ACK/Cmnt: [PATCH 0/3][C/Unstable] Switch back to out-of-tree virtualbox guest modules
Seth Forshee
seth.forshee at canonical.com
Thu Oct 11 12:13:01 UTC 2018
On Thu, Oct 11, 2018 at 10:43:19AM +0200, Stefan Bader wrote:
> On 10.10.2018 15:41, Seth Forshee wrote:
> > https://bugs.launchpad.net/bugs/1796647
> >
> > == SRU Justification ==
> >
> > Impact: Vagrant environments are no longer able to use virtualbox
> > feartures like shared folders. This is due to Ubuntu Cosmic kernels
> > currently shipping the upstream vboxguest module and not importing
> > modules from the virtualbox-guest-dkms package.
> >
> > Fix: One solution would be to use the dkms modules for Vagrant, but this
> > is not ideal. At this late stage the simplest solution is to go back to
> > importing the out-of-tree modules.
> >
> > Regression Potential: Minimal. Will have no impact outside of virtualbox
> > environments, and the drivers being imported come from the
> > virtualbox-guest-dkms package in cosmic and are thus expected to work
> > well.
> >
> > Thanks,
> > Seth
> >
> > ---
> >
> > Seth Forshee (3):
> > UBUNTU: [Config] CONFIG_VBOXGUEST=n
> > UBUNTU: ubuntu: vbox -- update to 5.2.18-dfsg-2
> > UBUNTU: ubuntu: enable vbox build
> >
> > .../abi/4.18.0-9.10/amd64/generic.modules | 1 -
> > .../abi/4.18.0-9.10/amd64/lowlatency.modules | 1 -
> > .../abi/4.18.0-9.10/i386/generic.modules | 1 -
> > .../abi/4.18.0-9.10/i386/lowlatency.modules | 1 -
> > debian.master/config/annotations | 5 +-
> > debian.master/config/config.common.ubuntu | 2 +-
> > ubuntu/Makefile | 4 +-
> > ubuntu/vbox/BOM | 4 +-
> > ubuntu/vbox/dkms.conf | 2 +-
> > ubuntu/vbox/vboxguest/VBoxGuest-linux.c | 56 +---
> > ubuntu/vbox/vboxguest/common/time/time.c | 310 +++++++++++++++++-
> > .../include/VBox/VBoxGuestCoreTypes.h | 3 +
> > ubuntu/vbox/vboxguest/include/VBox/cdefs.h | 21 +-
> > ubuntu/vbox/vboxguest/include/VBox/err.h | 3 +
> > ubuntu/vbox/vboxguest/include/VBox/log.h | 3 +
> > .../vbox/vboxguest/include/internal/magics.h | 10 +-
> > ubuntu/vbox/vboxguest/include/iprt/asm.h | 62 +++-
> > ubuntu/vbox/vboxguest/include/iprt/cdefs.h | 61 ++++
> > ubuntu/vbox/vboxguest/include/iprt/err.h | 2 +
> > ubuntu/vbox/vboxguest/include/iprt/mangling.h | 28 ++
> > ubuntu/vbox/vboxguest/include/iprt/string.h | 25 ++
> > ubuntu/vbox/vboxguest/include/iprt/time.h | 28 +-
> > ubuntu/vbox/vboxguest/include/iprt/x86.h | 42 ++-
> > .../r0drv/linux/waitqueue-r0drv-linux.h | 2 +-
> > ubuntu/vbox/vboxguest/revision-generated.h | 2 +-
> > ubuntu/vbox/vboxguest/version-generated.h | 6 +-
> > .../vboxsf/include/VBox/VBoxGuestCoreTypes.h | 3 +
> > ubuntu/vbox/vboxsf/include/VBox/cdefs.h | 21 +-
> > ubuntu/vbox/vboxsf/include/VBox/err.h | 3 +
> > ubuntu/vbox/vboxsf/include/VBox/log.h | 3 +
> > ubuntu/vbox/vboxsf/include/VBox/shflsvc.h | 2 +
> > ubuntu/vbox/vboxsf/include/iprt/asm.h | 62 +++-
> > ubuntu/vbox/vboxsf/include/iprt/cdefs.h | 61 ++++
> > ubuntu/vbox/vboxsf/include/iprt/err.h | 2 +
> > ubuntu/vbox/vboxsf/include/iprt/mangling.h | 28 ++
> > ubuntu/vbox/vboxsf/include/iprt/string.h | 25 ++
> > ubuntu/vbox/vboxsf/include/iprt/time.h | 28 +-
> > ubuntu/vbox/vboxsf/revision-generated.h | 2 +-
> > ubuntu/vbox/vboxsf/utils.c | 17 +-
> > ubuntu/vbox/vboxsf/version-generated.h | 6 +-
> > ubuntu/vbox/vboxvideo/hgsmi_base.c | 11 +-
> > ubuntu/vbox/vboxvideo/hgsmi_defs.h | 14 +-
> > ubuntu/vbox/vboxvideo/modesetting.c | 62 ++--
> > ubuntu/vbox/vboxvideo/revision-generated.h | 2 +-
> > ubuntu/vbox/vboxvideo/vbox_drv.c | 2 +-
> > ubuntu/vbox/vboxvideo/vbox_drv.h | 29 +-
> > ubuntu/vbox/vboxvideo/vbox_err.h | 2 +
> > ubuntu/vbox/vboxvideo/vbox_fb.c | 4 +-
> > ubuntu/vbox/vboxvideo/vbox_main.c | 6 +-
> > ubuntu/vbox/vboxvideo/vbox_mode.c | 52 +--
> > ubuntu/vbox/vboxvideo/vbox_ttm.c | 69 +++-
> > ubuntu/vbox/vboxvideo/vboxvideo.h | 117 ++++---
> > ubuntu/vbox/vboxvideo/vbva_base.c | 78 ++---
> > ubuntu/vbox/vboxvideo/version-generated.h | 6 +-
> > 54 files changed, 1070 insertions(+), 332 deletions(-)
> >
> >
> Just a minor question, the changes to the annotations don't seem to enforce this
> setting. Would this be reasonable to add? Otherwise
I opted not to because the expectation is that someday the upstream
driver will be good enough. The note gives us a more gentle reminder of
why it was turned off when doing config review.
Seth
More information about the kernel-team
mailing list