netdev - Re: [RFC PATCH net-next] net: ethernet: mediatek: support setting MTU

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 for Android: free password hash cracker in your pocket

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

Message-ID: <CALW65jaPV7sH8v3-K4X8gnzKTymxF=5dnatuKEuesDf64YKa=A@mail.gmail.com>
Date:   Fri, 30 Oct 2020 10:43:51 +0800
From:   DENG Qingfang <dqfext@...il.com>
To:     Andrew Lunn <andrew@...n.ch>
Cc:     netdev <netdev@...r.kernel.org>, Felix Fietkau <nbd@....name>,
        John Crispin <john@...ozen.org>,
        Sean Wang <sean.wang@...iatek.com>,
        Mark Lee <Mark-MC.Lee@...iatek.com>,
        Landen Chao <landen.chao@...iatek.com>,
        Jakub Kicinski <kuba@...nel.org>,
        "David S . Miller" <davem@...emloft.net>,
        Matthias Brugger <matthias.bgg@...il.com>,
        René van Dorst <opensource@...rst.com>,
        David Woodhouse <dwmw2@...radead.org>,
        Frank Wunderlich <frank-w@...lic-files.de>
Subject: Re: [RFC PATCH net-next] net: ethernet: mediatek: support setting MTU

On Thu, Oct 29, 2020 at 9:02 PM Andrew Lunn <andrew@...n.ch> wrote:
>
> On Thu, Oct 29, 2020 at 02:39:15PM +0800, DENG Qingfang wrote:
> > MT762x HW supports frame length up to 2048 (maximum length on GDM),
> > so allow setting MTU up to 2030.
> >
> > Signed-off-by: DENG Qingfang <dqfext@...il.com>
> > ---
> >
> > I only tested this on MT7621, no sure if it is applicable for other SoCs
> > especially MT7628, which has an old IP.
> >
> > ---
> >  drivers/net/ethernet/mediatek/mtk_eth_soc.c | 31 ++++++++++++++++++++-
> >  drivers/net/ethernet/mediatek/mtk_eth_soc.h | 11 ++++++--
> >  2 files changed, 38 insertions(+), 4 deletions(-)
> >
> > diff --git a/drivers/net/ethernet/mediatek/mtk_eth_soc.c b/drivers/net/ethernet/mediatek/mtk_eth_soc.c
> > index 6d2d60675ffd..a0c56d9be1d5 100644
> > --- a/drivers/net/ethernet/mediatek/mtk_eth_soc.c
> > +++ b/drivers/net/ethernet/mediatek/mtk_eth_soc.c
> > @@ -353,7 +353,7 @@ static void mtk_mac_config(struct phylink_config *config, unsigned int mode,
> >       /* Setup gmac */
> >       mcr_cur = mtk_r32(mac->hw, MTK_MAC_MCR(mac->id));
> >       mcr_new = mcr_cur;
> > -     mcr_new |= MAC_MCR_MAX_RX_1536 | MAC_MCR_IPG_CFG | MAC_MCR_FORCE_MODE |
> > +     mcr_new |= MAC_MCR_IPG_CFG | MAC_MCR_FORCE_MODE |
> >                  MAC_MCR_BACKOFF_EN | MAC_MCR_BACKPR_EN | MAC_MCR_FORCE_LINK;
>
> Since you no longer set MAC_MCR_MAX_RX_1536, what does the hardware
> default to?

2'b00, i.e. 1518 bytes.

>
> >       /* Only update control register when needed! */
> > @@ -2499,6 +2499,34 @@ static void mtk_uninit(struct net_device *dev)
> >       mtk_rx_irq_disable(eth, ~0);
> >  }
> >
> > +static int mtk_change_mtu(struct net_device *dev, int new_mtu)
> > +{
> > +     struct mtk_mac *mac = netdev_priv(dev);
> > +     u32 mcr_cur, mcr_new;
> > +     int length;
> > +
> > +     mcr_cur = mtk_r32(mac->hw, MTK_MAC_MCR(mac->id));
> > +     mcr_new = mcr_cur & ~MAC_MCR_MAX_RX_LEN_MASK;
> > +     length = new_mtu + MTK_RX_ETH_HLEN;
> > +
> > +     if (length <= 1518)
> > +             mcr_new |= MAC_MCR_MAX_RX_LEN(MAC_MCR_MAX_RX_LEN_1518);
> > +     else if (length <= 1536)
> > +             mcr_new |= MAC_MCR_MAX_RX_LEN(MAC_MCR_MAX_RX_LEN_1536);
> > +     else if (length <= 1552)
> > +             mcr_new |= MAC_MCR_MAX_RX_LEN(MAC_MCR_MAX_RX_LEN_1552);
> > +     else
> > +             mcr_new |= MAC_MCR_MAX_RX_LEN(MAC_MCR_MAX_RX_LEN_2048);
>
> You should have another if here, and return -EIVAL is the user asked
> for an MTU of 2049 of greater.

It is handled in net_device->max_mtu:
> eth->netdev[id]->max_mtu = MTK_MAX_RX_LENGTH - MTK_RX_ETH_HLEN;

As I tested:

> # ip link set eth0 mtu 2031
> RTNETLINK answers: Invalid argument

>
>     Andrew