All of lore.kernel.org
 help / color / mirror / Atom feed
* yocto-bsp and kconf-check
@ 2013-08-29 15:36 Jate Sujjavanich
  2013-08-29 15:51 ` Bruce Ashfield
  2013-08-29 19:46 ` Tom Zanussi
  0 siblings, 2 replies; 7+ messages in thread
From: Jate Sujjavanich @ 2013-08-29 15:36 UTC (permalink / raw)
  To: 'yocto@yoctoproject.org'

It appears that the yocto-bsp script generates a kernel configuration that creates some warnings during kern-tools' kconf_check. The {{machine}}.cfg file has many non-hardware options, therefore the script warns.

It seems like many of these should be in the standard kernel configuration. Is this correct, and does the yocto-bsp data need to be updated?

- Jate


^ permalink raw reply	[flat|nested] 7+ messages in thread

* Re: yocto-bsp and kconf-check
  2013-08-29 15:36 yocto-bsp and kconf-check Jate Sujjavanich
@ 2013-08-29 15:51 ` Bruce Ashfield
  2013-08-29 18:56   ` Jate Sujjavanich
  2013-08-29 19:46 ` Tom Zanussi
  1 sibling, 1 reply; 7+ messages in thread
From: Bruce Ashfield @ 2013-08-29 15:51 UTC (permalink / raw)
  To: Jate Sujjavanich; +Cc: 'yocto@yoctoproject.org'

On 13-08-29 11:36 AM, Jate Sujjavanich wrote:
> It appears that the yocto-bsp script generates a kernel configuration that creates some warnings during kern-tools' kconf_check. The {{machine}}.cfg file has many non-hardware options, therefore the script warns.
>
> It seems like many of these should be in the standard kernel configuration. Is this correct, and does the yocto-bsp data need to be updated?

Which kernel version ? But the answer is not necessarily, if a machine 
config
is specifying something that hasn't been tagged "hardware" or that
has a specified exception, you get a warning.

It's not about them being common or not, it's about BSPs following a
base policy versus having wild, per-board behaviour.

Cheers,

Bruce

>
> - Jate
> _______________________________________________
> yocto mailing list
> yocto@yoctoproject.org
> https://lists.yoctoproject.org/listinfo/yocto
>



^ permalink raw reply	[flat|nested] 7+ messages in thread

* Re: yocto-bsp and kconf-check
  2013-08-29 15:51 ` Bruce Ashfield
@ 2013-08-29 18:56   ` Jate Sujjavanich
  2013-08-29 19:45     ` Bruce Ashfield
  0 siblings, 1 reply; 7+ messages in thread
From: Jate Sujjavanich @ 2013-08-29 18:56 UTC (permalink / raw)
  To: 'Bruce Ashfield'; +Cc: 'yocto@yoctoproject.org'

> Which kernel version ? 

This is linux-yocto-3.8 and the dylan-9.0.0 yocto.

> It's not about them being common or not, it's about BSPs following a
> base policy versus having wild, per-board behaviour.

I've cleaned up many items from the yocto-bsp generated {{machine}}.cfg that are already added by KTYPE. I have two remaining warnings for CONFIG_NET and CONFIG_NETDEVICES. Nested includes eventually lead to the line

force kconf non-hardware base.cfg

which adds those config options. I thought the non-hardware would prevent the specified_non_hdw.cfg warning.


Any ideas?

> -----Original Message-----
> From: Bruce Ashfield [mailto:bruce.ashfield@windriver.com]
> Sent: Thursday, August 29, 2013 11:52 AM
> To: Jate Sujjavanich
> Cc: 'yocto@yoctoproject.org'
> Subject: Re: [yocto] yocto-bsp and kconf-check
> 
> On 13-08-29 11:36 AM, Jate Sujjavanich wrote:
> > It appears that the yocto-bsp script generates a kernel configuration
> that creates some warnings during kern-tools' kconf_check. The
> {{machine}}.cfg file has many non-hardware options, therefore the script
> warns.
> >
> > It seems like many of these should be in the standard kernel
> configuration. Is this correct, and does the yocto-bsp data need to be
> updated?
> 
> Which kernel version ? But the answer is not necessarily, if a machine
> config is specifying something that hasn't been tagged "hardware" or
> that has a specified exception, you get a warning.
> 
> It's not about them being common or not, it's about BSPs following a
> base policy versus having wild, per-board behaviour.
> 
> Cheers,
> 
> Bruce
> 
> >
> > - Jate
> > _______________________________________________
> > yocto mailing list
> > yocto@yoctoproject.org
> > https://lists.yoctoproject.org/listinfo/yocto
> >



^ permalink raw reply	[flat|nested] 7+ messages in thread

* Re: yocto-bsp and kconf-check
  2013-08-29 18:56   ` Jate Sujjavanich
@ 2013-08-29 19:45     ` Bruce Ashfield
  2013-08-29 21:39       ` Jate Sujjavanich
  0 siblings, 1 reply; 7+ messages in thread
From: Bruce Ashfield @ 2013-08-29 19:45 UTC (permalink / raw)
  To: Jate Sujjavanich; +Cc: 'yocto@yoctoproject.org'

On 13-08-29 02:56 PM, Jate Sujjavanich wrote:
>> Which kernel version ?
>
> This is linux-yocto-3.8 and the dylan-9.0.0 yocto.
>
>> It's not about them being common or not, it's about BSPs following a
>> base policy versus having wild, per-board behaviour.
>
> I've cleaned up many items from the yocto-bsp generated {{machine}}.cfg that are already added by KTYPE. I have two remaining warnings for CONFIG_NET and CONFIG_NETDEVICES. Nested includes eventually lead to the line
>
> force kconf non-hardware base.cfg
>
> which adds those config options. I thought the non-hardware would prevent the specified_non_hdw.cfg warning.

It should, unless there's a bug. Can you send me the steps to
reproduce the config ? i.e. just your generated BSP layer in a
.tgz should be enough.

Bruce

>
>
> Any ideas?
>
>> -----Original Message-----
>> From: Bruce Ashfield [mailto:bruce.ashfield@windriver.com]
>> Sent: Thursday, August 29, 2013 11:52 AM
>> To: Jate Sujjavanich
>> Cc: 'yocto@yoctoproject.org'
>> Subject: Re: [yocto] yocto-bsp and kconf-check
>>
>> On 13-08-29 11:36 AM, Jate Sujjavanich wrote:
>>> It appears that the yocto-bsp script generates a kernel configuration
>> that creates some warnings during kern-tools' kconf_check. The
>> {{machine}}.cfg file has many non-hardware options, therefore the script
>> warns.
>>>
>>> It seems like many of these should be in the standard kernel
>> configuration. Is this correct, and does the yocto-bsp data need to be
>> updated?
>>
>> Which kernel version ? But the answer is not necessarily, if a machine
>> config is specifying something that hasn't been tagged "hardware" or
>> that has a specified exception, you get a warning.
>>
>> It's not about them being common or not, it's about BSPs following a
>> base policy versus having wild, per-board behaviour.
>>
>> Cheers,
>>
>> Bruce
>>
>>>
>>> - Jate
>>> _______________________________________________
>>> yocto mailing list
>>> yocto@yoctoproject.org
>>> https://lists.yoctoproject.org/listinfo/yocto
>>>
>



^ permalink raw reply	[flat|nested] 7+ messages in thread

* Re: yocto-bsp and kconf-check
  2013-08-29 15:36 yocto-bsp and kconf-check Jate Sujjavanich
  2013-08-29 15:51 ` Bruce Ashfield
@ 2013-08-29 19:46 ` Tom Zanussi
  1 sibling, 0 replies; 7+ messages in thread
From: Tom Zanussi @ 2013-08-29 19:46 UTC (permalink / raw)
  To: Jate Sujjavanich; +Cc: 'yocto@yoctoproject.org'

On Thu, 2013-08-29 at 11:36 -0400, Jate Sujjavanich wrote:
> It appears that the yocto-bsp script generates a kernel configuration that creates some warnings during kern-tools' kconf_check. The {{machine}}.cfg file has many non-hardware options, therefore the script warns.
> 
> It seems like many of these should be in the standard kernel configuration. Is this correct, and does the yocto-bsp data need to be updated?
> 

The templates in yocto-bsp are derived from the reference and qemu BSPs,
and have basically inherited the kernel config from those.

Now that the new kernel has landed, I've started updating the templates
for 3.10, and yes, if all the reference and qemu BSPs have been fixed in
this regard, I need to update the yocto-bsp templates with the new
options as well.

BTW, which arch did you use to generate your BSP?

Tom

> - Jate
> _______________________________________________
> yocto mailing list
> yocto@yoctoproject.org
> https://lists.yoctoproject.org/listinfo/yocto




^ permalink raw reply	[flat|nested] 7+ messages in thread

* Re: yocto-bsp and kconf-check
  2013-08-29 19:45     ` Bruce Ashfield
@ 2013-08-29 21:39       ` Jate Sujjavanich
  2013-08-30  4:47         ` Bruce Ashfield
  0 siblings, 1 reply; 7+ messages in thread
From: Jate Sujjavanich @ 2013-08-29 21:39 UTC (permalink / raw)
  To: 'Bruce Ashfield'; +Cc: 'yocto@yoctoproject.org'

Some additional information: I noticed that the two CONFIG's are also defined in the fragment features/usb-net/usb-net.cfg. They are defined without using the non-hardware flag.


> -----Original Message-----
> From: Bruce Ashfield [mailto:bruce.ashfield@windriver.com]
> Sent: Thursday, August 29, 2013 3:45 PM
> To: Jate Sujjavanich
> Cc: 'yocto@yoctoproject.org'
> Subject: Re: [yocto] yocto-bsp and kconf-check
> 
> On 13-08-29 02:56 PM, Jate Sujjavanich wrote:
> >> Which kernel version ?
> >
> > This is linux-yocto-3.8 and the dylan-9.0.0 yocto.
> >
> >> It's not about them being common or not, it's about BSPs following a
> >> base policy versus having wild, per-board behaviour.
> >
> > I've cleaned up many items from the yocto-bsp generated
> > {{machine}}.cfg that are already added by KTYPE. I have two remaining
> > warnings for CONFIG_NET and CONFIG_NETDEVICES. Nested includes
> > eventually lead to the line
> >
> > force kconf non-hardware base.cfg
> >
> > which adds those config options. I thought the non-hardware would
> prevent the specified_non_hdw.cfg warning.
> 
> It should, unless there's a bug. Can you send me the steps to reproduce
> the config ? i.e. just your generated BSP layer in a .tgz should be
> enough.
> 
> Bruce
> 
> >
> >
> > Any ideas?
> >
> >> -----Original Message-----
> >> From: Bruce Ashfield [mailto:bruce.ashfield@windriver.com]
> >> Sent: Thursday, August 29, 2013 11:52 AM
> >> To: Jate Sujjavanich
> >> Cc: 'yocto@yoctoproject.org'
> >> Subject: Re: [yocto] yocto-bsp and kconf-check
> >>
> >> On 13-08-29 11:36 AM, Jate Sujjavanich wrote:
> >>> It appears that the yocto-bsp script generates a kernel
> >>> configuration
> >> that creates some warnings during kern-tools' kconf_check. The
> >> {{machine}}.cfg file has many non-hardware options, therefore the
> >> script warns.
> >>>
> >>> It seems like many of these should be in the standard kernel
> >> configuration. Is this correct, and does the yocto-bsp data need to
> >> be updated?
> >>
> >> Which kernel version ? But the answer is not necessarily, if a
> >> machine config is specifying something that hasn't been tagged
> >> "hardware" or that has a specified exception, you get a warning.
> >>
> >> It's not about them being common or not, it's about BSPs following a
> >> base policy versus having wild, per-board behaviour.
> >>
> >> Cheers,
> >>
> >> Bruce
> >>
> >>>
> >>> - Jate
> >>> _______________________________________________
> >>> yocto mailing list
> >>> yocto@yoctoproject.org
> >>> https://lists.yoctoproject.org/listinfo/yocto
> >>>
> >



^ permalink raw reply	[flat|nested] 7+ messages in thread

* Re: yocto-bsp and kconf-check
  2013-08-29 21:39       ` Jate Sujjavanich
@ 2013-08-30  4:47         ` Bruce Ashfield
  0 siblings, 0 replies; 7+ messages in thread
From: Bruce Ashfield @ 2013-08-30  4:47 UTC (permalink / raw)
  To: Jate Sujjavanich; +Cc: 'yocto@yoctoproject.org'

On 13-08-29 5:39 PM, Jate Sujjavanich wrote:
> Some additional information: I noticed that the two CONFIG's are also defined in the fragment features/usb-net/usb-net.cfg. They are defined without using the non-hardware flag.
>

That's could be a contributing factor, if they have their bucket changed
by multiple fragments, the auditing gets harder.

I'll do a run with your BSP and let you know what I find.

Bruce

>
>> -----Original Message-----
>> From: Bruce Ashfield [mailto:bruce.ashfield@windriver.com]
>> Sent: Thursday, August 29, 2013 3:45 PM
>> To: Jate Sujjavanich
>> Cc: 'yocto@yoctoproject.org'
>> Subject: Re: [yocto] yocto-bsp and kconf-check
>>
>> On 13-08-29 02:56 PM, Jate Sujjavanich wrote:
>>>> Which kernel version ?
>>>
>>> This is linux-yocto-3.8 and the dylan-9.0.0 yocto.
>>>
>>>> It's not about them being common or not, it's about BSPs following a
>>>> base policy versus having wild, per-board behaviour.
>>>
>>> I've cleaned up many items from the yocto-bsp generated
>>> {{machine}}.cfg that are already added by KTYPE. I have two remaining
>>> warnings for CONFIG_NET and CONFIG_NETDEVICES. Nested includes
>>> eventually lead to the line
>>>
>>> force kconf non-hardware base.cfg
>>>
>>> which adds those config options. I thought the non-hardware would
>> prevent the specified_non_hdw.cfg warning.
>>
>> It should, unless there's a bug. Can you send me the steps to reproduce
>> the config ? i.e. just your generated BSP layer in a .tgz should be
>> enough.
>>
>> Bruce
>>
>>>
>>>
>>> Any ideas?
>>>
>>>> -----Original Message-----
>>>> From: Bruce Ashfield [mailto:bruce.ashfield@windriver.com]
>>>> Sent: Thursday, August 29, 2013 11:52 AM
>>>> To: Jate Sujjavanich
>>>> Cc: 'yocto@yoctoproject.org'
>>>> Subject: Re: [yocto] yocto-bsp and kconf-check
>>>>
>>>> On 13-08-29 11:36 AM, Jate Sujjavanich wrote:
>>>>> It appears that the yocto-bsp script generates a kernel
>>>>> configuration
>>>> that creates some warnings during kern-tools' kconf_check. The
>>>> {{machine}}.cfg file has many non-hardware options, therefore the
>>>> script warns.
>>>>>
>>>>> It seems like many of these should be in the standard kernel
>>>> configuration. Is this correct, and does the yocto-bsp data need to
>>>> be updated?
>>>>
>>>> Which kernel version ? But the answer is not necessarily, if a
>>>> machine config is specifying something that hasn't been tagged
>>>> "hardware" or that has a specified exception, you get a warning.
>>>>
>>>> It's not about them being common or not, it's about BSPs following a
>>>> base policy versus having wild, per-board behaviour.
>>>>
>>>> Cheers,
>>>>
>>>> Bruce
>>>>
>>>>>
>>>>> - Jate
>>>>> _______________________________________________
>>>>> yocto mailing list
>>>>> yocto@yoctoproject.org
>>>>> https://lists.yoctoproject.org/listinfo/yocto
>>>>>
>>>
>



^ permalink raw reply	[flat|nested] 7+ messages in thread

end of thread, other threads:[~2013-08-30  4:47 UTC | newest]

Thread overview: 7+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2013-08-29 15:36 yocto-bsp and kconf-check Jate Sujjavanich
2013-08-29 15:51 ` Bruce Ashfield
2013-08-29 18:56   ` Jate Sujjavanich
2013-08-29 19:45     ` Bruce Ashfield
2013-08-29 21:39       ` Jate Sujjavanich
2013-08-30  4:47         ` Bruce Ashfield
2013-08-29 19:46 ` Tom Zanussi

This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.