bip39: scrub vector registers after mnemonic decoding#534
Open
odudex wants to merge 1 commit into
Open
Conversation
test_clear finds plaintext secret material ~11KB deep on the stack after bip39_mnemonic_validate returns. The residue comes from XMM registers xsaved by the dynamic linker on subsequent PLT calls, not from a source-level local buffer. Adding a zeroed stack scratch in bip39_mnemonic_to_bytes forces glibc memset to issue SSE/AVX zero stores that clobber the vector registers before the function returns, so later xsave operations no longer spill secret bytes.
Contributor
|
OK, I'm looking at this and having some trouble verifying it. Running locally, I see the same leak after I added Did this change resolve test_clear failing on your machine? |
|
yes it do on 2 different machines: |
Contributor
|
Hmm, what is your full build line (configure args and initial make)? |
Details |
Contributor
|
OK so there are a few things going on here:
Note that this issue disappears entirely if static linking, only dynamic linking is affected. I've addressed (1), I suspect for the remainder the answer is probably to replace memcpy/cmp/set with unoptimizable variants that prevent the "optimized" libc versions from being called, and therefore don't require clearing the registers at the end :( |
Contributor
|
@pythcoiner @odudex Can you please try my attempt at a fix https://github.com/ElementsProject/libwally-core/tree/multi_updates and let me know if this allows test_clear to pass for you? Thanks |
Author
I can confirm all tests, python and C, including test_clear are green for me in branch multi_updates @ 78499f0 |
|
@jgriffiths I can confirm |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
3 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
During the work proposed on #533 I separated this out of scope, yet likely relevant commit from @pythcoiner, so it can be evaluated individually:
"test_clear finds plaintext secret material ~11KB deep on the stack after bip39_mnemonic_validate returns. The residue comes from XMM registers xsaved by the dynamic linker on subsequent PLT calls, not from a source-level local buffer. Adding a zeroed stack scratch in bip39_mnemonic_to_bytes forces glibc memset to issue SSE/AVX zero stores that clobber the vector registers before the function returns, so later xsave operations no longer spill secret bytes."