hp300/hp300/machdep.c

/*
 * Copyright (c) 1988 University of Utah.
 * Copyright (c) 1982, 1986, 1990, 1993
 *	The Regents of the University of California.  All rights reserved.
 *
 * This code is derived from software contributed to Berkeley by
 * the Systems Programming Group of the University of Utah Computer
 * Science Department.
 *
 * %sccs.include.redist.c%
 *
 * from: Utah $Hdr: machdep.c 1.74 92/12/20$
 *
 *	@(#)machdep.c	8.6 (Berkeley) 01/12/94
 */

#include <sys/param.h>
#include <sys/systm.h>
#include <sys/signalvar.h>
#include <sys/kernel.h>
#include <sys/map.h>
#include <sys/proc.h>
#include <sys/buf.h>
#include <sys/reboot.h>
#include <sys/conf.h>
#include <sys/file.h>
#include <sys/clist.h>
#include <sys/callout.h>
#include <sys/malloc.h>
#include <sys/mbuf.h>
#include <sys/msgbuf.h>
#include <sys/ioctl.h>
#include <sys/tty.h>
#include <sys/mount.h>
#include <sys/user.h>
#include <sys/exec.h>
#include <sys/sysctl.h>
#ifdef SYSVSHM
#include <sys/shm.h>
#endif
#ifdef HPUXCOMPAT
#include <hp/hpux/hpux.h>
#endif

#include <machine/cpu.h>
#include <machine/reg.h>
#include <machine/psl.h>
#include <hp/dev/cons.h>
#include <hp300/hp300/isr.h>
#include <hp300/hp300/pte.h>
#include <net/netisr.h>

#define	MAXMEM	64*1024*CLSIZE	/* XXX - from cmap.h */
#include <vm/vm_kern.h>

/* the following is used externally (sysctl_hw) */
char machine[] = "hp300";		/* cpu "architecture" */

vm_map_t buffer_map;
extern vm_offset_t avail_end;

/*
 * Declare these as initialized data so we can patch them.
 */
int	nswbuf = 0;
#ifdef	NBUF
int	nbuf = NBUF;
#else
int	nbuf = 0;
#endif
#ifdef	BUFPAGES
int	bufpages = BUFPAGES;
#else
int	bufpages = 0;
#endif
int	msgbufmapped;		/* set when safe to use msgbuf */
int	maxmem;			/* max memory per process */
int	physmem = MAXMEM;	/* max supported memory, changes to actual */
/*
 * safepri is a safe priority for sleep to set for a spin-wait
 * during autoconfiguration or after a panic.
 */
int	safepri = PSL_LOWIPL;

extern	u_int lowram;
extern	short exframesize[];

/*
 * Console initialization: called early on from main,
 * before vm init or startup.  Do enough configuration
 * to choose and initialize a console.
 */
consinit()
{

	/*
	 * Set cpuspeed immediately since cninit() called routines
	 * might use delay.  Note that we only set it if a custom value
	 * has not already been specified.
	 */
	if (cpuspeed == 0) {
		switch (machineid) {
		case HP_320:
		case HP_330:
		case HP_340:
			cpuspeed = MHZ_16;
			break;
		case HP_350:
		case HP_360:
		case HP_380:
			cpuspeed = MHZ_25;
			break;
		case HP_370:
		case HP_433:
			cpuspeed = MHZ_33;
			break;
		case HP_375:
			cpuspeed = MHZ_50;
			break;
		default:	/* assume the fastest */
			cpuspeed = MHZ_50;
			break;
		}
		if (mmutype == MMU_68040)
			cpuspeed *= 2;	/* XXX */
	}
	/*
         * Find what hardware is attached to this machine.
         */
	find_devs();

	/*
	 * Initialize the console before we print anything out.
	 */
	cninit();
}

/*
 * cpu_startup: allocate memory for variable-sized tables,
 * initialize cpu, and do autoconfiguration.
 */
cpu_startup()
{
	register unsigned i;
	register caddr_t v, firstaddr;
	int base, residual;
	vm_offset_t minaddr, maxaddr;
	vm_size_t size;
#ifdef BUFFERS_UNMANAGED
	vm_offset_t bufmemp;
	caddr_t buffermem;
	int ix;
#endif
#ifdef DEBUG
	extern int pmapdebug;
	int opmapdebug = pmapdebug;

	pmapdebug = 0;
#endif

	/*
	 * Initialize error message buffer (at end of core).
	 * avail_end was pre-decremented in pmap_bootstrap to compensate.
	 */
	for (i = 0; i < btoc(sizeof (struct msgbuf)); i++)
		pmap_enter(kernel_pmap, (vm_offset_t)msgbufp,
		    avail_end + i * NBPG, VM_PROT_ALL, TRUE);
	msgbufmapped = 1;

	/*
	 * Good {morning,afternoon,evening,night}.
	 */
	printf(version);
	identifycpu();
	printf("real mem = %d\n", ctob(physmem));

	/*
	 * Allocate space for system data structures.
	 * The first available real memory address is in "firstaddr".
	 * The first available kernel virtual address is in "v".
	 * As pages of kernel virtual memory are allocated, "v" is incremented.
	 * As pages of memory are allocated and cleared,
	 * "firstaddr" is incremented.
	 * An index into the kernel page table corresponding to the
	 * virtual memory address maintained in "v" is kept in "mapaddr".
	 */
	/*
	 * Make two passes.  The first pass calculates how much memory is
	 * needed and allocates it.  The second pass assigns virtual
	 * addresses to the various data structures.
	 */
	firstaddr = 0;
again:
	v = (caddr_t)firstaddr;

#define	valloc(name, type, num) \
	    (name) = (type *)v; v = (caddr_t)((name)+(num))
#define	valloclim(name, type, num, lim) \
	    (name) = (type *)v; v = (caddr_t)((lim) = ((name)+(num)))
	valloc(cfree, struct cblock, nclist);
	valloc(callout, struct callout, ncallout);
	valloc(swapmap, struct map, nswapmap = maxproc * 2);
#ifdef SYSVSHM
	valloc(shmsegs, struct shmid_ds, shminfo.shmmni);
#endif

	/*
	 * Determine how many buffers to allocate.
	 * Since HPs tend to be long on memory and short on disk speed,
	 * we allocate more buffer space than the BSD standard of
	 * use 10% of memory for the first 2 Meg, 5% of remaining.
	 * We just allocate a flat 10%.  Insure a minimum of 16 buffers.
	 * We allocate 1/2 as many swap buffer headers as file i/o buffers.
	 */
	if (bufpages == 0)
		bufpages = physmem / 10 / CLSIZE;
	if (nbuf == 0) {
		nbuf = bufpages;
		if (nbuf < 16)
			nbuf = 16;
	}
	if (nswbuf == 0) {
		nswbuf = (nbuf / 2) &~ 1;	/* force even */
		if (nswbuf > 256)
			nswbuf = 256;		/* sanity */
	}
	valloc(swbuf, struct buf, nswbuf);
	valloc(buf, struct buf, nbuf);
	/*
	 * End of first pass, size has been calculated so allocate memory
	 */
	if (firstaddr == 0) {
		size = (vm_size_t)(v - firstaddr);
		firstaddr = (caddr_t) kmem_alloc(kernel_map, round_page(size));
		if (firstaddr == 0)
			panic("startup: no room for tables");
#ifdef BUFFERS_UNMANAGED
		buffermem = (caddr_t) kmem_alloc(kernel_map, bufpages*CLBYTES);
		if (buffermem == 0)
			panic("startup: no room for buffers");
#endif
		goto again;
	}
	/*
	 * End of second pass, addresses have been assigned
	 */
	if ((vm_size_t)(v - firstaddr) != size)
		panic("startup: table size inconsistency");
	/*
	 * Now allocate buffers proper.  They are different than the above
	 * in that they usually occupy more virtual memory than physical.
	 */
	size = MAXBSIZE * nbuf;
	buffer_map = kmem_suballoc(kernel_map, (vm_offset_t *)&buffers,
				   &maxaddr, size, TRUE);
	minaddr = (vm_offset_t)buffers;
	if (vm_map_find(buffer_map, vm_object_allocate(size), (vm_offset_t)0,
			&minaddr, size, FALSE) != KERN_SUCCESS)
		panic("startup: cannot allocate buffers");
	base = bufpages / nbuf;
	residual = bufpages % nbuf;
#ifdef BUFFERS_UNMANAGED
	bufmemp = (vm_offset_t) buffermem;
#endif
	for (i = 0; i < nbuf; i++) {
		vm_size_t curbufsize;
		vm_offset_t curbuf;

		/*
		 * First <residual> buffers get (base+1) physical pages
		 * allocated for them.  The rest get (base) physical pages.
		 *
		 * The rest of each buffer occupies virtual space,
		 * but has no physical memory allocated for it.
		 */
		curbuf = (vm_offset_t)buffers + i * MAXBSIZE;
		curbufsize = CLBYTES * (i < residual ? base+1 : base);
#ifdef BUFFERS_UNMANAGED
		/*
		 * Move the physical pages over from buffermem.
		 */
		for (ix = 0; ix < curbufsize/CLBYTES; ix++) {
			vm_offset_t pa;

			pa = pmap_extract(kernel_pmap, bufmemp);
			if (pa == 0)
				panic("startup: unmapped buffer");
			pmap_remove(kernel_pmap, bufmemp, bufmemp+CLBYTES);
			pmap_enter(kernel_pmap,
				   (vm_offset_t)(curbuf + ix * CLBYTES),
				   pa, VM_PROT_READ|VM_PROT_WRITE, TRUE);
			bufmemp += CLBYTES;
		}
#else
		vm_map_pageable(buffer_map, curbuf, curbuf+curbufsize, FALSE);
		vm_map_simplify(buffer_map, curbuf);
#endif
	}
#ifdef BUFFERS_UNMANAGED
#if 0
	/*
	 * We would like to free the (now empty) original address range
	 * but too many bad things will happen if we try.
	 */
	kmem_free(kernel_map, (vm_offset_t)buffermem, bufpages*CLBYTES);
#endif
#endif
	/*
	 * Allocate a submap for exec arguments.  This map effectively
	 * limits the number of processes exec'ing at any time.
	 */
	exec_map = kmem_suballoc(kernel_map, &minaddr, &maxaddr,
				 16*NCARGS, TRUE);
	/*
	 * Allocate a submap for physio
	 */
	phys_map = kmem_suballoc(kernel_map, &minaddr, &maxaddr,
				 VM_PHYS_SIZE, TRUE);

	/*
	 * Finally, allocate mbuf pool.  Since mclrefcnt is an off-size
	 * we use the more space efficient malloc in place of kmem_alloc.
	 */
	mclrefcnt = (char *)malloc(NMBCLUSTERS+CLBYTES/MCLBYTES,
				   M_MBUF, M_NOWAIT);
	bzero(mclrefcnt, NMBCLUSTERS+CLBYTES/MCLBYTES);
	mb_map = kmem_suballoc(kernel_map, (vm_offset_t *)&mbutl, &maxaddr,
			       VM_MBUF_SIZE, FALSE);
	/*
	 * Initialize callouts
	 */
	callfree = callout;
	for (i = 1; i < ncallout; i++)
		callout[i-1].c_next = &callout[i];
	callout[i-1].c_next = NULL;

#ifdef DEBUG
	pmapdebug = opmapdebug;
#endif
	printf("avail mem = %d\n", ptoa(cnt.v_free_count));
	printf("using %d buffers containing %d bytes of memory\n",
		nbuf, bufpages * CLBYTES);
	/*
	 * Set up CPU-specific registers, cache, etc.
	 */
	initcpu();

	/*
	 * Set up buffers, so they can be used to read disk labels.
	 */
	bufinit();

	/*
	 * Configure the system.
	 */
	configure();
}

/*
 * Set registers on exec.
 * XXX Should clear registers except sp, pc,
 * but would break init; should be fixed soon.
 */
setregs(p, entry, retval)
	register struct proc *p;
	u_long entry;
	int retval[2];
{
	struct frame *frame = (struct frame *)p->p_md.md_regs;

	frame->f_pc = entry & ~1;
#ifdef FPCOPROC
	/* restore a null state frame */
	p->p_addr->u_pcb.pcb_fpregs.fpf_null = 0;
	m68881_restore(&p->p_addr->u_pcb.pcb_fpregs);
#endif
#ifdef HPUXCOMPAT
	if (p->p_md.md_flags & MDP_HPUX) {

		frame->f_regs[A0] = 0; /* not 68010 (bit 31), no FPA (30) */
		retval[0] = 0;		/* no float card */
#ifdef FPCOPROC
		retval[1] = 1;		/* yes 68881 */
#else
		retval[1] = 0;		/* no 68881 */
#endif
	}
	/*
	 * XXX This doesn't have much to do with setting registers but
	 * I didn't want to muck up kern_exec.c with this code, so I
	 * stuck it here.
	 *
	 * Ensure we perform the right action on traps type 1 and 2:
	 * If our parent is an HPUX process and we are being traced, turn
	 * on HPUX style interpretation.  Else if we were using the HPUX
	 * style interpretation, revert to the BSD interpretation.
	 *
	 * Note that we do this by changing the trap instruction in the
	 * global "sigcode" array which then gets copied out to the user's
	 * sigcode in the stack.  Since we are changing it in the global
	 * array we must always reset it, even for non-HPUX processes.
	 *
	 * Note also that implementing it in this way creates a potential
	 * race where we could have tweaked it for process A which then
	 * blocks in the copyout to the stack and process B comes along
	 * and untweaks it causing A to wind up with the wrong setting
	 * when the copyout continues.  However, since we have already
	 * copied something out to this user stack page (thereby faulting
	 * it in), this scenerio is extremely unlikely.
	 */
	{
		extern short sigcodetrap[];

		if ((p->p_pptr->p_md.md_flags & MDP_HPUX) &&
		    (p->p_flag & P_TRACED)) {
			p->p_md.md_flags |= MDP_HPUXTRACE;
			*sigcodetrap = 0x4E42;
		} else {
			p->p_md.md_flags &= ~MDP_HPUXTRACE;
			*sigcodetrap = 0x4E41;
		}
	}
#endif
}

/*
 * Info for CTL_HW
 */
extern	char machine[];
char	cpu_model[120];
extern	char ostype[], osrelease[], version[];

identifycpu()
{
	char *t, *mc;
	int len;

	switch (machineid) {
	case HP_320:
		t = "320 (16.67MHz";
		break;
	case HP_330:
		t = "318/319/330 (16.67MHz";
		break;
	case HP_340:
		t = "340 (16.67MHz";
		break;
	case HP_350:
		t = "350 (25MHz";
		break;
	case HP_360:
		t = "360 (25MHz";
		break;
	case HP_370:
		t = "370 (33.33MHz";
		break;
	case HP_375:
		t = "345/375 (50MHz";
		break;
	case HP_380:
		t = "380/425 (25MHz";
		break;
	case HP_433:
		t = "433 (33MHz";
		break;
	default:
		printf("\nunknown machine type %d\n", machineid);
		panic("startup");
	}
	mc = (mmutype == MMU_68040 ? "40" :
	       (mmutype == MMU_68030 ? "30" : "20"));
	sprintf(cpu_model, "HP9000/%s MC680%s CPU", t, mc);
	switch (mmutype) {
	case MMU_68040:
	case MMU_68030:
		strcat(cpu_model, "+MMU");
		break;
	case MMU_68851:
		strcat(cpu_model, ", MC68851 MMU");
		break;
	case MMU_HP:
		strcat(cpu_model, ", HP MMU");
		break;
	default:
		printf("%s\nunknown MMU type %d\n", cpu_model, mmutype);
		panic("startup");
	}
	len = strlen(cpu_model);
	if (mmutype == MMU_68040)
		len += sprintf(cpu_model + len,
		    "+FPU, 4k on-chip physical I/D caches");
	else if (mmutype == MMU_68030)
		len += sprintf(cpu_model + len, ", %sMHz MC68882 FPU",
		       machineid == HP_340 ? "16.67" :
		       (machineid == HP_360 ? "25" :
			(machineid == HP_370 ? "33.33" : "50")));
	else
		len += sprintf(cpu_model + len, ", %sMHz MC68881 FPU",
		       machineid == HP_350 ? "20" : "16.67");
	switch (ectype) {
	case EC_VIRT:
		sprintf(cpu_model + len, ", %dK virtual-address cache",
		       machineid == HP_320 ? 16 : 32);
		break;
	case EC_PHYS:
		sprintf(cpu_model + len, ", %dK physical-address cache",
		       machineid == HP_370 ? 64 : 32);
		break;
	}
	strcat(cpu_model, ")");
	printf("%s\n", cpu_model);
	/*
	 * Now that we have told the user what they have,
	 * let them know if that machine type isn't configured.
	 */
	switch (machineid) {
	case -1:		/* keep compilers happy */
#if !defined(HP320) && !defined(HP350)
	case HP_320:
	case HP_350:
#endif
#ifndef HP330
	case HP_330:
#endif
#if !defined(HP360) && !defined(HP370)
	case HP_340:
	case HP_360:
	case HP_370:
#endif
#if !defined(HP380)
	case HP_380:
	case HP_433:
#endif
		panic("CPU type not configured");
	default:
		break;
	}
}

/*
 * machine dependent system variables.
 */
cpu_sysctl(name, namelen, oldp, oldlenp, newp, newlen, p)
	int *name;
	u_int namelen;
	void *oldp;
	size_t *oldlenp;
	void *newp;
	size_t newlen;
	struct proc *p;
{

	/* all sysctl names at this level are terminal */
	if (namelen != 1)
		return (ENOTDIR);		/* overloaded */

	switch (name[0]) {
	case CPU_CONSDEV:
		return (sysctl_rdstruct(oldp, oldlenp, newp, &cn_tty->t_dev,
		    sizeof cn_tty->t_dev));
	default:
		return (EOPNOTSUPP);
	}
	/* NOTREACHED */
}

#ifdef USELEDS
#include <hp300/hp300/led.h>

int inledcontrol = 0;	/* 1 if we are in ledcontrol already, cheap mutex */
char *ledaddr;

/*
 * Map the LED page and setup the KVA to access it.
 */
ledinit()
{
	extern caddr_t ledbase;

	pmap_enter(kernel_pmap, (vm_offset_t)ledbase, (vm_offset_t)LED_ADDR,
		   VM_PROT_READ|VM_PROT_WRITE, TRUE);
	ledaddr = (char *) ((int)ledbase | (LED_ADDR & PGOFSET));
}

/*
 * Do lights:
 *	`ons' is a mask of LEDs to turn on,
 *	`offs' is a mask of LEDs to turn off,
 *	`togs' is a mask of LEDs to toggle.
 * Note we don't use splclock/splx for mutual exclusion.
 * They are expensive and we really don't need to be that precise.
 * Besides we would like to be able to profile this routine.
 */
ledcontrol(ons, offs, togs)
	register int ons, offs, togs;
{
	static char currentleds;
	register char leds;

	inledcontrol = 1;
	leds = currentleds;
	if (ons)
		leds |= ons;
	if (offs)
		leds &= ~offs;
	if (togs)
		leds ^= togs;
	currentleds = leds;
	*ledaddr = ~leds;
	inledcontrol = 0;
}
#endif

#define SS_RTEFRAME	1
#define SS_FPSTATE	2
#define SS_USERREGS	4

struct sigstate {
	int	ss_flags;		/* which of the following are valid */
	struct	frame ss_frame;		/* original exception frame */
	struct	fpframe ss_fpstate;	/* 68881/68882 state info */
};

/*
 * WARNING: code in locore.s assumes the layout shown for sf_signum
 * thru sf_handler so... don't screw with them!
 */
struct sigframe {
	int	sf_signum;		/* signo for handler */
	int	sf_code;		/* additional info for handler */
	struct	sigcontext *sf_scp;	/* context ptr for handler */
	sig_t	sf_handler;		/* handler addr for u_sigc */
	struct	sigstate sf_state;	/* state of the hardware */
	struct	sigcontext sf_sc;	/* actual context */
};

#ifdef HPUXCOMPAT
struct	hpuxsigcontext {
	int	hsc_syscall;
	char	hsc_action;
	char	hsc_pad1;
	char	hsc_pad2;
	char	hsc_onstack;
	int	hsc_mask;
	int	hsc_sp;
	short	hsc_ps;
	int	hsc_pc;
/* the rest aren't part of the context but are included for our convenience */
	short	hsc_pad;
	u_int	hsc_magic;		/* XXX sigreturn: cookie */
	struct	sigcontext *hsc_realsc;	/* XXX sigreturn: ptr to BSD context */
};

/*
 * For an HP-UX process, a partial hpuxsigframe follows the normal sigframe.
 * Tremendous waste of space, but some HP-UX applications (e.g. LCL) need it.
 */
struct hpuxsigframe {
	int	hsf_signum;
	int	hsf_code;
	struct	sigcontext *hsf_scp;
	struct	hpuxsigcontext hsf_sc;
	int	hsf_regs[15];
};
#endif

#ifdef DEBUG
int sigdebug = 0;
int sigpid = 0;
#define SDB_FOLLOW	0x01
#define SDB_KSTACK	0x02
#define SDB_FPSTATE	0x04
#endif

/*
 * Send an interrupt to process.
 */
void
sendsig(catcher, sig, mask, code)
	sig_t catcher;
	int sig, mask;
	unsigned code;
{
	register struct proc *p = curproc;
	register struct sigframe *fp, *kfp;
	register struct frame *frame;
	register struct sigacts *psp = p->p_sigacts;
	register short ft;
	int oonstack, fsize;
	extern char sigcode[], esigcode[];

	frame = (struct frame *)p->p_md.md_regs;
	ft = frame->f_format;
	oonstack = psp->ps_sigstk.ss_flags & SA_ONSTACK;
	/*
	 * Allocate and validate space for the signal handler
	 * context. Note that if the stack is in P0 space, the
	 * call to grow() is a nop, and the useracc() check
	 * will fail if the process has not already allocated
	 * the space with a `brk'.
	 */
#ifdef HPUXCOMPAT
	if (p->p_md.md_flags & MDP_HPUX)
		fsize = sizeof(struct sigframe) + sizeof(struct hpuxsigframe);
	else
#endif
	fsize = sizeof(struct sigframe);
	if ((psp->ps_flags & SAS_ALTSTACK) &&
	    (psp->ps_sigstk.ss_flags & SA_ONSTACK) == 0 &&
	    (psp->ps_sigonstack & sigmask(sig))) {
		fp = (struct sigframe *)(psp->ps_sigstk.ss_base +
					 psp->ps_sigstk.ss_size - fsize);
		psp->ps_sigstk.ss_flags |= SA_ONSTACK;
	} else
		fp = (struct sigframe *)(frame->f_regs[SP] - fsize);
	if ((unsigned)fp <= USRSTACK - ctob(p->p_vmspace->vm_ssize))
		(void)grow(p, (unsigned)fp);
#ifdef DEBUG
	if ((sigdebug & SDB_KSTACK) && p->p_pid == sigpid)
		printf("sendsig(%d): sig %d ssp %x usp %x scp %x ft %d\n",
		       p->p_pid, sig, &oonstack, fp, &fp->sf_sc, ft);
#endif
	if (useracc((caddr_t)fp, fsize, B_WRITE) == 0) {
#ifdef DEBUG
		if ((sigdebug & SDB_KSTACK) && p->p_pid == sigpid)
			printf("sendsig(%d): useracc failed on sig %d\n",
			       p->p_pid, sig);
#endif
		/*
		 * Process has trashed its stack; give it an illegal
		 * instruction to halt it in its tracks.
		 */
		SIGACTION(p, SIGILL) = SIG_DFL;
		sig = sigmask(SIGILL);
		p->p_sigignore &= ~sig;
		p->p_sigcatch &= ~sig;
		p->p_sigmask &= ~sig;
		psignal(p, SIGILL);
		return;
	}
	kfp = (struct sigframe *)malloc((u_long)fsize, M_TEMP, M_WAITOK);
	/*
	 * Build the argument list for the signal handler.
	 */
	kfp->sf_signum = sig;
	kfp->sf_code = code;
	kfp->sf_scp = &fp->sf_sc;
	kfp->sf_handler = catcher;
	/*
	 * Save necessary hardware state.  Currently this includes:
	 *	- general registers
	 *	- original exception frame (if not a "normal" frame)
	 *	- FP coprocessor state
	 */
	kfp->sf_state.ss_flags = SS_USERREGS;
	bcopy((caddr_t)frame->f_regs,
	      (caddr_t)kfp->sf_state.ss_frame.f_regs, sizeof frame->f_regs);
	if (ft >= FMT7) {
#ifdef DEBUG
		if (ft > 15 || exframesize[ft] < 0)
			panic("sendsig: bogus frame type");
#endif
		kfp->sf_state.ss_flags |= SS_RTEFRAME;
		kfp->sf_state.ss_frame.f_format = frame->f_format;
		kfp->sf_state.ss_frame.f_vector = frame->f_vector;
		bcopy((caddr_t)&frame->F_u,
		      (caddr_t)&kfp->sf_state.ss_frame.F_u, exframesize[ft]);
		/*
		 * Leave an indicator that we need to clean up the kernel
		 * stack.  We do this by setting the "pad word" above the
		 * hardware stack frame to the amount the stack must be
		 * adjusted by.
		 *
		 * N.B. we increment rather than just set f_stackadj in
		 * case we are called from syscall when processing a
		 * sigreturn.  In that case, f_stackadj may be non-zero.
		 */
		frame->f_stackadj += exframesize[ft];
		frame->f_format = frame->f_vector = 0;
#ifdef DEBUG
		if (sigdebug & SDB_FOLLOW)
			printf("sendsig(%d): copy out %d of frame %d\n",
			       p->p_pid, exframesize[ft], ft);
#endif
	}
#ifdef FPCOPROC
	kfp->sf_state.ss_flags |= SS_FPSTATE;
	m68881_save(&kfp->sf_state.ss_fpstate);
#ifdef DEBUG
	if ((sigdebug & SDB_FPSTATE) && *(char *)&kfp->sf_state.ss_fpstate)
		printf("sendsig(%d): copy out FP state (%x) to %x\n",
		       p->p_pid, *(u_int *)&kfp->sf_state.ss_fpstate,
		       &kfp->sf_state.ss_fpstate);
#endif
#endif
	/*
	 * Build the signal context to be used by sigreturn.
	 */
	kfp->sf_sc.sc_onstack = oonstack;
	kfp->sf_sc.sc_mask = mask;
	kfp->sf_sc.sc_sp = frame->f_regs[SP];
	kfp->sf_sc.sc_fp = frame->f_regs[A6];
	kfp->sf_sc.sc_ap = (int)&fp->sf_state;
	kfp->sf_sc.sc_pc = frame->f_pc;
	kfp->sf_sc.sc_ps = frame->f_sr;
#ifdef HPUXCOMPAT
	/*
	 * Create an HP-UX style sigcontext structure and associated goo
	 */
	if (p->p_md.md_flags & MDP_HPUX) {
		register struct hpuxsigframe *hkfp;

		hkfp = (struct hpuxsigframe *)&kfp[1];
		hkfp->hsf_signum = bsdtohpuxsig(kfp->sf_signum);
		hkfp->hsf_code = kfp->sf_code;
		hkfp->hsf_scp = (struct sigcontext *)
			&((struct hpuxsigframe *)(&fp[1]))->hsf_sc;
		hkfp->hsf_sc.hsc_syscall = 0;		/* XXX */
		hkfp->hsf_sc.hsc_action = 0;		/* XXX */
		hkfp->hsf_sc.hsc_pad1 = hkfp->hsf_sc.hsc_pad2 = 0;
		hkfp->hsf_sc.hsc_onstack = kfp->sf_sc.sc_onstack;
		hkfp->hsf_sc.hsc_mask = kfp->sf_sc.sc_mask;
		hkfp->hsf_sc.hsc_sp = kfp->sf_sc.sc_sp;
		hkfp->hsf_sc.hsc_ps = kfp->sf_sc.sc_ps;
		hkfp->hsf_sc.hsc_pc = kfp->sf_sc.sc_pc;
		hkfp->hsf_sc.hsc_pad = 0;
		hkfp->hsf_sc.hsc_magic = 0xdeadbeef;
		hkfp->hsf_sc.hsc_realsc = kfp->sf_scp;
		bcopy((caddr_t)frame->f_regs, (caddr_t)hkfp->hsf_regs,
		      sizeof (hkfp->hsf_regs));

		kfp->sf_signum = hkfp->hsf_signum;
		kfp->sf_scp = hkfp->hsf_scp;
	}
#endif
	(void) copyout((caddr_t)kfp, (caddr_t)fp, fsize);
	frame->f_regs[SP] = (int)fp;
#ifdef DEBUG
	if (sigdebug & SDB_FOLLOW)
		printf("sendsig(%d): sig %d scp %x fp %x sc_sp %x sc_ap %x\n",
		       p->p_pid, sig, kfp->sf_scp, fp,
		       kfp->sf_sc.sc_sp, kfp->sf_sc.sc_ap);
#endif
	/*
	 * Signal trampoline code is at base of user stack.
	 */
	frame->f_pc = (int)PS_STRINGS - (esigcode - sigcode);
#ifdef DEBUG
	if ((sigdebug & SDB_KSTACK) && p->p_pid == sigpid)
		printf("sendsig(%d): sig %d returns\n",
		       p->p_pid, sig);
#endif
	free((caddr_t)kfp, M_TEMP);
}

/*
 * 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 priviledges or to cause
 * a machine fault.
 */
struct sigreturn_args {
	struct sigcontext *sigcntxp;
};
/* ARGSUSED */
sigreturn(p, uap, retval)
	struct proc *p;
	struct sigreturn_args *uap;
	int *retval;
{
	register struct sigcontext *scp;
	register struct frame *frame;
	register int rf;
	struct sigcontext tsigc;
	struct sigstate tstate;
	int flags;

	scp = uap->sigcntxp;
#ifdef DEBUG
	if (sigdebug & SDB_FOLLOW)
		printf("sigreturn: pid %d, scp %x\n", p->p_pid, scp);
#endif
	if ((int)scp & 1)
		return (EINVAL);
#ifdef HPUXCOMPAT
	/*
	 * Grab context as an HP-UX style context and determine if it
	 * was one that we contructed in sendsig.
	 */
	if (p->p_md.md_flags & MDP_HPUX) {
		struct hpuxsigcontext *hscp = (struct hpuxsigcontext *)scp;
		struct hpuxsigcontext htsigc;

		if (useracc((caddr_t)hscp, sizeof (*hscp), B_WRITE) == 0 ||
		    copyin((caddr_t)hscp, (caddr_t)&htsigc, sizeof htsigc))
			return (EINVAL);
		/*
		 * If not generated by sendsig or we cannot restore the
		 * BSD-style sigcontext, just restore what we can -- state
		 * will be lost, but them's the breaks.
		 */
		hscp = &htsigc;
		if (hscp->hsc_magic != 0xdeadbeef ||
		    (scp = hscp->hsc_realsc) == 0 ||
		    useracc((caddr_t)scp, sizeof (*scp), B_WRITE) == 0 ||
		    copyin((caddr_t)scp, (caddr_t)&tsigc, sizeof tsigc)) {
			if (hscp->hsc_onstack & 01)
				p->p_sigacts->ps_sigstk.ss_flags |= SA_ONSTACK;
			else
				p->p_sigacts->ps_sigstk.ss_flags &= ~SA_ONSTACK;
			p->p_sigmask = hscp->hsc_mask &~ sigcantmask;
			frame = (struct frame *) p->p_md.md_regs;
			frame->f_regs[SP] = hscp->hsc_sp;
			frame->f_pc = hscp->hsc_pc;
			frame->f_sr = hscp->hsc_ps &~ PSL_USERCLR;
			return (EJUSTRETURN);
		}
		/*
		 * Otherwise, overlay BSD context with possibly modified
		 * HP-UX values.
		 */
		tsigc.sc_onstack = hscp->hsc_onstack;
		tsigc.sc_mask = hscp->hsc_mask;
		tsigc.sc_sp = hscp->hsc_sp;
		tsigc.sc_ps = hscp->hsc_ps;
		tsigc.sc_pc = hscp->hsc_pc;
	} else
#endif
	/*
	 * Test and fetch the context structure.
	 * We grab it all at once for speed.
	 */
	if (useracc((caddr_t)scp, sizeof (*scp), B_WRITE) == 0 ||
	    copyin((caddr_t)scp, (caddr_t)&tsigc, sizeof tsigc))
		return (EINVAL);
	scp = &tsigc;
	if ((scp->sc_ps & (PSL_MBZ|PSL_IPL|PSL_S)) != 0)
		return (EINVAL);
	/*
	 * Restore the user supplied information
	 */
	if (scp->sc_onstack & 01)
		p->p_sigacts->ps_sigstk.ss_flags |= SA_ONSTACK;
	else
		p->p_sigacts->ps_sigstk.ss_flags &= ~SA_ONSTACK;
	p->p_sigmask = scp->sc_mask &~ sigcantmask;
	frame = (struct frame *) p->p_md.md_regs;
	frame->f_regs[SP] = scp->sc_sp;
	frame->f_regs[A6] = scp->sc_fp;
	frame->f_pc = scp->sc_pc;
	frame->f_sr = scp->sc_ps;
	/*
	 * Grab pointer to hardware state information.
	 * If zero, the user is probably doing a longjmp.
	 */
	if ((rf = scp->sc_ap) == 0)
		return (EJUSTRETURN);
	/*
	 * See if there is anything to do before we go to the
	 * expense of copying in close to 1/2K of data
	 */
	flags = fuword((caddr_t)rf);
#ifdef DEBUG
	if (sigdebug & SDB_FOLLOW)
		printf("sigreturn(%d): sc_ap %x flags %x\n",
		       p->p_pid, rf, flags);
#endif
	/*
	 * fuword failed (bogus sc_ap value).
	 */
	if (flags == -1)
		return (EINVAL);
	if (flags == 0 || copyin((caddr_t)rf, (caddr_t)&tstate, sizeof tstate))
		return (EJUSTRETURN);
#ifdef DEBUG
	if ((sigdebug & SDB_KSTACK) && p->p_pid == sigpid)
		printf("sigreturn(%d): ssp %x usp %x scp %x ft %d\n",
		       p->p_pid, &flags, scp->sc_sp, uap->sigcntxp,
		       (flags&SS_RTEFRAME) ? tstate.ss_frame.f_format : -1);
#endif
	/*
	 * Restore most of the users registers except for A6 and SP
	 * which were handled above.
	 */
	if (flags & SS_USERREGS)
		bcopy((caddr_t)tstate.ss_frame.f_regs,
		      (caddr_t)frame->f_regs, sizeof(frame->f_regs)-2*NBPW);
	/*
	 * Restore long stack frames.  Note that we do not copy
	 * back the saved SR or PC, they were picked up above from
	 * the sigcontext structure.
	 */
	if (flags & SS_RTEFRAME) {
		register int sz;

		/* grab frame type and validate */
		sz = tstate.ss_frame.f_format;
		if (sz > 15 || (sz = exframesize[sz]) < 0)
			return (EINVAL);
		frame->f_stackadj -= sz;
		frame->f_format = tstate.ss_frame.f_format;
		frame->f_vector = tstate.ss_frame.f_vector;
		bcopy((caddr_t)&tstate.ss_frame.F_u, (caddr_t)&frame->F_u, sz);
#ifdef DEBUG
		if (sigdebug & SDB_FOLLOW)
			printf("sigreturn(%d): copy in %d of frame type %d\n",
			       p->p_pid, sz, tstate.ss_frame.f_format);
#endif
	}
#ifdef FPCOPROC
	/*
	 * Finally we restore the original FP context
	 */
	if (flags & SS_FPSTATE)
		m68881_restore(&tstate.ss_fpstate);
#ifdef DEBUG
	if ((sigdebug & SDB_FPSTATE) && *(char *)&tstate.ss_fpstate)
		printf("sigreturn(%d): copied in FP state (%x) at %x\n",
		       p->p_pid, *(u_int *)&tstate.ss_fpstate,
		       &tstate.ss_fpstate);
#endif
#endif
#ifdef DEBUG
	if ((sigdebug & SDB_FOLLOW) ||
	    ((sigdebug & SDB_KSTACK) && p->p_pid == sigpid))
		printf("sigreturn(%d): returns\n", p->p_pid);
#endif
	return (EJUSTRETURN);
}

int	waittime = -1;

boot(howto)
	register int howto;
{
	/* take a snap shot before clobbering any registers */
	if (curproc && curproc->p_addr)
		savectx(curproc->p_addr, 0);

	boothowto = howto;
	if ((howto&RB_NOSYNC) == 0 && waittime < 0) {
		register struct buf *bp;
		int iter, nbusy;

		waittime = 0;
		(void) spl0();
		printf("syncing disks... ");
		/*
		 * Release vnodes held by texts before sync.
		 */
		if (panicstr == 0)
			vnode_pager_umount(NULL);
#ifdef notdef
#include "vn.h"
#if NVN > 0
		vnshutdown();
#endif
#endif
		sync(&proc0, (void *)NULL, (int *)NULL);

		for (iter = 0; iter < 20; iter++) {
			nbusy = 0;
			for (bp = &buf[nbuf]; --bp >= buf; )
				if ((bp->b_flags & (B_BUSY|B_INVAL)) == B_BUSY)
					nbusy++;
			if (nbusy == 0)
				break;
			printf("%d ", nbusy);
			DELAY(40000 * iter);
		}
		if (nbusy)
			printf("giving up\n");
		else
			printf("done\n");
		/*
		 * If we've been adjusting the clock, the todr
		 * will be out of synch; adjust it now.
		 */
		resettodr();
	}
	splhigh();			/* extreme priority */
	if (howto&RB_HALT) {
		printf("halted\n\n");
		asm("	stop	#0x2700");
	} else {
		if (howto & RB_DUMP)
			dumpsys();
		doboot();
		/*NOTREACHED*/
	}
	/*NOTREACHED*/
}

int	dumpmag = 0x8fca0101;	/* magic number for savecore */
int	dumpsize = 0;		/* also for savecore */
long	dumplo = 0;

dumpconf()
{
	int nblks;

	dumpsize = physmem;
	if (dumpdev != NODEV && bdevsw[major(dumpdev)].d_psize) {
		nblks = (*bdevsw[major(dumpdev)].d_psize)(dumpdev);
		if (dumpsize > btoc(dbtob(nblks - dumplo)))
			dumpsize = btoc(dbtob(nblks - dumplo));
		else if (dumplo == 0)
			dumplo = nblks - btodb(ctob(physmem));
	}
	/*
	 * Don't dump on the first CLBYTES (why CLBYTES?)
	 * in case the dump device includes a disk label.
	 */
	if (dumplo < btodb(CLBYTES))
		dumplo = btodb(CLBYTES);
}

/*
 * Doadump comes here after turning off memory management and
 * getting on the dump stack, either when called above, or by
 * the auto-restart code.
 */
dumpsys()
{

	msgbufmapped = 0;
	if (dumpdev == NODEV)
		return;
	/*
	 * For dumps during autoconfiguration,
	 * if dump device has already configured...
	 */
	if (dumpsize == 0)
		dumpconf();
	if (dumplo < 0)
		return;
	printf("\ndumping to dev %x, offset %d\n", dumpdev, dumplo);
	printf("dump ");
	switch ((*bdevsw[major(dumpdev)].d_dump)(dumpdev)) {

	case ENXIO:
		printf("device bad\n");
		break;

	case EFAULT:
		printf("device not ready\n");
		break;

	case EINVAL:
		printf("area improper\n");
		break;

	case EIO:
		printf("i/o error\n");
		break;

	default:
		printf("succeeded\n");
		break;
	}
}

initcpu()
{
#ifdef MAPPEDCOPY
	extern u_int mappedcopysize;

	/*
	 * Initialize lower bound for doing copyin/copyout using
	 * page mapping (if not already set).  We don't do this on
	 * VAC machines as it loses big time.
	 */
	if (mappedcopysize == 0) {
		if (ectype == EC_VIRT)
			mappedcopysize = (u_int) -1;
		else
			mappedcopysize = NBPG;
	}
#endif
	parityenable();
#ifdef USELEDS
	ledinit();
#endif
}

straytrap(pc, evec)
	int pc;
	u_short evec;
{
	printf("unexpected trap (vector offset %x) from %x\n",
	       evec & 0xFFF, pc);
}

int	*nofault;

badaddr(addr)
	register caddr_t addr;
{
	register int i;
	label_t	faultbuf;

#ifdef lint
	i = *addr; if (i) return(0);
#endif
	nofault = (int *) &faultbuf;
	if (setjmp((label_t *)nofault)) {
		nofault = (int *) 0;
		return(1);
	}
	i = *(volatile short *)addr;
	nofault = (int *) 0;
	return(0);
}

badbaddr(addr)
	register caddr_t addr;
{
	register int i;
	label_t	faultbuf;

#ifdef lint
	i = *addr; if (i) return(0);
#endif
	nofault = (int *) &faultbuf;
	if (setjmp((label_t *)nofault)) {
		nofault = (int *) 0;
		return(1);
	}
	i = *(volatile char *)addr;
	nofault = (int *) 0;
	return(0);
}

netintr()
{
#ifdef INET
	if (netisr & (1 << NETISR_ARP)) {
		netisr &= ~(1 << NETISR_ARP);
		arpintr();
	}
	if (netisr & (1 << NETISR_IP)) {
		netisr &= ~(1 << NETISR_IP);
		ipintr();
	}
#endif
#ifdef NS
	if (netisr & (1 << NETISR_NS)) {
		netisr &= ~(1 << NETISR_NS);
		nsintr();
	}
#endif
#ifdef ISO
	if (netisr & (1 << NETISR_ISO)) {
		netisr &= ~(1 << NETISR_ISO);
		clnlintr();
	}
#endif
#ifdef CCITT
	if (netisr & (1 << NETISR_CCITT)) {
		netisr &= ~(1 << NETISR_CCITT);
		ccittintr();
	}
#endif
}

intrhand(sr)
	int sr;
{
	register struct isr *isr;
	register int found = 0;
	register int ipl;
	extern struct isr isrqueue[];
	static int straycount;

	ipl = (sr >> 8) & 7;
	switch (ipl) {

	case 3:
	case 4:
	case 5:
		ipl = ISRIPL(ipl);
		isr = isrqueue[ipl].isr_forw;
		for (; isr != &isrqueue[ipl]; isr = isr->isr_forw) {
			if ((isr->isr_intr)(isr->isr_arg)) {
				found++;
				break;
			}
		}
		if (found)
			straycount = 0;
		else if (++straycount > 50)
			panic("intrhand: stray interrupt");
		else
			printf("stray interrupt, sr 0x%x\n", sr);
		break;

	case 0:
	case 1:
	case 2:
	case 6:
	case 7:
		if (++straycount > 50)
			panic("intrhand: unexpected sr");
		else
			printf("intrhand: unexpected sr 0x%x\n", sr);
		break;
	}
}

#if defined(DEBUG) && !defined(PANICBUTTON)
#define PANICBUTTON
#endif

#ifdef PANICBUTTON
int panicbutton = 1;	/* non-zero if panic buttons are enabled */
int crashandburn = 0;
int candbdelay = 50;	/* give em half a second */

void
candbtimer(arg)
	void *arg;
{

	crashandburn = 0;
}
#endif

/*
 * Level 7 interrupts can be caused by the keyboard or parity errors.
 */
nmihand(frame)
	struct frame frame;
{
	if (kbdnmi()) {
#ifdef PANICBUTTON
		static int innmihand = 0;

		/*
		 * Attempt to reduce the window of vulnerability for recursive
		 * NMIs (e.g. someone holding down the keyboard reset button).
		 */
		if (innmihand == 0) {
			innmihand = 1;
			printf("Got a keyboard NMI\n");
			innmihand = 0;
		}
		if (panicbutton) {
			if (crashandburn) {
				crashandburn = 0;
				panic(panicstr ?
				      "forced crash, nosync" : "forced crash");
			}
			crashandburn++;
			timeout(candbtimer, (void *)0, candbdelay);
		}
#endif
		return;
	}
	if (parityerror(&frame))
		return;
	/* panic?? */
	printf("unexpected level 7 interrupt ignored\n");
}

/*
 * Parity error section.  Contains magic.
 */
#define PARREG		((volatile short *)IIOV(0x5B0000))
static int gotparmem = 0;
#ifdef DEBUG
int ignorekperr = 0;	/* ignore kernel parity errors */
#endif

/*
 * Enable parity detection
 */
parityenable()
{
	label_t	faultbuf;

	nofault = (int *) &faultbuf;
	if (setjmp((label_t *)nofault)) {
		nofault = (int *) 0;
#ifdef DEBUG
		printf("No parity memory\n");
#endif
		return;
	}
	*PARREG = 1;
	nofault = (int *) 0;
	gotparmem = 1;
#ifdef DEBUG
	printf("Parity detection enabled\n");
#endif
}

/*
 * Determine if level 7 interrupt was caused by a parity error
 * and deal with it if it was.  Returns 1 if it was a parity error.
 */
parityerror(fp)
	struct frame *fp;
{
	if (!gotparmem)
		return(0);
	*PARREG = 0;
	DELAY(10);
	*PARREG = 1;
	if (panicstr) {
		printf("parity error after panic ignored\n");
		return(1);
	}
	if (!findparerror())
		printf("WARNING: transient parity error ignored\n");
	else if (USERMODE(fp->f_sr)) {
		printf("pid %d: parity error\n", curproc->p_pid);
		uprintf("sorry, pid %d killed due to memory parity error\n",
			curproc->p_pid);
		psignal(curproc, SIGKILL);
#ifdef DEBUG
	} else if (ignorekperr) {
		printf("WARNING: kernel parity error ignored\n");
#endif
	} else {
		regdump(fp, 128);
		panic("kernel parity error");
	}
	return(1);
}

/*
 * Yuk!  There has got to be a better way to do this!
 * Searching all of memory with interrupts blocked can lead to disaster.
 */
findparerror()
{
	static label_t parcatch;
	static int looking = 0;
	volatile int pg, o, s;
	register volatile int *ip;
	register int i;
	int found;

#ifdef lint
	i = o = pg = 0; if (i) return(0);
#endif
	/*
	 * If looking is true we are searching for a known parity error
	 * and it has just occured.  All we do is return to the higher
	 * level invocation.
	 */
	if (looking)
		longjmp(&parcatch);
	s = splhigh();
	/*
	 * If setjmp returns true, the parity error we were searching
	 * for has just occured (longjmp above) at the current pg+o
	 */
	if (setjmp(&parcatch)) {
		printf("Parity error at 0x%x\n", ctob(pg)|o);
		found = 1;
		goto done;
	}
	/*
	 * If we get here, a parity error has occured for the first time
	 * and we need to find it.  We turn off any external caches and
	 * loop thru memory, testing every longword til a fault occurs and
	 * we regain control at setjmp above.  Note that because of the
	 * setjmp, pg and o need to be volatile or their values will be lost.
	 */
	looking = 1;
	ecacheoff();
	for (pg = btoc(lowram); pg < btoc(lowram)+physmem; pg++) {
		pmap_enter(kernel_pmap, (vm_offset_t)vmmap, ctob(pg),
		    VM_PROT_READ, TRUE);
		ip = (int *)vmmap;
		for (o = 0; o < NBPG; o += sizeof(int))
			i = *ip++;
	}
	/*
	 * Getting here implies no fault was found.  Should never happen.
	 */
	printf("Couldn't locate parity error\n");
	found = 0;
done:
	looking = 0;
	pmap_remove(kernel_pmap, (vm_offset_t)vmmap, (vm_offset_t)&vmmap[NBPG]);
	ecacheon();
	splx(s);
	return(found);
}

regdump(fp, sbytes)
	struct frame *fp; /* must not be register */
	int sbytes;
{
	static int doingdump = 0;
	register int i;
	int s;
	extern char *hexstr();

	if (doingdump)
		return;
	s = splhigh();
	doingdump = 1;
	printf("pid = %d, pc = %s, ",
	       curproc ? curproc->p_pid : -1, hexstr(fp->f_pc, 8));
	printf("ps = %s, ", hexstr(fp->f_sr, 4));
	printf("sfc = %s, ", hexstr(getsfc(), 4));
	printf("dfc = %s\n", hexstr(getdfc(), 4));
	printf("Registers:\n     ");
	for (i = 0; i < 8; i++)
		printf("        %d", i);
	printf("\ndreg:");
	for (i = 0; i < 8; i++)
		printf(" %s", hexstr(fp->f_regs[i], 8));
	printf("\nareg:");
	for (i = 0; i < 8; i++)
		printf(" %s", hexstr(fp->f_regs[i+8], 8));
	if (sbytes > 0) {
		if (fp->f_sr & PSL_S) {
			printf("\n\nKernel stack (%s):",
			       hexstr((int)(((int *)&fp)-1), 8));
			dumpmem(((int *)&fp)-1, sbytes, 0);
		} else {
			printf("\n\nUser stack (%s):", hexstr(fp->f_regs[SP], 8));
			dumpmem((int *)fp->f_regs[SP], sbytes, 1);
		}
	}
	doingdump = 0;
	splx(s);
}

extern char kstack[];
#define KSADDR	((int *)&(kstack[(UPAGES-1)*NBPG]))

dumpmem(ptr, sz, ustack)
	register int *ptr;
	int sz, ustack;
{
	register int i, val;
	extern char *hexstr();

	for (i = 0; i < sz; i++) {
		if ((i & 7) == 0)
			printf("\n%s: ", hexstr((int)ptr, 6));
		else
			printf(" ");
		if (ustack == 1) {
			if ((val = fuword(ptr++)) == -1)
				break;
		} else {
			if (ustack == 0 &&
			    (ptr < KSADDR || ptr > KSADDR+(NBPG/4-1)))
				break;
			val = *ptr++;
		}
		printf("%s", hexstr(val, 8));
	}
	printf("\n");
}

char *
hexstr(val, len)
	register int val;
	int len;
{
	static char nbuf[9];
	register int x, i;

	if (len > 8)
		return("");
	nbuf[len] = '\0';
	for (i = len-1; i >= 0; --i) {
		x = val & 0xF;
		if (x > 9)
			nbuf[i] = x - 10 + 'A';
		else
			nbuf[i] = x + '0';
		val >>= 4;
	}
	return(nbuf);
}

#ifdef DEBUG
char oflowmsg[] = "k-stack overflow";
char uflowmsg[] = "k-stack underflow";

badkstack(oflow, fr)
	int oflow;
	struct frame fr;
{
	extern char kstackatbase[];

	printf("%s: sp should be %x\n",
	       oflow ? oflowmsg : uflowmsg,
	       kstackatbase - (exframesize[fr.f_format] + 8));
	regdump(&fr, 0);
	panic(oflow ? oflowmsg : uflowmsg);
}
#endif