⭐ 欢迎来到虫虫下载站! | 📦 资源下载 📁 资源专辑 ℹ️ 关于我们
⭐ 虫虫下载站

📄 kvm_hp300.c

📁 早期freebsd实现
💻 C
字号:
/*- * Copyright (c) 1989, 1992, 1993 *	The Regents of the University of California.  All rights reserved. * * This code is derived from software developed by the Computer Systems * Engineering group at Lawrence Berkeley Laboratory under DARPA contract * BG 91-66 and contributed to Berkeley. * * 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. */#if defined(LIBC_SCCS) && !defined(lint)static char sccsid[] = "@(#)kvm_hp300.c	8.1 (Berkeley) 6/4/93";#endif /* LIBC_SCCS and not lint *//* * Hp300 machine dependent routines for kvm.  Hopefully, the forthcoming  * vm code will one day obsolete this module. */#include <sys/param.h>#include <sys/user.h>#include <sys/proc.h>#include <sys/stat.h>#include <unistd.h>#include <nlist.h>#include <kvm.h>#include <vm/vm.h>#include <vm/vm_param.h>#include <limits.h>#include <db.h>#include "kvm_private.h"#if defined(hp300)#include <hp300/hp300/pte.h>#endif#if defined(luna68k)#include <luna68k/luna68k/pte.h>#endif#ifndef btop#define	btop(x)		(((unsigned)(x)) >> PGSHIFT)	/* XXX */#define	ptob(x)		((caddr_t)((x) << PGSHIFT))	/* XXX */#endifstruct vmstate {	u_long lowram;	int mmutype;	struct ste *Sysseg;};#define KREAD(kd, addr, p)\	(kvm_read(kd, addr, (char *)(p), sizeof(*(p))) != sizeof(*(p)))void_kvm_freevtop(kd)	kvm_t *kd;{	if (kd->vmst != 0)		free(kd->vmst);}int_kvm_initvtop(kd)	kvm_t *kd;{	struct vmstate *vm;	struct nlist nlist[4];	vm = (struct vmstate *)_kvm_malloc(kd, sizeof(*vm));	if (vm == 0)		return (-1);	kd->vmst = vm;	nlist[0].n_name = "_lowram";	nlist[1].n_name = "_mmutype";	nlist[2].n_name = "_Sysseg";	nlist[3].n_name = 0;	if (kvm_nlist(kd, nlist) != 0) {		_kvm_err(kd, kd->program, "bad namelist");		return (-1);	}	vm->Sysseg = 0;	if (KREAD(kd, (u_long)nlist[0].n_value, &vm->lowram)) {		_kvm_err(kd, kd->program, "cannot read lowram");		return (-1);	}	if (KREAD(kd, (u_long)nlist[1].n_value, &vm->mmutype)) {		_kvm_err(kd, kd->program, "cannot read mmutype");		return (-1);	}	if (KREAD(kd, (u_long)nlist[2].n_value, &vm->Sysseg)) {		_kvm_err(kd, kd->program, "cannot read segment table");		return (-1);	}	return (0);}static int_kvm_vatop(kd, sta, va, pa)	kvm_t *kd;	struct ste *sta;	u_long va;	u_long *pa;{	register struct vmstate *vm;	register u_long lowram;	register u_long addr;	int p, ste, pte;	int offset;	if (ISALIVE(kd)) {		_kvm_err(kd, 0, "vatop called in live kernel!");		return((off_t)0);	}	vm = kd->vmst;	offset = va & PGOFSET;	/*	 * If we are initializing (kernel segment table pointer not yet set)	 * then return pa == va to avoid infinite recursion.	 */	if (vm->Sysseg == 0) {		*pa = va;		return (NBPG - offset);	}	lowram = vm->lowram;	if (vm->mmutype == -2) {		struct ste *sta2;		addr = (u_long)&sta[va >> SG4_SHIFT1];		/*		 * Can't use KREAD to read kernel segment table entries.		 * Fortunately it is 1-to-1 mapped so we don't have to. 		 */		if (sta == vm->Sysseg) {			if (lseek(kd->pmfd, (off_t)addr, 0) == -1 ||			    read(kd->pmfd, (char *)&ste, sizeof(ste)) < 0)				goto invalid;		} else if (KREAD(kd, addr, &ste))			goto invalid;		if ((ste & SG_V) == 0) {			_kvm_err(kd, 0, "invalid level 1 descriptor (%x)",				 ste);			return((off_t)0);		}		sta2 = (struct ste *)(ste & SG4_ADDR1);		addr = (u_long)&sta2[(va & SG4_MASK2) >> SG4_SHIFT2];		/*		 * Address from level 1 STE is a physical address,		 * so don't use kvm_read.		 */		if (lseek(kd->pmfd, (off_t)(addr - lowram), 0) == -1 || 		    read(kd->pmfd, (char *)&ste, sizeof(ste)) < 0)			goto invalid;		if ((ste & SG_V) == 0) {			_kvm_err(kd, 0, "invalid level 2 descriptor (%x)",				 ste);			return((off_t)0);		}		sta2 = (struct ste *)(ste & SG4_ADDR2);		addr = (u_long)&sta2[(va & SG4_MASK3) >> SG4_SHIFT3];	} else {		addr = (u_long)&sta[va >> SEGSHIFT];		/*		 * Can't use KREAD to read kernel segment table entries.		 * Fortunately it is 1-to-1 mapped so we don't have to. 		 */		if (sta == vm->Sysseg) {			if (lseek(kd->pmfd, (off_t)addr, 0) == -1 ||			    read(kd->pmfd, (char *)&ste, sizeof(ste)) < 0)				goto invalid;		} else if (KREAD(kd, addr, &ste))			goto invalid;		if ((ste & SG_V) == 0) {			_kvm_err(kd, 0, "invalid segment (%x)", ste);			return((off_t)0);		}		p = btop(va & SG_PMASK);		addr = (ste & SG_FRAME) + (p * sizeof(struct pte));	}	/*	 * Address from STE is a physical address so don't use kvm_read.	 */	if (lseek(kd->pmfd, (off_t)(addr - lowram), 0) == -1 || 	    read(kd->pmfd, (char *)&pte, sizeof(pte)) < 0)		goto invalid;	addr = pte & PG_FRAME;	if (pte == PG_NV) {		_kvm_err(kd, 0, "page not valid");		return (0);	}	*pa = addr - lowram + offset;		return (NBPG - offset);invalid:	_kvm_err(kd, 0, "invalid address (%x)", va);	return (0);}int_kvm_kvatop(kd, va, pa)	kvm_t *kd;	u_long va;	u_long *pa;{	return (_kvm_vatop(kd, (u_long)kd->vmst->Sysseg, va, pa));}/* * Translate a user virtual address to a physical address. */int_kvm_uvatop(kd, p, va, pa)	kvm_t *kd;	const struct proc *p;	u_long va;	u_long *pa;{	register struct vmspace *vms = p->p_vmspace;	int kva;	/*	 * If this is a live kernel we just look it up in the kernel	 * virtually allocated flat 4mb page table (i.e. let the kernel	 * do the table walk).  In this way, we avoid needing to know	 * the MMU type.	 */	if (ISALIVE(kd)) {		struct pte *ptab;		int pte, offset;		kva = (int)&vms->vm_pmap.pm_ptab;		if (KREAD(kd, kva, &ptab)) {			_kvm_err(kd, 0, "invalid address (%x)", va);			return (0);		}		kva = (int)&ptab[btop(va)];		if (KREAD(kd, kva, &pte) || (pte & PG_V) == 0) {			_kvm_err(kd, 0, "invalid address (%x)", va);			return (0);		}		offset = va & PGOFSET;		*pa = (pte & PG_FRAME) | offset;		return (NBPG - offset);	}	/*	 * Otherwise, we just walk the table ourself.	 */	kva = (int)&vms->vm_pmap.pm_stab;	if (KREAD(kd, kva, &kva)) {		_kvm_err(kd, 0, "invalid address (%x)", va);		return (0);	}	return (_kvm_vatop(kd, kva, va, pa));}

⌨️ 快捷键说明

复制代码 Ctrl + C
搜索代码 Ctrl + F
全屏模式 F11
切换主题 Ctrl + Shift + D
显示快捷键 ?
增大字号 Ctrl + =
减小字号 Ctrl + -