Backport an upstream patch that adds support for ELFv2 ABI on big endian
ppc64. As musl only supports ELFv2 ABI on ppc64 regardless of
endianness, this is required to be able to build OpenSSL for ppc64be.
Modify our targets patch to add linux-powerpc64-openwrt, which will use
the linux64v2 perlasm scheme. This will probably break the combination
ppc64 with glibc, but as we really only want to support musl, this
shouldn't be a problem.
Signed-off-by: Stijn Tintel <stijn@linux-ipv6.be>
Acked-by: Rui Salvaterra <rsalvaterra@gmail.com>
This version fixes two vulnerabilities:
- SM2 Decryption Buffer Overflow (CVE-2021-3711)
Severity: High
- Read buffer overruns processing ASN.1 strings (CVE-2021-3712)
Severity: Medium
Signed-off-by: Eneas U de Queiroz <cotequeiroz@gmail.com>
Before this commit, it was assumed that mkhash is in the PATH. While
this was fine for the normal build workflow, this led to some issues if
make TOPDIR="$(pwd)" -C "$pkgdir" compile
was called manually. In most of the cases, I just saw warnings like this:
make: Entering directory '/home/.../package/gluon-status-page'
bash: line 1: mkhash: command not found
bash: line 1: mkhash: command not found
bash: line 1: mkhash: command not found
bash: line 1: mkhash: command not found
bash: line 1: mkhash: command not found
bash: line 1: mkhash: command not found
bash: line 1: mkhash: command not found
bash: line 1: mkhash: command not found
[...]
While these were only warnings and the package still compiled sucessfully,
I also observed that some package even fail to build because of this.
After applying this commit, the variable $(MKHASH) is introduced. This
variable points to $(STAGING_DIR_HOST)/bin/mkhash, which is always the
correct path.
Signed-off-by: Leonardo Mörlein <me@irrelefant.net>
This version fixes 2 security vulnerabilities, among other changes:
- CVE-2021-3450: problem with verifying a certificate chain when using
the X509_V_FLAG_X509_STRICT flag.
- CVE-2021-3449: OpenSSL TLS server may crash if sent a maliciously
crafted renegotiation ClientHello message from a client.
Signed-off-by: Eneas U de Queiroz <cotequeiroz@gmail.com>
The packages feed has a proposed package for a GOST engine, which needs
support from the main openssl library. It is a default option in
OpenSSL. All that needs to be done here is to not disable it.
Package increases by a net 1-byte, so it is not really really worth
keeping this optional.
This commit also includes a commented-out example engine configuration
in openssl.cnf, as it is done for other available engines.
Signed-off-by: Eneas U de Queiroz <cotequeiroz@gmail.com>
OpenSSL downloads itself are distributed using Akamai CDN, so use these
sources as the highest priority.
Remove a stale mirror which seems to be offline for a longer time
already.
Add fallbacks to the old release path also for the mirrors.
Signed-off-by: David Bauer <mail@david-bauer.net>
This fixes 4 security vulnerabilities/bugs:
- CVE-2021-2839 - SSLv2 vulnerability. Openssl 1.1.1 does not support
SSLv2, but the affected functions still exist. Considered just a bug.
- CVE-2021-2840 - calls EVP_CipherUpdate, EVP_EncryptUpdate and
EVP_DecryptUpdate may overflow the output length argument in some
cases where the input length is close to the maximum permissable
length for an integer on the platform. In such cases the return value
from the function call will be 1 (indicating success), but the output
length value will be negative.
- CVE-2021-2841 - The X509_issuer_and_serial_hash() function attempts to
create a unique hash value based on the issuer and serial number data
contained within an X509 certificate. However it was failing to
correctly handle any errors that may occur while parsing the issuer
field (which might occur if the issuer field is maliciously
constructed). This may subsequently result in a NULL pointer deref and
a crash leading to a potential denial of service attack.
- Fixed SRP_Calc_client_key so that it runs in constant time. This could
be exploited in a side channel attack to recover the password.
The 3 CVEs above are currently awaiting analysis.
Signed-off-by: Eneas U de Queiroz <cotequeiroz@gmail.com>
Fixes: CVE-2020-1971, defined as high severity, summarized as:
NULL pointer deref in GENERAL_NAME_cmp function can lead to a DOS
attack.
Signed-off-by: Eneas U de Queiroz <cotequeiroz@gmail.com>
This sets the --cross-compile-prefix option when running Configure, so
that that it will not use the host gcc to figure out, among other
things, compiler defines. It avoids errors, if the host 'gcc' is
handled by clang:
mips-openwrt-linux-musl-gcc: error: unrecognized command-line option
'-Qunused-arguments'
Signed-off-by: Eneas U de Queiroz <cotequeiroz@gmail.com>
Tested-by: Rosen Penev <rosenp@gmail.com>
Fixes NULL dereference in SSL_check_chain() for TLS 1.3, marked with
high severity, assigned CVE-2020-1967.
Ref: https://www.openssl.org/news/secadv/20200421.txt
Signed-off-by: Petr Štetiar <ynezz@true.cz>
There were two changes between 1.1.1e and 1.1.1f:
- a change in BN prime generation to avoid possible fingerprinting of
newly generated RSA modules
- the patch reversing EOF detection we had already applied.
Signed-off-by: Eneas U de Queiroz <cotequeiroz@gmail.com>
This adds patches to avoid possible application breakage caused by a
change in behavior introduced in 1.1.1e. It affects at least nginx,
which logs error messages such as:
nginx[16652]: [crit] 16675#0: *358 SSL_read() failed (SSL: error:
4095126:SSL routines:ssl3_read_n:unexpected eof while reading) while
keepalive, client: xxxx, server: [::]:443
Openssl commits db943f4 (Detect EOF while reading in libssl), and
22623e0 (Teach more BIOs how to handle BIO_CTRL_EOF) changed the
behavior when encountering an EOF in SSL_read(). Previous behavior was
to return SSL_ERROR_SYSCALL, but errno would still be 0. The commits
being reverted changed it to SSL_ERRO_SSL, and add an error to the
stack, which is correct. Unfortunately this affects a number of
applications that counted on the old behavior, including nginx.
The reversion was discussed in openssl/openssl#11378, and implemented as
PR openssl/openssl#11400.
Signed-off-by: Eneas U de Queiroz <cotequeiroz@gmail.com>
This version includes bug and security fixes, including medium-severity
CVE-2019-1551, affecting RSA1024, RSA1536, DSA1024 & DH512 on x86_64.
Signed-off-by: Eneas U de Queiroz <cotequeiroz@gmail.com>
This adds commented configuration help for the alternate, afalg-sync
engine to /etc/ssl/openssl.cnf.
Signed-off-by: Eneas U de Queiroz <cotequeiroz@gmail.com>
This adds engine configuration sections to openssl.cnf, with a commented
list of engines. To enable an engine, all you have to do is uncomment
the engine line.
It also adds some useful comments to the devcrypto engine configuration
section. Other engines currently don't have configuration commands.
Signed-off-by: Eneas U de Queiroz <cotequeiroz@gmail.com>
This version fixes 3 low-severity vulnerabilities:
- CVE-2019-1547: ECDSA remote timing attack
- CVE-2019-1549: Fork Protection
- CVE-2019-1563: Padding Oracle in PKCS7_dataDecode and
CMS_decrypt_set1_pkey
Patches were refreshed.
Signed-off-by: Eneas U de Queiroz <cotequeiroz@gmail.com>
Highlights of this version:
- Prevent over long nonces in ChaCha20-Poly1305 (CVE-2019-1543)
- Fix OPENSSL_config bug (patch removed)
- Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
- Enable SHA3 pre-hashing for ECDSA and DSA
Signed-off-by: Eneas U de Queiroz <cote2004-github@yahoo.com>
Signed-off-by: Christian Lamparter <chunkeey@gmail.com> [DMARC removal]
No target is using kernel 3.18 anymore, remove all the generic
support for kernel 3.18.
The removed packages are depending on kernel 3.18 only and are not used on
any recent kernel.
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Add a conditional to the individual package's for the kmods in DEPENDS.
This avoids the need to compile the kernel modules when the crypto
engine packages are not selected. The final binares are not affected by
this.
Signed-off-by: Eneas U de Queiroz <cote2004-github@yahoo.com>
Tested-by: Rosen Penev <rosenp@gmail.com>
This applies an upstream patch that fixes a OPENSSL_config() bug that
causes SSL initialization to fail when the openssl.cnf file is not
found. The config file is not installed by default.
Signed-off-by: Eneas U de Queiroz <cote2004-github@yahoo.com>
The sender domain has a DMARC Reject/Quarantine policy which disallows
sending mailing list messages using the original "From" header.
To mitigate this problem, the original message has been wrapped
automatically by the mailing list software.
Enable engine support by default. Right now, some packages require
this, so it is always enabled by the bots. Many packages will compile
differently when engine support is detected, needing engine symbols from
the libraries.
However, being off by default, a user compiling its own image will fail
to run some popular packages from the official repo.
Note that disabling engines did not work in 1.0.2, so this problem never
showed up before.
NPN support has been removed in major browsers & servers, and has become
a small bloat, so it does not make sense to leave it on by default.
Remove deprecated CONFIG_ENGINE_CRYPTO symbol that is no longer needed.
Signed-off-by: Eneas U de Queiroz <cote2004-github@yahoo.com>
Openssl 1.1.0 made wholesale changes to its building system.
Apparently, parallel builds are working now.
Signed-off-by: Eneas U de Queiroz <cote2004-github@yahoo.com>
Openssh uses digest contexts across forks, which is not supported by the
/dev/crypto engine. The speed of digests is usually not worth enabling
them anyway. This changes the default of the DIGESTS option to NONE, so
the user still has the option to enable them.
Added another patch related to the use of encryption contexts across
forks, that ignores a failure to close a previous open session when
reinitializing a context, instead of failing the reinitialization.
Added a link to the Cryptographic Hardware Accelerators document to the
engine pacakges description, to provide more detailed instructions to
configure the engines.
Revert the removal of the OPENSSL_ENGINE_CRYPTO symbol, currently used
by openssh. There is an open PR to update openssh; when merged, this
symbol can be safely removed.
Signed-off-by: Eneas U de Queiroz <cote2004-github@yahoo.com>
Signed-off-by: Hans Dedecker <dedeckeh@gmail.com> [refresh patches]
The patches to the /dev/crypto engine were commited to openssl master,
and will be in the next major version (3.0).
Changes:
- Optimization in computing a digest in one operation, saving an ioctl
- Runtime configuration options for the choice of algorithms to use
- Command to dump useful information about the algorithms supported by
the engine and the system.
- Build the devcrypto engine as a dynamic module, like other engines.
The devcrypto engine is built as a separate package by default, but
options were added to allow building the engines into the main library.
Signed-off-by: Eneas U de Queiroz <cote2004-github@yahoo.com>
[refresh patches]
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
This is bugfix release that incorporated all of the devcrypto engine
patches currently in the tree.
The cleaning procedure in Package/Configure was not removing the
dependency files, causing linking errors during a rebuild with
different options. It was replaced by a simple make clean.
Signed-off-by: Eneas U de Queiroz <cote2004-github@yahoo.com>
- Add the /etc/ssl/openssl.cnf as a separate package, to avoid breaking
the transitional mechanism, allowing libopenssl_1.0* and
libopenssl_1.1* to coexist.
- Remove the (selecting) dependency on @KERNEL_AIO
- Use global SOURCE_DATE_EPOCH
Signed-off-by: Eneas U de Queiroz <cote2004-github@yahoo.com>
Add a patch to enable the option to change the default ciphersuite list
ordering to prefer ChaCha20 over AES-GCM. This is used by default for
all platforms, except for x86_64 and aarch64. The assumption is that
only the latter have AES-specific CPU instructions and asm code that
uses them in openssl. Chacha20Poly1305 is 3x faster than AES-256 in
systems without AES instructions, with an equivalent strength.
Disable error messages by default except for devices with small flash or
RAM, to aid debugging.
Disable ASM by default on arm platform with small flash. Size
difference on mips and powerpc, the other platforms with small flash
devices, are not really relevant (using 100K as a threshold). All of
the affected platforms are source-only anyway.
Signed-off-by: Eneas U de Queiroz <cote2004-github@yahoo.com>
This version adds the following functionality:
* TLS 1.3
* AFALG engine support for hardware accelleration
* x25519 ECC curve support
* CRIME protection: disable use of compression by default
* Support for ChaCha20 and Poly1305
Patches fixing bugs in the /dev/crypto engine were applied, from
https://github.com/openssl/openssl/pull/7585
This increses the size of the ipk binray on MIPS32 by about 32%:
old:
693.941 bin/packages/mips_24kc/base/libopenssl1.0.0_1.0.2q-2_mips_24kc.ipk
193.827 bin/packages/mips_24kc/base/openssl-util_1.0.2q-2_mips_24kc.ipk
new:
912.493 bin/packages/mips_24kc/base/libopenssl1.1_1.1.1a-2_mips_24kc.ipk
239.316 bin/packages/mips_24kc/base/openssl-util_1.1.1a-2_mips_24kc.ipk
Signed-off-by: Eneas U de Queiroz <cote2004-github@yahoo.com>
Adds the following configuration options:
* using optimized assembler code (was always on before)
* use of x86 SSE2 instructions
* dyanic engine support
* include error messages
* Camellia, Gost, Idea, MDC2, Seed & Whirlpool algorithms
* RFC3779, CMS protocols
* VIA padlock hardware acceleration engine
Installs openssl.cnf with the library as it is used by engines
independent of the openssl util.
Fixes DTLS option that was innefective before.
Disables insecure SSL3 protocol and SHA0.
Adds openwrt-specific targets to Configure script, including asm support
for i386, ppc and mips64.
Strips building dirs from CFLAGS shown in binary.
Skips the fuzz directory during build.
Removed include/crypto/devcrypto.h that was included here, to use the
cryptodev-linux package, now that it was been moved from the packages
feed to the main openwrt repository.
This decreses the size of the ipk binray on MIPS32 by about 3.3%:
old:
706.957 bin/packages/mips_24kc/base/libopenssl1.0.0_1.0.2q-2_mips_24kc.ipk
199.294 bin/packages/mips_24kc/base/openssl-util_1.0.2q-2_mips_24kc.ipk
new:
693.941 bin/packages/mips_24kc/base/libopenssl1.0.0_1.0.2q-2_mips_24kc.ipk
193.827 bin/packages/mips_24kc/base/openssl-util_1.0.2q-2_mips_24kc.ipk
Signed-off-by: Eneas U de Queiroz <cote2004-github@yahoo.com>
Host "gd.tuwien.ac.at" does not exists anymore, so we replace it by "ftp.pca.dfn.de" from the official list of mirrors.
Signed-off-by: Sven Roederer <devel-sven@geroedel.de>
This fixes the following security problems:
* CVE-2018-5407: Microarchitecture timing vulnerability in ECC scalar multiplication
* CVE-2018-0734: Timing vulnerability in DSA signature generation
* Resolve a compatibility issue in EC_GROUP handling with the FIPS Object Module
Signed-off-by: Sven Roederer <freifunk@it-solutions.geroedel.de>
In the case of upstream libraries, set the ABI_VERSION variable to the
soname value of the first version version after the last backwards
incompatible change.
For custom OpenWrt libraries, set the ABI_VERSION to the date of the
last Git commit doing backwards incompatible changes to the source,
such as changing function singatures or dropping exported symbols.
The soname values have been determined by either checking
https://abi-laboratory.pro/index.php?view=tracker or - in the case
of OpenWrt libraries - by carefully reviewing the changes made to
header files thorough the corresponding Git history.
In the future, the ABI_VERSION values must be bumped whenever the
library is updated to an incpompatible version but not with every
package update, in order to reduce the dependency churn in the
binary package repository.
Signed-off-by: Jo-Philipp Wich <jo@mein.io>
Some package (e.g. libunbound) depend on OPENSSL_WITH_DEPRECATED. In some
situations it may happen that libunbound and openssl are only pulled in as
build dependencies, but are not enabled in .config.
In such cases, the defaults of symbols like OPENSSL_WITH_DEPRECATED are
ignored (as the whole symbol depends on PACKAGE_libopenssl), and config
symbol dependencies of libunbound aren't effective either (as libunbound
is not actually enabled).
This commit works around the issue by introducing a hidden negated symbol
OPENSSL_NO_DEPRECATED, which is always disabled when PACKAGE_libopenssl is
disabled, and ensures that OpenSSL is built with deprecated APIs in this
case. A user can still manage to break the build by explicitly enabling
libopenssl and disabling OPENSSL_WITH_DEPRECATED; the interaction between
build dependencies and config symbols will require further discussion.
Signed-off-by: Matthias Schiffer <mschiffer@universe-factory.net>
This fixes the following security problems:
* CVE-2018-0732: Client DoS due to large DH parameter
* CVE-2018-0737: Cache timing vulnerability in RSA Key Generation
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
It's not needed now since commit a621b8c ("include: clean package
staging dir files before configure")
Fixes FS#1309
Signed-off-by: Yousong Zhou <yszhou4tech@gmail.com>
So that it will not try to run c_rehash with the just built binaries on
certs/demo.
Fixesopenwrt/packages#5432
Reported-by: Val Kulkov <val.kulkov@gmail.com>
Signed-off-by: Yousong Zhou <yszhou4tech@gmail.com>
add no-ssl3-method again as 1.0.2n compiles without the ssl3-method(s)
Fixes CVEs: CVE-2017-3737, CVE-2017-3738
Signed-off-by: Peter Wagner <tripolar@gmx.at>
CPE ids helps to tracks CVE in packages.
https://cpe.mitre.org/specification/
Thanks to swalker for CPE to package mapping and
keep tracking CVEs.
Acked-by: Jo-Philipp Wich <jo@mein.io>
Signed-off-by: Alexander Couzens <lynxis@fe80.eu>
don't set no-ssl3-method when CONFIG_OPENSSL_WITH_SSL3 di disabled otherwise the compile breaks with this error:
../libssl.so: undefined reference to `SSLv3_client_method'
Fixes CVE: CVE-2017-3735, CVE-2017-3736
Signed-off-by: Peter Wagner <tripolar@gmx.at>
OpenSSL is built with the generic linux settings for most targets,
including aarch64. These generic settings are designed for 32-bit CPU and
provide no assembler optmization: this is widely suboptimal for aarch64.
This patch simply switches to the aarch64 settings that are already
available in OpenSSL.
Here is the output of "openssl speed" before the optimization, with
"(...)" representing build flags that didn't change:
OpenSSL 1.0.2l 25 May 2017
options:bn(64,32) rc4(ptr,char) des(idx,cisc,2,int) aes(partial) blowfish(ptr)
compiler: aarch64-openwrt-linux-musl-gcc (...)
And after this patch, OpenSSL uses 64 bit mode and assembler optimizations:
OpenSSL 1.0.2l 25 May 2017
options:bn(64,64) rc4(ptr,char) des(idx,cisc,2,int) aes(partial) blowfish(ptr)
compiler: aarch64-openwrt-linux-musl-gcc (...) -DSHA1_ASM -DSHA256_ASM -DSHA512_ASM
Here are some benchmarks on a pine64+ running latest LEDE master r5142-20d363aed3:
before# openssl speed sha aes blowfish
The 'numbers' are in 1000s of bytes per second processed.
type 16 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
sha1 3918.89k 9982.43k 19148.03k 24933.03k 27325.78k
sha256 4604.51k 10240.64k 17472.51k 21355.18k 22801.07k
sha512 3662.19k 14539.41k 21443.16k 29544.11k 33177.60k
blowfish cbc 16266.63k 16940.86k 17176.92k 17237.33k 17252.35k
aes-128 cbc 19712.95k 21447.40k 22091.09k 22258.35k 22304.09k
aes-192 cbc 17680.12k 19064.47k 19572.14k 19703.13k 19737.26k
aes-256 cbc 15986.67k 17132.48k 17537.28k 17657.17k 17689.26k
after# openssl speed sha aes blowfish
type 16 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
sha1 6770.87k 26172.80k 86878.38k 205649.58k 345978.20k
sha256 20913.93k 74663.85k 184658.18k 290891.09k 351032.66k
sha512 7633.10k 30110.14k 50083.24k 71883.43k 82485.25k
blowfish cbc 16224.93k 16933.55k 17173.76k 17234.94k 17252.35k
aes-128 cbc 19425.74k 21193.31k 22065.74k 22304.77k 22380.54k
aes-192 cbc 17452.29k 18883.84k 19536.90k 19741.70k 19800.06k
aes-256 cbc 15815.89k 17003.01k 17530.03k 17695.40k 17746.60k
For some reason AES and blowfish do not benefit, but SHA performance
improves between 1.7x and 15x. SHA256 clearly benefits the most from the
optimization (4.5x on small blocks, 15x on large blocks!).
When using EVP (with "openssl speed -evp <algo>"):
# Before, EVP mode
type 16 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
sha1 3824.46k 10049.66k 19170.56k 24947.03k 27325.78k
sha256 3368.33k 8511.15k 16061.44k 20772.52k 22721.88k
sha512 2845.23k 11381.57k 19467.69k 28512.26k 33008.30k
bf-cbc 15146.74k 16623.83k 17092.01k 17211.39k 17249.62k
aes-128-cbc 17873.03k 20870.61k 21933.65k 22216.36k 22301.35k
aes-192-cbc 16184.18k 18607.15k 19447.13k 19670.02k 19737.26k
aes-256-cbc 14774.06k 16757.25k 17457.58k 17639.42k 17686.53k
# After, EVP mode
type 16 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
sha1 7056.97k 27142.10k 89515.86k 209155.41k 347419.99k
sha256 7745.70k 29750.06k 95341.48k 211001.69k 332376.75k
sha512 4550.47k 18086.06k 39997.10k 65880.75k 81431.21k
bf-cbc 15129.20k 16619.03k 17090.56k 17212.76k 17246.89k
aes-128-cbc 99619.74k 269032.34k 450214.23k 567353.00k 613933.06k
aes-192-cbc 93180.74k 231017.79k 361766.66k 433671.51k 461731.16k
aes-256-cbc 89343.23k 209858.58k 310160.04k 362234.88k 380878.85k
Blowfish does not seem to have assembler optimization at all, and SHA
still benefits (between 1.6x and 14.5x) but is generally slower than in
non-EVP mode.
However, AES performance is improved between 5.5x and 27.5x, which is
really impressive! For aes-128-cbc on large blocks, a core i7-6600U
@2.60GHz is only twice as fast...
Signed-off-by: Baptiste Jonglez <git@bitsofnetworks.org>
Refresh mirror list, some doesn't offer OpenSSL and add main site as last resort.
Source: https://www.openssl.org/source/mirror.html
Signed-off-by: Daniel Engberg <daniel.engberg.lists@pyret.net>
The current way of creating a STAMP_CONFIGURED filename for OpenSSL can
lead to an extremely long filename that makes touch unable to create it,
and fail the build.
Use mkhash to produce a hash against OPENSSL_OPTIONS which creates a
shortert stamp file,
Fixes#572
Signed-off-by: Florian Fainelli <f.fainelli@gmail.com>
This fixes the following security problems:
CVE-2017-3731: Truncated packet could crash via OOB read
CVE-2017-3732: BN_mod_exp may produce incorrect results on x86_64
CVE-2016-7055: Montgomery multiplication may produce incorrect results
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
