Merged to master.
- Queries
- All Stories
- Search
- Advanced Search
- Transactions
- Transaction Logs
Feed Advanced Search
Advanced Search
Advanced Search
Jan 20 2021
Jan 20 2021
jukivili closed T5197: libgcrypt: s390x/zSeries AES acceleration, a subtask of T5196: libgcrypt: s390x/zSeries performance improvements, as Resolved.
jukivili closed T5198: libgcrypt: s390x/zSeries SHA256/SHA512 acceleration, a subtask of T5196: libgcrypt: s390x/zSeries performance improvements, as Resolved.
Merged to master.
jukivili closed T5199: libgcrypt: s390x/zSeries GHASH/AES-GCM acceleration, a subtask of T5196: libgcrypt: s390x/zSeries performance improvements, as Resolved.
jukivili closed T5200: libgcrypt: s390x/zSeries SHA3 acceleration, a subtask of T5196: libgcrypt: s390x/zSeries performance improvements, as Resolved.
Merged to master.
jukivili closed T5201: libgcrypt: s390x/zSeries 128-bit vector implementation of ChaCha20, a subtask of T5196: libgcrypt: s390x/zSeries performance improvements, as Resolved.
jukivili closed T5201: libgcrypt: s390x/zSeries 128-bit vector implementation of ChaCha20 as Resolved.
Merged to master.
jukivili closed T5202: libgcrypt: s390x/zSeries implementation of Poly1305 / ChaCha20-Poly1305 AEAD, a subtask of T5196: libgcrypt: s390x/zSeries performance improvements, as Resolved.
jukivili closed T5202: libgcrypt: s390x/zSeries implementation of Poly1305 / ChaCha20-Poly1305 AEAD as Resolved.
Merged to master.
jukivili closed T5203: libgcrypt: s390x/zSeries SHA1 acceleration, a subtask of T5196: libgcrypt: s390x/zSeries performance improvements, as Resolved.
Merged to master.
In T5257#141913, @jukivili wrote:You probably define CFLAGS=-m32 in your installation script for 32-bit build. I'd try with CFLAGS="-O2 -m32".
0001-Split-inline-assembly-blocks-with-many-memory-operan.patch11 KBDownload
Thanks for report. I reproduced this by building i386 with optimizations disabled "-O0" (gcc 10). With normal optimization level such as "-O2", the issue does not appear.
Sure. Thanks for testing. The problem with new versions is that ppl don't like to test release candidates and thus we need do real releases and wait for the outfall. ;-)
• werner triaged T5256: libgcrypt, convert Intel syntax x86_64 assembly files to AT&T syntax as Normal priority.
jukivili created T5256: libgcrypt, convert Intel syntax x86_64 assembly files to AT&T syntax in the S1 Public space.
jukivili added a comment to T5255: libgcrypt: build "error: invalid operand for instruction" when compiling with Clang & LTO.
Breakage appears to happen in configure.ac. When building with clang without LTO following check gives "no":
Fixed by jukvilli’s patch.
__outer
That works, thanks. So does that become part of the next release?
__outer
telans added a comment to T5255: libgcrypt: build "error: invalid operand for instruction" when compiling with Clang & LTO.
Both are affected. I updated to reflect that I tested the newer version
• werner triaged T5255: libgcrypt: build "error: invalid operand for instruction" when compiling with Clang & LTO as Normal priority.
FWIW, after the release I had some time and after some trouble with my Pi4B I ran into the same problem.
Jan 19 2021
Jan 19 2021
Confirmed working after applying your patch!
Thanks for you report.
• werner closed T4294: Release Libgcrypt 1.9.0, a subtask of T4068: libgcrypt 1.8.3 make check errors, as Resolved.
• werner closed T4294: Release Libgcrypt 1.9.0, a subtask of T4243: Test failure in libgcrypt-1.8.4, as Resolved.
• werner closed T4294: Release Libgcrypt 1.9.0, a subtask of T5243: libgcrypt "check if fips_is_operational and error return if not" patch for FIPS 140, as Resolved.
• werner closed T4294: Release Libgcrypt 1.9.0, a subtask of T4288: Add getrandom support for the BSDs, as Resolved.
• werner closed T4294: Release Libgcrypt 1.9.0, a subtask of T4869: constant-time mpi_invm, as Resolved.
• werner closed T4294: Release Libgcrypt 1.9.0, a subtask of T5182: libgcrypt self tests for FIPS 140, as Resolved.
• werner closed T4294: Release Libgcrypt 1.9.0, a subtask of T4274: Fail selftests when checksum file is missing in FIPS mode only, as Resolved.
• werner closed T4294: Release Libgcrypt 1.9.0, a subtask of T4293: Add dedicated X25519 function to Libcgrypt , as Resolved.
• werner removed a parent task for T4294: Release Libgcrypt 1.9.0: T5244: libgcrypt: Restrict MD5 use.
We plan this for 1.10 but it may also go into one of the next 1.9.x releases
Docs done.
Jan 18 2021
Jan 18 2021
• werner moved T5182: libgcrypt self tests for FIPS 140 from For 1.9 to For 1.10 on the libgcrypt board.
• werner moved T5195: Incorrect HWCAP2 check for AArch32 from For 1.9 to For 1.8 on the libgcrypt board.
• werner moved T4951: Support point compression in Libgcrypt from For 1.9 to For 1.10 on the libgcrypt board.
• werner removed a parent task for T4294: Release Libgcrypt 1.9.0: T1303: Please support GCRYSEXP_FMT_BASE64.
• werner removed a subtask for T1303: Please support GCRYSEXP_FMT_BASE64: T4294: Release Libgcrypt 1.9.0.
I am not sure. MD5 is still important for some applications, say CRAM-MD5. IIRC, back in 2008 we dis-allowed RMD160 and added separate RMD160 code directly to gpg to fulfill FIPS requirements.
• werner added a comment to T5243: libgcrypt "check if fips_is_operational and error return if not" patch for FIPS 140.
Okay for 1.9.
Jan 15 2021
Jan 15 2021
Note that even after rCce1cbe16992a: Disable non-allowed algorithms in FIPS mode, gcry_md_open won't return an error with disabled algo.
• gniibe added a parent task for T4294: Release Libgcrypt 1.9.0: T5244: libgcrypt: Restrict MD5 use.
• gniibe added a subtask for T5182: libgcrypt self tests for FIPS 140: T4294: Release Libgcrypt 1.9.0.
• gniibe added a parent task for T4294: Release Libgcrypt 1.9.0: T5182: libgcrypt self tests for FIPS 140.
• gniibe removed a parent task for T5182: libgcrypt self tests for FIPS 140: T4294: Release Libgcrypt 1.9.0.
• gniibe removed a subtask for T4294: Release Libgcrypt 1.9.0: T5182: libgcrypt self tests for FIPS 140.
• gniibe renamed T5243: libgcrypt "check if fips_is_operational and error return if not" patch for FIPS 140 from libgcrypt "check if fips_is_operational and error return if not fix" for FIPS 140 to libgcrypt "check if fips_is_operational and error return if not" patch for FIPS 140.
Jan 12 2021
Jan 12 2021
Note: The commit in master (1.9) is rCe0898d0628789414
and in 1.8 it is rC03e6d6597198ee
The commit which fixes this is rC761a1a0d30
Jan 11 2021
Jan 11 2021
• aheinecke moved T4294: Release Libgcrypt 1.9.0 from Restricted Project Column to Restricted Project Column on the Restricted Project board.
Jan 8 2021
Jan 8 2021
• werner added a comment to T4964: ecc: Discrepancy of handling MPI for the interpretation of signed and unsigned.
I agree to the sexp change - but it should not be backported to 1.8
• gniibe added a comment to T4964: ecc: Discrepancy of handling MPI for the interpretation of signed and unsigned.
For printing SEXP, it would be good to have this change:
• gniibe added a comment to T4956: agent: Discrepancy of handling MPI for the interpretation of signed and unsigned.
rG47c1c329ed82: agent,ecc: Use of opaque MPI for ECC, fixup 'd'. does the fixup when reading keys.
• gniibe added a comment to T4964: ecc: Discrepancy of handling MPI for the interpretation of signed and unsigned.
I describe about rC6f8b1d4cb798: ecc: Consistently handle parameters as unsigned value..
• gniibe updated the task description for T4964: ecc: Discrepancy of handling MPI for the interpretation of signed and unsigned.
Reading compressed point (in keys) is supported (except for NIST P-224). When curve point is represented in compressed format, it is correctly interpreted now. So, for example, I think that with 1.9.0, gpgsm can handle certificate which uses compressed format in its curve point representation.
Jan 7 2021
Jan 7 2021
Yes, bug is also in 1.8 branch.