Re: [PATCH] octeontx2-af: Send Link events one by one

[Subbaraya Sundeep <sbhatta@marvell.com>]

Hi Simon,

On 2025-05-12 at 10:09:54, Simon Horman (horms@kernel.org) wrote:
> On Wed, May 07, 2025 at 10:46:23PM +0530, Subbaraya Sundeep wrote:
> > Send link events one after another otherwise new message
> > is overwriting the message which is being processed by PF.
> > 
> > Fixes: a88e0f936ba9 ("octeontx2: Detect the mbox up or down message via register")
> > Signed-off-by: Subbaraya Sundeep <sbhatta@marvell.com>
> > ---
> >  drivers/net/ethernet/marvell/octeontx2/af/rvu_cgx.c | 2 ++
> >  1 file changed, 2 insertions(+)
> > 
> > diff --git a/drivers/net/ethernet/marvell/octeontx2/af/rvu_cgx.c b/drivers/net/ethernet/marvell/octeontx2/af/rvu_cgx.c
> > index 992fa0b..ebb56eb 100644
> > --- a/drivers/net/ethernet/marvell/octeontx2/af/rvu_cgx.c
> > +++ b/drivers/net/ethernet/marvell/octeontx2/af/rvu_cgx.c
> > @@ -272,6 +272,8 @@ static void cgx_notify_pfs(struct cgx_link_event *event, struct rvu *rvu)
> >  
> >  		otx2_mbox_msg_send_up(&rvu->afpf_wq_info.mbox_up, pfid);
> 
> Hi Subbaraya,
> 
> Are there other callers of otx2_mbox_msg_send_up()
> which also need this logic? If so, perhaps a helper is useful.
> If not, could you clarify why?
> 
UP messages are async notifications where we just send and forget.
There are other callers as I said we just send and forget everywhere
in the driver. Only this callsite has been modified because we have
seen an issue on customer setup where bunch of link events are queued
for a same device at one point of time.
> >  
> > +		otx2_mbox_wait_for_rsp(&rvu->afpf_wq_info.mbox_up, pfid);
> 
> This can return an error. Which is checked in otx2_sync_mbox_up_msg().
> Does it make sense to do so here too?
> 
Yes it makes sense to use otx2_sync_mbox_up_msg here. I will use it
here.

Thanks,
Sundeep
> > +
> >  		mutex_unlock(&rvu->mbox_lock);
> >  	} while (pfmap);
> >  }
> > -- 
> > 2.7.4
> >