machdep.c

早期freebsd实现

/*-
 * Copyright (c) 1982, 1987, 1990, 1993
 *	The Regents of the University of California.  All rights reserved.
 *
 * This code is derived from software contributed to Berkeley by
 * William Jolitz.
 *
 * 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.
 * 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. All advertising materials mentioning features or use of this software
 *    must display the following acknowledgement:
 *	This product includes software developed by the University of
 *	California, Berkeley and its contributors.
 * 4. Neither the name of the University nor the names of its contributors
 *    may be used to endorse or promote products derived from this software
 *    without specific prior written permission.
 *
 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``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 REGENTS OR CONTRIBUTORS 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.
 *
 *	@(#)machdep.c	8.2 (Berkeley) 1/12/94
 */

#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/user.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/sysctl.h>

#include <net/netisr.h>

#include <vm/vm.h>
#include <vm/vm_kern.h>
#include <vm/vm_page.h>

vm_map_t buffer_map;
extern vm_offset_t avail_end;

#include <machine/cpu.h>
#include <machine/reg.h>
#include <machine/psl.h>
#include <machine/specialreg.h>
#include <i386/isa/rtc.h>
#include <i386/i386/cons.h>

/*
 * 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 */

/*
 * Machine-dependent startup code
 */
int boothowto = 0, Maxmem = 0;
long dumplo;
int physmem, maxmem;
extern int bootdev;
#ifdef SMALL
extern int forcemaxmem;
#endif
int biosmem;

extern cyloffset;

cpu_startup(firstaddr)
	int firstaddr;
{
	register int unixsize;
	register unsigned i;
	register struct pte *pte;
	int mapaddr, j;
	register caddr_t v;
	int maxbufs, base, residual;
	extern long Usrptsize;
	vm_offset_t minaddr, maxaddr;
	vm_size_t size;

	/*
	 * 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, msgbufp, avail_end + i * NBPG,
			   VM_PROT_ALL, TRUE);
	msgbufmapped = 1;

#ifdef KDB
	kdb_init();			/* startup kernel debugger */
#endif
	/*
	 * Good {morning,afternoon,evening,night}.
	 */
	printf(version);
	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.
	 * Use 10% of memory for the first 2 Meg, 5% of the remaining
	 * memory. Insure a minimum of 16 buffers.
	 * We allocate 1/2 as many swap buffer headers as file i/o buffers.
	 */
	if (bufpages == 0)
		if (physmem < (2 * 1024 * 1024))
			bufpages = physmem / 10 / CLSIZE;
		else
			bufpages = ((2 * 1024 * 1024 + physmem) / 20) / CLSIZE;
	if (nbuf == 0) {
		nbuf = bufpages / 2;
		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 = (int)kmem_alloc(kernel_map, round_page(size));
		if (firstaddr == 0)
			panic("startup: no room for tables");
		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;
	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);
		vm_map_pageable(buffer_map, curbuf, curbuf+curbufsize, FALSE);
		vm_map_simplify(buffer_map, curbuf);
	}
	/*
	 * 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;

	/*printf("avail mem = %d\n", ptoa(vm_page_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();
}

#ifdef PGINPROF
/*
 * Return the difference (in microseconds)
 * between the  current time and a previous
 * time as represented  by the arguments.
 * If there is a pending clock interrupt
 * which has not been serviced due to high
 * ipl, return error code.
 */
/*ARGSUSED*/
vmtime(otime, olbolt, oicr)
	register int otime, olbolt, oicr;
{

	return (((time.tv_sec-otime)*60 + lbolt-olbolt)*16667);
}
#endif

struct sigframe {
	int	sf_signum;
	int	sf_code;
	struct	sigcontext *sf_scp;
	sig_t	sf_handler;
	int	sf_eax;	
	int	sf_edx;	
	int	sf_ecx;	
	struct	sigcontext sf_sc;
} ;

extern int kstack[];

/*
 * 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.
 */
void
sendsig(catcher, sig, mask, code)
	sig_t catcher;
	int sig, mask;
	unsigned code;
{
	register struct proc *p = curproc;
	register int *regs;
	register struct sigframe *fp;
	struct sigacts *psp = p->p_sigacts;
	int oonstack, frmtrap;

	regs = p->p_md.md_regs;
        oonstack = psp->ps_sigstk.ss_flags & SA_ONSTACK;
	frmtrap = curpcb->pcb_flags & FM_TRAP;
	/*
	 * 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'.
	 */
	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 - sizeof(struct sigframe));
		psp->ps_sigstk.ss_flags |= SA_ONSTACK;
	} else {
		if (frmtrap)
			fp = (struct sigframe *)(regs[tESP]
				- sizeof(struct sigframe));
		else
			fp = (struct sigframe *)(regs[sESP]
				- sizeof(struct sigframe));
	}

	if ((unsigned)fp <= USRSTACK - ctob(p->p_vmspace->vm_ssize)) 
		(void)grow(p, (unsigned)fp);

	if (useracc((caddr_t)fp, sizeof (struct sigframe), B_WRITE) == 0) {
		/*
		 * 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;
	}

	/* 
	 * Build the argument list for the signal handler.
	 */
	fp->sf_signum = sig;
	fp->sf_code = code;
	fp->sf_scp = &fp->sf_sc;
	fp->sf_handler = catcher;

	/* save scratch registers */
	if(frmtrap) {
		fp->sf_eax = regs[tEAX];
		fp->sf_edx = regs[tEDX];
		fp->sf_ecx = regs[tECX];
	} else {
		fp->sf_eax = regs[sEAX];
		fp->sf_edx = regs[sEDX];
		fp->sf_ecx = regs[sECX];
	}
	/*
	 * Build the signal context to be used by sigreturn.
	 */
	fp->sf_sc.sc_onstack = oonstack;
	fp->sf_sc.sc_mask = mask;
	if(frmtrap) {
		fp->sf_sc.sc_sp = regs[tESP];
		fp->sf_sc.sc_fp = regs[tEBP];
		fp->sf_sc.sc_pc = regs[tEIP];
		fp->sf_sc.sc_ps = regs[tEFLAGS];
		regs[tESP] = (int)fp;
		regs[tEIP] = (int)((struct pcb *)kstack)->pcb_sigc;
	} else {
		fp->sf_sc.sc_sp = regs[sESP];
		fp->sf_sc.sc_fp = regs[sEBP];
		fp->sf_sc.sc_pc = regs[sEIP];
		fp->sf_sc.sc_ps = regs[sEFLAGS];
		regs[sESP] = (int)fp;
		regs[sEIP] = (int)((struct pcb *)kstack)->pcb_sigc;
	}
}

/*
 * 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;
};
sigreturn(p, uap, retval)
	struct proc *p;
	struct sigreturn_args *uap;
	int *retval;
{
	register struct sigcontext *scp;
	register struct sigframe *fp;
	register int *regs = p->p_md.md_regs;


	fp = (struct sigframe *) regs[sESP] ;

	if (useracc((caddr_t)fp, sizeof (*fp), 0) == 0)
		return(EINVAL);

	/* restore scratch registers */
	regs[sEAX] = fp->sf_eax ;
	regs[sEDX] = fp->sf_edx ;
	regs[sECX] = fp->sf_ecx ;

	scp = fp->sf_scp;
	if (useracc((caddr_t)scp, sizeof (*scp), 0) == 0)
		return(EINVAL);
#ifdef notyet
	if ((scp->sc_ps & PSL_MBZ) != 0 || (scp->sc_ps & PSL_MBO) != PSL_MBO) {
		return(EINVAL);
	}
#endif
	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 &~
	    (sigmask(SIGKILL)|sigmask(SIGCONT)|sigmask(SIGSTOP));
	regs[sEBP] = scp->sc_fp;
	regs[sESP] = scp->sc_sp;
	regs[sEIP] = scp->sc_pc;
	regs[sEFLAGS] = scp->sc_ps;
	return(EJUSTRETURN);
}

int	waittime = -1;

boot(arghowto)
	int arghowto;
{
	register long dummy;		/* r12 is reserved */
	register int howto;		/* r11 == how to boot */
	register int devtype;		/* r10 == major of root dev */
	extern char *panicstr;
extern int cold;

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

		waittime = 0;
		(void) splnet();
		printf("syncing disks... ");
		/*
		 * Release inodes held by texts before update.
		 */
		if (panicstr == 0)
			vnode_pager_umount(NULL);
		sync((struct sigcontext *)0);

		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");
		DELAY(10000);			/* wait for printf to finish */
	}
	splhigh();
	devtype = major(rootdev);
	if (howto&RB_HALT) {
		printf("halting (in tight loop); hit reset\n\n");
		splx(0xfffd);	/* all but keyboard XXX */
		for (;;) ;
	} else {
		if (howto & RB_DUMP) {
			dumpsys();	
			/*NOTREACHED*/
		}
	}
#ifdef lint
	dummy = 0; dummy = dummy;
	printf("howto %d, devtype %d\n", arghowto, devtype);
#endif
#ifdef	notdef
	pg("pausing (hit any key to reset)");
#endif
	reset_cpu();
	for(;;) ;
	/*NOTREACHED*/
}

int	dumpmag = 0x8fca0101;	/* magic number for savecore */
int	dumpsize = 0;		/* also for savecore */
/*
 * 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()
{

	if (dumpdev == NODEV)
		return;
	if ((minor(dumpdev)&07) != 1)
		return;
	dumpsize = physmem;
	printf("\ndumping to dev %x, offset %d\n", dumpdev, dumplo);