qemu-devel
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Qemu-devel] Is network backend netmap worth keeping?


From: Giuseppe Lettieri
Subject: Re: [Qemu-devel] Is network backend netmap worth keeping?
Date: Mon, 2 Sep 2019 22:50:13 +0200
User-agent: Mozilla/5.0 (X11; Linux x86_64; rv:60.0) Gecko/20100101 Thunderbird/60.8.0

Hi all,

I have been thinking of the submodule suggestion and I have also prepared a patch for it (attached). However, I am not sure about what we want to achieve with it. In particular, I am not sure that the option is useful for end users. The problem is that netmap, unlike capstone and slirp, is a library plus a kernel module. If netmap is not installed in the system, the --enable-netmap=git option will successfully build qemu with netmap support, but then -netdev netmap will fail at runtime. On the other end, if netmap is installed in the system, using the =git option may cause qemu to be built against a different, possibly incompatible version.

If the option is only useful for developers to check that some qemu change does not break anything, then probably it should be enabled in some other, less visible way. What do you think?

Cheers,

Giuseppe

On 8/12/19 2:32 PM, Philippe Mathieu-Daudé wrote:

On 8/8/19 1:52 PM, Markus Armbruster wrote:
Giuseppe Lettieri <address@hidden> writes:

Dear Markus,

the netmap project is alive and well, if a bit understuffed. We have
moved to github:

https://github.com/luigirizzo/netmap

We have users from FreeBSD, where it is part of the official kernel,
and Linux, both from Academia and industry.

But you asked about the netmap backend in QEMU, in particular. When it
was merged, the decision was made to disable it by default because it
was not supported upstream in Linux. As Jason Wang says, this support
is even more unlikely now than it was then.

The fact the the backend has to be explicitly enabled and built from
the sources has obviously cut down the number of potential
users. However, we still think it is useful and we have pending
updates for it. If it's causing problems in the workflow, I am willing
to help as much as I can.
Could we make it a submodule, simililar to slirp and capstone?
Good idea, this would extend the coverage. Netmap users/developers are
probably best suited to do this.

     --enable-netmap=system      use the system's netmap
     --enable-netmap=git         use the git submodule
     --enable-netmap             use system's, else git, else fail
     --disable-netmap            disable netmap
     default                     use system's, else git, else disable

A fresh clone of https://github.com/luigirizzo/netmap clocks in at
14MiB, which is between libslirp's 1.5MiB and capstone's 72MiB.
In which directory should we clone it? As /netmap directly?

Should we start using a 3rd-party/ subdirectory?

Similarly, what about the virglrenderer component?

Its repository is: https://anongit.freedesktop.org/git/virglrenderer.git

Attachment: 0001-netmap-support-git-submodule-build-otption.patch
Description: Text Data


reply via email to

[Prev in Thread] Current Thread [Next in Thread]