releases/6.7.6/revert-drm-msm-gpu-push-gpu-lock-down-past-runpm.patch

From 917e9b7c2350e3e53162fcf5035e5f2d68e2cbed Mon Sep 17 00:00:00 2001
From: Rob Clark <robdclark@chromium.org>
Date: Tue, 9 Jan 2024 10:22:17 -0800
Subject: Revert "drm/msm/gpu: Push gpu lock down past runpm"

From: Rob Clark <robdclark@chromium.org>

commit 917e9b7c2350e3e53162fcf5035e5f2d68e2cbed upstream.

This reverts commit abe2023b4cea192ab266b351fd38dc9dbd846df0.

Changing the locking order means that scheduler/msm_job_run() can race
with the recovery kthread worker, with the result that the GPU gets an
extra runpm get when we are trying to power it off.  Leaving the GPU in
an unrecovered state.

I'll need to come up with a different scheme for appeasing lockdep.

Signed-off-by: Rob Clark <robdclark@chromium.org>
Patchwork: https://patchwork.freedesktop.org/patch/573835/
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
---
 drivers/gpu/drm/msm/msm_gpu.c        |   11 +++++------
 drivers/gpu/drm/msm/msm_ringbuffer.c |    7 +++++--
 2 files changed, 10 insertions(+), 8 deletions(-)

--- a/drivers/gpu/drm/msm/msm_gpu.c
+++ b/drivers/gpu/drm/msm/msm_gpu.c
@@ -749,12 +749,14 @@ void msm_gpu_submit(struct msm_gpu *gpu,
 	struct msm_ringbuffer *ring = submit->ring;
 	unsigned long flags;
 
-	pm_runtime_get_sync(&gpu->pdev->dev);
+	WARN_ON(!mutex_is_locked(&gpu->lock));
 
-	mutex_lock(&gpu->lock);
+	pm_runtime_get_sync(&gpu->pdev->dev);
 
 	msm_gpu_hw_init(gpu);
 
+	submit->seqno = submit->hw_fence->seqno;
+
 	update_sw_cntrs(gpu);
 
 	/*
@@ -779,11 +781,8 @@ void msm_gpu_submit(struct msm_gpu *gpu,
 	gpu->funcs->submit(gpu, submit);
 	gpu->cur_ctx_seqno = submit->queue->ctx->seqno;
 
-	hangcheck_timer_reset(gpu);
-
-	mutex_unlock(&gpu->lock);
-
 	pm_runtime_put(&gpu->pdev->dev);
+	hangcheck_timer_reset(gpu);
 }
 
 /*
--- a/drivers/gpu/drm/msm/msm_ringbuffer.c
+++ b/drivers/gpu/drm/msm/msm_ringbuffer.c
@@ -21,8 +21,6 @@ static struct dma_fence *msm_job_run(str
 
 	msm_fence_init(submit->hw_fence, fctx);
 
-	submit->seqno = submit->hw_fence->seqno;
-
 	mutex_lock(&priv->lru.lock);
 
 	for (i = 0; i < submit->nr_bos; i++) {
@@ -34,8 +32,13 @@ static struct dma_fence *msm_job_run(str
 
 	mutex_unlock(&priv->lru.lock);
 
+	/* TODO move submit path over to using a per-ring lock.. */
+	mutex_lock(&gpu->lock);
+
 	msm_gpu_submit(gpu, submit);
 
+	mutex_unlock(&gpu->lock);
+
 	return dma_fence_get(submit->hw_fence);
 }