nl802154 enum misspelled, just send patch?

nl802154 enum misspelled, just send patch?

[Christoffer Holmstedt]

Hi
While working on the CCA improvements to wpan-tools I noticed that one of the
enums has a value misspelled. As the header is not in UAPI yet I suggest we
fix this.

NL802154_CCA_UWB_MULTIPEXED is missing an l (ell) in multiplexed.

The header file is included in wpan-tools repository and available in the
kernel. Should I just send a patch for bluetooth-next and then the new header
file is copied over to wpan-tools repository or what is the procedure here?

I assume normal procedure for more "stable" user-space software is that the
kernel header file(s) should not be included in the user-space tool but instead
included and linked against the kernel headers (from running system or similiar
source). At the moment the nl802154 header is included in wpan-tools to make it
easier to build while developing.

So, should I send a patch for both bluetooth-next and wpan-tools or just
bluetooth-next?

Regards
-- 
Christoffer Holmstedt

Re: nl802154 enum misspelled, just send patch?

[Varka Bhadram]

On 06/10/2015 01:54 PM, Christoffer Holmstedt wrote:

> Hi
> While working on the CCA improvements to wpan-tools I noticed that one of the
> enums has a value misspelled. As the header is not in UAPI yet I suggest we
> fix this.
>
> NL802154_CCA_UWB_MULTIPEXED is missing an l (ell) in multiplexed.
>
> The header file is included in wpan-tools repository and available in the
> kernel. Should I just send a patch for bluetooth-next and then the new header
> file is copied over to wpan-tools repository or what is the procedure here?
>
> I assume normal procedure for more "stable" user-space software is that the
> kernel header file(s) should not be included in the user-space tool but instead
> included and linked against the kernel headers (from running system or similiar
> source). At the moment the nl802154 header is included in wpan-tools to make it
> easier to build while developing.
>
> So, should I send a patch for both bluetooth-next and wpan-tools or just
> bluetooth-next?

Send patch for both trees .  :-)

-- 
Varka Bhadram

Re: nl802154 enum misspelled, just send patch?

[Christoffer Holmstedt]

On Wed, Jun 10, 2015 at 02:00:24PM +0530, Varka Bhadram wrote:
> On 06/10/2015 01:54 PM, Christoffer Holmstedt wrote:
> 
> >Hi
> >While working on the CCA improvements to wpan-tools I noticed that one of the
> >enums has a value misspelled. As the header is not in UAPI yet I suggest we
> >fix this.
> >
> >NL802154_CCA_UWB_MULTIPEXED is missing an l (ell) in multiplexed.
> >
> >The header file is included in wpan-tools repository and available in the
> >kernel. Should I just send a patch for bluetooth-next and then the new header
> >file is copied over to wpan-tools repository or what is the procedure here?
> >
> >I assume normal procedure for more "stable" user-space software is that the
> >kernel header file(s) should not be included in the user-space tool but instead
> >included and linked against the kernel headers (from running system or similiar
> >source). At the moment the nl802154 header is included in wpan-tools to make it
> >easier to build while developing.
> >
> >So, should I send a patch for both bluetooth-next and wpan-tools or just
> >bluetooth-next?
> 
> Send patch for both trees .  :-)
> 
> -- 
> Varka Bhadram

Will do.

-- 
Christoffer Holmstedt

Re: nl802154 enum misspelled, just send patch?

[Alexander Aring]

On Wed, Jun 10, 2015 at 10:24:49AM +0200, Christoffer Holmstedt wrote:
> Hi
> While working on the CCA improvements to wpan-tools I noticed that one of the
> enums has a value misspelled. As the header is not in UAPI yet I suggest we
> fix this.
> 
> NL802154_CCA_UWB_MULTIPEXED is missing an l (ell) in multiplexed.
> 
> The header file is included in wpan-tools repository and available in the
> kernel. Should I just send a patch for bluetooth-next and then the new header
> file is copied over to wpan-tools repository or what is the procedure here?
> 

First send patches for kernel, when Marcel has applied the change. Then
you can send another patch for wpan-tools for updating the nl802154 header.

> I assume normal procedure for more "stable" user-space software is that the
> kernel header file(s) should not be included in the user-space tool but instead
> included and linked against the kernel headers (from running system or similiar
> source). At the moment the nl802154 header is included in wpan-tools to make it
> easier to build while developing.
> 

In this case, if we would change the uapi header then compiling will
fail, because the enum is different. In this case it's safer to put the
header in wpan-tools.

If the kernel get a enum value which is higher than the own nl802154
capabilities, this should be handled by nl802154 by checking on ATTR_MAX
value.

> So, should I send a patch for both bluetooth-next and wpan-tools or just
> bluetooth-next?
> 

API changes are always bad, but this doesn't change the API. It's just a
compile failure when we change it. So it's okay to fix that now. 

So first send patches to bluetooth-next then wpan-tools with updating
nl802154 header and also updating your cca print functionality. Maybe
mention the bluetooth-next commit id in the commit msg of wpan-tools patch.

- Alex