The current logic is naive in the case that there is more than
one RRset in an answer (Typically, when a non-CNAME query is answered
by one or more CNAME RRs, and then then an answer RRset.)
If all the RRsets validate, then they are cached and marked as validated,
but if any RRset doesn't validate, then the AD flag is not set (good) and
ALL the RRsets are cached marked as not validated.
This breaks when, eg, the answer contains a validated CNAME, pointing
to a non-validated answer. A subsequent query for the CNAME without do
will get an answer with the AD flag wrongly reset, and worse, the same
query with do will get a cached answer without RRSIGS, rather than
being forwarded.
The code now records the validation of individual RRsets and that
is used to correctly set the "validated" bits in the cache entries.
Further fix to 0549c73b7e
Handles case when RR name is not a pointer to the question,
only occurs for some auth-mode replies, therefore not
detected by fuzzing (?)
Fix heap overflow in DNS code. This is a potentially serious
security hole. It allows an attacker who can make DNS
requests to dnsmasq, and who controls the contents of
a domain, which is thereby queried, to overflow
(by 2 bytes) a heap buffer and either crash, or
even take control of, dnsmasq.
Some consider it good practice to obscure software version numbers to
clients. Compiling with -DNO_ID removes the *.bind info structure.
This includes: version, author, copyright, cachesize, cache insertions,
evictions, misses & hits, auth & servers.
The list of exceptions to being able to locally answer
cached data for validated records when DNSSEC data is requested
was getting too long, so don't ever do that. This means
that the cache no longer has to hold RRSIGS and allows
us to lose lots of code. Note that cached validated
answers are still returned as long as do=0
Fix off-by-one in code which checks for over-long domain names
in received DNS packets. This enables buffer overflow attacks
which can certainly crash dnsmasq and may allow for arbitrary
code execution. The problem was introduced in commit b8f16556d,
release 2.73rc6, so has not escaped into any stable release.
Note that the off-by-one was in the label length determination,
so the buffer can be overflowed by as many bytes as there are
labels in the name - ie, many.
Thanks to Ron Bowes, who used lcmatuf's afl-fuzz tool to find
the problem.
check_for_local_domain() was broken due to new code matching F_*
bits in cache entries for DNSSEC. Because F_DNSKEY | F_DS is
used to match RRSIG entries, cache_find_by_name() insists on an exact match
of those bits. So adding F_DS to the bits that check_for_local_domain()
sends to cache_find_by_name() won't result in DS records as well
as the others, it results in only DS records. Add a new bit, F_NSIGMATCH
which suitably changes the behaviour of cache_find_by_name().
Simon Kelley