[OE-core] [PATCH 0/5] linux-yocto: consolidated pull request

Bruce Ashfield bruce.ashfield at windriver.com
Sat Aug 24 14:59:09 UTC 2013


On 13-08-24 9:50 AM, Saul Wold wrote:
> On 08/24/2013 06:45 AM, Bruce Ashfield wrote:
>> On 13-08-24 9:34 AM, Bruce Ashfield wrote:
>>> On 13-08-24 2:33 AM, Saul Wold wrote:
>>>> On 08/23/2013 11:08 AM, Bruce Ashfield wrote:
>>>>> Richard/Saul,
>>>>>
>>>>> Here's my consolidated set of changes that represent the biggest chunk
>>>>> of work for the Yocto 1.5 release. The obvious change is that the 3.10
>>>>> kernel is now the latest linux-yocto kernel (and linux-yocto-dev will
>>>>> mark on). I've built and booted core-image-sato for all the oe-core
>>>>> qemu machines.
>>>>>
>>>>> 3.10 also the latest LTSI and will be updated with that content
>>>>> when it
>>>>> becomes available.
>>>>>
>>>>> Other features of the 3.10 kernel include:
>>>>>
>>>>>    - refreshed -rt support
>>>>>    - refreshed yaffs2, aufs3
>>>>>    - cryptodev
>>>>>    - bfs, edf, and OCF staged features
>>>>>    - scrubbed and updated meta data for v3.10
>>>>>    - improved tools support for meta data updates and queue
>>>>> maintenance
>>>>>    - patch carry forward from all previous linux-yocto kernels and
>>>>>      configuration.
>>>>>
>>>>> I've also tested linux-yocto-rt and linux-yocto-tiny for the new,
>>>>> shiny tree.
>>>>>
>>>>> I've also bumped the linux-libc-headers to 3.10 and have fixed
>>>>> gst-plugins-good
>>>>> to adapt to a changed #define in the exported headers.
>>>>>
>>>>> Other parts of this update include an uprev of guilt, and in fact, a
>>>>> restoration of "normal" guil operations, since the kernel tree is now
>>>>> manipulated and maintained via a smaller, more contained script. This
>>>>> was necessary for both usability reasons and the fact that all of
>>>>> guilts
>>>>> protections were being removed for the kernel's use .. which makes you
>>>>> ask the quesetion "why use it". It was time for a change.
>>>>>
>>>>> The patches are pretty self explanitory and all contain detailed long
>>>>> logs .. but the summary is:
>>>>>
>>>>>     [PATCH 1/5] kern-tools: usability, bug fixes and no guilt
>>>>>     [PATCH 2/5] guilt: update to latest git version
>>>>>     [PATCH 3/5] linux-libc-headers: update to v3.10
>>>> One of the changes in 3.10 seems to have introduced a conflicting
>>>> definition of ptrace_peeksiginfo_args in linux/ptrace.h
>>>>
>>>>>  In file included from ptrace.h:22:0,
>>>>> |                  from ptrace05.c:38:
>>>>> |
>>>>> /srv/ssd/sgw/builds/world/tmp/sysroots/qemux86-64/usr/include/linux/ptrace.h:58:8:
>>>>>
>>>>>
>>>>>
>>>>> error: redefinition of 'struct ptrace_peeksiginfo_args'
>>>>> |  struct ptrace_peeksiginfo_args {
>>>>> |         ^
>>>>> | In file included from ptrace.h:9:0,
>>>>> |                  from ptrace05.c:38:
>>>>> |
>>>>> /srv/ssd/sgw/builds/world/tmp/sysroots/qemux86-64/usr/include/sys/ptrace.h:191:8:
>>>>>
>>>>>
>>>>>
>>>>> note: originally defined here
>>>>> |  struct ptrace_peeksiginfo_args
>>>>> |         ^
>>>>
>>>> This affects ltp, valgrind and strace.
>>>>
>>>> Please check the autobuilder as I am building with this change set and
>>>> expect these failures.
>>>
>>> I'm unfortunately off for the weekend and won't have cycles to look into
>>> any failures until Monday.
>>
>> As it turns out, I may have some time to work on this tonight, since I
>> have to delete the 3.8 yocto recipe and get that patch out.
>>
>> So I've launched a LSB build in the background and will check on it
>> later.
>>
> Just FYI, it might not show up in an LSB build, since those tools
> valgrind, ltp and strace are more Dev tools, so a world build would be
> better.
>
> Also, not sure why but ARM failed to patch 3.8 kernel overnight on the
> AB. (lots of red).
>
> I just restarted AB with out the 3.10 headers.

ok .. this time I'm REALLY going to step away for a few hours.

The mismatch is between the eglibc 2.18 sys/ptrace.h and the kernel's
definition of the structure. I wouldn't really call it a mismatch,
since they are both the same size and field types .. both simply
define it now.

I'm adding Khem: What's the right thing to do here?

I can obviously hack the kernel to remove the definition, but that make
the linux-yocto 3.10 the only kernel that will work against yocto 1.5's
eglibc for building any applications that use ptrace.h.

Falling back to the 3.8 headers will work, but again, we are avoiding
the problem and any end user that does bump to a newer set of libc-headers
will see the problem.

I'll wait for consensus on the approach and add a change if required
into my linux-yocto 3.8 removal patch tonight.

Bruce

>
> Sau!
>
>> If anyone else does get a fix for this, make sure to cc' me on any
>> changes, so I won't spend time trying to fix it a second time! :)
>>
>> Bruce
>>
>>>
>>> So if you want to leave the old headers in place, that's an option.
>>>
>>> I fixed the one issue that I saw with Sato .. but I am after all one
>>> person and can't possibly build and boot all the qemu boards with
>>> all image combinations! This is after all, where I'd expect other
>>> people to help out and fix build issues.
>>>
>>> Cheers,
>>>
>>> Bruce
>>>
>>>>
>>>> Sau!
>>>>
>>>> Sau!
>>>>
>>>>>     [PATCH 4/5] gst-plugins-good: fix 3.10 libc-headers build failure
>>>>>     [PATCH 5/5] linux-yocto: introduce v3.10
>>>>>
>>>>> ** There is one pending activity. As part of thsi update to 3.10, we
>>>>> want
>>>>> to move to the 3.4 and 3.10 kernel, dropping 3.8 support. But the
>>>>> routerstationpro BSP is still using it as its preferred version. I'm
>>>>> a one man show on this update, and I wasn't able to get the 100+
>>>>> patches
>>>>> ported to 3.10 AND get this update out before the M4 cutoff.
>>>>>
>>>>> ** I'm still attempting to udpate all the reference boards, and
>>>>> once done, I'll send a follow up patch.
>>>>>
>>>>> ** Please test, I've done my best, but things can slip through. We'll
>>>>> work issues in the stabilization period.
>>>>>
>>>>> Cheers,
>>>>>
>>>>> Bruce
>>>>>
>>>>>
>>>>> The following changes since commit
>>>>> 501e1a321d15abd712e3a6ee6ccdb8e21c3d4ea4:
>>>>>
>>>>>    bitbake: prserv/serv: Fix pid file removal (2013-08-23 17:43:40
>>>>> +0100)
>>>>>
>>>>> are available in the git repository at:
>>>>>
>>>>>    git://git.pokylinux.org/poky-contrib zedd/kernel
>>>>>    http://git.pokylinux.org/cgit.cgi/poky-contrib/log/?h=zedd/kernel
>>>>>
>>>>> Bruce Ashfield (5):
>>>>>    kern-tools: usability, bug fixes and no guilt
>>>>>    guilt: update to latest git version
>>>>>    linux-libc-headers: update to v3.10
>>>>>    gst-plugins-good: fix 3.10 libc-headers build failure
>>>>>    linux-yocto: introduce v3.10
>>>>>
>>>>>   meta/classes/kernel-yocto.bbclass                  |    2 +
>>>>>   meta/conf/distro/include/tcmode-default.inc        |    2 +-
>>>>>   .../guilt/files/guilt-import-commit.patch          |   96 ------
>>>>>   meta/recipes-devtools/guilt/files/guilt-init.patch |   25 --
>>>>>   meta/recipes-devtools/guilt/files/guilt-pop.patch  |   73 -----
>>>>>   .../guilt/files/guilt-push-no-series.patch         |   29 --
>>>>>   meta/recipes-devtools/guilt/files/guilt-push.patch |   42 ---
>>>>>   .../guilt/files/guilt-set-git_exec_path.patch      |   29 --
>>>>>   ...lt-update-supported-git-versions-to-1.8.x.patch |   28 --
>>>>>   meta/recipes-devtools/guilt/files/guilt.patch      |  319
>>>>> --------------------
>>>>>   .../guilt/files/improve_auto_header_gen.patch      |   71 -----
>>>>>   .../files/make_git_commands_conditional.patch      |   48 ---
>>>>>   .../guilt/files/optional_head_check.patch          |   60 ----
>>>>>   .../guilt/files/uninstall_force.patch              |   14 -
>>>>>   .../{guilt-native_0.33.bb => guilt-native_git.bb}  |   21 +-
>>>>>   .../kern-tools/kern-tools-native_git.bb            |    6 +-
>>>>>   .../linux-libc-headers/linux-libc-headers_3.10.bb  |    7 +
>>>>>   .../linux-libc-headers/linux-libc-headers_3.8.bb   |    8 -
>>>>>   meta/recipes-kernel/linux/linux-yocto-rt_3.10.bb   |   30 ++
>>>>>   meta/recipes-kernel/linux/linux-yocto-tiny_3.10.bb |   25 ++
>>>>>   meta/recipes-kernel/linux/linux-yocto_3.10.bb      |   30 ++
>>>>>   ...define-V4L2_CID_HCENTER-and-V4L2_CID_VCEN.patch |   45 +++
>>>>>   .../gstreamer/gst-plugins-good_0.10.31.bb          |    1 +
>>>>>   23 files changed, 150 insertions(+), 861 deletions(-)
>>>>>   delete mode 100644
>>>>> meta/recipes-devtools/guilt/files/guilt-import-commit.patch
>>>>>   delete mode 100644
>>>>> meta/recipes-devtools/guilt/files/guilt-init.patch
>>>>>   delete mode 100644 meta/recipes-devtools/guilt/files/guilt-pop.patch
>>>>>   delete mode 100644
>>>>> meta/recipes-devtools/guilt/files/guilt-push-no-series.patch
>>>>>   delete mode 100644
>>>>> meta/recipes-devtools/guilt/files/guilt-push.patch
>>>>>   delete mode 100644
>>>>> meta/recipes-devtools/guilt/files/guilt-set-git_exec_path.patch
>>>>>   delete mode 100644
>>>>> meta/recipes-devtools/guilt/files/guilt-update-supported-git-versions-to-1.8.x.patch
>>>>>
>>>>>
>>>>>
>>>>>
>>>>>   delete mode 100644 meta/recipes-devtools/guilt/files/guilt.patch
>>>>>   delete mode 100644
>>>>> meta/recipes-devtools/guilt/files/improve_auto_header_gen.patch
>>>>>   delete mode 100644
>>>>> meta/recipes-devtools/guilt/files/make_git_commands_conditional.patch
>>>>>   delete mode 100644
>>>>> meta/recipes-devtools/guilt/files/optional_head_check.patch
>>>>>   delete mode 100644
>>>>> meta/recipes-devtools/guilt/files/uninstall_force.patch
>>>>>   rename meta/recipes-devtools/guilt/{guilt-native_0.33.bb =>
>>>>> guilt-native_git.bb} (45%)
>>>>>   create mode 100644
>>>>> meta/recipes-kernel/linux-libc-headers/linux-libc-headers_3.10.bb
>>>>>   delete mode 100644
>>>>> meta/recipes-kernel/linux-libc-headers/linux-libc-headers_3.8.bb
>>>>>   create mode 100644 meta/recipes-kernel/linux/linux-yocto-rt_3.10.bb
>>>>>   create mode 100644
>>>>> meta/recipes-kernel/linux/linux-yocto-tiny_3.10.bb
>>>>>   create mode 100644 meta/recipes-kernel/linux/linux-yocto_3.10.bb
>>>>>   create mode 100644
>>>>> meta/recipes-multimedia/gstreamer/gst-plugins-good-0.10.31/0001-v4l2_calls-define-V4L2_CID_HCENTER-and-V4L2_CID_VCEN.patch
>>>>>
>>>>>
>>>>>
>>>>>
>>>>>
>>>
>>
>>
>>




More information about the Openembedded-core mailing list