x86/64: Save/restore user pagetable pointer for x86/64 PV guests.

Signed-off-by: Keir Fraser <keir@xensource.com>

diff --git a/tools/libxc/xc_linux_restore.c b/tools/libxc/xc_linux_restore.c
index 7d1933c1ebe324e553b607e1b21ba7172f016099..6fa0388d3f5e0ba57b7cc43ad256c9c97b8e3855 100644 (file)
--- a/tools/libxc/xc_linux_restore.c
+++ b/tools/libxc/xc_linux_restore.c
@@ -19,7 +19,7 @@ static unsigned long max_mfn;
 /* virtual starting address of the hypervisor */
 static unsigned long hvirt_start;
 
-/* #levels of page tables used by the currrent guest */
+/* #levels of page tables used by the current guest */
 static unsigned int pt_levels;
 
 /* total number of pages used by the current guest */
@@ -857,6 +857,28 @@ int xc_linux_restore(int xc_handle, int io_fd,
 
         ctxt.ctrlreg[3] = xen_pfn_to_cr3(p2m[pfn]);
 
+        /* Guest pagetable (x86/64) stored in otherwise-unused CR1. */
+        if ( (pt_levels == 4) && ctxt.ctrlreg[1] )
+        {
+            pfn = xen_cr3_to_pfn(ctxt.ctrlreg[1]);
+
+            if (pfn >= max_pfn) {
+                ERROR("User PT base is bad: pfn=%lu max_pfn=%lu type=%08lx",
+                      pfn, max_pfn, pfn_type[pfn]);
+                goto out;
+            }
+
+            if ( (pfn_type[pfn] & XEN_DOMCTL_PFINFO_LTABTYPE_MASK) !=
+                 ((unsigned long)pt_levels<<XEN_DOMCTL_PFINFO_LTAB_SHIFT) ) {
+                ERROR("User PT base is bad. pfn=%lu nr=%lu type=%08lx %08lx",
+                      pfn, max_pfn, pfn_type[pfn],
+                      (unsigned long)pt_levels<<XEN_DOMCTL_PFINFO_LTAB_SHIFT);
+                goto out;
+            }
+
+            ctxt.ctrlreg[1] = xen_pfn_to_cr3(p2m[pfn]);
+        }
+
         domctl.cmd = XEN_DOMCTL_setvcpucontext;
         domctl.domain = (domid_t)dom;
         domctl.u.vcpucontext.vcpu = i;

diff --git a/tools/libxc/xc_linux_save.c b/tools/libxc/xc_linux_save.c
index e866699f6d072c23d866ea0b7743b53b97437a45..f9800c4d5f61e87dd98af4dbf72362fb585ccca6 100644 (file)
--- a/tools/libxc/xc_linux_save.c
+++ b/tools/libxc/xc_linux_save.c
@@ -34,7 +34,7 @@ static unsigned long max_mfn;
 /* virtual starting address of the hypervisor */
 static unsigned long hvirt_start;
 
-/* #levels of page tables used by the currrent guest */
+/* #levels of page tables used by the current guest */
 static unsigned int pt_levels;
 
 /* total number of pages used by the current guest */
@@ -491,7 +491,7 @@ static int canonicalize_pagetable(unsigned long type, unsigned long pfn,
     ** reserved hypervisor mappings. This depends on the current
     ** page table type as well as the number of paging levels.
     */
-    xen_start = xen_end = pte_last = PAGE_SIZE / ((pt_levels == 2)? 4 : 8);
+    xen_start = xen_end = pte_last = PAGE_SIZE / ((pt_levels == 2) ? 4 : 8);
 
     if (pt_levels == 2 && type == XEN_DOMCTL_PFINFO_L2TAB)
         xen_start = (hvirt_start >> L2_PAGETABLE_SHIFT);
@@ -1279,6 +1279,18 @@ int xc_linux_save(int xc_handle, int io_fd, uint32_t dom, uint32_t max_iters,
         ctxt.ctrlreg[3] = 
             xen_pfn_to_cr3(mfn_to_pfn(xen_cr3_to_pfn(ctxt.ctrlreg[3])));
 
+        /* Guest pagetable (x86/64) stored in otherwise-unused CR1. */
+        if ( (pt_levels == 4) && ctxt.ctrlreg[1] )
+        {
+            if ( !MFN_IS_IN_PSEUDOPHYS_MAP(xen_cr3_to_pfn(ctxt.ctrlreg[1])) ) {
+                ERROR("PT base is not in range of pseudophys map");
+                goto out;
+            }
+            /* Least-significant bit means 'valid PFN'. */
+            ctxt.ctrlreg[1] = 1 |
+                xen_pfn_to_cr3(mfn_to_pfn(xen_cr3_to_pfn(ctxt.ctrlreg[1])));
+        }
+
         if (!write_exact(io_fd, &ctxt, sizeof(ctxt))) {
             ERROR("Error when writing to state file (1) (errno %d)", errno);
             goto out;

diff --git a/xen/arch/x86/domain.c b/xen/arch/x86/domain.c
index 379adca5822ae3834867d1b0c02474b9ec7b2f37..57fc717915382be61c51476bbce6ef6811d1dba5 100644 (file)
--- a/xen/arch/x86/domain.c
+++ b/xen/arch/x86/domain.c
@@ -641,6 +641,31 @@ int arch_set_info_guest(
             }
 
             v->arch.guest_table = pagetable_from_pfn(cr3_pfn);
+
+#ifdef __x86_64__
+            if ( c.nat->ctrlreg[1] )
+            {
+                cr3_pfn = gmfn_to_mfn(d, xen_cr3_to_pfn(c.nat->ctrlreg[1]));
+
+                if ( !mfn_valid(cr3_pfn) ||
+                     (paging_mode_refcounts(d)
+                      ? !get_page(mfn_to_page(cr3_pfn), d)
+                      : !get_page_and_type(mfn_to_page(cr3_pfn), d,
+                                           PGT_base_page_table)) )
+                {
+                    cr3_pfn = pagetable_get_pfn(v->arch.guest_table);
+                    v->arch.guest_table = pagetable_null();
+                    if ( paging_mode_refcounts(d) )
+                        put_page(mfn_to_page(cr3_pfn));
+                    else
+                        put_page_and_type(mfn_to_page(cr3_pfn));
+                    destroy_gdt(v);
+                    return -EINVAL;
+                }
+
+                v->arch.guest_table_user = pagetable_from_pfn(cr3_pfn);
+            }
+#endif
         }
 #ifdef CONFIG_COMPAT
         else

diff --git a/xen/arch/x86/domctl.c b/xen/arch/x86/domctl.c
index c6a81d3a263bd316a66fc3939db325bc6a9fa2a1..d8ca0b0621614833ec17a33e56dc9f774da7863e 100644 (file)
--- a/xen/arch/x86/domctl.c
+++ b/xen/arch/x86/domctl.c
@@ -470,8 +470,15 @@ void arch_get_info_guest(struct vcpu *v, vcpu_guest_context_u c)
         c(user_regs.eflags |= v->arch.iopl << 12);
 
         if ( !IS_COMPAT(v->domain) )
+        {
             c.nat->ctrlreg[3] = xen_pfn_to_cr3(
                 pagetable_get_pfn(v->arch.guest_table));
+#ifdef __x86_64__
+            if ( !pagetable_is_null(v->arch.guest_table_user) )
+                c.nat->ctrlreg[1] = xen_pfn_to_cr3(
+                    pagetable_get_pfn(v->arch.guest_table_user));
+#endif
+        }
 #ifdef CONFIG_COMPAT
         else
         {

diff --git a/xen/include/public/arch-x86/xen.h b/xen/include/public/arch-x86/xen.h
index 22fcd69a65855747e1140f3cd7cb4e5b4fb963c8..f927130058735aa958765a51ea2f091d8522c903 100644 (file)
--- a/xen/include/public/arch-x86/xen.h
+++ b/xen/include/public/arch-x86/xen.h
@@ -132,6 +132,7 @@ struct vcpu_guest_context {
     unsigned long ldt_base, ldt_ents;       /* LDT (linear address, # ents) */
     unsigned long gdt_frames[16], gdt_ents; /* GDT (machine frames, # ents) */
     unsigned long kernel_ss, kernel_sp;     /* Virtual TSS (only SS1/SP1)   */
+    /* NB. User pagetable on x86/64 is placed in ctrlreg[1]. */
     unsigned long ctrlreg[8];               /* CR0-CR7 (control registers)  */
     unsigned long debugreg[8];              /* DB0-DB7 (debug registers)    */
 #ifdef __i386__