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:   Fri, 25 Oct 2019 11:02:36 +0200
From:   Stanislaw Gruszka <sgruszka@redhat.com>
To:     Lorenzo Bianconi <lorenzo@kernel.org>
Cc:     kvalo@codeaurora.org, linux-wireless@vger.kernel.org, nbd@nbd.name,
        lorenzo.bianconi@redhat.com, oleksandr@natalenko.name,
        netdev@vger.kernel.org
Subject: Re: [PATCH wireless-drivers 2/2] mt76: dma: fix buffer unmap with
 non-linear skbs
Message-ID: <20191025090236.GB14818@redhat.com>
References: <cover.1571868221.git.lorenzo@kernel.org>
 <1f7560e10edd517bfd9d3c0dd9820e6f420726b6.1571868221.git.lorenzo@kernel.org>
 <20191024081816.GA2440@redhat.com>
 <20191024090148.GC9346@localhost.localdomain>
MIME-Version: 1.0
In-Reply-To: <20191024090148.GC9346@localhost.localdomain>
User-Agent: Mutt/1.5.21 (2010-09-15)
Content-Type: text/plain; charset=WINDOWS-1252
Content-Transfer-Encoding: quoted-printable
Content-Disposition: inline
Sender: linux-wireless-owner@vger.kernel.org
Precedence: bulk

On Thu, Oct 24, 2019 at 11:01:48AM +0200, Lorenzo Bianconi wrote:
> > On Thu, Oct 24, 2019 at 12:23:16AM +0200, Lorenzo Bianconi wrote:
> > > mt76 dma layer is supposed to unmap skb data buffers while keep txwi
> > > mapped on hw dma ring. At the moment mt76 wrongly unmap txwi or does
> > > not unmap data fragments in even positions for non-linear skbs. This
> > > issue may result in hw hangs with A-MSDU if the system relies on IOMM=
U
> > > or SWIOTLB. Fix this behaviour properly unmapping data fragments on
> > > non-linear skbs.
> >=20
> > If we have to keep txwi mapped, before unmap fragments, when then
> > txwi is unmaped ?
>=20
> txwi are mapped when they are crated in mt76_alloc_txwi(). Whenever we ne=
ed to
> modify them we sync the DMA in mt76_dma_tx_queue_skb(). txwi are unmapped=
 in
> mt76_tx_free() at driver unload.

So not only txwi is wrongly unmapped on runtime, but we can call
dma_sync_single_for_cpu/device() after dma_unmap_single().
That serious bug, good you spotted it and fixed!

Stanislaw