From mboxrd@z Thu Jan 1 00:00:00 1970 From: Jason Wang Subject: RE: [PATCH net] hyperv: Fix the error processing in netvsc_send() Date: Mon, 02 Feb 2015 06:57:22 +0008 Message-ID: <1422859762.7028.2@smtp.corp.redhat.com> References: <1422563689-31036-1-git-send-email-haiyangz@microsoft.com> <1422613519.8840.0@smtp.corp.redhat.com> Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii"; Format="flowed" Content-Transfer-Encoding: 7bit Cc: "olaf@aepfle.de" , "netdev@vger.kernel.org" , "driverdev-devel@linuxdriverproject.org" , "linux-kernel@vger.kernel.org" , "davem@davemloft.net" To: Haiyang Zhang Return-path: In-Reply-To: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Errors-To: driverdev-devel-bounces@linuxdriverproject.org Sender: "devel" List-Id: netdev.vger.kernel.org On Fri, Jan 30, 2015 at 11:05 PM, Haiyang Zhang wrote: > > >> -----Original Message----- >> From: Jason Wang [mailto:jasowang@redhat.com] >> Sent: Friday, January 30, 2015 5:25 AM >> > + if (ret != 0) { >> > + if (section_index != NETVSC_INVALID_INDEX) >> > + netvsc_free_send_slot(net_device, section_index); >> >> What if ret is -EINVAL or -ENOSPC? Looks like we need free the skb >> in >> this case also. > > In these cases, skb is freed in netvsc_start_xmit(). > > >> > >> > + } else if (skb) { >> > + dev_kfree_skb_any(skb); >> >> The caller - netvsc_start_xmit() do this also, may be handle this in >> caller is better since netvsc_start_xmit() is the only user that >> tries >> to send a skb? > > When the packet is sent out normally, we frees it in netvsc_send() if > it's > copied to send-buffer. The free is done in netvsc_send(), because the > copy > is also in this function. If it's not copied, it will be freed in > another > function -- netvsc_xmit_completion(). > > netvsc_start_xmit() only does free skb in error case. Ok. > > >> btw, I find during netvsc_start_xmit(), ret was change to -ENOSPC >> when >> queue_sends[q_idx] < 1. But non of the caller check -ENOSPC in fact? > > In this case, we don't request re-send, so set ret to a value other > than > -EAGAIN. Why not? We have available slots for it to be sent now. Dropping the packet in this case may cause out of order sending. > It's handled in the same way as errors != -EAGAIN, so we don't > need to check this value specifically. Thanks