- Security: Fix double-free in server TCP listener cleanup A double-free
in the server could be triggered by an authenticated user if dropbear is
running with -a (Allow connections to forwarded ports from any host)
This could potentially allow arbitrary code execution as root by an
authenticated user. Affects versions 2013.56 to 2016.74. Thanks to Mark
Shepard for reporting the crash.
CVE-2017-9078 https://secure.ucc.asn.au/hg/dropbear/rev/c8114a48837c
- Security: Fix information disclosure with ~/.ssh/authorized_keys
symlink. Dropbear parsed authorized_keys as root, even if it were a
symlink. The fix is to switch to user permissions when opening
authorized_keys
A user could symlink their ~/.ssh/authorized_keys to a root-owned file
they couldn't normally read. If they managed to get that file to contain
valid authorized_keys with command= options it might be possible to read
other contents of that file.
This information disclosure is to an already authenticated user.
Thanks to Jann Horn of Google Project Zero for reporting this.
CVE-2017-9079 https://secure.ucc.asn.au/hg/dropbear/rev/0d889b068123
Refresh patches, rework 100-pubkey_path.patch to work with new
authorized_keys validation.
Signed-off-by: Kevin Darbyshire-Bryant <kevin@darbyshire-bryant.me.uk>
The only HMACs currently available use MD5 and SHA1, both of which have known
weaknesses. We already compile in the SHA256 code since we use Curve25519
by default, so there's no significant size penalty to enabling this.
Signed-off-by: Joseph C. Sible <josephcsible@users.noreply.github.com>
As security precaution and to limit the attack surface based on
the version reported by tools like nmap mask out the dropbear
version so the version is not visible anymore by snooping on the
wire. Version is still visible by 'dropbear -V'
Based on a patch by Hans Dedecker <dedeckeh@gmail.com>
Signed-off-by: Kevin Darbyshire-Bryant <kevin@darbyshire-bryant.me.uk>
Signed-off-by: Felix Fietkau <nbd@nbd.name>
Fix SIGSEGV in rfc1035.c answer_request() line 1228 where memset()
is called with header & limit pointing at the same address and thus
tries to clear memory from before the buffer begins.
answer_request() is called with an invalid edns packet size provided by
the client. Ensure the udp_size provided by the client is bounded by
512 and configured maximum as per RFC 6891 6.2.3 "Values lower than 512
MUST be treated as equal to 512"
The client that exposed the problem provided a payload udp size of 0.
Signed-off-by: Kevin Darbyshire-Bryant <kevin@darbyshire-bryant.me.uk>
Acked-by: Hans Dedecker <dedeckeh@gmail.com>
Some small tweaks and improvements :
9828ab1 Fix compiler warning.
f77700a Fix compiler warning.
0fbd980 Fix compiler warning.
43cdf1c Remove automatic IDN support when building i18n.
ff19b1a Fix &/&& confusion.
2aaea18 Add .gitattributes to substitute VERSION on export.
Signed-off-by: Hans Dedecker <dedeckeh@gmail.com>
Permit users of the full variant to disable the NO_ID *.bind pseudo
domain masking.
Defaulted 'on' in all variants.
Signed-off-by: Kevin Darbyshire-Bryant <kevin@darbyshire-bryant.me.uk>
Don't expose dnsmasq version & other data to clients via the *.bind
pseudo domain. This uses a new 'NO_ID' compile time option which has been
discussed and submitted upstream.
This is an alternate to replacing version with 'unknown' which affects
the version reported to syslog and 'dnsmasq --version'
Run time tested with & without NO_ID on Archer C7 v2
Signed-off-by: Kevin Darbyshire-Bryant <kevin@darbyshire-bryant.me.uk>
A number of small tweaks & improvements on the way to a final release.
Most notable:
Improve DHCPv4 address-in-use check.
Remove the recently introduced RFC-6842 (Client-ids in DHCP replies)
support as it turns out some clients are getting upset.
Signed-off-by: Kevin Darbyshire-Bryant <kevin@darbyshire-bryant.me.uk>
--bogus-priv now applies to IPv6 prefixes as specified in RFC6303 - this
is significantly friendlier to upstream servers.
CNAME fix in auth mode - A domain can only have a CNAME if it has no
other records
Drop 2 patches now included upstream.
Compile & run tested Archer C7 v2
Signed-off-by: Kevin Darbyshire-Bryant <kevin@darbyshire-bryant.me.uk>
New test release (since test1) includes 2 LEDE patches that are
upstream and may be dropped, along with many spelling fixes.
Add forthcoming 2017 root zone trust anchor to trust-anchors.conf.
Backport 2 patches that just missed test3:
Reduce logspam of those domains handled locally 'local addresses only'
Implement RFC-6842 (Client-ids in DHCP replies)
Compile & run tested Archer C7 v2
Signed-off-by: Kevin Darbyshire-Bryant <kevin@darbyshire-bryant.me.uk>
Bump to dnsmasq 2.77test1 - this includes a number of fixes since 2.76
and allows dropping of 2 LEDE carried patches.
Notable fix in rrfilter code when talking to Nominum's DNS servers
especially with DNSSEC.
A patch to switch dnsmasq back to 'soft fail' for SERVFAIL responses
from dns servers is also included. This mean dnsmasq tries all
configured servers before giving up.
A 'localise queries' enhancement has also been backported (it will
appear in test2/rc'n') this is especially important if using the
recently imported to LEDE 'use dnsmasq standalone' feature 9525743c
I have been following dnsmasq HEAD ever since 2.76 release.
Compile & Run tested: ar71xx, Archer C7 v2
Tested-by: Kevin Darbyshire-Bryant <kevin@darbyshire-bryant.me.uk>
Signed-off-by: Kevin Darbyshire-Bryant <kevin@darbyshire-bryant.me.uk>
Fixes a 100% cpu usage issue if using dhcp-script.
Signed-off-by: Kevin Darbyshire-Bryant <kevin@darbyshire-bryant.me.uk>
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Add a new variable TARGET_INIT_PATH which holds the default $PATH variable
value configured in menuconfig.
Signed-off-by: Jo-Philipp Wich <jow@openwrt.org>
- Security: Message printout was vulnerable to format string injection.
If specific usernames including "%" symbols can be created on a system
(validated by getpwnam()) then an attacker could run arbitrary code as root
when connecting to Dropbear server.
A dbclient user who can control username or host arguments could potentially
run arbitrary code as the dbclient user. This could be a problem if scripts
or webpages pass untrusted input to the dbclient program.
- Security: dropbearconvert import of OpenSSH keys could run arbitrary code as
the local dropbearconvert user when parsing malicious key files
- Security: dbclient could run arbitrary code as the local dbclient user if
particular -m or -c arguments are provided. This could be an issue where
dbclient is used in scripts.
- Security: dbclient or dropbear server could expose process memory to the
running user if compiled with DEBUG_TRACE and running with -v
The security issues were reported by an anonymous researcher working with
Beyond Security's SecuriTeam Secure Disclosure www.beyondsecurity.com/ssd.html
Signed-off-by: Jo-Philipp Wich <jo@mein.io>
Update the dropbear package to version 2016.73, refresh patches.
The measured .ipk sizes on an x86_64 build are:
94588 dropbear_2015.71-3_x86_64.ipk
95316 dropbear_2016.73-1_x86_64.ipk
This is an increase of roughly 700 bytes after compression.
Tested-by: Kevin Darbyshire-Bryant <kevin@darbyshire-bryant.me.uk>
Signed-off-by: Jo-Philipp Wich <jo@mein.io>
Utmp support tracks who is currenlty logged in by logging info to the file /var/run/utmp (supported by busybox)
Putuline support will use the utmp structure to write to the utmp file
Signed-off-by: Hans Dedecker <dedeckeh@gmail.com>
A dropbear instance having an interface config won't start if the interface is down as no
IP address is available.
Adding interface triggers for each configured interface executing the dropbear reload script
will start the dropbear instance when the interface is up.
Signed-off-by: Hans Dedecker <dedeckeh@gmail.com>
While technically required by the RFC, they are usually completely
unused (DSA), or have security issues (3DES, CBC)
Signed-off-by: Felix Fietkau <nbd@openwrt.org>
This enables passworldless login for root via SSH whenever no root
password is set (e.g. after reset, flashing without keeping config
or in failsafe) and removes telnet support alltogether.
Signed-off-by: Steven Barth <steven@midlink.org>
The dm9601 driver expects to receive a single encapsulated ethernet
frame from the device in one URB transfer, and it provides an URB
buffer of length 1,522 to receive it. This is not a round multiple
of USB transfer packets.
The device in question [1] provides a stream of such frames and it
does not conveniently slice them up as the dm9601 driver expects. We
can end up with 1,536 (0x600) bytes returned by the device in response
to the URB request. This may include several encapsulated ethernet
frames, and/or fragments thereof.
It seems to me that the kernel 'Oops' arises because the dwc_otg driver
does not notice that the destination buffer is too small to receive the
full 1,536 bytes. Comparing dwc_otg's update_urb_state_xfer_comp with
dwc2's dwc2_update_urb_state is suggestive.
More details: https://github.com/raspberrypi/linux/issues/1045
All Credits to: https://github.com/mw9
Signed-off-by: Marian Hello <marian.hello@gmail.com>
Reviewed-by: Zoltan HERPAI <wigyori@uid0.hu>
Patch 610- is updated as check_entry helper was killed in 3.18.37
Fixes CVE-2016-5195 (dirtycow)
Compile-tested on ar71xx, mxs, sunxi, imx6
Runtime-tested on ar71xx (PB42)
Signed-off-by: Zoltan HERPAI <wigyori@uid0.hu>
Zoltan HERPAI
Stijn Tintel
Kevin Darbyshire-Bryant
Joseph C. Sible
Hans Dedecker
Hauke Mehrtens
Steven Barth
Jo-Philipp Wich
Hannu Nyman
Felix Fietkau
Alexandru Ardelean
Steven Barth
Marian Hello
Gad Krumholz
Matthias Schiffer
Shaleen Jain
Jing Lin
Imre Kaloz