Re: [net-next PATCH 1/3] net: introduce napi_is_scheduled helper

From: Eric Dumazet
Date: Mon Oct 02 2023 - 08:56:43 EST


On Mon, Oct 2, 2023 at 2:55 PM Christian Marangi <ansuelsmth@xxxxxxxxx> wrote:
>
> On Mon, Oct 02, 2023 at 02:49:11PM +0200, Eric Dumazet wrote:
> > On Mon, Oct 2, 2023 at 2:43 PM Christian Marangi <ansuelsmth@xxxxxxxxx> wrote:
> > >
> > > On Mon, Oct 02, 2023 at 02:35:22PM +0200, Eric Dumazet wrote:
> > > > On Mon, Oct 2, 2023 at 2:29 PM Christian Marangi <ansuelsmth@xxxxxxxxx> wrote:
> > > >
> > > > > Ehhh the idea here was to reduce code duplication since the very same
> > > > > test will be done in stmmac. So I guess this code cleanup is a NACK and
> > > > > I have to duplicate the test in the stmmac driver.
> > > >
> > > > I simply wanted to add a comment in front of this function/helper,
> > > > advising not using it unless absolutely needed.
> > > >
> > > > Thus my question "In which context is it safe to call this helper ?"
> > > >
> > > > As long as it was private with a driver, I did not mind.
> > > >
> > > > But if made public in include/linux/netdevice.h, I would rather not
> > > > have to explain
> > > > to future users why it can be problematic.
> > >
> > > Oh ok!
> > >
> > > We have plenty of case similar to this. (example some clock API very
> > > internal that should not be used normally or regmap related)
> > >
> > > I will include some comments warning that this should not be used in
> > > normal circumstances and other warnings. If you have suggestion on what
> > > to add feel free to write them.
> > >
> > > Any clue on how to proceed with the sge driver?
> > >
> >
> > I would remove use of this helper for something with no race ?
> >
> > Feel free to submit this :
> >
> > (Alternative would be to change napi_schedule() to return a boolean)
> >
>
> Think mod napi_schedule() to return a bool would result in massive
> warning (actually error with werror) with return value not handled.
>

It should not, unless we added a __must_check

> I will submit with your Suggested-by. Ok for you?

Absolutely, thanks.

>
> > diff --git a/drivers/net/ethernet/chelsio/cxgb3/sge.c
> > b/drivers/net/ethernet/chelsio/cxgb3/sge.c
> > index 2e9a74fe0970df333226b80af8716f30865c01b7..09d0e6aa4db982e3488e0c28bed33e83453801d0
> > 100644
> > --- a/drivers/net/ethernet/chelsio/cxgb3/sge.c
> > +++ b/drivers/net/ethernet/chelsio/cxgb3/sge.c
> > @@ -2501,14 +2501,6 @@ static int napi_rx_handler(struct napi_struct
> > *napi, int budget)
> > return work_done;
> > }
> >
> > -/*
> > - * Returns true if the device is already scheduled for polling.
> > - */
> > -static inline int napi_is_scheduled(struct napi_struct *napi)
> > -{
> > - return test_bit(NAPI_STATE_SCHED, &napi->state);
> > -}
> > -
> > /**
> > * process_pure_responses - process pure responses from a response queue
> > * @adap: the adapter
> > @@ -2674,9 +2666,9 @@ static int rspq_check_napi(struct sge_qset *qs)
> > {
> > struct sge_rspq *q = &qs->rspq;
> >
> > - if (!napi_is_scheduled(&qs->napi) &&
> > - is_new_response(&q->desc[q->cidx], q)) {
> > - napi_schedule(&qs->napi);
> > + if (is_new_response(&q->desc[q->cidx], q) &&
> > + napi_schedule_prep(&qs->napi)) {
> > + __napi_schedule(&qs->napi);
> > return 1;
> > }
> > return 0;
>
> --
> Ansuel