On Thu, May 22, 2025 at 07:18:14PM +0200, Pablo Neira Ayuso wrote: > On Wed, May 21, 2025 at 03:12:38PM +0200, Phil Sutter wrote: > > Faced with unexpected values or missing attributes, most of the netlink > > deserialization code would complain, drop the nftables object being > > constructed and carry on. Some error paths though instead call BUG() or > > assert(0) instead. This series eliminates at least the most prominent > > ones among those (patches 1 and 3). > > > > Patch 4 prevents object deserialization from aborting upon the first one > > with unexpected data. If netlink_delinearize_obj() returns NULL, an > > error message has been emitted already so just return 0 to the foreach > > loop so it continues with the next object. > > > > Patch 2 is just preparation work for patch 3. > > Reviewed-by: Pablo Neira Ayuso <pablo@xxxxxxxxxxxxx> Applied after dropping two explicit newline characters leftover from BUG() to netlink_error() conversion. Thanks, Phil
