Re: [libvirt] [PATCH v2 3/9] hostdev: Use common reattach code in virHostdevPCINodeDeviceReAttach()
On 01/27/2016 12:39 PM, Andrea Bolognani wrote:
On Tue, 2016-01-26 at 18:56 -0500, John Ferlan wrote:
>> +/**
>> + * virHostdevPCINodeDeviceReAttach:
>
> ^^ Oy ReAttach vs. Reattach is an eye test ;-)
Maybe we should standardize on either one or the other? I personally
consider "ReAttach" to be quite an eyesore, but then again it's all
over the public API so it's not going anywhere...
Sadly, I think we're stuck with that CaMel case because it's a driver
function...
>> + * @hostdev_mgr: hostdev manager
>> + * @pci: PCI device
>
> Perhaps better to indicate a "new"ly generated PCI device that does not
> track the internal reattach states and other state information such as
> the stub driver.
>
> IOW: this is not a copy of an [in]activePCIHostdevs element
Great idea! Maybe even use a different name for the parameter, based
on whether the virPCIDevicePtr is going to be used for something other
than looking up the actual device?
This is I believe where I went back to patch 1 and started thinking
about what is passed in 'pci'... Anything to help make things more
obvious could be beneficial, especially considering what this code ends
up doing...
John
Cheers.
--
Andrea Bolognani
Software Engineer - Virtualization Team