On 03/12/2014 08:14 AM, terry white wrote:

    my reading of this 'thread' suggests that "pppd" and/or "serial" have
started "DEMANDING" bleeding edge kernel versions.  if that 'correct', for
me anyway, it's going to prove problematic.  that aside, it doesn't seem
like the most elegant solution, to problems it's meant to address ...

There have been _hundreds of changes_ to the tty core since 3.2, many
to specifically fix the type of error reported.

Very few of these changes are trivial; there are at least a dozen patchsets
with 10 or more patches. I doubt these will ever be backported to -stable;
the changes are just too extensive.

It makes sense to establish if these changes have indeed fixed the observed
error already or if more in-depth debugging is necessary to uncover the root
cause.

Regards,
Peter Hurley