On Tue, Jan 16, 2024 at 05:39:05AM -0500, Andrea Bolognani wrote:
On Wed, Jan 10, 2024 at 11:07:47AM +0800, Xianglai Li wrote:
> @@ -2080,6 +2081,11 @@ bool virQEMUCapsHasPCIMultiBus(const virDomainDef *def)
> return true;
> }
>
> + /* loongarch64 support PCI-multibus on all machine types
> + * since forever */
> + if (ARCH_IS_LOONGARCH(def->os.arch))
> + return true;
You can fold this into one of the existing architecture checks.
Actually, I'll look into posting a patch that unifies a bunch of
them, because the way the function is written today makes it way more
complicated than it really needs to be :)
[...]
You can squash in the diff below to take care of all of the above,
but note that, just like virQEMUCapsHasPCIMultiBus(),
qemuDomainDefaultNetModel() will likely see some cleaning up soon.
Patches here:
https://lists.libvirt.org/archives/list/devel@lists.libvirt.org/thread/MR...
--
Andrea Bolognani / Red Hat / Virtualization