From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from eggs.gnu.org ([2001:4830:134:3::10]:38740) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1XSpeN-0006gN-O6 for qemu-devel@nongnu.org; Sat, 13 Sep 2014 11:54:20 -0400 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1XSpeD-0001Ww-Vi for qemu-devel@nongnu.org; Sat, 13 Sep 2014 11:54:11 -0400 Received: from mail-pa0-x234.google.com ([2607:f8b0:400e:c03::234]:53349) by eggs.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1XSpeD-0001Wa-PA for qemu-devel@nongnu.org; Sat, 13 Sep 2014 11:54:01 -0400 Received: by mail-pa0-f52.google.com with SMTP id kq14so3427956pab.11 for ; Sat, 13 Sep 2014 08:54:00 -0700 (PDT) Date: Sat, 13 Sep 2014 23:53:58 +0800 From: Jun Li Message-ID: <20140913155331.GA22493@localhost.localdomain> References: <1409568798-2292-1-git-send-email-junmuzi@gmail.com> <20140905153343.GH27649@stefanha-thinkpad.redhat.com> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <20140905153343.GH27649@stefanha-thinkpad.redhat.com> Subject: Re: [Qemu-devel] [PATCH v2] qcow2: add update refcount table realization for update_refcount List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: Stefan Hajnoczi Cc: kwolf@redhat.com, juli@redhat.com, famz@redhat.com, qemu-devel@nongnu.org, stefanha@redhat.com On Fri, 09/05 16:33, Stefan Hajnoczi wrote: > On Mon, Sep 01, 2014 at 06:52:48PM +0800, Jun Li wrote: > > How does this patch handle self-describing refcount blocks? I think > they will keep the refcount block alive forever because your code will > not decide to free them. > Sorry, I have ignored self-describing refcount blocks. :) > This patch should also discard the refcount block if we decide to free > it (in the same way that we discard at cluster_offset). > > > diff --git a/block/qcow2-refcount.c b/block/qcow2-refcount.c > > index 43665b8..63f36e6 100644 > > --- a/block/qcow2-refcount.c > > +++ b/block/qcow2-refcount.c > > @@ -586,6 +586,37 @@ static int QEMU_WARN_UNUSED_RESULT update_refcount(BlockDriverState *bs, > > if (refcount == 0 && s->discard_passthrough[type]) { > > update_refcount_discard(bs, cluster_offset, s->cluster_size); > > } > > + > > + /* When refcount block is NULL, update refcount table */ > > + if (block_index == 0) { > > What is the purpose of block_index == 0? Here is want to reduce the probability of running the following code. Only when block_index == 0, we will run the following code to free refcount block. > > > + int k = block_index; > > + int refcount_block_entries = s->cluster_size / sizeof(uint16_t); > > + for (k = 0; k < refcount_block_entries; k++) { > > + if (refcount_block[k] != cpu_to_be16(0)) { > > + break; > > + } > > + } > > + > > + if (k == refcount_block_entries) { > > + qemu_vfree(refcount_block); > > You can't do this, the buffer belongs to the refcount block cache. > Please look at the cache get/put as well as qcow2_cache_create/destroy. ok, thx. Should add qcow2_cache_put(bs, s->refcount_block_cache, (void**) refcount_block); > > > + /* update refcount table */ > > + unsigned int refcount_table_index; > > + uint64_t data64 = cpu_to_be64(0); > > + refcount_table_index = cluster_index >> (s->cluster_bits - > > + REFCOUNT_SHIFT); > > + ret = bdrv_pwrite_sync(bs->file, > > + s->refcount_table_offset + > > + refcount_table_index * > > + sizeof(uint64_t), > > + &data64, sizeof(data64)); > > + if (ret < 0) { > > + goto fail; > > + } > > Plase use write_reftable_entry(). ok, got it. I will submit a new version when I submit v3 of qcow2 shrinking. Best Regards, Jun Li