Small fix for shadow_remove_all_access_in_page
PATCH 1/1 Small fix for shadow_remove_all_access_in_page
Here's a fairly simple patch to correct a bug I hit when experimenting with VM forking earlier. Let me know if there's anything wrong with the changeset; this is the first time I've tried exporting a changeset from Mercurial.
--Michael Vrable
# HG changeset patch
# User Michael Vrable <[email protected]> # Node ID df9609e384a4a819a69d01556b8231b842b04050 # Parent d8637529dafff3c610e4bb31349d4ae00a23a18e Fix a bug in shadow_remove_all_access_in_page.
Fixes a fairly obvious bug, in which an if statement was reversed. The bug has been present for a while, so apparently this code isn't well tested.
xen/arch/x86/shadow32.c +1 -1
@@ -2226,7 +2226,7 @@for (i =0 ; i <L1_PAGETABLE_ENTRIES ; i++) { -if (unlikely (!l1e_has_changed (pl1e[i], match, flags) ==0 ) ) +if (unlikely (!l1e_has_changed (pl1e[i], match, flags)) ) {l1_pgentry_t ol2e = pl1e[i]; pl1e[i] =l1e_empty ();
xen/arch/x86/shadow_public.c +1 -1
@@ -1634,7 +1634,7 @@for (i =0 ; i <L1_PAGETABLE_ENTRIES ; i++) { -if (unlikely (!l1e_has_changed (pl1e[i], match, flags) ==0 ) ) +if (unlikely (!l1e_has_changed (pl1e[i], match, flags)) ) {l1_pgentry_t ol2e = pl1e[i]; pl1e[i] =l1e_empty ();
Quoted 4 lines · Michael Vrable
Here's a fairly simple patch to correct a bug I hit when experimenting with VM forking earlier. Let me know if there's anything wrong with the changeset; this is the first time I've tried exporting a changeset from Mercurial.
Thanks. An embarrassing bug for whoever it was committed that code!
-- Keir