On Wed, May 24, 2017 at 10:46 PM, Khem Raj <raj.khem@gmail.com> wrote:
On Wed, May 24, 2017 at 7:40 PM, Bruce Ashfield
<bruce.ashfield@gmail.com> wrote:
>
>
> On Tue, May 23, 2017 at 3:36 PM, Bruce Ashfield <bruce.ashfield@gmail.com>
> wrote:
>>
>>
>>
>> On Mon, May 22, 2017 at 1:06 PM, Khem Raj <raj.khem@gmail.com> wrote:
>>>
>>> >> +SRC_URI += "file://drm.cfg"
>>> >
>>> >
>>> >
>>> > If anything gets added to the default configurations, they need to go
>>> > into
>>> > the kernel-cache
>>> > and then be included by the relevant kernel types there.
>>> >
>>> > If you want, I can take care of switching things to that format.
>>> >
>>>
>>> yes please.
>>
>>
>> Will do. I'll make it available to all active kernel versions + future
>> ones.
>
>
> As it turns out, we already have a similar fragment for vmware guests:
>
> cfg/vmware-guest.cfg
>
> --------------------
>
> CONFIG_VMWARE_BALLOON=m
>
> CONFIG_VMWARE_PVSCSI=y
>
> CONFIG_VMWARE_VMCI=m
>
> CONFIG_VSOCKETS=m
>
> CONFIG_VMWARE_VMCI_VSOCKETS=m
>
> CONFIG_VMXNET3=y
>
> CONFIG_DRM_VMWGFX=m
>
> CONFIG_DRM_VMWGFX_FBCON=y
>
> CONFIG_FUSION=y
>
> CONFIG_FUSION_SPI=y
>
> CONFIG_FUSION_FC=y
>
> CONFIG_FUSION_SAS=y
>
> CONFIG_FUSION_MAX_SGE=128
>
> CONFIG_FUSION_CTL=y
>
> CONFIG_FUSION_LAN=y
>
> CONFIG_FUSION_LOGGING=y
>
> CONFIG_SCSI_MPT2SAS=y
>
> CONFIG_SCSI_MPT2SAS_MAX_SGE=128
>
> CONFIG_SCSI_MPT2SAS_LOGGING=y
>
> CONFIG_SCSI_MPT3SAS=y
>
> CONFIG_SCSI_MPT3SAS_MAX_SGE=128
>
> CONFIG_SCSI_MPT3SAS_LOGGING=y
>
> -------------
>
> To stay consistent with that naming, I'll call this virtual-box-guest.cfg,
> and it can be enabled
> via KERNEL_FEATURES when needed.
>
> I'm hesitant to globally enable it, since we don't need those modules on all
> linux-yocto
> kernels, just ones that are going to run on virtual box.
>
> I could enable them for qemux86*, since those already have virtio configs
> enabled by
> default and these have a similar goal.
>
> Thoughts ? I'd like to make sure that sounds reasonable, before spending the
> time to
> make the changes and test.
>

I would suggest to enable them by default since these are modules they
dont impact the size of kernel itself. Then we can use recommends to
choose them in machine configs.

We've never done that before, there's hundreds of modules we could turn on 
just in case, but instead only enable them for machines that can actually use
them.

Many users use the meta package for modules and expect the build to be as
minimal as possible.

So I'm balancing both those use cases.

I'll complete the fragment and we can enable it in a second commit to keep the
options open.

Bruce
 

> Bruce
>
>>
>> Cheers,
>>
>> Bruce
>>
>>
>>
>> --
>> "Thou shalt not follow the NULL pointer, for chaos and madness await thee
>> at its end"
>
>
>
>
> --
> "Thou shalt not follow the NULL pointer, for chaos and madness await thee at
> its end"



--
"Thou shalt not follow the NULL pointer, for chaos and madness await thee at its end"