From: Ben Greear Date: Tue, 25 Jan 2011 14:01:22 -0800 Subject: [ath9k-devel] [PATCH RESEND 09/11] ath9k: Try all queues when looking for next packet to send. In-Reply-To: <20110125213710.GD2927@tuxdriver.com> References: <1294643513-18820-1-git-send-email-greearb@candelatech.com> <1294643513-18820-10-git-send-email-greearb@candelatech.com> <20110125213710.GD2927@tuxdriver.com> Message-ID: <4D3F4832.4080907@candelatech.com> List-Id: MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit To: ath9k-devel@lists.ath9k.org On 01/25/2011 01:37 PM, John W. Linville wrote: > On Sun, Jan 09, 2011 at 11:11:51PM -0800, greearb at candelatech.com wrote: >> From: Ben Greear >> >> There can be multiple struct ath_atx_ac entries for each >> txq, so if the first one doesn't have any packets to send >> now, try the rest of them. This should help keep xmit >> going when using multiple stations, especially with AMPDU >> enabled. >> >> Signed-off-by: Ben Greear > > This one doesn't seem to apply any more. Can I persuade you to rebase > and resubmit it? I think this is already taken care of, or at least the important parts. Felix did some initial work in this area, and then I followed on with this one: ath9k: Try more than one queue when scheduling new aggregate. The only ath9k patches I have in my queue that are not upstream are: (From Louis Rodriguez) ath9k: warn when we get a ATH9K_INT_TIM_TIMER and are idle and my: ath9k: Implement rx copy-break. Which appears DOA as folks seem to like paged-skbs instead. Thanks, Ben -- Ben Greear Candela Technologies Inc http://www.candelatech.com