netdev - Re: [PATCH net-next 1/4] net: Add SRIOV VGT+ support

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]

Date:   Wed, 30 Aug 2017 21:30:23 +0300
From:   Saeed Mahameed <saeedm@....mellanox.co.il>
To:     Sabrina Dubroca <sd@...asysnail.net>
Cc:     Saeed Mahameed <saeedm@...lanox.com>,
        "David S. Miller" <davem@...emloft.net>,
        Linux Netdev List <netdev@...r.kernel.org>,
        Eugenia Emantayev <eugenia@...lanox.com>,
        Mohamad Haj Yahia <mohamad@...lanox.com>,
        Hannes Frederic Sowa <hannes@...essinduktion.org>
Subject: Re: [PATCH net-next 1/4] net: Add SRIOV VGT+ support

On Tue, Aug 29, 2017 at 7:43 PM, Sabrina Dubroca <sd@...asysnail.net> wrote:
> 2017-08-29, 13:13:09 +0300, Saeed Mahameed wrote:
>> On Mon, Aug 28, 2017 at 6:52 PM, Sabrina Dubroca <sd@...asysnail.net> wrote:
>> > 2017-08-27, 14:06:15 +0300, Saeed Mahameed wrote:
>> > [...]
>> >> +#define VF_VLAN_BITMAP       DIV_ROUND_UP(VF_VLAN_N_VID, sizeof(__u64) * BITS_PER_BYTE)
>> >> +struct ifla_vf_vlan_trunk {
>> >> +     __u32 vf;
>> >> +     __u64 allowed_vlans_8021q_bm[VF_VLAN_BITMAP];
>> >> +     __u64 allowed_vlans_8021ad_bm[VF_VLAN_BITMAP];
>> >> +};
>> >
>> > This is huge (1032B). And you put one of these in the netlink message
>> > for each VF.  This means that with 51 VF (at least in my environment,
>> > where each VF takes 1296B), you're going to overflow the u16 size of a
>> > single attribute (IFLA_VFINFO_LIST), and you cannot dump the device
>> > anymore. I'm afraid this is going to break existing setups.
>> >
>>
>> Yes ! We will fix this,
>> we are considering to report only a boolean in VFINFO which indecates
>> if VGT+ is enable or not
>> and provide a new attribute per VF to report only the vlan list of specific VF.
>
> I don't see what this is going to look like. Maybe you can describe
> more precisely what you want to add to the netlink message? (otherwise
> I'll wait for the patches)
>

We are still looking for the best way from user experience perspective.
Currently we think we will only allow to query VF vlan list on one VF
at a time, with vlan list special ip route command.

Still under definition, we will let you know once ready.

> If you add large attributes for each VF, this is still going to break.
>
>
> Thanks.
>
> --
> Sabrina