netdev - Re: [PATCH net-next v2] Add Mellanox BlueField Gigabit Ethernet 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 for Android: free password hash cracker in your pocket

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

Message-ID: <20200820171431.194169ee@kicinski-fedora-PC1C0HJN>
Date:   Thu, 20 Aug 2020 17:14:31 -0700
From:   Jakub Kicinski <kuba@...nel.org>
To:     Vladimir Oltean <olteanv@...il.com>
Cc:     David Thompson <dthompson@...lanox.com>, netdev@...r.kernel.org,
        davem@...emloft.net, jiri@...lanox.com,
        Asmaa Mnebhi <asmaa@...lanox.com>
Subject: Re: [PATCH net-next v2] Add Mellanox BlueField Gigabit Ethernet
 driver

On Fri, 21 Aug 2020 02:04:39 +0300 Vladimir Oltean wrote:
> On Thu, Jul 30, 2020 at 05:30:59PM -0700, Jakub Kicinski wrote:
> > On Thu, 30 Jul 2020 18:53:58 -0400 David Thompson wrote:  
> > > +
> > > +	/* Tell networking subsystem to poll GigE driver */
> > > +	napi_schedule(&priv->napi);  
> > 
> > _irqoff  
> 
> Hmm, I wouldn't be so sure about this particular advice. With
> PREEMPT_RT, interrupt handlers are force-threaded and run in process
> context, therefore with hardirqs enabled. This driver doesn't call
> request_irq with IRQF_NO_THREAD, so calling napi_schedule_irqoff would
> create a bug that is very, very difficult to find.

Doesn't PREEMPT_RT take a local_lock or some form thereof around the
irq threads then? If it doesn't then we probably need one around NAPI.

Regardless even if that's the case this is an existing issue, and not
something that changes how the driver API would look.