linux-kernel - Re: [net-next v13 06/11] net: mtip: Add mtip_switch_{rx|tx} functions to the L2 switch driver

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: <20250630133436.71238a65@wsk>
Date: Mon, 30 Jun 2025 13:34:36 +0200
From: Lukasz Majewski <lukma@...x.de>
To: Paolo Abeni <pabeni@...hat.com>
Cc: Andrew Lunn <andrew+netdev@...n.ch>, davem@...emloft.net, Eric Dumazet
 <edumazet@...gle.com>, Jakub Kicinski <kuba@...nel.org>, Rob Herring
 <robh@...nel.org>, Krzysztof Kozlowski <krzk+dt@...nel.org>, Conor Dooley
 <conor+dt@...nel.org>, Shawn Guo <shawnguo@...nel.org>, Sascha Hauer
 <s.hauer@...gutronix.de>, Pengutronix Kernel Team <kernel@...gutronix.de>,
 Fabio Estevam <festevam@...il.com>, Richard Cochran
 <richardcochran@...il.com>, netdev@...r.kernel.org,
 devicetree@...r.kernel.org, linux-kernel@...r.kernel.org,
 imx@...ts.linux.dev, linux-arm-kernel@...ts.infradead.org, Stefan Wahren
 <wahrenst@....net>, Simon Horman <horms@...nel.org>, Łukasz Majewski <lukasz.majewski@...lbox.org>
Subject: Re: [net-next v13 06/11] net: mtip: Add mtip_switch_{rx|tx}
 functions to the L2 switch driver

Hi Paolo,

> >  static void mtip_switch_tx(struct net_device *dev)
> >  {
> > +	struct mtip_ndev_priv *priv = netdev_priv(dev);
> > +	struct switch_enet_private *fep = priv->fep;
> > +	unsigned short status;
> > +	struct sk_buff *skb;
> > +	unsigned long flags;
> > +	struct cbd_t *bdp;
> > +
> > +	spin_lock_irqsave(&fep->hw_lock, flags);  
> 
> This is called from napi (bh) context, and every other caller
> is/should be BH, too. You should use
> 
> 	spin_lock_bh()

I've double check the spin locking in the driver - I've also consult
the fec_main.c.

It looks like the mtip_switch_rx() and fec_enet_rx() are not using
explicit locks and rely on NAPI locking.

On the other hand - the fec_enet_tx (and corresponding MTIP variant)
use spin_lock(), not the _bh() variant.

> 
> Also please test your patches with CONFIG_LOCKDEP and
> CONFIG_DEBUG_SPINLOCK enabled, thet will help finding this king of
> issues.

This was enabled by default. By changing all locks to _bh() there were
deadlocks observed.

On the MTIP driver (due to this HW IP block) there are some locks which
must disable interrupts:

1. One is when mtip_adjust_link() is called - as it is the same for
both switch ports. Moreover, at some use cases it is required that the
switch IP block is reset.

2. The mtip_atable_dynamicms_learn_migration() - it changes the content
of dynamic switching table. IRQ from switch shall not be possible at
this time as it can be called from mtip_switch_rx() (from NAPI) and
from timer/kthread (at specified period).

To sum up:

I'm going to prepare the v14 with changes around the timer / kthread
running mtip_atable_dynamicms_learn_migration() and use time stamps
extracted from jiffies.

Locks will be optimized and following paradigm used with fec_main.c
driver.

> 
> /P

Best regards,

Lukasz Majewski

--

DENX Software Engineering GmbH,      Managing Director: Erika Unter
HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany
Phone: (+49)-8142-66989-59 Fax: (+49)-8142-66989-80 Email: lukma@...x.de

Content of type "application/pgp-signature" skipped