[S390] net: provide architecture specific NET_SKB_PAD

NET_SKB_PAD has been increased from 32 to 64 and later to
max(32, L1_CACHE_BYTES). This led to a 25% throughput decrease for
streaming workloads accompanied by a 37% CPU cost increase on s390.
Define a architecture specific NET_SKB_PAD with the old value of 32.

Signed-off-by: Horst Hartmann <horsth@linux.vnet.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>