[c-nsp] ATM CRC errors

[Antonio Querubin]

On Wed, 27 Oct 2004, Andy Dills wrote:

> On Wed, 27 Oct 2004, F. David Sinn wrote:
>
> > The CRC's can be related to lost cell's, since if a component cell (or
> > cells) of a AAL5 encapsulated frame are lost, the resultant frame will
> > no longer pass CRC checks when the re-assembly is attempted.
>
> Mea culpa, I was very incorrect. I don't know why I was thinking the CRC
> check was on L2 instead of L3.

Actually CRC failures do occur at L2 on ATM cards and are counted in the
interface stats.  However, they'd affect all VCs connected through the
same physical interface instead of just specific VCs as it appears in this
particular case.