1 |
|
Updating Information for MidnightBSD users. |
2 |
|
|
3 |
+ |
20171003: |
4 |
+ |
SQLite 3.20.1 |
5 |
+ |
|
6 |
+ |
20171001: |
7 |
+ |
Subversion 1.8.17 |
8 |
+ |
|
9 |
+ |
Perl 5.26.0 |
10 |
+ |
|
11 |
+ |
Change 0.10 version to 1.0. There are several compatibility issues |
12 |
+ |
with using 0.10 as the trailing zero is dropped in several utilities |
13 |
+ |
making it look like 0.1. |
14 |
+ |
|
15 |
+ |
20170918: |
16 |
+ |
Introduce nvme(4) and nvd(4) from FreeBSD. |
17 |
+ |
|
18 |
+ |
Fix build of boot code and rescue. |
19 |
+ |
|
20 |
+ |
20170819: |
21 |
+ |
Heimdal KDC-REP service name validation vulerability patched. |
22 |
+ |
|
23 |
+ |
Introduce a partial fix for AMD Ryzen issues. On Ryzen, move |
24 |
+ |
the lower shared page by one. |
25 |
+ |
|
26 |
+ |
20170326: |
27 |
+ |
sudo removed from base. Use doas(1) or install sudo from mports |
28 |
+ |
|
29 |
+ |
Stable 0.9 created, continue development on 0.10 |
30 |
+ |
|
31 |
+ |
20170305: |
32 |
+ |
Add hast module to bsnmpd |
33 |
+ |
|
34 |
+ |
20170302: |
35 |
+ |
add a callback to the ada(4) driver so that it knows when |
36 |
+ |
GEOM has released references to it. |
37 |
+ |
|
38 |
+ |
20170219: |
39 |
+ |
Add /dev/full device. |
40 |
+ |
|
41 |
+ |
The lindev device has been removed since /dev/full has been made a |
42 |
+ |
standard device. |
43 |
+ |
|
44 |
+ |
Serf 1.3.9 |
45 |
+ |
Subversion 1.8.10 |
46 |
+ |
apr 1.5.2 |
47 |
+ |
apr-util 1.5.4 |
48 |
+ |
|
49 |
+ |
20170129: |
50 |
+ |
add doas utility from OpenBSD. |
51 |
+ |
|
52 |
+ |
20161105: |
53 |
+ |
BIND 9.9.9-p4 |
54 |
+ |
|
55 |
+ |
OpenSSH 7.3p1 |
56 |
+ |
|
57 |
+ |
20161103: |
58 |
+ |
OpenSSL security patch |
59 |
+ |
|
60 |
+ |
Due to improper handling of alert packets, OpenSSL would consume an excessive |
61 |
+ |
amount of CPU time processing undefined alert messages. |
62 |
+ |
|
63 |
+ |
20161015: |
64 |
+ |
libarchive 3.2.1 |
65 |
+ |
|
66 |
+ |
xz 5.2.2 |
67 |
+ |
|
68 |
+ |
20161013: |
69 |
+ |
Sync ZFS code with Illuminos/FreeBSD 9.2. Added support for |
70 |
+ |
feature flags, pool version 5000. This also includes some |
71 |
+ |
bug fixes and performance optimizations. |
72 |
+ |
|
73 |
+ |
20160925: |
74 |
+ |
Import NetBSD vis(3) and unvis(3) as well as mtree. |
75 |
+ |
|
76 |
+ |
one-true-awk 20121220 |
77 |
+ |
|
78 |
+ |
inetd now honors kern.ipc.somaxconn value. |
79 |
+ |
|
80 |
+ |
netmap synced with FreeBSD 9.2 |
81 |
+ |
|
82 |
+ |
linuxolator now has dtrace probes. |
83 |
+ |
|
84 |
+ |
bsdgrep now correctly handles -m to exclude only one file. |
85 |
+ |
|
86 |
+ |
UFS file systems can now be resized in read-write mode due to the new |
87 |
+ |
write suspension feature. |
88 |
+ |
|
89 |
+ |
Basic support added for Intel Raid Recover Technology. |
90 |
+ |
|
91 |
+ |
GMIRROR & GRAID3 now mark volumes clean on shutdown earlier to help with ZFS issues. |
92 |
+ |
|
93 |
+ |
Highpoint hpt27xx now in GENERIC kernel. |
94 |
+ |
|
95 |
+ |
20160923: |
96 |
+ |
Security update for OpenSSL |
97 |
+ |
|
98 |
+ |
A malicious client can send an excessively large OCSP Status Request extension. |
99 |
+ |
If that client continually requests renegotiation, sending a large OCSP Status |
100 |
+ |
Request extension each time, then there will be unbounded memory growth on the |
101 |
+ |
server. [CVE-2016-6304] |
102 |
+ |
|
103 |
+ |
An overflow can occur in MDC2_Update() either if called directly or through |
104 |
+ |
the EVP_DigestUpdate() function using MDC2. If an attacker is able to supply |
105 |
+ |
very large amounts of input data after a previous call to EVP_EncryptUpdate() |
106 |
+ |
with a partial block then a length check can overflow resulting in a heap |
107 |
+ |
corruption. [CVE-2016-6303] |
108 |
+ |
|
109 |
+ |
If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a |
110 |
+ |
DoS attack where a malformed ticket will result in an OOB read which will |
111 |
+ |
ultimately crash. [CVE-2016-6302] |
112 |
+ |
|
113 |
+ |
The function BN_bn2dec() does not check the return value of BN_div_word(). |
114 |
+ |
This can cause an OOB write if an application uses this function with an |
115 |
+ |
overly large BIGNUM. This could be a problem if an overly large certificate |
116 |
+ |
or CRL is printed out from an untrusted source. TLS is not affected because |
117 |
+ |
record limits will reject an oversized certificate before it is parsed. |
118 |
+ |
[CVE-2016-2182] |
119 |
+ |
|
120 |
+ |
The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is |
121 |
+ |
the total length the OID text representation would use and not the amount |
122 |
+ |
of data written. This will result in OOB reads when large OIDs are presented. |
123 |
+ |
[CVE-2016-2180] |
124 |
+ |
|
125 |
+ |
Some calculations of limits in OpenSSL have used undefined pointer arithmetic. |
126 |
+ |
This could cause problems with some malloc implementations. [CVE-2016-2177] |
127 |
+ |
|
128 |
+ |
Operations in the DSA signing algorithm should run in constant time in order to |
129 |
+ |
avoid side channel attacks. A flaw in the OpenSSL DSA implementation means that |
130 |
+ |
a non-constant time codepath is followed for certain operations. [CVE-2016-2178] |
131 |
+ |
|
132 |
+ |
In a DTLS connection where handshake messages are delivered out-of-order those |
133 |
+ |
messages that OpenSSL is not yet ready to process will be buffered for later |
134 |
+ |
use. Under certain circumstances, a flaw in the logic means that those messages |
135 |
+ |
do not get removed from the buffer even though the handshake has been completed. |
136 |
+ |
An attacker could force up to approx. 15 messages to remain in the buffer when |
137 |
+ |
they are no longer required. These messages will be cleared when the DTLS |
138 |
+ |
connection is closed. The default maximum size for a message is 100k. Therefore |
139 |
+ |
the attacker could force an additional 1500k to be consumed per connection. |
140 |
+ |
[CVE-2016-2179] |
141 |
+ |
|
142 |
+ |
A flaw in the DTLS replay attack protection mechanism means that records that |
143 |
+ |
arrive for future epochs update the replay protection "window" before the MAC |
144 |
+ |
for the record has been validated. This could be exploited by an attacker by |
145 |
+ |
sending a record for the next epoch (which does not have to decrypt or have a |
146 |
+ |
valid MAC), with a very large sequence number. This means that all subsequent |
147 |
+ |
legitimate packets are dropped causing a denial of service for a specific |
148 |
+ |
DTLS connection. [CVE-2016-2181] |
149 |
+ |
|
150 |
+ |
In OpenSSL 1.0.2 and earlier some missing message length checks can result in |
151 |
+ |
OOB reads of up to 2 bytes beyond an allocated buffer. There is a theoretical |
152 |
+ |
DoS risk but this has not been observed in practice on common platforms. |
153 |
+ |
[CVE-2016-6306] |
154 |
+ |
|
155 |
+ |
20160918: |
156 |
+ |
With the addition of auditdistd(8), a new auditdistd user is now |
157 |
+ |
depended on during installworld. "mergemaster -p" can be used to add |
158 |
+ |
the user prior to installworld. |
159 |
+ |
|
160 |
+ |
The VFS KBI was changed with the merge of several nullfs |
161 |
+ |
optimizations and fixes. All filesystem modules must be |
162 |
+ |
recompiled. |
163 |
+ |
|
164 |
+ |
20160916: |
165 |
+ |
The random(4) support for the VIA hardware random number |
166 |
+ |
generator (`PADLOCK') is no longer enabled unconditionally. |
167 |
+ |
Add the PADLOCK_RNG option in the custom kernel config if |
168 |
+ |
needed. The GENERIC kernels on i386 and amd64 do include the |
169 |
+ |
option, so the change only affects the custom kernel |
170 |
+ |
configurations. |
171 |
+ |
|
172 |
+ |
A new version of ZFS (pool version 5000) has been merged. |
173 |
+ |
Starting with this version the old system of ZFS pool versioning |
174 |
+ |
is superseded by "feature flags". This concept enables forward |
175 |
+ |
compatibility against certain future changes in functionality of ZFS |
176 |
+ |
pools. The first two read-only compatible "feature flags" for ZFS |
177 |
+ |
pools are "com.delphix:async_destroy" and "com.delphix:empty_bpobj". |
178 |
+ |
For more information read the new zpool-features(7) manual page. |
179 |
+ |
Please refer to the "ZFS notes" section of this file for information |
180 |
+ |
on upgrading boot ZFS pools. |
181 |
+ |
|
182 |
+ |
20160906: |
183 |
+ |
Add support for the MosChip MCS9904 four serial ports |
184 |
+ |
controller. |
185 |
+ |
|
186 |
+ |
Add support for walltimestamp in DTrace. |
187 |
+ |
|
188 |
+ |
Various gdb improvments. |
189 |
+ |
|
190 |
+ |
ZFS |
191 |
+ |
Import the zio nop-write improvement from Illumos. To reduce I/O, |
192 |
+ |
nop-write omits overwriting data if the checksum (cryptographically |
193 |
+ |
secure) of new data matches the checksum of existing data. |
194 |
+ |
It also saves space if snapshots are in use. |
195 |
+ |
|
196 |
+ |
It currently works only on datasets with enabled compression, disabled |
197 |
+ |
deduplication and sha256 checksums. |
198 |
+ |
|
199 |
+ |
Add loader(8) tunable to enable/disable nopwrite functionality: |
200 |
+ |
vfs.zfs.nopwrite_enabled |
201 |
+ |
|
202 |
+ |
Introduce a new dataset aclmode setting "restricted" to protect ACL's |
203 |
+ |
being destroyed or corrupted by a drive-by chmod. |
204 |
+ |
|
205 |
+ |
New loader-only tunables: |
206 |
+ |
vfs.zfs.sync_pass_deferred_free |
207 |
+ |
vfs.zfs.sync_pass_dont_compress |
208 |
+ |
vfs.zfs.sync_pass_rewrite |
209 |
+ |
|
210 |
+ |
chkgrp(8) add support for q flag |
211 |
+ |
|
212 |
+ |
Fix problem with the Samsung 840 PRO series SSD detection. |
213 |
+ |
The device reports support for SATA Asynchronous Notification in its |
214 |
+ |
IDENTIFY data, but returns error on attempt to enable that feature. |
215 |
+ |
Make SATA XPT of CAM only report these errors, but not fail the device. |
216 |
+ |
|
217 |
+ |
20160905: |
218 |
+ |
Add a resource limit for the total number of kqueues |
219 |
+ |
available to the user. Kqueue now saves the ucred of the |
220 |
+ |
allocating thread, to correctly decrement the counter on close. |
221 |
+ |
Based on FreeBSD SVN 256849 |
222 |
+ |
|
223 |
+ |
Import netcat from OpenBSD 5.2 |
224 |
+ |
|
225 |
+ |
20160904: |
226 |
+ |
Introduced experimental TCP sysctls starting with |
227 |
+ |
net.inet.tcp.experimental.initcwnd10 |
228 |
+ |
|
229 |
+ |
20160814: |
230 |
+ |
switched default desktop port to midnightbsd-desktop. This gives us flexibility to change it |
231 |
+ |
in the release after the fact. |
232 |
+ |
|
233 |
+ |
tzdata 2016a |
234 |
+ |
|
235 |
+ |
20160811: |
236 |
+ |
libdispatch 210 |
237 |
+ |
|
238 |
+ |
Added quirks for several models of SSDs to enable advanced format/4k mode. List includes |
239 |
+ |
Samsung 830, 840, 850 and 750 series, Intel x25 and a few Toshiba models. Also |
240 |
+ |
added WD Red drives. |
241 |
+ |
|
242 |
+ |
Updated list of pci device vendors. |
243 |
+ |
|
244 |
+ |
Updated list of usb devices. |
245 |
+ |
|
246 |
+ |
20160807: |
247 |
+ |
Implement several changes to libmport to fix some memory corruption issues. |
248 |
+ |
|
249 |
+ |
20160806: |
250 |
+ |
sqlite3 3.13.0 |
251 |
+ |
|
252 |
+ |
20160805: |
253 |
+ |
Merged fixes for libmport that improve error handling when installing packages. Also |
254 |
+ |
support mkdir -p like behavior for plist entries. |
255 |
+ |
|
256 |
+ |
20160531: |
257 |
+ |
Fix four security issues with MidnightBSD. |
258 |
+ |
|
259 |
+ |
The implementation of TIOCGSERIAL ioctl(2) does not clear the output |
260 |
+ |
struct before sending to userland in the linux emulation layer. |
261 |
+ |
|
262 |
+ |
The compat 43 stat(2) system call exposes kernel stack to userland. |
263 |
+ |
|
264 |
+ |
libarchive - CVE-2015-2304 and CVE-2013-0211 fix issues with |
265 |
+ |
cpio directory traversal and an integer signedness error in the archive |
266 |
+ |
write zip data routine. |
267 |
+ |
|
268 |
+ |
|
269 |
+ |
20160528: |
270 |
+ |
Fixed minor issues with mined(1) and msearch(1). |
271 |
+ |
|
272 |
+ |
20160526: |
273 |
+ |
Add support for Ivybridge and Haswell Intel CPUs to hwpmc(4). |
274 |
+ |
|
275 |
+ |
Fix libpmc(3) build with clang compiler. |
276 |
+ |
|
277 |
+ |
20160519: |
278 |
+ |
Kernel Security updates |
279 |
+ |
|
280 |
+ |
atkbd(4) - Incorrect signedness comparison in the ioctl(2) handler allows a malicious |
281 |
+ |
local user to overwrite a portion of the kernel memory. |
282 |
+ |
|
283 |
+ |
Incorrect argument handling in sendmsg(2) |
284 |
+ |
|
285 |
+ |
Incorrect argument handling in the socket code allows malicious local |
286 |
+ |
user to overwrite large portion of the kernel memory. |
287 |
+ |
|
288 |
+ |
20160505: |
289 |
+ |
OpenSSL security patch |
290 |
+ |
|
291 |
+ |
The padding check in AES-NI CBC MAC was rewritten to be in constant time |
292 |
+ |
by making sure that always the same bytes are read and compared against |
293 |
+ |
either the MAC or padding bytes. But it no longer checked that there was |
294 |
+ |
enough data to have both the MAC and padding bytes. [CVE-2016-2107] |
295 |
+ |
|
296 |
+ |
An overflow can occur in the EVP_EncodeUpdate() function which is used for |
297 |
+ |
Base64 encoding of binary data. [CVE-2016-2105] |
298 |
+ |
|
299 |
+ |
An overflow can occur in the EVP_EncryptUpdate() function, however it is |
300 |
+ |
believed that there can be no overflows in internal code due to this problem. |
301 |
+ |
[CVE-2016-2106] |
302 |
+ |
|
303 |
+ |
When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio() |
304 |
+ |
a short invalid encoding can casuse allocation of large amounts of memory |
305 |
+ |
potentially consuming excessive resources or exhausting memory. |
306 |
+ |
[CVE-2016-2109] |
307 |
+ |
|
308 |
+ |
20160412: |
309 |
+ |
0.8 stable branch created. Continue development as 0.9. |
310 |
+ |
|
311 |
+ |
Fix several issues with wait6 system call addition. |
312 |
+ |
|
313 |
+ |
20160409: |
314 |
+ |
libmport now supports two new plist formats: |
315 |
+ |
@(root,wheel,4775) myfile |
316 |
+ |
@dir(root,wheel,775) mydir |
317 |
+ |
|
318 |
+ |
On delete, absoluate paths are now handled properly. |
319 |
+ |
|
320 |
+ |
20160317: |
321 |
+ |
OpenSSH doesn't have the luck of the Irish. |
322 |
+ |
|
323 |
+ |
Fix a security issue with OpenSSH X11 forwarding that can allow an attacker |
324 |
+ |
run shell commands on the call to xauth. |
325 |
+ |
|
326 |
+ |
Incorrect argument validation in sysarch(2) |
327 |
+ |
|
328 |
+ |
A special combination of sysarch(2) arguments, specify a request to |
329 |
+ |
uninstall a set of descriptors from the LDT. The start descriptor |
330 |
+ |
is cleared and the number of descriptors are provided. Due to invalid |
331 |
+ |
use of a signed intermediate value in the bounds checking during argument |
332 |
+ |
validity verification, unbound zero'ing of the process LDT and adjacent |
333 |
+ |
memory can be initiated from usermode. |
334 |
+ |
|
335 |
+ |
Patch obtained from FreeBSD. |
336 |
+ |
|
337 |
+ |
20160229: |
338 |
+ |
top now displays information on ZFS arc cache. |
339 |
+ |
|
340 |
+ |
20160228: |
341 |
+ |
llvm + clang 3.3 is now the default compiler in MidnightBSD. |
342 |
+ |
|
343 |
+ |
20160222: |
344 |
+ |
Introduce pipe2 to linux emulation layer. |
345 |
+ |
|
346 |
+ |
20160114: |
347 |
+ |
OpenSSL |
348 |
+ |
|
349 |
+ |
The signature verification routines will crash with a NULL pointer dereference |
350 |
+ |
if presented with an ASN.1 signature using the RSA PSS algorithm and absent |
351 |
+ |
mask generation function parameter. [CVE-2015-3194] |
352 |
+ |
|
353 |
+ |
When presented with a malformed X509_ATTRIBUTE structure, OpenSSL will leak |
354 |
+ |
memory. [CVE-2015-3195] |
355 |
+ |
|
356 |
+ |
If PSK identity hints are received by a multi-threaded client then the values |
357 |
+ |
are incorrectly updated in the parent SSL_CTX structure. [CVE-2015-3196] |
358 |
+ |
|
359 |
+ |
Fix security on bsnmpd configuration file during installation. |
360 |
+ |
|
361 |
+ |
TCP MD5 signature denial of service |
362 |
+ |
|
363 |
+ |
A programming error in processing a TCP connection with both TCP_MD5SIG |
364 |
+ |
and TCP_NOOPT socket options may lead to kernel crash. |
365 |
+ |
|
366 |
+ |
SCTP |
367 |
+ |
|
368 |
+ |
A lack of proper input checks in the ICMPv6 processing in the SCTP stack |
369 |
+ |
can lead to either a failed kernel assertion or to a NULL pointer |
370 |
+ |
dereference. In either case, a kernel panic will follow. |
371 |
+ |
|
372 |
+ |
20160102: |
373 |
+ |
Happy New Year |
374 |
+ |
|
375 |
+ |
20151101: |
376 |
+ |
Increase kern.ipc.somaxconn default to 256. |
377 |
+ |
|
378 |
+ |
20151017: |
379 |
+ |
Add initial statistics api to libmport and a driver to print |
380 |
+ |
it in mport(1). |
381 |
+ |
|
382 |
+ |
20151002: |
383 |
+ |
Revised rpcbind(8) patch to fix issues with NIS |
384 |
+ |
|
385 |
+ |
20150930: |
386 |
+ |
In rpcbind(8), netbuf structures are copied directly, which would result in |
387 |
+ |
two netbuf structures that reference to one shared address buffer. When one |
388 |
+ |
of the two netbuf structures is freed, access to the other netbuf structure |
389 |
+ |
would result in an undefined result that may crash the rpcbind(8) daemon. |
390 |
+ |
|
391 |
+ |
20150926: |
392 |
+ |
libmport now supports @preexec, @postexec, @preunexec and @postunexec |
393 |
+ |
to replace @exec and @unexec. |
394 |
+ |
|
395 |
+ |
pre exec runs afer pre-install scripts but before actual installation |
396 |
+ |
|
397 |
+ |
post exec runs after install but before post install scripts and |
398 |
+ |
pkg message. |
399 |
+ |
|
400 |
+ |
pre unexec runs before pre uninstall scripts |
401 |
+ |
|
402 |
+ |
post unexec runs before de-install scripts and after file removal. |
403 |
+ |
|
404 |
+ |
20150917: |
405 |
+ |
Fix kqueue write events for files > 2GB |
406 |
+ |
|
407 |
|
20150825: |
408 |
|
kernel: |
409 |
|
fix a security issue on amd64 where the GS segment CPU register can be changed via |