Re: [PATCH net] net: fec: avoid lock evasion when reading pps_enable

[Eric Dumazet]

On Mon, May 13, 2024 at 9:53 AM Wei Fang <wei.fang@xxxxxxx> wrote:
>
> > -----Original Message-----
> > From: Eric Dumazet <edumazet@xxxxxxxxxx>
> > Sent: 2024年5月13日 15:29
> > To: Wei Fang <wei.fang@xxxxxxx>
> > Cc: davem@xxxxxxxxxxxxx; kuba@xxxxxxxxxx; pabeni@xxxxxxxxxx; Shenwei
> > Wang <shenwei.wang@xxxxxxx>; Clark Wang <xiaoning.wang@xxxxxxx>;
> > richardcochran@xxxxxxxxx; andrew@xxxxxxx; netdev@xxxxxxxxxxxxxxx;
> > linux-kernel@xxxxxxxxxxxxxxx; imx@xxxxxxxxxxxxxxx
> > Subject: Re: [PATCH net] net: fec: avoid lock evasion when reading pps_enable
> >
> > On Mon, May 13, 2024 at 4:02 AM Wei Fang <wei.fang@xxxxxxx> wrote:
> > >
> > > The assignment of pps_enable is protected by tmreg_lock, but the read
> > > operation of pps_enable is not. So the Coverity tool reports a lock
> > > evasion warning which may cause data race to occur when running in a
> > > multithread environment. Although this issue is almost impossible to
> > > occur, we'd better fix it, at least it seems more logically
> > > reasonable, and it also prevents Coverity from continuing to issue warnings.
> > >
> > > Fixes: 278d24047891 ("net: fec: ptp: Enable PPS output based on ptp
> > > clock")
> > > Signed-off-by: Wei Fang <wei.fang@xxxxxxx>
> > > ---
> > >  drivers/net/ethernet/freescale/fec_ptp.c | 8 +++++---
> > >  1 file changed, 5 insertions(+), 3 deletions(-)
> > >
> > > diff --git a/drivers/net/ethernet/freescale/fec_ptp.c
> > > b/drivers/net/ethernet/freescale/fec_ptp.c
> > > index 181d9bfbee22..8d37274a3fb0 100644
> > > --- a/drivers/net/ethernet/freescale/fec_ptp.c
> > > +++ b/drivers/net/ethernet/freescale/fec_ptp.c
> > > @@ -104,14 +104,16 @@ static int fec_ptp_enable_pps(struct
> > fec_enet_private *fep, uint enable)
> > >         struct timespec64 ts;
> > >         u64 ns;
> > >
> > > -       if (fep->pps_enable == enable)
> > > -               return 0;
> > > -
> > >         fep->pps_channel = DEFAULT_PPS_CHANNEL;
> > >         fep->reload_period = PPS_OUPUT_RELOAD_PERIOD;
> >
> > Why are these writes left without the spinlock protection ?
> For fec driver, the pps_channel and the reload_period of PPS request
> are always fixed, and they were also not protected by the lock in the
> original code.

If this is the case, please move this initialization elsewhere, so
that we can be absolutely sure of the  claim.

I see fep->reload_period being overwritten in this file, I do not see
clear evidence this is all safe.