Return-path: Received: from mx1.redhat.com ([209.132.183.28]:58330 "EHLO mx1.redhat.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S932305AbcKNMoo (ORCPT ); Mon, 14 Nov 2016 07:44:44 -0500 Date: Mon, 14 Nov 2016 13:43:23 +0100 From: Stanislaw Gruszka To: Mathias Kresin Cc: linux-wireless@vger.kernel.org, Helmut Schaa , Felix Fietkau Subject: Re: [PATCH 05/10] rt2800: make ba_size depend on ampdu_factor Message-ID: <20161114124323.GA31857@redhat.com> (sfid-20161114_134447_767997_4BAAF23A) References: <1478095865-8651-1-git-send-email-sgruszka@redhat.com> <1478095865-8651-6-git-send-email-sgruszka@redhat.com> <20161114084536.GB12372@redhat.com> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii In-Reply-To: <20161114084536.GB12372@redhat.com> Sender: linux-wireless-owner@vger.kernel.org List-ID: On Mon, Nov 14, 2016 at 09:45:36AM +0100, Stanislaw Gruszka wrote: > Could you check below patch and see if it helps? If it does not, > could you printk sta->ht_cap.ampdu_density and ba_size values > and provide them here. Actually please print parameters from below patch. I think ba_size should be based on per TID buf_size instead of ampdu_factor, in case STA has buf size different for some TIDs. Also adding Felix to cc since my orginal patch: http://marc.info/?l=linux-wireless&m=147809595316289&w=2 was shamelessly stolen from mt76 driver. Perhaps Felix could provide us some expertise. Thanks Stanislaw diff --git a/drivers/net/wireless/ralink/rt2x00/rt2800lib.c b/drivers/net/wireless/ralink/rt2x00/rt2800lib.c index 2515702..35bc38c 100644 --- a/drivers/net/wireless/ralink/rt2x00/rt2800lib.c +++ b/drivers/net/wireless/ralink/rt2x00/rt2800lib.c @@ -7950,6 +7950,8 @@ int rt2800_ampdu_action(struct ieee80211_hw *hw, struct ieee80211_vif *vif, struct rt2x00_sta *sta_priv = (struct rt2x00_sta *)sta->drv_priv; int ret = 0; + printk("action %d sta %pM tid %u buf_size %u ampdu_factor %u\n", params->action, sta->addr, params->tid, params->buf_size, sta->ht_cap.ampdu_factor); + /* * Don't allow aggregation for stations the hardware isn't aware * of because tx status reports for frames to an unknown station