This solution does not keep snapshots cache because vz sdk lacks good support
for snapshot related events.

Libvirt and vz sdk has different approach to snapshot ids. vz sdk always
auto generate them while libvirt has ability to specify id by user.
Thus I have no other choice rather than simply ignore ids set by user
or generated by libvirt.
Signed-off-by: NNikolay Shirokovskiy <nshirokovskiy@virtuozzo.com>
Signed-off-by: NMaxim Nestratov <mnestratov@virtuozzo.com>

This reverts commit 071fe092.
It was committed by a mistake and correct patch was committed
earlier as baad90fb.

Signed-off-by: NNikolay Shirokovskiy <nshirokovskiy@virtuozzo.com>
Signed-off-by: NMaxim Nestratov <mnestratov@virtuozzo.com>

First we don't need to add disk in this case. Second flag should
be skipped.
Signed-off-by: NNikolay Shirokovskiy <nshirokovskiy@virtuozzo.com>

Signed-off-by: NNikolay Shirokovskiy <nshirokovskiy@virtuozzo.com>

SDK handles empty cdroms all right. We just need to
pass "" instead of NULL (not setting is good too).

However we can get problems here. Disk detaching treats source
as ids. Fortunately disk detaching is not supported for cdroms
yet and for hard disks we can not get empty source - this is prohibitited
by xml parsing code.
Signed-off-by: NNikolay Shirokovskiy <nshirokovskiy@virtuozzo.com>
Signed-off-by: NMaxim Nestratov <mnestratov@virtuozzo.com>

Signed-off-by: NYuri Chornoivan <yurchor@ukr.net>

If we want to delete all disks for container or vm
we should make a loop from 0 to NumberOfDisks and always
use zero index in PrlVmCfg_GetHardDisk to get disk handle.
When we delete first disk after that numbers of other disks
will be changed, start from 0 to NumberOfDisks-1.
That's why we should always use zero index.

We don't have input devices in SDK thus for define/dumpxml
operations to be consistent we need to:

1. on dumpxml: infer input devices from other parts of config.
It is already done in prlsdkLoadDomain.

2. on define: check that input devices are the same that
will be infer back on dumpxml operation.

The second part should be fixed.
Signed-off-by: NNikolay Shirokovskiy <nshirokovskiy@virtuozzo.com>

libvirt handles empty source as NULL, while vz sdk as
"" thus we need a bit of conversion.
Signed-off-by: NNikolay Shirokovskiy <nshirokovskiy@virtuozzo.com>

Current implementation does not detect all incompatible configurations.
For example if we have in vzsdk bootorder "cdrom1, cdrom0" (that is
"hdb, hda" in case of ide cdroms) and cdroms do not have disk
images inserted. In this case boot order check code fails to
distiguish them at all as for both PrlVmDev_GetFriendlyName gives "".
Well the consequences are only missing warnings but as
we just have introduced all the necessary tools to face the problem -
let's fix it.
Signed-off-by: NNikolay Shirokovskiy <nshirokovskiy@virtuozzo.com>

Actually using disk PrlVmDev_GetFriendlyName as id on
detaching volumes is not a problem. We can only detach
hard disks and these can not have empty friendly names.
But upcoming update device functionality for cdroms
can not use disk source as id at all as update operation
typically change this same source value. Thus we will need
to use cdrom bus and cdrom target name as cdrom id. So in attempt
to use same id scheme for all purpuses lets fix hard disk
detach function to use new id.
Signed-off-by: NNikolay Shirokovskiy <nshirokovskiy@virtuozzo.com>

Our intention is to use disk bus and disk target name pair
as disk id instead of name returned by PrlVmDev_GetFriendlyName.
We already have the code that extracts this pair from vzsdk
data. Let's factor it out into a function.
Signed-off-by: NNikolay Shirokovskiy <nshirokovskiy@virtuozzo.com>

SDK does not allocate memory when getting strings thus we
need to call every function that returns string twice.
First to obtain string length, second to obtain string
itself. It is tedious so let's create helper functions
for cases when we know length of the result beforehand
and we are not.
Signed-off-by: NNikolay Shirokovskiy <nshirokovskiy@virtuozzo.com>

Lock driver when a new domain is created in prlsdkNewDomainByHandle
and try to find it in the list under lock again because it can race
with vzDomainDefineXMLFlags when a domain with the same uuid is added
via vz dispatcher directly and libvirt define.
Signed-off-by: NMaxim Nestratov <mnestratov@virtuozzo.com>

This patch introduces a new 'vzDriver' lockable object and provides
helper functions to allocate/destroy it and we pass it to prlsdkXxx
functions instead of virConnectPtr.
Now we store domain related objects such as domain list, capabitilies
etc. within a single vz_driver vzDriver structure, which is shared by
all driver connections. It is allocated during daemon initialization or
in a lazy manner when a new connection to 'vz' driver is established.
When a connection to vz daemon drops, vzDestroyConnection is called,
which in turn relays disconnect event to all connection to 'vz' driver.
Signed-off-by: NMaxim Nestratov <mnestratov@virtuozzo.com>

GCC in RHEL-6 complains about listen:

../../src/conf/domain_conf.c:23718: error: declaration of 'listen' shadows a global declaration [-Wshadow]
/usr/include/sys/socket.h:204: error: shadowed declaration is here [-Wshadow]

This renames all the listen to gListen.
Signed-off-by: NPavel Hrdina <phrdina@redhat.com>

There is no point the use two different getters on the same listen
structure few lines apart.
Signed-off-by: NPavel Hrdina <phrdina@redhat.com>

we don't need to allocate macstr at all as it is an array
and already has the the space it needs.
Signed-off-by: NNikolay Shirokovskiy <nshirokovskiy@virtuozzo.com>

Quite straigthforward as vz sdk memory setting function makes
just what we want to that is set "amount of physical memory
allocated to a domain".

'useflags' is introduced for non flag function implementation.
We can't just use combination of flags like "live | config" or
we fail for inactive domains. Other combinations have drawbacks
too.
Signed-off-by: NNikolay Shirokovskiy <nshirokovskiy@virtuozzo.com>

As usual we try to deal correctly with vz domains that were
created by other means and thus can have all range of SDK domain
parameters. If vz domain boot order can't be represented
in libvirt os boot section let's give warning and make os boot section
represent SDK to some extent.

1. Os boot section supports up to 4 boot devices. Here we just
cut SDK boot order up to this limit. Not too bad.

2. If there is a floppy in boot order let's just skip it.
Anyway we don't show it in the xml. Not too bad too.

3. SDK boot order with unsupported disks order. Say we have "hdb, hda" in
SDK. We can not present this thru os boot order. Well let's just
give warning but leave double <boot dev='hd'/> in xml. It's
kind of misleading but we warn you!

SDK boot order have an extra parameters 'inUse' and 'sequenceIndex'
which makes our task more complicated. In realitly however 'inUse'
is always on and 'sequenceIndex' is not less than 'boot position index'
which simplifies out task back again! To be on a safe side let's explicitly
check for this conditions!

We have another exercise here. We want to check for unrepresentable
condition 3 (see above). The tricky part is that in contrast to
domains defined thru this driver 3-rd party defined domains can
have device ordering different from default. Thus we need
some id to check that N-th boot disk of os boot section is same as
N-th boot disk of SDK boot. This is what prlsdkBootOrderCheck
for. It uses disks sources paths as id for disks and iface names
for network devices.
Signed-off-by: NNikolay Shirokovskiy <nshirokovskiy@virtuozzo.com>
Signed-off-by: NMaxim Nestratov <mnestratov@virtuozzo.com>

We want to report boot order in dumpxml for vz domains.
Thus we want disks devices to be sorted in output compatible with boot
ordering specification. So let's just use virDomainDiskInsert
which makes appropriate sorting.
Signed-off-by: NNikolay Shirokovskiy <nshirokovskiy@virtuozzo.com>

The patch makes some refactoring of the existing code. Current boot order spec code
makes very simple thing in somewhat obscure way. In case of VMs
it sets the first hdd as the only bootable device. In case of CTs it
doesn't touch the boot order at all if one of the filesystems is mounted to root.
Otherwise like in case of VMs it sets the first hdd as the only bootable
device and additionally sets this device mount point to root. Refactored
code makes all this explicit.

The actual boot order support is simple. Common libvirt domain xml parsing
code makes the exact ordering of disks devices as described in docs
for boot ordering (disks are sorted by bus order first, device target
second. Bus order is the order of disk buses appearence in original
xml. Device targets order is alphabetical). We add devices in the
same order and SDK designates device indexes sequentially for each
device type. Thus device index is equal to its boot index. For
example N-th cdrom in boot specification refers to sdk cdrom with
it's device index N.

If there is no boot spec in xml the parsing code will add <boot dev='hdd'>
for HVMs automatically and we backward compatibly set fist hdd as
bootable.
Signed-off-by: NNikolay Shirokovskiy <nshirokovskiy@virtuozzo.com>
Signed-off-by: NMaxim Nestratov <mnestratov@virtuozzo.com>

It's just a combination of AddImplicitControllers, and AddConsoleCompat.
Every caller that wants ImplicitControllers also wants the ConsoleCompat
AFAICT, so lump them together. We also need it for future patches.

In prlsdkAddNet() Attach/DetachNet() functions
privconn should be the first argument

Each version of virtuozzo supports only one type of SCSI controller
So if we add disk on SCSI bus, we should set SCSI controller model.
We can take it from vzCapabilities structure.

We should report correct disk format depending on vz version and domain type.
Since we support only one disk format for each domain type, we can take it
from vzCapabilities structure.

As long as we have another function checking disk parameters correctness,
let's have them in one place. Here we change prefix of the moved function and
start to call it from vzCheckUnsupportedDisks rather than add disk.
Signed-off-by: NMaxim Nestratov <mnestratov@virtuozzo.com>

Now we check disk parameters correctness in DomainPostParse.
Signed-off-by: NMaxim Nestratov <mnestratov@virtuozzo.com>

Since commit 9c14a9ab we have broken active domain listing
because reworked prlsdkLoadDomain doesn't set dom->def->id
propely. It just looses it when a new def structure is set.
Now we make prlsdkConvertDomainState function return void
and move calling it after an old dom->def is replaces with
a new one within prlsdkLoadDomain function.
Signed-off-by: NMaxim Nestratov <mnestratov@virtuozzo.com>

Signed-off-by: NNikolay Shirokovskiy <nshirokovskiy@virtuozzo.com>

Race condition:
User calls defineXML to create new instance.
The main thread from vzDomainDefineXMLFlags() creates new instance by prlsdkCreateVm.
Then this thread calls prlsdkAddDomain to add new domain to domains list.
The second thread receives notification from hypervisor that new VM was created.
It calls prlsdkHandleVmAddedEvent() and also tries to add new domain to domains list.
These two threads call virDomainObjListFindByUUID() from prlsdkAddDomain() and don't find new domain.
So they add two domains with the same uuid to domains list.

This fix splits logic of prlsdkAddDomain() into two functions.
1. vzNewDomain() creates new empty domain in domains list with the specific uuid.
2. prlsdkLoadDomain() add data from VM to domain object.

New algorithm for creating an instance:
In vzDomainDefineXMLFlags() we add new domain to domain list by calling vzNewDomain()
and only after that we call CreateVm() to create VM.
It means that we "reserve" domain object with the specific uuid.
After creation of new VM we add info from this VM
to reserved domain object by calling prlsdkLoadDomain().

Before this patch prlsdkLoadDomain() worked in 2 different cases:
1. It creates and initializes new domain. Then updates it from sdk handle.
2. It updates existed domain from sdk handle.
In this patch we remove code which creates new domain from LoadDomain()
and move it to vzNewDomain().
Now prlsdkLoadDomain() only updates domain from skd handle.

In notification handler prlsdkHandleVmAddedEvent() we check
the existence of a domain and if it doesn't exist we add new domain by calling
vzNewDomain() and load info from sdk handle via prlsdkLoadDomain().

Bug cause:
Update the domain that is subscribed to hypervisor notification.
LoadDomain() rewrites notifications fields in vzDomObj structure and makes domain as "unsubscribed".
Fix:
Initialize notification fields in vzDomObj only if we create a new domain.
And do not reinitialize these fields if we update domain (by calling LoadDomain with olddom argument)

In commit 7039bb3c we have removed code that saves uuid to vzDomObj.uuid
So this field is no longer needed.

prlsdkGetDomainIds() returns name and uuid for specified instance.
Now output arguments can be NULL.
It allows to get only necessary info(name or uuid).

Now with the new struct the data can be stored in a much saner place.

This step can be omitted, so that drivers can decide what to do when the
user requests to use default vcpu pinning.

The array doesn't necessarily have the same cardinality as the count of
vCPUs for a domain. Iterating it can cause access beyond the end of the
array.

Implement VIR_DOMAIN_UNDEFINE_MANAGED_SAVE and
VIR_DOMAIN_UNDEFINE_SNAPSHOTS_METADATA flags support.
Signed-off-by: NMaxim Nestratov <mnestratov@virtuozzo.com>