Add a specific error to report that a netlink response had
NLM_F_DUMP_INTR set, indicating that the set of results may be
incomplete or inconsistent.
unix.EINTR was previously returned (with no results) when the
NLM_F_DUMP_INTR flag was set. Now, errors.Is(err, unix.EINTR) will
still work. But, this will be a breaking change for any code that's
checking for equality with unix.EINTR.
Return results with ErrDumpInterrupted. Results may be incomplete
or inconsistent, but give the caller the option of using them.
Look for NLM_F_DUMP_INTR in more places:
- linkSubscribeAt, neighSubscribeAt, routeSubscribeAt
- can do an initial dump, which may report inconsistent results
-> if there's an error callback, call it with ErrDumpInterrupted
- socketDiagXDPExecutor
- makes an NLM_F_DUMP request, without using Execute()
-> give it the same behaviour as functions that do use Execute()
Signed-off-by: Rob Murray <rob.murray@docker.com>
The package level var native (holding the native endianness) is
initialized at package load time. Thus there is no need to re-initalize
it in functions using it, e.g. (*Handle).filterModify, parseU32Data,
parseFwData, parseBpfData and parseMatchAllData.
This fixes a data race between these functions and any read access of
var native, e.g. in LinkDeserialize as reported in issue #633.
Also don't re-declare local variables shadowing the global package-level
var.
Fixes#633
Signed-off-by: Tobias Klauser <tobias@cilium.io>
iproute2's own netlink library asserts that the sockaddr sender pid
has to be the one of the kernel [0]. It also doesn't bail out on pid
mismatch but only skips the message instead. We've seen cases where
the latter had a pid 0; in such case we should skip to the next nl
message instead of hard bail out.
[0] https://git.kernel.org/pub/scm/network/iproute2/iproute2.git/tree/lib/libnetlink.c
rtnl_dump_filter_l(), __rtnl_talk_iov()
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>