From 0fcb9b7f2ede22f8c96b2a0e15d3a5a27d7d41ad Mon Sep 17 00:00:00 2001
From: Etienne Perot <eperot@google.com>
Date: Wed, 11 Sep 2024 15:13:53 -0700
Subject: [PATCH] Disable GPU sniffer on all non-smoke GPU tests.

The presubmit pipeline does not exercise these, so commit
a689c11a76539af18c432f942494c4a43f3857b7 broke them.

This change disables the sniffer on all the non-smoke tests to unbreak
the release pipeline. I will then send another change to re-enable them
on tests where the sniffer works fine after manual testing.

Updates #10885

PiperOrigin-RevId: 673555026
---
 test/gpu/cuda_test.go                       | 4 +++-
 test/gpu/nccl_test.go                       | 4 +++-
 test/gpu/ollama/ollama.go                   | 2 +-
 test/gpu/pytorch_test.go                    | 4 +++-
 test/gpu/sr_test.go                         | 4 +++-
 test/gpu/stablediffusion/stablediffusion.go | 2 +-
 test/gpu/vllm/vllm_test.go                  | 4 +++-
 7 files changed, 17 insertions(+), 7 deletions(-)

diff --git a/test/gpu/cuda_test.go b/test/gpu/cuda_test.go
index 9762f5527..2fb25f431 100644
--- a/test/gpu/cuda_test.go
+++ b/test/gpu/cuda_test.go
@@ -396,7 +396,9 @@ func (*FullyCompatible) IsExpectedFailure(ctx context.Context, env *TestEnvironm
 
 // getContainerOpts returns the container run options to run CUDA tests.
 func getContainerOpts() (dockerutil.RunOpts, error) {
-	opts, err := dockerutil.GPURunOpts(dockerutil.SniffGPUOpts{AllowIncompatibleIoctl: true})
+	opts, err := dockerutil.GPURunOpts(dockerutil.SniffGPUOpts{
+		DisableSnifferReason: "TODO(gvisor.dev/issue/10885): Verify that this test works",
+	})
 	if err != nil {
 		return dockerutil.RunOpts{}, fmt.Errorf("failed to get GPU run options: %w", err)
 	}
diff --git a/test/gpu/nccl_test.go b/test/gpu/nccl_test.go
index 74c7772b9..b82282711 100644
--- a/test/gpu/nccl_test.go
+++ b/test/gpu/nccl_test.go
@@ -27,7 +27,9 @@ import (
 func runNCCL(ctx context.Context, t *testing.T, testName string) {
 	t.Helper()
 	c := dockerutil.MakeContainer(ctx, t)
-	opts, err := dockerutil.GPURunOpts(dockerutil.SniffGPUOpts{AllowIncompatibleIoctl: true})
+	opts, err := dockerutil.GPURunOpts(dockerutil.SniffGPUOpts{
+		DisableSnifferReason: "TODO(gvisor.dev/issue/10885): Verify that this test works",
+	})
 	if err != nil {
 		t.Fatalf("Failed to get GPU run options: %v", err)
 	}
diff --git a/test/gpu/ollama/ollama.go b/test/gpu/ollama/ollama.go
index 172cf278e..7443d8682 100644
--- a/test/gpu/ollama/ollama.go
+++ b/test/gpu/ollama/ollama.go
@@ -150,7 +150,7 @@ type dockerServer struct {
 // in a local Docker container.
 func NewDocker(ctx context.Context, cont *dockerutil.Container, logger testutil.Logger) (*Ollama, error) {
 	opts, err := dockerutil.GPURunOpts(dockerutil.SniffGPUOpts{
-		AllowIncompatibleIoctl: true,
+		DisableSnifferReason: "TODO(gvisor.dev/issue/10885): Verify that this test works",
 	})
 	if err != nil {
 		return nil, fmt.Errorf("failed to get GPU run options: %w", err)
diff --git a/test/gpu/pytorch_test.go b/test/gpu/pytorch_test.go
index 15b5ffa3a..05d74a53a 100644
--- a/test/gpu/pytorch_test.go
+++ b/test/gpu/pytorch_test.go
@@ -26,7 +26,9 @@ import (
 func runPytorch(ctx context.Context, t *testing.T, scriptPath string, args ...string) {
 	t.Helper()
 	c := dockerutil.MakeContainer(ctx, t)
-	opts, err := dockerutil.GPURunOpts(dockerutil.SniffGPUOpts{AllowIncompatibleIoctl: true})
+	opts, err := dockerutil.GPURunOpts(dockerutil.SniffGPUOpts{
+		DisableSnifferReason: "TODO(gvisor.dev/issue/10885): Verify that this test works",
+	})
 	if err != nil {
 		t.Fatalf("Failed to get GPU run options: %v", err)
 	}
diff --git a/test/gpu/sr_test.go b/test/gpu/sr_test.go
index 5f9db3dd2..e8ca6cfc8 100644
--- a/test/gpu/sr_test.go
+++ b/test/gpu/sr_test.go
@@ -34,7 +34,9 @@ func TestGPUCheckpointRestore(t *testing.T) {
 	c := dockerutil.MakeContainer(ctx, t)
 	defer c.CleanUp(ctx)
 
-	opts, err := dockerutil.GPURunOpts(dockerutil.SniffGPUOpts{AllowIncompatibleIoctl: true})
+	opts, err := dockerutil.GPURunOpts(dockerutil.SniffGPUOpts{
+		DisableSnifferReason: "TODO(gvisor.dev/issue/10885): Verify that this test works",
+	})
 	if err != nil {
 		t.Fatalf("failed to get GPU run options: %v", err)
 	}
diff --git a/test/gpu/stablediffusion/stablediffusion.go b/test/gpu/stablediffusion/stablediffusion.go
index 511b00d53..44499221a 100644
--- a/test/gpu/stablediffusion/stablediffusion.go
+++ b/test/gpu/stablediffusion/stablediffusion.go
@@ -48,7 +48,7 @@ func (dr *dockerRunner) Run(ctx context.Context, image string, argv []string) ([
 	cont := dockerutil.MakeContainer(ctx, dr.logger)
 	defer cont.CleanUp(ctx)
 	opts, err := dockerutil.GPURunOpts(dockerutil.SniffGPUOpts{
-		AllowIncompatibleIoctl: true,
+		DisableSnifferReason: "TODO(gvisor.dev/issue/10885): Verify that this test works",
 	})
 	if err != nil {
 		return nil, fmt.Errorf("failed to get GPU run options: %w", err)
diff --git a/test/gpu/vllm/vllm_test.go b/test/gpu/vllm/vllm_test.go
index d4d55c138..2c84ce750 100644
--- a/test/gpu/vllm/vllm_test.go
+++ b/test/gpu/vllm/vllm_test.go
@@ -51,7 +51,9 @@ func doVLLMTest(b *testing.B) {
 	}
 
 	// Run vllm.
-	runOpts, err := dockerutil.GPURunOpts(dockerutil.SniffGPUOpts{AllowIncompatibleIoctl: true})
+	runOpts, err := dockerutil.GPURunOpts(dockerutil.SniffGPUOpts{
+		DisableSnifferReason: "TODO(gvisor.dev/issue/10885): Verify that this test works",
+	})
 	if err != nil {
 		b.Fatalf("failed to get GPU run options: %v", err)
 	}