[OE-core] [PATCH 2/4 V2] systemd: Upgrade 219 -> 224

Khem Raj raj.khem at gmail.com
Fri Aug 21 15:34:59 UTC 2015


On Fri, Aug 21, 2015 at 3:14 AM, ChenQi <Qi.Chen at windriver.com> wrote:
> Hi Khem,
>
> I built core-image-minimal for qemuarm64.
> There's a lot of failures and warnings at boot time and the system boots
> into rescue mode.

Can you paste the boot logs somewhere ?

> And I also verified 199 has no such problem.
>
> Best Regards,
> Chen Qi
>
>
> On 08/21/2015 09:46 AM, Randy MacLeod wrote:
>>
>> On 2015-08-20 09:17 PM, Khem Raj wrote:
>>>
>>> On Thu, Aug 20, 2015 at 6:04 AM, Philip Balister <philip at balister.org>
>>> wrote:
>>>>
>>>> On 08/19/2015 10:21 PM, Khem Raj wrote:
>>>>>
>>>>> On Wed, Aug 19, 2015 at 12:55 PM, Burton, Ross <ross.burton at intel.com>
>>>>> wrote:
>>>>>>
>>>>>>
>>>>>> On 17 August 2015 at 16:41, Khem Raj <raj.khem at gmail.com> wrote:
>>>>>>>
>>>>>>>
>>>>>>> There are many reasons, for me its overlay support for
>>>>>>> systemd-nspawn,
>>>>>>> networkd has got many new features that is now usable w.r.t. IP
>>>>>>> forwarding,
>>>>>>> vxlan etc.
>>>>>>> and it has many bug fixed in those 2000 odd commits since 219, no
>>>>>>> different then any other package upgrades we do in general it keep
>>>>>>> the
>>>>>>> upgrade workload lower as we roll the releases.
>>>>>>> Any specific concerns ?
>>>>>>
>>>>>>
>>>>>>
>>>>>> Can we get an updated patch with a clearer commit log?
>>>>
>>>>
>>>> I've also heard that as systemd evolves, more binaries are getting added
>>>> to the base package that should be packaged separately for people
>>>> interested in small images. I do not have personal experience here, but
>>>> wanted to pass along the feedback.
>>>>
>>>> We should look at buildhistory packaging differences when we do
>>>> upgrades.
>>>
>>>
>>> Here is the diff between files in 219 and 224, if you want to know
>>> more I can paste more info just let me know.
>>>
>>> https://gist.github.com/kraj/2a066973a5e5cf83ed24
>>>
>>> sizes have gone up on daemons, no new daemons besides some new service
>>> files
>>> and scripts are added
>>>
>>
>> ubu-15.10 will use v224 as well:
>> http://cdimage.ubuntu.com/daily-live/current/wily-desktop-amd64.manifest
>> and Arch and a couple other distro are using this version already:
>>    http://pkgs.org/download/systemd
>>
>> v224 was released 3 weeks ago:
>> ---
>> $ git log -1 v224
>> commit b2a0ac5e5b29c73ca7c0da23369a4769d5a91ddd
>> ...
>> Date:   Fri Jul 31 18:56:38 2015 +0200
>> ---
>>
>> Khem's summary of the binaries is reassuring given that there
>> has been lots of churn:
>>
>> $ git log --oneline v219..v224 |  wc -l
>> 2167
>> $ git diff v219..v224 | diffstat | tail -1
>>  1367 files changed, 109907 insertions(+), 166577 deletions(-)
>>
>> I'm skimming the NEWS file from 219->224 and I haven't seen
>> anything that concerns me yet:
>>    https://github.com/systemd/systemd/blob/master/NEWS
>> Qi,
>> Please take a closer look at the NEWS file,
>> review Khem's uprev and build and boot qemuarm64, qemuppc
>> when you have time. Send partial feedback today even if you
>> just review the commit and NEWS file.
>>
>> Qi may not be able to do that in the next couple of day
>> due to SDK work.
>>
>> Khem,
>> What testing have you done so far?
>> Any ptest?
>> What toolchain version are you building with, btw?
>>
>>
>> It's late in M3 and we still have the kernel and toolchain coming in
>> but unless we hear of known problems with v224, let's go for it
>> once the new toolchain and kernel have settled.
>>
>



More information about the Openembedded-core mailing list