Skip to content

gh-155811: Add a seqcount to gc_stats to prevent torn reads #155828

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
pablogsal merged 4 commits into from
Oct 5, 2026
Merged

gh-155811: Add a seqcount to gc_stats to prevent torn reads #155828

Show file tree
Hide file tree
Changes from 1 commit
Commits
Show all changes
4 commits
Select commit Hold shift + click to select a range
1291568
update_seq
maurycy Aug 15, 2026
0cb2510
no need for XCHGL, MOVL is enough?
maurycy Aug 15, 2026
424b95f
Merge branch 'main' into gc-mon-seq
maurycy Oct 4, 2026
577305e
gh-155811: Retry an inconsistent GC snapshot once
pablogsal Oct 5, 2026
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
Prev Previous commit
Next Next commit
no need for XCHGL, MOVL is enough?
  • Loading branch information
@maurycy
maurycy committed Aug 15, 2026
commit 0cb2510028c2c87c271ceb2efecfcecfa1986c80
7 changes: 4 additions & 3 deletions Python/gc.c
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1400,9 +1400,10 @@ static void
add_stats(GCState *gcstate, int gen, struct gc_generation_stats *stats)
{
struct gc_stats *generation_stats = gcstate->generation_stats;
uint32_t seq = _Py_atomic_load_uint32(&generation_stats->update_seq);
uint32_t seq = _Py_atomic_load_uint32_relaxed(&generation_stats->update_seq);
assert((seq & 1) == 0);
_Py_atomic_store_uint32(&generation_stats->update_seq, seq + 1);
/* Odd seq tells the reader that an update is in progress. */
_Py_atomic_store_uint32_relaxed(&generation_stats->update_seq, seq + 1);
_Py_atomic_fence_seq_cst();

struct gc_generation_stats *prev_stats = gc_get_prev_stats(gcstate, gen);
Expand All @@ -1419,7 +1420,7 @@ add_stats(GCState *gcstate, int gen, struct gc_generation_stats *stats)
cur_stats->duration += stats->duration;
cur_stats->heap_size = stats->heap_size;
cur_stats->ts_stop = stats->ts_stop;
_Py_atomic_store_uint32(&generation_stats->update_seq, seq + 2);
_Py_atomic_store_uint32_release(&generation_stats->update_seq, seq + 2);
}

/* This is the main function. Read this to understand how the
Expand Down
7 changes: 4 additions & 3 deletions Python/gc_free_threading.c
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -2284,9 +2284,10 @@ gc_collect_main(PyThreadState *tstate, int generation, _PyGC_Reason reason)
/* Update stats. */
PyMutex_Lock(&gcstate->stats_mutex);
struct gc_stats *generation_stats = gcstate->generation_stats;
uint32_t seq = _Py_atomic_load_uint32(&generation_stats->update_seq);
uint32_t seq = _Py_atomic_load_uint32_relaxed(&generation_stats->update_seq);
assert((seq & 1) == 0);
_Py_atomic_store_uint32(&generation_stats->update_seq, seq + 1);
/* Odd seq tells the reader that an update is in progress. */
_Py_atomic_store_uint32_relaxed(&generation_stats->update_seq, seq + 1);
_Py_atomic_fence_seq_cst();
struct gc_generation_stats *stats = get_stats(gcstate, generation);
stats->ts_start = start;
Expand All @@ -2296,7 +2297,7 @@ gc_collect_main(PyThreadState *tstate, int generation, _PyGC_Reason reason)
stats->uncollectable += n;
stats->duration += duration;
stats->candidates += state.candidates;
_Py_atomic_store_uint32(&generation_stats->update_seq, seq + 2);
_Py_atomic_store_uint32_release(&generation_stats->update_seq, seq + 2);
PyMutex_Unlock(&gcstate->stats_mutex);

GC_STAT_ADD(generation, objects_collected, m);
Expand Down
Loading