linux-kernel - Re: [PATCH net-next v2 1/1] net: dsa: microchip: Remove ineffective checks from ksz_set_mac

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 for Android: free password hash cracker in your pocket

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <aAubnUSDpwtfuCrm@pengutronix.de>
Date: Fri, 25 Apr 2025 16:26:37 +0200
From: Oleksij Rempel <o.rempel@...gutronix.de>
To: "Russell King (Oracle)" <linux@...linux.org.uk>
Cc: "David S. Miller" <davem@...emloft.net>, Andrew Lunn <andrew@...n.ch>,
	Eric Dumazet <edumazet@...gle.com>,
	Florian Fainelli <f.fainelli@...il.com>,
	Jakub Kicinski <kuba@...nel.org>, Paolo Abeni <pabeni@...hat.com>,
	Vladimir Oltean <olteanv@...il.com>,
	Woojung Huh <woojung.huh@...rochip.com>, kernel@...gutronix.de,
	linux-kernel@...r.kernel.org, netdev@...r.kernel.org,
	UNGLinuxDriver@...rochip.com
Subject: Re: [PATCH net-next v2 1/1] net: dsa: microchip: Remove ineffective
 checks from ksz_set_mac_eee()

On Fri, Apr 25, 2025 at 02:41:21PM +0100, Russell King (Oracle) wrote:
> On Fri, Apr 25, 2025 at 01:08:45PM +0200, Oleksij Rempel wrote:
> > KSZ switches handle EEE internally via PHY advertisement and do not
> > support MAC-level configuration. The ksz_set_mac_eee() handler previously
> > rejected Tx LPI disable and timer changes, but provided no real control.
> 
> Err what?
> 
> ksz does not set phylink_config->eee_enabled_default, so the default
> state in phylink is eee_enabled = false, tx_lpi_enabled = false. It
> doesn't set the default LPI timer, so tx_lpi_timer = 0.
> 
> As the driver does not implement the ability to change the LPI timer
> enable nor the timer value, this seemed reasonable as the values are
> not reported (being reported as zeros) and thus prevents modification
> thereof.
> 
> Why do you want to allow people to change parameters that have no
> effect?

The original ksz_get_mac_eee() used to report tx_lpi_enabled = true,
which correctly reflected the internal EEE/LPI activity of the hardware.

After commit [0945a7b44220 ("net: dsa: ksz: remove setting of tx_lpi
parameters")], ksz_get_mac_eee() was removed, and now tx_lpi_enabled defaults
to false via the phylink fallback.

This leads to two problems:

- Userspace is unable to disable EEE cleanly (ethtool --set-eee lan1 eee off
  fails with -EINVAL).

- At the same time, userspace sees tx_lpi_enabled = false, which does not match
  hardware behavior (EEE/LPI is active).

At the moment, keeping the old validation checks blocks userspace from
disabling EEE at all.

But removing all validation risks accepting nonsensical parameter combinations.

Right now, I am not fully sure what the best way forward is.
 
-- 
Pengutronix e.K.                           |                             |
Steuerwalder Str. 21                       | http://www.pengutronix.de/  |
31137 Hildesheim, Germany                  | Phone: +49-5121-206917-0    |
Amtsgericht Hildesheim, HRA 2686           | Fax:   +49-5121-206917-5555 |