> -----Original Message----- > From: Wei Liu <wei.liu@xxxxxxxxxx> > Sent: Wednesday, April 7, 2021 11:01 AM > To: Haiyang Zhang <haiyangz@xxxxxxxxxxxxx> > Cc: Wei Liu <wei.liu@xxxxxxxxxx>; Dexuan Cui <decui@xxxxxxxxxxxxx>; > davem@xxxxxxxxxxxxx; kuba@xxxxxxxxxx; KY Srinivasan > <kys@xxxxxxxxxxxxx>; Stephen Hemminger <sthemmin@xxxxxxxxxxxxx>; > Wei Liu <liuwe@xxxxxxxxxxxxx>; netdev@xxxxxxxxxxxxxxx; linux- > kernel@xxxxxxxxxxxxxxx; linux-hyperv@xxxxxxxxxxxxxxx > Subject: Re: [PATCH net-next] net: mana: Add a driver for Microsoft Azure > Network Adapter (MANA) > > On Wed, Apr 07, 2021 at 02:34:01PM +0000, Haiyang Zhang wrote: > > > > > > > -----Original Message----- > > > From: Wei Liu <wei.liu@xxxxxxxxxx> > > > Sent: Wednesday, April 7, 2021 9:17 AM > > > To: Dexuan Cui <decui@xxxxxxxxxxxxx> > > > Cc: davem@xxxxxxxxxxxxx; kuba@xxxxxxxxxx; KY Srinivasan > > > <kys@xxxxxxxxxxxxx>; Haiyang Zhang <haiyangz@xxxxxxxxxxxxx>; > Stephen > > > Hemminger <sthemmin@xxxxxxxxxxxxx>; wei.liu@xxxxxxxxxx; Wei Liu > > > <liuwe@xxxxxxxxxxxxx>; netdev@xxxxxxxxxxxxxxx; linux- > > > kernel@xxxxxxxxxxxxxxx; linux-hyperv@xxxxxxxxxxxxxxx > > > Subject: Re: [PATCH net-next] net: mana: Add a driver for Microsoft > > > Azure Network Adapter (MANA) > > > > > > On Tue, Apr 06, 2021 at 04:23:21PM -0700, Dexuan Cui wrote: > > > [...] > > > > +config MICROSOFT_MANA > > > > + tristate "Microsoft Azure Network Adapter (MANA) support" > > > > + default m > > > > + depends on PCI_MSI > > > > + select PCI_HYPERV > > > > > > OOI which part of the code requires PCI_HYPERV? > > > > > > Asking because I can't immediately find code that looks to be > > > Hyper-V specific (searching for vmbus etc). This device looks like > > > any other PCI devices to me. > > > > It depends on the VF nic's PCI config space which is presented by the > pci_hyperv driver. > > I think all it matters is the PCI bus is able to handle the configuration space > access, right? Assuming there is an emulated PCI root complex which > exposes the config space to the driver, will this driver still work? > > I'm trying to understand how tightly coupled with Hyper-V PCI this driver is. > In an alternative universe, Microsft may suddenly decide to sell this > hardware and someone wants to passthrough an VF via VFIO. I don't see > how this driver wouldn't work, hence the original question. > > There is no need to change the code. I'm just curious about a tiny detail in > the implementation. Currently, the PCI config space only comes from pci_hyperv, so we have this dependency. If the pci config space is presented from other ways in an "alternative universe", we may need to add other dependencies. And yes, the VF should continue to work:) Thanks, - Haiyang
