On 12/06/2017 05:54 PM, Dawid Zamirski wrote:
> The PHP7 variant of the macro wasn't safe if the hash key was not a
> string type. This was found when running php script with just
> libvirt_connect call under xdebug session which segfaulted. This patch
> makes the following changes:
>
> * make sur
On 06.12.2017 20:49, John Ferlan wrote:
>
>
> On 10/30/2017 07:49 AM, Nikolay Shirokovskiy wrote:
>> Deadlock scenario:
>>
>> 1. domain is being started and domain driver takes read lock
>> for nwfilter update lock thru virNWFilterReadLockFilterUpdates.
>>
>> 2. firewalld restarted and event lo
vif-* scripts support it for a long time, and expect addresses to be
separated by spaces. Add appropriate support to libxl driver.
---
changed in v2:
- moved libxlMakeIPList function to xenconfig/xen_common.c and renamed
to xenMakeIPList
---
src/libxl/libxl_conf.c | 5 +++--
src/libxl/libx
---
changed in v2:
- add tests for xenconfig driver too
---
tests/libxlxml2domconfigdata/multiple-ip.json | 37 +++-
tests/libxlxml2domconfigdata/multiple-ip.xml | 22 +-
tests/libxlxml2domconfigtest.c| 1 +-
tests/xlconfigdata/test-vif-multi-ip.cfg | 25
---
new patch in v2
---
src/xenconfig/xen_common.c | 32 +---
1 file changed, 17 insertions(+), 15 deletions(-)
diff --git a/src/xenconfig/xen_common.c b/src/xenconfig/xen_common.c
index 7f838b6..40b1483 100644
--- a/src/xenconfig/xen_common.c
+++ b/src/xenconfig/xen_c
From: Chen Hanxiao
Since commit 5e5019bf, we've no longer use
VIR_ERR_AGENT_UNSYNCED anymore.
Mark it as DEPRECATED.
Signed-off-by: Chen Hanxiao
---
include/libvirt/virterror.h | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/include/libvirt/virterror.h b/include/libvirt/vir
From: Chen Hanxiao
We lacked of hot unplugging redirdev device.
This patch add support for it.
We could use detach-device --live now.
Signed-off-by: Chen Hanxiao
---
v2: add news section
docs/news.xml | 9
src/qemu/qemu_driver.c | 4 +-
src/qemu/qemu_hotplug.c | 111
Move PCI validation checks out of qemu_command into the proper
qemu_domain validation helper.
Since there's a lot to move, we'll start slow by replicating the
pcie-root and pci-root avoidance from qemuBuildSkipController and
the first switch found in qemuBuildControllerDevStr.
Signed-off-by: John
Excluding the qemuCaps checks, move the remainder of the checks
that validate whether the PCI definition is valid or not into
qemuDomainControllerDefValidatePCI.
Signed-off-by: John Ferlan
---
src/qemu/qemu_command.c | 96 --
src/qemu/qemu_domain.c | 134 +++
Move the SATA controller check from command line building to
controller def validation. This includes copying the SATA
skip check found in qemuBuildSkipController.
Signed-off-by: John Ferlan
---
src/qemu/qemu_command.c | 6 --
src/qemu/qemu_domain.c | 25 -
2 files
Move the qemuCaps checks over to qemuDomainControllerDefValidatePCI
and add empty lines between cases so it's a bit easier on the eyes
to read.
This requires two test updates in order to set the correct capability
bit for an xml2xml test as well as setting up the similar capability
for the pseries
Move the call to qemuDomainCheckCCWS390AddressSupport from
qemuBuildControllerDevStr to qemuDomainDeviceDefValidateController.
This means we will get the qemuCaps from the driver opaque
variable passed to qemuDomainDeviceDefValidate.
Signed-off-by: John Ferlan
---
src/qemu/qemu_command.c | 4 -
Similar to the checking the modelName vs. NAME_NONE, let's make the
ModelNameTypeToString check more generic too within the checking done
in controller validation (with the same ignore certain models.
NB: We need to keep the ModelNameTypeToString fetch in command line
validation since we use it, b
From: Lin Ma
Move the IDE controller check from command line building to
controller def validation. Also explicitly include the avoidance
check for the implicit IDE controller from qemuBuildSkipController.
Cause the IDE case for command line building to generate a
failure if called to add an IDE
Move SCSI validation from qemu_command into qemu_domain.
This includes the @model reset/check when the controller
model hasn't yet been set. While at it modify the switch
to account for all virDomainControllerModelSCSI types
rather than using the default label.
Rename/reorder the args in qemuCheck
Shorten up a few characters and reference the pciopts pointer
Signed-off-by: John Ferlan
---
src/qemu/qemu_command.c | 115 +---
1 file changed, 59 insertions(+), 56 deletions(-)
diff --git a/src/qemu/qemu_command.c b/src/qemu/qemu_command.c
index e84
Move the non USB implicit controller checks into their own
helper to reduce the cruft in qemuBuildControllerDevCommandLine.
Signed-off-by: John Ferlan
---
src/qemu/qemu_command.c | 65 +++--
1 file changed, 41 insertions(+), 24 deletions(-)
diff --git
v2:
https://www.redhat.com/archives/libvir-list/2017-December/msg00173.html
Changes since v2:
* Push a few patches from v1 that were ACK'd
* Make the Skip helper private to qemu_command.c and just copied
the various skip checks from the new helper into the specific
controller validate he
Move the various modelName == NAME_NONE from the command line
generation into domain controller validation. Also rather than
have multiple cases with the same check, let's make the code
more generic, but also note that it was the modelName option
that caused the failure. We also have to be sure no
Introduce the bare bones helper to validate whether the controller
definition is valid.
Signed-off-by: John Ferlan
---
src/qemu/qemu_domain.c | 25 -
1 file changed, 24 insertions(+), 1 deletion(-)
diff --git a/src/qemu/qemu_domain.c b/src/qemu/qemu_domain.c
index 75cecb
On 12/06/2017 10:27 AM, John Ferlan wrote:
> v1:
> https://www.redhat.com/archives/libvir-list/2017-December/msg00079.html
>
> Changes since v1:
>
> * Remove USB from the pile of things to move. That simplifies the
>Skip helper to remove the @flags argument.
>
>Even though it seems t
On 12/06/2017 10:28 AM, Peter Krempa wrote:
> virStorageFileReportBrokenChain uses data from the driver private data
> pointer to print the user and group. This would lead to a crash in call
> paths where we did not initialize the storage backend as recently added
> in commit 24e47ee2b93 to qemuD
On 12/06/2017 01:06 PM, Daniel P. Berrange wrote:
> The Fedora mingw support is all merged in Fedora repos, so remove the
> outdated link.
>
> Signed-off-by: Daniel P. Berrange
> ---
> docs/windows.html.in | 4 +---
> 1 file changed, 1 insertion(+), 3 deletions(-)
>
Reviewed-by: John Ferlan
On 12/06/2017 01:05 PM, Daniel P. Berrange wrote:
> Change all links to https:// where the remote site supports it. Fix URLs for
> a few packages that moved, and delete entries which appear to be dead.
>
> Signed-off-by: Daniel P. Berrange
> ---
> docs/apps.html.in | 70
>
On 12/06/2017 12:50 PM, Daniel P. Berrange wrote:
> Currently we only describe setting the CN field for server certs. This leads
> to inevitable pain for users who set it to the fully qualified hostname and
> then use a unqualified hostname or IP address to connect in the URI. Describe
> the usag
On 12/06/2017 11:53 AM, Daniel P. Berrange wrote:
> The contents of a are a choice of exactly one union member. The
> RNG schema, however, was allowing an arbitrary number of instances of every
> union member at once.
>
> Signed-off-by: Daniel P. Berrange
> ---
> docs/schemas/nwfilter.rng | 9
On 12/06/2017 11:55 AM, Daniel P. Berrange wrote:
> Various example XML documents for arp/rarp filtering have a protocolid
> XML attribute defined. This is never parsed or output by the libvirt XML
> handling code, so shouldn't be present in example XML files either
>
> Signed-off-by: Daniel P.
The Fedora mingw support is all merged in Fedora repos, so remove the
outdated link.
Signed-off-by: Daniel P. Berrange
---
docs/windows.html.in | 4 +---
1 file changed, 1 insertion(+), 3 deletions(-)
diff --git a/docs/windows.html.in b/docs/windows.html.in
index df60940750..3f012d18da 100644
-
Change all links to https:// where the remote site supports it. Fix URLs for
a few packages that moved, and delete entries which appear to be dead.
Signed-off-by: Daniel P. Berrange
---
docs/apps.html.in | 70 ---
1 file changed, 30 insertions(
Currently we only describe setting the CN field for server certs. This leads
to inevitable pain for users who set it to the fully qualified hostname and
then use a unqualified hostname or IP address to connect in the URI. Describe
the usage of Subject Alt Name extensions, to provide multiple hostna
On 10/30/2017 07:49 AM, Nikolay Shirokovskiy wrote:
> Deadlock scenario:
>
> 1. domain is being started and domain driver takes read lock
> for nwfilter update lock thru virNWFilterReadLockFilterUpdates.
>
> 2. firewalld restarted and event loop hangs trying to take write lock
> for nwfilter up
Various example XML documents for arp/rarp filtering have a protocolid
XML attribute defined. This is never parsed or output by the libvirt XML
handling code, so shouldn't be present in example XML files either
Signed-off-by: Daniel P. Berrange
---
tests/nwfilterxml2firewalldata/arp.xml|
The PHP7 variant of the macro wasn't safe if the hash key was not a
string type. This was found when running php script with just
libvirt_connect call under xdebug session which segfaulted. This patch
makes the following changes:
* make sure that tmp_name is initialized to NULL
* set the key name
The contents of a are a choice of exactly one union member. The
RNG schema, however, was allowing an arbitrary number of instances of every
union member at once.
Signed-off-by: Daniel P. Berrange
---
docs/schemas/nwfilter.rng | 98 +--
1 file changed,
On Wed, 2017-12-06 at 11:40 -0500, Dawid Zamirski wrote:
> On Wed, 2017-12-06 at 16:21 +0100, Michal Privoznik wrote:
> > On 12/06/2017 12:17 AM, Dawid Zamirski wrote:
> > >
> > > &_info.length, &_info.index, 0, &_pos); \
> >
> > but not here. Does that work for you?
> >
>
> It does but with a
On Wed, 2017-12-06 at 16:21 +0100, Michal Privoznik wrote:
> On 12/06/2017 12:17 AM, Dawid Zamirski wrote:
> >
> > &_info.length, &_info.index, 0, &_pos); \
>
> but not here. Does that work for you?
>
It does but with a compiler warning :-) I did not realize PHP5 had
zend_ulong and I thought it
On 12/01/2017 05:50 PM, Ján Tomko wrote:
> Split out the logic of unlinking devices from
> qemuDomainNamespaceTeardownHostdev for reuse in other functions.
> ---
> src/qemu/qemu_domain.c | 55
> +++---
> 1 file changed, 39 insertions(+), 16 deletions(-)
On 12/01/2017 05:50 PM, Ján Tomko wrote:
> Rebased + rename path to paths in one more function.
>
> Ján Tomko (7):
> qemuDomainNamespaceTeardownHostdev: rename path to paths
> qemuDomainNamespaceSetupHostdev: rename path to paths
> Introduce qemuDomainNamespaceMknodPaths
> Introduce qemuDo
On Wed, Dec 06, 2017 at 10:18:53AM -0500, John Ferlan wrote:
On 12/05/2017 09:18 AM, Ján Tomko wrote:
[...]
We should not skip validation for implicit devices. Some of the checks
added later are for implicit devices (PCI root, SATA controller).
It would be enough to split out the logic of f
virStorageFileReportBrokenChain uses data from the driver private data
pointer to print the user and group. This would lead to a crash in call
paths where we did not initialize the storage backend as recently added
in commit 24e47ee2b93 to qemuDomainDetermineDiskChain.
Resolves: https://bugzilla.r
Move the SATA controller check from command line building to
controller def validation.
Signed-off-by: John Ferlan
---
src/qemu/qemu_command.c | 6 --
src/qemu/qemu_domain.c | 17 -
2 files changed, 16 insertions(+), 7 deletions(-)
diff --git a/src/qemu/qemu_command.c b/sr
Move the qemuCaps checks over to qemuDomainControllerDefValidatePCI
and add empty lines between cases so it's a bit easier on the eyes
to read.
This requires two test updates in order to set the correct capability
bit for an xml2xml test as well as setting up the similar capability
for the pseries
Similar to the checking the modelName vs. NAME_NONE, let's make the
ModelNameTypeToString check more generic too within the checking done
in controller validation (with the same ignore certain models.
NB: We need to keep the ModelNameTypeToString fetch in command line
validation since we use it, b
Move the various modelName == NAME_NONE from the command line
generation into domain controller validation. Also rather than
have multiple cases with the same check, let's make the code
more generic, but also note that it was the modelName option
that caused the failure. We also have to be sure no
From: Lin Ma
Adding an IDE controller for a machinetype that has no built-in IDE
controller, libvirt will log an error. Currently the machinetype list
which returns by qemuDomainMachineHasBuiltinIDE only includes 440fx,
malta, sun4u and g3beige.
Remove the disk and the .args file since the expec
From: Lin Ma
Adding an IDE controller for a machinetype that has no built-in IDE
controller, libvirt will log an error. Currently the machinetype list
which returns by qemuDomainMachineHasBuiltinIDE only includes 440fx,
malta, sun4u and g3beige.
Signed-off-by: Lin Ma
Signed-off-by: John Ferlan
Move the call to qemuDomainCheckCCWS390AddressSupport from
qemuBuildControllerDevStr to qemuDomainDeviceDefValidateController.
This means we will get the qemuCaps from the driver opaque
variable passed to qemuDomainDeviceDefValidate.
Signed-off-by: John Ferlan
---
src/qemu/qemu_command.c | 4 -
v1:
https://www.redhat.com/archives/libvir-list/2017-December/msg00079.html
Changes since v1:
* Remove USB from the pile of things to move. That simplifies the
Skip helper to remove the @flags argument.
Even though it seems there was some sort of objection to Skip'ing
during Validatio
Move SCSI validation from qemu_command into qemu_domain.
This includes the @model reset/check when the controller
model hasn't yet been set. While at it modify the switch
to account for all virDomainControllerModelSCSI types
rather than using the default label.
Rename/reorder the args in qemuCheck
Make sure all types of virDomainControllerType are handled in the
switch statement.
Signed-off-by: John Ferlan
---
src/qemu/qemu_command.c | 5 +++--
1 file changed, 3 insertions(+), 2 deletions(-)
diff --git a/src/qemu/qemu_command.c b/src/qemu/qemu_command.c
index d7696fb93..c5066b16b 100644
In order to be able to reuse some code for both controller def
validation and command line building, create a convenience helper
that will skip the same devices for validation that are being
skipped for command line building. This is in preference to copying
the individual checks in subsequent patc
From: Lin Ma
Move the IDE controller check from command line building to
controller def validation.
Cause the IDE case for command line building to generate a
failure if called to add an IDE since that shouldn't happen
if the Validate code did the right thing.
Signed-off-by: John Ferlan
---
s
Introduce the bare bones helper to validate whether the controller
definition is valid.
Signed-off-by: John Ferlan
---
src/qemu/qemu_domain.c | 26 +-
1 file changed, 25 insertions(+), 1 deletion(-)
diff --git a/src/qemu/qemu_domain.c b/src/qemu/qemu_domain.c
index cd213
Move PCI validation checks out of qemu_command into the proper
qemu_domain validation helper.
Since there's a lot to move, we'll start slow by replicating the
check in qemuBuildControllerDevCommandLine to skip the pcie-root
and pci-root tests for not pseries guests.
Next we'll move the controller
Excluding the qemuCaps checks, move the remainder of the checks
that validate whether the PCI definition is valid or not into
qemuDomainControllerDefValidatePCI.
Signed-off-by: John Ferlan
---
src/qemu/qemu_command.c | 96 --
src/qemu/qemu_domain.c | 134 +++
Shorten up a few characters and reference the pciopts pointer
Signed-off-by: John Ferlan
---
src/qemu/qemu_command.c | 115 +---
1 file changed, 59 insertions(+), 56 deletions(-)
diff --git a/src/qemu/qemu_command.c b/src/qemu/qemu_command.c
index 3ec
On 12/06/2017 12:17 AM, Dawid Zamirski wrote:
> The PHP7 variant of the macro wasn't safe if the hash key was not a
> string type. This was found when running php script with just
> libvirt_connect call under xdebug session which segfaulted. This patch
> makes the following changes:
>
> * make sur
On 12/05/2017 09:18 AM, Ján Tomko wrote:
> On Mon, Dec 04, 2017 at 08:38:52PM -0500, John Ferlan wrote:
>> In order to be able to reuse some code for both controller def
>> validation and command line building, create a convenience helper
>> that will perform the "skip" avoidance checks. It will
On 12/04/2017 04:21 PM, David Vrabel wrote:
> There are use cases where it is useful to use the support in libvirt
> for file-backed guest memory, but without using hugetlbfs but tmpfs
> instead (for example, to run tests on hosts that do not have hugepages
> configured, or to use Linux's idle page
On Wed, Dec 06, 2017 at 01:56:23PM +0100, Cedric Bosdonnat wrote:
> On Tue, 2017-12-05 at 16:52 +, Daniel P. Berrange wrote:
> > On Tue, Dec 05, 2017 at 10:53:17AM +0100, Cédric Bosdonnat wrote:
> > > libvirt-sandbox-init-common is expecting -d and -v parameters to
> > > set it in debug or verb
When qemuDomainFindOrCreateSCSIDiskController adds a controller,
let's use the same model as a currently found controller under the
assumption that the reason to add the controller in hotplug is
because virDomainHostdevAssignAddress determined that there were
too many devices on the existing contro
Details in the various patches.
John Ferlan (4):
qemu: Tolerate storage source private data being NULL for hotplug SCSI
hostdev
qemu: Use same model when adding hostdev SCSI controller
conf: Use existing SCSI hostdev model to create new
conf: Fix generating addresses for SCSI hostdev
Commit id 'c5c96545' neglected to validate that the srcPriv was
non-NULL before dereferencing. Similar problem to what was fixed
by commit id '8056721c' but missed during multiple rebases and
code reworks.
Signed-off-by: John Ferlan
---
src/qemu/qemu_hotplug.c | 5 +++--
1 file changed, 3 insert
In virDomainDefMaybeAddHostdevSCSIcontroller when we add a new
controller because someone neglected to add one or we're adding
one because the existing one is full, we should copy over the
model number from the existing controller since whatever we
create should at least have the same characteristi
https://bugzilla.redhat.com/show_bug.cgi?id=1519130
Commit id 'dc692438' reverted the automagic addition of a SCSI
controller attempt during virDomainHostdevAssignAddress; however,
the logic to determine where to place the next_unit depended upon
the "new" controller being added. Without the new
On Tue, 2017-12-05 at 16:52 +, Daniel P. Berrange wrote:
> On Tue, Dec 05, 2017 at 10:53:17AM +0100, Cédric Bosdonnat wrote:
> > libvirt-sandbox-init-common is expecting -d and -v parameters to
> > set it in debug or verbose mode... but those will never be passed
> > by the launcher program.
>
On Wed, Dec 06, 2017 at 01:50:56PM +0100, Cedric Bosdonnat wrote:
> On Tue, 2017-12-05 at 16:57 +, Daniel P. Berrange wrote:
> > On Tue, Dec 05, 2017 at 10:53:21AM +0100, Cédric Bosdonnat wrote:
> > > Python2 is going to die soon, convert to python3.
> >
> > I'm unclear whether this change dro
On Tue, 2017-12-05 at 16:57 +, Daniel P. Berrange wrote:
> On Tue, Dec 05, 2017 at 10:53:21AM +0100, Cédric Bosdonnat wrote:
> > Python2 is going to die soon, convert to python3.
>
> I'm unclear whether this change drops py2 support, or whether it makes it
> work with py2+3 in parallel.
>
> T
On Wed, 2017-12-06 at 10:18 +, Daniel P. Berrange wrote:
> Although the core code in libvirt-go-xml doesn't depend on libvirt, the test
> suite pulls in libvirt.git to validate XML parsing against all XML files found
> under libvirt.git/tests. We should thus have a dependancy in jenkins to
> t
On Wed, 2017-11-22 at 17:40 +0100, Pavel Hrdina wrote:
> Signed-off-by: Pavel Hrdina
> ---
> guests/vars/mappings.yml | 14 ++
> guests/vars/projects/base.yml | 4
> 2 files changed, 18 insertions(+)
I guess this is intended mostly for the 'test' flavor?
In any case,
R
Although the core code in libvirt-go-xml doesn't depend on libvirt, the test
suite pulls in libvirt.git to validate XML parsing against all XML files found
under libvirt.git/tests. We should thus have a dependancy in jenkins to trigger
rebuilds when libvirt changes.
Signed-off-by: Daniel P. Berran
We format the 'chassis' and 'port' properties on the QEMU command
line later on, so we should make sure they've been set.
Signed-off-by: Andrea Bolognani
---
src/qemu/qemu_command.c | 4 +++-
1 file changed, 3 insertions(+), 1 deletion(-)
diff --git a/src/qemu/qemu_command.c b/src/qemu/qemu_com
Most of the options are only applicable to one or two controller
types, so they should be filtered out everywhere else.
This will reduce user confusion and, in at least one corner case,
prevent guests from disappearing on daemon restart.
Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=14838
The input configurations set all existing options for all PCI
controllers, to see what ends up sticking in the output.
Signed-off-by: Andrea Bolognani
---
.../i440fx-controllers-pciopts.args| 24 +++
.../i440fx-controllers-pciopts.xml | 36 ++
.../pser
See patch 3/3 for an explanation.
Andrea Bolognani (3):
qemu: Add missing checks for pcie-root-port options
tests: Add some tests for PCI controller options
qemu: Clean up PCI controller options
src/qemu/qemu_command.c| 4 +-
src/qemu/qemu_domain.c
On Tue, Dec 05, 2017 at 16:48:36 -0500, John Ferlan wrote:
>
>
> On 12/05/2017 11:41 AM, Daniel P. Berrange wrote:
> > The unprivileged libvirtd does not support nwfilter config, by leaves the
> > driver active. It is supposed to result in all APIs being an effective
> > no-op, but several APIs r
76 matches
Mail list logo