Re: A new network interface send queue API
Jason Thorpe <[email protected]>
| Newsgroups | gmane.os.netbsd.devel.network |
|---|---|
| Message-ID | <[email protected]> |
> On Nov 12, 2023, at 10:36 PM, Kengo NAKAHARA <[email protected]> wrote: > > Hi, > > What do you think about Tx NIC multiqueue? > > Currently, drivers which support Tx NIC multiqueue avoid driver global lock > to be MP-scalable. I'm worried about sc->sc_txlock might be required in > ifp->if_transmit() driver implementation. It does not currently impact the few drivers that use Tx multi-queue. In my example, “sc_txlock” is really just a placeholder for “hardware has single Tx queue, and this is the mutex for it”. My work here has no impact on the drivers that use the driver-specific ifp->if_transmit facility at this time, although I would like to harmonize them at some point so common sets of assertions and queue manipulation routines can be used. -- thorpej