On 02/09/2016 06:00 PM, Konstantin Olchanski wrote:
With luck this will settle soon enough, but I did not see any explanation for
this binary incompatibility,
should I think that this is the new way of things - each new release junks all
the old driver kernel modules
and I must harrass the hardware vendors to issue new drivers?
As I recall, similar things have happened with certain EL5 and EL6
update releases, especially early on in what Red Hat calls the
'Production 1' phase (see:
https://access.redhat.com/support/policy/updates/errata ). After the
first four years this is much less likely to happen.
There appear to be clues in the 7.2 release notes section on new
features for the kernel, as well as a section on new hardware drivers.
And it looks like, according to the Virtualbox bug report previously
mentioned by S A, that a driver update did the breakage (VBox needs to
emulate certain ethernet cards and uses the Linux headers during build,
and it appears that is the source of this particular breakage). I see a
lot of updates to network drivers in the kernel changelog between the
'229 kernels and the fist '327 kernels (you can find them yourself; rpm
-q --changelog kernel will do that, but you might want to specify which
kernel when you run it, like I did using: rpm -q --changelog
kernel-3.10.0-327.4.5.el7 ). It looks like this particular issue may
have been initially triggered at the (unreleased) '309 kernel level.
This change would not affect *all* kernel modules, just ones that use
that particular include file.
It probably came down to 'this driver breaks a few minor (from RH's
point of view) things, but fixes bugs X, Y, and Z on certain new
hardware that we have to support.'
