On Thu, Jan 25, 2024 at 01:07:10PM +0100, Peter Krempa wrote:
On Wed, Jan 24, 2024 at 20:37:35 +0100, Andrea Bolognani wrote:
> @@ -876,10 +876,9 @@ qemuDomainFindOrCreateSCSIDiskController(virDomainObj *vm,
> cont = g_new0(virDomainControllerDef, 1);
> cont->type = VIR_DOMAIN_CONTROLLER_TYPE_SCSI;
> cont->idx = controller;
> - if (model == VIR_DOMAIN_CONTROLLER_MODEL_SCSI_DEFAULT)
> - cont->model = qemuDomainGetSCSIControllerModel(vm->def, cont,
priv->qemuCaps);
> - else
> - cont->model = model;
> + cont->model = model;
> +
> + cont->model = qemuDomainGetSCSIControllerModel(vm->def, cont,
priv->qemuCaps);
This makes no sense at first look. Commit message doesn't explain it,
nothing in the code explains it. You have to look into
qemuDomainGetSCSIControllerModel to see why it can be done.
I've looked ahead to see how you've refactored it, at which point it
makes sense, but I'm not really persuaded that this can't be part of the
commit that is actually fixing the function to not depend on the value
in the 'model' struct.
IMO this patch makes the code strictly worse, even when it's short
lived.
That's why I've justified in the commit message as being something
that would pay off later on ;)
I can roll it into the later patch. I think I had it there initially,
but it seemed that it would make reviewing the change harder. Since
you're the one reviewing, I'm happy to go with your preference.
--
Andrea Bolognani / Red Hat / Virtualization