Fix invalidate if memory requested was not bucket aligned

Submitted by Frediano Ziglio on Aug. 7, 2012, 12:23 p.m.

Details

Message ID 7CE799CC0E4DE04B88D5FDF226E18AC2CDFFB08D19@LONPMAILBOX01.citrite.net
State New
Headers show

Commit Message

Frediano Ziglio Aug. 7, 2012, 12:23 p.m.
When memory is mapped in qemu_map_cache with lock != 0 a reverse mapping
is created pointing to the virtual address of location requested.
The cached mapped entry is saved in last_address_vaddr with the memory
location of the base virtual address (without bucket offset).
However when this entry is invalidated the virtual address saved in the
reverse mapping is used. This cause that the mapping is freed but the
last_address_vaddr is not reset.

Signed-off-by: Frediano Ziglio <frediano.ziglio@citrix.com>
---
 xen-mapcache.c |    9 +++++----
 1 files changed, 5 insertions(+), 4 deletions(-)

Patch hide | download patch | download mbox

diff --git a/xen-mapcache.c b/xen-mapcache.c
index 59ba085..9cd6db3 100644
--- a/xen-mapcache.c
+++ b/xen-mapcache.c
@@ -320,10 +320,6 @@  void xen_invalidate_map_cache_entry(uint8_t *buffer)
     target_phys_addr_t size;
     int found = 0;
 
-    if (mapcache->last_address_vaddr == buffer) {
-        mapcache->last_address_index = -1;
-    }
-
     QTAILQ_FOREACH(reventry, &mapcache->locked_entries, next) {
         if (reventry->vaddr_req == buffer) {
             paddr_index = reventry->paddr_index;
@@ -342,6 +338,11 @@  void xen_invalidate_map_cache_entry(uint8_t *buffer)
     QTAILQ_REMOVE(&mapcache->locked_entries, reventry, next);
     g_free(reventry);
 
+    if (mapcache->last_address_index == paddr_index) {
+        mapcache->last_address_index = -1;
+        mapcache->last_address_vaddr = NULL;
+    }
+
     entry = &mapcache->entry[paddr_index % mapcache->nr_buckets];
     while (entry && (entry->paddr_index != paddr_index || entry->size != size)) {
         pentry = entry;