Date:   Thu, 16 Mar 2023 11:12:04 +0100
From:   Michal Swiatkowski <michal.swiatkowski@linux.intel.com>
To:     Daniil Tatianin <d-tatianin@yandex-team.ru>
Cc:     Ariel Elior <aelior@marvell.com>,
        Manish Chopra <manishc@marvell.com>,
        "David S. Miller" <davem@davemloft.net>,
        Eric Dumazet <edumazet@google.com>,
        Jakub Kicinski <kuba@kernel.org>,
        Paolo Abeni <pabeni@redhat.com>,
        Yuval Mintz <Yuval.Mintz@qlogic.com>, netdev@vger.kernel.org,
        linux-kernel@vger.kernel.org
Subject: Re: [PATCH] qed/qed_sriov: avoid a possible NULL deref in
 configure_min_tx_rate
Message-ID: <ZBLrb6C1mEjgAGHr@localhost.localdomain>
References: <20230315194809.579756-1-d-tatianin@yandex-team.ru>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <20230315194809.579756-1-d-tatianin@yandex-team.ru>
Precedence: bulk

On Wed, Mar 15, 2023 at 10:48:09PM +0300, Daniil Tatianin wrote:
> We have to make sure that the info returned by qed_iov_get_vf_info is
> valid before using it.
> 
> Found by Linux Verification Center (linuxtesting.org) with the SVACE
> static analysis tool.
> 
> Fixes: 733def6a04bf ("qed*: IOV link control")
> Signed-off-by: Daniil Tatianin <d-tatianin@yandex-team.ru>
> ---
>  drivers/net/ethernet/qlogic/qed/qed_sriov.c | 3 +++
>  1 file changed, 3 insertions(+)
> 
> diff --git a/drivers/net/ethernet/qlogic/qed/qed_sriov.c b/drivers/net/ethernet/qlogic/qed/qed_sriov.c
> index 2bf18748581d..cd43f1b23eb1 100644
> --- a/drivers/net/ethernet/qlogic/qed/qed_sriov.c
> +++ b/drivers/net/ethernet/qlogic/qed/qed_sriov.c
> @@ -4404,6 +4404,9 @@ qed_iov_configure_min_tx_rate(struct qed_dev *cdev, int vfid, u32 rate)
>  	}
>  
>  	vf = qed_iov_get_vf_info(QED_LEADING_HWFN(cdev), (u16)vfid, true);
> +	if (!vf)
> +		return -EINVAL;
> +
>  	vport_id = vf->vport_id;
>  
>  	return qed_configure_vport_wfq(cdev, vport_id, rate);
> -- 
> 2.25.1
> 

There is also potential NULL pointer dereference in:
qed_iov_handle_trust_change()
Should be:
if (!vf || !vf->vf->vport_instance)

I think it can be a part of this fix.

Thanks,
Michal