From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from eggs.gnu.org ([2001:4830:134:3::10]:38541) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1Xe6IP-0007So-Dt for qemu-devel@nongnu.org; Tue, 14 Oct 2014 13:54:14 -0400 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1Xe6IG-000895-15 for qemu-devel@nongnu.org; Tue, 14 Oct 2014 13:54:05 -0400 Received: from mail-pa0-x231.google.com ([2607:f8b0:400e:c03::231]:59312) by eggs.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1Xe6IF-00088k-QW for qemu-devel@nongnu.org; Tue, 14 Oct 2014 13:53:55 -0400 Received: by mail-pa0-f49.google.com with SMTP id hz1so8188080pad.22 for ; Tue, 14 Oct 2014 10:53:54 -0700 (PDT) From: Jun Li Date: Wed, 15 Oct 2014 01:53:34 +0800 Message-Id: <1413309216-18294-1-git-send-email-junmuzi@gmail.com> Subject: [Qemu-devel] [PATCH v4 0/2] qcow2: Patch for shrinking qcow2 disk image List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: kwolf@redhat.com, stefanha@redhat.com, famz@redhat.com, juli@redhat.com Cc: Jun Li , qemu-devel@nongnu.org v4: Add deal with COW clusters in l2 table. When using COW, some of (l2_entry >> s->cluster_bits) will larger than s->refcount_table_size, so need to discard this l2_entry. v3: Fixed host cluster leak. A simple testing: Step 1, # /opt/qemu-git-arm/bin/qemu-img info /home/lijun/Work/tmp/shrink.qcow2 image: /home/lijun/Work/tmp/shrink.qcow2 file format: qcow2 virtual size: 2.0G (2147483648 bytes) disk size: 2.0G cluster_size: 65536 Format specific information: compat: 1.1 lazy refcounts: false corrupt: false Step 2, # /opt/qemu-git-arm/bin/qemu-img resize /home/lijun/Work/tmp/shrink.qcow2 -100M Image resized. Step 3, # /opt/qemu-git-arm/bin/qemu-img info /home/lijun/Work/tmp/shrink.qcow2 image: /home/lijun/Work/tmp/shrink.qcow2 file format: qcow2 virtual size: 1.9G (2042626048 bytes) disk size: 1.9G cluster_size: 65536 Format specific information: compat: 1.1 lazy refcounts: false corrupt: false Step 4, # /opt/qemu-git-arm/bin/qemu-img check /home/lijun/Work/tmp/shrink.qcow2 No errors were found on the image. 24576/31168 = 78.85% allocated, 0.80% fragmented, 0.00% compressed clusters Image end offset: 1631911936 As above show, in step 1, "disk size" and "virtual size" are all 2.0G. After Step 3, "disk size" and "virtual size" are all 1.9G. BTW, above is also a simple testing. I will submit a test case for qemu-iotests later. But I am not so familar with howto write qemu-iotests now. In file block/qcow2.c, just using ftruncate to fix host cluster leak. In file block/qcow2-cluster.c, just re-copy qcow2_grow_l1_table to realize qcow2_shrink_l1_and_l2_table. In file block/qcow2-refcount.c, also update the realization to handle self-describing refcount blocks in function update_refcount. Thanks. Jun Li (2): qcow2: Add qcow2_shrink_l1_and_l2_table for qcow2 shrinking qcow2: add update refcount table realization for update_refcount block/qcow2-cluster.c | 186 +++++++++++++++++++++++++++++++++++++++++++++++++ block/qcow2-refcount.c | 46 +++++++++++- block/qcow2.c | 40 +++++++++-- block/qcow2.h | 2 + 4 files changed, 269 insertions(+), 5 deletions(-) -- 1.9.3