Re: [Qemu-devel] [PATCH v3 0/5] Support building qemu-user powered docker test images

["Alex Bennée" <alex.bennee@linaro.org>]

Fam Zheng <famz@redhat.com> writes:

> On Fri, 07/08 08:53, Alex Bennée wrote:
>>
>> Fam Zheng <famz@redhat.com> writes:
>>
>> > On Tue, 06/28 16:42, Alex Bennée wrote:
>> >> This is the latest iteration of my qemu-user support inside Docker.
>> >> They apply directly on top of master. I've made the changes suggested
>> >> in the last review and split apart another patch. I've also added a
>> >> new update command so a tagged image can be updated with the latest
>> >> QEMU.
>> >>
>> >> To use, first ensure you build the debian-bootstrap image:
>> >>
>> >>     DEB_ARCH=armhf DEB_TYPE=testing \
>> >>       ./tests/docker/docker.py build qemu:debian-bootstrap \
>> >>       ./tests/docker/dockerfiles/debian-bootstrap.docker \
>> >>       --include-executable=./arm-linux-user/qemu-arm
>> >
>> > Alex,
>> >
>> > Sorry for the late reply.
>> >
>> > Is there a way to setup or detect the binfmt_misc configuration on the
>> > system?
>>
>> The trick is doing this in a distribution agnostic way. Certainly
>> Debian-a-like systems have scripts that allow you to query the state of
>> binmfmt_misc.
>
> What about changing the docker file from "FROM scratch" to "FROM debian"? Then
> I think you can query the state in the container.

But that would pull in the host debian build so you'd get clashes.

> (What is the query command
> called, BTW?)

11:08 alex@zen/x86_64  [qemu.git/mttcg/base-patches-v4] >/usr/sbin/update-binfmts --display qemu-arm
qemu-arm (enabled):
     package = qemu-user-binfmt
        type = magic
      offset = 0
       magic = \x7f\x45\x4c\x46\x01\x01\x01\x00\x00\x00\x00\x00\x00\x00\x00\x00\x02\x00\x28\x00
        mask = \xff\xff\xff\xff\xff\xff\xff\x00\xff\xff\xff\xff\xff\xff\xff\xff\xfe\xff\xff\xff
 interpreter = /usr/bin/qemu-arm
    detector =

> In addition, what about the idea I proposed in v2: moving the debootstrap
> command to a dockerfile "RUN" directive? Does it work?  I can experiment with
> that in this weekend, but if you know that already, that's even better. I think
> one hurdle in this series is the required build step.

The initial run needs a working shell so that's why we do the two stage
seed and then setup.

>
> Fam


--
Alex Bennée