it and the OS has enabled XGETBV for application use.
It might need to also check XCR0[2] (having executed XGETBV) to check that
the kernel actually supports saving the YMM registers, but I suspect the
kernel might defer setting that until the first fault.
See vol 1 section 13.5 of the Intel SDM (intel_x86_325462.pdf).
Fixes toolchain/45673
- The postscreen_access_list feature failed to ignore case in the first
character of a command (e.g., permit, reject, etc.). Reported by Francis
Picabia. (This fix is incorrectly listed in the HISTORY files of earlier
releases, and will be removed with a future patch.)
- Strip the datalink suffix (e.g., %eth0) from IPv6 addresses returned by
the system getaddrinfo() routine. Such suffixes break the default
mynetworks value, the Postfix SMTP server's reverse/forward DNS
name/address mapping check, and possibly more.
- To eliminate the possibility of collisions with connection cache lookup
keys, the Postfix LDAP client now computes those lookup keys by joining
the number-valued connection properties with ASCII null, just like it
already did with the string-valued connection properties.
- There was a memory leak during one-time TLS library initialization
(introduced with Postfix 2.5). Reported by Coverity.
- There was a memory leak in the unused oqmgr(8) program (introduced with
Postfix 2.3). Reported by Coverity.
* man page improvements
* dnsmasq + dbus + IPv6 linklocal support (requires dnsmasq-2.64)
* sort interface protocols as well as interface (bge0, bge0:ra, bge0:dhcp6)
* add interface scope to exported IPv6 link local addresses
* DUID feature works again
* If we receive >1 MTU in the RA we use the last one
* Expire RA options separately from the RA
* Merge multiple RA options together except when a single is needed
* Some memory leaks plugged
* Don't overwrite invalid memory when DNSSL options begin with a NULL
* Ensure we have a large enough buffer for an escaped DNSSL option from RA
--- 9.9.2-P1 released ---
3407. [security] Named could die on specific queries with dns64 enabled.
[Addressed in change #3388 for BIND 9.8.5 and 9.9.3.]
--- 9.9.2 released ---
3383. [security] A certain combination of records in the RBT could
cause named to hang while populating the additional
section of a response. [RT #31090]
3373. [bug] win32: open raw files in binary mode. [RT #30944]
3364. [security] Named could die on specially crafted record.
[RT #30416]
--- 9.9.2rc1 released ---
3370. [bug] Address use after free while shutting down. [RT #30241]
3369. [bug] nsupdate terminated unexpectedly in interactive mode
if built with readline support. [RT #29550]
3368. [bug] <dns/iptable.h>, <dns/private.h> and <dns/zone.h>
were not C++ safe.
3367. [bug] dns_dnsseckey_create() result was not being checked.
[RT #30685]
3366. [bug] Fixed Read-After-Write dependency violation for IA64
atomic operations. [RT #25181]
3365. [bug] Removed spurious newlines from log messages in
zone.c [RT #30675]
3363. [bug] Need to allow "forward" and "fowarders" options
in static-stub zones; this had been overlooked.
[RT #30482]
3362. [bug] Setting some option values to 0 in named.conf
could trigger an assertion failure on startup.
[RT #27730]
3361. [bug] "rndc signing -nsec3param" didn't work correctly
when salt was set to '-' (no salt). [RT #30099]
3360. [bug] 'host -w' could die. [RT #18723]
3359. [bug] An improperly-formed TSIG secret could cause a
memory leak. [RT #30607]
3357. [port] Add support for libxml2-2.8.x [RT #30440]
3356. [bug] Cap the TTL of signed RRsets when RRSIGs are
approaching their expiry, so they don't remain
in caches after expiry. [RT #26429]
3355. [port] Use more portable awk in verify system test.
3354. [func] Improve OpenSSL error logging. [RT #29932]
--- 9.9.2b1 released ---
3353. [bug] Use a single task for task exclusive operations.
[RT #29872]
3352. [bug] Ensure that learned server attributes timeout of the
adb cache. [RT #29856]
3351. [bug] isc_mem_put and isc_mem_putanddetach didn't report
caller if either ISC_MEM_DEBUGSIZE or ISC_MEM_DEBUGCTX
memory debugging flags are set. [RT #30243]
3350. [bug] Memory read overrun in isc___mem_reallocate if
ISC_MEM_DEBUGCTX memory debugging flag is set.
[RT #30240]
3349. [bug] Change #3345 was incomplete. [RT #30233]
3348. [bug] Prevent RRSIG data from being cached if a negative
record matching the covering type exists at a higher
trust level. Such data already can't be retrieved from
the cache since change 3218 -- this prevents it
being inserted into the cache as well. [RT #26809]
3347. [bug] dnssec-settime: Issue a warning when writing a new
private key file would cause a change in the
permissions of the existing file. [RT #27724]
3346. [security] Bad-cache data could be used before it was
initialized, causing an assert. [RT #30025]
3345. [bug] Addressed race condition when removing the last item
or inserting the first item in an ISC_QUEUE.
[RT #29539]
3344. [func] New "dnssec-checkds" command checks a zone to
determine which DS records should be published
in the parent zone, or which DLV records should be
published in a DLV zone, and queries the DNS to
ensure that it exists. (Note: This tool depends
on python; it will not be built or installed on
systems that do not have a python interpreter.)
[RT #28099]
3342. [bug] Change #3314 broke saving of stub zones to disk
resulting in excessive cpu usage in some cases.
[RT #29952]
3341. [func] New "dnssec-verify" command checks a signed zone
to ensure correctness of signatures and of NSEC/NSEC3
chains. [RT #23673]
3339. [func] Allow the maximum supported rsa exponent size to be
specified: "max-rsa-exponent-size <value>;" [RT #29228]
3338. [bug] Address race condition in units tests: asyncload_zone
and asyncload_zt. [RT #26100]
3337. [bug] Change #3294 broke support for the multiple keys
in controls. [RT #29694]
3335. [func] nslookup: return a nonzero exit code when unable
to get an answer. [RT #29492]
3334. [bug] Hold a zone table reference while performing a
asyncronous load of a zone. [RT #28326]
3333. [bug] Setting resolver-query-timeout too low can cause
named to not recover if it loses connectivity.
[RT #29623]
3332. [bug] Re-use cached DS rrsets if possible. [RT #29446]
3331. [security] dns_rdataslab_fromrdataset could produce bad
rdataslabs. [RT #29644]
3330. [func] Fix missing signatures on NOERROR results despite
RPZ rewriting. Also
- add optional "recursive-only yes|no" to the
response-policy statement
- add optional "max-policy-ttl" to the response-policy
statement to limit the false data that
"recursive-only no" can introduce into
resolvers' caches
- add a RPZ performance test to bin/tests/system/rpz
when queryperf is available.
- the encoding of PASSTHRU action to "rpz-passthru".
(The old encoding is still accepted.)
[RT #26172]
3329. [bug] Handle RRSIG signer-name case consistently: We
generate RRSIG records with the signer-name in
lower case. We accept them with any case, but if
they fail to validate, we try again in lower case.
[RT #27451]
3328. [bug] Fixed inconsistent data checking in dst_parse.c.
[RT #29401]
3317. [func] Add ECDSA support (RFC 6605). [RT #21918]
the TOOLDIR version of libnbcompat, associated include files,
and associated defs.mk file, instead of the version from the
.OBJDIR of src/tools/compat. This should fix PR 47188.
ppc64 toolchain where function names are prefixed with a period, so "data"
becomes ".data" and ".data" is confused by the assembler with the segment
directive with the same name. Clearly this is a toolchain issue; we should
be able to call functions "text" and "data" but it is simpler to fix the
code rather than the toolchain.
