Fix setjmp/longjmp table matching to survive raw jmp_buf save/restore#1297
Open
qianxichen233 wants to merge 2 commits into
Open
Fix setjmp/longjmp table matching to survive raw jmp_buf save/restore#1297qianxichen233 wants to merge 2 commits into
qianxichen233 wants to merge 2 commits into
Conversation
buf[0] previously stored the jmp_buf's own fixed address as the key testSetjmp matches against. For a jmp_buf that lives in a long-lived buffer and gets saved/restored as raw bytes by an unwind-protect-style helper (memcpy(&saved, &buf, sizeof(buf)) and back), the address never changes, so every registration on that buffer was indistinguishable. A caller that restores an older snapshot before re-throwing therefore kept matching its own already-fired registration, causing an unbounded catch/rethrow loop that silently exhausted the null-collected GC heap. Replace the address-based key with a fresh per-call monotonic token stored in buf[0], and stop clearing matched table entries (which broke legitimate repeated longjmps to a still-live buffer with no intervening setjmp). Tokens make raw byte restoration meaningful again: an outer, still-live registration's token correctly falls through the inner frame's table and resolves in the outer frame's own table, matching native register/stack snapshot restoration semantics.
Member
|
Can you trigger another ci run? |
Contributor
End-to-End Test ReportTest Previewgrate harnessGrate Test Report
Cases
static harnessTest ReportDeterministic TestsSummary
Test Results by Category
Fail TestsSummary
wasm harnessTest ReportDeterministic TestsSummary
Test Results by Category
Fail TestsSummary
Test Results by Category
C++ harnessSummary
Cases
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Yaxuan-w
reviewed
Jul 6, 2026
Yaxuan-w
left a comment
Yaxuan-w
left a comment
Member
There was a problem hiding this comment.
Is that possible to have a unit test case for this?
Comment on lines
+75
to
+76
| int token = ++g_setjmp_token; | ||
| if (token == 0) token = ++g_setjmp_token; /* skip the 0 sentinel on wraparound */ |
Member
There was a problem hiding this comment.
In C, signed integer overflow is UB, so the program has already entered UB before we get to the if (token == 0) sentinel check. Maybe we can consider uint32_t in this case?
Contributor
Author
There was a problem hiding this comment.
theoritically this should never happen, gc heap overflow would happen before token overflow, as roughly ten million setjmp invocation would exhaust gc heap with even 4GB size (maximum possible size), an overflow of token would take 2 billion invocation
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
2 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.
buf[0] previously stored the jmp_buf's own fixed address as the key testSetjmp matches against. For a jmp_buf that lives in a long-lived buffer and gets saved/restored as raw bytes by an unwind-protect-style helper (memcpy(&saved, &buf, sizeof(buf)) and back), the address never changes, so every registration on that buffer was indistinguishable. A caller that restores an older snapshot before re-throwing therefore kept matching its own already-fired registration, causing an unbounded catch/rethrow loop that silently exhausted the null-collected GC heap.
Replace the address-based key with a fresh per-call monotonic token stored in buf[0], and stop clearing matched table entries (which broke legitimate repeated longjmps to a still-live buffer with no intervening setjmp). Tokens make raw byte restoration meaningful again: an outer, still-live registration's token correctly falls through the inner frame's table and resolves in the outer frame's own table, matching native register/stack snapshot restoration semantics.