Re: [PATCH iwl-next v4 3/3] igc: add support for forcing link speed without autonegotiation

[Simon Horman <horms@kernel.org>]

On Wed, May 06, 2026 at 02:25:23PM +0800, Abdul Rahim, Faizal wrote:
> 
> 
> On 30/4/2026 9:50 pm, Simon Horman wrote:
> > From: 'Simon Horman' <horms@kernel.org>
> > 
> > This is an AI-generated review of your patch. The human sending this
> > email has considered the AI review valid, or at least plausible.
> > Full review at: https://sashiko.dev
> > ---
> > igc: add support for forcing link speed without autonegotiation
> > 
> > This commit allows users to force 10/100 Mb/s link speed and duplex via
> > ethtool when autonegotiation is disabled. Previously, the driver rejected
> > these requests.
> > 
> >> diff --git a/drivers/net/ethernet/intel/igc/igc_ethtool.c b/drivers/net/ethernet/intel/igc/igc_ethtool.c
> >> index cfcbf2fdad6e..9997ebbdf778 100644
> >> --- a/drivers/net/ethernet/intel/igc/igc_ethtool.c
> >> +++ b/drivers/net/ethernet/intel/igc/igc_ethtool.c
> > [ ... ]
> >> @@ -2000,6 +2013,41 @@ static int igc_ethtool_get_link_ksettings(struct net_device *netdev,
> >>  	return 0;
> >>  }
> >>  
> >> +/**
> >> + * igc_handle_autoneg_disabled - Configure forced speed/duplex settings
> >> + * @adapter: private driver structure
> >> + * @speed: requested speed (must be SPEED_10 or SPEED_100)
> >> + * @duplex: requested duplex
> >> + *
> >> + * Records forced speed/duplex when autoneg is disabled.
> >> + * Caller must validate speed before calling this function.
> >> + */
> >> +static void igc_handle_autoneg_disabled(struct igc_adapter *adapter, u32 speed,
> >> +					u8 duplex)
> >> +{
> >> +	struct igc_mac_info *mac = &adapter->hw.mac;
> >> +
> >> +	switch (speed) {
> >> +	case SPEED_10:
> >> +		mac->forced_speed_duplex = (duplex == DUPLEX_FULL) ?
> >> +			IGC_FORCED_10F : IGC_FORCED_10H;
> >> +		break;
> >> +	case SPEED_100:
> >> +		mac->forced_speed_duplex = (duplex == DUPLEX_FULL) ?
> >> +			IGC_FORCED_100F : IGC_FORCED_100H;
> >> +		break;
> >> +	default:
> >> +		WARN_ONCE(1, "Unsupported speed %u\n", speed);
> >> +		return;
> >> +	}
> >> +
> >> +	mac->autoneg_enabled = false;
> >> +
> >> +	/* Half-duplex cannot support flow control per IEEE 802.3 */
> >> +	if (duplex != DUPLEX_FULL)
> >> +		adapter->hw.fc.requested_mode = igc_fc_none;
> > 
> > Does this permanently overwrite the user's flow control preference?
> > 
> > If half-duplex is requested, requested_mode is set to igc_fc_none. Because
> > requested_mode represents the persistent state for the user's flow control
> > preference, the user might permanently lose their configuration. If they later
> > switch back to full-duplex, requested_mode remains igc_fc_none, and flow
> > control stays disabled.
> > 
> > Would it be better to dynamically override hw->fc.current_mode to
> > igc_fc_none during link configuration instead of mutating requested_mode?
> > 
> 
> Yes you're right, thanks, will update.
> 
> >> +}
> > [ ... ]
> >> @@ -2071,14 +2126,20 @@ igc_ethtool_set_link_ksettings(struct net_device *netdev,
> >>  		}
> >>  	}
> >>  
> >> +	if (cmd->base.autoneg == AUTONEG_DISABLE &&
> >> +	    cmd->base.speed != SPEED_10 && cmd->base.speed != SPEED_100) {
> >> +		netdev_info(dev, "Unsupported speed for forced link\n");
> >> +		return -EINVAL;
> >> +	}
> >> +
> > 
> > Should we also validate cmd->base.duplex here to prevent a fallback to
> > half-duplex?
> > 
> > If a user runs a command like "ethtool -s eth0 speed 100 autoneg off" while
> > the link is down, ethtool queries and preserves the current duplex value,
> > which is DUPLEX_UNKNOWN (0xff).
> > 
> > In igc_handle_autoneg_disabled(), the code checks if duplex == DUPLEX_FULL,
> > which evaluates to false, and forces the interface to 100 Mbps half-duplex.
> > 
> > Could we explicitly check that cmd->base.duplex is either DUPLEX_HALF or
> > DUPLEX_FULL and return -EINVAL otherwise?
> 
> Will update.
> 
> > 
> >>  	while (test_and_set_bit(__IGC_RESETTING, &adapter->state))
> >>  		usleep_range(1000, 2000);
> > 
> > This isn't a bug introduced by this patch, but does this ad-hoc
> > synchronization loop reimplement locking without fairness or lockdep support?
> > 
> > According to the Networking Subsystem Guidelines, using bit operations as a
> > substitute for real locks violates core kernel locking rules and introduces
> > potential races.
> > 
> > Could this be replaced with a proper concurrency mechanism such as a mutex?
> > 
> 
> Hi Simon, you've raised this issue in v2, and after discussion, you've
> agreed that this change doesn't belong in this patch series.
> 
> Not sure if I missed anything?

Sorry, my bad. I missed that we'd already covered this one.