[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <CAEP_g=-dwrhuJddX9Jq1yLuUrH0XKwuEdknOHipdZx8dB64xdg@mail.gmail.com>
Date: Wed, 10 Aug 2011 08:16:20 +0800
From: Jesse Gross <jesse@...ira.com>
To: Shreyas Bhatewara <sbhatewara@...are.com>
Cc: David Miller <davem@...emloft.net>,
"netdev@...r.kernel.org" <netdev@...r.kernel.org>,
Scott Goldman <scottjg@...are.com>,
VMware PV-Drivers <pv-drivers@...are.com>
Subject: Re: [PATCH net] vmxnet3: Don't enable vlan filters in promiscuous mode.
On Wed, Aug 10, 2011 at 5:32 AM, Shreyas Bhatewara
<sbhatewara@...are.com> wrote:
>
>
> On Mon, 8 Aug 2011, Jesse Gross wrote:
>
>> @@ -1929,14 +1929,17 @@ static void
>> vmxnet3_vlan_rx_add_vid(struct net_device *netdev, u16 vid)
>> {
>> struct vmxnet3_adapter *adapter = netdev_priv(netdev);
>> - u32 *vfTable = adapter->shared->devRead.rxFilterConf.vfTable;
>> - unsigned long flags;
>>
>> - VMXNET3_SET_VFTABLE_ENTRY(vfTable, vid);
>> - spin_lock_irqsave(&adapter->cmd_lock, flags);
>> - VMXNET3_WRITE_BAR1_REG(adapter, VMXNET3_REG_CMD,
>> - VMXNET3_CMD_UPDATE_VLAN_FILTERS);
>> - spin_unlock_irqrestore(&adapter->cmd_lock, flags);
>> + if (!(netdev->flags & IFF_PROMISC)) {
>> + u32 *vfTable = adapter->shared->devRead.rxFilterConf.vfTable;
>> + unsigned long flags;
>> +
>> + VMXNET3_SET_VFTABLE_ENTRY(vfTable, vid);
>> + spin_lock_irqsave(&adapter->cmd_lock, flags);
>> + VMXNET3_WRITE_BAR1_REG(adapter, VMXNET3_REG_CMD,
>> + VMXNET3_CMD_UPDATE_VLAN_FILTERS);
>> + spin_unlock_irqrestore(&adapter->cmd_lock, flags);
>> + }
>>
>
> If this is done, the driver will ignore all vlan tag registrations (and
> deletions) while the interface is in promiscuous mode. Better solution
> would be to send UPDATE_VLAN_FILTERS command alone inside the promiscuous
> condition. vfTable can be set/unset unconditionally as before. By doing
> this, when the interface comes out of promiscuous mode, the restored vlan
> state will have all the added/removed vlan tags into effect.
Adds and removes are not ignored when in promiscuous mode because the
active_vlans bitfield is still being updated (this is in the context
that you clipped out). When we come out of promiscuous mode,
vmxnet3_restore_vlan() is called which will update the hardware with
the vlans that have been registered. This is much safer than directly
updating vfTable because it will get overwritten if vmxnet3_set_mc()
is called again and is potentially very error prone to have it not
reflect what we actually want programmed in the hardware.
--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Powered by blists - more mailing lists