[Qemu-devel] [PATCH for-2.11] nbd/client: Don't hard-disconnect on ESHUTDOWN from server

[Qemu-devel] [PATCH for-2.11] nbd/client: Don't hard-disconnect on ESHUTDOWN from server

[Eric Blake]

The NBD spec says that a server may fail any transmission request
with ESHUTDOWN when it is apparent that no further request from
the client can be successfully honored.  The client is supposed
to then initiate a soft shutdown (wait for all remaining in-flight
requests to be answered, then send NBD_CMD_DISC).  However, since
qemu's server never uses ESHUTDOWN errors, this code was mostly
untested since its introduction in commit b6f5d3b5.

More recently, I learned that nbdkit as the NBD server is able to
send ESHUTDOWN errors, so I finally tested this code, and noticed
that our client was special-casing ESHUTDOWN to cause a hard
shutdown (immediate disconnect, with no NBD_CMD_DISC), but only
if the server sends this error as a simple reply.  Further
investigation found that commit d2febedb introduced a regression
where structured replies behave differently than simple replies -
but that the structured reply behavior is more in line with the
spec (even if we still lack code in nbd-client.c to properly quit
sending further requests).  So this patch reverts the portion of
b6f5d3b5 that introduced an improper hard-disconnect special-case
at the lower level, and leaves the future enhancement of a nicer
soft-disconnect at the higher level for another day.

CC: qemu-trivial@nongnu.org
Signed-off-by: Eric Blake <eblake@redhat.com>
---
 nbd/client.c | 6 ------
 1 file changed, 6 deletions(-)

diff --git a/nbd/client.c b/nbd/client.c
index 4e15fc484d..eea236ca06 100644
--- a/nbd/client.c
+++ b/nbd/client.c
@@ -996,15 +996,9 @@ int nbd_receive_reply(QIOChannel *ioc, NBDReply *reply, Error **errp)
         if (ret < 0) {
             break;
         }
-
         trace_nbd_receive_simple_reply(reply->simple.error,
                                        nbd_err_lookup(reply->simple.error),
                                        reply->handle);
-        if (reply->simple.error == NBD_ESHUTDOWN) {
-            /* This works even on mingw which lacks a native ESHUTDOWN */
-            error_setg(errp, "server shutting down");
-            return -EINVAL;
-        }
         break;
     case NBD_STRUCTURED_REPLY_MAGIC:
         ret = nbd_receive_structured_reply_chunk(ioc, &reply->structured, errp);
-- 
2.13.6

Re: [Qemu-devel] [PATCH for-2.11] nbd/client: Don't hard-disconnect on ESHUTDOWN from server

[Vladimir Sementsov-Ogievskiy]

13.11.2017 22:48, Eric Blake wrote:
> The NBD spec says that a server may fail any transmission request
> with ESHUTDOWN when it is apparent that no further request from
> the client can be successfully honored.  The client is supposed
> to then initiate a soft shutdown (wait for all remaining in-flight
> requests to be answered, then send NBD_CMD_DISC).  However, since
> qemu's server never uses ESHUTDOWN errors, this code was mostly
> untested since its introduction in commit b6f5d3b5.
>
> More recently, I learned that nbdkit as the NBD server is able to
> send ESHUTDOWN errors, so I finally tested this code, and noticed
> that our client was special-casing ESHUTDOWN to cause a hard
> shutdown (immediate disconnect, with no NBD_CMD_DISC), but only
> if the server sends this error as a simple reply.  Further
> investigation found that commit d2febedb introduced a regression
> where structured replies behave differently than simple replies -
> but that the structured reply behavior is more in line with the
> spec (even if we still lack code in nbd-client.c to properly quit
> sending further requests).  So this patch reverts the portion of
> b6f5d3b5 that introduced an improper hard-disconnect special-case
> at the lower level, and leaves the future enhancement of a nicer
> soft-disconnect at the higher level for another day.
>
> CC: qemu-trivial@nongnu.org
> Signed-off-by: Eric Blake <eblake@redhat.com>

Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>

> ---
>   nbd/client.c | 6 ------
>   1 file changed, 6 deletions(-)
>
> diff --git a/nbd/client.c b/nbd/client.c
> index 4e15fc484d..eea236ca06 100644
> --- a/nbd/client.c
> +++ b/nbd/client.c
> @@ -996,15 +996,9 @@ int nbd_receive_reply(QIOChannel *ioc, NBDReply *reply, Error **errp)
>           if (ret < 0) {
>               break;
>           }
> -
>           trace_nbd_receive_simple_reply(reply->simple.error,
>                                          nbd_err_lookup(reply->simple.error),
>                                          reply->handle);
> -        if (reply->simple.error == NBD_ESHUTDOWN) {
> -            /* This works even on mingw which lacks a native ESHUTDOWN */
> -            error_setg(errp, "server shutting down");
> -            return -EINVAL;
> -        }
>           break;
>       case NBD_STRUCTURED_REPLY_MAGIC:
>           ret = nbd_receive_structured_reply_chunk(ioc, &reply->structured, errp);


-- 
Best regards,
Vladimir

Re: [Qemu-devel] [PATCH for-2.11] nbd/client: Don't hard-disconnect on ESHUTDOWN from server

[Eric Blake]

[-- Attachment #1: Type: text/plain, Size: 1656 bytes --]

On 11/13/2017 01:48 PM, Eric Blake wrote:
> The NBD spec says that a server may fail any transmission request
> with ESHUTDOWN when it is apparent that no further request from
> the client can be successfully honored.  The client is supposed
> to then initiate a soft shutdown (wait for all remaining in-flight
> requests to be answered, then send NBD_CMD_DISC).  However, since
> qemu's server never uses ESHUTDOWN errors, this code was mostly
> untested since its introduction in commit b6f5d3b5.
> 
> More recently, I learned that nbdkit as the NBD server is able to
> send ESHUTDOWN errors, so I finally tested this code, and noticed
> that our client was special-casing ESHUTDOWN to cause a hard
> shutdown (immediate disconnect, with no NBD_CMD_DISC), but only
> if the server sends this error as a simple reply.  Further
> investigation found that commit d2febedb introduced a regression
> where structured replies behave differently than simple replies -
> but that the structured reply behavior is more in line with the
> spec (even if we still lack code in nbd-client.c to properly quit
> sending further requests).  So this patch reverts the portion of
> b6f5d3b5 that introduced an improper hard-disconnect special-case
> at the lower level, and leaves the future enhancement of a nicer
> soft-disconnect at the higher level for another day.
> 
> CC: qemu-trivial@nongnu.org

I think I meant qemu-stable, not qemu-trivial.  At any rate, this is now
queued on my NBD tree for 2.11-rc2.

-- 
Eric Blake, Principal Software Engineer
Red Hat, Inc.           +1-919-301-3266
Virtualization:  qemu.org | libvirt.org


[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 619 bytes --]