On 11/22/2017 02:03 PM, Eric Blake wrote:
On 11/22/2017 04:19 AM, Vladimir Sementsov-Ogievskiy wrote:
Signed-off-by: Vladimir Sementsov-Ogievskiy <[email protected]>
---
nbd/server.c | 34 ++++++++++++++++++++++------------
1 file changed, 22 insertions(+), 12 deletions(-)
I think that means we need a tri-state return from nbd_opt_read(): < 0
on I/O failure, 0 if the EINVAL message has been sent, and 1 if the read
was successful; I also think it means that the handler for
NBD_OPT_EXPORT_NAME does not really fit the bill for using the same
handler. Hopefully this explanation will give you more insight into the
counter-proposal patch I'm about to post.
Hmm, I never posted the counter-proposal, because I got caught up in
handling the CVEs that I spotted in the same area, then the 2.11
release. I'll get back to this series shortly.
--
Eric Blake, Principal Software Engineer
Red Hat, Inc. +1-919-301-3266
Virtualization: qemu.org | libvirt.org
