From a3a3339c97329634a42aaa5e61c2b028bd861ab1 Mon Sep 17 00:00:00 2001
From: Benedykt Bela <benedykt.bela@intel.com>
Date: Tue, 12 May 2026 15:59:37 +0000
Subject: [PATCH] Fix flaky reentrant backward test by passing explicit grad
 for non-scalar output (#179704)

Fixes #179703

More detailed description of the problem is in the issue #179703

## What this PR changes

In `test/test_autograd.py`, `_test_reentrant_parent_error_on_cpu` [(THIS LINE)](https://github.com/pytorch/pytorch/blob/3b39608061aac9da750ad3954197400b191286c3/test/test_autograd.py#L12574) defines a reentrant autograd function whose backward currently does:

```python
reentrant_root.backward()
```

This PR changes it to:

```python
reentrant_root.backward(grad)
```

## Why this solution

`reentrant_root` is non-scalar (`[3, 3]`). For non-scalar outputs, `.backward()` requires an explicit gradient seed. Calling `.backward()` without that seed can raise:

`RuntimeError: grad can be implicitly created only for scalar outputs`

## Effect on test behavior

Before this change, the test had two competing error sources:

1. Intended error: `"Simulate error"` from `SimulateBackwardError`
2. Accidental error: non-scalar `.backward()` without explicit grad

Because execution order is nondeterministic, CI could intermittently fail when (2) happened first.

After this change, the accidental error source is removed. The test now consistently validates only the intended behavior.

Pull Request resolved: https://github.com/pytorch/pytorch/pull/179704
Approved by: https://github.com/jeffdaily

(cherry picked from commit 24cdfbdd4ce3d3b080e64e6a8b03a8ecda547bd0)
---
 test/test_autograd.py | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/test/test_autograd.py b/test/test_autograd.py
index dbd1454ff7459..76fe65ea86b17 100644
--- a/test/test_autograd.py
+++ b/test/test_autograd.py
@@ -11755,7 +11755,7 @@ def forward(ctx, inp):
             @staticmethod
             def backward(ctx, grad):
                 # Reentrant backward in child will take much longer.
-                reentrant_root.backward()
+                reentrant_root.backward(grad)
                 return grad
 
         # Parent gpu graph.