summaryrefslogtreecommitdiff
diff options
context:
space:
mode:
authorDavid Gibson <david@gibson.dropbear.id.au>2012-12-04 11:38:39 +1100
committerDoug Goldstein <cardoe@cardoe.com>2013-01-18 21:02:09 -0600
commit8a4cb5fcb0f746bb8d7b8ac7c50327a5c70abb98 (patch)
tree1c95fff4b43fa914400a56363be75dbdc20da32b
parenttarget-xtensa: fix ITLB/DTLB page protection flags (diff)
downloadqemu-kvm-8a4cb5fcb0f746bb8d7b8ac7c50327a5c70abb98.tar.gz
qemu-kvm-8a4cb5fcb0f746bb8d7b8ac7c50327a5c70abb98.tar.bz2
qemu-kvm-8a4cb5fcb0f746bb8d7b8ac7c50327a5c70abb98.zip
migration: Fix madvise breakage if host and guest have different page sizes
madvise(DONTNEED) will throw away the contents of the whole page at the given address, even if the given length is less than the page size. One can argue about whether that's the correct behaviour, but that's what it's done for a long time in Linux at least. That means that the madvise() in ram_load(), on a setup where TARGET_PAGE_SIZE is smaller than the host page size, can throw away data in guest pages adjacent to the one it's actually processing right now, leading to guest memory corruption on an incoming migration. This patch therefore, disables the madvise() if the host page size is larger than TARGET_PAGE_SIZE. This means we don't get the benefits of that madvise() in this case, but a more complete fix is more difficult to accomplish. This at least fixes the guest memory corruption. Signed-off-by: David Gibson <david@gibson.dropbear.id.au> Reported-by: Alexey Kardashevskiy <aik@ozlabs.ru> Signed-off-by: Anthony Liguori <aliguori@us.ibm.com> (cherry picked from commit 45e6cee42b98d10e2e14885ab656541a9ffd5187)
-rw-r--r--arch_init.c3
1 files changed, 2 insertions, 1 deletions
diff --git a/arch_init.c b/arch_init.c
index f849f9b87..5c912942b 100644
--- a/arch_init.c
+++ b/arch_init.c
@@ -764,7 +764,8 @@ static int ram_load(QEMUFile *f, void *opaque, int version_id)
memset(host, ch, TARGET_PAGE_SIZE);
#ifndef _WIN32
if (ch == 0 &&
- (!kvm_enabled() || kvm_has_sync_mmu())) {
+ (!kvm_enabled() || kvm_has_sync_mmu()) &&
+ getpagesize() <= TARGET_PAGE_SIZE) {
qemu_madvise(host, TARGET_PAGE_SIZE, QEMU_MADV_DONTNEED);
}
#endif