Messages in this thread | | | Date | Sat, 2 Dec 2023 10:58:45 -0800 | From | Jakub Kicinski <> | Subject | Re: [PATCH net-next v13 01/13] rtase: Add pci table supported in this module |
| |
On Sat, 2 Dec 2023 17:27:57 +0100 Andrew Lunn wrote: > > > + * The block of the Realtek RTL90xx series is our entire chip architecture, > > > + * the GMAC is connected to the switch core, and there is no PHY in between. > > > + * In addition, this driver is mainly used to control GMAC, but does not > > > + * control the switch core, so it is not the same as DSA. > > > > Okay, but you seem to only register one netdev. > > > > Which MAC is it for? > > The GMAC one. This is going to be a DSA system, and this driver is for > the conduit MAC the CPU uses. At some point, i hope there is a DSA > driver added, or the existing realtek driver is extended to support > this switch.
Oh, thanks, it even says so in the comment. I blame it on late night reviewing. I was confused by the "driver [...] does not control the switch core, so it is not the same as DSA." Looking at the discussion in v3 it sounds like the switch is controlled by a different PCI function? In which case it very much sounds like DSA. Or maybe there was a minor misunderstanding there, and the driver will need MFD / auxbus, which is still close to DSA.
I'm mainly asking to make sure we avoid "implicitly programming" the switch, like, IIUC, one of the TI drivers did. And we ended up with multiple versions / modes of operation :( Sounds like this driver doesn't touch switch registers yet, tho, so all good.
| |