common.c

来自「xen虚拟机源代码安装包」· C语言 代码 · 共 1,973 行 · 第 1/5 页

/******************************************************************************
 * arch/x86/mm/shadow/common.c
 *
 * Shadow code that does not need to be multiply compiled.
 * Parts of this code are Copyright (c) 2006 by XenSource Inc.
 * Parts of this code are Copyright (c) 2006 by Michael A Fetterman
 * Parts based on earlier work by Michael A Fetterman, Ian Pratt et al.
 * 
 * This program is free software; you can redistribute it and/or modify
 * it under the terms of the GNU General Public License as published by
 * the Free Software Foundation; either version 2 of the License, or
 * (at your option) any later version.
 *
 * This program is distributed in the hope that it will be useful,
 * but WITHOUT ANY WARRANTY; without even the implied warranty of
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
 * GNU General Public License for more details.
 *
 * You should have received a copy of the GNU General Public License
 * along with this program; if not, write to the Free Software
 * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
 */

#include <xen/config.h>
#include <xen/types.h>
#include <xen/mm.h>
#include <xen/trace.h>
#include <xen/sched.h>
#include <xen/perfc.h>
#include <xen/irq.h>
#include <xen/domain_page.h>
#include <xen/guest_access.h>
#include <xen/keyhandler.h>
#include <asm/event.h>
#include <asm/page.h>
#include <asm/current.h>
#include <asm/flushtlb.h>
#include <asm/shadow.h>
#include <xen/numa.h>
#include "private.h"


/* Set up the shadow-specific parts of a domain struct at start of day.
 * Called for every domain from arch_domain_create() */
void shadow_domain_init(struct domain *d)
{
    int i;
    shadow_lock_init(d);
    for ( i = 0; i <= SHADOW_MAX_ORDER; i++ )
        INIT_LIST_HEAD(&d->arch.paging.shadow.freelists[i]);
    INIT_LIST_HEAD(&d->arch.paging.shadow.p2m_freelist);
    INIT_LIST_HEAD(&d->arch.paging.shadow.pinned_shadows);

    /* Use shadow pagetables for log-dirty support */
    paging_log_dirty_init(d, shadow_enable_log_dirty, 
                          shadow_disable_log_dirty, shadow_clean_dirty_bitmap);

#if (SHADOW_OPTIMIZATIONS & SHOPT_OUT_OF_SYNC)
    d->arch.paging.shadow.oos_active = 0;
#endif
}

/* Setup the shadow-specfic parts of a vcpu struct. Note: The most important
 * job is to initialize the update_paging_modes() function pointer, which is
 * used to initialized the rest of resources. Therefore, it really does not
 * matter to have v->arch.paging.mode pointing to any mode, as long as it can
 * be compiled.
 */
void shadow_vcpu_init(struct vcpu *v)
{
#if (SHADOW_OPTIMIZATIONS & SHOPT_OUT_OF_SYNC)
    int i, j;

    for ( i = 0; i < SHADOW_OOS_PAGES; i++ )
    {
        v->arch.paging.shadow.oos[i] = _mfn(INVALID_MFN);
        v->arch.paging.shadow.oos_snapshot[i] = _mfn(INVALID_MFN);
        for ( j = 0; j < SHADOW_OOS_FIXUPS; j++ )
            v->arch.paging.shadow.oos_fixup[i].smfn[j] = _mfn(INVALID_MFN);
    }
#endif

    v->arch.paging.mode = &SHADOW_INTERNAL_NAME(sh_paging_mode, 3);
}

#if SHADOW_AUDIT
int shadow_audit_enable = 0;

static void shadow_audit_key(unsigned char key)
{
    shadow_audit_enable = !shadow_audit_enable;
    printk("%s shadow_audit_enable=%d\n",
           __func__, shadow_audit_enable);
}

static int __init shadow_audit_key_init(void)
{
    register_keyhandler(
        'O', shadow_audit_key,  "toggle shadow audits");
    return 0;
}
__initcall(shadow_audit_key_init);
#endif /* SHADOW_AUDIT */

int _shadow_mode_refcounts(struct domain *d)
{
    return shadow_mode_refcounts(d);
}


/**************************************************************************/
/* x86 emulator support for the shadow code
 */

struct segment_register *hvm_get_seg_reg(
    enum x86_segment seg, struct sh_emulate_ctxt *sh_ctxt)
{
    struct segment_register *seg_reg = &sh_ctxt->seg_reg[seg];
    if ( !__test_and_set_bit(seg, &sh_ctxt->valid_seg_regs) )
        hvm_get_segment_register(current, seg, seg_reg);
    return seg_reg;
}

static int hvm_translate_linear_addr(
    enum x86_segment seg,
    unsigned long offset,
    unsigned int bytes,
    enum hvm_access_type access_type,
    struct sh_emulate_ctxt *sh_ctxt,
    unsigned long *paddr)
{
    struct segment_register *reg = hvm_get_seg_reg(seg, sh_ctxt);
    int okay;

    okay = hvm_virtual_to_linear_addr(
        seg, reg, offset, bytes, access_type, sh_ctxt->ctxt.addr_size, paddr);

    if ( !okay )
    {
        hvm_inject_exception(TRAP_gp_fault, 0, 0);
        return X86EMUL_EXCEPTION;
    }

    return 0;
}

static int
hvm_read(enum x86_segment seg,
         unsigned long offset,
         void *p_data,
         unsigned int bytes,
         enum hvm_access_type access_type,
         struct sh_emulate_ctxt *sh_ctxt)
{
    unsigned long addr;
    int rc;

    rc = hvm_translate_linear_addr(
        seg, offset, bytes, access_type, sh_ctxt, &addr);
    if ( rc )
        return rc;

    if ( access_type == hvm_access_insn_fetch )
        rc = hvm_fetch_from_guest_virt(p_data, addr, bytes, 0);
    else
        rc = hvm_copy_from_guest_virt(p_data, addr, bytes, 0);

    switch ( rc )
    {
    case HVMCOPY_okay:
        return X86EMUL_OKAY;
    case HVMCOPY_bad_gva_to_gfn:
        return X86EMUL_EXCEPTION;
    default:
        break;
    }

    return X86EMUL_UNHANDLEABLE;
}

static int
hvm_emulate_read(enum x86_segment seg,
                 unsigned long offset,
                 void *p_data,
                 unsigned int bytes,
                 struct x86_emulate_ctxt *ctxt)
{
    if ( !is_x86_user_segment(seg) )
        return X86EMUL_UNHANDLEABLE;
    return hvm_read(seg, offset, p_data, bytes, hvm_access_read,
                    container_of(ctxt, struct sh_emulate_ctxt, ctxt));
}

static int
hvm_emulate_insn_fetch(enum x86_segment seg,
                       unsigned long offset,
                       void *p_data,
                       unsigned int bytes,
                       struct x86_emulate_ctxt *ctxt)
{
    struct sh_emulate_ctxt *sh_ctxt =
        container_of(ctxt, struct sh_emulate_ctxt, ctxt);
    unsigned int insn_off = offset - sh_ctxt->insn_buf_eip;

    ASSERT(seg == x86_seg_cs);

    /* Fall back if requested bytes are not in the prefetch cache. */
    if ( unlikely((insn_off + bytes) > sh_ctxt->insn_buf_bytes) )
        return hvm_read(seg, offset, p_data, bytes,
                        hvm_access_insn_fetch, sh_ctxt);

    /* Hit the cache. Simple memcpy. */
    memcpy(p_data, &sh_ctxt->insn_buf[insn_off], bytes);
    return X86EMUL_OKAY;
}

static int
hvm_emulate_write(enum x86_segment seg,
                  unsigned long offset,
                  void *p_data,
                  unsigned int bytes,
                  struct x86_emulate_ctxt *ctxt)
{
    struct sh_emulate_ctxt *sh_ctxt =
        container_of(ctxt, struct sh_emulate_ctxt, ctxt);
    struct vcpu *v = current;
    unsigned long addr;
    int rc;

    if ( !is_x86_user_segment(seg) )
        return X86EMUL_UNHANDLEABLE;

    /* How many emulations could we save if we unshadowed on stack writes? */
    if ( seg == x86_seg_ss )
        perfc_incr(shadow_fault_emulate_stack);

    rc = hvm_translate_linear_addr(
        seg, offset, bytes, hvm_access_write, sh_ctxt, &addr);
    if ( rc )
        return rc;

    return v->arch.paging.mode->shadow.x86_emulate_write(
        v, addr, p_data, bytes, sh_ctxt);
}

static int 
hvm_emulate_cmpxchg(enum x86_segment seg,
                    unsigned long offset,
                    void *p_old,
                    void *p_new,
                    unsigned int bytes,
                    struct x86_emulate_ctxt *ctxt)
{
    struct sh_emulate_ctxt *sh_ctxt =
        container_of(ctxt, struct sh_emulate_ctxt, ctxt);
    struct vcpu *v = current;
    unsigned long addr, old[2], new[2];
    int rc;

    if ( !is_x86_user_segment(seg) )
        return X86EMUL_UNHANDLEABLE;

    rc = hvm_translate_linear_addr(
        seg, offset, bytes, hvm_access_write, sh_ctxt, &addr);
    if ( rc )
        return rc;

    old[0] = new[0] = 0;
    memcpy(old, p_old, bytes);
    memcpy(new, p_new, bytes);

    if ( bytes <= sizeof(long) )
        return v->arch.paging.mode->shadow.x86_emulate_cmpxchg(
            v, addr, old[0], new[0], bytes, sh_ctxt);

#ifdef __i386__
    if ( bytes == 8 )
        return v->arch.paging.mode->shadow.x86_emulate_cmpxchg8b(
            v, addr, old[0], old[1], new[0], new[1], sh_ctxt);
#endif

    return X86EMUL_UNHANDLEABLE;
}

static struct x86_emulate_ops hvm_shadow_emulator_ops = {
    .read       = hvm_emulate_read,
    .insn_fetch = hvm_emulate_insn_fetch,
    .write      = hvm_emulate_write,
    .cmpxchg    = hvm_emulate_cmpxchg,
};

static int
pv_emulate_read(enum x86_segment seg,
                unsigned long offset,
                void *p_data,
                unsigned int bytes,
                struct x86_emulate_ctxt *ctxt)
{
    unsigned int rc;

    if ( !is_x86_user_segment(seg) )
        return X86EMUL_UNHANDLEABLE;

    if ( (rc = copy_from_user(p_data, (void *)offset, bytes)) != 0 )
    {
        propagate_page_fault(offset + bytes - rc, 0); /* read fault */
        return X86EMUL_EXCEPTION;
    }

    return X86EMUL_OKAY;
}

static int
pv_emulate_write(enum x86_segment seg,
                 unsigned long offset,
                 void *p_data,
                 unsigned int bytes,
                 struct x86_emulate_ctxt *ctxt)
{
    struct sh_emulate_ctxt *sh_ctxt =
        container_of(ctxt, struct sh_emulate_ctxt, ctxt);
    struct vcpu *v = current;
    if ( !is_x86_user_segment(seg) )
        return X86EMUL_UNHANDLEABLE;
    return v->arch.paging.mode->shadow.x86_emulate_write(
        v, offset, p_data, bytes, sh_ctxt);
}

static int 
pv_emulate_cmpxchg(enum x86_segment seg,
                   unsigned long offset,
                   void *p_old,
                   void *p_new,
                   unsigned int bytes,
                   struct x86_emulate_ctxt *ctxt)
{
    struct sh_emulate_ctxt *sh_ctxt =
        container_of(ctxt, struct sh_emulate_ctxt, ctxt);
    unsigned long old[2], new[2];
    struct vcpu *v = current;

    if ( !is_x86_user_segment(seg) )
        return X86EMUL_UNHANDLEABLE;

    old[0] = new[0] = 0;
    memcpy(old, p_old, bytes);
    memcpy(new, p_new, bytes);

    if ( bytes <= sizeof(long) )
        return v->arch.paging.mode->shadow.x86_emulate_cmpxchg(
            v, offset, old[0], new[0], bytes, sh_ctxt);

#ifdef __i386__
    if ( bytes == 8 )
        return v->arch.paging.mode->shadow.x86_emulate_cmpxchg8b(
            v, offset, old[0], old[1], new[0], new[1], sh_ctxt);
#endif

    return X86EMUL_UNHANDLEABLE;
}

static struct x86_emulate_ops pv_shadow_emulator_ops = {
    .read       = pv_emulate_read,
    .insn_fetch = pv_emulate_read,
    .write      = pv_emulate_write,
    .cmpxchg    = pv_emulate_cmpxchg,
};

struct x86_emulate_ops *shadow_init_emulation(
    struct sh_emulate_ctxt *sh_ctxt, struct cpu_user_regs *regs)
{
    struct segment_register *creg, *sreg;
    struct vcpu *v = current;
    unsigned long addr;

    sh_ctxt->ctxt.regs = regs;
    sh_ctxt->ctxt.force_writeback = 0;

    if ( !is_hvm_vcpu(v) )
    {
        sh_ctxt->ctxt.addr_size = sh_ctxt->ctxt.sp_size = BITS_PER_LONG;
        return &pv_shadow_emulator_ops;
    }

    /* Segment cache initialisation. Primed with CS. */
    sh_ctxt->valid_seg_regs = 0;
    creg = hvm_get_seg_reg(x86_seg_cs, sh_ctxt);

    /* Work out the emulation mode. */
    if ( hvm_long_mode_enabled(v) && creg->attr.fields.l )
    {
        sh_ctxt->ctxt.addr_size = sh_ctxt->ctxt.sp_size = 64;
    }
    else
    {