Palacios Public Git Repository

To checkout Palacios execute

  git clone http://v3vee.org/palacios/palacios.web/palacios.git
This will give you the master branch. You probably want the devel branch or one of the release branches. To switch to the devel branch, simply execute
  cd palacios
  git checkout --track -b devel origin/devel
The other branches are similar.


*** empty log message ***
[palacios.git] / palacios / src / geekos / vmm_paging.c
index 2a8575f..4c1469d 100644 (file)
@@ -6,9 +6,215 @@
 
 extern struct vmm_os_hooks * os_hooks;
 
+void delete_page_tables_pde32(vmm_pde_t * pde) {
+  int i, j;
+
+  if (pde==NULL) { 
+    return ;
+  }
+
+  for (i = 0; (i < MAX_PAGE_DIR_ENTRIES); i++) {
+    if (pde[i].present) {
+      vmm_pte_t * pte = (vmm_pte_t *)(pde[i].pt_base_addr << PAGE_POWER);
+      
+      for (j = 0; (j < MAX_PAGE_TABLE_ENTRIES); j++) {
+       if ((pte[j].present) && (pte[j].vmm_info & GUEST_PAGE)){
+         os_hooks->free_page((void *)(pte[j].page_base_addr  << PAGE_POWER));
+       }
+      }
+      
+      os_hooks->free_page(pte);
+    }
+  }
+
+  os_hooks->free_page(pde);
+}
+
+
+int init_shadow_paging_state(shadow_paging_state_t *state)
+{
+  state->guest_page_directory_type=state->shadow_page_directory_type=PDE32;
+  
+  state->guest_page_directory=state->shadow_page_directory=NULL;
+
+  init_shadow_map(&(state->shadow_map));
+  return 0;
+}
+  
+
+int wholesale_update_shadow_paging_state(shadow_paging_state_t *state)
+{
+  unsigned i, j;
+  vmm_pde_t *cur_guest_pde, *cur_shadow_pde;
+  vmm_pte_t *cur_guest_pte, *cur_shadow_pte;
+
+  // For now, we'll only work with PDE32
+  if (state->guest_page_directory_type!=PDE32) { 
+    return -1;
+  }
+  
+  cur_shadow_pde=(vmm_pde_t*)(state->shadow_page_directory);
+  
+  cur_guest_pde = (vmm_pde_t*)(os_hooks->physical_to_virtual(state->guest_page_directory));
+
+  // Delete the current page table
+  delete_page_tables_pde32(cur_shadow_pde);
+
+  cur_shadow_pde = os_hooks->allocate_pages(1);
+
+  state->shadow_page_directory = cur_shadow_pde;
+  state->shadow_page_directory_type=PDE32;
+
+  
+  for (i=0;i<MAX_PAGE_DIR_ENTRIES;i++) { 
+    cur_shadow_pde[i] = cur_guest_pde[i];
+    // The shadow can be identical to the guest if it's not present
+    if (!cur_shadow_pde[i].present) { 
+      continue;
+    }
+    if (cur_shadow_pde[i].large_pages) { 
+      // large page - just map it through shadow map to generate its physical location
+      addr_t guest_addr = PAGE_ADDR(cur_shadow_pde[i].pt_base_addr);
+      addr_t host_addr;
+      shadow_map_entry_t *ent;
+
+      ent = get_shadow_map_region_by_addr(&(state->shadow_map),guest_addr);
+      
+      if (!ent) { 
+       // FIXME Panic here - guest is trying to map to physical memory
+       // it does not own in any way!
+       return -1;
+      }
+
+      // FIXME Bounds check here to see if it's trying to trick us
+      
+      switch (ent->host_type) { 
+      case HOST_REGION_PHYSICAL_MEMORY:
+       // points into currently allocated physical memory, so we just
+       // set up the shadow to point to the mapped location
+       if (map_guest_physical_to_host_physical(ent,guest_addr,&host_addr)) { 
+         // Panic here
+         return -1;
+       }
+       cur_shadow_pde[i].pt_base_addr = PAGE_ALIGNED_ADDR(host_addr);
+       // FIXME set vmm_info bits here
+       break;
+      case HOST_REGION_UNALLOCATED:
+       // points to physical memory that is *allowed* but that we
+       // have not yet allocated.  We mark as not present and set a
+       // bit to remind us to allocate it later
+       cur_shadow_pde[i].present=0;
+       // FIXME Set vminfo bits here so that we know that we will be
+       // allocating it later
+       break;
+      case HOST_REGION_NOTHING:
+       // points to physical memory that is NOT ALLOWED.   
+       // We will mark it as not present and set a bit to remind
+       // us that it's bad later and insert a GPF then
+       cur_shadow_pde[i].present=0;
+       break;
+      case HOST_REGION_MEMORY_MAPPED_DEVICE:
+      case HOST_REGION_REMOTE:
+      case HOST_REGION_SWAPPED:
+      default:
+       // Panic.  Currently unhandled
+       return -1;
+       break;
+      }
+    } else {
+      addr_t host_addr;
+      addr_t guest_addr;
+
+      // small page - set PDE and follow down to the child table
+      cur_shadow_pde[i] = cur_guest_pde[i];
+      
+      // Allocate a new second level page table for the shadow
+      cur_shadow_pte = os_hooks->allocate_pages(1);
+
+      // make our first level page table in teh shadow point to it
+      cur_shadow_pde[i].pt_base_addr = PAGE_ALIGNED_ADDR(cur_shadow_pte);
+
+      shadow_map_entry_t *ent;
+      
+      guest_addr=PAGE_ADDR(cur_guest_pde[i].pt_base_addr);
+
+      ent = get_shadow_map_region_by_addr(&(state->shadow_map),guest_addr);
+      
+      if (!ent) { 
+       // FIXME Panic here - guest is trying to map to physical memory
+       // it does not own in any way!
+       return -1;
+      }
+
+      // Address of the relevant second level page table in the guest
+      if (map_guest_physical_to_host_physical(ent,guest_addr,&host_addr)) { 
+       // Panic here
+       return -1;
+      }
+      // host_addr now contains the host physical address for the guest's 2nd level page table
+
+      // Now we transform it to relevant virtual address
+      cur_guest_pte = os_hooks->physical_to_virtual((void*)host_addr);
 
+      // Now we walk through the second level guest page table
+      // and clone it into the shadow
+      for (j=0;j<MAX_PAGE_TABLE_ENTRIES;j++) { 
+       cur_shadow_pte[j] = cur_guest_pte[j];
+
+       addr_t guest_addr = PAGE_ADDR(cur_shadow_pte[j].page_base_addr);
+       
+       shadow_map_entry_t *ent;
+
+       ent = get_shadow_map_region_by_addr(&(state->shadow_map),guest_addr);
+      
+       if (!ent) { 
+         // FIXME Panic here - guest is trying to map to physical memory
+         // it does not own in any way!
+         return -1;
+       }
+
+       switch (ent->host_type) { 
+       case HOST_REGION_PHYSICAL_MEMORY:
+         // points into currently allocated physical memory, so we just
+         // set up the shadow to point to the mapped location
+         if (map_guest_physical_to_host_physical(ent,guest_addr,&host_addr)) { 
+           // Panic here
+           return -1;
+         }
+         cur_shadow_pte[j].page_base_addr = PAGE_ALIGNED_ADDR(host_addr);
+         // FIXME set vmm_info bits here
+         break;
+       case HOST_REGION_UNALLOCATED:
+         // points to physical memory that is *allowed* but that we
+         // have not yet allocated.  We mark as not present and set a
+         // bit to remind us to allocate it later
+         cur_shadow_pte[j].present=0;
+         // FIXME Set vminfo bits here so that we know that we will be
+         // allocating it later
+         break;
+       case HOST_REGION_NOTHING:
+         // points to physical memory that is NOT ALLOWED.   
+         // We will mark it as not present and set a bit to remind
+         // us that it's bad later and insert a GPF then
+         cur_shadow_pte[j].present=0;
+         break;
+       case HOST_REGION_MEMORY_MAPPED_DEVICE:
+       case HOST_REGION_REMOTE:
+       case HOST_REGION_SWAPPED:
+       default:
+         // Panic.  Currently unhandled
+         return -1;
+       break;
+       }
+      }
+    }
+  }
+  return 0;
+}
+      
 
 
+#if 0
 /* We generate a page table to correspond to a given memory layout
  * pulling pages from the mem_list when necessary
  * If there are any gaps in the layout, we add them as unmapped pages
@@ -124,27 +330,8 @@ vmm_pde_t * generate_guest_page_tables(vmm_mem_layout_t * layout, vmm_mem_list_t
   return pde;
 }
 
+#endif
 
-void free_guest_page_tables(vmm_pde_t * pde) {
-  int i, j;
-
-
-  for (i = 0; (i < MAX_PAGE_DIR_ENTRIES); i++) {
-    if (pde[i].present) {
-      vmm_pte_t * pte = (vmm_pte_t *)(pde[i].pt_base_addr << PAGE_POWER);
-      
-      for (j = 0; (j < MAX_PAGE_TABLE_ENTRIES); j++) {
-       if ((pte[j].present) && (pte[j].vmm_info & GUEST_PAGE)){
-         os_hooks->free_page((void *)(pte[j].page_base_addr  << PAGE_POWER));
-       }
-      }
-      
-      os_hooks->free_page(pte);
-    }
-  }
-
-  os_hooks->free_page(pde);
-}
 
 
 
@@ -217,7 +404,7 @@ void PrintDebugPageTables(vmm_pde_t * pde)
     
     
 
-
+#if 0
 
 pml4e64_t * generate_guest_page_tables_64(vmm_mem_layout_t * layout, vmm_mem_list_t * list) {
   pml4e64_t * pml = os_hooks->allocate_pages(1);
@@ -409,3 +596,5 @@ pml4e64_t * generate_guest_page_tables_64(vmm_mem_layout_t * layout, vmm_mem_lis
   }
   return pml;
 }
+
+#endif