netdev - Re: [PATCH v5 net-next 5/5] net: ethernet: mtk_wed: add reset/reset

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <CAKgT0UeX0n0b887MWUXiO54-PBMhxgSKPTab9AX3LPk3R4fS+w@mail.gmail.com>
Date:   Thu, 12 Jan 2023 10:13:30 -0800
From:   Alexander Duyck <alexander.duyck@...il.com>
To:     Lorenzo Bianconi <lorenzo@...nel.org>
Cc:     netdev@...r.kernel.org, davem@...emloft.net, edumazet@...gle.com,
        kuba@...nel.org, pabeni@...hat.com, lorenzo.bianconi@...hat.com,
        nbd@....name, john@...ozen.org, sean.wang@...iatek.com,
        Mark-MC.Lee@...iatek.com, sujuan.chen@...iatek.com,
        daniel@...rotopia.org, leon@...nel.org
Subject: Re: [PATCH v5 net-next 5/5] net: ethernet: mtk_wed: add
 reset/reset_complete callbacks

On Thu, Jan 12, 2023 at 8:26 AM Lorenzo Bianconi <lorenzo@...nel.org> wrote:
>
> > On Wed, 2023-01-11 at 18:22 +0100, Lorenzo Bianconi wrote:
> > > Introduce reset and reset_complete wlan callback to schedule WLAN driver
> > > reset when ethernet/wed driver is resetting.
> > >
> > > Tested-by: Daniel Golle <daniel@...rotopia.org>
> > > Co-developed-by: Sujuan Chen <sujuan.chen@...iatek.com>
> > > Signed-off-by: Sujuan Chen <sujuan.chen@...iatek.com>
> > > Signed-off-by: Lorenzo Bianconi <lorenzo@...nel.org>
> > > ---
> > >  drivers/net/ethernet/mediatek/mtk_eth_soc.c |  7 ++++
> > >  drivers/net/ethernet/mediatek/mtk_wed.c     | 40 +++++++++++++++++++++
> > >  drivers/net/ethernet/mediatek/mtk_wed.h     |  8 +++++
> > >  include/linux/soc/mediatek/mtk_wed.h        |  2 ++
> > >  4 files changed, 57 insertions(+)
> > >
> >
> > Do we have any updates on the implementation that would be making use
> > of this? It looks like there was a discussion for the v2 of this set to
> > include a link to an RFC posting that would make use of this set.
>
> I posted the series to linux-wireless mailing list adding netdev one in cc:
> https://lore.kernel.org/linux-wireless/cover.1673103214.git.lorenzo@kernel.org/T/#md34b4ffcb07056794378fa4e8079458ecca69109

Thanks. It would be useful to include this link in the next revision
to make it easier to review.

> >
> > > diff --git a/drivers/net/ethernet/mediatek/mtk_eth_soc.c b/drivers/net/ethernet/mediatek/mtk_eth_soc.c
> > > index 1af74e9a6cd3..0147e98009c2 100644
> > > --- a/drivers/net/ethernet/mediatek/mtk_eth_soc.c
> > > +++ b/drivers/net/ethernet/mediatek/mtk_eth_soc.c
> > > @@ -3924,6 +3924,11 @@ static void mtk_pending_work(struct work_struct *work)
> > >     set_bit(MTK_RESETTING, &eth->state);
> > >
> > >     mtk_prepare_for_reset(eth);
> > > +   mtk_wed_fe_reset();
> > > +   /* Run again reset preliminary configuration in order to avoid any
> > > +    * possible race during FE reset since it can run releasing RTNL lock.
> > > +    */
> > > +   mtk_prepare_for_reset(eth);
> > >
> > >     /* stop all devices to make sure that dma is properly shut down */
> > >     for (i = 0; i < MTK_MAC_COUNT; i++) {
> > > @@ -3961,6 +3966,8 @@ static void mtk_pending_work(struct work_struct *work)
> > >
> > >     clear_bit(MTK_RESETTING, &eth->state);
> > >
> > > +   mtk_wed_fe_reset_complete();
> > > +
> > >     rtnl_unlock();
> > >  }
> > >
> > > diff --git a/drivers/net/ethernet/mediatek/mtk_wed.c b/drivers/net/ethernet/mediatek/mtk_wed.c
> > > index a6271449617f..4854993f2941 100644
> > > --- a/drivers/net/ethernet/mediatek/mtk_wed.c
> > > +++ b/drivers/net/ethernet/mediatek/mtk_wed.c
> > > @@ -206,6 +206,46 @@ mtk_wed_wo_reset(struct mtk_wed_device *dev)
> > >     iounmap(reg);
> > >  }
> > >
> > > +void mtk_wed_fe_reset(void)
> > > +{
> > > +   int i;
> > > +
> > > +   mutex_lock(&hw_lock);
> > > +
> > > +   for (i = 0; i < ARRAY_SIZE(hw_list); i++) {
> > > +           struct mtk_wed_hw *hw = hw_list[i];
> > > +           struct mtk_wed_device *dev = hw->wed_dev;
> > > +
> > > +           if (!dev || !dev->wlan.reset)
> > > +                   continue;
> > > +
> > > +           /* reset callback blocks until WLAN reset is completed */
> > > +           if (dev->wlan.reset(dev))
> > > +                   dev_err(dev->dev, "wlan reset failed\n");
> >
> > The reason why having the consumer would be useful are cases like this.
> > My main concern is if the error value might be useful to actually
> > expose rather than just treating it as a boolean. Usually for things
> > like this I prefer to see the result captured and if it indicates error
> > we return the error value since this could be one of several possible
> > causes for the error assuming this returns an int and not a bool.
>
> we can have 2 independent wireless chips connected here so, if the first one
> fails, should we exit or just log the error?

I would think you should log the error. I notice in your wireless
implementation you can return BUSY or TIMEOUT. Rather than doing the
dev_err in your reset function to distinguish between the two you
could just return the error and leave the printing of the error to
this dev_err message.

Also a follow-on question I had. It looks like reset_complete returns
an int but it is being ignored and in your implementation it is just
returning 0. Should that be a void instead of an int?