[OE-core] efficacy of defining overriding linux-yocto...bb recipe file?

[Robert P. J. Day]

  recently began looking at someone's else layer and noticed that that
layer defined a new kernel recipe by (yeesh) creating its very own
"linux-yocto_3.14.bb" file. not a bbappend file -- its very own .bb
file, which i thought was kind of weird.

  just to make sure i'm not missing any subtleties, there are only two
*recommended* ways i know of to define/extend kernel recipes. first,
just a regular .bbappend file ... nuff said.

  the second is how, say, the meta-fsl people do it (which i like), by
defining totally new, fsl-specific recipes:

linux-fslc_3.18.bb
linux-imx_2.6.35.3.bb
linux-imx_3.10.53.bb
linux-imx-mfgtool_3.10.53.bb
linux-imx-rt_3.10.31.bb
linux-ls1_3.12.bb
linux-timesys_3.0.15.bb

and then having their machine definition files set something like:

PREFERRED_PROVIDER_virtual/kernel ?= "linux-timesys"

  i've just never seen a layer flat out create its own base-level
"linux-yocto" kernel recipe .bb file. am i safe in suggesting that
that's just not the way things are done? (i'm guessing the only way to
guarantee that *that* recipe file is used is to bump up the priority
value of the layer, yes?)

rday

-- 

========================================================================
Robert P. J. Day                                 Ottawa, Ontario, CANADA
                        http://crashcourse.ca

Twitter:                                       http://twitter.com/rpjday
LinkedIn:                               http://ca.linkedin.com/in/rpjday
========================================================================