Michal Privoznik [Wed, 28 Aug 2019 09:41:49 +0000 (11:41 +0200)]
domain_conf: Relax SCSI addr used check
In domain_conf.c we have virDomainSCSIDriveAddressIsUsed()
function which returns true or false if given drive address is
already in use for given domain config or not. However, it also
takes a shortcut and returns true (meaning address in use) if the
unit number equals 7. This is because for some controllers this
is reserved address. The limitation comes mostly from vmware and
applies to lsilogic, buslogic, spapr-vscsi and vmpvscsi models.
On the other hand, we were not checking for the maximum unit
number (aka LUN number) which is also relevant and differs from
model to model.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com> Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Michal Privoznik [Tue, 10 Sep 2019 14:45:47 +0000 (16:45 +0200)]
domain_conf: Make virDomainDeviceFindSCSIController accept virDomainDeviceDriveAddress struct
So far, the virDomainDeviceFindSCSIController() takes
virDomainDeviceInfo structure which is an overkill. It assumes
that the passed structure is type of
VIR_DOMAIN_DEVICE_ADDRESS_TYPE_DRIVE which is not obvious.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com> Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Ján Tomko [Fri, 18 Oct 2019 21:12:19 +0000 (23:12 +0200)]
docs: hacking: mention GLib alternatives of libvirt string allocation macros
Document the preferred alternatives to existing libvirt macros for
allocating strings. These cannot be deleted just yet because
converting them will require a lot of work.
Signed-off-by: Ján Tomko <jtomko@redhat.com> Reviewed-by: Daniel P. Berrangé <berrange@redhat.com> Reviewed-by: Andrea Bolognani <abologna@redhat.com>
Ján Tomko [Fri, 18 Oct 2019 21:12:19 +0000 (23:12 +0200)]
docs: hacking: mention GLib alternatives of libvirt allocation macros
Document the preferred alternatives to existing libvirt macros for
memory allocation. These cannot be deleted just yet because
converting them will require a lot of work.
Signed-off-by: Ján Tomko <jtomko@redhat.com> Reviewed-by: Andrea Bolognani <abologna@redhat.com>
Ján Tomko [Fri, 18 Oct 2019 20:57:23 +0000 (22:57 +0200)]
docs: hacking: emphasize some sections
Namely:
* holding up the first-time patch submissions for moderation,
which might cause first-time submitters to question the process
* not CC-ing individual developers
Signed-off-by: Ján Tomko <jtomko@redhat.com> Reviewed-by: Daniel P. Berrangé <berrange@redhat.com> Reviewed-by: Andrea Bolognani <abologna@redhat.com>
Peter Krempa [Thu, 10 Oct 2019 13:02:00 +0000 (15:02 +0200)]
qemu: block: Don't query monitor in qemuBlockStorageSourceCreateDetectSize
Calling the monitor was convenient for the implementation in
qemuDomainBlockCopyCommon, but causes the snapshot code to call
query-named-block-nodes for every disk.
Fix this by removing the monitor call from
qemuBlockStorageSourceCreateDetectSize so that the data can be reused in
loops.
Signed-off-by: Peter Krempa <pkrempa@redhat.com> ACKed-by: Eric Blake <eblake@redhat.com>
Peter Krempa [Wed, 9 Oct 2019 12:13:10 +0000 (14:13 +0200)]
qemu: monitor: Introduce new interface to query-named-block-nodes
Retrieve data for individual block nodes in a hash table. Currently only
capacity and allocation data is extracted but this will be extended in
the future.
Signed-off-by: Peter Krempa <pkrempa@redhat.com> ACKed-by: Eric Blake <eblake@redhat.com>
Peter Krempa [Wed, 9 Oct 2019 13:34:54 +0000 (15:34 +0200)]
util: hash: Add new constructor 'virHashNew'
Add a simpler constructor for hash tables which specifically does not
require specifying the initial hash size and uses simpler freeing
function.
The initial hash table size usually is not important as the hash table
is growing when it reaches certain number of entries in one bucket.
Additionally many callers pass in a random small number for ad-hoc table
use so using a central one will simplify things.
Signed-off-by: Peter Krempa <pkrempa@redhat.com> ACKed-by: Eric Blake <eblake@redhat.com>
Peter Krempa [Wed, 9 Oct 2019 13:26:37 +0000 (15:26 +0200)]
util: hash: Add possibility to use simpler data free function in virHash
Introduce a new type virHashDataFreeSimple which has only a void * as
argument for cases when knowing the name of the entry when freeing the
hash entry is not required.
Signed-off-by: Peter Krempa <pkrempa@redhat.com> ACKed-by: Eric Blake <eblake@redhat.com>
Peter Krempa [Mon, 14 Oct 2019 15:17:16 +0000 (17:17 +0200)]
qemu: Replace use of virDomainDiskFindByBusAndDst with virDomainDiskByTarget
In both replaced cases we have other code that verifies that the bus
can't be changed or that the target is unique, so limiting the search to
disks with same bus makes no sense.
Signed-off-by: Peter Krempa <pkrempa@redhat.com> Reviewed-by: Daniel Henrique Barboza <danielhb413@gmail.com>
Peter Krempa [Tue, 8 Oct 2019 13:17:32 +0000 (15:17 +0200)]
qemu: domain: Tolerate NULL @disk in qemuDomainPrepareDiskSourceData
In some cases we want to prepare a @src which is not meant to belong to
a disk and thus does not require us to copy the data. Allow passing in
NULL @disk into qemuDomainPrepareDiskSourceData.
Signed-off-by: Peter Krempa <pkrempa@redhat.com> Reviewed-by: Daniel Henrique Barboza <danielhb413@gmail.com>
Peter Krempa [Tue, 8 Oct 2019 12:58:48 +0000 (14:58 +0200)]
qemu: domain: Split out setup of virStorageSource from qemu driver config
qemuDomainPrepareDiskSourceData historically prepared everything but
we've split out the majority of the functionality so that it sets up
predominately only according to the configuration of the disk. There
was one leftover bit of setting the gluster debug level from the config.
Split this out into a separate function so that
qemuDomainPrepareDiskSourceData only prepares based on the disk.
Signed-off-by: Peter Krempa <pkrempa@redhat.com> Reviewed-by: Daniel Henrique Barboza <danielhb413@gmail.com> ACKed-by: Eric Blake <eblake@redhat.com>
Peter Krempa [Tue, 15 Oct 2019 13:21:21 +0000 (15:21 +0200)]
conf: Reset disk type if <source> element is completely missing
The disk type is not part of source and thus it's parsed earlier. This
bypasses the checks when parsing a disk type='network' if it's
completely missing the source.
Since there are possible active users of this (it was reported as a
problem with openstack) fix it by resetting the disk type to '_FILE' for
an empty cdrom which is handled correctly.
Signed-off-by: Peter Krempa <pkrempa@redhat.com> Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
Michal Privoznik [Tue, 17 Sep 2019 09:25:03 +0000 (11:25 +0200)]
qemuDomainGetHostdevPath: Drop @freeTmpPath
The @freeTmpPath boolean is used to determine if @tmpPath holds
an allocated memory or is a pointer to a constant string and
therefore if it needs to be freed or not when returning from the
function. Well, we can unify the way we set @tmpPath so that it
always holds an allocated memory and thus always must be freed.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com> Reviewed-by: Cole Robinson <crobinso@redhat.com>
Michal Privoznik [Tue, 17 Sep 2019 09:20:42 +0000 (11:20 +0200)]
qemu_domain: Drop few useless checks in qemuDomainGetHostdevPath
There are three cases where vir*DeviceGetPath() returns a const
string. In these cases, the string is initialized in
corresponding vir*DeviceNew() calls which fail if string couldn't
be allocated. There's no point in checking the second time if the
string is NULL.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com> Reviewed-by: Cole Robinson <crobinso@redhat.com>
Michal Privoznik [Tue, 17 Sep 2019 08:39:00 +0000 (10:39 +0200)]
qemu_cgroup: Teardown Cgroup for more host device types
Since its introduction in v1.0.5-rc1-19-g6e13860cb4 the
qemuTeardownHostdevCgroup() does nothing unless the passed
hostdev is a PCI device with VFIO backend. This seems
unnecessary.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com> Reviewed-by: Cole Robinson <crobinso@redhat.com>
virhostdev: Introduce and use virHostdevIsVFIODevice
In some places we need to check if a hostdev has VFIO backend.
Because of how complicated virDomainHostdevDef structure is, the
check consists of three lines. Move them to a function and
replace all checks with the function call.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com> Reviewed-by: Cole Robinson <crobinso@redhat.com>
qemu: fix CPU model error probing capabilities for ppc
The CPU driver only supports CPU models for PPC64 architecture, not
plain PPC.
Failed to probe capabilities for /usr/bin/qemu-system-ppc:
this function is not supported by the connection driver:
'ppc' architecture is not supported by CPU driver
Reviewed-by: Daniel Henrique Barboza <danielhb413@gmail.com> Reviewed-by: Andrea Bolognani <abologna@redhat.com> Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
network: allow DHCP/DNS/TFTP explicitly in OUTPUT rules
While the default iptables setup used by Fedora/RHEL distros
only restricts traffic on the INPUT and/or FORWARD rules,
some users might have custom firewalls that restrict the
OUTPUT rules too.
These can prevent DHCP/DNS/TFTP responses from dnsmasq
from reaching the guest VMs. We should thus whitelist
these protocols in the OUTPUT chain, as well as the
INPUT chain.
Signed-off-by: Malina Salina <malina.salina@protonmail.com>
Initial patch then modified to add unit tests and IPv6
support
Reviewed-by: Michal Privoznik <mprivozn@redhat.com> Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
Some people from IBM does not use 'ibm.com' domain emails.
Suggested-by: Leonardo Augusto Guimarães Garcia <lagarcia@br.ibm.com> Signed-off-by: Daniel Henrique Barboza <danielhb413@gmail.com> Reviewed-by: Andrea Bolognani <abologna@redhat.com>
Allow stack traces to be included with log messages
Sometimes it is useful to see the callpath for log messages.
This change enhances the log filter syntax so that stack traces
can be show by setting '1:+NAME' instead of '1:NAME'.
With the huge & ever increasing number of logging statements per file,
this will be incredibly verbose and have a major performance penalty.
This makes the feature impractical to use widely and as such it is not
worth the code maint cost.
Removing this seldom used feature allows us to drop the 'execinfo'
module in gnulib which provides the backtrace() function which doesn't
exist on non-Linux.
Users who want to get stack traces of parts of libvirt can use GDB,
or systemtap for live tracing with minimal perf impact.
Reviewed-by: Ján Tomko <jtomko@redhat.com> Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
Michal Privoznik [Thu, 17 Oct 2019 15:17:01 +0000 (17:17 +0200)]
glibcompat: Reimplement g_strdup_printf() and g_strdup_vprintf()
These functions don't really abort() on OOM. The fix was merged
upstream, but not in the minimal version we require. Provide our
own implementation which can be removed once we bump the minimal
version.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com> Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
As part of an goal to eliminate Perl from libvirt build tools,
rewrite the minimize-po.pl tool in Python.
This was a straight conversion, manually going line-by-line to
change the syntax from Perl to Python. Thus the overall structure
of the file and approach is the same.
Reviewed-by: Ján Tomko <jtomko@redhat.com> Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
projects / libvirt.git / log