John Ferlan [Mon, 2 May 2016 10:43:32 +0000 (06:43 -0400)]
qemu: Introduce qemuDomainSecretIV
Add the data structure and infrastructure to support an initialization
vector (IV) secrets. The IV secret generation will need to have access
to the domain private master key, so let's make sure the prepare disk
and hostdev functions can accept that now.
Anywhere that needs to make a decision over which secret type to use
in order to fill in or use the IV secret has a switch added.
John Ferlan [Mon, 2 May 2016 17:31:47 +0000 (13:31 -0400)]
qemu: Move qemuDomainSecretDestroy to qemuProcessLaunch
Rather than need to call qemuDomainSecretDestroy after any call to
qemuProcessLaunch, let's do the destroy in qemuProcessLaunch since
that's where command line is eventually generated and processed. Once
it's generated, we can clear out the secrets.
John Ferlan [Mon, 2 May 2016 17:20:55 +0000 (13:20 -0400)]
qemu: Move qemuDomainSecretPrepare to qemuProcessPrepareDomain
Commit id '40d8e2ba3' added the function to qemuProcessStart because
in order to set up some secrets in the future we will need the master
key. However, since the previous patch split the master key creation
into two parts (create just the key and create the file), we can now
call qemuDomainSecretPrepare from qemuProcessPrepareDomain since the
file is not necessary.
John Ferlan [Mon, 2 May 2016 17:11:24 +0000 (13:11 -0400)]
qemu: Split out the master key create and write
A recent review of related changes noted that we should split the creation
(or generation) of the master key into the qemuProcessPrepareDomain and leave
the writing of the master key for qemuProcessPrepareHost.
Made the adjustment and modified some comments to functions that have
changed calling parameters, but didn't change the intro doc.
This removes the opencoded payload freeing in the client, to use
the shared virNetMessageClearPayload call. Two changes:
- ClearPayload sets nfds=0, which fixes a potential crash if
an error path called virNetMessageFree/Clear on the message
after fds was free'd
- We drop the inner loop VIR_FORCE_CLOSE... this may mean fds are
kept open a little bit longer if the call is blocking but in
practice I don't think it will have any effect
I've noticed this while trying to compile libvirt on my arm box.
CC rpc/libvirt_net_rpc_server_la-virnetserverclient.lo
rpc/virnetserverclient.c: In function 'virNetServerClientNewPostExecRestart':
rpc/virnetserverclient.c:516:45: error: cast increases required alignment of target type [-Werror=cast-align]
(long long *) ×tamp) < 0) {
^
cc1: all warnings being treated as errors
Problem is, @timestap is defined as time_t which is 32 bits long,
and we are typecasting it to long long which is 64bits long.
Solution is to make @timestamp type of long long. At the same
time, we can make @conn_time in _virNetServerClient struct long
long too. There is no need for it to be type of time_t.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
virNetServerClientNewPostExecRestart: Drop useless typecasts
In this function, @id is defined as unsigned long long. When
passing this variable to virJSONValueObjectGetNumberUlong(),
well address of this variable, it's typecasted to ull*. There
is no need for that. It's a same story with @nrequests_max.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Adding new *.replies files for qemucapabilitiestest or updating the
files when libvirt adds an additional QMP command into the probing
process is quite painful. The goal of the new qemucapsprobe command is
to make this process as easy as
json_reformat uses two spaces for when indenting nested objects, let's
do the same. The result of virJSONValueToString will be exactly the same
as json_reformat would produce.
qemu: Make virQEMUCapsNewForBinary usable from tests
virQEMUCapsNewForBinary unconditionally loads data from cache and probes
using both QMP and -help parsing, which is suboptimal when we want to
use it in tests.
Currently all qemu driver tests are statically linked to qemu driver
library, which makes it impossible to mock any API from the library.
This patch creates a shared qemu driver library which can be used
instead of the static one.
NB we can't use libvirt_driver_qemu.so directly since it is linked with
-module and it is supposed to be dlopened.
Bhyve supports ACPI shutdown by issuing SIGTERM signal to a bhyve
process.
Add the bhyveDomainShutdown() function and virBhyveProcessShutdown()
helper function that just sends SIGTERM to VM's bhyve process. If
a guest supports ACPI shutdown then process will be terminated and
this event will be noticed by the bhyve monitor code that will handle
setting proper status and clean up VM's resources by calling
virBhyveProcessStop().
bhyve: drop virProcessKillPainfully() from destroy
Current implementation of domainDestroy for bhyve calls
virProcessKillPainfully() for the bhyve process and then
executes "bhyvectl --destroy".
This is wrong for two reasons:
* bhyvectl --destroy alone is sufficient because it terminates
the process
* virProcessKillPainfully() first sends SIGTERM and after few
attempts sends SIGKILL. As SIGTERM triggers ACPI shutdown that
we're not interested in, it creates an unwanted side effect in
domainDestroy.
Also, destroy monitor only after "bhyvectl --destroy" command succeeded
to avoid a case when the command fails and domain remains running, but
not being monitored anymore.
Eric Blake [Wed, 4 May 2016 19:15:11 +0000 (13:15 -0600)]
build: fix 32-bit build of admin
We can't guarantee which 64-bit type will be used in an RPC struct;
while %lu worked on 64-bit Linux, that won't always be the type
used on all 64-bit platforms; and certainly is not right for 32-bit:
admin.c: In function 'adminDispatchClientGetInfo':
admin.c:265:25: error: format '%lu' expects argument of type 'long unsigned int', but argument 7 has type 'uint64_t {aka long long unsigned int}' [-Werror=format=]
Erik Skultety [Tue, 3 May 2016 15:04:42 +0000 (17:04 +0200)]
admin: Add a check to reject negative argument for number of typed params
Since nparams can be technically negative, it is a good practice throughout
our code to check if nparams actually has a non-negative value. The same effect
would be achieved by converting our internal typed params serializer argument
to 'unsigned' type, but it definitely would not be the path of least resistance.
Signed-off-by: Erik Skultety <eskultet@redhat.com>
Now that we have all the pieces in place, we can add the 'iothread=#' to
the command line for the (two) controllers that support it (virtio-scsi-pci
and virtio-scsi-ccw). Add the tests as well...
What was not realized is that the rule 'libvirt-docs/*'
and ' docs/*.html docs/html docs/*.gif' actually point
to the exact same content. ie, we had previously included
the website HTML in *both* the -docs and -devel packages.
So this change ended up listing the files twice, which
caused RPM to print a load of warnings:
Erik Skultety [Wed, 4 May 2016 07:30:58 +0000 (09:30 +0200)]
daemon: Add VIR_ERR_NO_SERVER and VIR_ERR_NO_CLIENT to daemonErrorLogFilter
Commits 52a2eef9 and 62be5486 forgot to add these errors to daemon's error
whitelist, i.e. in order to avoid log file pollution with errors like "Domain
not found" or "Server not found" in this case, since these events are valid
and expected to occur.
Signed-off-by: Erik Skultety <eskultet@redhat.com>
John Ferlan [Wed, 20 Apr 2016 18:43:04 +0000 (14:43 -0400)]
qemu: Add capability for virtio-scsi iothreads
An iothread for virtio-scsi is a property of the controller. Add a lookup
of the 'virtio-scsi-pci' and 'virtio-scsi-ccw' device properties and parse
the output. For both, support for the iothread was added in qemu 2.4
while support for virtio-scsi in general was added in qemu 1.4.
Modify the various mock capabilities replies (by hand) to reflect the
when virtio-scsi was supported and then specifically when the iothread
property was added. For versions prior to 1.4, use the no device error
return for virtio-scsi. For versions 1.4 to before 2.4, add some data
for virtio-scsi-pci even though it isn't complete we're not looking for
anything specific there anyway. For 2.4 to 2.6, add a more complete reply.
configure: Remove nested conditionals in LIBVIRT_CHECK_INIT_SCRIPT
We don't need them any longer; moreover, the previous structure
made it very easy for bugs to slip in, by having the result of one
check influence the following one.
By placing the check for "$with_init_script" = check front and
center, hopefully this won't happen (as easily) again.
Erik Skultety [Fri, 22 Apr 2016 11:05:42 +0000 (13:05 +0200)]
admin: Introduce virAdmClientGetInfo API
Expose a public API to retrieve some identity and connection information about
a client connected to the specified server on daemon. The identity info
retrieved is mostly connection transport dependent, i.e. there won't be any
socket address returned for a local (UNIX socket) connection, while on the
other hand, when connected through TLS or unencrypted TCP, obviously no UNIX
process identification will be present in the returned data. All supported
values that can be returned in typed params are exposed and documented in
include/libvirt/libvirt-admin.h
Signed-off-by: Erik Skultety <eskultet@redhat.com>
Erik Skultety [Fri, 22 Apr 2016 11:08:10 +0000 (13:08 +0200)]
virnetserverclient: Add an internal method to retrieve client's identity
This method just aggregates various client object attributes, like socket
address, connection type (RO/RW), and some TCP/TLS/UNIX identity in an atomic
manner.
Signed-off-by: Erik Skultety <eskultet@redhat.com>
We do have a similar method, serving the same purpose, for TLS, but we lack
one for SASL. So introduce one, in order for other modules to be able to find
out, if a SASL session is active, or better said, that a SASL session exists
at all.
Signed-off-by: Erik Skultety <eskultet@redhat.com>
Erik Skultety [Tue, 26 Apr 2016 11:39:06 +0000 (13:39 +0200)]
virnetsocket: Provide socket address format in a more standard form
Our socket address format is in a rather non-standard format and that is
because sasl library requires the IP address and service to be delimited by a
semicolon. The string form is a completely internal matter, however once the
admin interfaces to retrieve client identity information are merged, we should
return the socket address string in a common format, e.g. format defined by
URI rfc-3986, i.e. the IP address and service are delimited by a colon and
in case of an IPv6 address, square brackets are added:
Examples:
127.0.0.1:1234
[::1]:1234
This patch changes our default format to the one described above, while adding
separate methods to request the non-standard SASL format using semicolon as a
delimiter.
Signed-off-by: Erik Skultety <eskultet@redhat.com>
Erik Skultety [Fri, 22 Apr 2016 09:00:50 +0000 (11:00 +0200)]
admin: include: Introduce some client's identity related typed params macros
This patch could easily be squashed with the virAdmClientGetInfo method
introduced later one, but the idea was to split the logic to as many preferably
independent patches as possible. As the subject hints, this patch defines
some public typed params-related macros used within virAdmClientGetInfo
method. The thing is, there's one identity attribute missing in the set --
SELinux context, which libvirt internally supports in virIdentity, but it
doesn't seem to do much (or anything at all), so there's some room to extend
the set in the future.
Signed-off-by: Erik Skultety <eskultet@redhat.com>
Erik Skultety [Fri, 22 Apr 2016 07:47:09 +0000 (09:47 +0200)]
admin: Introduce virAdmServerLookupClient
Just like with server-related APIs, before any of client-based APIs can be
called, a reference to a client-side client object needs to be obtained. For
this purpose, a lookup method should exist. Apart from the client retrieval
logic, a new error code for non-existent client had to be added as well.
Signed-off-by: Erik Skultety <eskultet@redhat.com>
In majority of our functions we have this variable @ret that is
overwritten a lot. In other areas of the code we use 'goto
cleanup;' just so that this wouldn't happen. But here.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
In these functions I'm fixing here, we do call
qemuMonitorJSONCheckError() followed by another check if qemu
reply contains 'return' object. If it wouldn't, the former
CheckError() function would error out and the flow would not even
get to the latter.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Ján Tomko [Tue, 3 May 2016 10:29:26 +0000 (12:29 +0200)]
Remove useless os.machine NULL check
In qemuDomainDefAddDefaultDevices we check for a non-NULL
def->os.machine for x86 archs, but not the others.
Moreover, the only caller - qemuDomainDefPostParse
already checks for it and even then it can happen only
if /etc/libvirt contains an XML without a machine type.
Thing is, for some reasons you can have your domain's RTC to be
in something different than UTC. More weirdly, it's not only time
zone what you can shift it of, but an arbitrary value. So, if
domain is configured that way, libvirt will correctly put it onto
qemu cmd line and moreover track it as this offset changes during
domain's life time (e.g. because guest OS decides the best thing
to do is set new time to RTC). Anyway, they way in which this
tracking is implemented is events. But we've got a problem if
change in guest's RTC occurs and the daemon is not running. The
event is lost and we end up reporting invalid value in domain
XML. Therefore, when the daemon is starting up again and it is
reconnecting to all running domains, re-fetch their RTC so the
correct offset value can be computed.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Erik Skultety [Sat, 16 Apr 2016 07:13:21 +0000 (09:13 +0200)]
virt-admin: Introduce srv-clients-list command
Wire-up the public client listing API. Along with this change, a private time
simple conversion method to interpret client's timestamp obtained from server
has been added as well. Format used to for time output is as follows:
YYYY-mm-DD HH:MM:SS+ZZZZ.
Although libvirt exposes methods time-related methods through virtime.h
internally, it utilizes millisecond precision which we don't need in this case,
especially when connection timestamps use precision to seconds only.
This is just a convenience int to string conversion method.
To reflect the new API, man page has been adjusted accordingly.
Signed-off-by: Erik Skultety <eskultet@redhat.com>
Erik Skultety [Thu, 14 Apr 2016 22:25:04 +0000 (00:25 +0200)]
rpc: virnetserverclient: Implement client connection transport retrieval
Although we document 6 types of transport that we support, internally we can
only differentiate between TCP, TLS, and UNIX transports only, since both SSH
and libssh2 transports, due to using netcat, behave in the exactly the same
way as a UNIX socket.
Signed-off-by: Erik Skultety <eskultet@redhat.com>
Erik Skultety [Thu, 14 Apr 2016 22:21:05 +0000 (00:21 +0200)]
rpc: virnetserver: Support retrieval of a list of clients
For now, the list copy is done simply by locking the whole server, walking the
original and increasing the refcount on each object. We may want to change
the list to a lockable object (like list of domains) later in the future if
we discover some performance issues related to locking the whole server in
order to walk the whole list of clients, possibly issuing some 'ForEach'
callback.
Signed-off-by: Erik Skultety <eskultet@redhat.com>
Erik Skultety [Thu, 14 Apr 2016 22:22:59 +0000 (00:22 +0200)]
rpc: gendispatch: Tune it to support client structure
Now that libvirt-admin supports another client-side object and provided that
we want to generate as many both client-side and server-side RPC dispatchers,
support for this needs to be added to gendispatch.
Signed-off-by: Erik Skultety <eskultet@redhat.com>
Erik Skultety [Wed, 13 Apr 2016 08:35:26 +0000 (10:35 +0200)]
admin: Introduce virAdmClient client-side object
Besides ID, the object also stores static data like connection transport and
connection timestamp, since once obtained a list of all clients connected to a
server, from user's perspective, it would be nice to know whether a given
client is remote or local only and when did it connect to the daemon.
Along with the object introduction, all necessary client-side methods necessary
to work with the object are added as well.
Signed-off-by: Erik Skultety <eskultet@redhat.com>
Erik Skultety [Tue, 12 Apr 2016 17:21:43 +0000 (19:21 +0200)]
rpc: virnetserverclient: Introduce new attribute conn_time to client
Besides ID, libvirt should provide several parameters to help the user
distinguish two clients from each other. One of them is the connection
timestamp. This patch also adds a testcase for proper JSON formatting of the
new attribute too (proper formatting of older clients that did not support
this attribute yet is included in the existing tests) - in order to
testGenerateJSON to work, a mock of time_t time(time_t *timer) needed to be
created.
Signed-off-by: Erik Skultety <eskultet@redhat.com>
Erik Skultety [Wed, 13 Apr 2016 18:54:40 +0000 (20:54 +0200)]
rpc: virnetserverclient: Identify clients by an integer ID
Admin API needs a way of addressing specific clients. Unlike servers, which we
are happy to address by names both because its name reflects its purpose (to
some extent) and we only have two of them (so far), naming clients doesn't make
any sense, since a) each client is an anonymous, i.e. not recognized after a
disconnect followed by a reconnect, b) we can't predict what kind of requests
it's going to send to daemon, and c) the are loads of them comming and going,
so the only viable option is to use an ID which is of a reasonably wide data
type.
Signed-off-by: Erik Skultety <eskultet@redhat.com>
Andrea Bolognani [Thu, 28 Apr 2016 15:19:29 +0000 (17:19 +0200)]
configure: Add systemd detection to --with-init-script=check
Most distributions, including RHEL, have switched to systemd,
so we should detect it and act accordingly. This also means
that 'systemd+redhat' should be preferred to legacy 'redhat'.
Our witness for the check is the availability of the systemctl
command on the host.
Boris Fiuczynski [Fri, 29 Apr 2016 13:23:41 +0000 (15:23 +0200)]
qemu: add panic device support for S390
If a panic device is being defined without a model in a domain
the default value is always overwritten with model ISA. An ISA
bus does not exist on S390 and therefore specifying a panic device
results in an unsupported configuration.
Since the S390 architecture inherently provides a crash detection
capability the panic device should be defined in the domain xml.
This patch adds an s390 panic device model and prevents setting a
device address on it.
Signed-off-by: Boris Fiuczynski <fiuczy@linux.vnet.ibm.com> Reviewed-by: Cornelia Huck <cornelia.huck@de.ibm.com>
projects / libvirt.git / log