archive/ruby
1
0
Fork 0
mirror of https://github.com/ruby/ruby.git synced 2025-08-15 13:39:04 +02:00
Code Issues Projects Releases Packages Wiki Activity Actions 4

YJIT: Use raw memory write to update pointers in code

Browse source 
Because we have set all code memory to writable before the reference
updating phase, we can use raw memory writes directly.
This commit is contained in:
Kunshan Wang 2025-07-19 16:22:46 +08:00 • committed by Alan Wu
parent fd492a45eb
commit 5ef20b3a27
1 changed files with 3 additions and 5 deletions
Download patch file Download diff file Expand all files Collapse all files

8
yjit/src/core.rs
View file

@ -2091,11 +2091,9 @@ pub extern "C" fn rb_yjit_iseq_update_references(iseq: IseqPtr) {
// Only write when the VALUE moves, to be copy-on-write friendly.
if new_addr != object {
for (byte_idx, &byte) in new_addr.as_u64().to_le_bytes().iter().enumerate() {
let byte_code_ptr = value_code_ptr.add_bytes(byte_idx);
cb.write_mem(byte_code_ptr, byte)
.expect("patching existing code should be within bounds");
}
// SAFETY: Since we already set code memory writable before the compacting phase,
// we can use raw memory accesses directly.
unsafe { value_ptr.write_unaligned(new_addr); }
}
}
}