mm.c

xen 3.2.2 源码

 * Update ref counts to shadow tables appropriately.
 */
void update_cr3(struct vcpu *v)
{
    unsigned long cr3_mfn=0;

    if ( paging_mode_enabled(v->domain) )
    {
        paging_update_cr3(v);
        return;
    }

#if CONFIG_PAGING_LEVELS == 4
    if ( !(v->arch.flags & TF_kernel_mode) )
        cr3_mfn = pagetable_get_pfn(v->arch.guest_table_user);
    else
#endif
        cr3_mfn = pagetable_get_pfn(v->arch.guest_table);

    make_cr3(v, cr3_mfn);
}


static void invalidate_shadow_ldt(struct vcpu *v)
{
    int i;
    unsigned long pfn;
    struct page_info *page;
    
    if ( v->arch.shadow_ldt_mapcnt == 0 )
        return;

    v->arch.shadow_ldt_mapcnt = 0;

    for ( i = 16; i < 32; i++ )
    {
        pfn = l1e_get_pfn(v->arch.perdomain_ptes[i]);
        if ( pfn == 0 ) continue;
        l1e_write(&v->arch.perdomain_ptes[i], l1e_empty());
        page = mfn_to_page(pfn);
        ASSERT_PAGE_IS_TYPE(page, PGT_ldt_page);
        ASSERT_PAGE_IS_DOMAIN(page, v->domain);
        put_page_and_type(page);
    }

    /* Dispose of the (now possibly invalid) mappings from the TLB.  */
    if ( v == current )
        queue_deferred_ops(v->domain, DOP_FLUSH_TLB | DOP_RELOAD_LDT);
    else
        flush_tlb_mask(v->domain->domain_dirty_cpumask);
}


static int alloc_segdesc_page(struct page_info *page)
{
    struct desc_struct *descs;
    int i;

    descs = map_domain_page(page_to_mfn(page));

    for ( i = 0; i < 512; i++ )
        if ( unlikely(!check_descriptor(page_get_owner(page), &descs[i])) )
            goto fail;

    unmap_domain_page(descs);
    return 1;

 fail:
    unmap_domain_page(descs);
    return 0;
}


/* Map shadow page at offset @off. */
int map_ldt_shadow_page(unsigned int off)
{
    struct vcpu *v = current;
    struct domain *d = v->domain;
    unsigned long gmfn, mfn;
    l1_pgentry_t l1e, nl1e;
    unsigned long gva = v->arch.guest_context.ldt_base + (off << PAGE_SHIFT);
    int okay;

    BUG_ON(unlikely(in_irq()));

    guest_get_eff_kern_l1e(v, gva, &l1e);
    if ( unlikely(!(l1e_get_flags(l1e) & _PAGE_PRESENT)) )
        return 0;

    gmfn = l1e_get_pfn(l1e);
    mfn = gmfn_to_mfn(d, gmfn);
    if ( unlikely(!mfn_valid(mfn)) )
        return 0;

    okay = get_page_and_type(mfn_to_page(mfn), d, PGT_ldt_page);
    if ( unlikely(!okay) )
        return 0;

    nl1e = l1e_from_pfn(mfn, l1e_get_flags(l1e) | _PAGE_RW);

    l1e_write(&v->arch.perdomain_ptes[off + 16], nl1e);
    v->arch.shadow_ldt_mapcnt++;

    return 1;
}


static int get_page_from_pagenr(unsigned long page_nr, struct domain *d)
{
    struct page_info *page = mfn_to_page(page_nr);

    if ( unlikely(!mfn_valid(page_nr)) || unlikely(!get_page(page, d)) )
    {
        MEM_LOG("Could not get page ref for pfn %lx", page_nr);
        return 0;
    }

    return 1;
}


static int get_page_and_type_from_pagenr(unsigned long page_nr, 
                                         unsigned long type,
                                         struct domain *d)
{
    struct page_info *page = mfn_to_page(page_nr);

    if ( unlikely(!get_page_from_pagenr(page_nr, d)) )
        return 0;

    if ( unlikely(!get_page_type(page, type)) )
    {
        put_page(page);
        return 0;
    }

    return 1;
}

/*
 * We allow root tables to map each other (a.k.a. linear page tables). It
 * needs some special care with reference counts and access permissions:
 *  1. The mapping entry must be read-only, or the guest may get write access
 *     to its own PTEs.
 *  2. We must only bump the reference counts for an *already validated*
 *     L2 table, or we can end up in a deadlock in get_page_type() by waiting
 *     on a validation that is required to complete that validation.
 *  3. We only need to increment the reference counts for the mapped page
 *     frame if it is mapped by a different root table. This is sufficient and
 *     also necessary to allow validation of a root table mapping itself.
 */
#define define_get_linear_pagetable(level)                                  \
static int                                                                  \
get_##level##_linear_pagetable(                                             \
    level##_pgentry_t pde, unsigned long pde_pfn, struct domain *d)         \
{                                                                           \
    unsigned long x, y;                                                     \
    struct page_info *page;                                                 \
    unsigned long pfn;                                                      \
                                                                            \
    if ( (level##e_get_flags(pde) & _PAGE_RW) )                             \
    {                                                                       \
        MEM_LOG("Attempt to create linear p.t. with write perms");          \
        return 0;                                                           \
    }                                                                       \
                                                                            \
    if ( (pfn = level##e_get_pfn(pde)) != pde_pfn )                         \
    {                                                                       \
        /* Make sure the mapped frame belongs to the correct domain. */     \
        if ( unlikely(!get_page_from_pagenr(pfn, d)) )                      \
            return 0;                                                       \
                                                                            \
        /*                                                                  \
         * Ensure that the mapped frame is an already-validated page table. \
         * If so, atomically increment the count (checking for overflow).   \
         */                                                                 \
        page = mfn_to_page(pfn);                                            \
        y = page->u.inuse.type_info;                                        \
        do {                                                                \
            x = y;                                                          \
            if ( unlikely((x & PGT_count_mask) == PGT_count_mask) ||        \
                 unlikely((x & (PGT_type_mask|PGT_validated)) !=            \
                          (PGT_##level##_page_table|PGT_validated)) )       \
            {                                                               \
                put_page(page);                                             \
                return 0;                                                   \
            }                                                               \
        }                                                                   \
        while ( (y = cmpxchg(&page->u.inuse.type_info, x, x + 1)) != x );   \
    }                                                                       \
                                                                            \
    return 1;                                                               \
}


int is_iomem_page(unsigned long mfn)
{
    return (!mfn_valid(mfn) || (page_get_owner(mfn_to_page(mfn)) == dom_io));
}


int
get_page_from_l1e(
    l1_pgentry_t l1e, struct domain *d)
{
    unsigned long mfn = l1e_get_pfn(l1e);
    struct page_info *page = mfn_to_page(mfn);
    uint32_t l1f = l1e_get_flags(l1e);
    struct vcpu *curr = current;
    int okay;

    if ( !(l1f & _PAGE_PRESENT) )
        return 1;

    if ( unlikely(l1f & l1_disallow_mask(d)) )
    {
        MEM_LOG("Bad L1 flags %x", l1f & l1_disallow_mask(d));
        return 0;
    }

    if ( is_iomem_page(mfn) )
    {
        /* DOMID_IO reverts to caller for privilege checks. */
        if ( d == dom_io )
            d = curr->domain;

        if ( !iomem_access_permitted(d, mfn, mfn) )
        {
            if ( mfn != (PADDR_MASK >> PAGE_SHIFT) ) /* INVALID_MFN? */
                MEM_LOG("Non-privileged (%u) attempt to map I/O space %08lx", 
                        d->domain_id, mfn);
            return 0;
        }

        return 1;
    }

    /* Foreign mappings into guests in shadow external mode don't
     * contribute to writeable mapping refcounts.  (This allows the
     * qemu-dm helper process in dom0 to map the domain's memory without
     * messing up the count of "real" writable mappings.) */
    okay = (((l1f & _PAGE_RW) && 
             !(unlikely(paging_mode_external(d) && (d != curr->domain))))
            ? get_page_and_type(page, d, PGT_writable_page)
            : get_page(page, d));
    if ( !okay )
    {
        MEM_LOG("Error getting mfn %lx (pfn %lx) from L1 entry %" PRIpte
                " for dom%d",
                mfn, get_gpfn_from_mfn(mfn),
                l1e_get_intpte(l1e), d->domain_id);
    }
    else if ( pte_flags_to_cacheattr(l1f) !=
              ((page->count_info >> PGC_cacheattr_base) & 7) )
    {
        uint32_t x, nx, y = page->count_info;
        uint32_t cacheattr = pte_flags_to_cacheattr(l1f);

        if ( is_xen_heap_page(page) )
        {
            if ( (l1f & _PAGE_RW) &&
                 !(unlikely(paging_mode_external(d) &&
                            (d != curr->domain))) )
                put_page_type(page);
            put_page(page);
            MEM_LOG("Attempt to change cache attributes of Xen heap page");
            return 0;
        }

        while ( ((y >> PGC_cacheattr_base) & 7) != cacheattr )
        {
            x  = y;
            nx = (x & ~PGC_cacheattr_mask) | (cacheattr << PGC_cacheattr_base);
            y  = cmpxchg(&page->count_info, x, nx);
        }

#ifdef __x86_64__
        map_pages_to_xen((unsigned long)mfn_to_virt(mfn), mfn, 1,
                         PAGE_HYPERVISOR | cacheattr_to_pte_flags(cacheattr));
#endif
    }

    return okay;
}


/* NB. Virtual address 'l2e' maps to a machine address within frame 'pfn'. */
define_get_linear_pagetable(l2);
static int
get_page_from_l2e(
    l2_pgentry_t l2e, unsigned long pfn, struct domain *d)
{
    int rc;

    if ( !(l2e_get_flags(l2e) & _PAGE_PRESENT) )
        return 1;

    if ( unlikely((l2e_get_flags(l2e) & L2_DISALLOW_MASK)) )
    {
        MEM_LOG("Bad L2 flags %x", l2e_get_flags(l2e) & L2_DISALLOW_MASK);
        return 0;
    }

    rc = get_page_and_type_from_pagenr(l2e_get_pfn(l2e), PGT_l1_page_table, d);
    if ( unlikely(!rc) )
        rc = get_l2_linear_pagetable(l2e, pfn, d);

    return rc;
}


#if CONFIG_PAGING_LEVELS >= 3
define_get_linear_pagetable(l3);
static int
get_page_from_l3e(
    l3_pgentry_t l3e, unsigned long pfn, struct domain *d)
{
    int rc;

    if ( !(l3e_get_flags(l3e) & _PAGE_PRESENT) )
        return 1;

    if ( unlikely((l3e_get_flags(l3e) & l3_disallow_mask(d))) )
    {
        MEM_LOG("Bad L3 flags %x", l3e_get_flags(l3e) & l3_disallow_mask(d));
        return 0;
    }

    rc = get_page_and_type_from_pagenr(l3e_get_pfn(l3e), PGT_l2_page_table, d);
    if ( unlikely(!rc) )
        rc = get_l3_linear_pagetable(l3e, pfn, d);

    return rc;
}
#endif /* 3 level */

#if CONFIG_PAGING_LEVELS >= 4
define_get_linear_pagetable(l4);
static int
get_page_from_l4e(
    l4_pgentry_t l4e, unsigned long pfn, struct domain *d)
{
    int rc;

    if ( !(l4e_get_flags(l4e) & _PAGE_PRESENT) )
        return 1;

    if ( unlikely((l4e_get_flags(l4e) & L4_DISALLOW_MASK)) )
    {
        MEM_LOG("Bad L4 flags %x", l4e_get_flags(l4e) & L4_DISALLOW_MASK);
        return 0;
    }

    rc = get_page_and_type_from_pagenr(l4e_get_pfn(l4e), PGT_l3_page_table, d);
    if ( unlikely(!rc) )
        rc = get_l4_linear_pagetable(l4e, pfn, d);

    return rc;
}
#endif /* 4 level */

#ifdef __x86_64__

#ifdef USER_MAPPINGS_ARE_GLOBAL
#define adjust_guest_l1e(pl1e, d)                                            \
    do {                                                                     \
        if ( likely(l1e_get_flags((pl1e)) & _PAGE_PRESENT) &&                \
             likely(!is_pv_32on64_domain(d)) )                               \
        {                                                                    \
            /* _PAGE_GUEST_KERNEL page cannot have the Global bit set. */    \
            if ( (l1e_get_flags((pl1e)) & (_PAGE_GUEST_KERNEL|_PAGE_GLOBAL)) \
                 == (_PAGE_GUEST_KERNEL|_PAGE_GLOBAL) )                      \
                MEM_LOG("Global bit is set to kernel page %lx",              \
                        l1e_get_pfn((pl1e)));                                \
            if ( !(l1e_get_flags((pl1e)) & _PAGE_USER) )                     \
                l1e_add_flags((pl1e), (_PAGE_GUEST_KERNEL|_PAGE_USER));      \
            if ( !(l1e_get_flags((pl1e)) & _PAGE_GUEST_KERNEL) )             \
                l1e_add_flags((pl1e), (_PAGE_GLOBAL|_PAGE_USER));            \
        }                                                                    \
    } while ( 0 )
#else
#define adjust_guest_l1e(pl1e, d)                               \
    do {                                                        \
        if ( likely(l1e_get_flags((pl1e)) & _PAGE_PRESENT) &&   \
             likely(!is_pv_32on64_domain(d)) )                  \
            l1e_add_flags((pl1e), _PAGE_USER);                  \
    } while ( 0 )
#endif

#define adjust_guest_l2e(pl2e, d)                               \
    do {                                                        \
        if ( likely(l2e_get_flags((pl2e)) & _PAGE_PRESENT) &&   \
             likely(!is_pv_32on64_domain(d)) )                  \
            l2e_add_flags((pl2e), _PAGE_USER);                  \
    } while ( 0 )

#define adjust_guest_l3e(pl3e, d)                                   \
    do {                                                            \
        if ( likely(l3e_get_flags((pl3e)) & _PAGE_PRESENT) )        \
            l3e_add_flags((pl3e), likely(!is_pv_32on64_domain(d)) ? \
                                         _PAGE_USER :               \
                                         _PAGE_USER|_PAGE_RW);      \
    } while ( 0 )

#define adjust_guest_l4e(pl4e, d)                               \
    do {                                                        \
        if ( likely(l4e_get_flags((pl4e)) & _PAGE_PRESENT) &&   \
             likely(!is_pv_32on64_domain(d)) )                  \
            l4e_add_flags((pl4e), _PAGE_USER);                  \
    } while ( 0 )

#else /* !defined(__x86_64__) */

#define adjust_guest_l1e(_p, _d) ((void)(_d))