The incrementing and decrementing of 'ns_statscounter_recursclients'
were not properly balanced: for example, it would be incremented for
a prefetch query but not decremented if the query failed.
This commit ensures that the recursion quota and the recursive clients
counter are always in sync with each other.
(cherry picked from commit 82991451b4)
The duration_fromtext() function is truncating large numbers
to 32 bits instead of capping or rejecting them, i.e. 64424509445,
which is 0xf00000005, gets parsed as 32-bit value 5 (0x00000005).
Fail parsing a duration if any of its components is bigger than
32 bits. Using those kind of big numbers has no practical use case
for a duration.
The cfg_obj_asduration() function can overflow the 32 bit
seconds variable when calculating the duration from its component
parts.
To avoid that, use 64-bit calculation and return UINT32_MAX if the
calculated value is bigger than UINT32_MAX. Again, a number this big
has no practical use case anyway.
The buffer for the generated duration string is limited to 64 bytes,
which, in theory, is smaller than the longest possible generated
duration string.
Use 80 bytes instead, calculated by the '7 x (10 + 1) + 3' formula,
where '7' is the count of the duration's parts (year, month, etc.), '10'
is their maximum length when printed as a decimal number, '1' is their
indicator character (Y, M, etc.), and 3 is two more indicators (P and T)
and the terminating NUL character.
(cherry picked from commit fddaebb285)
The cfg_print_duration() checks added previously in the 'duration_test'
unit test uncovered a bug in cfg_print_duration().
When calculating the current 'str' pointer of the generated text in the
buffer 'buf', it erroneously adds 1 byte to compensate for that part's
indicator character. For example, to add 12 minutes, it needs to add
2 + 1 = 3 characters, where 2 is the length of "12", and 1 is the length
of "M" (for minute). The mistake was that the length of the indicator
is already included in 'durationlen[i]', so there is no need to
calculate it again.
In the result of this mistake the current pointer can advance further
than needed and end up after the zero-byte instead of right on it, which
essentially cuts off any further generated text. For example, for a
5 minutes and 30 seconds duration, instead of having this:
'P', 'T', '5', 'M', '3', '0', 'S', '\0'
The function generates this:
'P', 'T', '5', 'M', '\0', '3', '0', 'S', '\0'
Fix the bug by adding to 'str' just 'durationlen[i]' instead of
'durationlen[i] + 1'.
(cherry picked from commit dc55f1ebb9)
Currently the 'duration_test' unit test checks only the
cfg_obj_asduration() function.
Extend the test so it checks also the reverse operation using the
cfg_print_duration() function, which is used in named-checkconf.
(cherry picked from commit 39290bb7cd)
The cfg_print_duration() function prints a ISO 8601 duration value
converted from an array of integers, where the parts of the date and
time are stored.
durationlen[6], which holds the "seconds" part of the duration, has
a special case in cfg_print_duration() to ensure that when there are
no values in the duration, the result still can be printed as "PT0S",
instead of just "P", so it can be a valid ISO 8601 duration value.
There is a logical error in one of the two special case code paths,
when it checks that no value from the "date" part is defined, and no
"hour" or "minute" from the "time" part are defined.
Because of the error, durationlen[6] can be used uninitialized, in
which case the second parameter passed to snprintf() (which is the
maximum allowed length) can contain a garbage value.
This can not be exploited because the buffer is still big enough to
hold the maximum possible amount of characters generated by the "%u%c"
format string.
Fix the logical bug, and initialize the 'durationlen' array to zeros
to be a little safer from other similar errors.
(cherry picked from commit 9440910187)
clear the compression buffer before use. this eliminates the
possibility of a latent bug that, when combined with other changes,
allowed an overread in a later version of BIND.
dohpath is specfied in draft-ietf-add-svcb-dns and has a value
of 7. It must be a relative path (start with a /), be encoded
as UTF8 and contain the variable dns ({?dns}).
(cherry picked from commit 6d561d3886)
When looking for changes in a catalog zone member zone we need to
also check if the TSIG key name associated with a primary server
has be added, removed or changed.
(cherry picked from commit 9172bd9b5a)
if ISC_BUFFER_USEINLINE is defined, then macros are used to implement
isc_buffer primitives (isc_buffer_init(), isc_buffer_region(), etc).
otherwise, functions are used. previously, only the functions had
DbC assertions, which made it possible for coding errors to go
undetected. this commit makes the macro versions enforce the same
requirements.
dns_rdata_tostruct doesn't need a mctx passed to it for SIG (the signer
is already expanded at this point). About the only time when mctx is
needed is when the structure is to be used after the rdata has been
destroyed.
(cherry picked from commit d6ad56bd9e)
Impact should be visible only in tests or tools because named never
uses view == NULL, which is a necessary condition to trigger this leak.
(cherry picked from commit 69256b3553)
Don't attempt to resolve DNS responses for intermediate results. This
may create multiple refreshes and can cause a crash.
One scenario is where for the query there is a CNAME and canonical
answer in cache that are both stale. This will trigger a refresh of
the RRsets because we encountered stale data and we prioritized it over
the lookup. It will trigger a refresh of both RRsets. When we start
recursing, it will detect a recursion loop because the recursion
parameters will eventually be the same. In 'dns_resolver_destroyfetch'
the sanity check fails, one of the callers did not get its event back
before trying to destroy the fetch.
Move the call to 'query_refresh_rrset' to 'ns_query_done', so that it
is only called once per client request.
Another scenario is where for the query there is a stale CNAME in the
cache that points to a record that is also in cache but not stale. This
will trigger a refresh of the RRset (because we encountered stale data
and we prioritized it over the lookup).
We mark RRsets that we add to the message with
DNS_RDATASETATTR_STALE_ADDED to prevent adding a duplicate RRset when
a stale lookup and a normal lookup conflict with each other. However,
the other non-stale RRset when following a CNAME chain will be added to
the message without setting that attribute, because it is not stale.
This is a variant of the bug in #2594. The fix covered the same crash
but for stale-answer-client-timeout > 0.
Fix this by clearing all RRsets from the message before refreshing.
This requires the refresh to happen after the query is send back to
the client.
(cherry picked from commit d939d2ecde)
Limit the amount of database lookups that can be triggered in
fctx_getaddresses() (i.e. when determining the name server addresses to
query next) by setting a hard limit on the number of NS RRs processed
for any delegation encountered. Without any limit in place, named can
be forced to perform large amounts of database lookups per each query
received, which severely impacts resolver performance.
The limit used (20) is an arbitrary value that is considered to be big
enough for any sane DNS delegation.
(cherry picked from commit 3a44097fd6)
It is possible to bypass Response Rate Limiting (RRL)
`responses-per-second` limitation using specially crafted wildcard
names, because the current implementation, when encountering a found
DNS name generated from a wildcard record, just strips the leftmost
label of the name before making a key for the bucket.
While that technique helps with limiting random requests like
<random>.example.com (because all those requests will be accounted
as belonging to a bucket constructed from "example.com" name), it does
not help with random names like subdomain.<random>.example.com.
The best solution would have been to strip not just the leftmost
label, but as many labels as necessary until reaching the suffix part
of the wildcard record from which the found name is generated, however,
we do not have that information readily available in the context of RRL
processing code.
Fix the issue by interpreting all valid wildcard domain names as
the zone's origin name concatenated to the "*" name, so they all will
be put into the same bucket.
(cherry picked from commit baa9698c9d)
previously, if ns_clientmgr_create() failed, the interface was not
cleaned up correctly and an assertion or segmentation fault could
follow. this has been fixed.
Having implicit inline-signing set for dnssec-policy when there is no
update policy is confusing, so lets make this explicit.
(cherry picked from commit 5ca02fe6e7e591d1fb85936ea4dda720c3d741ef)
This should make sure that the memory context is not destroyed
before the memory pool, which is using the context.
(cherry picked from commit e97c3eea95)
The dnstap query_message field was in some cases being filled in
with response messages, along with the response_message field.
The query_message field should only be used when logging requests,
and the response_message field only when logging responses.
There is one case in 'dns_nsec3_activex()' where it returns but forgets
to detach the db node. Add the missing 'dns_db_detachnode()' call.
This case only triggers if 'sig-signing-type' (privatetype) is set to 0
(which by default is not), or if the function is called with 'complete'
is set to 'true' (which at this moment do not exist).
(cherry picked from commit 0cf6c18ccb2205a1fc81431f908c8310f6136bbb)
When dumping an ADB address entry associated with a name,
the name bucket lock was held, but the entry bucket lock was
not; this could cause data races when other threads were updating
address entry info. (These races are probably not operationally
harmful, but they triggered TSAN error reports.)
(cherry picked from commit f841f545b7)
When checking if we should enable serve-stale, add an early out case
when the result is an error signalling a duplicate query or a query
that would be dropped.
(cherry picked from commit 059a4c2f4d)
When initially hitting the `fetches-per-zone` value, a log message
is being generated for the event of dropping the first fetch, then
any further log events occur only when another fetch is being dropped
and 60 seconds have been passed since the last logged message.
That logic isn't ideal because when the counter of the outstanding
fetches reaches zero, the structure holding the counters' values will
get deleted, and the information about the dropped fetches accumulated
during the last minute will not be logged.
Improve the fcount_logspill() function to makie sure that the final
values are getting logged before the counter object gets destroyed.
(cherry picked from commit 039871ceb7)
The command 'rndc dumpdb -expired' will include expired RRsets in the
output, but only for the RBTDB_VIRTUAL time (of 5 minutes). This means
that if there is a cache cleaning problem and contents are not cleaned
up, the rndc command has little diagnostic value. Fix this by including
all RRsets in the dumpdb output if the '-expired' flag is set.
(cherry picked from commit 930ba2c914)
Fedora 33 doesn't support RSASHA1 in future mode. There is no easy
check for this other than by attempting to perform a verification
using known good signatures. We don't attempt to sign with RSASHA1
as that would not work in FIPS mode. RSASHA1 is verify only.
The test vectors were generated using OpenSSL 3.0 and
util/gen-rsa-sha-vectors.c. Rerunning will generate a new set of
test vectors as the private key is not preserved.
e.g.
cc util/gen-rsa-sha-vectors.c -I /opt/local/include \
-L /opt/local/lib -lcrypto
(cherry picked from commit cd3f00874f)
The BUFSIZ value varies between platforms, it could be 8K on Linux and
512 bytes on mingw. Make sure the buffers are always big enough for the
output data to prevent truncation of the output by appropriately
enlarging or sizing the buffers.
(cherry picked from commit b19d932262)
Contrary to what the documentation states, memory filling is only
enabled by --enable-developer (or by setting -DISC_MEM_DEFAULTFILL=1) if
the internal memory allocator is used. However, the internal memory
allocator is disabled by default, so just using the --enable-developer
build-time option does not enable memory filling (passing "-M fill" on
the named command line is necessary to actually enable it). As memory
filling is a useful tool for troubleshooting certain types of bugs, it
should also be enabled by --enable-developer when the system allocator
is used.
Furthermore, memory-related preprocessor macros are handled in two
distinct locations: lib/isc/include/isc/mem.h and bin/named/main.c.
This makes the logic hard to follow.
Move all code handling the ISC_MEM_DEFAULTFILL preprocessor macro to
lib/isc/include/isc/mem.h, ensuring memory filling is enabled by the
--enable-developer build-time switch, no matter which memory allocator
is used.
Commit c96b6eb5ec changed the way mempool
code handles freed allocations that cannot be retained for later use as
"free list" items: it no longer uses different logic depending on
whether the internal allocator is used or the system one. However, that
commit did not update a relevant piece of code in isc_mempool_destroy(),
causing memory context statistics to always be off upon shutdown when
BIND 9 is built with -DISC_MEM_USE_INTERNAL_MALLOC=1. This causes
assertion failures. Update isc_mempool_destroy() accordingly in order
to prevent this issue from being triggered.
free_namelist could be passed names with associated rdatasets
when handling errors. These need to be disassociated before
calling dns_message_puttemprdataset.
(cherry picked from commit 745d5edc3a)
messages indicating the reason for a fallback to AXFR (i.e, because
the requested serial number is not present in the journal, or because
the size of the IXFR response would exceeed "max-ixfr-ratio") are now
logged at level info instead of debug(4).
(cherry picked from commit df1d81cf96)
fctx_decreference() may call fctx_destroy(), which in turn may free the
fetch context by calling isc_mem_putanddetach(). This means that
whenever fctx_decreference() is called, the fetch context pointer should
be assumed to point to garbage after that call. Meanwhile, the
following pattern is used in several places in lib/dns/resolver.c:
LOCK(&res->buckets[fctx->bucketnum].lock);
bucket_empty = fctx_decreference(fctx);
UNLOCK(&res->buckets[fctx->bucketnum].lock);
Given that 'fctx' may be freed by the fctx_decreference() call, there is
no guarantee that the value of fctx->bucketnum will be the same before
and after the fctx_decreference() call. This can cause all kinds of
locking issues as LOCK() calls no longer match up with their UNLOCK()
counterparts.
Fix by always using a helper variable to hold the bucket number when the
pattern above is used.
Note that fctx_try() still uses 'fctx' after calling fctx_decreference()
(it calls fctx_done()). This is safe to do because the reference count
for 'fctx' is increased a few lines earlier and it also cannot be zero
right before that increase happens, so the fctx_decreference() call in
that particular location never invokes fctx_destroy(). Nevertheless,
use a helper variable for that call site as well, to retain consistency
and to prevent copy-pasted code from causing similar problems in the
future.
The original sscanf processing allowed for a number of syntax errors
to be accepted. This included missing the closing brace in
${modifiers}
Look for both comma and right brace as intermediate seperators as
well as consuming the final right brace in the sscanf processing
for ${modifiers}. Check when we got right brace to determine if
the sscanf consumed more input than expected and if so behave as
if it had stopped at the first right brace.
(cherry picked from commit 7be64c0e94)
$GENERATE uses 'int' for its computations and some constructions
can overflow values that can be represented by an 'int' resulting
in undefined behaviour. Detect these conditions and return a
range error.
(cherry picked from commit 5327b9708f)
After refactoring of `validated()`, the `maybe_destroy()` function is
no longer expected to actually destroy the fetch context when it is
being called, so effectively it only ensures that the validators are
canceled when the context has no more queries and pending events, but
that is redundant, because `maybe_destroy()` `REQUIRE`s that the context
should be in the shutting down state, and the function which sets that
state is already canceling the validators in its own turn.
As a failsafe, to make sure that no validators will be created after
`fctx_doshutdown()` is called, add an early return from `valcreate()` if
the context is in the shutting down state.
The `resolver.c:validated()` function unlinks the current validator from
the fetch's validators list, which can leave it empty, then unlocks
the bucket lock. If, by a chance, the fetch was timed out just before
the `validated()` call, the final timeout callback running in parallel
with `validated()` can find the fetch context with no active fetches
and with an empty validators list and destroy it, which is unexpected
for the `validated()` function and can lead to a crash.
Increase the fetch context's reference count in the beginning of
`validated()` and decrease it when it finishes its work to avoid the
unexpected destruction of the fetch context.
