netdev - RE: [PATCH net 1/1] net: ethtool: mm: Allow Verify Enabled before Tx Enabled

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] [day] [month] [year] [list]

Message-ID: <SA1PR11MB5827C20FB9D7E7DE38148E26DA1A2@SA1PR11MB5827.namprd11.prod.outlook.com>
Date: Thu, 16 Jan 2025 07:00:32 +0000
From: "Choong, Chwee Lin" <chwee.lin.choong@...el.com>
To: Vladimir Oltean <vladimir.oltean@....com>, Furong Xu <0x1207@...il.com>
CC: "David S . Miller" <davem@...emloft.net>, Eric Dumazet
	<edumazet@...gle.com>, Jakub Kicinski <kuba@...nel.org>, Paolo Abeni
	<pabeni@...hat.com>, Simon Horman <horms@...nel.org>,
	"netdev@...r.kernel.org" <netdev@...r.kernel.org>,
	"linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>
Subject: RE: [PATCH net 1/1] net: ethtool: mm: Allow Verify Enabled before Tx
 Enabled


On Wednesday, January 15, 2025 5:49 PM, Vladimir Oltean <vladimir.oltean@....com> wrote:
>Thanks for copying me, Furong.

Hi Vladimir,

>On Wed, Jan 15, 2025 at 05:42:04PM +0800, Furong Xu wrote:
>> On Wed, 15 Jan 2025 14:59:31 +0800, Chwee-Lin Choong
><chwee.lin.choong@...el.com> wrote:
>>
>> > The current implementation of ethtool --set-mm restricts enabling
>> > the "verify_enabled" flag unless Tx preemption
>> > (tx_enabled) is already enabled. By default, verification is
>> > disabled, and enabling Tx preemption immediately activates
>> > preemption.
>> >
>> > When verification is intended, users can only enable verification
>> > after enabling tx_enabled, which temporarily deactivates preemption
>> > until verification completes. This creates an inconsistent and
>> > restrictive workflow.
>
>Where the premise of the patch is wrong is here. Users don't have to enable
>verification _after_ enabling TX. They can also enable verification _at the same
>time_ as TX, aka within the same netlink message. They just can't enable TX
>verification while TX in general is disabled. It just doesn't make sense.
>
The intent of this patch is tied to the earlier discussion about avoiding LLDP operations 
that forcefully enable verification whenever preemption support is advertised by the link partner.

Since we are currently seeking clarification from Avnu on their test requirements regarding 
scenarios with LLDP + verification disabled, I’ll provide an update once we hear back from Avnu.

>> > This patch modifies ethtool --set-mm to allow users to pre-enable
>> > verification locally using ethtool before Tx preemption is enabled
>> > via ethtool or negotiated through LLDP with a link partner.
>> >
>> > Current Workflow:
>> > 1. Enable pmac_enabled → Preemption supported 2. Enable tx_enabled →
>> > Preemption Tx enabled 3. verify_enabled defaults to off → Preemption
>> > active 4. Enable verify_enabled → Preemption deactivates →
>> > Verification starts
>> >                          → Verification success → Preemption active.
>> >
>> > Proposed Workflow:
>> > 1. Enable pmac_enabled → Preemption supported 2. Enable
>> > verify_enabled → Preemption supported and Verify enabled 3. Enable
>> > tx_enabled → Preemption Tx enabled → Verification starts
>> >                      → Verification success → Preemption active.
>> >
>>
>> Maybe you misunderstand the parameters of ethtool --set-mm.
>>
>> tools/testing/selftests/drivers/net/hw/ethtool_mm.sh will help you :)
>
>Yes, see manual_with_verification() there:
>
>ethtool --set-mm $tx verify-enabled on tx-enabled on

Thanks,
CL