While all the code that deals with qemu namespaces correctly
detects whether we are running as root (and turn into NO-OP for
qemu:///session) the actual unshare() call is not guarded with
such check. Therefore any attempt to start a domain under
qemu:///session shall fail as unshare() is reserved for root.
The fix consists of moving unshare() call (for which we have a
wrapper called virProcessSetupPrivateMountNS) into
qemuDomainBuildNamespace() where the proper check is performed.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com> Tested-by: Richard W.M. Jones <rjones@redhat.com>
John Ferlan [Fri, 13 Jan 2017 12:33:02 +0000 (07:33 -0500)]
storage: Alter error message in probe/empty checks
For case VIR_STORAGE_BLKID_PROBE_DIFFERENT, clean up the message to
avoid using the virsh like --overwrite syntax. Additionally provide
a different error message when not writing the label to avoid confusion.
John Ferlan [Fri, 13 Jan 2017 12:34:03 +0000 (07:34 -0500)]
storage: Clean up return value checking
Rather than special casing the VIR_STORAGE_BLKID_PROBE_UNKNOWN after
calling virStorageBackendBLKIDFindPart, just allow the switch statement
handle setting ret = -2.
John Ferlan [Fri, 13 Jan 2017 00:29:40 +0000 (19:29 -0500)]
storage: Alter logic when both BLKID and PARTED unavailable
If neither BLKID or PARTED is available and we're not writing, then
just return 0 which allows the underlying storage pool to generate
a failure. If both are unavailable and we're writing, then generate
a more generic error message.
qemu-capabilities: Fix query-cpu-model-expansion on s390 with older kernel
When running on s390 with a kernel that does not support cpu model checking and
with a Qemu new enough to support query-cpu-model-expansion, the gathering of qemu
capabilities will fail. Qemu responds to the query-cpu-model-expansion qmp
command with an error because the needed kernel ioct does not exist. When this
happens a guest cannot even be defined due to missing qemu capabilities data.
This patch fixes the problem by silently ignoring generic errors stemming from
calls to query-cpu-model-expansion.
Reported-by: Farhan Ali <alifm@linux.vnet.ibm.com> Signed-off-by: Collin L. Walling <walling@linux.vnet.ibm.com> Signed-off-by: Jason J. Herne <jjherne@linux.vnet.ibm.com>
Michal Privoznik [Fri, 13 Jan 2017 09:03:23 +0000 (10:03 +0100)]
qemu: Copy SELinux labels for namespace too
When creating new /dev/* for qemu, we do chown() and copy ACLs to
create the exact copy from the original /dev. I though that
copying SELinux labels is not necessary as SELinux will chose the
sane defaults. Surprisingly, it does not leaving namespace with
the following labels:
error: internal error: process exited while connecting to monitor: Error in GnuTLS initialization: Failed to acquire random data.
qemu-kvm: cannot initialize crypto: Unable to initialize GNUTLS library: Failed to acquire random data.
The solution is to copy the SELinux labels as well.
Reported-by: Andrea Bolognani <abologna@redhat.com> Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Peter Krempa [Thu, 12 Jan 2017 13:04:09 +0000 (14:04 +0100)]
Revert "storage: Validate the device formats at logical startup"
The check is pointless since LVM is capable to detect it's own members
and the check is flawed as it would fail if neither libblkid nor parted
is installed.
Peter Krempa [Thu, 12 Jan 2017 12:50:19 +0000 (13:50 +0100)]
Revert "storage: For FS pool check for properly formatted target volume"
The check does not work properly (crashes) with netfs filesystems and
also checking that a device is not empty when attempting to mount a
filesystem is not very usefull since the mount will fail anyways.
As the code would improve only a very minor corner case I don't really
see a reason to have this code at all.
This code would also fail if libvirt is compiled without support for
blkid and without parted.
Jim Fehlig [Thu, 12 Jan 2017 00:42:42 +0000 (17:42 -0700)]
libxl: always enable pae for x86_64 HVM
For HVM domains, pae is only set in libxl_domain_build_info when
explicitly specified in the hypervisor <features> config. This is
fine for i686 machines, but is incorrect behavior for x86_64 machines
where pae must always be enabled. See the following discussion for
additional details
Jiri Denemark [Wed, 11 Jan 2017 13:35:52 +0000 (14:35 +0100)]
qemu: Ignore non-boolean CPU model properties
The query-cpu-model-expansion is currently implemented for s390(x) only
and all CPU properties it returns are booleans. However, x86
implementation will report more types of properties. Without making the
code more tolerant older libvirt would fail to probe newer QEMU
versions.
Jiri Denemark [Wed, 11 Jan 2017 13:29:58 +0000 (14:29 +0100)]
qemu: Don't check CPU model property key
The qemuMonitorJSONParseCPUModelProperty function is a callback for
virJSONValueObjectForeachKeyValue and is called for each key/value pair,
thus it doesn't really make sense to check whether key is NULL.
Michal Privoznik [Wed, 11 Jan 2017 16:32:46 +0000 (17:32 +0100)]
qemuDomainCreateDevice: Canonicalize paths
So far the decision whether /dev/* entry is created in the qemu
namespace is really simple: does the path starts with "/dev/"?
This can be easily fooled by providing path like the following
(for any considered device like disk, rng, chardev, ..):
/dev/../var/lib/libvirt/images/disk.qcow2
Therefore, before making the decision the path should be
canonicalized.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Michal Privoznik [Wed, 11 Jan 2017 10:05:50 +0000 (11:05 +0100)]
qemu: Use namespaces iff available on the host kernel
So far the namespaces were turned on by default unconditionally.
For all non-Linux platforms we provided stub functions that just
ignored whatever namespaces setting there was in qemu.conf and
returned 0 to indicate success. Moreover, we didn't really check
if namespaces are available on the host kernel.
This is suboptimal as we might have ignored user setting.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Michal Privoznik [Wed, 11 Jan 2017 10:29:00 +0000 (11:29 +0100)]
util: Introduce virFileMoveMount
This is a simple wrapper over mount(). However, not every system
out there is capable of moving a mount point. Therefore, instead
of having to deal with this fact in all the places of our code we
can have a simple wrapper and deal with this fact at just one
place.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Michal Privoznik [Wed, 11 Jan 2017 09:23:39 +0000 (10:23 +0100)]
lxc: Move lxcContainerAvailable to virprocess
Other drivers (like qemu) would like to know if the namespaces
are available therefore it makes sense to move this function to
a shared module.
At the same time, this function had some default namespaces that
are checked with every call. It is not necessary - let callers
pass just those namespaces they are interested in.
With the move the function is renamed to
virProcessNamespaceAvailable.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
John Ferlan [Wed, 11 Jan 2017 00:35:10 +0000 (19:35 -0500)]
docs: Add NPIV/vHBA change description to news.xml
Add "Improvements" for commit id 'bb74a7ffe' and '78be2e8b7' which add
support for using the parent wwnn/wwpn or fabric_name rather than just
using the parent by scsi_hostX name.
Pino Toscano [Tue, 10 Jan 2017 18:43:20 +0000 (19:43 +0100)]
remote: do not check for an existing config dir
When composing the path to the default known_hosts file (for the libssh
and libssh2 drivers), do not check whether the configuration directory
(determined by virGetUserConfigDirectory()) exists: both the drivers can
handle non-existing files, and are able to create them (and their
directories) in that case.
This adds a small behaviour change: before, the key for an unknown host,
and manually accepted, was saved only if the configuration directory
existed -- a bit incoherent behaviour though.
Pino Toscano [Tue, 10 Jan 2017 18:43:19 +0000 (19:43 +0100)]
remote: fix logic for known_hosts and keyfile checks
If any of them is specified for the libssh and libssh2 drivers, there is
no need to depend on checks based on other paths: in particular, a
specified path for known_hosts was ignored if the local config directory
could not be determined, and the path for keyfile was ignored if the
home could not be determined.
Instead, lazily determine and use these two paths only in case they are
needed.
Pino Toscano [Tue, 10 Jan 2017 18:43:18 +0000 (19:43 +0100)]
rpc: libssh: allow a NULL known_hosts file
Make sure that virNetLibsshSessionSetHostKeyVerification accepts a NULL
value for the path to the known_hosts file:
- call ssh_options_set(SSH_OPTIONS_KNOWNHOSTS) anyway, using /dev/null,
otherwise libssh will use its default path
- do not call ssh_write_knownhost when no known hosts file was set
Laine Stump [Mon, 24 Oct 2016 15:25:53 +0000 (11:25 -0400)]
conf: eliminate virDomainPCIAddressReleaseSlot() in favor of ...Addr()
Surprisingly there was a virDomainPCIAddressReleaseAddr() function
already, but it was completely unused. Since we don't reserve entire
slots at once any more, there is no need to release entire slots
either, so we just replace the single call to
virDomainPCIAddressReleaseSlot() with a call to
virDomainPCIAddressReleaseAddr() and remove the now unused function.
The keen observer may be concerned that ...Addr() doesn't call
virDomainPCIAddressValidate(), as ...Slot() did. But really the
validation was pointless anyway - if the device hadn't been suitable
to be connected at that address, it would have failed validation
before every being reserved in the first place, so by definition it
will pass validation when it is being unplugged. (And anyway, even if
something "bad" happened and we managed to have a device incorrectly
at the given address, we would still want to be able to free it up for
use by a device that *did* validate properly).
This function is only called in two places, and the function itself is
just adding a single argument and calling
virDomainPCIAddressReserveNextAddr(), so we can remove it and instead
call virDomainPCIAddressReserveNextAddr() directly. (The main
motivation for doing this is to free up the name so that
qemuDomainPCIAddressReserveNextSlot() can be renamed in the next
patch, as its current name is now inaccurate and misleading).
Laine Stump [Sun, 23 Oct 2016 16:07:20 +0000 (12:07 -0400)]
conf: rename virDomainPCIAddressReserveAddr() to ...Internal()
This is in preparation for renaming virDomainPCIAddressReserveSlot()
to virDomainPCIAddressReserveAddr(), which is a better description of
what it does.
Laine Stump [Sun, 23 Oct 2016 15:57:55 +0000 (11:57 -0400)]
qemu: replace virDomainPCIAddressReserveAddr with virDomainPCIAddressReserveSlot
All occurences of the former use fromConfig=true, and that's exactly
how virDomainPCIAddressReserveSlot() calls
virDomainPCIaddressReserveAddr(), so just use *Slot() so that *Addr()
can be made static to conf/domain_addr.c (both functions will be
renamed in upcoming patches).
Since we don't actually reserve an entire slot at a time anymore, the
name of this function is just confusing, and it's almost identical in
operation to virDomainPCIAddressReserveNextAddr() anyway, so remove
the *Slot() function and replace calls to it with calls to *Addr(...,
-1).
Laine Stump [Sat, 22 Oct 2016 17:10:10 +0000 (13:10 -0400)]
conf: rename virDomainPCIAddressGetNextSlot() to ...GetNextAddr()
With the advent of VIR_PCI_CONNECT_AGGREGATE_SLOT, the new name is
more appropriate, since the address returned may be another address
on the same slot as last time, not necessarily a new slot.
Laine Stump [Sat, 22 Oct 2016 16:53:17 +0000 (12:53 -0400)]
conf: fix fromConfig argument to virDomainPCIAddressValidate()
fromConfig should be true if the caller wants
virDomainPCIAddressValidate() to loosen restrictions on its
interpretation of the pciConnectFlags. In particular, either
PCI_DEVICE or PCIE_DEVICE will be counted as equivalent to both, and
HOTPLUG will be ignored. In a few cases where libvirt was manually
overriding automatic address assignment, it was setting fromConfig to
false when validating the hardcoded manual override. This patch
changes those to fromConfig=true as a preemptive strike against any
future bugs that might otherwise surface.
Laine Stump [Sat, 22 Oct 2016 16:35:38 +0000 (12:35 -0400)]
conf: fix fromConfig argument to virDomainPCIAddressReserveAddr()
Although setting virDomainPCIAddressReserveAddr()'s fromConfig=true is
correct when a PCI addres is coming from a domain's config, the *true*
purpose of the fromConfig argument is to lower restrictions on what
kind of device can plug into what kind of controller - if fromConfig
is true, then a PCIE_DEVICE can plug into a slot that is marked as
only compatible with PCI_DEVICE (and vice versa), and the HOTPLUG flag
is ignored.
For a long time there have been several calls to
virDomainPCIAddressReserveAddr() that have fromConfig incorrectly set
to false - it's correct that the addresses aren't coming from user
config, but they are coming from hardcoded exceptions in libvirt that
should, if anything, pay *even less* attention to following the
pciConnectFlags (under the assumption that the libvirt programmer knew
what they were doing).
See commit b87703cf7 for an example of an actual bug caused by the
incorrect setting of the "fromConfig" argument to
virDomainPCIAddressReserveAddr(). Although they haven't resulted in
any reported bugs, this patch corrects all the other incorrect
settings of fromConfig in calls to virDomainPCIAddressReserveAddr().
Laine Stump [Wed, 19 Oct 2016 18:15:01 +0000 (14:15 -0400)]
conf: aggregate multiple devices on a slot when assigning PCI addresses
If a PCI device has VIR_PCI_CONNECT_AGGREGATE_SLOT set in its
pciConnectFlags, then during address assignment we allow multiple
instances of this type of device to be auto-assigned to multiple
functions on the same device. A slot is used for aggregating multiple
devices only if the first device assigned to that slot had
VIR_PCI_CONNECT_AGGREGATE_SLOT set. but any device types that have
AGGREGATE_SLOT set might be mix/matched on the same slot.
(NB: libvirt should never set the AGGREGATE_SLOT flag for a device
type that might need to be hotplugged. Currently it is only planned
for pcie-root-port and possibly other PCI controller types, and none
of those are hotpluggable anyway)
There aren't yet any devices that use this flag. That will be in a
later patch.
Laine Stump [Tue, 10 Jan 2017 05:20:11 +0000 (00:20 -0500)]
qemu: use virDomainPCIAddressSetAllMulti() to set multi when needed
If there are multiple devices assigned to the different functions of a
single PCI slot, they will not work properly if the device at function
0 doesn't have its "multi" attribute turned on, so it makes sense for
libvirt to turn it on during PCI address assignment. Setting multi
then assures that the new setting is stored in the config (so it will
be used next time the domain is started), preventing any potential
problems in the case that a future change in the configuration
eliminates the devices on all non-0 functions (multi will still be set
for function 0 even though it is the only function in use on the slot,
which has no useful purpose, but also doesn't cause any problems).
(NB: If we were to instead just decide on the setting for
multifunction at runtime, a later removal of the non-0 functions of a
slot would result in a silent change in the guest ABI for the
remaining device on function 0 (although it may seem like an
inconsequential guest ABI change, it *is* a guest ABI change to turn
off the multi bit).)
Laine Stump [Tue, 10 Jan 2017 05:02:40 +0000 (00:02 -0500)]
conf: new function virDomainPCIAddressSetAllMulti()
This utility function iterates through all devices looking for any
with a PCI address that has function != 0 (which implies that multiple
functions are in use on that slot), then uses an inner iterator to
find the device that's on function 0 of that same slot and sets the
"multi" in its virDomainDeviceInfo (as long as it hasn't already been
set explicitly by someone who presumably has better information than
we do).
It isn't yet called from anywhere, so will have no functional effect.
Laine Stump [Fri, 21 Oct 2016 17:05:33 +0000 (13:05 -0400)]
conf: start search for next unused PCI address at same slot as previous find
There is a very slight time advantage to beginning the search for the
next unused PCI address at the slot *after* the previous find (which
is now used), but if we do that, we will miss allocating the other
functions of the same slot (when we implement a
VIR_PCI_CONNECT_AGGREGATE_SLOT flag to support that).
Laine Stump [Wed, 19 Oct 2016 17:58:42 +0000 (13:58 -0400)]
conf: eliminate repetitive code in virDomainPCIAddressGetNextSlot()
virDomainPCIAddressGetNextSlot() starts searching from the last
allocated address and goes to the end of all the buses, then goes back
to the first bus and searches from there up to the starting point (in
case any address has been freed since the last time an address was
allocated. The loops for these two are almost, but not exactly, the
same, so they have remained as separate loops with the same code
inside the loop. To lessen maintenance headaches, the identical code
has been moved out into the function
virDomainPCIAddressFindUnusedFunctionOnBus(), which is called in place
of the loop contents.
Laine Stump [Wed, 19 Oct 2016 16:43:04 +0000 (12:43 -0400)]
conf: eliminate concept of "reserveEntireSlot"
setting reserveEntireSlot really accomplishes nothing - instead of
going to the trouble of computing the value for reserveEntireSlot and
then possibly setting *all* functions of the slot as in-use, we can
just set the in-use bit only for the specific function being used by a
device. Later we will know from the context (the PCI connect flags,
and whether we are reserving a specific address or asking for "the
next available") whether or not it is okay to allocate other functions
on the same slot.
Although it's not used yet, we allow specifying "-1" for the function
number when looking for the "next available slot" - this is going to
end up meaning "return the lowest available function in the slot, but
since we currently only provide a function from an otherwise unused
slot, "-1" ends up meaning "0".
Laine Stump [Sun, 16 Oct 2016 21:14:25 +0000 (17:14 -0400)]
conf: use struct instead of int for each slot in virDomainPCIAddressBus
When keeping track of which functions of which slots are allocated, we
will need to have more information than just the current bitmap with a
bit for each function that is currently stored for each slot in a
virDomainPCIAddressBus. To prepare for adding more per-slot info, this
patch changes "uint8_t slots" into "virDomainPCIAddressSlot slot", which
currently has a single member named "functions" that serves the same
purpose previously served directly by "slots".
libxl doesn't provide a way to write one log for each domain. Thus
we need to demux the messages. If our logger doesn't know to which
domain to attribute a message, then it will write it to the default
log file.
Starting with Xen 4.9 (commit f9858025 and following), libxl will
write the domain ID in an easy to grab manner. The logger introduced
by this commit will use it to demux the libxl log messages.
Thanks to the default log file, this logger will also work with older
versions of Xen.
Dawid Zamirski [Thu, 29 Dec 2016 20:34:32 +0000 (15:34 -0500)]
vbox: consolidate vbox IID structures.
* remove _vboxIID_v2_x and _vboxIID_v3_x structs and repalce with one
_vboxIID as all supprted vbox versions have the same IID structure.
* remove vboxIIDUnion that was used to abstract version depended IID
differences.
* remove IID_MEMBER macro and use the new vboxIID directly.
Dawid Zamirski [Thu, 29 Dec 2016 20:34:31 +0000 (15:34 -0500)]
vbox: fix _displayTakeScreenShotPNGToArray
This function was not implemented for vbox 5+ which removed
TakeScreenShotPNGToArray but provides TakeScreenShotToArray with
BitmapFormat_PNG argument which is the same thing.
Dawid Zamirski [Thu, 29 Dec 2016 20:34:30 +0000 (15:34 -0500)]
vbox: IVRDxServer to IVRDEServer.
The IVRDxServer was used because vbox < 4 used to have IVRDPServer
whereas vbox >= 4 has IVRDEServer. Now that support for legacy
versions is being removed, we can use IVRDEServer.
Dawid Zamirski [Thu, 29 Dec 2016 20:34:29 +0000 (15:34 -0500)]
vbox: remove code dealing with oldMediumInterface
* removed oldMediumInterface flag and related code that was used for
vbox 2.x
* remove accelerate2DVideo and networkRemoveInterface flags which were
also conditionals for handling legacy vbox versions.
Dawid Zamirski [Thu, 29 Dec 2016 20:34:27 +0000 (15:34 -0500)]
vbox: remove getMachineForSession flag.
* the getMachineForSession is always true for 4.0+. This also means that
checkflag argument in openSessionForMachine no longer has any meaning
because it was or'ed with getMachineForSession (always true)
* remove supportScreenshot flag - vbox 4.0+ supports it
* remove detachDevicesExplicitly flag only relevant for < 4.0
Dawid Zamirski [Thu, 29 Dec 2016 20:34:24 +0000 (15:34 -0500)]
vbox: remove code for old API versions.
This removes most of the code wrapped in VBOX_API_VERSION < 4000000
preprocessor checks. Those are the ones that can be safely removed
without needing to update driver code to accomodate it.
Dawid Zamirski [Tue, 29 Nov 2016 22:23:27 +0000 (17:23 -0500)]
vbox: remove SDK header files for vbox 3 and older.
* delete SDK header files for vbox older than 4.0
* delete .c files for vbox older than 4.0
* update vbox_XPCOMCGlue to use oldest supported header file, that is 4.0
going forward.
* remove deleted files from Makefile.am
Eric Blake [Tue, 10 Jan 2017 18:39:30 +0000 (12:39 -0600)]
maint: update to latest gnulib
Among other recent changes, this includes a workaround to avoid
Clang compiler bug https://llvm.org/bugs/show_bug.cgi?id=16404
having spurious link failures.
* .gnulib: Update.
* bootstrap: Synchronize to upstream.
docs: Document the release notes process for contributors
Now that we have built a fairly solid process for dealing with
release notes, we should start pushing for contributors to
provide the relevant information along with their code:
documenting the process is clearly a requirement for this to
happen.
Instead of encoding formatting information inside the
corresponding XSLT stylesheet, use a Python script to reformat
the text appropriately based on a few simple markers.
Splitting the task between the XSLT stylesheet and the Python
script allows us to keep both parts very simple.
Currently, building the NEWS file involves using a XSLT stylesheet
to extract information from the same HTML file that's used on the
libvirt website.
The process works, but it's quite fiddly in that it requires the
source HTML to be formatted in a very precise way, and a single
missing newline can mess up the resulting plain text considerably.
Moreover, the XSLT stylesheet itself encodes a lot of the details
of converting to plain text in a way that's not necessarily easy
to understand, tweak or fix.
To improve the process, move all existing entries to a new XML
file that contains exactly the information we care about in a
simple structured format, and start generating both the HTML and
plain text versions of the release notes using XSLT stylesheets
that can now afford to be almost trivial.
Michal Privoznik [Tue, 10 Jan 2017 15:33:58 +0000 (16:33 +0100)]
virSecuritySELinuxSetFileconHelper: Fix build with broken selinux.h
There are still some systems out there that have broken
setfilecon*() prototypes. Instead of taking 'const char *tcon' it
is taking 'char *tcon'. The function should just set the context,
not modify it.
We had been bitten with this problem before which resulted in 292d3f2d and subsequently b109c09765. However, with one my latest
commits (4674fc6afd6d) I've changed the type of @tcon variable to
'const char *' which results in build failure on the systems from
above.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Peter Krempa [Fri, 16 Dec 2016 14:45:26 +0000 (15:45 +0100)]
qemu: snapshot: Properly handle image locking
Images that became the backing chain of the current image due to the
snapshot need to be unlocked in the lock manager. Also if qemu was
paused during the snapshot the current top level images need to be
released until qemu is resumed so that they can be acquired properly.
Peter Krempa [Fri, 16 Dec 2016 05:12:34 +0000 (06:12 +0100)]
qemu: snapshot: Refactor snapshot rollback on failure
The code at first changed the definition and then rolled it back in case
of failure. This was ridiculous. Refactor the code so that the image in
the definition is changed only when the snapshot is successful.
The refactor will also simplify further fix of image locking when doing
snapshots.
Peter Krempa [Fri, 16 Dec 2016 06:10:46 +0000 (07:10 +0100)]
qemu: snapshot: Don't redetect backing chain after snapshot
Libvirt is able to properly model what happens to the backing chain
after a snapshot so there's no real need to redetect the data.
Additionally with the _REUSE_EXT flag this might end up in redetecting
wrong data if the user puts wrong backing chain reference into the
snapshot image.
John Ferlan [Tue, 10 Jan 2017 15:15:19 +0000 (10:15 -0500)]
storage: Fix storage_backend probing when PARTED not installed.
Commit id 'a48c674f' caused problems for systems without PARTED installed.
So move the PARTED probing code back to storage_backend_disk.c and create
a shim within storage_backend.c to call it if WITH_STORAGE_DISK is true;
otherwise, just return -1 with the error.
John Ferlan [Thu, 15 Dec 2016 14:19:49 +0000 (09:19 -0500)]
storage: Validate the device formats at logical startup
At startup time, rather than blindly trusting the target devices are
still properly formatted, let's check to make sure the pool's target
devices are all properly formatted before attempting to start the pool.
John Ferlan [Thu, 15 Dec 2016 11:39:42 +0000 (06:39 -0500)]
storage: Clean up logical pool devices on build failure
If the build fails, then we need to ensure that we've run pvremove
on any devices which we've run pvcreate on; otherwise, a subsequent
build could fail since running pvcreate twice on a device requires
special force arguments.
John Ferlan [Thu, 15 Dec 2016 20:45:01 +0000 (15:45 -0500)]
storage: Adjust disk label found to match labels
Currently as long as the disk is formatted using a known parted format
type, the algorithm is happy to continue. However, that leaves a scenario
whereby a disk formatted using "pc98" could be used by a pool that's defined
using "dvh" (or vice versa). Alter the check to be match and different
and adjust the caller.
John Ferlan [Wed, 14 Dec 2016 23:34:24 +0000 (18:34 -0500)]
storage: Move and rename disk backend label checking
Rather than have the Disk code having to use PARTED to determine if
there's something on the device, let's use the virStorageBackendDeviceProbe.
and only fallback to the PARTED probing if the BLKID code isn't built in.
This will also provide a mechanism for the other current caller (File
System Backend) to utilize a PARTED parsing algorithm in the event that
BLKID isn't built in to at least see if *something* exists on the disk
before blindly trying to use. The PARTED error checking will not find
file system types, but if there is a partition table set on the device,
it will at least cause a failure.
Move virStorageBackendDiskValidLabel and virStorageBackendDiskFindLabel
to storage_backend and rename/rework the code to fit the new model.
Update the virsh.pod description to provide a more generic description
of the process since we could now use either blkid or parted to find
data on the target device.
John Ferlan [Wed, 14 Dec 2016 23:03:20 +0000 (18:03 -0500)]
storage: Add writelabel bool for virStorageBackendDeviceProbe
It's possible that the API could be called from a startup path in
order to check whether the label on the device matches what our
format is. In order to handle that condition, add a 'writelabel'
boolean to the API in order to indicate whether a write or just
read is about to happen.
This alters two "error" conditions that would care about knowing.
John Ferlan [Wed, 14 Dec 2016 22:45:20 +0000 (17:45 -0500)]
storage: Add partition type checks for BLKID probing
A device may be formatted using some sort of disk partition format type.
We can check that using the blkid_ API's as well - so alter the logic to
allow checking the device for both a filesystem and a disk partition.
Commit id '27758859' introduced the "NO_OVERWRITE" flag check for
file system backends; however, the implementation, documentation,
and algorithm was inconsistent. For the "flag" description for the
API the flag was described as "Do not overwrite existing pool";
however, within the storage backend code the flag is described
as "it probes to determine if filesystem already exists on the
target device, renurning an error if exists".
The code itself was implemented using the paradigm to set up the
superblock probe by creating a filter that would cause the code
to only search for the provided format type. If that type wasn't
found, then the algorithm would return success allowing the caller
to format the device. If the format type already existed on the
device, then the code would fail indicating that the a filesystem
of the same type existed on the device.
The result is that if someone had a file system of one type on the
device, it was possible to overwrite it if a different format type
was specified in updated XML effectively trashing whatever was on
the device already.
This patch alters what NO_OVERWRITE does for a file system backend
to be more realistic and consistent with what should be expected when
the caller requests to not overwrite the data on the disk.
Rather than filter results based on the expected format type, the
code will allow success/failure be determined solely on whether the
blkid_do_probe calls finds some known format on the device. This
adjustment also allows removal of the virStoragePoolProbeResult
enum that was under utilized.
If it does find a formatted file system different errors will be
generated indicating a file system of a specific type already exists
or a file system of some other type already exists.
In the original virsh support commit id 'ddcd5674', the description
for '--no-overwrite' within the 'pool-build' command help output
has an ambiguous "of this type" included in the short description.
Compared to the longer description within the "Build a given pool."
section of the virsh.pod file it's more apparent that the meaning
of this flag would cause failure if a probe of the target already
has a filesystem.
So this patch also modifies the short description to just be the
antecedent of the 'overwrite' flag, which matches the API description.
This patch also modifies the grammar in virsh.pod for no-overwrite
as well as reworking the paragraph formats to make it easier to read.
qemuDomainCreateNamespace: move mkdir to qemuDomainBuildNamespace
Again, there is no need to create /var/lib/libvirt/$domain.*
directories in CreateNamespace(). It is sufficient to create them
as soon as we need them which is in BuildNamespace. This way we
don't leave them around for the whole lifetime of domain.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
qemuDomainGetPreservedMounts: Do not special case /dev
The c1140eb9e got me thinking. We don't want to special case /dev
in qemuDomainGetPreservedMounts(), but in all other places in the
code we special case it anyway. I mean,
/var/run/libvirt/$domain.dev path is constructed separately just
so that it is not constructed here. It makes only a little sense
(if any at all).
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
If something goes wrong in this function we try a rollback. That
is unlink all the directories we created earlier. For some weird
reason unlink() was called instead of rmdir().
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Michal Privoznik [Thu, 15 Dec 2016 15:47:15 +0000 (16:47 +0100)]
qemu: Use transactions from security driver
So far if qemu is spawned under separate mount namespace in order
to relabel everything it needs an access to the security driver
to run in that namespace too. This has a very nasty down side -
it is being run in a separate process, so any internal state
transition is NOT reflected in the daemon. This can lead to many
sleepless nights. Therefore, use the transaction APIs so that
libvirt developers can sleep tight again.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Michal Privoznik [Wed, 14 Dec 2016 09:14:49 +0000 (10:14 +0100)]
security driver: Introduce transaction APIs
With our new qemu namespace code in place, the relabelling of
devices is done not as good is it could: a child process is
spawned, it enters the mount namespace of the qemu process and
then runs desired API of the security driver.
Problem with this approach is that internal state transition of
the security driver done in the child process is not reflected in
the parent process. While currently it wouldn't matter that much,
it is fairly easy to forget about that. We should take the extra
step now while this limitation is still fresh in our minds.
Three new APIs are introduced here:
virSecurityManagerTransactionStart()
virSecurityManagerTransactionCommit()
virSecurityManagerTransactionAbort()
The Start() is going to be used to let security driver know that
we are starting a new transaction. During a transaction no
security labels are actually touched, but rather recorded and
only at Commit() phase they are actually updated. Should
something go wrong Abort() aborts the transaction freeing up all
memory allocated by transaction.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
projects / libvirt.git / log