netdev - Re: [PATCH 00/16] virtio-net: split virtio-net.c

Open Source and information security mailing list archives

Message-ID: <1680250214.4751976-2-xuanzhuo@linux.alibaba.com>
Date:   Fri, 31 Mar 2023 16:10:14 +0800
From:   Xuan Zhuo <xuanzhuo@...ux.alibaba.com>
To:     "Michael S. Tsirkin" <mst@...hat.com>
Cc:     Jason Wang <jasowang@...hat.com>,
        Jesper Dangaard Brouer <hawk@...nel.org>,
        Daniel Borkmann <daniel@...earbox.net>, netdev@...r.kernel.org,
        John Fastabend <john.fastabend@...il.com>,
        Alexei Starovoitov <ast@...nel.org>,
        virtualization@...ts.linux-foundation.org,
        Eric Dumazet <edumazet@...gle.com>,
        Jakub Kicinski <kuba@...nel.org>, bpf@...r.kernel.org,
        Paolo Abeni <pabeni@...hat.com>,
        "David S. Miller" <davem@...emloft.net>
Subject: Re: [PATCH 00/16] virtio-net: split virtio-net.c

On Fri, 31 Mar 2023 04:00:22 -0400, "Michael S. Tsirkin" <mst@...hat.com> wrote:
> On Fri, Mar 31, 2023 at 03:48:00PM +0800, Xuan Zhuo wrote:
> > On Fri, 31 Mar 2023 15:35:14 +0800, Jason Wang <jasowang@...hat.com> wrote:
> > > On Fri, Mar 31, 2023 at 3:31 PM Xuan Zhuo <xuanzhuo@...ux.alibaba.com> wrote:
> > > >
> > > > On Thu, 30 Mar 2023 02:17:43 -0400, "Michael S. Tsirkin" <mst@...hat.com> wrote:
> > > > > On Tue, Mar 28, 2023 at 05:28:31PM +0800, Xuan Zhuo wrote:
> > > > > > Considering the complexity of virtio-net.c and the new features we want
> > > > > > to add, it is time to split virtio-net.c into multiple independent
> > > > > > module files.
> > > > > >
> > > > > > This is beneficial to the maintenance and adding new functions.
> > > > > >
> > > > > > And AF_XDP support will be added later, then a separate xsk.c file will
> > > > > > be added.
> > > > > >
> > > > > > This patchset split virtio-net.c into these parts:
> > > > > >
> > > > > > * virtnet.c:         virtio net device ops (napi, tx, rx, device ops, ...)
> > > > > > * virtnet_common.c:  virtio net common code
> > > > > > * virtnet_ethtool.c: virtio net ethtool callbacks
> > > > > > * virtnet_ctrl.c:    virtio net ctrl queue command APIs
> > > > > > * virtnet_virtio.c:  virtio net virtio callbacks/ops (driver register, virtio probe, virtio free, ...)
> > > > > >
> > > > > > Please review.
> > > > > >
> > > > > > Thanks.
> > > > >
> > > > >
> > > > > I don't feel this is an improvement as presented, will need more work
> > > > > to make code placement more logical.
> > > >
> > > > Yes, this does need some time and energy. But I think this always need to do,
> > > > just when to do it. I think it is currently an opportunity.
> > > >
> > > >
> > > > >
> > > > > For example where do I find code to update rq stats?
> > > > > Rx data path should be virtnet.c?
> > > > > No it's in virtnet_ethtool.c because rq stats can be
> > > > > accessed by ethtool.
> > > >
> > > > That's what I do.
> > > >
> > > > > A bunch of stuff seems to be in headers just because of technicalities.
> > > > > virtio common seems to be a dumping ground with no guiding principle at
> > > > > all.
> > > >
> > > > Yes, I agree, with the development of time, common will indeed become a dumping
> > > > group. This is something we should pay attention to after this.
> > > >
> > > >
> > > > > drivers/net/virtio/virtnet_virtio.c is weird with
> > > > > virt repeated three times in the path.
> > > >
> > > > Any good idea.
> > > >
> > > > >
> > > > > These things only get murkier with time, at the point of reorg
> > > > > I would expect very logical placement, since
> > > > > without clear guiding rule finding where something is becomes harder but
> > > > > more importantly we'll now get endless heartburn about where does each new
> > > > > function go.
> > > > >
> > > > >
> > > > > The reorg is unfortunately not free - for example git log --follow will
> > > > > no longer easily match virtio because --follow works with exactly one
> > > > > path.
> > > >
> > > > One day we will face this problem.
> > > >
> > > > > It's now also extra work to keep headers self-consistent.
> > > >
> > > > Can we make it simpler, first complete the split.
> > > >
> > > >
> > > > > So it better be a big improvement to be worth it.
> > > >
> > > >
> > > > Or about split, do you have any better thoughts? Or do you think we have always
> > > > been like this and make Virtio-Net more and more complicated?
> > >
> > > My feeling is that maybe it's worth it to start using a separate file
> > > for xsk support.
> >
> > I agree.
> >
> > @Michael at this point, what is your thought?
> >
> >
> > Thanks.
> >
>
> I am fine with either adding just xsk in a new file or even
> just adding in same file working on a split later.

Okay, we are on the same page with xsk.

So the remaining problem is whether do we need to split ethtool, ctrl command,
and virtio out?

I think at least ctrl command and ethtool are very clear. The split of virtio
has indeed encountered some problems. I have independent modifications into a
separate commit. The patches to move the code do not include any other
modifications.

Thanks.

>
>
> > >
> > > Thanks
> > >
> > > >
> > > >
> > > > Thanks.
> > > >
> > > > >
> > > > >
> > > > >
> > > > >
> > > > > > Xuan Zhuo (16):
> > > > > >   virtio_net: add a separate directory for virtio-net
> > > > > >   virtio_net: move struct to header file
> > > > > >   virtio_net: add prefix to the struct inside header file
> > > > > >   virtio_net: separating cpu-related funs
> > > > > >   virtio_net: separate virtnet_ctrl_set_queues()
> > > > > >   virtio_net: separate virtnet_ctrl_set_mac_address()
> > > > > >   virtio_net: remove lock from virtnet_ack_link_announce()
> > > > > >   virtio_net: separating the APIs of cq
> > > > > >   virtio_net: introduce virtnet_rq_update_stats()
> > > > > >   virtio_net: separating the funcs of ethtool
> > > > > >   virtio_net: introduce virtnet_dev_rx_queue_group()
> > > > > >   virtio_net: introduce virtnet_get_netdev()
> > > > > >   virtio_net: prepare for virtio
> > > > > >   virtio_net: move virtnet_[en/dis]able_delayed_refill to header file
> > > > > >   virtio_net: add APIs to register/unregister virtio driver
> > > > > >   virtio_net: separating the virtio code
> > > > > >
> > > > > >  MAINTAINERS                                   |    2 +-
> > > > > >  drivers/net/Kconfig                           |    8 +-
> > > > > >  drivers/net/Makefile                          |    2 +-
> > > > > >  drivers/net/virtio/Kconfig                    |   11 +
> > > > > >  drivers/net/virtio/Makefile                   |   10 +
> > > > > >  .../net/{virtio_net.c => virtio/virtnet.c}    | 2368 ++---------------
> > > > > >  drivers/net/virtio/virtnet.h                  |  213 ++
> > > > > >  drivers/net/virtio/virtnet_common.c           |  138 +
> > > > > >  drivers/net/virtio/virtnet_common.h           |   14 +
> > > > > >  drivers/net/virtio/virtnet_ctrl.c             |  272 ++
> > > > > >  drivers/net/virtio/virtnet_ctrl.h             |   45 +
> > > > > >  drivers/net/virtio/virtnet_ethtool.c          |  578 ++++
> > > > > >  drivers/net/virtio/virtnet_ethtool.h          |    8 +
> > > > > >  drivers/net/virtio/virtnet_virtio.c           |  880 ++++++
> > > > > >  drivers/net/virtio/virtnet_virtio.h           |    8 +
> > > > > >  15 files changed, 2366 insertions(+), 2191 deletions(-)
> > > > > >  create mode 100644 drivers/net/virtio/Kconfig
> > > > > >  create mode 100644 drivers/net/virtio/Makefile
> > > > > >  rename drivers/net/{virtio_net.c => virtio/virtnet.c} (50%)
> > > > > >  create mode 100644 drivers/net/virtio/virtnet.h
> > > > > >  create mode 100644 drivers/net/virtio/virtnet_common.c
> > > > > >  create mode 100644 drivers/net/virtio/virtnet_common.h
> > > > > >  create mode 100644 drivers/net/virtio/virtnet_ctrl.c
> > > > > >  create mode 100644 drivers/net/virtio/virtnet_ctrl.h
> > > > > >  create mode 100644 drivers/net/virtio/virtnet_ethtool.c
> > > > > >  create mode 100644 drivers/net/virtio/virtnet_ethtool.h
> > > > > >  create mode 100644 drivers/net/virtio/virtnet_virtio.c
> > > > > >  create mode 100644 drivers/net/virtio/virtnet_virtio.h
> > > > > >
> > > > > > --
> > > > > > 2.32.0.3.g01195cf9f
> > > > >
> > > >
> > >
> > > _______________________________________________
> > > Virtualization mailing list
> > > Virtualization@...ts.linux-foundation.org
> > > https://lists.linuxfoundation.org/mailman/listinfo/virtualization
>

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