Instead of downloading a 4+GB archive and uncompressing it in /tmp,
let's uncompress the archive via NFS and feed the data straight into
cpio. This should be faster and avoids curl's error handling that's
happy to abort instead of re-trying.

Change-Id: Ifa5d842c50111ccdbd56e795b270ee6433a63adf
Reviewed-by: Frederik Gladhorn <frederik.gladhorn@qt.io>
Reviewed-by: Tony Sarajärvi <tony.sarajarvi@qt.io>