From mboxrd@z Thu Jan 1 00:00:00 1970 From: Arnd Bergmann Subject: Re: bug in virtio network driver? Date: Tue, 21 Aug 2007 17:06:29 +0200 Message-ID: <200708211706.29676.arnd@arndb.de> References: <200708211048.33534.borntraeger@de.ibm.com> <1187687377.19435.176.camel@localhost.localdomain> Mime-Version: 1.0 Content-Type: text/plain; charset="iso-8859-15" Content-Transfer-Encoding: quoted-printable Return-path: In-Reply-To: <1187687377.19435.176.camel@localhost.localdomain> Content-Disposition: inline List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Sender: virtualization-bounces@lists.linux-foundation.org Errors-To: virtualization-bounces@lists.linux-foundation.org To: virtualization@lists.linux-foundation.org List-Id: virtualization@lists.linuxfoundation.org On Tuesday 21 August 2007, Rusty Russell wrote: > > One solution would be to use the xmit_done interrupt. Unfortunately t= his would=20 > > require additional locking as multiple interrupts can happen at two o= r more=20 > > cpus. Do you have any better ideas? >=20 > The only reason that we don't do it in skb_xmit_done() is because > kfree_skb() isn't supposed to be called from an interrupt. =A0But there= 's > dev_kfree_skb_any() which can be used. That will simply trigger a softirq, right? If you need the softirq anyway= , why not do all the tx handling in the poll() method like other drivers? Arnd <><