In a few cases we might set seclabels on a path outside of
namespaces. For instance, when restoring a domain from a file,
the file is opened, relabelled and only then the namespace is
created and the FD is passed to QEMU (see v6.3.0-rc1~108 for more
info). Therefore, when restoring the label on the restore file,
we must ignore domain namespaces and restore the label directly
in the host.
This bug demonstrates itself when restoring a domain from a block
device. We don't create the block device inside the domain
namespace and thus the following error is reported at the end of
(otherwise successful) restore:
error : virProcessRunInFork:1236 : internal error: child reported (status=125): unable to stat: /dev/sda: No such file or directory
error : virProcessRunInFork:1240 : unable to stat: /dev/sda: No such file or directory
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Erik Skultety <eskultet@redhat.com>
The function calls virSecurityManagerDomainRestorePathLabel()
after all.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Erik Skultety <eskultet@redhat.com>
The new name is virSecurityManagerDomainRestorePathLabel().
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Erik Skultety <eskultet@redhat.com>
After previous commit this function is used no more.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Erik Skultety <eskultet@redhat.com>
After previous commit this function is used no more.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Erik Skultety <eskultet@redhat.com>
There are two places within qemu driver that misuse
qemuSecuritySetSavedStateLabel() to set seclabels on tempfiles
that are not state files: qemuDomainScreenshot() and
qemuDomainMemoryPeek(). They are doing so because of lack of
qemuSecurityDomainSetPathLabel() at the time of their
introduction.
In all three secdrivers (well, four if you count NOP driver) the
implementation of .domainSetSavedStateLabel and
.domainSetPathLabel callbacks is the same anyway.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Erik Skultety <eskultet@redhat.com>
In v6.4.0-72-g3dda889a44 I've introduced parsing and formatting
of new sysinfo type 'fwcfg'. However, I've forgot to introduce
code that would free parsed data.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Daniel Henrique Barboza <danielhb413@gmail.com>
In v6.4.0-rc1~143 I've introduced a check that is supposed to
return from the function early, if given path is not a dm target.
While the idea is still valid, the implementation had a flaw.
It calls stat() over given path and the uses major(sb.st_dev) to
learn the major of the device. This is then passed to
dm_is_dm_major() which returns true or false depending whether
the device is under devmapper's control or not.
The problem with this approach is in how the major of the device
is obtained - paths managed by devmapper are special files and
thus we want to be using st_rdev instead of st_dev to obtain the
major number. Well, that's what virIsDevMapperDevice() does
already so might as well us that.
Fixes: 01626c668e
Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=1839992
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Jiri Denemark <jdenemar@redhat.com>
When introducing virdevmapper.c (in v4.3.0-rc1~427) I didn't
realize there is a function that calls in devmapper. The function
is called virIsDevMapperDevice() and lives in virutil.c. Now that
we have a special file for handling devmapper move it there.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Jiri Denemark <jdenemar@redhat.com>
For the case where -fw_cfg uses a file, we need to set the
seclabels on it to allow QEMU the access. While QEMU allows
writing into the file (if specified on the command line), so far
we are enabling reading only and thus we can use read only label
(in case of SELinux).
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
This capability tracks whether QEMU supports -fw_cfg command line
option, more specifically whether it allows specifying filename.
There are some releases of QEMU which support -fw_cfg but not
filename. If this is ever a problem we can refine the capability
later on.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
There are recommendations and limitations to the name of the
config blobs we need to follow [1].
We don't want users to change any value only add new blobs. This
means, that the name must have "opt/" prefix and at the same time
must not begin with "opt/ovmf" nor "opt/org.qemu" as these are
reserved for OVMF or QEMU respectively.
1: docs/specs/fw_cfg.txt from qemu.git
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
QEMU has -fw_cfg which allows users to tweak how firmware
configures itself and/or provide new configuration blobs.
Introduce new <sysinfo/> type "fwcfg" that will hold these
new blobs.
It's possible to either specify new value as a string or
provide a filename which contents then serve as the value.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Setting OEM strings for a domain was introduced in
v4.1.0-rc1~315. However, any application that wanted to use them
(e.g. to point to an URL where a config file is stored) had to
'dmidecode -u --oem-string N' (where N is index of the string).
Well, we can expose them under our <sysinfo/> XML and if the
domain is running Libvirt inside it can be obtained using
virConnectGetSysinfo() API.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Since nobody sets custom dmidecode path anymore, we can drop all
code that exists only because of that.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Problem with custom dmidecode scripts is that they are hard to
modify, especially if we will want them to act differently based
on passed arguments. So far, we have two scripts which do no more
than 'cat $sysinfo' where $sysinfo is saved dmidecode output.
The virCommandSetDryRun() can be used to trick
virSysinfoReadDMI() thinking it executed real dmidecode.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
There is no real need to have two separate functions. They can be
merged together which not only saves couple of lines, but
prepares the structure of the code for future expansion. See next
commits.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Some variables defined in the function can be freed
automatically when going out of scope. This renders @result
variable and cleanup label needless.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
When trying to decode DMI table, just before constructing
virCommand() the decoder is looked for in PATH using
virFindFileInPath(). Well, this is not necessary because
virCommandRun() will do this too (in virExec()).
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Virtually every variable defined in the function can be freed
automatically when going out of scope.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
There's no need to set ctxt->node outside of the function. The
function can set it itself - it has all the info needed.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
I think that since <qemu:commandline/> is kind of a hack, it
doesn't deserve place in the front row.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
This is convenience macro, use it more. This commit was generated
using the following spatch:
@@
symbol node;
identifier old;
identifier ctxt;
type xmlNodePtr;
@@
- xmlNodePtr old;
+ VIR_XPATH_NODE_AUTORESTORE(ctxt);
...
- old = ctxt->node;
... when != old
- ctxt->node = old;
@@
symbol node;
identifier old;
identifier ctxt;
type xmlNodePtr;
@@
- xmlNodePtr old = ctxt->node;
+ VIR_XPATH_NODE_AUTORESTORE(ctxt);
... when != old
- ctxt->node = old;
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
This reverts b897973f2e
Even though it may have been the case in the past, relative
XPaths don't overwrite the ctxt->node. Thus, there's no need to
save it.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
Upon migration with disks, libvirt determines if each disk exists
on the destination and tries to pre-create missing ones. Well,
NVMe disks can't be pre-created, but they can be checked for
presence.
Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=1823639
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
Currently, the @vm is passed in as an argument and
testCompareXMLToArgvCreateArgs() is called over it which means
under the hood qemuProcessPrepareDomain() is called. But at the
point where ValidateSchema() is called, the domain object is
already 'prepared', i.e. all device aliases are assigned and so
on. But our code is not prepared to 'prepare' a domain twice - it
simply overwrites all the pointers leading to a memory leak.
Fortunately, this is only the problem of this test.
Resolve this by constructing the domain object from scratch.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Peter Krempa <pkrempa@redhat.com>
If the mirror destination is not a file but a NVMe disk, then
call qemuHostdevReAttachOneNVMeDisk() to reattach the NVMe back
to the host.
This would be done by blockjob code when the job finishes, but in
this case the job won't finish - QEMU is killed meanwhile.
Fixes: https://bugzilla.redhat.com/show_bug.cgi?id=1825785
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Peter Krempa <pkrempa@redhat.com>
In v5.10.0-rc1~42 (which was later fixed in v6.0.0-rc1~487) I am
removing XATTRs for a file that QEMU is mirroring a disk into but
it is killed meanwhile. Well, we can call
qemuSecurityRestoreImageLabel() which will not only remove XATTRs
but also use them to restore the original owner of the file.
This would be done by blockjob code when the job finishes, but in
this case the job won't finish - QEMU is killed meanwhile
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Peter Krempa <pkrempa@redhat.com>
In previous commit we started tracking whether QEMU supports
'-numa mem='. This is tied to the machine type because migration
from '-numa mem=' to '-numa memdev' is impossible (or vice
versa). But since it's tied to a machine type (where migration
from one to another is also unsupported) we can allow QEMU to get
rid of the deprecated command line.
Fixes: https://bugzilla.redhat.com/show_bug.cgi?id=1783355
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
When building -numa command line there is a for() loop that
builds '-numa memdev=' for each guest NUMA node. And also
records in a local variable whether any of memory-object-*
backends must be used to satisfy desired config. Well, instead of
checking in each iteration whether corresponding capabilities are
set, we can do swap if() and for() and check only once.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
There is 'numa-mem-supported' machine attribute which specifies
whether '-numa mem=' is supported. Store it in our capabilities
as it will be used in later commits when building the command
line.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
The AppArmor secdriver does not use labels to grant access to
resources. Therefore, it doesn't use XATTRs and hence it lacks
implementation of .domainMoveImageMetadata callback. This leads
to a harmless but needless error message appearing in the logs:
virSecurityManagerMoveImageMetadata:476 : this function is not
supported by the connection driver: virSecurityManagerMoveImageMetadata
Closes: https://gitlab.com/libvirt/libvirt/-/issues/25
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Erik Skultety <eskultet@redhat.com>
==179663== 35 (24 direct, 11 indirect) bytes in 1 blocks are definitely lost in loss record 205 of 461
==179663== at 0x4839EC6: calloc (vg_replace_malloc.c:762)
==179663== by 0x5791AC0: g_malloc0 (in /usr/lib64/libglib-2.0.so.0.6400.1)
==179663== by 0x190C79: qemuDomainObjPrivateXMLParseBlockjobDataCommit (qemu_domain.c:3295)
==179663== by 0x190DF7: qemuDomainObjPrivateXMLParseBlockjobDataSpecific (qemu_domain.c:3331)
==179663== by 0x19157D: qemuDomainObjPrivateXMLParseBlockjobData (qemu_domain.c:3469)
==179663== by 0x1918E8: qemuDomainObjPrivateXMLParseBlockjobs (qemu_domain.c:3498)
==179663== by 0x193841: qemuDomainObjPrivateXMLParse (qemu_domain.c:3944)
==179663== by 0x4A1BA9D: virDomainObjParseXML (domain_conf.c:22306)
==179663== by 0x4A1BFE9: virDomainObjParseNode (domain_conf.c:22429)
==179663== by 0x4A1C0B4: virDomainObjParseFile (domain_conf.c:22443)
==179663== by 0x1431E1: testCompareStatusXMLToXMLFiles (qemuxml2xmltest.c:61)
==179663== by 0x177722: virTestRun (testutils.c:142)
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Peter Krempa <pkrempa@redhat.com>
==156803== 58 (40 direct, 18 indirect) bytes in 1 blocks are definitely lost in loss record 306 of 463
==156803== at 0x4839EC6: calloc (vg_replace_malloc.c:762)
==156803== by 0x5791AC0: g_malloc0 (in /usr/lib64/libglib-2.0.so.0.6400.1)
==156803== by 0x48F60DC: virAlloc (viralloc.c:48)
==156803== by 0x18DD74: qemuStorageSourcePrivateDataAssignSecinfo (qemu_domain.c:2384)
==156803== by 0x18DFD5: qemuStorageSourcePrivateDataParse (qemu_domain.c:2433)
==156803== by 0x49EC884: virDomainStorageSourceParse (domain_conf.c:9857)
==156803== by 0x49ECBA3: virDomainDiskBackingStoreParse (domain_conf.c:9909)
==156803== by 0x49F129D: virDomainDiskDefParseXML (domain_conf.c:10785)
==156803== by 0x4A1804E: virDomainDefParseXML (domain_conf.c:21543)
==156803== by 0x4A1B60C: virDomainObjParseXML (domain_conf.c:22254)
==156803== by 0x4A1BFE9: virDomainObjParseNode (domain_conf.c:22429)
==156803== by 0x4A1C0B4: virDomainObjParseFile (domain_conf.c:22443
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Peter Krempa <pkrempa@redhat.com>
The @tmpIfname is a pointer into a const string. To avoid
mistakenly changing the const string via the pointer, make the
pointer const too.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
As suggested in the linked bug, libvirt should firstly check
whether the major number of the device is device mapper major.
Because if it isn't subsequent DM_DEVICE_DEPS task may not only
fail, but also yield different results. In the bugzilla this is
demonstrated by creating a devmapper target named 'loop0' and
then creating loop target /dev/loop0. When the latter is then
passed to a domain, our virDevMapperGetTargetsImpl() function
blindly asks devmapper to provide target dependencies for
/dev/loop0 and because of the way devmapper APIs work, it will
'sanitize' the input by using the last component only which is
'loop0' and thus return different results than expected.
Fixes: https://bugzilla.redhat.com/show_bug.cgi?id=1823976
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
We have a framework to register cleanup callbacks that are run
when a domain is shut down. The idea is to run callbacks in
reverse order than they were registered. However, looking at the
code this is not the case. Fortunately, this framework is used to
register a single callback and a single callback only -
qemuMigrationDstPrepareCleanup() - therefore there was no problem
just yet.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
This partially reverts fe65e9c8b5.
In the referenced commit I removed @ret from
virHostValidateBhyve() thinking it wasn't used when in fact it is
- it's usage is hidden under MODULE_STATUS_WARN(). Reintroduce
the variable back.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
In a few places we use 0 and false, or 1 and true interchangeably
even though the variable or return type in question is boolean.
Fix those places.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
There are few places where a return variable is introduced (ret
or retval), but then is never changed and is then passed to
return. Well, we can return the value that the variable is
initialized to directly.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
There are few functions that currently return an integer but in
fact they always return the same integer (zero). Make them void.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Since it's introduction in v0.9.7-147-gf4324e3292 the
virNetServerClientInitKeepAlive() function returned nothing than
a negative one. Fortunately, this did not pose any problem
because we ignored the retval happily. Well, it's time to check
for the retval because the function might fail regularly.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Instead of the following pattern:
type ret;
...
ret = func();
return ret;
we can use:
return func()
directly.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Erik Skultety <eskultet@redhat.com>
From time to time we are asked which PCI addresses are reserved
in QEMU. Let's document them in one place, it's easier than
reconstructing the list from the code each time.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Laine Stump <laine@redhat.com>
In 97a0aa2467 the @actual variable was renamed to @confactual.
However, the commit missed non-Linux case resulting in a broken
build.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
When a device is "move"-d (this basically means it was renamed),
we add the new device onto our list but keep the old there too.
Fortunately, udev sets this DEVPATH_OLD property which points to
the old device path. We can use it to remove the old instance.
To test this try renaming an interface, for instance:
# ip link set tunl0 name tunl1
# ip link set tunl1 name tunl0
One problem with udev is that it sends old ifname in INTERFACE
property, which creates a problem for us, the property is where
we get the ifname from and use it then to query all kind of info
about the interface. Well, if it is non-existent then we can't
query anything. This happens if ifname rename is suppressed
(net.ifnames=0 on kernel cmd line for instance). Fortunately, we
can use "kernel" source for udev events which has always the
fresh info.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Martin Kletzander <mkletzan@redhat.com>
Move internals of udevRemoveOneDevice() into a separate function
which accepts sysfs path as an argument and actually removes the
device from the internal list. It will be reused later.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Martin Kletzander <mkletzan@redhat.com>
When removing a node device object from the internal list the
udevRemoveOneDevice() function does plain unref over the object.
This is not sufficient. If there is another thread that's waiting
for the object lock it will wait forever.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
While I'm at it, use more g_autofree and g_autoptr() in this
file. This also fixes a possible mem-leak in
virNetDevGetVirtualFunctions().
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
I've just got a new machine and I'm still converging on the
kernel config. Anyway, since I don't have enabled any of SRIO-V
drivers, my kernel doesn't have NET_DEVLINK enabled (i.e.
virNetDevGetFamilyId() returns 0). But this makes nodedev driver
ignore all interfaces, because when enumerating all devices via
udev, the control reaches virNetDevSwitchdevFeature() eventually
and subsequently virNetDevGetFamilyId() which 'fails'. Well, it's
not really a failure - the virNetDevSwitchdevFeature() stub
simply returns 0.
Also, move the call a few lines below, just around the place
where it's needed.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
Introduced in v3.8.0-rc1~96, the virNetDevGetFamilyId() gets
netlink family ID for passed family name (even though it's used
only for getting "devlink" ID). Nevertheless, the function
returns 0 on an error or if no family ID was found. This makes it
harder for a caller to distinguish these two. Change the retval
so that a negative value is returned upon error, zero is no ID
found (but no error encountered) and a positive value is returned
on successful translation.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
As explained in the previous commit, we need to relabel the file
we are restoring the domain from. That is the FD that is passed
to QEMU. If the file is not under /dev then the file inside the
namespace is the very same as the one in the host. And regardless
of using transactions, the file will be relabeled. But, if the
file is under /dev then when using transactions only the copy
inside the namespace is relabeled and the one in the host is not.
But QEMU is reading from the one in the host, actually.
Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=1772838
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Erik Skultety <eskultet@redhat.com>
This API allows drivers to separate out handling of @stdin_path
of virSecurityManagerSetAllLabel(). The thing is, the QEMU driver
uses transactions for virSecurityManagerSetAllLabel() which
relabels devices from inside of domain's namespace. This is what
we usually want. Except when resuming domain from a file. The
file is opened before any namespace is set up and the FD is
passed to QEMU to read the migration stream from. Because of
this, the file lives outside of the namespace and if it so
happens that the file is a block device (i.e. it lives under
/dev) its copy will be created in the namespace. But the FD that
is passed to QEMU points to the original living in the host and
not in the namespace. So relabeling the file inside the namespace
helps nothing.
But if we have a separate API for relabeling the restore file
then the QEMU driver can continue calling
virSecurityManagerSetAllLabel() with transactions enabled and
call this new API without transactions.
We already have an API for relabeling a single file
(virSecurityManagerDomainSetPathLabel()) but in case of SELinux
it uses @imagelabel (which allows RW access) and we want to use
@content_context (which allows RO access).
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Erik Skultety <eskultet@redhat.com>
Previously, we used virCapabilitiesDomainDataLookup() to fill
machine type in post parse callback if none was provided in the
domain XML. If machine type couldn't be filled in an error was
reported. After 4a4132b462 we've changed it to
virQEMUCapsGetPreferredMachine() which returns NULL, but we no
longer report an error and proceed with the post parse callbacks
processing. This may lead to a crash because the code later on
assumes def->os.machine is not NULL.
Fixes: 4a4132b462
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Pavel Mores <pmores@redhat.com>
When preparing to do a blockcopy, the mirror image is modified so
that QEMU can access it. For instance, the mirror has seclabels
set, if it is a NVMe disk it is detached from the host and so on.
And usually, the restore is done upon successful finish of the
blockcopy operation. But, if something fails then we need to
explicitly revoke the access to the mirror image (and thus
reattach NVMe disk back to the host).
Fixes: https://bugzilla.redhat.com/show_bug.cgi?id=1822538
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Pavel Mores <pmores@redhat.com>
So far, libvirt generates the following path for automatic dumps:
$autoDumpPath/$id-$shortName-$timestamp
where $autoDumpPath is where libvirt stores dumps of guests (e.g.
/var/lib/libvirt/qemu/dump), $id is domain ID and $shortName is
shortened version of domain name. So for instance, the generated
path may look something like this:
/var/lib/libvirt/qemu/dump/1-QEMUGuest-2020-03-25-10:40:50
While in case of embed driver the following path would be
generated by default:
$root/lib/libvirt/qemu/dump/1-QEMUGuest-2020-03-25-10:40:50
which is not clashing with other embed drivers, we allow users to
override the default and have all embed drivers use the same
prefix. This can create clashing paths. Fortunately, we can reuse
the approach for machined name generation
(v6.1.0-178-gc9bd08ee35) and include part of hash of the root in
the generated path.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Andrea Bolognani <abologna@redhat.com>
Reviewed-by: Daniel Henrique Barboza <danielhb413@gmail.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
So far, libvirt generates the following path for memory:
$memoryBackingDir/$id-$shortName/ram-nodeN
where $memoryBackingDir is the path where QEMU mmaps() memory for
the guest (e.g. /var/lib/libvirt/qemu/ram), $id is domain ID
and $shortName is shortened version of domain name. So for
instance, the generated path may look something like this:
/var/lib/libvirt/qemu/ram/1-QEMUGuest/ram-node0
While in case of embed driver the following path would be
generated by default:
$root/lib/qemu/ram/1-QEMUGuest/ram-node0
which is not clashing with other embed drivers, we allow users to
override the default and have all embed drivers use the same
prefix. This can create clashing paths. Fortunately, we can reuse
the approach for machined name generation
(v6.1.0-178-gc9bd08ee35) and include part of hash of the root in
the generated path.
Note, the important change is in qemuGetMemoryBackingBasePath().
The rest is needed to pass driver around.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Andrea Bolognani <abologna@redhat.com>
Reviewed-by: Daniel Henrique Barboza <danielhb413@gmail.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
So far, libvirt generates the following path for hugepages:
$mnt/libvirt/qemu/$id-$shortName
where $mnt is the mount point of hugetlbfs corresponding to
hugepages of desired size (e.g. /dev/hugepages), $id is domain ID
and $shortName is shortened version of domain name. So for
instance, the generated path may look something like this:
/dev/hugepages/libvirt/qemu/1-QEMUGuest
But this won't work with embed driver really, because if there
are two instances of embed driver, and they both want to start a
domain with the same name and with hugepages, both drivers will
generate the same path which is not desired. Fortunately, we can
reuse the approach for machined name generation
(v6.1.0-178-gc9bd08ee35) and include part of hash of the root in
the generated path.
Note, the important change is in qemuGetBaseHugepagePath(). The
rest is needed to pass driver around.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Andrea Bolognani <abologna@redhat.com>
Reviewed-by: Daniel Henrique Barboza <danielhb413@gmail.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
This reverts commit 06a19921b6.
What I haven't realized when writing this ^^ commit is that the
virQEMUDriver structure already stores the root directory path.
And since the pointer is immutable it can be accessed right from
the structure and thus there is no need to duplicate it in the
driver config.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Andrea Bolognani <abologna@redhat.com>
Reviewed-by: Daniel Henrique Barboza <danielhb413@gmail.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
The code that generates "qemu-embed-$hash" is going to be useful
in more places. Separate it out into a function.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Andrea Bolognani <abologna@redhat.com>
Reviewed-by: Daniel Henrique Barboza <danielhb413@gmail.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
The virDomainGenerateMachineName() function doesn't belong in
src/conf/ really, because it has nothing to do with domain XML
parsing. It landed there because of lack of better place in the
past. But now that we have src/hypervisor/ the function should
live there. At the same time, the function name is changed to
match new location.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Andrea Bolognani <abologna@redhat.com>
Reviewed-by: Daniel Henrique Barboza <danielhb413@gmail.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Initially introduced in v3.10.0-rc1~172.
When generating a path for memory-backend-file or -mem-path, qemu
driver will use the following pattern:
$memoryBackingDir/libvirt/qemu/$id-$shortName
where $memoryBackingDir defaults to /var/lib/libvirt/qemu/ram but
can be overridden in qemu.conf. Anyway, the "/libvirt/qemu/" part
looks redundant, because it's already contained in the default,
or creates unnecessary nesting if overridden in qemu.conf.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Andrea Bolognani <abologna@redhat.com>
Introduced in v1.2.17-rc1~121, the assumption was that the
driver->privileged is immutable at the time but it might change
in the future. Well, it did not ever since. It is still immutable
variable. Drop the needless accessor then.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Andrea Bolognani <abologna@redhat.com>
The virQEMUDriverConfigNew() accepts path to root directory for
embed mode as an argument. If the argument is not NULL it uses
the passed value as prefix for some internal paths (e.g.
cfg->libDir). If it is NULL though, it looks if the other
argument, @privileged is true or false and generates internal
paths accordingly. But when calling the function from the test
suite, instead of passing NULL for @root, an empty string is
passed. Fortunately, this doesn't create a problem because in
both problematic cases the generated paths are "fixed" to point
somewhere into build dir or the code which is tested doesn't
access them.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Andrea Bolognani <abologna@redhat.com>
Reviewed-by: Daniel Henrique Barboza <danielhb413@gmail.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
While it is impossible for VIR_ALLOC() to return an error, we
should be consistent with the rest of the code and not continue
initializing the virSecurityDeviceLabelDef structure.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Even with GLib it is still possible for virQEMUCapsNew() to
return NULL because it calls virQEMUCapsInitialize() which is a
wrapper over pthread_once() which may fail. At least, we still
check for its retval. If it so happens that the virQEMUCapsNew()
fails and returns NULL, we should not dereference it.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Unfortunately, yajl_free() is not NOP on NULL. It really does
expect a valid pointer. Therefore, check whether the pointer we
want to pass to it is NULL or not.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
The last usage of plain fork() was removed in v0.9.7-rc1~50, but
we forgot to update the syntax-check exemption list accordingly.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
In the network driver code there's networkKillDaemon() which is
the same as virProcessKillPainfully(). Replace the former with
the later and drop what becomes unused function.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Now, that we know that the virtiofsd will have the pidfile open
and locked we can use virPidFileForceCleanupPath() to kill it and
unlink the pidfile.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Now, that we know that the slirp helper will have the pidfile
open and locked we can use virPidFileForceCleanupPath() to kill
it and unlink the pidfile.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Now, that our virCommandSetPidFile() is more intelligent we don't
need to rely on the daemon to create and lock the pidfile and use
virCommandSetPidFile() at the same time.
NOTE that as advertised in the previous commit, this was
temporarily broken, because both virCommand and
qemuProcessStartManagedPRDaemon() would try to lock the pidfile.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Our virCommand module allows us to set a pidfile for commands we
want to spawn. The caller constructs the string of pidfile path
and then uses virCommandSetPidFile() to tell the module to write
the pidfile once the command is ran. This usually works, but has
two flaws:
1) the child process does not hold the pidfile open & locked.
Therefore, the caller (or anybody else) can't use our fancy
virPidFileForceCleanupPath() function to kill the command
afterwards. Also, for everybody else on the system it's
needlessly harder to check if the pid from the pidfile is still
alive or not.
2) if the caller ever makes a mistake and passes the same pidfile
path for two different commands, the start of the second command
will overwrite the pidfile even though the first command might
still be running.
NOTE that this temporarily renders some command spawning
unusable, specifically those code patterns where both
virCommandSetPidFile() is used together with instructing spawned
command to acquire pidfile itself. Fortunately, there is only one
occurrence of such pattern and it is in
qemuProcessStartManagedPRDaemon(). This is fixed in next commit.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>
As a part of c799d150d5 I've introduced a test case that
tests whether passing error object between processes works. The
test spawns a child which reports a system error, parent process
then reads the error and compares with expected output. Problem
with this approach is that error message contains stringified
errno which is not portable. FreeBSD has generally different
messages than Linux. Therefore, use g_strerror() to do the errno
to string translation for us.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
In previous patches virKeyFile was replaced with its GLib
counterpart which created an incompatible change: comments can
now begin only with a number sign (#). While this won't probably
affect anyone, mention it in the release notes.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
When a QEMU process dies in the middle of a hotplug, then we fail
to restore the seclabels on the device. The problem is that if
the thread doing hotplug locks the domain object first and thus
blocks the thread that wants to do qemuProcessStop(), the
seclabel cleanup code will see vm->pid still set and mount
namespace used and therefore try to enter the namespace
represented by the PID. But the PID is gone really and thus
entering will fail and no restore is done. What we can do is to
try enter the namespace (if requested to do so) but if entering
fails, fall back to no NS mode.
Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=1814481
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Pavel Mores <pmores@redhat.com>
When running a function in a forked child, so far the only thing
we could report is exit status of the child and the error
message. However, it may be beneficial to the caller to know the
actual error that happened in the child.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Pavel Mores <pmores@redhat.com>
The @src is not always a file. It may also be a directory (for
instance qemuDomainCreateDeviceRecursive() assumes that) - even
though it doesn't happen usually. Anyway, mount() can mount only
a dir onto a dir and a file onto a file.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Pavel Mores <pmores@redhat.com>
The @devPath variable is not modifiable. It merely just points to
string containing path where private devtmpfs is being
constructed. Make it const so it doesn't look weird that it's not
freed.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Pavel Mores <pmores@redhat.com>
If building namespace fails somewhere in the middle (that is some
files exists under devMountsSavePath[i]), then plain rmdir() is
not enough to remove dir. Umount the temp location and use
virFileDeleteTree() to remove the directory.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Pavel Mores <pmores@redhat.com>
The virFileMakePathWithMode() which is our recursive version of
mkdir() fails, it simply just returns a negative value with errno
set. No error is reported (as compared to virFileTouch() for
instance).
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Pavel Mores <pmores@redhat.com>
In one of my previous commits I've introduced code that creates
all devices for given (possible) multipath target. But I've made
a mistake there - the code accesses 'next->path' without checking
if the disk source is local. Note that the 'next->path' is
NULL/doesn't make sense for VIR_STORAGE_TYPE_NVME.
Fixes: a30078cb83
Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=1814947
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
So far, when using the qemu:///embed driver, management
applications can't chose whether they want to register their
domains in machined or not. While having that option is certainly
desired, it will require more work. What we can do meanwhile is
to generate names that include part of hash of the root
directory. This is to ensure that if two applications using
different roots but the same domain name (and ID) start the
domain no clashing name for machined is generated.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Andrea Bolognani <abologna@redhat.com>
When initializing virQEMUDriverConfig structure we are given the
root directory for possible embed connection. Save it for future
use. While we could get it later from @uri member, it's not as
easy as dereferencing a pointer (virURIParse() +
virURIGetParam() + error reporting).
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Andrea Bolognani <abologna@redhat.com>
According to the linked BZ, machined expects either valid
hostname or valid FQDN (see systemd commit
v239-3092-gd65652f1f2). While in case of multiple dots, a
trailing one doesn't violate FQDN, it does violate the rule in
case of something simple, like "domain.". But it's safe to remove
it in both cases.
Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=1808499
Fixes: 45464db8ba
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
The virQEMUCaps structure has usedQMP member which in the past
used to tell if qemu we are dealing with is capable of QMP. Well,
we don't support HMP anymore (minus a few HMP passthrough
commands, which are wrapped into QMP anyways) and the member is
not used really.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Daniel Henrique Barboza <danielhb413@gmail.com>
If a disk has persistent reservations enabled, qemu-pr-helper
might open not only /dev/mapper/control but also individual
targets of the multipath device. We are already querying for them
in CGroups, but now we have to create them in the namespace too.
This was brought up in [1].
1: https://bugzilla.redhat.com/show_bug.cgi?id=1711045#c61
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Tested-by: Lin Ma <LMa@suse.com>
Reviewed-by: Jim Fehlig <jfehlig@suse.com>
Michal Privoznik