linux-kernel - Re: [PATCH net] mvpp2: enable rxhash only on the first port

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:   Sat, 9 May 2020 21:23:55 +0100
From:   Russell King - ARM Linux admin <linux@...linux.org.uk>
To:     Matteo Croce <mcroce@...hat.com>
Cc:     netdev@...r.kernel.org, linux-kernel@...r.kernel.org,
        "David S. Miller" <davem@...emloft.net>,
        Thomas Petazzoni <thomas.petazzoni@...tlin.com>,
        Stefan Chulski <stefanc@...vell.com>
Subject: Re: [PATCH net] mvpp2: enable rxhash only on the first port

On Sat, May 09, 2020 at 04:15:46PM +0200, Matteo Croce wrote:
> Currently rxhash only works on the first port of the CP (Communication
> Processor). Enabling it on other ports completely blocks packet reception.
> This patch only adds rxhash as supported feature to the first port,
> so rxhash can't be enabled on other ports:
> 
> 	# ethtool -K eth0 rxhash on
> 	# ethtool -K eth1 rxhash on
> 	# ethtool -K eth2 rxhash on
> 	Cannot change receive-hashing
> 	Could not change any device features
> 	# ethtool -K eth3 rxhash on
> 	Cannot change receive-hashing
> 	Could not change any device features
> 
> Fixes: 895586d5dc32 ("net: mvpp2: cls: Use RSS contexts to handle RSS tables")
> Signed-off-by: Matteo Croce <mcroce@...hat.com>
> ---
>  drivers/net/ethernet/marvell/mvpp2/mvpp2_main.c | 3 ++-
>  1 file changed, 2 insertions(+), 1 deletion(-)
> 
> diff --git a/drivers/net/ethernet/marvell/mvpp2/mvpp2_main.c b/drivers/net/ethernet/marvell/mvpp2/mvpp2_main.c
> index 2b5dad2ec650..ba71583c7ae3 100644
> --- a/drivers/net/ethernet/marvell/mvpp2/mvpp2_main.c
> +++ b/drivers/net/ethernet/marvell/mvpp2/mvpp2_main.c
> @@ -5423,7 +5423,8 @@ static int mvpp2_port_probe(struct platform_device *pdev,
>  			    NETIF_F_HW_VLAN_CTAG_FILTER;
>  
>  	if (mvpp22_rss_is_supported()) {
> -		dev->hw_features |= NETIF_F_RXHASH;
> +		if (port->id == 0)
> +			dev->hw_features |= NETIF_F_RXHASH;
>  		dev->features |= NETIF_F_NTUPLE;
>  	}

I seem to have discovered the cause of the problem in the old thread,
so I suggest we wait and see whether anyone offers up a proper
solution to this regression before we rush to completely disable
this feature.

I would suggest with a high degress of confidence based on my
research that prior to the offending commit (895586d5dc32), rx
hashing was working fine, distributing interrupts across the cores.

-- 
RMK's Patch system: https://www.armlinux.org.uk/developer/patches/
FTTC broadband for 0.8mile line in suburbia: sync at 10.2Mbps down 587kbps up