Hi all, Ok, as I mentioned before, I am happy to help here, but as I have a no cycles to dedicate, let's try to efficiently handle this patch first and then look at the GCM patch from Mamone. I don't think we would take the assembly for Poly1305, but fortunately no patch was provided and there is vectorization support in progress on the HACL* side, so I am sure that this code will be pretty fast... I had a brief look at the patches on Sunday and at first sight, these changes look ok, but I'd like to do proper review in Phabricator. @aoeuh, could you please split this work in two phabricator patches, one for blapi, one for chacha20? (you can push two commits with `moz-phab submit` and it will separate in two stacked phabricator patches). @George, once aoeuh is done, can you please make sure to have at least one or two person from IBM (or someone trusted) to: - have another look at the code to make sure it is correct - make sure this patch builds both with gyp (`build.sh`) and Makefiles Also, I'd like to make sure that some minimal help will be available from IBM if we hit issues, we won't have support for these machines in our CI, so we won't be able to test this. If we have reports that the code breaks something, for some reason, and there is no support, the only available option will be to revert that patch... :/ I am happy to discuss how we can see IBM engage more with us on supporting their platform in the long run, after we handle these changes. When we are done with the changes, I'll push the changes to our CI to see if everything passes, it might require a bit more work but if not I'll land the changes to NSS. Does that work for you all? Heads up to you all and to @glaubitz who is handling separate changes there: https://bugzilla.mozilla.org/show_bug.cgi?id=1687164 (these seem pretty much disconnected for now, but we might collide at some point)
Bug 1613235 Comment 32 Edit History
Note: The actual edited comment in the bug view page will always show the original commenter’s name and original timestamp.
Hi all, Ok, as I mentioned before, I am happy to help here, but as I have a no cycles to dedicate, let's try to efficiently handle this patch first and then look at the GCM patch from Mamone. I don't think we would take the assembly for Poly1305, but fortunately no patch was provided and there is vectorization support in progress on the HACL* side, so I am sure that this code will be pretty fast... I had a brief look at the patches on Sunday and at first sight, these changes look ok, but I'd like to do proper review in Phabricator. @aoeuh, could you please split this work in two phabricator patches, one for blapi, one for chacha20? (you can push two commits with `moz-phab submit` and it will separate in two stacked phabricator patches). @George, once aoeuh is done, can you please make sure to have at least one or two person from IBM (or someone trusted) to: - have another look at the code to make sure it is correct - make sure this patch builds both with gyp (`build.sh`) and Makefiles on IBM platforms Also, I'd like to make sure that some minimal help will be available from IBM if we hit issues, we won't have support for these machines in our CI, so we won't be able to test this. If we have reports that the code breaks something, for some reason, and there is no support, the only available option will be to revert that patch... :/ I am happy to discuss how we can see IBM engage more with us on supporting their platform in the long run, after we handle these changes. When we are done with the changes, I'll push the changes to our CI to see if everything passes, it might require a bit more work but if not I'll land the changes to NSS. Does that work for you all? Heads up to you all and to @glaubitz who is handling separate changes there: https://bugzilla.mozilla.org/show_bug.cgi?id=1687164 (these seem pretty much disconnected for now, but we might collide at some point)