Re: [cip-dev] Generic x86 kernel configration (RE: [isar-cip-core][PATCH] Add support Plat'Home OpenBlocks IoT VX2)

[Gylstorff Quirin <quirin.gylstorff@siemens.com>]

On 4/23/24 3:18 PM, Nobuhiro Iwamatsu via lists.cip-project.org wrote:
> Hi Jan,
> 
>> This script will likely prioritize the choice of later-provided configs over those
>> earlier listed. But we rather generally want that features enabled in one config
>> are also enabled in the merged ones (superset).
>> How could we achieve this? Probably most of the comments I would have on
>> the ipc227e diff would relate to this.
> 
> I see. I think we need to check the configuration of each device to achieve that.
> First I create a list of configs.

I look into using [merge_config.py]  instead of [merge_config.sh].
This looks promising as it takes the Kconfig of the kernel into account.
It still requires an make savedefconfig afterwards. `merge_config.py` 
shows warnings like:

Merged configuration 
'./cip-kernel-config/5.10.y-cip/x86/siemens_ipc227e_defconfig'
Merged configuration 
'./cip-kernel-config/5.10.y-cip/x86/plathome_obsvx2_defconfig'
warning: The int symbol BASE_SMALL (defined at init/Kconfig:2043) is 
being evaluated in a logical context somewhere. It will always evaluate 
to n.
warning: The int symbol ARCH_MMAP_RND_BITS_DEFAULT (defined at 
arch/Kconfig:732) is being evaluated in a logical context somewhere. It 
will always evaluate to n.
warning: The int symbol ARCH_MMAP_RND_COMPAT_BITS_DEFAULT (defined at 
arch/Kconfig:766) is being evaluated in a logical context somewhere. It 
will always evaluate to n.
warning: NFSD_V2_ACL (defined at fs/nfsd/Kconfig:34) has direct 
dependencies NFSD && NETWORK_FILESYSTEMS with value m, but is currently 
being y-selected by the following symbols:
  - NFSD_V3_ACL (defined at fs/nfsd/Kconfig:47), with value y, direct 
dependencies NFSD_V3 && NETWORK_FILESYSTEMS (value: y), and select 
condition NFSD_V3 && NETWORK_FILESYSTEMS (value: y)
No change to configuration in 
'/home/gylstorff/02_repos/cip-kernel-config/5.10.y-cip/x86/ipc227e_obsvx2.merged'
warning: PREEMPT_VOLUNTARY (defined at kernel/Kconfig.preempt:20) was 
assigned the value 'y' but got the value 'n' -- check dependencies
warning: KVM (defined at arch/x86/kvm/Kconfig:21) was assigned the value 
'm' but got the value 'n' -- check dependencies
warning: KVM_INTEL (defined at arch/x86/kvm/Kconfig:77) was assigned the 
value 'm' but got the value 'n' -- check dependencies


[merge_config.py]: 
https://github.com/ulfalizer/Kconfiglib/blob/master/examples/merge_config.py
[merge_config.sh]: 
https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/scripts/kconfig/merge_config.sh

Best regards

Quirin

> 
> Best regards,
>    Nobuhiro
> 
>> -----Original Message-----
>> From: Jan Kiszka <jan.kiszka@siemens.com>
>> Sent: Friday, April 12, 2024 1:51 AM
>> To: iwamatsu nobuhiro(岩松 信洋 ○ＤＩＴＣ□ＤＩＴ○ＯＳＴ)
>> <nobuhiro1.iwamatsu@toshiba.co.jp>; schilling.benjamin@siemens.com;
>> cip-dev@lists.cip-project.org
>> Cc: Chris.Paterson2@renesas.com; minmin@plathome.co.jp
>> Subject: Re: Generic x86 kernel configration (RE: [isar-cip-core][PATCH] Add
>> support Plat'Home OpenBlocks IoT VX2)
>>
>> Hi Nobuhiro,
>>
>> thanks for picking this up!
>>
>> On 05.04.24 09:39, nobuhiro1.iwamatsu@toshiba.co.jp wrote:
>>> Hi all,
>>>
>>> First of all I created generic kernel configuration for x86 CIP reference
>> hardware and performed boot tests.
>>> The configurations of each device in cip-kernel-config are merged using
>> KERNEL/scripts/kconfig/merge_config.sh.
>>
>> This script will likely prioritize the choice of later-provided configs over those
>> earlier listed. But we rather generally want that features enabled in one config
>> are also enabled in the merged ones (superset).
>> How could we achieve this? Probably most of the comments I would have on
>> the ipc227e diff would relate to this.
>>
>> Jan
>>
>>> Also, booting is fine[2][3]. But there are a lot of differences in the
>>> configurations[0][1], so I think we need to decide what to do with them.
>>>
>>> About M-COM x86, but I haven't investigated it yet.
>>>
>>> Best regards,
>>>    Nobuhiro
>>>
>>> Merged config:
>>>
>>> https://gitlab.com/iwamatsu/cip-merge-config-work/-/raw/master/config
>>> Diff with ipc227e original config [0]:
>>>
>>>
>> https://gitlab.com/iwamatsu/cip-merge-config-work/-/raw/master/siemens
>>> _ipc227e_config.v6.1.83-cip18.diff
>>> Diff with OpenBlocks IoT VX2 original config[1]:
>>>
>>>
>> https://gitlab.com/iwamatsu/cip-merge-config-work/-/blob/master/platho
>>> me_obsvx2_config.v6.1.83-cip18.diff
>>> Kernel size:
>>>      7760544 bytes
>>>
>>> ipc227e boot log (merged)[2]
>>>      https://lava.ciplatform.org/scheduler/job/1113259
>>>
>>>
>> https://gitlab.com/iwamatsu/cip-merge-config-work/-/raw/master/ipc227e
>>> -merged.log
>>> ipc227e boot log (NO merged)
>>>      https://lava.ciplatform.org/scheduler/job/1113356
>>>
>> https://gitlab.com/iwamatsu/cip-merge-config-work/-/raw/master/ipc227e-
>> NO-merged.log
>>>      kernel size: 7251072 bytes
>>> ipc227e boot log (diff)
>>>
>>>
>> https://gitlab.com/iwamatsu/cip-merge-config-work/-/blob/master/ipc227
>>> e.log.diff
>>>
>>> OpenBlocks IoT VX2 boot log (merged)[3]
>>>      https://lava.ciplatform.org/scheduler/job/1113301
>>>
>>>
>> https://gitlab.com/iwamatsu/cip-merge-config-work/-/raw/master/obs-mer
>>> ged.log OpenBlocks IoT VX2 boot log (NO merged)
>>>      https://lava.ciplatform.org/scheduler/job/1113303
>>>
>> https://gitlab.com/iwamatsu/cip-merge-config-work/-/raw/master/obs-no-
>> merge.log
>>>      kernel size: 9203296 bytes
>>> OpenBlocks IoT VX2 boot log (diff)
>>>
>>>
>> https://gitlab.com/iwamatsu/cip-merge-config-work/-/raw/master/obs.log
>>> .diff
>>>
>>>
>>>> -----Original Message-----
>>>> From: Jan Kiszka <jan.kiszka@siemens.com>
>>>> Sent: Tuesday, March 26, 2024 3:52 PM
>>>> To: Benjamin Schilling <schilling.benjamin@siemens.com>;
>>>> cip-dev@lists.cip-project.org; iwamatsu nobuhiro(岩松 信洋 ○ＤＩＴＣ□
>> Ｄ
>>>> ＩＴ○ＯＳＴ) <nobuhiro1.iwamatsu@toshiba.co.jp>
>>>> Subject: Re: [isar-cip-core][PATCH] Add support Plat'Home OpenBlocks
>>>> IoT
>>>> VX2
>>>>
>>>> On 22.03.24 12:39, Benjamin Schilling wrote:
>>>>> Hi all,
>>>>>
>>>>> I'm joining the discussion to start integrating the M-COM x86 support.
>>>>> What's the status on the generic x86-64 target?
>>>>>
>>>>
>>>> Nobuhiro wrote recently he has a prototype. Maybe he can share to
>>>> finalize it together.
>>>>
>>>> Jan
>>>>
>>>> --
>>>> Siemens AG, Technology
>>>> Linux Expert Center
>>
>> --
>> Siemens AG, Technology
>> Linux Expert Center
>>
>>
>> -=-=-=-=-=-=-=-=-=-=-=-
>> Links: You receive all messages sent to this group.
>> View/Reply Online (#15741): https://lists.cip-project.org/g/cip-dev/message/15741
>> Mute This Topic: https://lists.cip-project.org/mt/105344534/1753640
>> Group Owner: cip-dev+owner@lists.cip-project.org
>> Unsubscribe: https://lists.cip-project.org/g/cip-dev/leave/8129121/1753640/1405269326/xyzzy [quirin.gylstorff@siemens.com]
>> -=-=-=-=-=-=-=-=-=-=-=-
>>