[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
[Qemu-devel] [PATCH 2/5] qcow2: Limit COW to where it's needed
From: |
Kevin Wolf |
Subject: |
[Qemu-devel] [PATCH 2/5] qcow2: Limit COW to where it's needed |
Date: |
Mon, 7 May 2012 19:55:51 +0200 |
This fixes a regression introduced in commit 250196f1. The bug leads to
data corruption, found during an Autotest run with a Fedora 8 guest.
Consider a write request whose first part is covered by an already
allocated cluster, but additional clusters need to be newly allocated.
When counting the number of clusters to allocate, the qcow2 code would
decide to do COW for all remaining clusters of the write request, even
if some of them are already allocated.
If during this COW operation another write request is issued that touches
the same cluster, it will still refer to the old cluster. When the COW
completes, the first request will update the L2 table and the second
write request will be lost. Note that the requests need not overlap, it's
enough for them to touch the same cluster.
This patch ensures that only clusters that really require COW are
considered for allocation. In this case any other request writing to the
same cluster will be an allocating write and gets serialised.
Reported-by: Marcelo Tosatti <address@hidden>
Tested-by: Marcelo Tosatti <address@hidden>
Signed-off-by: Kevin Wolf <address@hidden>
---
block/qcow2-cluster.c | 14 +++++++++-----
1 files changed, 9 insertions(+), 5 deletions(-)
diff --git a/block/qcow2-cluster.c b/block/qcow2-cluster.c
index 353889d..10c22fe 100644
--- a/block/qcow2-cluster.c
+++ b/block/qcow2-cluster.c
@@ -883,15 +883,19 @@ again:
assert(keep_clusters <= nb_clusters);
nb_clusters -= keep_clusters;
} else {
+ keep_clusters = 0;
+ cluster_offset = 0;
+ }
+
+ if (nb_clusters > 0) {
/* For the moment, overwrite compressed clusters one by one */
- if (cluster_offset & QCOW_OFLAG_COMPRESSED) {
+ uint64_t entry = be64_to_cpu(l2_table[l2_index + keep_clusters]);
+ if (entry & QCOW_OFLAG_COMPRESSED) {
nb_clusters = 1;
} else {
- nb_clusters = count_cow_clusters(s, nb_clusters, l2_table,
l2_index);
+ nb_clusters = count_cow_clusters(s, nb_clusters, l2_table,
+ l2_index + keep_clusters);
}
-
- keep_clusters = 0;
- cluster_offset = 0;
}
cluster_offset &= L2E_OFFSET_MASK;
--
1.7.6.5
- [Qemu-devel] [PULL 0/5] Block patches for 1.1, Kevin Wolf, 2012/05/07
- [Qemu-devel] [PATCH 1/5] sheepdog: switch to writethrough mode if cluster doesn't support flush, Kevin Wolf, 2012/05/07
- [Qemu-devel] [PATCH 2/5] qcow2: Limit COW to where it's needed,
Kevin Wolf <=
- [Qemu-devel] [PATCH 3/5] block: make bdrv_create adopt coroutine, Kevin Wolf, 2012/05/07
- [Qemu-devel] [PATCH 4/5] qcow2: lock on prealloc, Kevin Wolf, 2012/05/07
- [Qemu-devel] [PATCH 5/5] fdc: simplify media change handling, Kevin Wolf, 2012/05/07
- Re: [Qemu-devel] [PULL 0/5] Block patches for 1.1, Anthony Liguori, 2012/05/08
- [Qemu-devel] [PULL 0/5] Block patches for 1.1, Kevin Wolf, 2012/05/30
- [Qemu-devel] [PATCH 2/5] sheepdog: fix return value of do_load_save_vm_state, Kevin Wolf, 2012/05/30
- [Qemu-devel] [PATCH 3/5] block: prevent snapshot mode $TMPDIR symlink attack, Kevin Wolf, 2012/05/30
- [Qemu-devel] [PATCH 5/5] ahci: SATA FIS is 20 bytes, not 0x20, Kevin Wolf, 2012/05/30
- [Qemu-devel] [PATCH 1/5] virtio: Fix compiler warning for non Linux hosts, Kevin Wolf, 2012/05/30
- [Qemu-devel] [PATCH 4/5] virtio-blk: Fix geometry sector calculation, Kevin Wolf, 2012/05/30