Re: [PATCH v5 net-next 9/9] octeontx2: switch: add TC flow offload path for switch flows

[Jakub Kicinski]

On Tue, 28 Jul 2026 12:45:40 +0530 Ratheesh Kannoth wrote:
> On 2026-07-28 at 02:20:59, Jakub Kicinski (kuba@xxxxxxxxxx) wrote:
> > On Fri, 24 Jul 2026 15:10:18 +0530 Ratheesh Kannoth wrote:  
> > > Register an ingress flow-table offload callback that translates TC
> > > flower rules into fl_tuple state, resolves ingress and egress
> > > pcifunc via FIB for accelerated ports, and notifies the RVU AF over
> > > the PF mailbox.  The AF forwards flow updates to switchdev and
> > > keeps per-cookie packet counters in sync using NPC MCAM multi-stats
> > > when the switch requests SWDEV2AF refresh.  
> >
> > drivers/net/ethernet/marvell/octeontx2/nic/otx2_pf.c:682:35: warning: incorrect type in argument 1 (different address spaces)
> > drivers/net/ethernet/marvell/octeontx2/nic/otx2_pf.c:682:35:    expected void volatile [noderef] __iomem *addr
> > drivers/net/ethernet/marvell/octeontx2/nic/otx2_pf.c:682:35:    got void *hwbase
> > --
> > pw-bot: cr  
> otx2_pf.c is not modified in this patch. Could you please help me understand why this warning
>  is popping up for this patch ?

Not sure, the builds in NIPA are incremental so it's possible that 
a race or some IDK random sun particle caused make to rebuild it.
Maybe you could send a quick little series to clean up all the W=1 C=1
warnings in this driver on modern compilers? Then such randomness will
not matter. Looks like there is only a handful of warnings.