netdev - Re: [RFC PATCH 02/13] net: dsa: implement a central TX reallocation procedure

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:   Mon, 19 Oct 2020 14:19:20 +0200
From:   Andrew Lunn <andrew@...n.ch>
To:     Vladimir Oltean <vladimir.oltean@....com>
Cc:     David Laight <David.Laight@...LAB.COM>,
        'Florian Fainelli' <f.fainelli@...il.com>,
        "netdev@...r.kernel.org" <netdev@...r.kernel.org>,
        "vivien.didelot@...il.com" <vivien.didelot@...il.com>,
        "kuba@...nel.org" <kuba@...nel.org>,
        Christian Eggers <ceggers@...i.de>,
        Kurt Kanzenbach <kurt@...utronix.de>
Subject: Re: [RFC PATCH 02/13] net: dsa: implement a central TX reallocation
 procedure

On Mon, Oct 19, 2020 at 10:30:47AM +0000, Vladimir Oltean wrote:
> On Mon, Oct 19, 2020 at 08:33:27AM +0000, David Laight wrote:
> > Is it possible to send the extra bytes from a separate buffer fragment?
> > The entire area could be allocated (coherent) when the rings are
> > allocated.
> > That would save having to modify the skb at all.
> > 
> > Even if some bytes of the frame header need 'adjusting' transmitting
> > from a copy may be faster - especially on systems with an iommu.
> > 
> > Many (many) moons ago we found the cutoff point for copying frames
> > on a system with an iommu to be around 1k bytes.
> 
> Please help me understand better how to implement what you're suggesting.
> DSA switches have 3 places where they might insert a tag:
> 1. Between the source MAC address and the EtherType (this is the most
>    common)
> 2. Before the destination MAC address
> 3. Before the FCS
> 
> I imagine that the most common scenario (1) is also the most difficult
> to implement using fragments, since I would need to split the Ethernet
> header from the rest of the skb data area, which might defeat the
> purpose.

We also have length issues. Most scatter/gather DMA engines require
the fragments are multiple of 4 bytes. Only the last segment does not
have this length restriction. And some of the DSA tag headers are 2
bytes, or 1 byte. So some master devices are going to have to convert
the fragments back to a linear buffer.

    Andrew