Tested with: cryptocheck -d ossl0 -a all -z on amd64
Details
Diff Detail
- Repository
- rG FreeBSD src repository
- Lint
Lint Skipped - Unit
Tests Skipped - Build Status
Buildable 53277 Build 50168: arc lint + arc unit
Event Timeline
This appears to revert commit 454c425dbed1d341025c71a8645011a980e20ccd.
The underlying problem has since been fixed upstream: https://github.com/openssl/openssl/commit/89661d41e44090ab244652418e449eb1ceec6128
Could we perhaps cherry-pick that change into our tree?
I remembered that I had clobbered this when I woke up this morning, I will reapply it as a fixup so this commit doesn't undo it. I do think cherry-picking that change so we don't go regress in the future might be a good idea.
Ping @markj. Are you ok with this now? I could maybe reorder these commits so that I commit all the Makefile.asm fixes plus the cherry-pick of the OpenSSL fix first, then this regen commit, and lastly the commit to switch libcrypto over to the sys versions.
Yes, this looks ok. D41304 will need an update because a symbol name changed, but that's easy.
I think it'd make sense to regenerate asm files after the makefile changes, but so long as the output doesn't change I think it doesn't matter too much?
I could maybe reorder these commits so that I commit all the Makefile.asm fixes plus the cherry-pick of the OpenSSL fix first, then this regen commit, and lastly the commit to switch libcrypto over to the sys versions.
That seems like the logical ordering to me