fix(octet): prevent o_free from freeing Lua-allocated userdata struct - #1195
Merged
Conversation
o_free() called zfree(t) on the octet struct after ref reached 0. For userdata-backed octets (from o_new via lua_newuserdata), the struct is Lua-allocated, not zmalloc'd. Under pcall, incremental GC fires between o_arg and o_free, causing ref to reach 0 in o_free which then corrupts the Lua heap via wrong-allocator free. Changes: - o_alloc: explicitly initialize ref=0 (was relying on zeroed malloc) - o_free: only zfree(t) when ref < 0 (heap-allocated, o_alloc path); userdata-backed octets (ref==0 after GC+o_free) only free val. Set val=NULL after freeing for defence in depth. - o_destroy: set val=NULL after GC finalizer frees it
…ncrypt path The HEDLEY_ASSUME(h != NULL) tells the compiler it can elide null checks, but the immediately following SAFE_GOTO(h, ...) is precisely a null check. The decrypt path (line 697) never had this assume; this aligns the encrypt path to match.
Every expansion of these macros is an error path. HEDLEY_UNLIKELY tells the compiler to lay out the happy path contiguously and move error handlers out of line. Affects thousands of call sites across all native modules. Zero behavioral change.
Every *_new function either returns a valid pointer or errors out. HEDLEY_RETURNS_NON_NULL on the declaration propagates that knowledge to every caller, letting the compiler eliminate redundant null checks after SAFE guards. Affects octet, big, ecp, ecp2, secp, hash, fp12, float, and time constructors.
o_val and o_len are stateless field accessors with no side effects. Annotating them as HEDLEY_PURE lets the compiler CSE repeated calls in hot loops, particularly o_len in ZK witness generation paths.
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
1 participant
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.
o_free() called zfree(t) on the octet struct after ref reached 0. For userdata-backed octets (from o_new via lua_newuserdata), the struct is Lua-allocated, not zmalloc'd. Under pcall, incremental GC fires between o_arg and o_free, causing ref to reach 0 in o_free which then corrupts the Lua heap via wrong-allocator free.
Changes: