qemuAgentGetInterfaceAddresses: turn ifname into char*

We only care about the first part of the 'ifname' string,
splitting it is overkill.

Instead, just replace the ':' with a '\0' in a copy of the string.
This reduces the count of the varaibles containing some form
of the interface name to two.

Signed-off-by: Ján Tomko <jtomko@redhat.com>
Reviewed-by: Jonathon Jongsma <jjongsma@redhat.com>
Reviewed-by: Neal Gompa <ngompa13@gmail.com>

qemu: agent: remove cleanup in qemuAgentGetInterfaces

Signed-off-by: Ján Tomko <jtomko@redhat.com>
Reviewed-by: Jonathon Jongsma <jjongsma@redhat.com>
Reviewed-by: Neal Gompa <ngompa13@gmail.com>

qemu: agent: use g_auto in qemuAgentGetInterfaces

Signed-off-by: Ján Tomko <jtomko@redhat.com>
Reviewed-by: Jonathon Jongsma <jjongsma@redhat.com>
Reviewed-by: Neal Gompa <ngompa13@gmail.com>

qemu: agent: use GetArray to remove a check

The error check for ValueObjectGet("return") is redundant,
qemuAgentCommand already checked for us that the reply contains
a "return" object.

It does not guarantee, that it is an array.

Use virJSONValueObjectGetArray that combines getting the object
with checking for its type and return the more helpful of
the two error messages.

Signed-off-by: Ján Tomko <jtomko@redhat.com>
Reviewed-by: Jonathon Jongsma <jjongsma@redhat.com>
Reviewed-by: Neal Gompa <ngompa13@gmail.com>

qemu: agent: split out qemuAgentGetInterfaceAddresses

Convert one interface from the "return" array returned by
"guest-network-get-interfaces" to virDomainInterface.

Due to the functionality of squashing interface aliases together,
this is not a pure function - it either:
* Adds the interface to ifaces_ret, incrementing ifaces_count
  and adds a pointer to it into the ifaces_store hash table.
* Adds the additional IP addresses from the interface alias
  to the existing interface entry, found through the hash table.
  This does not increment ifaces_count or extend the array.

Signed-off-by: Ján Tomko <jtomko@redhat.com>
Reviewed-by: Jonathon Jongsma <jjongsma@redhat.com>
Reviewed-by: Neal Gompa <ngompa13@gmail.com>

qemu: agent: simplify access to ifaces_ret

We have a local 'iface' variable that contains the same value
eventually. Initialize it early instead of indexing two more
times.

Signed-off-by: Ján Tomko <jtomko@redhat.com>
Reviewed-by: Jonathon Jongsma <jjongsma@redhat.com>
Reviewed-by: Neal Gompa <ngompa13@gmail.com>

qemu: agent: use virHashNew

We're passing 'ifaces_count' to virHashCreate as the initial
hash table size just after we've initialized it to zero.
This translates to a default of 256 inside virHashCreateFull.

Instead of this obfuscation, use virHashNew (default of 32),
to make it obvious we don't care about the initial hash size.

Also remove the error handling, since neither of the functions
return any errors after switching to g_new0.

Signed-off-by: Ján Tomko <jtomko@redhat.com>
Reviewed-by: Jonathon Jongsma <jjongsma@redhat.com>
Reviewed-by: Neal Gompa <ngompa13@gmail.com>

qemu: agent: use g_auto for ifname

This lets us conveniently reduce its scope to the outer loop.

Signed-off-by: Ján Tomko <jtomko@redhat.com>
Reviewed-by: Jonathon Jongsma <jjongsma@redhat.com>
Reviewed-by: Neal Gompa <ngompa13@gmail.com>

qemu: agent: expand addrs upfront

qemuAgentGetInterfaceOneAddress returns exactly one address
for every iteration of the loop (and we error out if not).

Instead of expanding the addrs by one on every iteration,
do it upfront since we know how many times the loop will
execute.

Signed-off-by: Ján Tomko <jtomko@redhat.com>
Reviewed-by: Jonathon Jongsma <jjongsma@redhat.com>
Reviewed-by: Neal Gompa <ngompa13@gmail.com>

qemu: agent: reduce scope of addrs_count

Signed-off-by: Ján Tomko <jtomko@redhat.com>
Reviewed-by: Jonathon Jongsma <jjongsma@redhat.com>
Reviewed-by: Neal Gompa <ngompa13@gmail.com>

qemu: agent: remove impossible errors

For both 'ip_addr_arr' and 'ret_array', we:
1) already checked that they are arrays
2) only iterate up to the array size

Remove the duplicate checks.

Signed-off-by: Ján Tomko <jtomko@redhat.com>
Reviewed-by: Jonathon Jongsma <jjongsma@redhat.com>
Reviewed-by: Neal Gompa <ngompa13@gmail.com>

qemu: agent: split out qemuAgentGetInterfaceOneAddress

A function that takes one entry from the "ip-addresses" array
returned by "guest-network-get-interfaces" and converts it
into virDomainIPAddress.

Signed-off-by: Ján Tomko <jtomko@redhat.com>
Reviewed-by: Jonathon Jongsma <jjongsma@redhat.com>
Reviewed-by: Neal Gompa <ngompa13@gmail.com>

qemu: agent: remove redundant checks

virJSONValueObjectGetArray returns NULL if the object with
the supplied key is not an array.

Calling virJSONValueIsArray right after is redundant.

Signed-off-by: Ján Tomko <jtomko@redhat.com>
Reviewed-by: Jonathon Jongsma <jjongsma@redhat.com>
Reviewed-by: Neal Gompa <ngompa13@gmail.com>

tools: spell out the type instead of using typeof

Signed-off-by: Ján Tomko <jtomko@redhat.com>
Reviewed-by: Daniel Henrique Barboza <danielhb413@gmail.com>

conf: virDomainUSBAddressHubNew: refactor

Remove the pointless variable and pointer stealing.

Signed-off-by: Ján Tomko <jtomko@redhat.com>
Reviewed-by: Neal Gompa <ngompa13@gmail.com>
Reviewed-by: Erik Skultety <eskultet@redhat.com>

conf: use g_new0

Signed-off-by: Ján Tomko <jtomko@redhat.com>
Reviewed-by: Neal Gompa <ngompa13@gmail.com>
Reviewed-by: Erik Skultety <eskultet@redhat.com>

conf: cpu: use g_new0

Signed-off-by: Ján Tomko <jtomko@redhat.com>
Reviewed-by: Neal Gompa <ngompa13@gmail.com>
Reviewed-by: Erik Skultety <eskultet@redhat.com>

conf: nwfilter_params: use g_new0

Signed-off-by: Ján Tomko <jtomko@redhat.com>
Reviewed-by: Neal Gompa <ngompa13@gmail.com>
Reviewed-by: Erik Skultety <eskultet@redhat.com>

conf: virsecretobj: use g_new0

Signed-off-by: Ján Tomko <jtomko@redhat.com>
Reviewed-by: Neal Gompa <ngompa13@gmail.com>
Reviewed-by: Erik Skultety <eskultet@redhat.com>

conf: storage: use g_new0

Signed-off-by: Ján Tomko <jtomko@redhat.com>
Reviewed-by: Neal Gompa <ngompa13@gmail.com>
Reviewed-by: Erik Skultety <eskultet@redhat.com>

conf: numa: use g_new0

Signed-off-by: Ján Tomko <jtomko@redhat.com>
Reviewed-by: Neal Gompa <ngompa13@gmail.com>
Reviewed-by: Erik Skultety <eskultet@redhat.com>

conf: interface: use g_new0

Signed-off-by: Ján Tomko <jtomko@redhat.com>
Reviewed-by: Neal Gompa <ngompa13@gmail.com>
Reviewed-by: Erik Skultety <eskultet@redhat.com>

conf: domain_addr: use g_new0

Signed-off-by: Ján Tomko <jtomko@redhat.com>
Reviewed-by: Neal Gompa <ngompa13@gmail.com>
Reviewed-by: Erik Skultety <eskultet@redhat.com>

conf: node_device: use g_new0

Signed-off-by: Ján Tomko <jtomko@redhat.com>
Reviewed-by: Neal Gompa <ngompa13@gmail.com>
Reviewed-by: Erik Skultety <eskultet@redhat.com>

conf: network: use g_new0

Signed-off-by: Ján Tomko <jtomko@redhat.com>
Reviewed-by: Neal Gompa <ngompa13@gmail.com>
Reviewed-by: Erik Skultety <eskultet@redhat.com>

conf: capabilities: use g_new0

Signed-off-by: Ján Tomko <jtomko@redhat.com>
Reviewed-by: Neal Gompa <ngompa13@gmail.com>
Reviewed-by: Erik Skultety <eskultet@redhat.com>

conf: domain: use g_new0

Signed-off-by: Ján Tomko <jtomko@redhat.com>
Reviewed-by: Neal Gompa <ngompa13@gmail.com>
Reviewed-by: Erik Skultety <eskultet@redhat.com>

schema: Make element "topology" in host CPU definition optional

This element is not always present, see e.g.
x86_64-cpuid-Xeon-X5460-host.xml, x86_64-cpuid-Pentium-P6100-host.xml,
or x86_64-cpuid-EPYC-7601-32-Core-ibpb-host.xml.

Signed-off-by: Tim Wiederhake <twiederh@redhat.com>
Reviewed-by: Jiri Denemark <jdenemar@redhat.com>

spec: Install all schemas

The main spec file was missing basictypes.rng and mingw did not install
cpu.rng. Let's just install all *.rng files in the schemas directory.

Signed-off-by: Jiri Denemark <jdenemar@redhat.com>
Reviewed-by: Peter Krempa <pkrempa@redhat.com>

cpu_map: Remove monitor feature from EPYC-Rome

The feature is filtered by KVM and never automatically enabled. So even
though QEMU definition of EPYC-Rome contains this feature, the guest
won't see it. Also domain capabilities will show it as disabled for KVM
domains. Thus the feature should not really be included in our
definition of EPYC-Rome.

Signed-off-by: Jiri Denemark <jdenemar@redhat.com>
Reviewed-by: Peter Krempa <pkrempa@redhat.com>

spec: Add cpu.rng to %files

Fixes: 51v5d325240c645ea6c1a0902c695cf299410b1f90c
Signed-off-by: Cole Robinson <crobinso@redhat.com>

wireshark: Fix missing bool_t on macOS

The header has to be explicitly added to pull definition of bool_t and a
few other types. Otherwise packet-libvirt.c can't be compiled.

Signed-off-by: Roman Bolshakov <r.bolshakov@yadro.com>
Reviewed-by: Andrea Bolognani <abologna@redhat.com>

cputest: Update QEMU data for Ryzen 9 3900X

The CPU should be identified as EPYC-Rome, but the QEMU binary used to
gather the original test data did not support this model. Let's update
the supported models to QEMU 5.1.0.

Signed-off-by: Jiri Denemark <jdenemar@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>

cpu_map: Defined and enable EPYC-Rome model

Signed-off-by: Markus Schade <markus.schade@hetzner.com>
Reviewed-by: Jiri Denemark <jdenemar@redhat.com>

Add testdata for AMD EPYC 7502

Signed-off-by: Markus Schade <markus.schade@hetzner.com>
Reviewed-by: Jiri Denemark <jdenemar@redhat.com>

tests: Rename some test files in cputestdata

The files contained the "-invalid" marker in their filename, marking
them as test cases that are supposed to fail in the virschematest.
Unfortunately, the "-invalid" marker does not discriminate between
different tests the files might be used in.

A later patch will introduce a new test validating the XML. This
test is not supposed to fail, as the files contain valid XML.

Signed-off-by: Tim Wiederhake <twiederh@redhat.com>
Reviewed-by: Peter Krempa <pkrempa@redhat.com>

virsh: Add "validate" argument to [hypervisor-]cpu-compare

Signed-off-by: Tim Wiederhake <twiederh@redhat.com>
Reviewed-by: Peter Krempa <pkrempa@redhat.com>

cpu: Wire in XML validation

This adds a new value to virConnectCompareCPUFlags,
"VIR_CONNECT_CPU_VALIDATE_XML", that governs XML document validation in
virCPUDefParseXML.

In src/conf/cpu_conf.c, include configmake.h for PKGDATADIR and
virfile.h for virFileFindResource.

Signed-off-by: Tim Wiederhake <twiederh@redhat.com>
Reviewed-by: Peter Krempa <pkrempa@redhat.com>

util: Allow validation for single XML node

Validation is usually performed on an entire document. If we are only
interested in validating a single nested node that can occur in
different contexts, this would require writing different schemas for
any of those different contexts.

By temporarily replacing the document's root node, we can validate the
relevant node only.

Signed-off-by: Tim Wiederhake <twiederh@redhat.com>
Reviewed-by: Peter Krempa <pkrempa@redhat.com>

schema: Add schema for guest or host cpu definition

`virsh cpu-compare` and `virsh hypervisor-cpu-compare` both accept
guest and host cpu definitions. This schema is able to validate both
possibilities.

Signed-off-by: Tim Wiederhake <twiederh@redhat.com>
Reviewed-by: Peter Krempa <pkrempa@redhat.com>

schema: Move guest cpu definition to cputypes.rng

Signed-off-by: Tim Wiederhake <twiederh@redhat.com>
Reviewed-by: Peter Krempa <pkrempa@redhat.com>

schema: Move host cpu definition to cputypes.rng

This also inlines the defintions for "cpufeature", "cpuspec",
"featureName" and "pagesHost", as "cpu" was the only user.

Doing so avoids a naming collision when cputypes.rng is included in
other schemas in a later patch.

Signed-off-by: Tim Wiederhake <twiederh@redhat.com>
Reviewed-by: Peter Krempa <pkrempa@redhat.com>

schema: Unify apostrophe and quotation mark usage

Quotation marks were used ~ 7000 times, apostrophes ~ 3000 times.

Signed-off-by: Tim Wiederhake <twiederh@redhat.com>
Reviewed-by: Peter Krempa <pkrempa@redhat.com>

examples: fix storage volume size schema

There is no such <storage> element, <capacity> and <allocation> exist at
the top level.

Reviewed-by: Peter Krempa <pkrempa@redhat.com>
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>

Fix incorrect g_new0 conversions

I left in a 'return' or 'goto cleanup' in a few places
where I did the conversion manually.

Signed-off-by: Ján Tomko <jtomko@redhat.com>
Reported-by: John Ferlan <jferlan@redhat.com>
Reviewed-by: John Ferlan <jferlan@redhat.com>

util: o-z: use g_new0

Signed-off-by: Ján Tomko <jtomko@redhat.com>
Reviewed-by: Erik Skultety <eskultet@redhat.com>

util: a-n: use g_new0

Signed-off-by: Ján Tomko <jtomko@redhat.com>
Reviewed-by: Erik Skultety <eskultet@redhat.com>

util: conf: use g_new0

Signed-off-by: Ján Tomko <jtomko@redhat.com>
Reviewed-by: Erik Skultety <eskultet@redhat.com>

util: firewall: use g_new0

Signed-off-by: Ján Tomko <jtomko@redhat.com>
Reviewed-by: Erik Skultety <eskultet@redhat.com>

util: systemd: use g_new0

Signed-off-by: Ján Tomko <jtomko@redhat.com>
Reviewed-by: Erik Skultety <eskultet@redhat.com>

util: netdev: use g_new0

Signed-off-by: Ján Tomko <jtomko@redhat.com>
Reviewed-by: Erik Skultety <eskultet@redhat.com>

util: command: use g_new0

Signed-off-by: Ján Tomko <jtomko@redhat.com>
Reviewed-by: Erik Skultety <eskultet@redhat.com>

util: sysinfo: use g_new0

Signed-off-by: Ján Tomko <jtomko@redhat.com>
Reviewed-by: Erik Skultety <eskultet@redhat.com>

util: storagefile: use g_new0

Signed-off-by: Ján Tomko <jtomko@redhat.com>
Reviewed-by: Erik Skultety <eskultet@redhat.com>

util: resctrl: use g_new0

Signed-off-by: Ján Tomko <jtomko@redhat.com>
Reviewed-by: Erik Skultety <eskultet@redhat.com>

util: split out VIR_ALLOC calls

To make the following commits simpler.

Signed-off-by: Ján Tomko <jtomko@redhat.com>
Reviewed-by: Erik Skultety <eskultet@redhat.com>

util: resctrl fix spacing in comment

Signed-off-by: Ján Tomko <jtomko@redhat.com>
Reviewed-by: Erik Skultety <eskultet@redhat.com>

mailmap: consolidate my email addresses

Signed-off-by: Matt Coleman <matt@datto.com>
Reviewed-by: Neal Gompa <ngompa13@gmail.com>
Reviewed-by: Andrea Bolognani <abologna@redhat.com>

virsh: network-port: remove pointless comment

We do not have a legacy API for listing network ports
so there's nothing to fall back on.

Signed-off-by: Ján Tomko <jtomko@redhat.com>
Reviewed-by: Erik Skultety <eskultet@redhat.com>

virsh: delete vsh[CM]alloc

Signed-off-by: Ján Tomko <jtomko@redhat.com>
Reviewed-by: Erik Skultety <eskultet@redhat.com>

virsh: use g_new0 instead of vsh[CM]alloc

Signed-off-by: Ján Tomko <jtomko@redhat.com>
Reviewed-by: Erik Skultety <eskultet@redhat.com>

virsh: do not add bools into size calculations

Switch the allocation in virshSnapshotListCollect and
its cargo-culted Checkpoint counterpart to two separate
g_new0 calls and move the boolean expression to
the if condition that chooses between them.

Signed-off-by: Ján Tomko <jtomko@redhat.com>
Reviewed-by: Erik Skultety <eskultet@redhat.com>

rpm: Enable Xen support on AArch64

Starting with Linux 5.9, Xen Dom0 works on commonly available
AArch64 devices, such as the Raspberry Pi 4.

Reference: https://xenproject.org/2020/09/29/xen-on-raspberry-pi-4-adventures/

Signed-off-by: Neal Gompa <ngompa13@gmail.com>
Reviewed-by: Andrea Bolognani <abologna@redhat.com>

tests: cover disk, interface <backenddomain>

There is present no XML test coverage for this.
Add genericxml parse + formatting coverage.

Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
Signed-off-by: Cole Robinson <crobinso@redhat.com>

cpu: Modify virCPUarmCompare to perform compare actions

Modify virCPUarmCompare in cpu_arm.c to perform compare action.
This patch only adds host to host CPU compare, the rest cases
remains the same. This is useful for source and destination host
compare during migrations to avoid migration between different
CPU models that have different CPU freatures.

Signed-off-by: Zhenyu Zheng <zheng.zhenyu@outlook.com>
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Daniel Henrique Barboza <danielhb413@gmail.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>

rpm: include aarch64 & riscv64 in systemtap 64-bit arch tapset rename

Reviewed-by: Andrea Bolognani <abologna@redhat.com>
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>

qemu: use g_new0

Signed-off-by: Ján Tomko <jtomko@redhat.com>
Reviewed-by: Erik Skultety <eskultet@redhat.com>

qemu: firmware: use g_new0

Signed-off-by: Ján Tomko <jtomko@redhat.com>
Reviewed-by: Erik Skultety <eskultet@redhat.com>

qemu: command: use g_new0

Signed-off-by: Ján Tomko <jtomko@redhat.com>
Reviewed-by: Erik Skultety <eskultet@redhat.com>

qemu: process: use g_new0

Signed-off-by: Ján Tomko <jtomko@redhat.com>
Reviewed-by: Erik Skultety <eskultet@redhat.com>

qemu: monitor: json: use g_new0

Signed-off-by: Ján Tomko <jtomko@redhat.com>
Reviewed-by: Erik Skultety <eskultet@redhat.com>

qemu: driver: use g_new0

Signed-off-by: Ján Tomko <jtomko@redhat.com>
Reviewed-by: Erik Skultety <eskultet@redhat.com>

qemu: domain: use g_new0

Signed-off-by: Ján Tomko <jtomko@redhat.com>
Reviewed-by: Erik Skultety <eskultet@redhat.com>

qemu: capabilities: use g_new0

Signed-off-by: Ján Tomko <jtomko@redhat.com>
Reviewed-by: Erik Skultety <eskultet@redhat.com>

qemu: separate out VIR_ALLOC calls

Move them to separate conditions to reduce churn
in following patches.

Signed-off-by: Ján Tomko <jtomko@redhat.com>
Reviewed-by: Erik Skultety <eskultet@redhat.com>

esx: switch esxUtil_ResolveHostname to return a new string

Change the interface of esxUtil_ResolveHostname() to return a newly
allocated string with the result address, instead of forcing the callers
to provide a buffer and its size. This way we can simply (auto)free the
string, and make the function stacks smaller.

Signed-off-by: Pino Toscano <ptoscano@redhat.com>
Reviewed-by: Laine Stump <laine@redhat.com>

esx: call freeaddrinfo earlier in esxUtil_ResolveHostname

Call freeaddrinfo() as soon as @result is not needed anymore, i.e. right
after getnameinfo(); this avoids calling freeaddrinfo() in two branches.

Signed-off-by: Pino Toscano <ptoscano@redhat.com>
Reviewed-by: Laine Stump <laine@redhat.com>

qemu: Taint cpu host-passthrough only after migration

From a discussion last year[1], Dan recommended libvirt drop the tain
flag for cpu host-passthrough, unless the VM has been migrated.

This repurposes the existing host-cpu taint flag to do just that.

[1]: https://www.redhat.com/archives/virt-tools-list/2019-February/msg00041.html

https://bugzilla.redhat.com/show_bug.cgi?id=1673098

Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Signed-off-by: Cole Robinson <crobinso@redhat.com>

docs: formatdomain: add spicevmc <redirdev> example

spicevmc is the most common <redirdev> usage. This adds an XML example
for it.

Reviewed-by: Ján Tomko <jtomko@redhat.com>
Signed-off-by: Cole Robinson <crobinso@redhat.com>

docs: formatdomain: fix net downscript 'since'

It was added in 6.4.0, not 5.1.0

Fixes: 61ba6f09b181ec8c1591dc1df8c3cd8e093ca67b
Reviewed-by: Ján Tomko <jtomko@redhat.com>
Signed-off-by: Cole Robinson <crobinso@redhat.com>

docs: formatdomain: fix incorrect 'Vsock' heading indent

Currently it is visually at the same indent as <seclabel>. This
fixes it to be grouped it with <devices>

Fixes: d4abb7b45d48aba77c2464dbea95f02a43a60834
Reviewed-by: Ján Tomko <jtomko@redhat.com>
Signed-off-by: Cole Robinson <crobinso@redhat.com>

docs: formatdomain: remove doubled filesystem <driver>

libvirt doesn't reject this but only one <driver> element takes
effect.

Drop the instance that is already referenced in the previous example

Reviewed-by: Ján Tomko <jtomko@redhat.com>
Signed-off-by: Cole Robinson <crobinso@redhat.com>

qemu: migration_cookie: s/VIR_FREE/g_free/

Previous refactors allow us to plainly replace all VIR_FREE by g_free to
finish the modernization of the file.

Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>

qemuMigrationCookieXMLParse: Avoid VIR_FREE when parsing lockstate

Restructure the control-flow a bit using an temporary variable to avoid
the need to use VIR_FREE.

Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>

qemuMigrationCookieAddCaps: Use 'g_new0' instead of VIR_ALLOC

Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>

qemuMigrationCookieCapsXMLParse: Refactor memory handling

Use modern allocators, automatic memory feeing, and decrease the scope
of some variables to remove the 'cleanup' label.

Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>

qemuMigrationCookieNBDXMLParse: Refactor memory handling

Use modern allocators, automatic memory feeing, and decrease the scope
of some variables to remove the 'error' and 'cleanup' labels.

Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>

qemuMigrationCookieNetworkXMLParse: Refactor memory handling

Use modern allocators, automatic memory feeing, and decrease the scope
of some variables to remove the 'error' label.

Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>

qemuMigrationCookieXMLFormat: Refactor memory handling

Use automatic memory freeing to get rid of the 'error' label. Since the
'tmp' variable was used only in one instance, rename it appropriately.

Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>

qemuMigrationCookieNetworkAlloc: Refactor memory handling

Use modern memory handling approach to simplify the code.

Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>

qemuMigrationCookieGraphicsSpiceAlloc: Refactor memory handling

Use modern memory handling approach to simplify the code.

Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>

qemu: migration_cookie: Register 'autoptr' functions for internal types

Register the the cleanup functions for 'qemuMigrationCookieGraphics',
'qemuMigrationCookieNetwork', 'qemuMigrationCookieNBD', and
'qemuMigrationCookieCaps'.

Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>

qemuDomainExtractTLSSubject: Refactor memory handling

Switch to automatic memory cleaning, use g_new0 for allocation and get
rid of the 'error' label.

Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>

qemuMigrationCookieXMLFormat: Extract formatting of NBD

Move the code into 'qemuMigrationCookieNBDXMLFormat' and use modern XML
formatting code patterns.

Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>

qemuMigrationCookieNetworkXMLFormat: Refactor XML formatting

Use 'virXMLFormatElement' both for formating the whole <network> element
but also for formatting the <interface> subelements. This alows to
remove the crazy logic which was determining which element was already
formatted.

Additional simplification is achieved by switching to skipping the loop
using 'continue' rather than putting everything in a giant block.

Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>

qemuMigrationCookieGraphicsXMLFormat: Use 'virXMLFormatElement'

Switch to the two buffer approach to simplify the logic for terminating
the element.

Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>

qemuMigrationCookieXMLParse: Remove 'error' label

Now it only returns -1 so we can do that directly.

Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>

qemuMigrationCookieXMLParse: Remove comment mentioning that error was already set

Most of our functions report errors so there's no need to mention it
here again.

Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>

qemuMigrationCookieXMLParse: Decrease scope of 'nodes' and use automatic freeing

Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>

qemuMigrationCookieXMLParse: Check domain element count more defensively

Make sure that 'virXPathNodeSet' returns '1' as the only expected value
rather than relying on the fact that the previous check for the number
of elements ensures success of the subsequent call.

The error message no longer mentions the number of <domain> elements in
the cookie, but this is a very unlikely internal error anyways.

Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>