Received-SPF: pass (google.com: best guess record for domain of linux-wireless-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) client-ip=209.132.180.67;
Date:   Mon, 12 Aug 2019 13:50:52 +0200
From:   Stanislaw Gruszka <sgruszka@redhat.com>
To:     Lorenzo Bianconi <lorenzo.bianconi@redhat.com>
Cc:     linux-wireless@vger.kernel.org, Felix Fietkau <nbd@nbd.name>,
        Ryder Lee <ryder.lee@mediatek.com>, Roy Luo <royluo@google.com>
Subject: Re: [RFC] mt76: fix tx hung regression on MT7630E
Message-ID: <20190812115051.GB6333@redhat.com>
References: <20190729125351.GA3086@redhat.com>
 <20190729140241.GC4030@localhost.localdomain>
 <20190730135450.GA2361@redhat.com>
 <20190730145531.GA3813@localhost.localdomain>
 <20190731081957.GA4096@redhat.com>
 <20190731085147.GB4096@redhat.com>
 <20190731090927.GA3665@localhost.localdomain>
 <20190805100110.GA17889@redhat.com>
 <20190805112719.GA12280@localhost.localdomain>
 <20190805123916.GA24209@redhat.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <20190805123916.GA24209@redhat.com>
User-Agent: Mutt/1.5.21 (2010-09-15)
Sender: linux-wireless-owner@vger.kernel.org
Precedence: bulk

On Mon, Aug 05, 2019 at 02:39:16PM +0200, Stanislaw Gruszka wrote:
> On Mon, Aug 05, 2019 at 01:27:19PM +0200, Lorenzo Bianconi wrote:
> > > ... but I think we have bug when do mt76_txq_schedule_all() in
> > > tx_tasklet, because we can schedule on queues that are stoped.
> > > So reverting 41634aa8d6db and then optimize by removing tx_tasklet
> > > for mmio and remove not needed mt76_txq_schedule_all() calls looks
> > > more reasoneble to me.
> > 
> > schedule a stopped queue seems not harmful at a first glance since we do not
> > copy pending skbs if we have not enough room in the dma ring.
> 
> mac80211 stop queues for various other reasons than 
> IEEE80211_QUEUE_STOP_REASON_DRIVER .

After looking in more details, we check if queue is stopped in 
ieee80211_tx_dequeue(). But we do not check that for skb's queued
in mtxq->retry_q .

> > diff --git a/drivers/net/wireless/mediatek/mt76/tx.c b/drivers/net/wireless/mediatek/mt76/tx.c
> > index 5397827668b9..bd2d34c4f326 100644
> > --- a/drivers/net/wireless/mediatek/mt76/tx.c
> > +++ b/drivers/net/wireless/mediatek/mt76/tx.c
> > @@ -495,6 +495,9 @@ mt76_txq_schedule_list(struct mt76_dev *dev, enum mt76_txq_id qid)
> >  	while (1) {
> >  		bool empty = false;
> >  
> > +		if (hwq->stopped)
> > +			break;
> > +
> >  		if (sq->swq_queued >= 4)
> >  			break;
> > 
> > Does it fix the issue you are facing?
> 
> I'll not be able to test this patch this week. Will have access to
> the hardware next week. 
> 
> I checeked before, if
> 'q->queued + (tx_info.nbuf + 1) / 2 >= q->ndesc - 1' is triggered
> when MT7630E hangs and it is not. But maybe second part of the patch
> will help.

Patch did not help.

I debugged problem a bit more and issue is related with HW encryption.
Using full mac80211 SW encyption by retuning -EOPNOTSUPP
mt76x02_set_key() make the problem gone as well. Looks there is
some race between setting HW keys and transmitting.

Stanislaw