| Age | Commit message (Collapse) | Author |
|---|
|
A malicious client can attempt to connect to an NBD server, and then
intentionally delay progress in the handshake, including if it does
not know the TLS secrets. Although the previous two patches reduce
this behavior by capping the default max-connections parameter and
killing slow clients, they did not eliminate the possibility of a
client waiting to close the socket until after the QMP nbd-server-stop
command is executed, at which point qemu would SEGV when trying to
dereference the NULL nbd_server global which is no longer present.
This amounts to a denial of service attack. Worse, if another NBD
server is started before the malicious client disconnects, I cannot
rule out additional adverse effects when the old client interferes
with the connection count of the new server (although the most likely
is a crash due to an assertion failure when checking
nbd_server->connections > 0).
For environments without this patch, the CVE can be mitigated by
ensuring (such as via a firewall) that only trusted clients can
connect to an NBD server. Note that using frameworks like libvirt
that ensure that TLS is used and that nbd-server-stop is not executed
while any trusted clients are still connected will only help if there
is also no possibility for an untrusted client to open a connection
but then stall on the NBD handshake.
Given the previous patches, it would be possible to guarantee that no
clients remain connected by having nbd-server-stop sleep for longer
than the default handshake deadline before finally freeing the global
nbd_server object, but that could make QMP non-responsive for a long
time. So intead, this patch fixes the problem by tracking all client
sockets opened while the server is running, and forcefully closing any
such sockets remaining without a completed handshake at the time of
nbd-server-stop, then waiting until the coroutines servicing those
sockets notice the state change. nbd-server-stop now has a second
AIO_WAIT_WHILE_UNLOCKED (the first is indirectly through the
blk_exp_close_all_type() that disconnects all clients that completed
handshakes), but forced socket shutdown is enough to progress the
coroutines and quickly tear down all clients before the server is
freed, thus finally fixing the CVE.
This patch relies heavily on the fact that nbd/server.c guarantees
that it only calls nbd_blockdev_client_closed() from the main loop
(see the assertion in nbd_client_put() and the hoops used in
nbd_client_put_nonzero() to achieve that); if we did not have that
guarantee, we would also need a mutex protecting our accesses of the
list of connections to survive re-entrancy from independent iothreads.
Although I did not actually try to test old builds, it looks like this
problem has existed since at least commit 862172f45c (v2.12.0, 2017) -
even back when that patch started using a QIONetListener to handle
listening on multiple sockets, nbd_server_free() was already unaware
that the nbd_blockdev_client_closed callback can be reached later by a
client thread that has not completed handshakes (and therefore the
client's socket never got added to the list closed in
nbd_export_close_all), despite that patch intentionally tearing down
the QIONetListener to prevent new clients.
Reported-by: Alexander Ivanov <alexander.ivanov@virtuozzo.com>
Fixes: CVE-2024-7409
CC: qemu-stable@nongnu.org
Signed-off-by: Eric Blake <eblake@redhat.com>
Message-ID: <20240807174943.771624-14-eblake@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
|
|
Allowing an unlimited number of clients to any web service is a recipe
for a rudimentary denial of service attack: the client merely needs to
open lots of sockets without closing them, until qemu no longer has
any more fds available to allocate.
For qemu-nbd, we default to allowing only 1 connection unless more are
explicitly asked for (-e or --shared); this was historically picked as
a nice default (without an explicit -t, a non-persistent qemu-nbd goes
away after a client disconnects, without needing any additional
follow-up commands), and we are not going to change that interface now
(besides, someday we want to point people towards qemu-storage-daemon
instead of qemu-nbd).
But for qemu proper, and the newer qemu-storage-daemon, the QMP
nbd-server-start command has historically had a default of unlimited
number of connections, in part because unlike qemu-nbd it is
inherently persistent until nbd-server-stop. Allowing multiple client
sockets is particularly useful for clients that can take advantage of
MULTI_CONN (creating parallel sockets to increase throughput),
although known clients that do so (such as libnbd's nbdcopy) typically
use only 8 or 16 connections (the benefits of scaling diminish once
more sockets are competing for kernel attention). Picking a number
large enough for typical use cases, but not unlimited, makes it
slightly harder for a malicious client to perform a denial of service
merely by opening lots of connections withot progressing through the
handshake.
This change does not eliminate CVE-2024-7409 on its own, but reduces
the chance for fd exhaustion or unlimited memory usage as an attack
surface. On the other hand, by itself, it makes it more obvious that
with a finite limit, we have the problem of an unauthenticated client
holding 100 fds opened as a way to block out a legitimate client from
being able to connect; thus, later patches will further add timeouts
to reject clients that are not making progress.
This is an INTENTIONAL change in behavior, and will break any client
of nbd-server-start that was not passing an explicit max-connections
parameter, yet expects more than 100 simultaneous connections. We are
not aware of any such client (as stated above, most clients aware of
MULTI_CONN get by just fine on 8 or 16 connections, and probably cope
with later connections failing by relying on the earlier connections;
libvirt has not yet been passing max-connections, but generally
creates NBD servers with the intent for a single client for the sake
of live storage migration; meanwhile, the KubeSAN project anticipates
a large cluster sharing multiple clients [up to 8 per node, and up to
100 nodes in a cluster], but it currently uses qemu-nbd with an
explicit --shared=0 rather than qemu-storage-daemon with
nbd-server-start).
We considered using a deprecation period (declare that omitting
max-parameters is deprecated, and make it mandatory in 3 releases -
then we don't need to pick an arbitrary default); that has zero risk
of breaking any apps that accidentally depended on more than 100
connections, and where such breakage might not be noticed under unit
testing but only under the larger loads of production usage. But it
does not close the denial-of-service hole until far into the future,
and requires all apps to change to add the parameter even if 100 was
good enough. It also has a drawback that any app (like libvirt) that
is accidentally relying on an unlimited default should seriously
consider their own CVE now, at which point they are going to change to
pass explicit max-connections sooner than waiting for 3 qemu releases.
Finally, if our changed default breaks an app, that app can always
pass in an explicit max-parameters with a larger value.
It is also intentional that the HMP interface to nbd-server-start is
not changed to expose max-connections (any client needing to fine-tune
things should be using QMP).
Suggested-by: Daniel P. Berrangé <berrange@redhat.com>
Signed-off-by: Eric Blake <eblake@redhat.com>
Message-ID: <20240807174943.771624-12-eblake@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
[ericb: Expand commit message to summarize Dan's argument for why we
break corner-case back-compat behavior without a deprecation period]
Signed-off-by: Eric Blake <eblake@redhat.com>
|
|
Upcoming patches to fix a CVE need to track an opaque pointer passed
in by the owner of a client object, as well as request for a time
limit on how fast negotiation must complete. Prepare for that by
changing the signature of nbd_client_new() and adding an accessor to
get at the opaque pointer, although for now the two servers
(qemu-nbd.c and blockdev-nbd.c) do not change behavior even though
they pass in a new default timeout value.
Suggested-by: Vladimir Sementsov-Ogievskiy <vsementsov@yandex-team.ru>
Signed-off-by: Eric Blake <eblake@redhat.com>
Message-ID: <20240807174943.771624-11-eblake@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
[eblake: s/LIMIT/MAX_SECS/ as suggested by Dan]
Signed-off-by: Eric Blake <eblake@redhat.com>
|
|
The has_FOO for pointer-valued FOO are redundant, except for arrays.
They are also a nuisance to work with. Recent commit "qapi: Start to
elide redundant has_FOO in generated C" provided the means to elide
them step by step. This is the step for qapi/block*.json.
Said commit explains the transformation in more detail.
There is one instance of the invariant violation mentioned there:
qcow2_signal_corruption() passes false, "" when node_name is an empty
string. Take care to pass NULL then.
The previous two commits cleaned up two more.
Additionally, helper bdrv_latency_histogram_stats() loses its output
parameters and returns a value instead.
Cc: Kevin Wolf <kwolf@redhat.com>
Cc: Hanna Reitz <hreitz@redhat.com>
Cc: qemu-block@nongnu.org
Signed-off-by: Markus Armbruster <armbru@redhat.com>
Message-Id: <20221104160712.3005652-11-armbru@redhat.com>
[Fixes for #ifndef LIBRBD_SUPPORTS_ENCRYPTION and MacOS squashed in]
|
|
According to the NBD spec, a server that advertises
NBD_FLAG_CAN_MULTI_CONN promises that multiple client connections will
not see any cache inconsistencies: when properly separated by a single
flush, actions performed by one client will be visible to another
client, regardless of which client did the flush.
We always satisfy these conditions in qemu - even when we support
multiple clients, ALL clients go through a single point of reference
into the block layer, with no local caching. The effect of one client
is instantly visible to the next client. Even if our backend were a
network device, we argue that any multi-path caching effects that
would cause inconsistencies in back-to-back actions not seeing the
effect of previous actions would be a bug in that backend, and not the
fault of caching in qemu. As such, it is safe to unconditionally
advertise CAN_MULTI_CONN for any qemu NBD server situation that
supports parallel clients.
Note, however, that we don't want to advertise CAN_MULTI_CONN when we
know that a second client cannot connect (for historical reasons,
qemu-nbd defaults to a single connection while nbd-server-add and QMP
commands default to unlimited connections; but we already have
existing means to let either style of NBD server creation alter those
defaults). This is visible by no longer advertising MULTI_CONN for
'qemu-nbd -r' without -e, as in the iotest nbd-qemu-allocation.
The harder part of this patch is setting up an iotest to demonstrate
behavior of multiple NBD clients to a single server. It might be
possible with parallel qemu-io processes, but I found it easier to do
in python with the help of libnbd, and help from Nir and Vladimir in
writing the test.
Signed-off-by: Eric Blake <eblake@redhat.com>
Suggested-by: Nir Soffer <nsoffer@redhat.com>
Suggested-by: Vladimir Sementsov-Ogievskiy <v.sementsov-og@mail.ru>
Message-Id: <20220512004924.417153-3-eblake@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
|
|
The next patch wants to adjust whether the NBD server code advertises
MULTI_CONN based on whether it is known if the server limits to
exactly one client. For a server started by QMP, this information is
obtained through nbd_server_start (which can support more than one
export); but for qemu-nbd (which supports exactly one export), it is
controlled only by the command-line option -e/--shared. Since we
already have a hook function used by qemu-nbd, it's easiest to just
alter its signature to fit our needs.
Signed-off-by: Eric Blake <eblake@redhat.com>
Message-Id: <20220512004924.417153-2-eblake@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
|
|
Hi all! Current logic of relying on search through backing chain is not
safe neither convenient.
Sometimes it leads to necessity of extra bitmap copying. Also, we are
going to add "snapshot-access" driver, to access some snapshot state
through NBD. And this driver is not formally a filter, and of course
it's not a COW format driver. So, searching through backing chain will
not work. Instead of widening the workaround of bitmap searching, let's
extend the interface so that user can select bitmap precisely.
Note, that checking for bitmap active status is not copied to the new
API, I don't see a reason for it, user should understand the risks. And
anyway, bitmap from other node is unrelated to this export being
read-only or read-write.
Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@openvz.org>
Message-Id: <20220314213226.362217-3-v.sementsov-og@mail.ru>
[eblake: Adjust S-o-b to Vladimir's new email, with permission]
Reviewed-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Eric Blake <eblake@redhat.com>
|
|
The TLS usage for NBD was restricted to IP sockets because validating
x509 certificates requires knowledge of the hostname that the client
is connecting to.
TLS does not have to use x509 certificates though, as PSK (pre-shared
keys) provide an alternative credential option. These have no
requirement for a hostname and can thus be trivially used for UNIX
sockets.
Furthermore, with the ability to overide the default hostname for
TLS validation in the previous patch, it is now also valid to want
to use x509 certificates with FD passing and UNIX sockets.
Reviewed-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
Message-Id: <20220304193610.3293146-6-berrange@redhat.com>
Signed-off-by: Eric Blake <eblake@redhat.com>
|
|
Avoid accessing QCryptoTLSCreds internals by using
the qcrypto_tls_creds_check_endpoint() helper.
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
|
|
Our default of a backlog of 1 connection is rather puny; it gets in
the way when we are explicitly allowing multiple clients (such as
qemu-nbd -e N [--shared], or nbd-server-start with its default
"max-connections":0 for unlimited), but is even a problem when we
stick to qemu-nbd's default of only 1 active client but use -t
[--persistent] where a second client can start using the server once
the first finishes. While the effects are less noticeable on TCP
sockets (since the client can poll() to learn when the server is ready
again), it is definitely observable on Unix sockets, where on Linux, a
client will fail with EAGAIN and no recourse but to sleep an arbitrary
amount of time before retrying if the server backlog is already full.
Since QMP nbd-server-start is always persistent, it now always
requests a backlog of SOMAXCONN; meanwhile, qemu-nbd will request
SOMAXCONN if persistent, otherwise its backlog should be based on the
expected number of clients.
See https://bugzilla.redhat.com/1925045 for a demonstration of where
our low backlog prevents libnbd from connecting as many parallel
clients as it wants.
Reported-by: Richard W.M. Jones <rjones@redhat.com>
Signed-off-by: Eric Blake <eblake@redhat.com>
CC: qemu-stable@nongnu.org
Message-Id: <20210209152759.209074-2-eblake@redhat.com>
Tested-by: Richard W.M. Jones <rjones@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Signed-off-by: Eric Blake <eblake@redhat.com>
|
|
Since 'block-export-add' is new to 5.2, we can still tweak the
interface; there, allowing 'bitmaps':['str'] is nicer than
'bitmap':'str'. This wires up the qapi and qemu-nbd changes to permit
passing multiple bitmaps as distinct metadata contexts that the NBD
client may request, but the actual support for more than one will
require a further patch to the server.
Note that there are no changes made to the existing deprecated
'nbd-server-add' command; this required splitting the QAPI type
BlockExportOptionsNbd, which fortunately does not affect QMP
introspection.
Signed-off-by: Eric Blake <eblake@redhat.com>
Message-Id: <20201027050556.269064-5-eblake@redhat.com>
Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Reviewed-by: Peter Krempa <pkrempa@redhat.com>
|
|
Rather than open-coding the translation from the deprecated
NbdServerAddOptions type to the preferred BlockExportOptionsNbd, it's
better to utilize QAPI_CLONE_MEMBERS. This solves a couple of issues:
first, if we do any more refactoring of the base type (which an
upcoming patch plans to do), we don't have to revisit the open-coding.
Second, our assignment to arg->name is fishy: the generated QAPI code
for qapi_free_NbdServerAddOptions does not visit arg->name if
arg->has_name is false, but if it DID visit it, we would have
introduced a double-free situation when arg is finally freed.
Signed-off-by: Eric Blake <eblake@redhat.com>
Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Reviewed-by: Markus Armbruster <armbru@redhat.com>
Message-Id: <20201027050556.269064-4-eblake@redhat.com>
|
|
There is no real reason any more why nbd_export_new() and
nbd_export_create() should be separate functions. The latter only
performs a few checks before it calls the former.
What makes the current state stand out is that it's the only function in
BlockExportDriver that is not a static function inside nbd/server.c, but
a small wrapper in blockdev-nbd.c that then calls back into nbd/server.c
for the real functionality.
Move all the checks to nbd/server.c and make the resulting function
static to improve readability.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
Message-Id: <20200924152717.287415-27-kwolf@redhat.com>
Acked-by: Stefan Hajnoczi <stefanha@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
|
|
The 'writable' option is a basic option that will probably be applicable
to most if not all export types that we will implement. Move it from NBD
to the generic BlockExport layer.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
Message-Id: <20200924152717.287415-26-kwolf@redhat.com>
Acked-by: Stefan Hajnoczi <stefanha@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
|
|
Every export type will need a BlockBackend, so creating it centrally in
blk_exp_add() instead of the .create driver callback avoids duplication.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
Message-Id: <20200924152717.287415-24-kwolf@redhat.com>
Acked-by: Stefan Hajnoczi <stefanha@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
|
|
Implement a new QMP command block-export-del and make nbd-server-remove
a wrapper around it.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
Message-Id: <20200924152717.287415-21-kwolf@redhat.com>
Acked-by: Stefan Hajnoczi <stefanha@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
|
|
The reference owned by the user/monitor that is created when adding the
export and dropped when removing it was tied to the 'exports' list in
nbd/server.c. Every block export will have a user reference, so move it
to the block export level and tie it to the 'block_exports' list in
block/export/export.c instead. This is necessary for introducing a QMP
command for removing exports.
Note that exports are present in block_exports even after the user has
requested shutdown. This is different from NBD's exports where exports
are immediately removed on a shutdown request, even if they are still in
the process of shutting down. In order to avoid that the user still
interacts with an export that is shutting down (and possibly removes it
a second time), we need to remember if the user actually still owns it.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
Message-Id: <20200924152717.287415-20-kwolf@redhat.com>
Acked-by: Stefan Hajnoczi <stefanha@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
|
|
We'll need an id to identify block exports in monitor commands. This
adds one.
Note that this is different from the 'name' option in the NBD server,
which is the externally visible export name. While block export ids need
to be unique in the whole process, export names must be unique only for
the same server. Different export types or (potentially in the future)
multiple NBD servers can have the same export name externally, but still
need different block export ids internally.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
Message-Id: <20200924152717.287415-19-kwolf@redhat.com>
Acked-by: Stefan Hajnoczi <stefanha@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
|
|
This adds a function to shut down all block exports, and another one to
shut down the block exports of a single type. The latter is used for now
when stopping the NBD server. As soon as we implement support for
multiple NBD servers, we'll need a per-server list of exports and it
will be replaced by a function using that.
As a side effect, the BlockExport layer has a list tracking all existing
exports now. closed_exports loses its only user and can go away.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
Message-Id: <20200924152717.287415-18-kwolf@redhat.com>
Acked-by: Stefan Hajnoczi <stefanha@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
|
|
Instead of letting the driver allocate and return the BlockExport
object, allocate it already in blk_exp_add() and pass it. This allows us
to initialise the generic part before calling into the driver so that
the driver can just use these values instead of having to parse the
options a second time.
For symmetry, move freeing the BlockExport to blk_exp_unref().
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
Message-Id: <20200924152717.287415-17-kwolf@redhat.com>
Acked-by: Stefan Hajnoczi <stefanha@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
|
|
Every block export needs a block node to export, so add a 'node-name'
option to BlockExportOptions and remove the replaced option 'device'
from BlockExportOptionsNbd.
To maintain compatibility in nbd-server-add, BlockExportOptionsNbd needs
to be wrapped by a new type NbdServerAddOptions that adds 'device' back
because nbd-server-add doesn't use the BlockExportOptions base type at
all (so even without changing it to a 'node-name' option in
block-export-add, this compatibility code would be necessary).
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
Message-Id: <20200924152717.287415-16-kwolf@redhat.com>
Acked-by: Stefan Hajnoczi <stefanha@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
|
|
Having a refcount makes sense for all types of block exports. It is also
a prerequisite for keeping a list of all exports at the BlockExport
level.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
Message-Id: <20200924152717.287415-14-kwolf@redhat.com>
Acked-by: Stefan Hajnoczi <stefanha@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
|
|
With this change, NBD exports are now only created through the
BlockExport interface. This allows us finally to move things from the
NBD layer to the BlockExport layer if they make sense for other export
types, too.
blk_exp_add() returns only a weak reference, so the explicit
nbd_export_put() goes away.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
Message-Id: <20200924152717.287415-12-kwolf@redhat.com>
Acked-by: Stefan Hajnoczi <stefanha@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
|
|
The export close callback is unused by the built-in NBD server. qemu-nbd
uses it only during shutdown to wait for the unrefed export to actually
go away. It can just use nbd_export_close_all() instead and do without
the callback.
This removes the close callback from nbd_export_new() and makes both
callers of it more similar.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Message-Id: <20200924152717.287415-11-kwolf@redhat.com>
Acked-by: Stefan Hajnoczi <stefanha@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
|
|
qemu-nbd allows use of writethrough cache modes, which mean that write
requests made through NBD will cause a flush before they complete.
Expose the same functionality in block-export-add.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
Message-Id: <20200924152717.287415-10-kwolf@redhat.com>
Acked-by: Stefan Hajnoczi <stefanha@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
|
|
This is a QMP equivalent of qemu-nbd's --shared option, limiting the
maximum number of clients that can attach at the same time.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Message-Id: <20200924152717.287415-9-kwolf@redhat.com>
Acked-by: Stefan Hajnoczi <stefanha@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
|
|
nbd-server-add tries to be convenient and adds two questionable
features that we don't want to share in block-export-add, even for NBD
exports:
1. When requesting a writable export of a read-only device, the export
is silently downgraded to read-only. This should be an error in the
context of block-export-add.
2. When using a BlockBackend name, unplugging the device from the guest
will automatically stop the NBD server, too. This may sometimes be
what you want, but it could also be very surprising. Let's keep
things explicit with block-export-add. If the user wants to stop the
export, they should tell us so.
Move these things into the nbd-server-add QMP command handler so that
they apply only there.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
Message-Id: <20200924152717.287415-8-kwolf@redhat.com>
Acked-by: Stefan Hajnoczi <stefanha@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
|
|
Instead of implementing qemu-nbd --offset in the NBD code, just put a
raw block node with the requested offset on top of the user image and
rely on that doing the job.
This does not only simplify the nbd_export_new() interface and bring it
closer to the set of options that the nbd-server-add QMP command offers,
but in fact it also eliminates a potential source for bugs in the NBD
code which previously had to add the offset manually in all relevant
places.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Message-Id: <20200924152717.287415-7-kwolf@redhat.com>
Acked-by: Stefan Hajnoczi <stefanha@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
|
|
We want to have a common set of commands for all types of block exports.
Currently, this is only NBD, but we're going to add more types.
This patch adds the basic BlockExport and BlockExportDriver structs and
a QMP command block-export-add that creates a new export based on the
given BlockExportOptions.
qmp_nbd_server_add() becomes a wrapper around qmp_block_export_add().
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
Message-Id: <20200924152717.287415-5-kwolf@redhat.com>
Acked-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
|
|
The name BlockExport will be used for the struct containing the runtime
state of block exports, so change the name of export creation options.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Message-Id: <20200924152717.287415-4-kwolf@redhat.com>
Acked-by: Stefan Hajnoczi <stefanha@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
|
|
Move all block export related types and commands from block-core to the
new QAPI module block-export.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Message-Id: <20200924152717.287415-3-kwolf@redhat.com>
Acked-by: Stefan Hajnoczi <stefanha@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
|
|
Move the arguments of nbd-server-add to a new struct BlockExportNbd and
convert the command to 'boxed': true. This makes it easier to share code
with the storage daemon.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Message-Id: <20200224143008.13362-11-kwolf@redhat.com>
Acked-by: Stefan Hajnoczi <stefanha@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
|
|
Add a --nbd-server option to qemu-storage-daemon to start the built-in
NBD server right away. It maps the arguments for nbd-server-start to the
command line, with the exception that it uses SocketAddress instead of
SocketAddressLegacy: New interfaces shouldn't use legacy types, and the
additional nesting would be nasty on the command line.
Example (only with required options):
--nbd-server addr.type=inet,addr.host=localhost,addr.port=10809
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Message-Id: <20200224143008.13362-10-kwolf@redhat.com>
Acked-by: Stefan Hajnoczi <stefanha@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
|
|
Allow blockdevs to match the feature already present in qemu-nbd -D.
Enhance iotest 223 to cover it.
Signed-off-by: Eric Blake <eblake@redhat.com>
Reviewed-by: Maxim Levitsky <mlevitsk@redhat.com>
Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Message-Id: <20191114024635.11363-5-eblake@redhat.com>
|
|
Qemu as server currently won't accept export names larger than 256
bytes, nor create dirty bitmap names longer than 1023 bytes, so most
uses of qemu as client or server have no reason to get anywhere near
the NBD spec maximum of a 4k limit per string.
However, we weren't actually enforcing things, ignoring when the
remote side violates the protocol on input, and also having several
code paths where we send oversize strings on output (for example,
qemu-nbd --description could easily send more than 4k). Tighten
things up as follows:
client:
- Perform bounds check on export name and dirty bitmap request prior
to handing it to server
- Validate that copied server replies are not too long (ignoring
NBD_INFO_* replies that are not copied is not too bad)
server:
- Perform bounds check on export name and description prior to
advertising it to client
- Reject client name or metadata query that is too long
- Adjust things to allow full 4k name limit rather than previous
256 byte limit
Signed-off-by: Eric Blake <eblake@redhat.com>
Message-Id: <20191114024635.11363-4-eblake@redhat.com>
Reviewed-by: Maxim Levitsky <mlevitsk@redhat.com>
Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
|
|
When iothreads are in use, the failure to grab the aio context results
in an assertion failure when trying to unlock things during blk_unref,
when trying to unlock a mutex that was not locked. In short, all
calls to nbd_export_put need to done while within the correct aio
context. But since nbd_export_put can recursively reach itself via
nbd_export_close, and recursively grabbing the context would deadlock,
we can't do the context grab directly in those functions, but must do
so in their callers.
Hoist the use of the correct aio_context from nbd_export_new() to its
caller qmp_nbd_server_add(). Then tweak qmp_nbd_server_remove(),
nbd_eject_notifier(), and nbd_esport_close_all() to grab the right
context, so that all callers during qemu now own the context before
nbd_export_put() can call blk_unref().
Remaining uses in qemu-nbd don't matter (since that use case does not
support iothreads).
Suggested-by: Kevin Wolf <kwolf@redhat.com>
Signed-off-by: Eric Blake <eblake@redhat.com>
Message-Id: <20190917023917.32226-1-eblake@redhat.com>
Reviewed-by: Sergio Lopez <slp@redhat.com>
|
|
When creating a read-only image, we are still advertising support for
TRIM and WRITE_ZEROES to the client, even though the client should not
be issuing those commands. But seeing this requires looking across
multiple functions:
All callers to nbd_export_new() passed a single flag based solely on
whether the export allows writes. Later, we then pass a constant set
of flags to nbd_negotiate_options() (namely, the set of flags which we
always support, at least for writable images), which is then further
dynamically modified with NBD_FLAG_SEND_DF based on client requests
for structured options. Finally, when processing NBD_OPT_EXPORT_NAME
or NBD_OPT_EXPORT_GO we bitwise-or the original caller's flag with the
runtime set of flags we've built up over several functions.
Let's refactor things to instead compute a baseline of flags as soon
as possible which gets shared between multiple clients, in
nbd_export_new(), and changing the signature for the callers to pass
in a simpler bool rather than having to figure out flags. We can then
get rid of the 'myflags' parameter to various functions, and instead
refer to client for everything we need (we still have to perform a
bitwise-OR for NBD_FLAG_SEND_DF during NBD_OPT_EXPORT_NAME and
NBD_OPT_EXPORT_GO, but it's easier to see what is being computed).
This lets us quit advertising senseless flags for read-only images, as
well as making the next patch for exposing FAST_ZERO support easier to
write.
Signed-off-by: Eric Blake <eblake@redhat.com>
Message-Id: <20190823143726.27062-2-eblake@redhat.com>
Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
[eblake: improve commit message, update iotest 223]
|
|
The NBD specification defines NBD_FLAG_CAN_MULTI_CONN, which can be
advertised when the server promises cache consistency between
simultaneous clients (basically, rules that determine what FUA and
flush from one client are able to guarantee for reads from another
client). When we don't permit simultaneous clients (such as qemu-nbd
without -e), the bit makes no sense; and for writable images, we
probably have a lot more work before we can declare that actions from
one client are cache-consistent with actions from another. But for
read-only images, where flush isn't changing any data, we might as
well advertise multi-conn support. What's more, advertisement of the
bit makes it easier for clients to determine if 'qemu-nbd -e' was in
use, where a second connection will succeed rather than hang until the
first client goes away.
This patch affects qemu as server in advertising the bit. We may want
to consider patches to qemu as client to attempt parallel connections
for higher throughput by spreading the load over those connections
when a server advertises multi-conn, but for now sticking to one
connection per nbd:// BDS is okay.
See also: https://bugzilla.redhat.com/1708300
Signed-off-by: Eric Blake <eblake@redhat.com>
Message-Id: <20190815185024.7010-1-eblake@redhat.com>
[eblake: tweak blockdev-nbd.c to not request shared when writable,
fix iotest 233]
Reviewed-by: John Snow <jsnow@redhat.com>
|
|
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Signed-off-by: Juan Quintela <quintela@redhat.com>
|
|
In my "build everything" tree, changing sysemu/sysemu.h triggers a
recompile of some 5400 out of 6600 objects (not counting tests and
objects that don't depend on qemu/osdep.h).
Almost a third of its inclusions are actually superfluous. Delete
them. Downgrade two more to qapi/qapi-types-run-state.h, and move one
from char/serial.h to char/serial.c.
hw/semihosting/config.c, monitor/monitor.c, qdev-monitor.c, and
stubs/semihost.c define variables declared in sysemu/sysemu.h without
including it. The compiler is cool with that, but include it anyway.
This doesn't reduce actual use much, as it's still included into
widely included headers. The next commit will tackle that.
Signed-off-by: Markus Armbruster <armbru@redhat.com>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Message-Id: <20190812052359.30071-27-armbru@redhat.com>
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
|
|
As with the previous patch to qemu-nbd, the nbd-server-start QMP command
also needs to be able to specify authorization when enabling TLS encryption.
First the client must create a QAuthZ object instance using the
'object-add' command:
{
'execute': 'object-add',
'arguments': {
'qom-type': 'authz-list',
'id': 'authz0',
'parameters': {
'policy': 'deny',
'rules': [
{
'match': '*CN=fred',
'policy': 'allow'
}
]
}
}
}
They can then reference this in the new 'tls-authz' parameter when
executing the 'nbd-server-start' command:
{
'execute': 'nbd-server-start',
'arguments': {
'addr': {
'type': 'inet',
'host': '127.0.0.1',
'port': '9000'
},
'tls-creds': 'tls0',
'tls-authz': 'authz0'
}
}
Reviewed-by: Eric Blake <eblake@redhat.com>
Reviewed-by: Juan Quintela <quintela@redhat.com>
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
Message-Id: <20190227162035.18543-3-berrange@redhat.com>
Signed-off-by: Eric Blake <eblake@redhat.com>
|
|
We only had two callers to nbd_export_new; qemu-nbd.c always
passed a valid offset/length pair (because it already checked
the file length, to ensure that offset was in bounds), while
blockdev-nbd.c always passed 0/-1. Then nbd_export_new reduces
the size to a multiple of BDRV_SECTOR_SIZE (can only happen
when offset is not sector-aligned, since bdrv_getlength()
currently rounds up) (someday, it would be nice to have
byte-accurate lengths - but not today).
However, I'm finding it easier to work with the code if we are
consistent on having both callers pass in a valid length, and
just assert that things are sane in nbd_export_new, meaning
that no negative values were passed, and that offset+size does
not exceed 63 bits (as that really is a fundamental limit to
later operations, whether we use off_t or uint64_t).
Signed-off-by: Eric Blake <eblake@redhat.com>
Message-Id: <20190117193658.16413-6-eblake@redhat.com>
Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
|
|
We only have one caller that wants to export a bitmap name,
which it does right after creation of the export. But there is
still a brief window of time where an NBD client could see the
export but not the dirty bitmap, which a robust client would
have to interpret as meaning the entire image should be treated
as dirty. Better is to eliminate the window entirely, by
inlining nbd_export_bitmap() into nbd_export_new(), and refusing
to create the bitmap in the first place if the requested bitmap
can't be located.
We also no longer need logic for setting a different bitmap
name compared to the bitmap being exported.
Signed-off-by: Eric Blake <eblake@redhat.com>
Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Message-Id: <20190111194720.15671-8-eblake@redhat.com>
|
|
Now that nbd-server-add can do the same functionality (well, other
than making the exported bitmap name different than the underlying
bitamp - but we argued that was not essential, since it is just as
easy to create a new non-persistent bitmap with the desired name),
we no longer need the experimental separate command.
Signed-off-by: Eric Blake <eblake@redhat.com>
Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Message-Id: <20190111194720.15671-7-eblake@redhat.com>
|
|
With the experimental x-nbd-server-add-bitmap command, there was
a window of time where an NBD client could see the export but not
the associated dirty bitmap, which can cause a client that planned
on using the dirty bitmap to be forced to treat the entire image
as dirty as a safety fallback. Furthermore, if the QMP client
successfully exports a disk but then fails to add the bitmap, it
has to take on the burden of removing the export. Since we don't
allow changing the exposed dirty bitmap (whether to a different
bitmap, or removing advertisement of the bitmap), it is nicer to
make the bitmap tied to the export at the time the export is
created, with automatic failure to export if the bitmap is not
available.
The experimental command included an optional 'bitmap-export-name'
field for remapping the name exposed over NBD to be different from
the bitmap name stored on disk. However, my libvirt demo code
for implementing differential backups on top of persistent bitmaps
did not need to take advantage of that feature (it is instead
possible to create a new temporary bitmap with the desired name,
use block-dirty-bitmap-merge to merge one or more persistent
bitmaps into the temporary, then associate the temporary with the
NBD export, if control is needed over the exported bitmap name).
Hence, I'm not copying that part of the experiment over to the
stable addition. For more details on the libvirt demo, see
https://www.redhat.com/archives/libvir-list/2018-October/msg01254.html,
https://kvmforum2018.sched.com/event/FzuB/facilitating-incremental-backup-eric-blake-red-hat
This patch focuses on the user interface, and reduces (but does
not completely eliminate) the window where an NBD client can see
the export but not the dirty bitmap, with less work to clean up
after errors. Later patches will add further cleanups now that
this interface is declared stable via a single QMP command,
including removing the race window.
Update test 223 to use the new interface.
Signed-off-by: Eric Blake <eblake@redhat.com>
Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Message-Id: <20190111194720.15671-6-eblake@redhat.com>
|
|
The existing NBD code had a weird split where nbd_export_new()
created an export but did not add it to the list of exported
names until a later nbd_export_set_name() came along and grabbed
a second reference on the object; later, the first call to
nbd_export_close() drops the second reference while removing
the export from the list. This is in part because the QAPI
NbdServerRemoveNode enum documents the possibility of adding a
mode where we could do a soft disconnect: preventing new clients,
but waiting for existing clients to gracefully quit, based on
the mode used when calling nbd_export_close().
But in spite of all that, note that we never change the name of
an NBD export while it is exposed, which means it is easier to
just inline the process of setting the name as part of creating
the export.
Inline the contents of nbd_export_set_name() and
nbd_export_set_description() into the two points in an export
lifecycle where they matter, then adjust both callers to pass
the name up front. Note that for creation, all callers pass a
non-NULL name, (passing NULL at creation was for old style
servers, but we removed support for that in commit 7f7dfe2a),
so we can add an assert and do things unconditionally; but for
cleanup, because of the dual nature of nbd_export_close(), we
still have to be careful to avoid use-after-free. Along the
way, add a comment reminding ourselves of the potential of
adding a middle mode disconnect.
Signed-off-by: Eric Blake <eblake@redhat.com>
Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Message-Id: <20190111194720.15671-5-eblake@redhat.com>
|
|
Since we already forbid other nbd-server commands when not
in the right state, it is unlikely that any caller was relying
on a second stop to behave as a silent no-op. Update iotest
223 to show the improved behavior.
Signed-off-by: Eric Blake <eblake@redhat.com>
Message-Id: <20190111194720.15671-3-eblake@redhat.com>
Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
|
|
After the previous commit, nbd_client_new's first parameter is always
NULL. Let's drop it with all corresponding old-style negotiation code
path which is unreachable now.
Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Message-Id: <20181003170228.95973-3-vsementsov@virtuozzo.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
[eblake: re-wrap short line]
Signed-off-by: Eric Blake <eblake@redhat.com>
|
|
For now, the actual command ix x-nbd-server-add-bitmap, reflecting
the fact that we are still working on libvirt code that proves the
command works as needed, and also the fact that we may remove
bitmap-export-name (and just require that the exported name be the
bitmap name).
Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Message-Id: <20180609151758.17343-6-vsementsov@virtuozzo.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
[eblake: make the command experimental by adding x- prefix]
Signed-off-by: Eric Blake <eblake@redhat.com>
|
|
In my "build everything" tree, a change to the types in
qapi-schema.json triggers a recompile of about 4800 out of 5100
objects.
The previous commit split up qmp-commands.h, qmp-event.h, qmp-visit.h,
qapi-types.h. Each of these headers still includes all its shards.
Reduce compile time by including just the shards we actually need.
To illustrate the benefits: adding a type to qapi/migration.json now
recompiles some 2300 instead of 4800 objects. The next commit will
improve it further.
Signed-off-by: Markus Armbruster <armbru@redhat.com>
Message-Id: <20180211093607.27351-24-armbru@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>
[eblake: rebase to master]
Signed-off-by: Eric Blake <eblake@redhat.com>
|
