libkcapi/libkcapi.changes
Marcus Meissner 742ba25f6d Accepting request 830821 from home:dirkmueller:branches:security
- update to 1.2.0:
 * enhancement: kcapi-hasher: add madvise and 64 bit support by Brandur Simonsen
 * fix: fix clang warnding in KDF implementation by Khem Raj
 * fix: fix inverted logic in kcapi-main test logic reported by Ondrej Mosnáček
 * fix: return error when iteration count is zero for PBKDF as reported by
   Guido Vranken
 * enhancement: add function kcapi_cipher_stream_update_last to indicate the
   last block of a symmetric cipher stream operation
 * disable XTS multithreaded tests as it triggers a race discussed in
   https://github.com/smuellerDD/libkcapi/issues/92. The conclusion is
   the following: xts(aes) doesn't support chaining requests like for other
   ciphers such as CBC (at least as implemented in the kernel Crypto API).
   That can be seen in `crypto/testmgr.h` - the ciphers that are expected to
   return IVs usable for chaining have the `.iv_out` entries filled in in their
   test vectors (and those that don't support it do not). One can see that only
   CTR and CBC test vectors have them, not XTS.
   Looking again at how XTS is defined, it seems one could implement
   transparent chaining by simply decrypting the final tweak using the tweak
   key and return it as the output IV... but I believe this has never been
   mandated nor implemented in the Crypto API (likely because of the overhead
   of the final tweak decryption, which would be pointless if you're not going
   to use the output IV - and there is currently no way to signal to the driver
   that you are going to need it).
 * disable AIO parallel tests due to undefined behavior

OBS-URL: https://build.opensuse.org/request/show/830821
OBS-URL: https://build.opensuse.org/package/show/security/libkcapi?expand=0&rev=35
2020-09-03 06:23:31 +00:00

190 lines
11 KiB
Plaintext
Raw Blame History

This file contains invisible Unicode characters

This file contains invisible Unicode characters that are indistinguishable to humans but may be processed differently by a computer. If you think that this is intentional, you can safely ignore this warning. Use the Escape button to reveal them.

-------------------------------------------------------------------
Mon Aug 31 13:30:58 UTC 2020 - Dirk Mueller <dmueller@suse.com>
- update to 1.2.0:
* enhancement: kcapi-hasher: add madvise and 64 bit support by Brandur Simonsen
* fix: fix clang warnding in KDF implementation by Khem Raj
* fix: fix inverted logic in kcapi-main test logic reported by Ondrej Mosnáček
* fix: return error when iteration count is zero for PBKDF as reported by
Guido Vranken
* enhancement: add function kcapi_cipher_stream_update_last to indicate the
last block of a symmetric cipher stream operation
* disable XTS multithreaded tests as it triggers a race discussed in
https://github.com/smuellerDD/libkcapi/issues/92. The conclusion is
the following: xts(aes) doesn't support chaining requests like for other
ciphers such as CBC (at least as implemented in the kernel Crypto API).
That can be seen in `crypto/testmgr.h` - the ciphers that are expected to
return IVs usable for chaining have the `.iv_out` entries filled in in their
test vectors (and those that don't support it do not). One can see that only
CTR and CBC test vectors have them, not XTS.
Looking again at how XTS is defined, it seems one could implement
transparent chaining by simply decrypting the final tweak using the tweak
key and return it as the output IV... but I believe this has never been
mandated nor implemented in the Crypto API (likely because of the overhead
of the final tweak decryption, which would be pointless if you're not going
to use the output IV - and there is currently no way to signal to the driver
that you are going to need it).
* disable AIO parallel tests due to undefined behavior
-------------------------------------------------------------------
Wed Jan 8 07:23:22 UTC 2020 - Marcus Meissner <meissner@suse.com>
- updated to 1.1.5:
- Fix invocation of ansi_cprng in FIPS mode during testing
- Fix testing on kernels >= 5.0
- Add virtualization test for kernel 5.1
- Fix the limit between vmsplice() and sendmsg() by Christophe Leroy
- Fix remove code duplication by Ondrej Mosnáček
- Fix potential memleak in speed-test
- updated to 1.1.4:
- Fix: use sendmsg when processing more than 1<<16 bytes input data which improves performance on some architectures
- updated to 1.1.3:
- Fix: default location of FIPS 140-2 HMAC control file is .<orig file>.hmac (was accidentally moved to <orig file>.hmac with 1.1.0)
- updated to 1.1.2:
- Fix: Bug fixes for GCC 8.1.0 regarding string length checks by Krzysztof Kozlowski
- Enhancement: ensure that tests execute on architectures other than X86 by Ondrej Mosnáček
- Fix: Bug fix to initialize FDs at the correct time in kcapi-kernel-if.c by Ondrej Mosnáček
- Test fix: Support test execution outside build environment by Ondrej Mosnáček
- updated to 1.1.1:
- Fix: Bug fixes for kcapi_hasher by Ondrej Mosnáček
- updated to 1.1.0:
- API Enhancement: Addition of kcapi_handle_reinit
- Fix: simplify code by removing the internal *_fd functions from kcapi-kernel-if.c
- Test enhancement: add IIV speed testing
- Fix: add a loop around the read system call to always obtain all generated data
- Fix: use host compiler for compiling docproc (reported by Christophe LEROY, fixed by Björn Esser)
- Fix: make error handling of hashing applications consistent with coreutils applications (reported by Christophe LEROY)
- Fix: support for zero length files (patched by Ondrej Mosnáček)
- Fix: support for zero message hashes on kernels <= 4.9 (patched by Ondrej Mosnáček)
- Fix: Add Travis CI test system provided by Ondrej Mosnáček
- Fix: Add several fixes to kcapi-hasher by Ondrej Mosnáček
- Fix: Add additional tests for kcapi-hasher by Ondrej Mosnáček
- Fix: Apply unpadding only to last block of data by Ondrej Mosnáček
- Fix: Fix resource leaks in error code paths suggested by Ondrej Mosnáček
- Enhancement: achieve hmaccalc CLI equivalence by Ondrej Mosnáček
- updated to 1.0.3:
- Fix: support STDIN and --tag of sha*sum applications
- Enhancement: Add small enhancements to support integration with distros -- reported by Björn Esser
- updated to 1.0.2:
- Fix: hasher-test.sh on 32-bit systems
- Fix: AIO return code handling on large number of requests -- reported by Jonathan Cameron
- Enhancement: disable coredumps of library
- Fix: remove unchecked -fstack-protector-strong from Makefile -- reported by Mathieu Malaterre
- Fix: document that kcapi_cipher_stream_op must be called in a loop to collect all data in a multhreaded environment.
- Test Fix: Update symmetric multithreaded stream test to invoke kcapi_cipher_stream_op in a loop to collect all data.
- Fix: Initialize the cipher handle on stack with zeros as the library expects a zero-initialized cipher handle. This fixes a possible segfault where free() is called on a non-initialized memory location.
- Fix: port algif_kpp and algif_akcipher to 4.15-rc3
- updated to 1.0.1:
- Fix: constify AEAD cipher input data
- Fix: use GCC byte swapping acceleration if present
- Fix: KDF counter handling on little endian systems when generating more than 255 blocks
- Use LD_PRELOAD for execution of test cases to force using of the freshly compiled binaries
- Fix: return code handling of _kcapi_common_vmsplice_chunk_fd as reported by Christophe Leroy
- Fix: return code handling in _kcapi_md_update
- Fix: kcapi-hasher now supports files larger than 2GB
- Fix: kcapi-dgst now supports files larger than 2GB
- Fix: use stack protector
- Fix: rename header guards to remove leading underscore as pointed out by Markus Elfring
- Test Fix: Allow compiing the test code without asymmetric and KPP support
- updated to 1.0.0:
- Fix: Small compile fixes for new checks of GCC 7
- API Change: Rename all LOG_* enums to KCAPI_LOG_* to prevent namespace poisoning
- Fix: soname and file name of library now compiles with conventions (thanks to Marcus Meissner)
- Fix: kcapi-rng.c: unify FD/syscall read code and fix __NR_getrandom resolution
- Enhancement: add kcapi-enc application to access symmetric encryption on command line
- Fix: consolidate duplicate code in kcapi-hasher
- Enhancement: add kcapi-dgst application to access hashes on command line
- Enhancement: add kcapi-rng man page
- Enhancement: add kcapi-rng --hex command line option
- Fix: enable full symmetric AIO support
- Fix: consolidate all test code into test/ and invoke all tests with test-invocation.sh
- Fix: fix memleaks in error code paths as reported by clang
- Fix: reduce memory footprint by rearranging data structures
- Fix: kcapi-hasher is now fully FIPS 140-2 compliant as it now includes the integrity test for libkcapi.so
- Enhancement: Add speed tests for MV-CESA accelerated ciphers and hash algorithms (thanks to Bastian Stender)
- Test Enhancement: add kcapi-enc-test-large.c test testing edge conditions of AF_ALG
- Test Enhancement: add virttest.sh - use of test system based on eudyptula-boot to test on linux-4.3.6, linux-4.4.86, linux-4.5, linux-4.7, linux-4.10, linux-4.12
- Test Enhancement: add kcapi-fuzz-test.sh to support fuzzing the AF_ALG interfaces
- Enhancement: add RPM SPEC file (tested with Fedora 26)
- API Change: replace --disable-lib-asym with --enable-lib-asym as the algif_akcipher.c kernel interface is not likely to be added to the kernel anytime soon
- API Enhancement: add KPP API which is not compiled by default, use --enable-lib-kpp (the algif_kpp.c kernel interface is not likely to be added to the Linux kernel any time soon)
- Test Enhancement: Add KPP tests
- Enhancement: Re-enable AIO support for symmetric and AEAD ciphers down to Linux kernels 4.1 and 4.7, respectively. This is due to integrating a fix against a kernel crash when using AIO.
- Fix: simply KDF code base
- API Enhancement: add message digest convenience functions kcapi_md_*sha*
- API Enhancement: add cipher convenience functions kcapi_cipher_*_aes_*
- API Enhancement: add rng convenience function kcapi_rng_get_bytes
- API Change: remove kcapi_aead_getdata, use kcapi_aead_getdata_input and kcapi_aead_getdata_output instead
- API Change: remove kcapi_aead_outbuflen, use kcapi_aead_outbuflen_enc and kcapi_aead_outbuflen_dec instead
- updated to 0.14.0:
- AIO: fix tracking of completed IOCBs
- speed-test: fix AEAD handling
- speed-test: fix time calculation
- compiler now warns a user of deprecated API calls
- AIO: handle kernel errors for algif_skcipher gracefully
- AIO: using multiple IOCB if algif_aead interface supports it
- ASYM: add PKCS1 tests
- AIO: add ASYM AIO support
- AIO: fix AEAD AIO fallback
- AIO: add AIO fallback testing
- replace enforcement of symmetric cipher limits with a log message only (the underlying kernel implementations should catch any errors)
- add fuzzing tests
- use autotools build system as provided by Georges Savoundararadj with additional considerations from Marcin Nowakowski (thanks a lot)
- ALG_MAX_PAGES restriction is gone with current AF_ALG interface
- add HKDF (RFC5869)
- add apps/kcapi-rng
- add support for multiple accepts where the caller maintains the opfd
- fix memleak in error case in PBKDF
- add multithreaded symmetric cipher tests
- enable full AIO support for kernels 4.13 and higher (fallback AIO implementation using synchronous support for earlier kernels) -- this is due to the broken AIO support for earlier kernels
- Add tests for the AAD copy operation to be supported for kernel 4.13
- dropped libkcapi-use-external-fipshmac.patch (done differently in upstream)
- dropped reproduciblesort.patch (done differently upstream)
- dropped reproducibledate.patch: merged upstream
- libkcapi.keyring imported
-------------------------------------------------------------------
Thu Dec 5 10:10:41 UTC 2019 - Martin Liška <mliska@suse.cz>
- Use %make_build and respect %optflags.
-------------------------------------------------------------------
Fri Sep 27 16:40:49 UTC 2019 - Stefan Brüns <stefan.bruens@rwth-aachen.de>
- Remove docbook-utils BuildRequires, xmlto is sufficient
- Spec file cleanup, use license macro, drop defattr, drop BuildRoot
-------------------------------------------------------------------
Wed Jul 12 14:51:26 UTC 2017 - meissner@suse.com
- Change the signing to use openssl sha256/sha512 directly, to
avoid fipscheck / hmaccalc.
-------------------------------------------------------------------
Sat Jul 8 14:04:41 UTC 2017 - bwiedemann@suse.com
- Add reproduciblesort.patch to always link .o files in the same order and
- Add reproducibledate.patch to not add current time to man-pages to fix build-compare
-------------------------------------------------------------------
Thu Jun 29 08:13:54 UTC 2017 - meissner@suse.com
- libkcapi-use-external-fipshmac.patch: use external fipshmac,
our chroots / vm builds do not necessarily have the right kernel.
-------------------------------------------------------------------
Wed Jun 28 08:03:30 UTC 2017 - jengelh@inai.de
- Compact descriptions a bit
- Remove libkcapi provide/requires
- Use %_libdir throughout and avoid /lib
-------------------------------------------------------------------
Thu Dec 22 14:03:43 UTC 2016 - abergmann@suse.com
- Initial release 0.13.0.
A library and tools to access the kernel crypto api.
FATE#323554 bsc#1045948