include <sys/errno.h>
#include <../bsd/sys/sysent.h>
-extern struct proc *current_proc(void);
-
kern_return_t
thread_userstack(
thread_t,
mach_vm_offset_t *
);
+void * find_user_regs(thread_t);
+
unsigned int get_msr_exportmask(void);
unsigned int get_msr_nbits(void);
void IOSleep(int);
+void thread_set_cthreadself(thread_t thread, uint64_t pself, int isLP64);
+
/*
* thread_userstack:
*
__unused thread_t thread,
int flavor,
thread_state_t tstate,
- unsigned int count,
+ __unused unsigned int count,
user_addr_t *user_stack,
int *customstack
)
{
- struct i386_saved_state *state;
- i386_thread_state_t *state25;
- vm_offset_t uesp;
-
- if (customstack)
- *customstack = 0;
-
- switch (flavor) {
- case i386_THREAD_STATE: /* FIXME */
- state25 = (i386_thread_state_t *) tstate;
- if (state25->esp)
- *user_stack = state25->esp;
- else
- *user_stack = USRSTACK;
- if (customstack && state25->esp)
- *customstack = 1;
- else
- *customstack = 0;
- break;
+ if (customstack)
+ *customstack = 0;
+
+ switch (flavor) {
+ case x86_THREAD_STATE32:
+ {
+ x86_thread_state32_t *state25;
+
+ state25 = (x86_thread_state32_t *) tstate;
+
+ if (state25->esp)
+ *user_stack = state25->esp;
+ else
+ *user_stack = VM_USRSTACK32;
+ if (customstack && state25->esp)
+ *customstack = 1;
+ else
+ *customstack = 0;
+ break;
+ }
+
+ case x86_THREAD_STATE64:
+ {
+ x86_thread_state64_t *state25;
+
+ state25 = (x86_thread_state64_t *) tstate;
+
+ if (state25->rsp)
+ *user_stack = state25->rsp;
+ else
+ *user_stack = VM_USRSTACK64;
+ if (customstack && state25->rsp)
+ *customstack = 1;
+ else
+ *customstack = 0;
+ break;
+ }
+
+ default:
+ return (KERN_INVALID_ARGUMENT);
+ }
+
+ return (KERN_SUCCESS);
+}
- case i386_NEW_THREAD_STATE:
- if (count < i386_NEW_THREAD_STATE_COUNT)
- return (KERN_INVALID_ARGUMENT);
- else {
- state = (struct i386_saved_state *) tstate;
- uesp = state->uesp;
- }
-
- /* If a valid user stack is specified, use it. */
- if (uesp)
- *user_stack = uesp;
- else
- *user_stack = USRSTACK;
- if (customstack && uesp)
- *customstack = 1;
- else
- *customstack = 0;
- break;
- default :
- return (KERN_INVALID_ARGUMENT);
- }
-
- return (KERN_SUCCESS);
-}
kern_return_t
thread_entrypoint(
__unused thread_t thread,
int flavor,
thread_state_t tstate,
- unsigned int count,
+ __unused unsigned int count,
mach_vm_offset_t *entry_point
)
{
- struct i386_saved_state *state;
- i386_thread_state_t *state25;
-
- /*
- * Set a default.
- */
- if (*entry_point == 0)
- *entry_point = VM_MIN_ADDRESS;
-
- switch (flavor) {
- case i386_THREAD_STATE:
- state25 = (i386_thread_state_t *) tstate;
- *entry_point = state25->eip ? state25->eip: VM_MIN_ADDRESS;
- break;
-
- case i386_NEW_THREAD_STATE:
- if (count < i386_THREAD_STATE_COUNT)
- return (KERN_INVALID_ARGUMENT);
- else {
- state = (struct i386_saved_state *) tstate;
-
- /*
- * If a valid entry point is specified, use it.
- */
- *entry_point = state->eip ? state->eip: VM_MIN_ADDRESS;
- }
- break;
- }
-
- return (KERN_SUCCESS);
-}
-
-struct i386_saved_state *
-get_user_regs(thread_t th)
-{
- if (th->machine.pcb)
- return(USER_REGS(th));
- else {
- printf("[get_user_regs: thread does not have pcb]");
- return NULL;
+ /*
+ * Set a default.
+ */
+ if (*entry_point == 0)
+ *entry_point = VM_MIN_ADDRESS;
+
+ switch (flavor) {
+ case x86_THREAD_STATE32:
+ {
+ x86_thread_state32_t *state25;
+
+ state25 = (i386_thread_state_t *) tstate;
+ *entry_point = state25->eip ? state25->eip: VM_MIN_ADDRESS;
+ break;
+ }
+
+ case x86_THREAD_STATE64:
+ {
+ x86_thread_state64_t *state25;
+
+ state25 = (x86_thread_state64_t *) tstate;
+ *entry_point = state25->rip ? state25->rip: VM_MIN_ADDRESS64;
+ break;
+ }
}
+ return (KERN_SUCCESS);
}
+
/*
* Duplicate parent state in child
* for U**X fork.
thread_t child
)
{
- struct i386_float_state floatregs;
-
-#ifdef XXX
- /* Save the FPU state */
- if ((pcb_t)(per_proc_info[cpu_number()].fpu_pcb) == parent->machine.pcb) {
- fp_state_save(parent);
- }
-#endif
+
+ pcb_t parent_pcb;
+ pcb_t child_pcb;
- if (child->machine.pcb == NULL || parent->machine.pcb == NULL)
+ if ((child_pcb = child->machine.pcb) == NULL ||
+ (parent_pcb = parent->machine.pcb) == NULL)
return (KERN_FAILURE);
+ /*
+ * Copy over the x86_saved_state registers
+ */
+ if (cpu_mode_is64bit()) {
+ if (thread_is_64bit(parent))
+ bcopy(USER_REGS64(parent), USER_REGS64(child), sizeof(x86_saved_state64_t));
+ else
+ bcopy(USER_REGS32(parent), USER_REGS32(child), sizeof(x86_saved_state_compat32_t));
+ } else
+ bcopy(USER_REGS32(parent), USER_REGS32(child), sizeof(x86_saved_state32_t));
- /* Copy over the i386_saved_state registers */
- child->machine.pcb->iss = parent->machine.pcb->iss;
-
- /* Check to see if parent is using floating point
+ /*
+ * Check to see if parent is using floating point
* and if so, copy the registers to the child
- * FIXME - make sure this works.
*/
+ fpu_dup_fxstate(parent, child);
- if (parent->machine.pcb->ims.ifps) {
- if (fpu_get_state(parent, &floatregs) == KERN_SUCCESS)
- fpu_set_state(child, &floatregs);
- }
-
- /* FIXME - should a user specified LDT, TSS and V86 info
+#ifdef MACH_BSD
+ /*
+ * Copy the parent's cthread id and USER_CTHREAD descriptor, if 32-bit.
+ */
+ child_pcb->cthread_self = parent_pcb->cthread_self;
+ if (!thread_is_64bit(parent))
+ child_pcb->cthread_desc = parent_pcb->cthread_desc;
+
+ /*
+ * FIXME - should a user specified LDT, TSS and V86 info
* be duplicated as well?? - probably not.
*/
- // duplicate any use LDT entry that was set I think this is appropriate.
-#ifdef MACH_BSD
- if (parent->machine.pcb->uldt_selector!= 0) {
- child->machine.pcb->uldt_selector = parent->machine.pcb->uldt_selector;
- child->machine.pcb->uldt_desc = parent->machine.pcb->uldt_desc;
- }
+ // duplicate any use LDT entry that was set I think this is appropriate.
+ if (parent_pcb->uldt_selector!= 0) {
+ child_pcb->uldt_selector = parent_pcb->uldt_selector;
+ child_pcb->uldt_desc = parent_pcb->uldt_desc;
+ }
#endif
-
-
+
return (KERN_SUCCESS);
}
void
thread_set_child(thread_t child, int pid)
{
- child->machine.pcb->iss.eax = pid;
- child->machine.pcb->iss.edx = 1;
- child->machine.pcb->iss.efl &= ~EFL_CF;
-}
-void thread_set_parent(thread_t parent, int pid);
-void
-thread_set_parent(thread_t parent, int pid)
-{
- parent->machine.pcb->iss.eax = pid;
- parent->machine.pcb->iss.edx = 0;
- parent->machine.pcb->iss.efl &= ~EFL_CF;
-}
-
-
-
-/*
- * System Call handling code
- */
-
-#define ERESTART -1 /* restart syscall */
-#define EJUSTRETURN -2 /* don't modify regs, just return */
+ if (thread_is_64bit(child)) {
+ x86_saved_state64_t *iss64;
+ iss64 = USER_REGS64(child);
-#define NO_FUNNEL 0
-#define KERNEL_FUNNEL 1
+ iss64->rax = pid;
+ iss64->rdx = 1;
+ iss64->isf.rflags &= ~EFL_CF;
+ } else {
+ x86_saved_state32_t *iss32;
-extern funnel_t * kernel_flock;
+ iss32 = USER_REGS32(child);
-extern int set_bsduthreadargs (thread_t, struct i386_saved_state *, void *);
-extern void * get_bsduthreadarg(thread_t);
-extern int * get_bsduthreadrval(thread_t th);
-extern int * get_bsduthreadlowpridelay(thread_t th);
+ iss32->eax = pid;
+ iss32->edx = 1;
+ iss32->efl &= ~EFL_CF;
+ }
+}
-extern long fuword(vm_offset_t);
-extern void unix_syscall(struct i386_saved_state *);
-extern void unix_syscall_return(int);
+void thread_set_parent(thread_t parent, int pid);
-/* following implemented in bsd/dev/i386/unix_signal.c */
-int __pthread_cset(struct sysent *);
+void
+thread_set_parent(thread_t parent, int pid)
+{
+ if (thread_is_64bit(parent)) {
+ x86_saved_state64_t *iss64;
-void __pthread_creset(struct sysent *);
+ iss64 = USER_REGS64(parent);
+ iss64->rax = pid;
+ iss64->rdx = 0;
+ iss64->isf.rflags &= ~EFL_CF;
+ } else {
+ x86_saved_state32_t *iss32;
-void
-unix_syscall_return(int error)
-{
- thread_t thread;
- volatile int *rval;
- struct i386_saved_state *regs;
- struct proc *p;
- unsigned short code;
- vm_offset_t params;
- struct sysent *callp;
- volatile int *lowpri_delay;
-
- thread = current_thread();
- rval = get_bsduthreadrval(thread);
- lowpri_delay = get_bsduthreadlowpridelay(thread);
- p = current_proc();
-
- regs = USER_REGS(thread);
-
- /* reconstruct code for tracing before blasting eax */
- code = regs->eax;
- params = (vm_offset_t) ((caddr_t)regs->uesp + sizeof (int));
- callp = (code >= nsysent) ? &sysent[63] : &sysent[code];
- if (callp == sysent) {
- code = fuword(params);
- }
+ iss32 = USER_REGS32(parent);
- if (error == ERESTART) {
- regs->eip -= 7;
- }
- else if (error != EJUSTRETURN) {
- if (error) {
- regs->eax = error;
- regs->efl |= EFL_CF; /* carry bit */
- } else { /* (not error) */
- regs->eax = rval[0];
- regs->edx = rval[1];
- regs->efl &= ~EFL_CF;
- }
+ iss32->eax = pid;
+ iss32->edx = 0;
+ iss32->efl &= ~EFL_CF;
}
+}
- ktrsysret(p, code, error, rval[0], (callp->sy_funnel & FUNNEL_MASK));
-
- __pthread_creset(callp);
- if ((callp->sy_funnel & FUNNEL_MASK) != NO_FUNNEL)
- (void) thread_funnel_set(current_thread()->funnel_lock, FALSE);
+/*
+ * System Call handling code
+ */
- if (*lowpri_delay) {
- /*
- * task is marked as a low priority I/O type
- * and the I/O we issued while in this system call
- * collided with normal I/O operations... we'll
- * delay in order to mitigate the impact of this
- * task on the normal operation of the system
- */
- IOSleep(*lowpri_delay);
- *lowpri_delay = 0;
- }
- KERNEL_DEBUG_CONSTANT(BSDDBG_CODE(DBG_BSD_EXCP_SC, code) | DBG_FUNC_END,
- error, rval[0], rval[1], 0, 0);
+extern long fuword(vm_offset_t);
- thread_exception_return();
- /* NOTREACHED */
-}
void
-unix_syscall(struct i386_saved_state *regs)
+machdep_syscall(x86_saved_state_t *state)
{
- thread_t thread;
- void *vt;
- unsigned short code;
- struct sysent *callp;
- int nargs;
- int error;
- int *rval;
- int funnel_type;
- vm_offset_t params;
- struct proc *p;
- volatile int *lowpri_delay;
-
- thread = current_thread();
- p = current_proc();
- rval = get_bsduthreadrval(thread);
- lowpri_delay = get_bsduthreadlowpridelay(thread);
-
- thread->task->syscalls_unix++; /* MP-safety ignored */
-
- //printf("[scall : eax %x]", regs->eax);
- code = regs->eax;
- params = (vm_offset_t) ((caddr_t)regs->uesp + sizeof (int));
- callp = (code >= nsysent) ? &sysent[63] : &sysent[code];
- if (callp == sysent) {
- code = fuword(params);
- params += sizeof (int);
- callp = (code >= nsysent) ? &sysent[63] : &sysent[code];
- }
+ int args[machdep_call_count];
+ int trapno;
+ int nargs;
+ machdep_call_t *entry;
+ x86_saved_state32_t *regs;
+
+ assert(is_saved_state32(state));
+ regs = saved_state32(state);
- vt = get_bsduthreadarg(thread);
+ trapno = regs->eax;
+#if DEBUG_TRACE
+ kprintf("machdep_syscall(0x%08x) code=%d\n", regs, trapno);
+#endif
- if ((nargs = (callp->sy_narg * sizeof (int))) &&
- (error = copyin((user_addr_t) params, (char *) vt, nargs)) != 0) {
- regs->eax = error;
- regs->efl |= EFL_CF;
- thread_exception_return();
- /* NOTREACHED */
- }
-
- rval[0] = 0;
- rval[1] = regs->edx;
+ if (trapno < 0 || trapno >= machdep_call_count) {
+ regs->eax = (unsigned int)kern_invalid(NULL);
- if ((error = __pthread_cset(callp))) {
- /* cancelled system call; let it returned with EINTR for handling */
- regs->eax = error;
- regs->efl |= EFL_CF;
thread_exception_return();
/* NOTREACHED */
}
+ entry = &machdep_call_table[trapno];
+ nargs = entry->nargs;
- funnel_type = (callp->sy_funnel & FUNNEL_MASK);
- if(funnel_type == KERNEL_FUNNEL)
- (void) thread_funnel_set(kernel_flock, TRUE);
-
- (void) set_bsduthreadargs(thread, regs, NULL);
-
- if (callp->sy_narg > 8)
- panic("unix_syscall max arg count exceeded (%d)", callp->sy_narg);
+ if (nargs != 0) {
+ if (copyin((user_addr_t) regs->uesp + sizeof (int),
+ (char *) args, (nargs * sizeof (int)))) {
+ regs->eax = KERN_INVALID_ADDRESS;
- ktrsyscall(p, code, callp->sy_narg, vt, funnel_type);
-
- {
- int *ip = (int *)vt;
- KERNEL_DEBUG_CONSTANT(BSDDBG_CODE(DBG_BSD_EXCP_SC, code) | DBG_FUNC_START,
- *ip, *(ip+1), *(ip+2), *(ip+3), 0);
- }
-
- error = (*(callp->sy_call))((void *) p, (void *) vt, &rval[0]);
-
-#if 0
- /* May be needed with vfork changes */
- regs = USER_REGS(thread);
-#endif
- if (error == ERESTART) {
- regs->eip -= 7;
+ thread_exception_return();
+ /* NOTREACHED */
+ }
}
- else if (error != EJUSTRETURN) {
- if (error) {
- regs->eax = error;
- regs->efl |= EFL_CF; /* carry bit */
- } else { /* (not error) */
- regs->eax = rval[0];
- regs->edx = rval[1];
- regs->efl &= ~EFL_CF;
- }
- }
-
- ktrsysret(p, code, error, rval[0], funnel_type);
-
- __pthread_creset(callp);
-
- if(funnel_type != NO_FUNNEL)
- (void) thread_funnel_set(current_thread()->funnel_lock, FALSE);
+ switch (nargs) {
+ case 0:
+ regs->eax = (*entry->routine.args_0)();
+ break;
+ case 1:
+ regs->eax = (*entry->routine.args_1)(args[0]);
+ break;
+ case 2:
+ regs->eax = (*entry->routine.args_2)(args[0],args[1]);
+ break;
+ case 3:
+ if (!entry->bsd_style)
+ regs->eax = (*entry->routine.args_3)(args[0],args[1],args[2]);
+ else {
+ int error;
+ uint32_t rval;
+
+ error = (*entry->routine.args_bsd_3)(&rval, args[0], args[1], args[2]);
+ if (error) {
+ regs->eax = error;
+ regs->efl |= EFL_CF; /* carry bit */
+ } else {
+ regs->eax = rval;
+ regs->efl &= ~EFL_CF;
+ }
+ }
+ break;
+ case 4:
+ regs->eax = (*entry->routine.args_4)(args[0], args[1], args[2], args[3]);
+ break;
- if (*lowpri_delay) {
- /*
- * task is marked as a low priority I/O type
- * and the I/O we issued while in this system call
- * collided with normal I/O operations... we'll
- * delay in order to mitigate the impact of this
- * task on the normal operation of the system
- */
- IOSleep(*lowpri_delay);
- *lowpri_delay = 0;
+ default:
+ panic("machdep_syscall: too many args");
}
- KERNEL_DEBUG_CONSTANT(BSDDBG_CODE(DBG_BSD_EXCP_SC, code) | DBG_FUNC_END,
- error, rval[0], rval[1], 0, 0);
+ if (current_thread()->funnel_lock)
+ (void) thread_funnel_set(current_thread()->funnel_lock, FALSE);
- thread_exception_return();
- /* NOTREACHED */
+ thread_exception_return();
+ /* NOTREACHED */
}
void
-machdep_syscall( struct i386_saved_state *regs)
+machdep_syscall64(x86_saved_state_t *state)
{
- int trapno, nargs;
- machdep_call_t *entry;
-
- trapno = regs->eax;
- if (trapno < 0 || trapno >= machdep_call_count) {
- regs->eax = (unsigned int)kern_invalid(NULL);
+ int trapno;
+ machdep_call_t *entry;
+ x86_saved_state64_t *regs;
- thread_exception_return();
- /* NOTREACHED */
- }
+ assert(is_saved_state64(state));
+ regs = saved_state64(state);
- entry = &machdep_call_table[trapno];
- nargs = entry->nargs;
-
- if (nargs > 0) {
- int args[nargs];
+ trapno = regs->rax & SYSCALL_NUMBER_MASK;
- if (copyin((user_addr_t) regs->uesp + sizeof (int),
- (char *) args,
- nargs * sizeof (int))) {
+ if (trapno < 0 || trapno >= machdep_call_count) {
+ regs->rax = (unsigned int)kern_invalid(NULL);
- regs->eax = KERN_INVALID_ADDRESS;
-
- thread_exception_return();
- /* NOTREACHED */
+ thread_exception_return();
+ /* NOTREACHED */
}
+ entry = &machdep_call_table64[trapno];
- switch (nargs) {
- case 1:
- regs->eax = (*entry->routine.args_1)(args[0]);
- break;
- case 2:
- regs->eax = (*entry->routine.args_2)(args[0],args[1]);
- break;
- case 3:
- regs->eax = (*entry->routine.args_3)(args[0],args[1],args[2]);
+ switch (entry->nargs) {
+ case 0:
+ regs->rax = (*entry->routine.args_0)();
break;
- case 4:
- regs->eax = (*entry->routine.args_4)(args[0],args[1],args[2],args[3]);
+ case 1:
+ regs->rax = (*entry->routine.args64_1)(regs->rdi);
break;
- default:
- panic("machdep_syscall(): too many args");
+ default:
+ panic("machdep_syscall64: too many args");
}
- }
- else
- regs->eax = (*entry->routine.args_0)();
+ if (current_thread()->funnel_lock)
+ (void) thread_funnel_set(current_thread()->funnel_lock, FALSE);
- if (current_thread()->funnel_lock)
- (void) thread_funnel_set(current_thread()->funnel_lock, FALSE);
-
- thread_exception_return();
- /* NOTREACHED */
+ thread_exception_return();
+ /* NOTREACHED */
}
kern_return_t
thread_set_cthread_self(uint32_t self)
{
- current_thread()->machine.pcb->cthread_self = self;
-
- return (KERN_SUCCESS);
+ current_thread()->machine.pcb->cthread_self = (uint64_t) self;
+
+ return (KERN_SUCCESS);
}
kern_return_t
thread_get_cthread_self(void)
{
- return ((kern_return_t)current_thread()->machine.pcb->cthread_self);
+ return ((kern_return_t)current_thread()->machine.pcb->cthread_self);
}
kern_return_t
thread_fast_set_cthread_self(uint32_t self)
{
- pcb_t pcb;
- pcb = (pcb_t)current_thread()->machine.pcb;
- thread_compose_cthread_desc(self, pcb);
- pcb->cthread_self = self; /* preserve old func too */
- return (USER_CTHREAD);
+ pcb_t pcb;
+ x86_saved_state32_t *iss;
+
+ pcb = (pcb_t)current_thread()->machine.pcb;
+ thread_compose_cthread_desc(self, pcb);
+ pcb->cthread_self = (uint64_t) self; /* preserve old func too */
+ iss = saved_state32(pcb->iss);
+ iss->gs = USER_CTHREAD;
+
+ return (USER_CTHREAD);
+}
+
+void
+thread_set_cthreadself(thread_t thread, uint64_t pself, int isLP64)
+{
+ if (isLP64 == 0) {
+ pcb_t pcb;
+ x86_saved_state32_t *iss;
+
+ pcb = (pcb_t)thread->machine.pcb;
+ thread_compose_cthread_desc(pself, pcb);
+ pcb->cthread_self = (uint64_t) pself; /* preserve old func too */
+ iss = saved_state32(pcb->iss);
+ iss->gs = USER_CTHREAD;
+ } else {
+ pcb_t pcb;
+ x86_saved_state64_t *iss;
+
+ pcb = thread->machine.pcb;
+
+ /* check for canonical address, set 0 otherwise */
+ if (!IS_USERADDR64_CANONICAL(pself))
+ pself = 0ULL;
+ pcb->cthread_self = pself;
+
+ /* XXX for 64-in-32 */
+ iss = saved_state64(pcb->iss);
+ iss->gs = USER_CTHREAD;
+ thread_compose_cthread_desc((uint32_t) pself, pcb);
+ }
+}
+
+
+kern_return_t
+thread_fast_set_cthread_self64(uint64_t self)
+{
+ pcb_t pcb;
+ x86_saved_state64_t *iss;
+
+ pcb = current_thread()->machine.pcb;
+
+ /* check for canonical address, set 0 otherwise */
+ if (!IS_USERADDR64_CANONICAL(self))
+ self = 0ULL;
+ pcb->cthread_self = self;
+ current_cpu_datap()->cpu_uber.cu_user_gs_base = self;
+
+ /* XXX for 64-in-32 */
+ iss = saved_state64(pcb->iss);
+ iss->gs = USER_CTHREAD;
+ thread_compose_cthread_desc((uint32_t) self, pcb);
+
+ return (USER_CTHREAD);
}
/*
kern_return_t
thread_set_user_ldt(uint32_t address, uint32_t size, uint32_t flags)
{
- pcb_t pcb;
- struct fake_descriptor temp;
- int mycpu;
+ pcb_t pcb;
+ struct fake_descriptor temp;
+ int mycpu;
- if (flags != 0)
- return -1; // flags not supported
- if (size > 0xFFFFF)
- return -1; // size too big, 1 meg is the limit
+ if (flags != 0)
+ return -1; // flags not supported
+ if (size > 0xFFFFF)
+ return -1; // size too big, 1 meg is the limit
- mp_disable_preemption();
- mycpu = cpu_number();
-
- // create a "fake" descriptor so we can use fix_desc()
- // to build a real one...
- // 32 bit default operation size
- // standard read/write perms for a data segment
- pcb = (pcb_t)current_thread()->machine.pcb;
- temp.offset = address;
- temp.lim_or_seg = size;
- temp.size_or_wdct = SZ_32;
- temp.access = ACC_P|ACC_PL_U|ACC_DATA_W;
+ mp_disable_preemption();
+ mycpu = cpu_number();
- // turn this into a real descriptor
- fix_desc(&temp,1);
+ // create a "fake" descriptor so we can use fix_desc()
+ // to build a real one...
+ // 32 bit default operation size
+ // standard read/write perms for a data segment
+ pcb = (pcb_t)current_thread()->machine.pcb;
+ temp.offset = address;
+ temp.lim_or_seg = size;
+ temp.size_or_wdct = SZ_32;
+ temp.access = ACC_P|ACC_PL_U|ACC_DATA_W;
- // set up our data in the pcb
- pcb->uldt_desc = *(struct real_descriptor*)&temp;
- pcb->uldt_selector = USER_SETTABLE; // set the selector value
+ // turn this into a real descriptor
+ fix_desc(&temp,1);
- // now set it up in the current table...
- *ldt_desc_p(USER_SETTABLE) = *(struct real_descriptor*)&temp;
+ // set up our data in the pcb
+ pcb->uldt_desc = *(struct real_descriptor*)&temp;
+ pcb->uldt_selector = USER_SETTABLE; // set the selector value
- mp_enable_preemption();
+ // now set it up in the current table...
+ *ldt_desc_p(USER_SETTABLE) = *(struct real_descriptor*)&temp;
- return USER_SETTABLE;
-}
-void
-mach25_syscall(struct i386_saved_state *regs)
-{
- printf("*** Atttempt to execute a Mach 2.5 system call at EIP=%x EAX=%x(%d)\n",
- regs->eip, regs->eax, -regs->eax);
- panic("FIXME!");
+ mp_enable_preemption();
+
+ return USER_SETTABLE;
}
+
#endif /* MACH_BSD */
-/* This routine is called from assembly before each and every mach trap.
- */
+typedef kern_return_t (*mach_call_t)(void *);
-extern unsigned int mach_call_start(unsigned int, unsigned int *);
+struct mach_call_args {
+ syscall_arg_t arg1;
+ syscall_arg_t arg2;
+ syscall_arg_t arg3;
+ syscall_arg_t arg4;
+ syscall_arg_t arg5;
+ syscall_arg_t arg6;
+ syscall_arg_t arg7;
+ syscall_arg_t arg8;
+ syscall_arg_t arg9;
+};
-__private_extern__
-unsigned int
-mach_call_start(unsigned int call_number, unsigned int *args)
-{
- int i, argc;
- unsigned int kdarg[3];
+static kern_return_t
+mach_call_arg_munger32(uint32_t sp, int nargs, int call_number, struct mach_call_args *args);
- current_thread()->task->syscalls_mach++; /* MP-safety ignored */
-/* Always prepare to trace mach system calls */
+static kern_return_t
+mach_call_arg_munger32(uint32_t sp, int nargs, int call_number, struct mach_call_args *args)
+{
+ unsigned int args32[9];
- kdarg[0]=0;
- kdarg[1]=0;
- kdarg[2]=0;
+ if (copyin((user_addr_t)(sp + sizeof(int)), (char *)args32, nargs * sizeof (int)))
+ return KERN_INVALID_ARGUMENT;
- argc = mach_trap_table[call_number>>4].mach_trap_arg_count;
-
- if (argc > 3)
- argc = 3;
-
- for (i=0; i < argc; i++)
- kdarg[i] = (int)*(args + i);
-
- KERNEL_DEBUG_CONSTANT(MACHDBG_CODE(DBG_MACH_EXCP_SC, (call_number>>4)) | DBG_FUNC_START,
- kdarg[0], kdarg[1], kdarg[2], 0, 0);
+ switch (nargs) {
+ case 9: args->arg9 = args32[8];
+ case 8: args->arg8 = args32[7];
+ case 7: args->arg7 = args32[6];
+ case 6: args->arg6 = args32[5];
+ case 5: args->arg5 = args32[4];
+ case 4: args->arg4 = args32[3];
+ case 3: args->arg3 = args32[2];
+ case 2: args->arg2 = args32[1];
+ case 1: args->arg1 = args32[0];
+ }
+ if (call_number == 90) {
+ /* munge_l for mach_wait_until_trap() */
+ args->arg1 = (((uint64_t)(args32[0])) | ((((uint64_t)(args32[1]))<<32)));
+ }
+ if (call_number == 93) {
+ /* munge_wl for mk_timer_arm_trap() */
+ args->arg2 = (((uint64_t)(args32[1])) | ((((uint64_t)(args32[2]))<<32)));
+ }
- return call_number; /* pass this back thru */
+ return KERN_SUCCESS;
}
-/* This routine is called from assembly after each mach system call
- */
-extern unsigned int mach_call_end(unsigned int, unsigned int);
+__private_extern__ void mach_call_munger(x86_saved_state_t *state);
-__private_extern__
-unsigned int
-mach_call_end(unsigned int call_number, unsigned int retval)
+void
+mach_call_munger(x86_saved_state_t *state)
{
- KERNEL_DEBUG_CONSTANT(MACHDBG_CODE(DBG_MACH_EXCP_SC,(call_number>>4)) | DBG_FUNC_END,
- retval, 0, 0, 0, 0);
- return retval; /* pass this back thru */
+ int argc;
+ int call_number;
+ mach_call_t mach_call;
+ kern_return_t retval;
+ struct mach_call_args args = { 0, 0, 0, 0, 0, 0, 0, 0, 0 };
+ x86_saved_state32_t *regs;
+
+ assert(is_saved_state32(state));
+ regs = saved_state32(state);
+
+ call_number = -(regs->eax);
+#if DEBUG_TRACE
+ kprintf("mach_call_munger(0x%08x) code=%d\n", regs, call_number);
+#endif
+
+ if (call_number < 0 || call_number >= mach_trap_count) {
+ i386_exception(EXC_SYSCALL, call_number, 1);
+ /* NOTREACHED */
+ }
+ mach_call = (mach_call_t)mach_trap_table[call_number].mach_trap_function;
+
+ if (mach_call == (mach_call_t)kern_invalid) {
+ i386_exception(EXC_SYSCALL, call_number, 1);
+ /* NOTREACHED */
+ }
+
+ argc = mach_trap_table[call_number].mach_trap_arg_count;
+ if (argc) {
+ retval = mach_call_arg_munger32(regs->uesp, argc, call_number, &args);
+ if (retval != KERN_SUCCESS) {
+ regs->eax = retval;
+
+ thread_exception_return();
+ /* NOTREACHED */
+ }
+ }
+ KERNEL_DEBUG_CONSTANT(MACHDBG_CODE(DBG_MACH_EXCP_SC, (call_number)) | DBG_FUNC_START,
+ (int) args.arg1, (int) args.arg2, (int) args.arg3, (int) args.arg4, 0);
+
+ retval = mach_call(&args);
+
+ KERNEL_DEBUG_CONSTANT(MACHDBG_CODE(DBG_MACH_EXCP_SC,(call_number)) | DBG_FUNC_END,
+ retval, 0, 0, 0, 0);
+ regs->eax = retval;
+
+ thread_exception_return();
+ /* NOTREACHED */
}
-typedef kern_return_t (*mach_call_t)(void *);
-extern __attribute__((regparm(1))) kern_return_t
-mach_call_munger(unsigned int call_number,
- unsigned int arg1,
- unsigned int arg2,
- unsigned int arg3,
- unsigned int arg4,
- unsigned int arg5,
- unsigned int arg6,
- unsigned int arg7,
- unsigned int arg8,
- unsigned int arg9
-);
+__private_extern__ void mach_call_munger64(x86_saved_state_t *regs);
-struct mach_call_args {
- unsigned int arg1;
- unsigned int arg2;
- unsigned int arg3;
- unsigned int arg4;
- unsigned int arg5;
- unsigned int arg6;
- unsigned int arg7;
- unsigned int arg8;
- unsigned int arg9;
-};
-__private_extern__
-__attribute__((regparm(1))) kern_return_t
-mach_call_munger(unsigned int call_number,
- unsigned int arg1,
- unsigned int arg2,
- unsigned int arg3,
- unsigned int arg4,
- unsigned int arg5,
- unsigned int arg6,
- unsigned int arg7,
- unsigned int arg8,
- unsigned int arg9
-)
+void
+mach_call_munger64(x86_saved_state_t *state)
{
+ int call_number;
int argc;
mach_call_t mach_call;
- kern_return_t retval;
- struct mach_call_args args = { 0, 0, 0, 0, 0, 0, 0, 0, 0 };
-
- current_thread()->task->syscalls_mach++; /* MP-safety ignored */
- call_number >>= 4;
+ x86_saved_state64_t *regs;
+
+ assert(is_saved_state64(state));
+ regs = saved_state64(state);
+
+ call_number = regs->rax & SYSCALL_NUMBER_MASK;
+
+ KERNEL_DEBUG_CONSTANT(MACHDBG_CODE(DBG_MACH_EXCP_SC,
+ (call_number)) | DBG_FUNC_START,
+ (int) regs->rdi, (int) regs->rsi,
+ (int) regs->rdx, (int) regs->r10, 0);
- argc = mach_trap_table[call_number].mach_trap_arg_count;
- switch (argc) {
- case 9: args.arg9 = arg9;
- case 8: args.arg8 = arg8;
- case 7: args.arg7 = arg7;
- case 6: args.arg6 = arg6;
- case 5: args.arg5 = arg5;
- case 4: args.arg4 = arg4;
- case 3: args.arg3 = arg3;
- case 2: args.arg2 = arg2;
- case 1: args.arg1 = arg1;
+ if (call_number < 0 || call_number >= mach_trap_count) {
+ i386_exception(EXC_SYSCALL, regs->rax, 1);
+ /* NOTREACHED */
}
-
- KERNEL_DEBUG_CONSTANT(MACHDBG_CODE(DBG_MACH_EXCP_SC, (call_number)) | DBG_FUNC_START,
- args.arg1, args.arg2, args.arg3, 0, 0);
-
mach_call = (mach_call_t)mach_trap_table[call_number].mach_trap_function;
- retval = mach_call(&args);
- KERNEL_DEBUG_CONSTANT(MACHDBG_CODE(DBG_MACH_EXCP_SC,(call_number)) | DBG_FUNC_END,
- retval, 0, 0, 0, 0);
+ if (mach_call == (mach_call_t)kern_invalid) {
+ i386_exception(EXC_SYSCALL, regs->rax, 1);
+ /* NOTREACHED */
+ }
+ argc = mach_trap_table[call_number].mach_trap_arg_count;
+
+ if (argc > 6) {
+ int copyin_count;
- return retval;
+ copyin_count = (argc - 6) * sizeof(uint64_t);
+
+ if (copyin((user_addr_t)(regs->isf.rsp + sizeof(user_addr_t)), (char *)®s->v_arg6, copyin_count)) {
+ regs->rax = KERN_INVALID_ARGUMENT;
+
+ thread_exception_return();
+ /* NOTREACHED */
+ }
+ }
+ regs->rax = (uint64_t)mach_call((void *)(®s->rdi));
+
+ KERNEL_DEBUG_CONSTANT(MACHDBG_CODE(DBG_MACH_EXCP_SC,
+ (call_number)) | DBG_FUNC_END,
+ (int)regs->rax, 0, 0, 0, 0);
+
+ thread_exception_return();
+ /* NOTREACHED */
}
+
/*
* thread_setuserstack:
*
thread_t thread,
mach_vm_address_t user_stack)
{
- struct i386_saved_state *ss = get_user_regs(thread);
+ if (thread_is_64bit(thread)) {
+ x86_saved_state64_t *iss64;
+
+ iss64 = USER_REGS64(thread);
+
+ iss64->isf.rsp = (uint64_t)user_stack;
+ } else {
+ x86_saved_state32_t *iss32;
+
+ iss32 = USER_REGS32(thread);
- ss->uesp = CAST_DOWN(unsigned int,user_stack);
+ iss32->uesp = CAST_DOWN(unsigned int, user_stack);
+ }
}
/*
thread_t thread,
int adjust)
{
- struct i386_saved_state *ss = get_user_regs(thread);
+ if (thread_is_64bit(thread)) {
+ x86_saved_state64_t *iss64;
+
+ iss64 = USER_REGS64(thread);
+
+ iss64->isf.rsp += adjust;
+
+ return iss64->isf.rsp;
+ } else {
+ x86_saved_state32_t *iss32;
- ss->uesp += adjust;
- return CAST_USER_ADDR_T(ss->uesp);
+ iss32 = USER_REGS32(thread);
+
+ iss32->uesp += adjust;
+
+ return CAST_USER_ADDR_T(iss32->uesp);
+ }
}
/*
* dependent thread state info.
*/
void
-thread_setentrypoint(
- thread_t thread,
- mach_vm_address_t entry)
+thread_setentrypoint(thread_t thread, mach_vm_address_t entry)
{
- struct i386_saved_state *ss = get_user_regs(thread);
+ if (thread_is_64bit(thread)) {
+ x86_saved_state64_t *iss64;
+
+ iss64 = USER_REGS64(thread);
+
+ iss64->isf.rip = (uint64_t)entry;
+ } else {
+ x86_saved_state32_t *iss32;
+
+ iss32 = USER_REGS32(thread);
+
+ iss32->eip = CAST_DOWN(unsigned int, entry);
+ }
+}
+
+
+kern_return_t
+thread_setsinglestep(thread_t thread, int on)
+{
+ if (thread_is_64bit(thread)) {
+ x86_saved_state64_t *iss64;
+
+ iss64 = USER_REGS64(thread);
+
+ if (on)
+ iss64->isf.rflags |= EFL_TF;
+ else
+ iss64->isf.rflags &= ~EFL_TF;
+ } else {
+ x86_saved_state32_t *iss32;
- ss->eip = CAST_DOWN(unsigned int,entry);
-}
+ iss32 = USER_REGS32(thread);
+ if (on)
+ iss32->efl |= EFL_TF;
+ else
+ iss32->efl &= ~EFL_TF;
+ }
+
+ return (KERN_SUCCESS);
+}
+
+
+
+/* XXX this should be a struct savearea so that CHUD will work better on x86 */
+void *
+find_user_regs(thread_t thread)
+{
+ return USER_STATE(thread);
+}
+
+void *
+get_user_regs(thread_t th)
+{
+ if (th->machine.pcb)
+ return(USER_STATE(th));
+ else {
+ printf("[get_user_regs: thread does not have pcb]");
+ return NULL;
+ }
+}
+
+#if CONFIG_DTRACE
+/*
+ * DTrace would like to have a peek at the kernel interrupt state, if available.
+ * Based on osfmk/chud/i386/chud_thread_i386.c:chudxnu_thread_get_state(), which see.
+ */
+x86_saved_state32_t *find_kern_regs(thread_t);
+
+x86_saved_state32_t *
+find_kern_regs(thread_t thread)
+{
+ if (thread == current_thread() &&
+ NULL != current_cpu_datap()->cpu_int_state &&
+ !(USER_STATE(thread) == current_cpu_datap()->cpu_int_state &&
+ current_cpu_datap()->cpu_interrupt_level == 1)) {
+
+ return saved_state32(current_cpu_datap()->cpu_int_state);
+ } else {
+ return NULL;
+ }
+}
+
+vm_offset_t dtrace_get_cpu_int_stack_top(void);
+
+vm_offset_t
+dtrace_get_cpu_int_stack_top(void)
+{
+ return current_cpu_datap()->cpu_int_stack_top;
+}
+#endif