On 18.08.2016 20:42, Laine Stump wrote:
> On Aug 18, 2016 5:01 AM, "Michal Privoznik" wrote:
>>
>> On 12.08.2016 04:41, Laine Stump wrote:
>>> If you define a libvirt virtual network with one or more IP addresses,
>>> it starts up an instance of dnsmasq. It's always been possible to
>>> avoid dnsm
In order to simplify the code and make it possible to recalculate the
PCI address set, I incorporated qemuDomainValidateDevicePCISlotsChipsets
into qemuDomainPCIAddressSetCreate.
Three functions had to be moved up to be seen in
qemuDomainPCIAddressSetCreate.
---
src/qemu/qemu_domain_address.c | 1
With a parameter, we can show whether we want
qemuDomainValidateDevicePCISlotsChipsets to assign new addresses or
simply verify the existing ones and reserve slots for potential future
devices. If the parameter is false and some device is missing an address,
a warning will be displayed.
The PCI add
We can as well generate the number of buses inside that function. Fewer
parameters will be useful when recalculating the pci address set on
demand.
---
src/qemu/qemu_domain_address.c | 17 +
1 file changed, 13 insertions(+), 4 deletions(-)
diff --git a/src/qemu/qemu_domain_address
The cached pci address set is not required anymore, because the set
is now being recalculated from the domain definition on demand,
so the cache can be deleted.
---
src/qemu/qemu_domain.c | 1 -
src/qemu/qemu_domain.h | 1 -
src/qemu/qemu_domain_address.c | 14 ++
3 f
Dropping the caching of pci address set.
Instead of using the cached address set, functions in qemu_hotplug.c
now recalculate it on demand.
---
src/qemu/qemu_domain_address.c | 8 +--
src/qemu/qemu_domain_address.h | 6 +
src/qemu/qemu_hotplug.c| 50 ++
Link to the previous version:
https://www.redhat.com/archives/libvir-list/2016-August/msg00721.html
Changes in v4:
* instead of failing, qemuDomainValidateDevicePCISlotsChipsets now simply
displays a warning when a new address is being assigned during pci address
set recalculation
Tomasz Flen
On Wed, Aug 17, 2016 at 10:22:46AM +0200, Erik Skultety wrote:
On 17/08/16 09:56, Andrea Bolognani wrote:
On Tue, 2016-08-16 at 22:31 +0200, Martin Kletzander wrote:
This change means we no longer have to cast arrays of
immutable strings to arrays of mutable strings; we still
have to do the opp
On Aug 18, 2016 5:01 AM, "Michal Privoznik" wrote:
>
> On 12.08.2016 04:41, Laine Stump wrote:
> > If you define a libvirt virtual network with one or more IP addresses,
> > it starts up an instance of dnsmasq. It's always been possible to
> > avoid dnsmasq's dhcp server (simply don't include a e
17-Aug-16 16:12, Maxim Nestratov пишет:
17-Aug-16 13:43, Mikhail Feoktistov пишет:
Ack the series, except [PATCH v2 7/7] vz: relax disk bus controller and device
indices check
On 16.08.2016 15:24, Maxim Nestratov wrote:
v1-v2 changes:
- fixed "vz: reset errors after ignoring return values"
18-Aug-16 14:43, Mikhail Feoktistov пишет:
This patch fixes a bug which occurs when we check a bus and unit number
for a new attached disk. We should do this check in ValidadionCallback,
not in PostParse callback. Because in PostParse we have not initialized
disk->info.addr.drive struct yet.
Mov
18-Aug-16 17:04, Mikhail Feoktistov пишет:
Ack
On 18.08.2016 15:37, Olga Krishtal wrote:
While dettaching/attaching device in OpenStack, nova
calls vzDomainDettachDevice twice, because the update of the internal
configuration of the ct comes a bit latter than the update event.
As the result, w
Hi libvirt experts,
I am starting this email thread to discuss the potential solution / proposal of
integrating vGPU support into libvirt for QEMU.
Some quick background, NVIDIA is implementing a VFIO based mediated device
framework to allow people to virtualize their devices without SR-IOV, for
Ack
On 18.08.2016 15:37, Olga Krishtal wrote:
While dettaching/attaching device in OpenStack, nova
calls vzDomainDettachDevice twice, because the update of the internal
configuration of the ct comes a bit latter than the update event.
As the result, we suffer from the second call to dettach the
While dettaching/attaching device in OpenStack, nova
calls vzDomainDettachDevice twice, because the update of the internal
configuration of the ct comes a bit latter than the update event.
As the result, we suffer from the second call to dettach the same device.
Signed-off-by: Olga Krishtal
---
While dettaching/attaching device in OpenStack, nova
calls vzDomainDettachDevice twice, because the update of the internal
configuration of the ct comes a bit latter than the update event.
As the result, we suffer from the second call to dettach the same device.
Signed-off-by: Olga Krishtal
---
On 18.08.2016 14:57, Olga Krishtal wrote:
While dettaching/attaching device in OpenStack, nova
calls vzDomainDettachDevice twice, because the update of the internal
configuration of the ct comes a bit latter than the update event.
As the result, we suffer from the second call to dettach the same
From: Chen Hanxiao
We may overwrite existed save image by
'virsh save' command.
This patch will check qemu save headers
before trying to rewrite it.
Signed-off-by: Chen Hanxiao
---
src/qemu/qemu_driver.c | 141 +++--
1 file changed, 89 insertions(+)
Hi Andrea,
I'm using the Debian Jessie with kernel Linux u202154 4.4.0-135-arm64.
But I am not sure whether the virtio-pci module has been built into the
kernel.
not as a module.
So I am changing to Ubuntu and RHEL73. As far as I know, ubuntu Wily
has the driver of virtio-pci.
Thanks~
On Thu, Aug 18, 2016 at 01:18:59AM -0400, Laine Stump wrote:
On 08/03/2016 12:20 PM, Ján Tomko wrote:
Pick up the new syntax-check speedups.
---
I think eblake usually pushes those without review under the "gnulib
updates are immune from review | can't be effectively reviewed" rule :-)
My i
Hi Cole,
Greate ! Thanks for your help.
On 17 August 2016 at 21:18, Cole Robinson wrote:
> On 08/17/2016 07:49 AM, Kevin Zhao wrote:
> > Hi Cole,
> > Long time no see~
> > Thanks for giving help me about AArch64 of virtio-pci. You have rich
> > experience about this,
> > I am gree
This is the first implementation of fspool with directory backend(dir).
Fspool is some host directory that holds another directories (items).
In an item fs of container can be placed.
Examle and usage:
Define:
virsh -c qemu:///system fspool-define-as fs_pool_name dir --target /path/on/host
Build
v
Continuing with the effort to split output parsing and defining, these new
functions return a logging object reference instead of defining the output.
Eventually, these functions will replace the existing ones (virLogAddOutputTo*)
which will then be dropped. Also, make the new functions non-static,
On 18.08.2016 14:47, Maxim Nestratov wrote:
18-Aug-16 14:43, Mikhail Feoktistov пишет:
This patch fixes a bug which occurs when we check a bus and unit number
for a new attached disk. We should do this check in ValidadionCallback,
not in PostParse callback. Because in PostParse we have not ini
Same as for outputs, introduce a new method, that is basically the same as
virLogParseAndDefineFilter with the difference that it does not define the
filter. It rather returns a newly created object that needs to be inserted into
a list and then defined separately.
Signed-off-by: Erik Skultety
--
The reason for this is to be later able to split parsing from defining in a
convenient and transparent way, so eventually, the original virLogParse*
(virLogParseAndDefine* after this patch) functions which were named a bit
poorly will be dropped completely.
Signed-off-by: Erik Skultety
---
daemo
18-Aug-16 14:43, Mikhail Feoktistov пишет:
This patch fixes a bug which occurs when we check a bus and unit number
for a new attached disk. We should do this check in ValidadionCallback,
not in PostParse callback. Because in PostParse we have not initialized
disk->info.addr.drive struct yet.
Mov
Similar to outputs, parser should do parsing only, thus the 'define' logic
is going to be stripped from virLogParseAndDefineFilters by replacing calls to
this method to virLogSetFilters instead.
Signed-off-by: Erik Skultety
---
daemon/libvirtd.c | 2 +-
src/locking/lock_daemon.c | 2 +-
Now that we're in the critical section, syslog connection can be re-opened
by issuing openlog, which is something that cannot be done beforehand, since
syslog keeps its file descriptor private and changing the tag earlier might
introduce a log inconsistency if something went wrong with preparing a
Now that previous patches tried to refactor the code and split parsing and
defining logic of logging primitives, there is no reason why we could not keep
journald's fd within the journald output object the same way as we do for
regular file-based outputs. Quite the opposite, by doing that we gain t
This is mainly virLogAddOutputTo* which were replaced by virLogNewOutputTo* and
the previously poorly named ones virLogParseAndDefine* functions. All of these
are unnecessary now, since all the original callers were transparently switched
to the new model of separate parsing and defining logic.
Si
Prepare a method that only defines a set of outputs. It takes a list of
outputs, preferably created by virLogParseOutputs. The original set of outputs
is reset and replaced by the new user-provided set of outputs.
Signed-off-by: Erik Skultety
---
src/libvirt_private.syms | 1 +
src/util/virlog.
Another abstraction added on the top of parsing a single logging output. This
method takes and parses the whole set of outputs, adding each single output
that has already been parsed into a caller-provided array. If the user-supplied
string contained duplicate outputs, only the last occurrence is t
Since virLogParseAndDefineOutputs is going to be stripped from 'output defining'
logic, replace all relevant occurrences with virLogSetOutputs call to make the
change transparent to all original callers (daemons mostly).
Signed-off-by: Erik Skultety
---
daemon/libvirtd.c | 6 +++---
src
This method will eventually replace virLogParseAndDefineFilters which
currently does both parsing and defining.
Signed-off-by: Erik Skultety
---
src/libvirt_private.syms | 1 +
src/util/virlog.c| 31 +++
src/util/virlog.h| 1 +
3 files changed, 33 in
Outputs are a bit trickier than filters, because if the user(config)-specified
set of outputs does contain duplicates. So, not only we would log twice, but
we would also leek FD for journald, since that one is global and is overwritten
every time a journald output was specified. For compatibility r
Several methods that were introduced in previous patches were not made static
on purpose so they could be introduced without any occurrence without the
compiler complaining about not using them. Since the scope of the methods is
local to the virlog module only, they can now be turned to static.
Si
Same as with outputs. Another layer of abstraction, this provides support for
parsing a set of filters (preferred way).
Signed-off-by: Erik Skultety
---
src/libvirt_private.syms | 1 +
src/util/virlog.c| 65
src/util/virlog.h| 1
Continuing with the refactor, in order to later split output parsing and output
defining, introduce a new function which will create a new virLogOutput object
which parser will insert into a list with the list being eventually defined.
Signed-off-by: Erik Skultety
---
src/libvirt_private.syms |
This method allocated a new filter object which it then returns back to caller.
Signed-off-by: Erik Skultety
---
src/libvirt_private.syms | 2 ++
src/util/virlog.c| 46 ++
src/util/virlog.h| 3 +++
3 files changed, 51 insertions(+)
d
This API is the entry point to output modification of the logger. Currently,
everything is done by virLogParseAndDefineOutputs. Parsing and defining will be
split into two operations both handled by this method transparently.
Signed-off-by: Erik Skultety
---
src/libvirt_private.syms | 1 +
src/
Introduce a method to parse an individual logging output. The difference
compared to the virLogParseAndDefineOutput is that this method does not define
the output, instead it makes use of the virLogNewOutputTo* methods introduced
in the previous patch and just returns the virLogOutput object that h
Prepare a method that only defines a set of filters. It takes a list of
filters, preferably created by virLogParseFilters. The original set of filters
is reset and replaced by the new user-provided set of filters.
Signed-off-by: Erik Skultety
---
src/libvirt_private.syms | 1 +
src/util/virlog.
Handling of outputs and filters has been changed in a way that splits
parsing and defining. Do the same thing for logging priority as well, this
however, doesn't need much of a preparation.
---
src/util/virlog.c | 21 +
tests/eventtest.c | 3 ++-
2 files changed, 11 insertions
v2 of the original series
https://www.redhat.com/archives/libvir-list/2016-May/msg00229.html
since v1:
- as Cole pointed out in 20/38 of the original series, the patches were not
designed in an elegant way and they were hard to review, so this series reworked
the whole series:
-> first the exi
This patch fixes a bug which occurs when we check a bus and unit number
for a new attached disk. We should do this check in ValidadionCallback,
not in PostParse callback. Because in PostParse we have not initialized
disk->info.addr.drive struct yet.
Move part of code from domainPostParseCallback to
==18324== 32 bytes in 1 blocks are still reachable in loss record 41 of 114
==18324==at 0x4C2C070: calloc (vg_replace_malloc.c:623)
==18324==by 0x4EA479B: virAlloc (viralloc.c:144)
==18324==by 0x4EA674A: virBitmapNewQuiet (virbitmap.c:77)
==18324==by 0x4EA67F7: virBitmapNew (virbitm
On 12.08.2016 04:41, Laine Stump wrote:
> For some unknown reason the original implementation of the
> element only took advantage of part of the functionality in the
> dnsmasq feature it exposes - it allowed specifying the ip address of a
> DNS server which *all* DNS requests would be forwarded t
On 12.08.2016 04:41, Laine Stump wrote:
> The new forward mode 'open' is just like mode='route', except that no
> firewall rules are added to assure that any traffic does or doesn't
> pass. It is assumed that either they aren't necessary, or they will be
> setup outside the scope of libvirt.
>
> R
On 12.08.2016 04:41, Laine Stump wrote:
> If you define a libvirt virtual network with one or more IP addresses,
> it starts up an instance of dnsmasq. It's always been possible to
> avoid dnsmasq's dhcp server (simply don't include a element),
> but until now it wasn't possible to avoid having th
On Wed, 2016-08-17 at 17:04 -0400, John Ferlan wrote:
> > @@ -6999,17 +7000,22 @@ qemuBuildMachineCommandLine(virCommandPtr cmd,
> >virTristateSwitchTypeToString(vmport));
> > }
> >
> > -if (def->mem.dump_core) {
> > -if (!virQEMUCapsGe
51 matches
Mail list logo