[PATCH 0/5] libvirt fixes and a kernel update

[PATCH 0/5] libvirt fixes and a kernel update

[Jonas Eriksson]

Hi,

The ultimate goal of this patch series was to use a DISTRO_FEATURE to include
the lxc.scc configuration in the kernel to be built. This is done to avoid that
the kernel configuration gets additional code paths ( =y options ) when adding
the meta-virtualization layer. The configurations that enables modules are left
untouched.

In this process, I stumbled upon some bugs/issues with the libvirt recipe,
which are updated as well.

Best regards,
/Jonas

Jonas Eriksson (5):
  libvirt: Change license to LGPLv2.1+
  libvirt: Fix selinux DISTRO_FEATURE check
  Use a DISTRO_FEATURE to toggle LXC support
  libvirt: Use DISTRO_FEATURES to enable Xen support
  libvirt: Disable polkit driver when there is no x11

 recipes-extended/libvirt/libvirt_1.2.1.bb      | 14 +++++++++-----
 recipes-kernel/linux/linux-yocto_3.10.bbappend |  4 +++-
 recipes-kernel/linux/linux-yocto_3.4.bbappend  |  4 +++-
 recipes-kernel/linux/linux-yocto_3.8.bbappend  |  4 +++-
 4 files changed, 18 insertions(+), 8 deletions(-)

-- 
1.9.rc1

[PATCH 1/5] libvirt: Change license to LGPLv2.1+

[Jonas Eriksson]

libvirt is to a large extent licensed under LGPLv2.1 (or later) rather
than GPLv2. From my investigations, there are only some tests and a
couple of files under tools are licensed as GPL. Other than that, files
are generally licensed as LGPL. My recommendation is therefore to set
the default license to LGPLv2.1+, and to add a LICENSE_${PN}-ptest that
is set to LGPLv2.1+ & GPLv2+ if a ptest package is added. This is
preferred to avoid having to specify explicit licenses for every other
package, such as the auto generated libvirt-locale-XX packages.

Signed-off-by: Jonas Eriksson <jonas.eriksson@enea.com>
---
 recipes-extended/libvirt/libvirt_1.2.1.bb | 5 +++--
 1 file changed, 3 insertions(+), 2 deletions(-)

diff --git a/recipes-extended/libvirt/libvirt_1.2.1.bb b/recipes-extended/libvirt/libvirt_1.2.1.bb
index 3b95486..d30f09d 100644
--- a/recipes-extended/libvirt/libvirt_1.2.1.bb
+++ b/recipes-extended/libvirt/libvirt_1.2.1.bb
@@ -1,7 +1,8 @@
 DESCRIPTION = "A toolkit to interact with the virtualization capabilities of recent versions of Linux." 
 HOMEPAGE = "http://libvirt.org"
-LICENSE = "GPLv2+"
-LIC_FILES_CHKSUM = "file://COPYING;md5=b234ee4d69f5fce4486a80fdaf4a4263"
+LICENSE = "LGPLv2.1+"
+LIC_FILES_CHKSUM = "file://COPYING;md5=b234ee4d69f5fce4486a80fdaf4a4263 \
+                    file://COPYING.LESSER;md5=4fbd65380cdd255951079008b364516c"
 SECTION = "console/tools"
 PR = "r1"
 
-- 
1.9.rc1

[PATCH 2/5] libvirt: Fix selinux DISTRO_FEATURE check

[Jonas Eriksson]

It seems that a $ was forgotten in the DISTRO_FEATURE check for selinux.

Signed-off-by: Jonas Eriksson <jonas.eriksson@enea.com>
---
 recipes-extended/libvirt/libvirt_1.2.1.bb | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/recipes-extended/libvirt/libvirt_1.2.1.bb b/recipes-extended/libvirt/libvirt_1.2.1.bb
index d30f09d..76d3f42 100644
--- a/recipes-extended/libvirt/libvirt_1.2.1.bb
+++ b/recipes-extended/libvirt/libvirt_1.2.1.bb
@@ -111,7 +111,7 @@ INITSCRIPT_PARAMS_${PN}-libvirtd = "defaults 72"
 # full config
 PACKAGECONFIG ??= "qemu yajl xen libxl xen-inotify uml openvz vmware vbox esx iproute2 \
                    polkit lxc test remote macvtap libvirtd netcf udev python ebtables \
-                   {@base_contains('DISTRO_FEATURES', 'selinux', 'selinux', '', d)} \
+                   ${@base_contains('DISTRO_FEATURES', 'selinux', 'selinux', '', d)} \
                   "
 
 # enable,disable,depends,rdepends
-- 
1.9.rc1

[PATCH 3/5] Use a DISTRO_FEATURE to toggle LXC support

[Jonas Eriksson]

LXC requires a number of cgroups and name spaces to be supported by the
kernel, which are not possible to enable as modules. This means that an
eventual performance hit would affect all linux-yocto kernels build with
the meta-virtualization layer added. To remove this overhead, require
LXC support to be explicitly enabled using a DISTRO_FEATURE, just like
Xen support is managed today as it suffers from the same problems.

At the same time as solving this for linux-yocto, we can use the
DISTRO_FEATURE variable to affect the default PKGCONFIG in libvirt.

Signed-off-by: Jonas Eriksson <jonas.eriksson@enea.com>
---
 recipes-extended/libvirt/libvirt_1.2.1.bb      | 3 ++-
 recipes-kernel/linux/linux-yocto_3.10.bbappend | 4 +++-
 recipes-kernel/linux/linux-yocto_3.4.bbappend  | 4 +++-
 recipes-kernel/linux/linux-yocto_3.8.bbappend  | 4 +++-
 4 files changed, 11 insertions(+), 4 deletions(-)

diff --git a/recipes-extended/libvirt/libvirt_1.2.1.bb b/recipes-extended/libvirt/libvirt_1.2.1.bb
index 76d3f42..dbc6ef3 100644
--- a/recipes-extended/libvirt/libvirt_1.2.1.bb
+++ b/recipes-extended/libvirt/libvirt_1.2.1.bb
@@ -110,8 +110,9 @@ INITSCRIPT_PARAMS_${PN}-libvirtd = "defaults 72"
 
 # full config
 PACKAGECONFIG ??= "qemu yajl xen libxl xen-inotify uml openvz vmware vbox esx iproute2 \
-                   polkit lxc test remote macvtap libvirtd netcf udev python ebtables \
+                   polkit test remote macvtap libvirtd netcf udev python ebtables \
                    ${@base_contains('DISTRO_FEATURES', 'selinux', 'selinux', '', d)} \
+                   ${@base_contains('DISTRO_FEATURES', 'lxc', 'lxc', '', d)} \
                   "
 
 # enable,disable,depends,rdepends
diff --git a/recipes-kernel/linux/linux-yocto_3.10.bbappend b/recipes-kernel/linux/linux-yocto_3.10.bbappend
index c029c05..584035b 100644
--- a/recipes-kernel/linux/linux-yocto_3.10.bbappend
+++ b/recipes-kernel/linux/linux-yocto_3.10.bbappend
@@ -3,7 +3,6 @@ FILESEXTRAPATHS_prepend := "${THISDIR}/${PN}:"
 SRC_URI += "file://xt-checksum.scc \
             file://ebtables.scc \
 	    file://vswitch.scc \
-	    file://lxc.scc \
 	    "
 module_autoload_openvswitch = "openvswitch"
 KERNEL_FEATURES_append = " features/kvm/qemu-kvm-enable.scc"
@@ -17,3 +16,6 @@ KERNEL_FEATURES_append += "${@base_contains('DISTRO_FEATURES', 'aufs', ' feature
 
 # xen kernel support
 SRC_URI += "${@base_contains('DISTRO_FEATURES', 'xen', ' file://xen.scc', '', d)}"
+
+# lxc support
+SRC_URI += "${@base_contains('DISTRO_FEATURES', 'lxc', ' file://lxc.scc', '', d)}"
diff --git a/recipes-kernel/linux/linux-yocto_3.4.bbappend b/recipes-kernel/linux/linux-yocto_3.4.bbappend
index bf84725..c1aaebe 100644
--- a/recipes-kernel/linux/linux-yocto_3.4.bbappend
+++ b/recipes-kernel/linux/linux-yocto_3.4.bbappend
@@ -3,7 +3,6 @@ FILESEXTRAPATHS_prepend := "${THISDIR}/${PN}:"
 SRC_URI += "file://xt-checksum.scc \
             file://ebtables.scc \
 	    file://vswitch.scc \
-	    file://lxc.scc \
 	    "
 module_autoload_openvswitch = "openvswitch"
 KERNEL_FEATURES_append = " features/kvm/qemu-kvm-enable.scc"
@@ -14,3 +13,6 @@ module_autoload_kvm-intel = "kvm-intel"
 
 # xen kernel support
 SRC_URI += "${@base_contains('DISTRO_FEATURES', 'xen', ' file://xen.scc', '', d)}"
+
+# lxc support
+SRC_URI += "${@base_contains('DISTRO_FEATURES', 'lxc', ' file://lxc.scc', '', d)}"
diff --git a/recipes-kernel/linux/linux-yocto_3.8.bbappend b/recipes-kernel/linux/linux-yocto_3.8.bbappend
index c029c05..584035b 100644
--- a/recipes-kernel/linux/linux-yocto_3.8.bbappend
+++ b/recipes-kernel/linux/linux-yocto_3.8.bbappend
@@ -3,7 +3,6 @@ FILESEXTRAPATHS_prepend := "${THISDIR}/${PN}:"
 SRC_URI += "file://xt-checksum.scc \
             file://ebtables.scc \
 	    file://vswitch.scc \
-	    file://lxc.scc \
 	    "
 module_autoload_openvswitch = "openvswitch"
 KERNEL_FEATURES_append = " features/kvm/qemu-kvm-enable.scc"
@@ -17,3 +16,6 @@ KERNEL_FEATURES_append += "${@base_contains('DISTRO_FEATURES', 'aufs', ' feature
 
 # xen kernel support
 SRC_URI += "${@base_contains('DISTRO_FEATURES', 'xen', ' file://xen.scc', '', d)}"
+
+# lxc support
+SRC_URI += "${@base_contains('DISTRO_FEATURES', 'lxc', ' file://lxc.scc', '', d)}"
-- 
1.9.rc1

[PATCH 4/5] libvirt: Use DISTRO_FEATURES to enable Xen support

[Jonas Eriksson]

The Xen drivers in libvirt is needed when the build file system image
and kernel are running as dom0. The parts of the kernel needed to run as
dom0 are controlled by DISTRO_FEATURES, and thus there is no need for
the Xen drivers if DISTRO_FEATURES does not contain 'xen'.

Signed-off-by: Jonas Eriksson <jonas.eriksson@enea.com>
---
 recipes-extended/libvirt/libvirt_1.2.1.bb | 5 +++--
 1 file changed, 3 insertions(+), 2 deletions(-)

diff --git a/recipes-extended/libvirt/libvirt_1.2.1.bb b/recipes-extended/libvirt/libvirt_1.2.1.bb
index dbc6ef3..a75cc45 100644
--- a/recipes-extended/libvirt/libvirt_1.2.1.bb
+++ b/recipes-extended/libvirt/libvirt_1.2.1.bb
@@ -109,10 +109,11 @@ INITSCRIPT_PARAMS_${PN}-libvirtd = "defaults 72"
 #PACKAGECONFIG ??= "xen libxl xen-inotify test remote libvirtd"
 
 # full config
-PACKAGECONFIG ??= "qemu yajl xen libxl xen-inotify uml openvz vmware vbox esx iproute2 \
-                   polkit test remote macvtap libvirtd netcf udev python ebtables \
+PACKAGECONFIG ??= "qemu yajl uml openvz vmware vbox esx iproute2 polkit test \
+                   remote macvtap libvirtd netcf udev python ebtables \
                    ${@base_contains('DISTRO_FEATURES', 'selinux', 'selinux', '', d)} \
                    ${@base_contains('DISTRO_FEATURES', 'lxc', 'lxc', '', d)} \
+                   ${@base_contains('DISTRO_FEATURES', 'xen', 'xen libxl xen-inotify', '', d)} \
                   "
 
 # enable,disable,depends,rdepends
-- 
1.9.rc1

[PATCH 5/5] libvirt: Disable polkit driver when there is no x11

[Jonas Eriksson]

To avoid having the default build of libvirt depend on virtual/x11
through polkit and consolekit, check for x11 in DISTRO_FEATURES before
enabling the polkit support in the default PKG_CONFIG.

Signed-off-by: Jonas Eriksson <jonas.eriksson@enea.com>
---
 recipes-extended/libvirt/libvirt_1.2.1.bb | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/recipes-extended/libvirt/libvirt_1.2.1.bb b/recipes-extended/libvirt/libvirt_1.2.1.bb
index a75cc45..2841fcb 100644
--- a/recipes-extended/libvirt/libvirt_1.2.1.bb
+++ b/recipes-extended/libvirt/libvirt_1.2.1.bb
@@ -109,11 +109,12 @@ INITSCRIPT_PARAMS_${PN}-libvirtd = "defaults 72"
 #PACKAGECONFIG ??= "xen libxl xen-inotify test remote libvirtd"
 
 # full config
-PACKAGECONFIG ??= "qemu yajl uml openvz vmware vbox esx iproute2 polkit test \
+PACKAGECONFIG ??= "qemu yajl uml openvz vmware vbox esx iproute2 test \
                    remote macvtap libvirtd netcf udev python ebtables \
                    ${@base_contains('DISTRO_FEATURES', 'selinux', 'selinux', '', d)} \
                    ${@base_contains('DISTRO_FEATURES', 'lxc', 'lxc', '', d)} \
                    ${@base_contains('DISTRO_FEATURES', 'xen', 'xen libxl xen-inotify', '', d)} \
+                   ${@base_contains('DISTRO_FEATURES', 'x11', 'polkit', '', d)} \
                   "
 
 # enable,disable,depends,rdepends
-- 
1.9.rc1

Re: [PATCH 0/5] libvirt fixes and a kernel update

[Bruce Ashfield]

On Wed, Feb 26, 2014 at 5:07 AM, Jonas Eriksson <jonas.eriksson@enea.com> wrote:
> Hi,
>
> The ultimate goal of this patch series was to use a DISTRO_FEATURE to include
> the lxc.scc configuration in the kernel to be built. This is done to avoid that
> the kernel configuration gets additional code paths ( =y options ) when adding
> the meta-virtualization layer. The configurations that enables modules are left
> untouched.

I've never been particularly concerned about the extra functionality that is
introduced when meta-virt is added, since container support (and arguably
lxc if you are doing "virtualization") is a fairly basic feature.

I'm concerned that within meta-virt we are creating an undocumented set of
DISTRO_FEATURES, and over using the concept. The items are already
controlled via PACKAGECONFIG, so users have flexibility from that point
of view.

We could also use IMAGE_FEATURES as an alternative to distro features,
since we are really only coordinating libvirt and the kernel when "lxc" is a
distro feature .. which arguably isn't distro wide.

Outside of the lxc change, everything looks pretty good, I'm going to ack
the individual patches to make it clear :)

Don't misunderstand my comments, there's no white/black answer here, I'm
just trying to spark a discussion to make sure the direction is clear
and we have
a pseudo-consensus :)

Cheers,

Bruce

>
> In this process, I stumbled upon some bugs/issues with the libvirt recipe,
> which are updated as well.
>
> Best regards,
> /Jonas
>
> Jonas Eriksson (5):
>   libvirt: Change license to LGPLv2.1+
>   libvirt: Fix selinux DISTRO_FEATURE check
>   Use a DISTRO_FEATURE to toggle LXC support
>   libvirt: Use DISTRO_FEATURES to enable Xen support
>   libvirt: Disable polkit driver when there is no x11
>
>  recipes-extended/libvirt/libvirt_1.2.1.bb      | 14 +++++++++-----
>  recipes-kernel/linux/linux-yocto_3.10.bbappend |  4 +++-
>  recipes-kernel/linux/linux-yocto_3.4.bbappend  |  4 +++-
>  recipes-kernel/linux/linux-yocto_3.8.bbappend  |  4 +++-
>  4 files changed, 18 insertions(+), 8 deletions(-)
>
> --
> 1.9.rc1
>
> _______________________________________________
> meta-virtualization mailing list
> meta-virtualization@yoctoproject.org
> https://lists.yoctoproject.org/listinfo/meta-virtualization



-- 
"Thou shalt not follow the NULL pointer, for chaos and madness await
thee at its end"

Re: [PATCH 4/5] libvirt: Use DISTRO_FEATURES to enable Xen support

[Bruce Ashfield]

On Wed, Feb 26, 2014 at 5:07 AM, Jonas Eriksson <jonas.eriksson@enea.com> wrote:
> The Xen drivers in libvirt is needed when the build file system image
> and kernel are running as dom0. The parts of the kernel needed to run as
> dom0 are controlled by DISTRO_FEATURES, and thus there is no need for
> the Xen drivers if DISTRO_FEATURES does not contain 'xen'.

Acked-by: Bruce Ashfield <bruce.ashfield@windriver.com>

>
> Signed-off-by: Jonas Eriksson <jonas.eriksson@enea.com>
> ---
>  recipes-extended/libvirt/libvirt_1.2.1.bb | 5 +++--
>  1 file changed, 3 insertions(+), 2 deletions(-)
>
> diff --git a/recipes-extended/libvirt/libvirt_1.2.1.bb b/recipes-extended/libvirt/libvirt_1.2.1.bb
> index dbc6ef3..a75cc45 100644
> --- a/recipes-extended/libvirt/libvirt_1.2.1.bb
> +++ b/recipes-extended/libvirt/libvirt_1.2.1.bb
> @@ -109,10 +109,11 @@ INITSCRIPT_PARAMS_${PN}-libvirtd = "defaults 72"
>  #PACKAGECONFIG ??= "xen libxl xen-inotify test remote libvirtd"
>
>  # full config
> -PACKAGECONFIG ??= "qemu yajl xen libxl xen-inotify uml openvz vmware vbox esx iproute2 \
> -                   polkit test remote macvtap libvirtd netcf udev python ebtables \
> +PACKAGECONFIG ??= "qemu yajl uml openvz vmware vbox esx iproute2 polkit test \
> +                   remote macvtap libvirtd netcf udev python ebtables \
>                     ${@base_contains('DISTRO_FEATURES', 'selinux', 'selinux', '', d)} \
>                     ${@base_contains('DISTRO_FEATURES', 'lxc', 'lxc', '', d)} \
> +                   ${@base_contains('DISTRO_FEATURES', 'xen', 'xen libxl xen-inotify', '', d)} \
>                    "
>
>  # enable,disable,depends,rdepends
> --
> 1.9.rc1
>
> _______________________________________________
> meta-virtualization mailing list
> meta-virtualization@yoctoproject.org
> https://lists.yoctoproject.org/listinfo/meta-virtualization



-- 
"Thou shalt not follow the NULL pointer, for chaos and madness await
thee at its end"

Re: [PATCH 1/5] libvirt: Change license to LGPLv2.1+

[Bruce Ashfield]

On Wed, Feb 26, 2014 at 5:07 AM, Jonas Eriksson <jonas.eriksson@enea.com> wrote:
> libvirt is to a large extent licensed under LGPLv2.1 (or later) rather
> than GPLv2. From my investigations, there are only some tests and a
> couple of files under tools are licensed as GPL. Other than that, files
> are generally licensed as LGPL. My recommendation is therefore to set
> the default license to LGPLv2.1+, and to add a LICENSE_${PN}-ptest that
> is set to LGPLv2.1+ & GPLv2+ if a ptest package is added. This is
> preferred to avoid having to specify explicit licenses for every other
> package, such as the auto generated libvirt-locale-XX packages.

Acked-by: Bruce Ashfield <bruce.ashfield@windriver.com>

>
> Signed-off-by: Jonas Eriksson <jonas.eriksson@enea.com>
> ---
>  recipes-extended/libvirt/libvirt_1.2.1.bb | 5 +++--
>  1 file changed, 3 insertions(+), 2 deletions(-)
>
> diff --git a/recipes-extended/libvirt/libvirt_1.2.1.bb b/recipes-extended/libvirt/libvirt_1.2.1.bb
> index 3b95486..d30f09d 100644
> --- a/recipes-extended/libvirt/libvirt_1.2.1.bb
> +++ b/recipes-extended/libvirt/libvirt_1.2.1.bb
> @@ -1,7 +1,8 @@
>  DESCRIPTION = "A toolkit to interact with the virtualization capabilities of recent versions of Linux."
>  HOMEPAGE = "http://libvirt.org"
> -LICENSE = "GPLv2+"
> -LIC_FILES_CHKSUM = "file://COPYING;md5=b234ee4d69f5fce4486a80fdaf4a4263"
> +LICENSE = "LGPLv2.1+"
> +LIC_FILES_CHKSUM = "file://COPYING;md5=b234ee4d69f5fce4486a80fdaf4a4263 \
> +                    file://COPYING.LESSER;md5=4fbd65380cdd255951079008b364516c"
>  SECTION = "console/tools"
>  PR = "r1"
>
> --
> 1.9.rc1
>
> _______________________________________________
> meta-virtualization mailing list
> meta-virtualization@yoctoproject.org
> https://lists.yoctoproject.org/listinfo/meta-virtualization



-- 
"Thou shalt not follow the NULL pointer, for chaos and madness await
thee at its end"

Re: [PATCH 2/5] libvirt: Fix selinux DISTRO_FEATURE check

[Bruce Ashfield]

On Wed, Feb 26, 2014 at 5:07 AM, Jonas Eriksson <jonas.eriksson@enea.com> wrote:
> It seems that a $ was forgotten in the DISTRO_FEATURE check for selinux.

ooops! :)


Acked-by: Bruce Ashfield <bruce.ashfield@windriver.com>

>
> Signed-off-by: Jonas Eriksson <jonas.eriksson@enea.com>
> ---
>  recipes-extended/libvirt/libvirt_1.2.1.bb | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/recipes-extended/libvirt/libvirt_1.2.1.bb b/recipes-extended/libvirt/libvirt_1.2.1.bb
> index d30f09d..76d3f42 100644
> --- a/recipes-extended/libvirt/libvirt_1.2.1.bb
> +++ b/recipes-extended/libvirt/libvirt_1.2.1.bb
> @@ -111,7 +111,7 @@ INITSCRIPT_PARAMS_${PN}-libvirtd = "defaults 72"
>  # full config
>  PACKAGECONFIG ??= "qemu yajl xen libxl xen-inotify uml openvz vmware vbox esx iproute2 \
>                     polkit lxc test remote macvtap libvirtd netcf udev python ebtables \
> -                   {@base_contains('DISTRO_FEATURES', 'selinux', 'selinux', '', d)} \
> +                   ${@base_contains('DISTRO_FEATURES', 'selinux', 'selinux', '', d)} \
>                    "
>
>  # enable,disable,depends,rdepends
> --
> 1.9.rc1
>
> _______________________________________________
> meta-virtualization mailing list
> meta-virtualization@yoctoproject.org
> https://lists.yoctoproject.org/listinfo/meta-virtualization



-- 
"Thou shalt not follow the NULL pointer, for chaos and madness await
thee at its end"

Re: [PATCH 5/5] libvirt: Disable polkit driver when there is no x11

[Bruce Ashfield]

On Wed, Feb 26, 2014 at 5:07 AM, Jonas Eriksson <jonas.eriksson@enea.com> wrote:
> To avoid having the default build of libvirt depend on virtual/x11
> through polkit and consolekit, check for x11 in DISTRO_FEATURES before
> enabling the polkit support in the default PKG_CONFIG.

Acked-by: Bruce Ashfield <bruce.ashfield@windriver.com>

>
> Signed-off-by: Jonas Eriksson <jonas.eriksson@enea.com>
> ---
>  recipes-extended/libvirt/libvirt_1.2.1.bb | 3 ++-
>  1 file changed, 2 insertions(+), 1 deletion(-)
>
> diff --git a/recipes-extended/libvirt/libvirt_1.2.1.bb b/recipes-extended/libvirt/libvirt_1.2.1.bb
> index a75cc45..2841fcb 100644
> --- a/recipes-extended/libvirt/libvirt_1.2.1.bb
> +++ b/recipes-extended/libvirt/libvirt_1.2.1.bb
> @@ -109,11 +109,12 @@ INITSCRIPT_PARAMS_${PN}-libvirtd = "defaults 72"
>  #PACKAGECONFIG ??= "xen libxl xen-inotify test remote libvirtd"
>
>  # full config
> -PACKAGECONFIG ??= "qemu yajl uml openvz vmware vbox esx iproute2 polkit test \
> +PACKAGECONFIG ??= "qemu yajl uml openvz vmware vbox esx iproute2 test \
>                     remote macvtap libvirtd netcf udev python ebtables \
>                     ${@base_contains('DISTRO_FEATURES', 'selinux', 'selinux', '', d)} \
>                     ${@base_contains('DISTRO_FEATURES', 'lxc', 'lxc', '', d)} \
>                     ${@base_contains('DISTRO_FEATURES', 'xen', 'xen libxl xen-inotify', '', d)} \
> +                   ${@base_contains('DISTRO_FEATURES', 'x11', 'polkit', '', d)} \
>                    "
>
>  # enable,disable,depends,rdepends
> --
> 1.9.rc1
>
> _______________________________________________
> meta-virtualization mailing list
> meta-virtualization@yoctoproject.org
> https://lists.yoctoproject.org/listinfo/meta-virtualization



-- 
"Thou shalt not follow the NULL pointer, for chaos and madness await
thee at its end"

Re: [PATCH 0/5] libvirt fixes and a kernel update

[Jonas Eriksson]

Hi,

On Wed, Feb 26, 2014 at 14:33:21 -0500 Bruce Ashfield wrote:
> On Wed, Feb 26, 2014 at 5:07 AM, Jonas Eriksson <jonas.eriksson@enea.com> wrote:
> > The ultimate goal of this patch series was to use a DISTRO_FEATURE to include
> > the lxc.scc configuration in the kernel to be built. This is done to avoid that
> > the kernel configuration gets additional code paths ( =y options ) when adding
> > the meta-virtualization layer. The configurations that enables modules are left
> > untouched.
> 
> I've never been particularly concerned about the extra functionality that is
> introduced when meta-virt is added, since container support (and arguably
> lxc if you are doing "virtualization") is a fairly basic feature.

I can see your concern, and it was something I weighed in when
adding lxc to DISTRO_FEATURES. My train of thought ended up
something like this:

- The LXC kernel conf parameters has some performance impact

- The Xen kernel conf parameters does too, and those are enabled
  through DISTRO_FEATURES

- DISTRO_FEATURES it is

> I'm concerned that within meta-virt we are creating an undocumented set of
> DISTRO_FEATURES, and over using the concept. The items are already
> controlled via PACKAGECONFIG, so users have flexibility from that point
> of view.

I'm not sure what your point is here, since you Ack:ed the other
PACKAGECONFIG-@base_contains patches. But I would guess that it
is to discourage from creating DISTRO_FEATURES just to create a
nice default PACKAGECONFIG. As an answer to that, my concern was
mainly rooted in the performance impact of the LXC kernel
configuration. After that, I just went with "It's there now,
might as well use it" in the PACKAGECONFIG, like for the other
patches.

> We could also use IMAGE_FEATURES as an alternative to distro features,
> since we are really only coordinating libvirt and the kernel when "lxc" is a
> distro feature .. which arguably isn't distro wide.

True enough. Would this motivate a change for the current Xen
DISTRO_FEATURE because of consistency? If anyone on the list
thinks it's obvious why Xen should be a DISTRO_FEATURE and not an
IMAGE_FEATURE, here would be a great place to explain it :-)

> Outside of the lxc change, everything looks pretty good, I'm going to ack
> the individual patches to make it clear :)

Great, thanks!

> Don't misunderstand my comments, there's no white/black answer here, I'm
> just trying to spark a discussion to make sure the direction is clear
> and we have a pseudo-consensus :)

Sounds like a good plan.

So, the way forward. I can create a new IMAGE_FEATURES-based LXC
patch and repost patches 3-5 in the series, unless there are some
other comments before ~end of business at UTC+1.

Thanks,
/Jonas

Re: [PATCH 0/5] libvirt fixes and a kernel update

[Jonas Eriksson]

Hi again,

On Thu, Feb 27, 2014 at 09:00:39 +0100 Jonas Eriksson wrote:
> On Wed, Feb 26, 2014 at 14:33:21 -0500 Bruce Ashfield wrote:
> > I've never been particularly concerned about the extra functionality that is
> > introduced when meta-virt is added, since container support (and arguably
> > lxc if you are doing "virtualization") is a fairly basic feature.
> 
> I can see your concern, and it was something I weighed in when
> adding lxc to DISTRO_FEATURES. My train of thought ended up
> something like this:
> 
> - The LXC kernel conf parameters has some performance impact
> 
> - The Xen kernel conf parameters does too, and those are enabled
>   through DISTRO_FEATURES
> 
> - DISTRO_FEATURES it is

So, a follow-up on this. It seems that there were some confusion
around whether or not extra kernel configuration parameters where
needed to run LXC. It turns out that the answer is no, and that
the linux-yocto kernel by default has all the bells and whistles
needed for LXC as it is. This of course means that the need for
an extra *_FEATURES parameter simply disappeared.

Now the only question is why you did not yell at me for missing
this fact, Bruce ;)

I've now reposted the series without the offending patch.

Thanks and sorry for the confusion,
/Jonas

Re: [PATCH 0/5] libvirt fixes and a kernel update

[Bruce Ashfield]

On Thu, Feb 27, 2014 at 3:00 AM, Jonas Eriksson <jonas.eriksson@enea.com> wrote:
> Hi,
>
> On Wed, Feb 26, 2014 at 14:33:21 -0500 Bruce Ashfield wrote:
>> On Wed, Feb 26, 2014 at 5:07 AM, Jonas Eriksson <jonas.eriksson@enea.com> wrote:
>> > The ultimate goal of this patch series was to use a DISTRO_FEATURE to include
>> > the lxc.scc configuration in the kernel to be built. This is done to avoid that
>> > the kernel configuration gets additional code paths ( =y options ) when adding
>> > the meta-virtualization layer. The configurations that enables modules are left
>> > untouched.
>>
>> I've never been particularly concerned about the extra functionality that is
>> introduced when meta-virt is added, since container support (and arguably
>> lxc if you are doing "virtualization") is a fairly basic feature.
>
> I can see your concern, and it was something I weighed in when
> adding lxc to DISTRO_FEATURES. My train of thought ended up
> something like this:

Glad to hear that it was thought about .. that's the important thing!

>
> - The LXC kernel conf parameters has some performance impact
>
> - The Xen kernel conf parameters does too, and those are enabled
>   through DISTRO_FEATURES

If you search the archives, you'll see where I've both suggested image
features and different machines to break up the Xen functionality. So
at least I'm consistent on that front :)

>
> - DISTRO_FEATURES it is
>
>> I'm concerned that within meta-virt we are creating an undocumented set of
>> DISTRO_FEATURES, and over using the concept. The items are already
>> controlled via PACKAGECONFIG, so users have flexibility from that point
>> of view.
>
> I'm not sure what your point is here, since you Ack:ed the other
> PACKAGECONFIG-@base_contains patches. But I would guess that it
> is to discourage from creating DISTRO_FEATURES just to create a
> nice default PACKAGECONFIG. As an answer to that, my concern was

Yes, that's pretty much it. And the existing ones are grandfathered in at
the moment :)

> mainly rooted in the performance impact of the LXC kernel
> configuration. After that, I just went with "It's there now,
> might as well use it" in the PACKAGECONFIG, like for the other
> patches.

I'd have done the same thing.

>
>> We could also use IMAGE_FEATURES as an alternative to distro features,
>> since we are really only coordinating libvirt and the kernel when "lxc" is a
>> distro feature .. which arguably isn't distro wide.
>
> True enough. Would this motivate a change for the current Xen
> DISTRO_FEATURE because of consistency? If anyone on the list
> thinks it's obvious why Xen should be a DISTRO_FEATURE and not an
> IMAGE_FEATURE, here would be a great place to explain it :-)

I'm hoping to hear as well, I've locally converted Xen to an image feature
and it worked fine :)

Bruce

>
>> Outside of the lxc change, everything looks pretty good, I'm going to ack
>> the individual patches to make it clear :)
>
> Great, thanks!
>
>> Don't misunderstand my comments, there's no white/black answer here, I'm
>> just trying to spark a discussion to make sure the direction is clear
>> and we have a pseudo-consensus :)
>
> Sounds like a good plan.
>
> So, the way forward. I can create a new IMAGE_FEATURES-based LXC
> patch and repost patches 3-5 in the series, unless there are some
> other comments before ~end of business at UTC+1.
>
> Thanks,
> /Jonas



-- 
"Thou shalt not follow the NULL pointer, for chaos and madness await
thee at its end"

Re: [PATCH 0/5] libvirt fixes and a kernel update

[Bruce Ashfield]

On Thu, Feb 27, 2014 at 11:18 AM, Jonas Eriksson
<jonas.eriksson@enea.com> wrote:
> Hi again,
>
> On Thu, Feb 27, 2014 at 09:00:39 +0100 Jonas Eriksson wrote:
>> On Wed, Feb 26, 2014 at 14:33:21 -0500 Bruce Ashfield wrote:
>> > I've never been particularly concerned about the extra functionality that is
>> > introduced when meta-virt is added, since container support (and arguably
>> > lxc if you are doing "virtualization") is a fairly basic feature.
>>
>> I can see your concern, and it was something I weighed in when
>> adding lxc to DISTRO_FEATURES. My train of thought ended up
>> something like this:
>>
>> - The LXC kernel conf parameters has some performance impact
>>
>> - The Xen kernel conf parameters does too, and those are enabled
>>   through DISTRO_FEATURES
>>
>> - DISTRO_FEATURES it is
>
> So, a follow-up on this. It seems that there were some confusion
> around whether or not extra kernel configuration parameters where
> needed to run LXC. It turns out that the answer is no, and that
> the linux-yocto kernel by default has all the bells and whistles
> needed for LXC as it is. This of course means that the need for
> an extra *_FEATURES parameter simply disappeared.

Right you are. We've beefed it up by default in many cases .. if
they are common enough.

>
> Now the only question is why you did not yell at me for missing
> this fact, Bruce ;)

:) we'll queue up some karma :)

>
> I've now reposted the series without the offending patch.

The series looks good, if no one objects by tomorrow .. I can do the merge.

Bruce

>
> Thanks and sorry for the confusion,
> /Jonas



-- 
"Thou shalt not follow the NULL pointer, for chaos and madness await
thee at its end"