Fix an alpha inheritance bug:

On alpha, PAL is involved in context management and after wiring
the CPU (in alpha_init()) a context switch was performed to tell
PAL about the context. This was bogusly brought over to ia64
where it introduced bugs, because we restored the context from
a mostly uninitialized PCB.

The cleanup constitutes:
o  Remove the unused arguments from ia64_init().
o  Don't return from ia64_init(), but instead call mi_startup()
   directly. This reduces the amount of muckery in assembly and
   also allows for the next bullet:
o  Save our currect context prior to calling mi_startup(). The
   reason for this is that many threads are created from thread0
   by cloning the PCB. By saving our context in the PCB, we have
   something sane to clone. It also ensures that a cloned thread
   that does not alter the context in any way will return to
   the saved context, where we're ready for the eventuality with
   a nice, user unfriendly panic().

The cleanup fixes at least the following bugs:
o  Entering mi_startup() with the RSE in enforced lazy mode.
o  Re-execution of ia64_init() in certain "lab" conditions.

While here, add proper unwind directives to __start() so that
the unwind knows it has reached the bottom of the (call) stack.

Approved by: re@ (blanket)

sys/ia64/ia64/locore.S

@@ -51,6 +51,9 @@ kstack:	.space KSTACK_PAGES * PAGE_SIZE
  * register r8.
  */
 ENTRY(__start, 1)
+	.prologue
+	.save	rp,r0
+	.body
 {	.mlx
 	mov	ar.rsc=0
 	movl	r16=ia64_vector_table	// set up IVT early
@@ -90,41 +93,10 @@ ENTRY(__start, 1)
 	br.call.sptk.many rp=_reloc
 	;;
 	br.call.sptk.many rp=ia64_init
-
-	/*
-	 * switch to thread0 and then initialise the rest of the kernel.
-	 */
-	alloc	r16=ar.pfs,0,0,1,0
 	;;
-	movl	out0=thread0
-	;;
-	add	out0=TD_PCB,out0
-	;;
-	ld8	out0=[out0]
-	;; 
-	add	r16=PCB_SPECIAL_RP,out0	// return to mi_startup_trampoline
-	movl	r17=mi_startup_trampoline
-	;;
-	st8	[r16]=r17
-	;;
-	br.call.sptk.many rp=restorectx
-
 	/* NOTREACHED */
-	
-END(__start)
-
-
-ENTRY(mi_startup_trampoline, 0)
-	.prologue
-	.save	rp,r0
-	.body
-
-	br.call.sptk.many rp=mi_startup
-
-	// Should never happen
 1:	br.cond.sptk.few 1b
-
+END(__start)
-END(mi_startup_trampoline)
 
 /*
  * fork_trampoline()

sys/ia64/ia64/locore.s

@@ -51,6 +51,9 @@ kstack:	.space KSTACK_PAGES * PAGE_SIZE
  * register r8.
  */
 ENTRY(__start, 1)
+	.prologue
+	.save	rp,r0
+	.body
 {	.mlx
 	mov	ar.rsc=0
 	movl	r16=ia64_vector_table	// set up IVT early
@@ -90,41 +93,10 @@ ENTRY(__start, 1)
 	br.call.sptk.many rp=_reloc
 	;;
 	br.call.sptk.many rp=ia64_init
-
-	/*
-	 * switch to thread0 and then initialise the rest of the kernel.
-	 */
-	alloc	r16=ar.pfs,0,0,1,0
 	;;
-	movl	out0=thread0
-	;;
-	add	out0=TD_PCB,out0
-	;;
-	ld8	out0=[out0]
-	;; 
-	add	r16=PCB_SPECIAL_RP,out0	// return to mi_startup_trampoline
-	movl	r17=mi_startup_trampoline
-	;;
-	st8	[r16]=r17
-	;;
-	br.call.sptk.many rp=restorectx
-
 	/* NOTREACHED */
-	
-END(__start)
-
-
-ENTRY(mi_startup_trampoline, 0)
-	.prologue
-	.save	rp,r0
-	.body
-
-	br.call.sptk.many rp=mi_startup
-
-	// Should never happen
 1:	br.cond.sptk.few 1b
-
+END(__start)
-END(mi_startup_trampoline)
 
 /*
  * fork_trampoline()

sys/ia64/ia64/machdep.c

@@ -149,6 +149,8 @@ vm_offset_t phys_avail[100];
 /* must be 2 less so 0 0 can signal end of chunks */
 #define PHYS_AVAIL_ARRAY_END ((sizeof(phys_avail) / sizeof(vm_offset_t)) - 2)
 
+void mi_startup(void);		/* XXX should be in a MI header */
+
 static void identifycpu(void);
 
 struct kva_md_info kmi;
@@ -441,7 +443,7 @@ calculate_frequencies(void)
 }
 
 void
-ia64_init(u_int64_t arg1, u_int64_t arg2)
+ia64_init(void)
 {
 	int phys_avail_cnt;
 	vm_offset_t kernstart, kernend;
@@ -762,6 +764,20 @@ ia64_init(u_int64_t arg1, u_int64_t arg2)
 	}
 #endif
 	ia64_set_tpr(0);
+
+	/*
+	 * Save our current context so that we have a known (maybe even
+	 * sane) context as the initial context for new threads that are
+	 * forked from us. If any of those threads (including thread0)
+	 * does something wrong, we may be lucky and return here where
+	 * we're ready for them with a nice panic.
+	 */
+	if (!savectx(thread0.td_pcb))
+		mi_startup();
+
+	/* We should not get here. */
+	panic("ia64_init: Whooaa there!");
+	/* NOTREACHED */
 }
 
 void

sys/ia64/include/cpu.h

@@ -114,7 +114,7 @@ long	fswintrberr(void);				/* MAGIC */
 int	ia64_highfp_drop(struct thread *);
 int	ia64_highfp_load(struct thread *);
 int	ia64_highfp_save(struct thread *);
-void	ia64_init(u_int64_t, u_int64_t);
+void	ia64_init(void);
 int	ia64_pa_access(u_long);
 void	init_prom_interface(struct rpb*);
 void	interrupt(u_int64_t, struct trapframe *);