From mboxrd@z Thu Jan 1 00:00:00 1970 Return-path: Received: from bu3sch.de ([62.75.166.246]:58973 "EHLO vs166246.vserver.de" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1753892AbYKUTCK (ORCPT ); Fri, 21 Nov 2008 14:02:10 -0500 From: Michael Buesch To: "Luis R. Rodriguez" Subject: Re: What should wireless drivers do under -ENOMEM Date: Fri, 21 Nov 2008 20:01:46 +0100 Cc: linux-wireless References: <43e72e890811211055x19882fa1l5ced86b6a031a203@mail.gmail.com> In-Reply-To: <43e72e890811211055x19882fa1l5ced86b6a031a203@mail.gmail.com> MIME-Version: 1.0 Content-Type: text/plain; charset="utf-8" Message-Id: <200811212001.46336.mb@bu3sch.de> (sfid-20081121_200214_093066_D8EA6618) Sender: linux-wireless-owner@vger.kernel.org List-ID: On Friday 21 November 2008 19:55:12 Luis R. Rodriguez wrote: > Under heavy stress you can sometimes run into -ENOMEM and you inform > the callers. Now ath5k/ath9k both require dev_alloc_skb()'ing some > buffers when RXing. Right now we simply move on, but I think it'd be > nicer to inform mac80211 and let mac80211 decide. If the system is Well, what should mac80211 decide? If you are under such heavy load that memory allocations fail, we really have only one chance. That is drop the network traffic until memory allocation works again. I don't see anything else we can do. -- Greetings Michael.