Received-SPF: pass (google.com: best guess record for domain of linux-kernel-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) client-ip=209.132.180.67;
Date:   Fri, 29 Mar 2019 12:48:47 +0200
From:   Aaro Koskinen <aaro.koskinen@iki.fi>
To:     Jose Abreu <jose.abreu@synopsys.com>
Cc:     "David S. Miller" <davem@davemloft.net>,
        Alexandre Torgue <alexandre.torgue@st.com>,
        Giuseppe Cavallaro <peppe.cavallaro@st.com>,
        netdev@vger.kernel.org, linux-kernel@vger.kernel.org,
        Aaro Koskinen <aaro.koskinen@nokia.com>
Subject: Re: [PATCH 1/6] net: stmmac: use correct DMA buffer size in the RX
 descriptor
Message-ID: <20190329104847.GF16484@darkstar.musicnaut.iki.fi>
References: <20190327203540.1884-1-aaro.koskinen@iki.fi>
 <20190327203540.1884-2-aaro.koskinen@iki.fi>
 <5c480a83-dfcf-c20c-f277-4449a7b41403@synopsys.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <5c480a83-dfcf-c20c-f277-4449a7b41403@synopsys.com>
User-Agent: Mutt/1.5.24 (2015-08-30)
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

Hi,

On Fri, Mar 29, 2019 at 10:02:20AM +0000, Jose Abreu wrote:
> On 3/27/2019 8:35 PM, Aaro Koskinen wrote:
> > From: Aaro Koskinen <aaro.koskinen@nokia.com>
> > 
> > We always program the maximum DMA buffer size into the receive descriptor,
> > although the allocated size may be less. E.g. with the default MTU size
> > we allocate only 1536 bytes. If somebody sends us a bigger frame, then
> > memory may get corrupted.
> > 
> > Fix by using exact buffer sizes.
> > 
> > Signed-off-by: Aaro Koskinen <aaro.koskinen@nokia.com>
> 
> So, I'm seeing that the maximum allowed buffer size that can be
> put in the descriptor changes between enhanced descriptors to
> normal descriptors (8KB vs. 2KB right ?).

There are two size fields, so it's 16KB vs 4 KB.

> Shouldn't stmmac_main know about this limit before trying to init
> descriptors ?

Yes, and that is not a problem.

> We do limit the MTU according to HW version but I would rather
> prefer not having to need to calculate min() values in the
> descriptor code and just use the value as is ...

I don't think the calculation can be avoided, as the passed value can
be bigger than a single field can hold, or less than the maximum.

A.