On Tue, Jan 23, 2024 at 15:25:05 -0600, Jonathon Jongsma wrote:
> When trying to start nbdkit-backed disks in backing chains, we were
> accidentally always checking the private data of the top of the chain
> instead of using the loop variable.
>
> Signed-off-by: Jonathon Jongsma
> ---
> src/qemu
On Wed, Jan 17, 2024 at 07:04:58PM +0300, Alexandra Diupina wrote:
In commit 4af3cbafdd0e31c5c5b20d57c4aaeb19efcb98bc
(Laine Stump , Tue Jul 31 14:36:51 2012 -0400)
the return -1 was removed and since then
virNetDevVPortProfileFormat() always returns 0,
so it is possible to make this function voi
On Wed, Jan 17, 2024 at 04:13:30PM +0100, Peter Krempa wrote:
Locks in following text:
A: virNetServer
B: virNetServerClient
C: daemonClientPrivate
'virNetServerSetClientAuthenticated' locks A then B
'remoteDispatchAuthPolkit' calls 'virNetServerSetClientAuthenticated'
while holding C.
If a cl
On Tue, Jan 23, 2024 at 12:28:34PM +0300, Egor Makrushin wrote:
Division between integers will also be integer.
Thus, to preserve fractional part explicitly
convert first operand to double.
Found by Linux Verification Center (linuxtesting.org) with SVACE.
Fixes: 28d54aab05 ("examples: Introduce
On 12/8/23 11:38, Michal Privoznik wrote:
>
Polite ping.
___
Devel mailing list -- devel@lists.libvirt.org
To unsubscribe send an email to devel-le...@lists.libvirt.org
On a Friday in 2023, Michal Privoznik wrote:
There are PCI devices with pretty large non-prefetchable memory,
for instance:
Memory at 9d80 (64-bit, non-prefetchable) [size=8M]
Memory at a680 (64-bit, non-prefetchable) [size=16K]
For cold plugged devices this is not a problem, because
For q35 guests, we normally add a USB controller by default,
but there's a scenario in which we can decide to skip it. Add
test coverage for it.
Signed-off-by: Andrea Bolognani
---
...tomatic-unavailable-q35.x86_64-latest.args | 33 +++
...utomatic-unavailable-q35.x86_64-latest.x
This is a significantly expanded upon follow up to [1], and
specifically the last 5 patches in that series.
While looking at implementing the improvements suggested by Peter, I
realized that there were many additional areas in which our handling
of defaults was suboptimal, with the relevant code s
This demonstrates that on aarch64, where a native panic device
doesn't exist, it's necessary for the user to specify the model
explicitly.
Signed-off-by: Andrea Bolognani
---
.../aarch64-panic-no-model.aarch64-latest.err | 1 +
tests/qemuxmlconfdata/aarch64-panic-no-model.xml| 13
We have a few test cases that cover the ability to set
and for a guest as a side effect. Introduce an
explicit test case for the functionality.
Signed-off-by: Andrea Bolognani
---
.../title-and-description.x86_64-latest.args | 31 ++
.../title-and-description.x86_64-latest.xml
Now that we have an explicit test case for the feature, we can
drop a bunch of duplicated accidental coverage.
Signed-off-by: Andrea Bolognani
---
tests/qemuxmlconfdata/440fx-wrong-root.xml | 5 -
.../cpu-host-model-features.x86_64-latest.xml| 5 -
tests
Signed-off-by: Andrea Bolognani
---
tests/qemuxmlconfdata/minimal-no-memory.xml | 25 ---
...latest.err => no-memory.x86_64-latest.err} | 0
tests/qemuxmlconfdata/no-memory.xml | 11
tests/qemuxmlconftest.c | 2 +-
4 files changed, 12 i
We have just added a number of test cases that supersede it.
Signed-off-by: Andrea Bolognani
---
.../minimal.x86_64-latest.args| 36 -
.../qemuxmlconfdata/minimal.x86_64-latest.xml | 40 ---
tests/qemuxmlconfdata/minimal.xml | 29 --
They reference functions that have since been renamed.
Signed-off-by: Andrea Bolognani
---
src/qemu/qemu_command.c | 6 +++---
src/qemu/qemu_domain.c | 2 +-
2 files changed, 4 insertions(+), 4 deletions(-)
diff --git a/src/qemu/qemu_command.c b/src/qemu/qemu_command.c
index b8f071ff2a..4aad70
The current defaults, that can be altered on a per-architecture
basis, are derived from the historical x86 behavior.
Every time support for a new architecture is added to libvirt,
care must be taken to override these default: if that doesn't
happen, guests will end up with additional hardware, whi
Group things together where it makes sense, avoid unnecessary
uses of 'else if', plus other tweaks.
Signed-off-by: Andrea Bolognani
---
src/qemu/qemu_domain.c | 18 --
1 file changed, 8 insertions(+), 10 deletions(-)
diff --git a/src/qemu/qemu_domain.c b/src/qemu/qemu_domain.c
i
It only has a single caller.
Signed-off-by: Andrea Bolognani
---
src/qemu/qemu_alias.c | 13 +++--
src/qemu/qemu_domain_address.c | 26 --
src/qemu/qemu_domain_address.h | 3 ---
3 files changed, 11 insertions(+), 31 deletions(-)
diff --git a/src/qemu/q
It only has a single caller.
Signed-off-by: Andrea Bolognani
---
src/qemu/qemu_domain.c | 4 +++-
src/qemu/qemu_domain_address.c | 25 -
src/qemu/qemu_domain_address.h | 4
3 files changed, 3 insertions(+), 30 deletions(-)
diff --git a/src/qemu/qemu_domain
qemuDomainGetSCSIControllerModel() can return -1 on failure,
but qemuDomainFindOrCreateSCSIDiskController() didn't implement
any handling for this scenario.
Signed-off-by: Andrea Bolognani
---
src/qemu/qemu_hotplug.c | 5 +
1 file changed, 5 insertions(+)
diff --git a/src/qemu/qemu_hotplug.
Not a massive difference, but it will make upcoming changes
nicer.
Signed-off-by: Andrea Bolognani
---
src/qemu/qemu_hotplug.c | 7 +++
1 file changed, 3 insertions(+), 4 deletions(-)
diff --git a/src/qemu/qemu_hotplug.c b/src/qemu/qemu_hotplug.c
index afb720fc0b..c883fef0e9 100644
--- a/sr
We currently have a single test case called "minimal", which
suffers from two big flaws:
* it's limited to the x86_64/pc machine type;
* it explicitly enables a number of devices.
Add several test cases, one for each of the architectures and
machine types that we have good support for.
Unlik
It has nothing to do with assigning addresses, so it makes more
sense to have it in qemu_domain.
Signed-off-by: Andrea Bolognani
---
src/qemu/qemu_domain.c | 37 ++
src/qemu/qemu_domain.h | 3 +++
src/qemu/qemu_domain_address.c | 36 --
Signed-off-by: Andrea Bolognani
---
src/qemu/qemu_domain.c | 14 --
1 file changed, 12 insertions(+), 2 deletions(-)
diff --git a/src/qemu/qemu_domain.c b/src/qemu/qemu_domain.c
index 97336d5002..7475fb4f39 100644
--- a/src/qemu/qemu_domain.c
+++ b/src/qemu/qemu_domain.c
@@ -4123,16
The function is modified to be stateless, which is more
consistent with existing helpers that deal with figuring out
default models for devices, and its name needs to change
accordingly.
Signed-off-by: Andrea Bolognani
---
src/qemu/qemu_domain.c | 21 +++--
src/qemu/qemu_domain.
Extract the logic from qemuDomainControllerDefPostParse().
The code is mostly unchanged, but there's a subtle difference:
the piix3-uhci has been moved from the top of the chunk to the
bottom. This is because the original code set cont->model
directly, which made it okay to start with a suboptimal
These are similar to the minimal cases that we just introduced,
but are intended to demonstrate what device or controller model
libvirt will choose when one is not provided by the user.
Signed-off-by: Andrea Bolognani
---
...64-virt-default-models.aarch64-latest.args | 44 +++
...h64-vir
In addition to the code in qemuDomainControllerDefPostParse(),
which we have just factored into its own function, we also have
some code in qemuDomainDefAddDefaultDevices() that deals with
choosing the model for a USB controller, specifically for q35
guests. Integrate it into the newly-created func
Mostly reduce the number of 'else if' and improve comments.
Signed-off-by: Andrea Bolognani
---
src/qemu/qemu_domain.c | 62 --
1 file changed, 36 insertions(+), 26 deletions(-)
diff --git a/src/qemu/qemu_domain.c b/src/qemu/qemu_domain.c
index d992b51877
This function is tightly coupled with
qemuDomainDefaultUSBControllerModel(), as whether or not the
legacy '-usb' option is going to be used depends on the return
value of both. As such, it makes sense for them to be close to
one another.
Signed-off-by: Andrea Bolognani
---
src/qemu/qemu_command.
Currently, we have special handling for USB controllers of
s390x guests hardcoded into the command line generator. This is
not great from a layering point of view and, given the complex
interactions between the various parts, just makes things very
confusing.
In order to make things easier to reas
Factor out the existing code.
Signed-off-by: Andrea Bolognani
---
src/qemu/qemu_domain.c | 30 +-
1 file changed, 21 insertions(+), 9 deletions(-)
diff --git a/src/qemu/qemu_domain.c b/src/qemu/qemu_domain.c
index 5b93529655..b3e63b3648 100644
--- a/src/qemu/qemu_dom
Factor out the existing code.
Signed-off-by: Andrea Bolognani
---
src/qemu/qemu_domain.c | 66 ++
1 file changed, 35 insertions(+), 31 deletions(-)
diff --git a/src/qemu/qemu_domain.c b/src/qemu/qemu_domain.c
index b3e63b3648..c9d213dd7b 100644
--- a/src/
Factor out the existing code.
Signed-off-by: Andrea Bolognani
---
src/qemu/qemu_domain.c | 23 +++
1 file changed, 15 insertions(+), 8 deletions(-)
diff --git a/src/qemu/qemu_domain.c b/src/qemu/qemu_domain.c
index c9d213dd7b..8a37cda464 100644
--- a/src/qemu/qemu_domain.c
+
qemuDomainDefAddDefaultDevices() contained what is essentialy
an open-coded implementation of the helper. Get rid of it.
Signed-off-by: Andrea Bolognani
---
src/qemu/qemu_domain.c | 7 +++
1 file changed, 3 insertions(+), 4 deletions(-)
diff --git a/src/qemu/qemu_domain.c b/src/qemu/qemu_do
Follow the established naming convention for consistency.
Signed-off-by: Andrea Bolognani
---
src/qemu/qemu_domain.c | 9 ++---
1 file changed, 6 insertions(+), 3 deletions(-)
diff --git a/src/qemu/qemu_domain.c b/src/qemu/qemu_domain.c
index 62a847d47f..22980c25a9 100644
--- a/src/qemu/qem
Most of the functions responsible for choosing architecture and
machine specific defaults are already close to one another, with
just a couple of strays. Having everything in one place will
hopefully make it harder to miss updating any of the functions
when new architectures are being introduced.
Other architectures generally don't have an ISA bus, so this
default never worked for them.
Since it's now possible for a specific model not to be chosen
during postparse, something that couldn't happen until now,
we need to handle the scenario by presenting the user with a
reasonable error messag
Currently we fall back to the x86-derived default of piix3-uhci,
which is a USB1 controller that's not virtualization-friendly
and overall a terrible choice for a modern architecture. The
fact that we didn't choose a better default when RISC-V support
was introduced was an oversight which is now ad
The idea of adding devices such as USB controllers or memory
balloons by default comes from attempting to match QEMU's own
defaults at a time when x86 was the only game in town.
The unfortunate consequence of this is that, if the user does
NOT want the device in question to be present, they have t
Just like piix3-uhci for USB, the choice of lsilogic for SCSI
was dictated entirely by that being the default for legacy x86
guests. Use virtio-scsi instead.
Signed-off-by: Andrea Bolognani
---
src/qemu/qemu_domain.c| 3 ++-
...64-virt-default-models.riscv64-latest.args
On Monday, January 22 2024, I wrote:
> On Wednesday, November 22 2023, Jiri Denemark wrote:
>
>> On Thu, Nov 09, 2023 at 15:30:46 +0100, Tim Wiederhake wrote:
>>> For rationale, see patch 3 (cpu_map: No longer ignore vmx- features in
>>> sync_qemu_features_i386.py).
>>>
>>> Adding features in bun
41 matches
Mail list logo
