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.


HVM synchronous operation facility
[palacios.git] / linux_usr / v3_guest_mem.h
index 907502c..cacddd1 100644 (file)
@@ -4,8 +4,16 @@
 #include <stdint.h>
 #include "v3_ctrl.h"
 
+
+#include "memtrack.h"
+
+/**************************/
+/* Access to guest memory */
+/**************************/
+
 struct v3_guest_mem_block {
-  void     *gpa;      // guest physical address is
+  void     *gpa;      // guest physical address this region starts at
+  void     *cumgpa;   // cumulative GPA in the VM including this block
   void     *hpa;      // mapped to this host physical address
   void     *uva;      // which is mapped here in this process
   uint64_t numpages;  // this many 4K pages
@@ -27,6 +35,79 @@ struct v3_guest_mem_map * v3_guest_mem_get_map(char *vmdev);
 int v3_map_guest_mem(struct v3_guest_mem_map *map);
 // This function unmaps it - it assumes myaddrs are valid
 int v3_unmap_guest_mem(struct v3_guest_mem_map *map);
+// What is the first GPA of this guest that contains memory ?
+void *v3_gpa_start(struct v3_guest_mem_map *map);
+// What is the last GPA of this guest that contains memory ?
+void *v3_gpa_end(struct v3_guest_mem_map *map);
+// map from a gpa to a uva, optionally giving the number of 
+// subsequent bytes for which the mapping is continguous
+static inline void *v3_gpa_to_uva(struct v3_guest_mem_map *map, void *gpa, uint64_t *num_bytes)
+{
+    uint64_t left, right, middle;
+
+    if (!map->fd) { 
+       // not mapped
+       if (num_bytes) { *num_bytes=0;}
+       return 0;
+    }
+    
+    left = 0; right=map->numblocks-1;
+    
+    while (right>=left) { 
+       middle = (right+left)/2;
+       if (gpa > map->block[middle].cumgpa) { 
+           left=middle+1;
+       } else if (gpa < map->block[middle].gpa) { 
+           right=middle-1;
+       } else {
+           break;
+       }
+    }
+
+    if (right>=left) { 
+       if (num_bytes) { 
+           *num_bytes = map->block[middle].cumgpa - gpa + 1;
+       }
+       return map->block[middle].uva + (gpa - map->block[middle].gpa);
+    } else {
+       if (num_bytes) { 
+           *num_bytes=0;
+       } 
+       return 0;
+    }
+}
+
+// efficiently map function over the specified guest memory
+int v3_guest_mem_apply(void (*func)(void *data, uint64_t num_bytes, void *priv),
+                      struct v3_guest_mem_map *map, void *gpa, uint64_t num_bytes, void *priv);
+
+// read data out of guest
+int v3_guest_mem_read(struct v3_guest_mem_map *map, void *gpa, uint64_t num_bytes, char *data);
+
+// write data into  guest
+int v3_guest_mem_write(struct v3_guest_mem_map *map, void *gpa, uint64_t num_bytes, char *data);
+
+// hash the guest's data
+int v3_guest_mem_hash(struct v3_guest_mem_map *map, void *gpa, uint64_t num_bytes, uint64_t *hash);
+
+
+/********************************/
+/* Guest memory access tracking */
+/********************************/
+
+#include "../linux_module/memtrack.h"
+
+int v3_guest_mem_track_start(char *vmdev, 
+                            v3_mem_track_access_t access, 
+                            v3_mem_track_reset_t reset, 
+                            uint64_t period);
+
+int v3_guest_mem_track_stop(char *vmdev);
+
+v3_mem_track_snapshot *v3_guest_mem_track_snapshot(char *vmdev);
+
+void v3_guest_mem_track_free_snapshot(v3_mem_track_snapshot *snap);
+
 
 
 #endif