netdev - Re: [RFC] NAPI behavior when (work

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] [day] [month] [year] [list]

Date:	Wed, 04 Feb 2015 05:37:09 -0800
From:	Eric Dumazet <eric.dumazet@...il.com>
To:	Vladimir Kondratiev <QCA_vkondrat@....qualcomm.com>
Cc:	netdev@...r.kernel.org, "David S. Miller" <davem@...emloft.net>,
	Herbert Xu <herbert@...dor.apana.org.au>,
	Eric Dumazet <edumazet@...gle.com>,
	Tom Herbert <therbert@...gle.com>,
	Vlad Yasevich <vyasevic@...hat.com>,
	Jiri Pirko <jiri@...nulli.us>,
	John Fastabend <john.fastabend@...il.com>,
	Joe Perches <joe@...ches.com>
Subject: Re: [RFC] NAPI behavior when (work < weight)

On Wed, 2015-02-04 at 15:22 +0200, Vladimir Kondratiev wrote:
> Hi
> 
> I am observing that if driver does not calls napi_complete() when
> less than full budget consumed, various strange things happen.
> 
> For example, if driver going to napi_disable(), it may wait forever
> in the loop:
> 
> 	while (test_and_set_bit(NAPI_STATE_SCHED, &n->state))
> 		msleep(1);
> 
> if last NAPI consumed less then full budget, because bit
> NAPI_STATE_SCHED never cleared.
> 
> Is it expected that driver calls napi_complete() when less than full
> budget consumed? Documentation does not say so.
> 
> What is intended behavior in case above?

A driver must call napi_complete() (or better : napi_complete_done() )
when/if returning work_done < budget

Thats because of the way we have to rearm interrupts in the driver,
_after_ calling napi_complete()



--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html