Re: [Qemu-devel] [PATCH V18 5/6] add-cow file format core code.

[Stefan Hajnoczi <stefanha@redhat.com>]

On Thu, May 09, 2013 at 02:24:20PM +0800, Dong Xu Wang wrote:
> On Thu, Apr 18, 2013 at 6:03 PM, Stefan Hajnoczi <stefanha@gmail.com> wrote:
> > On Wed, Apr 10, 2013 at 04:11:52PM +0800, Dong Xu Wang wrote:
> >> +static coroutine_fn int add_cow_co_flush(BlockDriverState *bs)
> >> +{
> >> +    BDRVAddCowState *s = bs->opaque;
> >> +    int ret;
> >> +
> >> +    qemu_co_mutex_lock(&s->lock);
> >> +    if (s->bitmap_cache) {
> >> +        ret = block_cache_flush(bs, s->bitmap_cache);
> >> +        if (ret < 0) {
> >> +            return ret;
> >> +        }
> >> +    }
> >> +    ret = bdrv_flush(s->image_hd);
> >
> > This is the wrong way around.  We must flush image_hd first so that
> > valid data is on disk.  Then we can flush bitmap_cache to mark the
> > clusters allocated.
> >
> > Beyond explicit flushes you also need to make sure that image_hd is
> > flushed *before* bitmap_cache tables are written out (e.g. cache
> > eviction when the cache becomes full).  It seems this code is missing.
> >
> Hi  Stefan, how can I make sure image_hd "flush" operation completed before
> I write bitmap_cache tables? Is there any functions I can use?

You could extend the block cache like this:

block_cache_add_dependency_func(BlockCache *c,
                                int (*dependency_fn)(BlockCache *c, void *opaque),
				void *opaque);

Then add-cow would register a custom dependency function that flushes
image_hd.

A hackier way of achieving the same thing is to instantiate an
image_hd_cache and set a dependency from the bitmap_cache on the
image_hd_cache.  The image_hd_cache actually has no cache entries, but
it uses image_hd and will flush it.

Stefan