On 9/13/21 11:38 AM, Peter Krempa wrote:
On Fri, Sep 10, 2021 at 11:18:20 -0400, Laine Stump wrote:
On 9/10/21 3:30 AM, Peter Krempa wrote:
On Thu, Sep 09, 2021 at 13:46:41 -0400, Laine Stump wrote:
The ACPI index of a device in a running guest can't be modified, and
libvirt doesn't actually
16.09.2021 15:57, Markus Armbruster wrote:
Vladimir Sementsov-Ogievskiy writes:
Great! Thanks for working on this!
15.09.2021 22:24, Markus Armbruster wrote:
PATCH 1+2 add feature flags to enum members. Awkward due to an
introspection design mistake; see PATCH 1 for details. Feedback
Wrap 'vshReadlineCommandGenerator' into a function with proper prototype
to provide a completer for the help command.
Signed-off-by: Peter Krempa
---
tools/vsh.c | 11 +++
1 file changed, 11 insertions(+)
diff --git a/tools/vsh.c b/tools/vsh.c
index 9c6783dad1..cf24586b25 100644
---
'--pool' of the 'pool-event' command and '--inputpool' of
'vol-create-from' use the above mentioned completer.
Signed-off-by: Peter Krempa
---
tools/virsh-pool.c | 1 +
tools/virsh-volume.c | 1 +
2 files changed, 2 insertions(+)
diff --git a/tools/virsh-pool.c b/tools/virsh-pool.c
index
For now the completion does the correct thing of completing a local path
if NULL is returned.
Introduce 'virshCompletePathLocalExisting' and use it in the
'VIRSH_COMMON_OPT_FILE' macro.
This for now serves as an annotation for the function which want to read
a file on the host running virsh. In
For now this serves just as an annotation because readline and also the
bash completion script insist on completing local paths when an empty
list is returned.
This will serve for future reference once we'll be able to properly
refuse to suggest anything.
The completer is used for fields such as
In cases such as the APIs for managed save management, the file path
provided via the '--file' option is passed to the API.
We'll need to make them distinct from cases for when virsh is using the
file so that different completers can be used.
Signed-off-by: Peter Krempa
---
'--storage' of the 'undefine' command and '--migrate-disks' of the
'migrate' command take a list of disk targets as an argument.
We can simply combine 'virshDomainDiskTargetCompleter' with
'virshCommaStringListComplete' to provide the completions.
Signed-off-by: Peter Krempa
---
When listing a snapshot tree, the '--from' option takes a name of a
snapshot to limit the subset. Use virshSnapshotNameCompleter as
completer for the option.
Signed-off-by: Peter Krempa
---
tools/virsh-snapshot.c | 1 +
1 file changed, 1 insertion(+)
diff --git a/tools/virsh-snapshot.c
Make it simple to spot which options of which commands are missing
autocompletion functions by introducing this hidden option.
In the future when we'll have completers for everything this can be also
used as a hard fail so that completers are always added.
Signed-off-by: Peter Krempa
---
We don't need to validate the real command twice, but it's better to
check that the real command name exists and it's not an alias to prevent
loops.
Signed-off-by: Peter Krempa
---
tools/vsh.c | 18 +-
1 file changed, 17 insertions(+), 1 deletion(-)
diff --git a/tools/vsh.c
Signed-off-by: Peter Krempa
---
tools/virsh-backup.c | 2 ++
tools/virsh-checkpoint.c | 1 +
tools/virsh-domain.c | 10 ++
tools/virsh-network.c| 1 +
tools/virsh-pool.c | 1 +
tools/virsh-secret.c | 1 +
tools/virsh-snapshot.c | 1 +
7 files changed, 17
Complete with the indexed targets (e.g. vda[3]) based on existing
indexes.
Signed-off-by: Peter Krempa
---
tools/virsh-completer-domain.c | 100 +
tools/virsh-completer-domain.h | 5 ++
tools/virsh-domain.c | 3 +
3 files changed, 108 insertions(+)
Include the proper header instead of duplicating the declaration.
Signed-off-by: Peter Krempa
---
tools/virsh-completer.h | 5 +
1 file changed, 1 insertion(+), 4 deletions(-)
diff --git a/tools/virsh-completer.h b/tools/virsh-completer.h
index 7edb8e2f7e..874813335d 100644
---
Prevent the need to edit the function declarations to put them into the
header. There was even inconsistent use.
Signed-off-by: Peter Krempa
---
tools/virsh-completer-checkpoint.h | 7 +-
tools/virsh-completer-domain.h | 155 +
tools/virsh-completer-host.h
When inspecting whether https://gitlab.com/libvirt/libvirt/-/issues/9 is
still valid I wrote a tool which outputs command options missing
completers. Now that I had a bit of time with lot of interruptions which
is ideal for going through such a thing I decided to clean up
the tool and post it
Introduce a proper flag 'VSH_CMD_FLAG_HIDDEN' for hiding commands from
output so that we can validate that there aren't any loops or
misconfigured commands.
Signed-off-by: Peter Krempa
---
tools/virsh.c | 3 ++-
tools/virt-admin.c | 3 ++-
tools/vsh.c| 9 ++---
tools/vsh.h
16.09.2021 13:48, Markus Armbruster wrote:
Vladimir Sementsov-Ogievskiy writes:
We'll need this for realizing qdev_find_child() in the next commit.
Signed-off-by: Vladimir Sementsov-Ogievskiy
---
softmmu/qdev-monitor.c | 48 +-
1 file changed, 33
[...]
>
> > > One thing you've missed though is that deletion of snapshots now
> becomes
> > > quite a challenge.
> > >
> >
> > Yeah I did not consider implementing deleting at that moment. However to
> > make
> > external snapshots usable it should be implemented as well.
>
> Yeah, both need to
On 9/15/21 1:07 PM, Kristina Hanicova wrote:
> This is v2 of:
> https://listman.redhat.com/archives/libvir-list/2021-August/msg00871.html
>
> diff to v1:
> - rebased onto the current master
> - changed the way of working with flags to make the code more readable
> (suggested by Jano and
On Thu, Sep 16, 2021 at 4:02 PM Michal Prívozník
wrote:
> On 9/15/21 1:07 PM, Kristina Hanicova wrote:
> > Signed-off-by: Kristina Hanicova
> > ---
> > include/libvirt/libvirt-network.h | 4
> > src/libvirt-network.c | 2 +-
> > 2 files changed, 5 insertions(+), 1 deletion(-)
16.09.2021 11:34, Markus Armbruster wrote:
Vladimir Sementsov-Ogievskiy writes:
Add a class that will unify block parents for blockdev-replace
functionality we are going to add.
Signed-off-by: Vladimir Sementsov-Ogievskiy
---
include/block/block-parent.h | 32 +
Vladimir Sementsov-Ogievskiy writes:
> 16.09.2021 15:57, Markus Armbruster wrote:
>> Vladimir Sementsov-Ogievskiy writes:
>>
>>> Great! Thanks for working on this!
>>>
>>> 15.09.2021 22:24, Markus Armbruster wrote:
PATCH 1+2 add feature flags to enum members. Awkward due to an
On 9/15/21 1:07 PM, Kristina Hanicova wrote:
> Signed-off-by: Kristina Hanicova
> ---
> include/libvirt/libvirt-network.h | 4
> src/libvirt-network.c | 2 +-
> 2 files changed, 5 insertions(+), 1 deletion(-)
>
> diff --git a/include/libvirt/libvirt-network.h
>
On 9/15/21 4:27 PM, Kristina Hanicova wrote:
> I think these functions look much more readable with just simple
> if conditions.
>
> Signed-off-by: Kristina Hanicova
> ---
> tools/virsh-pool.c | 68 --
> 1 file changed, 30 insertions(+), 38
Vladimir Sementsov-Ogievskiy writes:
> Great! Thanks for working on this!
>
> 15.09.2021 22:24, Markus Armbruster wrote:
>> PATCH 1+2 add feature flags to enum members. Awkward due to an
>> introspection design mistake; see PATCH 1 for details. Feedback
>> welcome, in particular from
The element for part of nodedev
XML is optional (see udevProcessSCSIDevice()) and as such might
not be formatted into nodedev XML (see
virNodeDeviceCapSCSIDefFormat()). Reflect this in our RNG.
Signed-off-by: Michal Privoznik
---
docs/schemas/nodedev.rng | 8 +---
1 file changed, 5
A nodedev can have 'scsi_generic' capabilities but corresponding
RNG is missing. Fortunately, it's very simple - there's only one
mandatory child element .
Signed-off-by: Michal Privoznik
---
docs/schemas/nodedev.rng | 11 +++
1 file changed, 11 insertions(+)
diff --git
The value of zero is valid (see
virNodeDeviceGetSCSIHostCaps()) but our RNG does not think so.
Switching the type to 'unsignedInt' does allow value of zero.
Signed-off-by: Michal Privoznik
---
docs/schemas/nodedev.rng | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git
These three patches address the first three problems raised in
https://bugzilla.redhat.com/show_bug.cgi?id=1999420
The fourth one is a bit different because I'm failing to find a path in
the code that would allow empty for vdpa.
Michal Prívozník (3):
schemas: Make for capscsi type
Great! Thanks for working on this!
15.09.2021 22:24, Markus Armbruster wrote:
PATCH 1+2 add feature flags to enum members. Awkward due to an
introspection design mistake; see PATCH 1 for details. Feedback
welcome, in particular from management application guys.
PATCH 3+4 implement policy
Vladimir Sementsov-Ogievskiy writes:
> We'll need this for realizing qdev_find_child() in the next commit.
>
> Signed-off-by: Vladimir Sementsov-Ogievskiy
> ---
> softmmu/qdev-monitor.c | 48 +-
> 1 file changed, 33 insertions(+), 15 deletions(-)
>
>
Vladimir Sementsov-Ogievskiy writes:
> Add a class that will unify block parents for blockdev-replace
> functionality we are going to add.
>
> Signed-off-by: Vladimir Sementsov-Ogievskiy
> ---
> include/block/block-parent.h | 32 +
> block/block-parent.c | 66
On 9/15/21 10:16 PM, Jonathon Jongsma wrote:
> Add up-to-date information about creating and defining mediated devices
> in libvirt.
>
> Signed-off-by: Jonathon Jongsma
> ---
> docs/drvnodedev.html.in | 83 ++---
> 1 file changed, 62 insertions(+), 21
When guest has NUMA nodes and QEMU is new enough to report
default RAM ID then ideally we would use -numa memdev= combined
with memory-backend-* combo becasue -mem-path/-mem-prealloc/-numa
mem are deprecated. Well, there is one problem - the .memdev=
attribute is machine type dependent (just look
This is a result of mu discussion with Jing:
https://listman.redhat.com/archives/libvir-list/2021-September/msg00407.html
Turns out, the problem has nothing to do with my virtio-mem patches so
I'm sending fix separately.
Michal Prívozník (2):
qemuxml2argvtest: Introduce another numa-topology
This test shows a bug we have: even though the XML says:
there is no -mem-prealloc nor .prealloc=yes anywhere on the cmd
line. This will be fixed in the next commit.
Signed-off-by: Michal Privoznik
---
...d-memory-numa-topology4.x86_64-latest.args | 33 +
The function should be named qemuBuildNumaCellCache (note the
missing 'd' in Build).
Signed-off-by: Michal Privoznik
---
Pushed as trivial.
src/qemu/qemu_command.c | 8
1 file changed, 4 insertions(+), 4 deletions(-)
diff --git a/src/qemu/qemu_command.c b/src/qemu/qemu_command.c
38 matches
Mail list logo