netdev - Re: [Intel-wired-lan] [PATCH linux-next] e1000: Remove redundant statement

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: <CAEuXFEzXSU-Ws6T_8TBVfgskh4VA14LmirFYSjdQpwtndfeeww@mail.gmail.com>
Date:   Wed, 20 Oct 2021 11:08:11 -0700
From:   Jesse Brandeburg <jesse.brandeburg@...il.com>
To:     Simon Horman <horms@...nel.org>
Cc:     luo penghao <cgel.zte@...il.com>,
        NetDEV list <netdev@...r.kernel.org>,
        Zeal Robot <zealci@....com.cn>,
        "linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>,
        intel-wired-lan@...ts.osuosl.org, Jakub Kicinski <kuba@...nel.org>,
        luo penghao <luo.penghao@....com.cn>,
        "David S . Miller" <davem@...emloft.net>
Subject: Re: [Intel-wired-lan] [PATCH linux-next] e1000: Remove redundant statement

Apologies for the duplicates, mail from my intel account going out
through outlook.com is not being delivered.

On Wed, Oct 20, 2021 at 7:00 AM Simon Horman <horms@...nel.org> wrote:

> > Value stored to 'ctrl_reg' is never read.
>
> I agree this does seem to be the case.
>
> > Reported-by: Zeal Robot <zealci@....com.cn>
> > Signed-off-by: luo penghao <luo.penghao@....com.cn>
>
> Reviewed-by: Simon Horman <horms@...nel.org>

Thanks for the review, but (davem/kuba) please do not apply.

> > @@ -1215,8 +1215,6 @@ static int e1000_integrated_phy_loopback(struct e1000_adapter *adapter)
> >               e1000_write_phy_reg(hw, PHY_CTRL, 0x8140);
> >       }
> >
> > -     ctrl_reg = er32(CTRL);

Thanks for your patch, but this change is not safe. you're removing a
read that could do two things. The first is that the read "flushes"
the write just above to PCI (it's a PCI barrier), and the second is
that the read can have some side effects.

If this change must be done, the code should be to remove the
assignment to ctrl_reg, but leave the read, so the line would just
look like:
        er32(CTRL);

This will get rid of the warning and not change the flow from the
hardware perspective.

> > -
> >       /* force 1000, set loopback */
> >       e1000_write_phy_reg(hw, PHY_CTRL, 0x4140);
> >

Please do not apply this.