[Libguestfs] [nbdkit PATCH 3/6] retry: Check next_ops->can_FOO on retry

[Eric Blake]

On 9/30/19 10:17 PM, Eric Blake wrote:
> After a retry, if the second connection has fewer permissions than the
> first, but we blindly call next_ops->FOO, we end up triggering an
> assertion failure in backend.c.  This is particularly noticeable when
> the force_readonly flag is in effect, as that makes it much easier for
> there to be fewer permissions than before.
> 
> Add testsuite coverage of pwrite to demonstrate.
> 
> Signed-off-by: Eric Blake <eblake at redhat.com>
> ---
>   filters/retry/retry.c | 58 ++++++++++++++++++++++++++++++++++++++-----
>   1 file changed, 52 insertions(+), 6 deletions(-)

> @@ -307,7 +348,12 @@ retry_cache (struct nbdkit_next_ops *next_ops, void *nxdata,
>     int r;
> 
>    again:
> -  r = next_ops->cache (nxdata, count, offset, flags, err);
> +  if (next_ops->can_cache (nxdata) != 1) {

This actually needs to be <= NBDKIT_CACHE_NONE (otherwise, it penalizes 
a change between EMULATE and NATIVE).

I'm also suspecting that I'll have to check can_fua and can_fast_zero 
for proper usage, but that will be a followup patch.

-- 
Eric Blake, Principal Software Engineer
Red Hat, Inc.           +1-919-301-3226
Virtualization:  qemu.org | libvirt.org