netdev - Re: [PATCH net-next] xsk: introduce xsk_dma

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:   Wed, 19 Apr 2023 23:12:28 -0700
From:   Christoph Hellwig <hch@...radead.org>
To:     Alexander Lobakin <aleksander.lobakin@...el.com>
Cc:     Xuan Zhuo <xuanzhuo@...ux.alibaba.com>, netdev@...r.kernel.org,
        Björn Töpel <bjorn@...nel.org>,
        Magnus Karlsson <magnus.karlsson@...el.com>,
        Maciej Fijalkowski <maciej.fijalkowski@...el.com>,
        Jonathan Lemon <jonathan.lemon@...il.com>,
        "David S. Miller" <davem@...emloft.net>,
        Eric Dumazet <edumazet@...gle.com>,
        Jakub Kicinski <kuba@...nel.org>,
        Paolo Abeni <pabeni@...hat.com>,
        Alexei Starovoitov <ast@...nel.org>,
        Daniel Borkmann <daniel@...earbox.net>,
        Jesper Dangaard Brouer <hawk@...nel.org>,
        John Fastabend <john.fastabend@...il.com>, bpf@...r.kernel.org,
        virtualization@...ts.linux-foundation.org,
        Jason Wang <jasowang@...hat.com>,
        "Michael S. Tsirkin" <mst@...hat.com>,
        Guenter Roeck <linux@...ck-us.net>,
        Gerd Hoffmann <kraxel@...hat.com>,
        Christoph Hellwig <hch@...radead.org>
Subject: Re: [PATCH net-next] xsk: introduce xsk_dma_ops

On Wed, Apr 19, 2023 at 03:22:39PM +0200, Alexander Lobakin wrote:
> If DMA syncs are not needed on your x86_64 DMA-coherent system, it
> doesn't mean we all don't need it.

If the DMA isn't actually a DMA (as in the virtio case, or other
cases that instead have to do their own dma mapping at much lower
layers) syncs generally don't make sense.

> Instead of filling pointers with
> "default" callbacks, you could instead avoid indirect calls at all when
> no custom DMA ops are specified. Pls see how for example Christoph did
> that for direct DMA. It would cost only one if-else for case without
> custom DMA ops here instead of an indirect call each time.

So yes, I think the abstraction here should not be another layer of
DMA ops, but the option to DMA map or not at all.