[PATCH 02/11] drm/xe: Introduce intel_runtime_pm_get_noresume at compat-i915-headers for display

[PATCH 02/11] drm/xe: Introduce intel_runtime_pm_get_noresume at compat-i915-headers for display

[Rodrigo Vivi]

The i915-display will start using the intel_runtime_pm_noresume.
So we need to add the compat header before it.

Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
---
 drivers/gpu/drm/xe/compat-i915-headers/i915_drv.h | 7 +++++++
 1 file changed, 7 insertions(+)

diff --git a/drivers/gpu/drm/xe/compat-i915-headers/i915_drv.h b/drivers/gpu/drm/xe/compat-i915-headers/i915_drv.h
index fef969112b1d..84fda436faab 100644
--- a/drivers/gpu/drm/xe/compat-i915-headers/i915_drv.h
+++ b/drivers/gpu/drm/xe/compat-i915-headers/i915_drv.h
@@ -176,6 +176,13 @@ static inline intel_wakeref_t intel_runtime_pm_get_if_in_use(struct xe_runtime_p
 	return xe_pm_runtime_get_if_in_use(xe);
 }
 
+static inline intel_wakeref_t intel_runtime_pm_get_noresume(struct xe_runtime_pm *pm)
+{
+	struct xe_device *xe = container_of(pm, struct xe_device, runtime_pm);
+	xe_pm_runtime_get_noresume(xe);
+	return true;
+}
+
 static inline void intel_runtime_pm_put_unchecked(struct xe_runtime_pm *pm)
 {
 	struct xe_device *xe = container_of(pm, struct xe_device, runtime_pm);
-- 
2.44.0

[PATCH 01/11] drm/xe: Introduce xe_pm_runtime_get_noresume for inner callers

[Rodrigo Vivi]

Let's ensure that we have an option for inner callers that will
raise WARN if device is not active and not protected by outer callers.

Make this also a void function forcing every caller to unconditionally
put the reference back afterwards.

This will be very important for cases where we want to hold the
reference before scheduling a work in a queue. Then the work job
will be responsible for putting it back.

While at this, already convert a case from mem_access_ongoing where
it is not checking for the reference and put it back, what would
cause the underflow.

v2: Fix identation.

Reviewed-by: Matthew Auld <matthew.auld@intel.com>
Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
---
 drivers/gpu/drm/xe/xe_exec_queue.c |  2 +-
 drivers/gpu/drm/xe/xe_pm.c         | 20 ++++++++++++++++++++
 drivers/gpu/drm/xe/xe_pm.h         |  1 +
 3 files changed, 22 insertions(+), 1 deletion(-)

diff --git a/drivers/gpu/drm/xe/xe_exec_queue.c b/drivers/gpu/drm/xe/xe_exec_queue.c
index 6a83bc57826a..f69a9c99329c 100644
--- a/drivers/gpu/drm/xe/xe_exec_queue.c
+++ b/drivers/gpu/drm/xe/xe_exec_queue.c
@@ -128,7 +128,7 @@ static int __xe_exec_queue_init(struct xe_exec_queue *q)
 	 * already grabbed the rpm ref outside any sensitive locks.
 	 */
 	if (!(q->flags & EXEC_QUEUE_FLAG_PERMANENT) && (q->flags & EXEC_QUEUE_FLAG_VM || !q->vm))
-		drm_WARN_ON(&xe->drm, !xe_device_mem_access_get_if_ongoing(xe));
+		xe_pm_runtime_get_noresume(xe);
 
 	return 0;
 
diff --git a/drivers/gpu/drm/xe/xe_pm.c b/drivers/gpu/drm/xe/xe_pm.c
index 9fbb6f6c598a..2e1362cf8deb 100644
--- a/drivers/gpu/drm/xe/xe_pm.c
+++ b/drivers/gpu/drm/xe/xe_pm.c
@@ -477,6 +477,26 @@ bool xe_pm_runtime_get_if_in_use(struct xe_device *xe)
 	return pm_runtime_get_if_in_use(xe->drm.dev) > 0;
 }
 
+/**
+ * xe_pm_runtime_get_noresume - Bump runtime PM usage counter without resuming
+ * @xe: xe device instance
+ *
+ * This function should be used in inner places where it is surely already
+ * protected by outer-bound callers of `xe_pm_runtime_get`.
+ * It will warn if not protected.
+ * The reference should be put back after this function regardless, since it
+ * will always bump the usage counter, regardless.
+ */
+void xe_pm_runtime_get_noresume(struct xe_device *xe)
+{
+	bool ref;
+
+	ref = xe_pm_runtime_get_if_in_use(xe);
+
+	if (drm_WARN(&xe->drm, !ref, "Missing outer runtime PM protection\n"))
+		pm_runtime_get_noresume(xe->drm.dev);
+}
+
 /**
  * xe_pm_runtime_resume_and_get - Resume, then get a runtime_pm ref if awake.
  * @xe: xe device instance
diff --git a/drivers/gpu/drm/xe/xe_pm.h b/drivers/gpu/drm/xe/xe_pm.h
index 0cb38ca244fe..119b630ad1d1 100644
--- a/drivers/gpu/drm/xe/xe_pm.h
+++ b/drivers/gpu/drm/xe/xe_pm.h
@@ -31,6 +31,7 @@ int xe_pm_runtime_get_ioctl(struct xe_device *xe);
 void xe_pm_runtime_put(struct xe_device *xe);
 int xe_pm_runtime_get_if_active(struct xe_device *xe);
 bool xe_pm_runtime_get_if_in_use(struct xe_device *xe);
+void xe_pm_runtime_get_noresume(struct xe_device *xe);
 bool xe_pm_runtime_resume_and_get(struct xe_device *xe);
 void xe_pm_assert_unbounded_bridge(struct xe_device *xe);
 int xe_pm_set_vram_threshold(struct xe_device *xe, u32 threshold);
-- 
2.44.0

[PATCH 02/11] drm/xe: Introduce intel_runtime_pm_get_noresume at compat-i915-headers for display

[Rodrigo Vivi]

The i915-display will start using the intel_runtime_pm_noresume.
So we need to add the compat header before it.

Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
---
 drivers/gpu/drm/xe/compat-i915-headers/i915_drv.h | 8 ++++++++
 1 file changed, 8 insertions(+)

diff --git a/drivers/gpu/drm/xe/compat-i915-headers/i915_drv.h b/drivers/gpu/drm/xe/compat-i915-headers/i915_drv.h
index fef969112b1d..ecaaef3df4bf 100644
--- a/drivers/gpu/drm/xe/compat-i915-headers/i915_drv.h
+++ b/drivers/gpu/drm/xe/compat-i915-headers/i915_drv.h
@@ -176,6 +176,14 @@ static inline intel_wakeref_t intel_runtime_pm_get_if_in_use(struct xe_runtime_p
 	return xe_pm_runtime_get_if_in_use(xe);
 }
 
+static inline intel_wakeref_t intel_runtime_pm_get_noresume(struct xe_runtime_pm *pm)
+{
+	struct xe_device *xe = container_of(pm, struct xe_device, runtime_pm);
+
+	xe_pm_runtime_get_noresume(xe);
+	return true;
+}
+
 static inline void intel_runtime_pm_put_unchecked(struct xe_runtime_pm *pm)
 {
 	struct xe_device *xe = container_of(pm, struct xe_device, runtime_pm);
-- 
2.44.0

[PATCH 03/11] drm/i915/display: convert inner wakeref get towards get_if_in_use

[Rodrigo Vivi]

This patch brings no functional change. Since at this point of
the code we are already asserting a wakeref was held, it means
that we are with runtime_pm 'in_use' and in practical terms we
are only bumping the pm_runtime usage counter and moving on.

However, xe driver has a lockdep annotation that warned us that
if a sync resume was actually called at this point, we could have
a deadlock because we are inside the power_domains->lock locked
area and the resume would call the irq_reset, which would also
try to get the power_domains->lock.

For this reason, let's convert this call to a safer option and
calm lockdep on.

v2: use _noresume variant instead of get_in_use (Ville, Imre)

Cc: Ville Syrjälä <ville.syrjala@linux.intel.com>
Acked-by: Imre Deak <imre.deak@intel.com>
Cc: Matthew Auld <matthew.auld@intel.com>
Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
---
 drivers/gpu/drm/i915/display/intel_display_power.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/gpu/drm/i915/display/intel_display_power.c b/drivers/gpu/drm/i915/display/intel_display_power.c
index 6fd4fa52253a..048943d0a881 100644
--- a/drivers/gpu/drm/i915/display/intel_display_power.c
+++ b/drivers/gpu/drm/i915/display/intel_display_power.c
@@ -646,7 +646,7 @@ release_async_put_domains(struct i915_power_domains *power_domains,
 	 * power well disabling.
 	 */
 	assert_rpm_raw_wakeref_held(rpm);
-	wakeref = intel_runtime_pm_get(rpm);
+	wakeref = intel_runtime_pm_get_noresume(rpm);
 
 	for_each_power_domain(domain, mask) {
 		/* Clear before put, so put's sanity check is happy. */
-- 
2.44.0

[PATCH 04/11] drm/xe: Move lockdep protection from mem_access to xe_pm_runtime

[Rodrigo Vivi]

The mem_access itself is not holding any lock, but attempting
to train lockdep with possible scarring locks happening during
runtime pm. We are going soon to kill the mem_access get and put
helpers in favor of direct xe_pm_runtime calls, so let's just
move this lock around to where it now belongs.

v2: s/lockdep_training/lockdep_prime (Matt Auld)

Reviewed-by: Matthew Auld <matthew.auld@intel.com>
Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
---
 drivers/gpu/drm/xe/xe_device.c | 23 -----------------
 drivers/gpu/drm/xe/xe_device.h |  4 ---
 drivers/gpu/drm/xe/xe_pm.c     | 45 ++++++++++++++++++++++++++++------
 3 files changed, 37 insertions(+), 35 deletions(-)

diff --git a/drivers/gpu/drm/xe/xe_device.c b/drivers/gpu/drm/xe/xe_device.c
index b0bfe75eb59f..82f686595b16 100644
--- a/drivers/gpu/drm/xe/xe_device.c
+++ b/drivers/gpu/drm/xe/xe_device.c
@@ -45,12 +45,6 @@
 #include "xe_vm.h"
 #include "xe_wait_user_fence.h"
 
-#ifdef CONFIG_LOCKDEP
-struct lockdep_map xe_device_mem_access_lockdep_map = {
-	.name = "xe_device_mem_access_lockdep_map"
-};
-#endif
-
 static int xe_file_open(struct drm_device *dev, struct drm_file *file)
 {
 	struct xe_device *xe = to_xe_device(dev);
@@ -702,23 +696,6 @@ void xe_device_mem_access_get(struct xe_device *xe)
 	if (xe_pm_read_callback_task(xe) == current)
 		return;
 
-	/*
-	 * Since the resume here is synchronous it can be quite easy to deadlock
-	 * if we are not careful. Also in practice it might be quite timing
-	 * sensitive to ever see the 0 -> 1 transition with the callers locks
-	 * held, so deadlocks might exist but are hard for lockdep to ever see.
-	 * With this in mind, help lockdep learn about the potentially scary
-	 * stuff that can happen inside the runtime_resume callback by acquiring
-	 * a dummy lock (it doesn't protect anything and gets compiled out on
-	 * non-debug builds).  Lockdep then only needs to see the
-	 * mem_access_lockdep_map -> runtime_resume callback once, and then can
-	 * hopefully validate all the (callers_locks) -> mem_access_lockdep_map.
-	 * For example if the (callers_locks) are ever grabbed in the
-	 * runtime_resume callback, lockdep should give us a nice splat.
-	 */
-	lock_map_acquire(&xe_device_mem_access_lockdep_map);
-	lock_map_release(&xe_device_mem_access_lockdep_map);
-
 	xe_pm_runtime_get(xe);
 	ref = atomic_inc_return(&xe->mem_access.ref);
 
diff --git a/drivers/gpu/drm/xe/xe_device.h b/drivers/gpu/drm/xe/xe_device.h
index 14be34d9f543..2653f53bee4e 100644
--- a/drivers/gpu/drm/xe/xe_device.h
+++ b/drivers/gpu/drm/xe/xe_device.h
@@ -16,10 +16,6 @@ struct xe_file;
 #include "xe_force_wake.h"
 #include "xe_macros.h"
 
-#ifdef CONFIG_LOCKDEP
-extern struct lockdep_map xe_device_mem_access_lockdep_map;
-#endif
-
 static inline struct xe_device *to_xe_device(const struct drm_device *dev)
 {
 	return container_of(dev, struct xe_device, drm);
diff --git a/drivers/gpu/drm/xe/xe_pm.c b/drivers/gpu/drm/xe/xe_pm.c
index 2e1362cf8deb..9d87a68ba6eb 100644
--- a/drivers/gpu/drm/xe/xe_pm.c
+++ b/drivers/gpu/drm/xe/xe_pm.c
@@ -68,6 +68,12 @@
  * management (RPS).
  */
 
+#ifdef CONFIG_LOCKDEP
+struct lockdep_map xe_pm_runtime_lockdep_map = {
+	.name = "xe_pm_runtime_lockdep_map"
+};
+#endif
+
 /**
  * xe_pm_suspend - Helper for System suspend, i.e. S0->S3 / S0->S2idle
  * @xe: xe device instance
@@ -297,11 +303,11 @@ int xe_pm_runtime_suspend(struct xe_device *xe)
 	xe_pm_write_callback_task(xe, current);
 
 	/*
-	 * The actual xe_device_mem_access_put() is always async underneath, so
+	 * The actual xe_pm_runtime_put() is always async underneath, so
 	 * exactly where that is called should makes no difference to us. However
 	 * we still need to be very careful with the locks that this callback
 	 * acquires and the locks that are acquired and held by any callers of
-	 * xe_device_mem_access_get(). We already have the matching annotation
+	 * xe_runtime_pm_get(). We already have the matching annotation
 	 * on that side, but we also need it here. For example lockdep should be
 	 * able to tell us if the following scenario is in theory possible:
 	 *
@@ -309,15 +315,15 @@ int xe_pm_runtime_suspend(struct xe_device *xe)
 	 * lock(A)                       |
 	 *                               | xe_pm_runtime_suspend()
 	 *                               |      lock(A)
-	 * xe_device_mem_access_get()    |
+	 * xe_pm_runtime_get()           |
 	 *
 	 * This will clearly deadlock since rpm core needs to wait for
 	 * xe_pm_runtime_suspend() to complete, but here we are holding lock(A)
 	 * on CPU0 which prevents CPU1 making forward progress.  With the
-	 * annotation here and in xe_device_mem_access_get() lockdep will see
+	 * annotation here and in xe_pm_runtime_get() lockdep will see
 	 * the potential lock inversion and give us a nice splat.
 	 */
-	lock_map_acquire(&xe_device_mem_access_lockdep_map);
+	lock_map_acquire(&xe_pm_runtime_lockdep_map);
 
 	/*
 	 * Applying lock for entire list op as xe_ttm_bo_destroy and xe_bo_move_notify
@@ -343,7 +349,7 @@ int xe_pm_runtime_suspend(struct xe_device *xe)
 
 	xe_irq_suspend(xe);
 out:
-	lock_map_release(&xe_device_mem_access_lockdep_map);
+	lock_map_release(&xe_pm_runtime_lockdep_map);
 	xe_pm_write_callback_task(xe, NULL);
 	return err;
 }
@@ -363,7 +369,7 @@ int xe_pm_runtime_resume(struct xe_device *xe)
 	/* Disable access_ongoing asserts and prevent recursive pm calls */
 	xe_pm_write_callback_task(xe, current);
 
-	lock_map_acquire(&xe_device_mem_access_lockdep_map);
+	lock_map_acquire(&xe_pm_runtime_lockdep_map);
 
 	/*
 	 * It can be possible that xe has allowed d3cold but other pcie devices
@@ -400,11 +406,31 @@ int xe_pm_runtime_resume(struct xe_device *xe)
 			goto out;
 	}
 out:
-	lock_map_release(&xe_device_mem_access_lockdep_map);
+	lock_map_release(&xe_pm_runtime_lockdep_map);
 	xe_pm_write_callback_task(xe, NULL);
 	return err;
 }
 
+/*
+ * For places where resume is synchronous it can be quite easy to deadlock
+ * if we are not careful. Also in practice it might be quite timing
+ * sensitive to ever see the 0 -> 1 transition with the callers locks
+ * held, so deadlocks might exist but are hard for lockdep to ever see.
+ * With this in mind, help lockdep learn about the potentially scary
+ * stuff that can happen inside the runtime_resume callback by acquiring
+ * a dummy lock (it doesn't protect anything and gets compiled out on
+ * non-debug builds).  Lockdep then only needs to see the
+ * xe_pm_runtime_lockdep_map -> runtime_resume callback once, and then can
+ * hopefully validate all the (callers_locks) -> xe_pm_runtime_lockdep_map.
+ * For example if the (callers_locks) are ever grabbed in the
+ * runtime_resume callback, lockdep should give us a nice splat.
+ */
+static void pm_runtime_lockdep_prime(void)
+{
+	lock_map_acquire(&xe_pm_runtime_lockdep_map);
+	lock_map_release(&xe_pm_runtime_lockdep_map);
+}
+
 /**
  * xe_pm_runtime_get - Get a runtime_pm reference and resume synchronously
  * @xe: xe device instance
@@ -416,6 +442,7 @@ void xe_pm_runtime_get(struct xe_device *xe)
 	if (xe_pm_read_callback_task(xe) == current)
 		return;
 
+	pm_runtime_lockdep_prime();
 	pm_runtime_resume(xe->drm.dev);
 }
 
@@ -445,6 +472,7 @@ int xe_pm_runtime_get_ioctl(struct xe_device *xe)
 	if (WARN_ON(xe_pm_read_callback_task(xe) == current))
 		return -ELOOP;
 
+	pm_runtime_lockdep_prime();
 	return pm_runtime_get_sync(xe->drm.dev);
 }
 
@@ -511,6 +539,7 @@ bool xe_pm_runtime_resume_and_get(struct xe_device *xe)
 		return true;
 	}
 
+	pm_runtime_lockdep_prime();
 	return pm_runtime_resume_and_get(xe->drm.dev) >= 0;
 }
 
-- 
2.44.0

[PATCH 05/11] drm/xe: Convert GSC HDCP from mem_access to direct xe_pm_runtime calls

[Rodrigo Vivi]

We need to convert so we can continue to kill the mem_access.

At this point we should be protected by the display wakerefs already,
so let's use the noresume variant.

Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
---
 drivers/gpu/drm/xe/display/xe_hdcp_gsc.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/drivers/gpu/drm/xe/display/xe_hdcp_gsc.c b/drivers/gpu/drm/xe/display/xe_hdcp_gsc.c
index 25c73602ef55..a221f0cf4bac 100644
--- a/drivers/gpu/drm/xe/display/xe_hdcp_gsc.c
+++ b/drivers/gpu/drm/xe/display/xe_hdcp_gsc.c
@@ -217,7 +217,7 @@ ssize_t intel_hdcp_gsc_msg_send(struct xe_device *xe, u8 *msg_in,
 	addr_out_off = PAGE_SIZE;
 
 	host_session_id = xe_gsc_create_host_session_id();
-	xe_device_mem_access_get(xe);
+	xe_pm_runtime_get_noresume(xe);
 	addr_in_wr_off = xe_gsc_emit_header(xe, &hdcp_message->hdcp_bo->vmap,
 					    addr_in_wr_off, HECI_MEADDRESS_HDCP,
 					    host_session_id, msg_in_len);
@@ -249,6 +249,6 @@ ssize_t intel_hdcp_gsc_msg_send(struct xe_device *xe, u8 *msg_in,
 			   msg_out_len);
 
 out:
-	xe_device_mem_access_put(xe);
+	xe_pm_runtime_put(xe);
 	return ret;
 }
-- 
2.44.0

[PATCH 06/11] drm/xe: Remove useless mem_access during probe

[Rodrigo Vivi]

xe_pm_init is the very last thing during the xe_pci_probe(),
hence these protections are useless from the point of view
of ensuring that the device is awake.

Let's remove it so we continue towards the goal of killing
xe_device_mem_access.

v2: Adding more cases
v3: Provide a separate fix for xe_tile_init_noalloc return (Matt)
    Adding a new case where display HDCP init calls which
    are also called at display probe time.

Cc: Matthew Auld <matthew.auld@intel.com>
Reviewed-by: Matthew Auld <matthew.auld@intel.com>
Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
---
 drivers/gpu/drm/xe/display/xe_hdcp_gsc.c |  2 --
 drivers/gpu/drm/xe/xe_ggtt.c             |  2 --
 drivers/gpu/drm/xe/xe_gt.c               |  9 ---------
 drivers/gpu/drm/xe/xe_tile.c             | 15 +++++----------
 drivers/gpu/drm/xe/xe_uc.c               | 11 -----------
 5 files changed, 5 insertions(+), 34 deletions(-)

diff --git a/drivers/gpu/drm/xe/display/xe_hdcp_gsc.c b/drivers/gpu/drm/xe/display/xe_hdcp_gsc.c
index a221f0cf4bac..b2bd56a9b76d 100644
--- a/drivers/gpu/drm/xe/display/xe_hdcp_gsc.c
+++ b/drivers/gpu/drm/xe/display/xe_hdcp_gsc.c
@@ -70,7 +70,6 @@ static int intel_hdcp_gsc_initialize_message(struct xe_device *xe,
 	int ret = 0;
 
 	/* allocate object of two page for HDCP command memory and store it */
-	xe_device_mem_access_get(xe);
 	bo = xe_bo_create_pin_map(xe, xe_device_get_root_tile(xe), NULL, PAGE_SIZE * 2,
 				  ttm_bo_type_kernel,
 				  XE_BO_CREATE_SYSTEM_BIT |
@@ -90,7 +89,6 @@ static int intel_hdcp_gsc_initialize_message(struct xe_device *xe,
 	hdcp_message->hdcp_cmd_in = cmd_in;
 	hdcp_message->hdcp_cmd_out = cmd_out;
 out:
-	xe_device_mem_access_put(xe);
 	return ret;
 }
 
diff --git a/drivers/gpu/drm/xe/xe_ggtt.c b/drivers/gpu/drm/xe/xe_ggtt.c
index 325337c38961..0f96b7db6dcc 100644
--- a/drivers/gpu/drm/xe/xe_ggtt.c
+++ b/drivers/gpu/drm/xe/xe_ggtt.c
@@ -206,14 +206,12 @@ static void xe_ggtt_initial_clear(struct xe_ggtt *ggtt)
 	u64 start, end;
 
 	/* Display may have allocated inside ggtt, so be careful with clearing here */
-	xe_device_mem_access_get(tile_to_xe(ggtt->tile));
 	mutex_lock(&ggtt->lock);
 	drm_mm_for_each_hole(hole, &ggtt->mm, start, end)
 		xe_ggtt_clear(ggtt, start, end - start);
 
 	xe_ggtt_invalidate(ggtt);
 	mutex_unlock(&ggtt->lock);
-	xe_device_mem_access_put(tile_to_xe(ggtt->tile));
 }
 
 int xe_ggtt_init(struct xe_ggtt *ggtt)
diff --git a/drivers/gpu/drm/xe/xe_gt.c b/drivers/gpu/drm/xe/xe_gt.c
index 85408e7a932b..063b710a8c7b 100644
--- a/drivers/gpu/drm/xe/xe_gt.c
+++ b/drivers/gpu/drm/xe/xe_gt.c
@@ -347,7 +347,6 @@ static int gt_fw_domain_init(struct xe_gt *gt)
 {
 	int err, i;
 
-	xe_device_mem_access_get(gt_to_xe(gt));
 	err = xe_force_wake_get(gt_to_fw(gt), XE_FW_GT);
 	if (err)
 		goto err_hw_fence_irq;
@@ -389,7 +388,6 @@ static int gt_fw_domain_init(struct xe_gt *gt)
 
 	err = xe_force_wake_put(gt_to_fw(gt), XE_FW_GT);
 	XE_WARN_ON(err);
-	xe_device_mem_access_put(gt_to_xe(gt));
 
 	return 0;
 
@@ -399,7 +397,6 @@ static int gt_fw_domain_init(struct xe_gt *gt)
 err_hw_fence_irq:
 	for (i = 0; i < XE_ENGINE_CLASS_MAX; ++i)
 		xe_hw_fence_irq_finish(&gt->fence_irq[i]);
-	xe_device_mem_access_put(gt_to_xe(gt));
 
 	return err;
 }
@@ -408,7 +405,6 @@ static int all_fw_domain_init(struct xe_gt *gt)
 {
 	int err, i;
 
-	xe_device_mem_access_get(gt_to_xe(gt));
 	err = xe_force_wake_get(gt_to_fw(gt), XE_FORCEWAKE_ALL);
 	if (err)
 		goto err_hw_fence_irq;
@@ -474,7 +470,6 @@ static int all_fw_domain_init(struct xe_gt *gt)
 
 	err = xe_force_wake_put(gt_to_fw(gt), XE_FORCEWAKE_ALL);
 	XE_WARN_ON(err);
-	xe_device_mem_access_put(gt_to_xe(gt));
 
 	return 0;
 
@@ -483,7 +478,6 @@ static int all_fw_domain_init(struct xe_gt *gt)
 err_hw_fence_irq:
 	for (i = 0; i < XE_ENGINE_CLASS_MAX; ++i)
 		xe_hw_fence_irq_finish(&gt->fence_irq[i]);
-	xe_device_mem_access_put(gt_to_xe(gt));
 
 	return err;
 }
@@ -496,7 +490,6 @@ int xe_gt_init_hwconfig(struct xe_gt *gt)
 {
 	int err;
 
-	xe_device_mem_access_get(gt_to_xe(gt));
 	err = xe_force_wake_get(gt_to_fw(gt), XE_FW_GT);
 	if (err)
 		goto out;
@@ -519,8 +512,6 @@ int xe_gt_init_hwconfig(struct xe_gt *gt)
 out_fw:
 	xe_force_wake_put(gt_to_fw(gt), XE_FW_GT);
 out:
-	xe_device_mem_access_put(gt_to_xe(gt));
-
 	return err;
 }
 
diff --git a/drivers/gpu/drm/xe/xe_tile.c b/drivers/gpu/drm/xe/xe_tile.c
index 0650b2fa75ef..74ecb5f39438 100644
--- a/drivers/gpu/drm/xe/xe_tile.c
+++ b/drivers/gpu/drm/xe/xe_tile.c
@@ -160,24 +160,19 @@ int xe_tile_init_noalloc(struct xe_tile *tile)
 {
 	int err;
 
-	xe_device_mem_access_get(tile_to_xe(tile));
-
 	err = tile_ttm_mgr_init(tile);
 	if (err)
-		goto err_mem_access;
+		return err;
 
 	tile->mem.kernel_bb_pool = xe_sa_bo_manager_init(tile, SZ_1M, 16);
-	if (IS_ERR(tile->mem.kernel_bb_pool)) {
-		err = PTR_ERR(tile->mem.kernel_bb_pool);
-		goto err_mem_access;
-	}
+	if (IS_ERR(tile->mem.kernel_bb_pool))
+		return PTR_ERR(tile->mem.kernel_bb_pool);
+
 	xe_wa_apply_tile_workarounds(tile);
 
 	xe_tile_sysfs_init(tile);
 
-err_mem_access:
-	xe_device_mem_access_put(tile_to_xe(tile));
-	return err;
+	return 0;
 }
 
 void xe_tile_migrate_wait(struct xe_tile *tile)
diff --git a/drivers/gpu/drm/xe/xe_uc.c b/drivers/gpu/drm/xe/xe_uc.c
index 7033f8c1b431..4feb35c95a1c 100644
--- a/drivers/gpu/drm/xe/xe_uc.c
+++ b/drivers/gpu/drm/xe/xe_uc.c
@@ -32,11 +32,8 @@ uc_to_xe(struct xe_uc *uc)
 /* Should be called once at driver load only */
 int xe_uc_init(struct xe_uc *uc)
 {
-	struct xe_device *xe = uc_to_xe(uc);
 	int ret;
 
-	xe_device_mem_access_get(xe);
-
 	/*
 	 * We call the GuC/HuC/GSC init functions even if GuC submission is off
 	 * to correctly move our tracking of the FW state to "disabled".
@@ -65,16 +62,8 @@ int xe_uc_init(struct xe_uc *uc)
 		goto err;
 
 	ret = xe_guc_db_mgr_init(&uc->guc.dbm, ~0);
-	if (ret)
-		goto err;
-
-	xe_device_mem_access_put(xe);
-
-	return 0;
 
 err:
-	xe_device_mem_access_put(xe);
-
 	return ret;
 }
 
-- 
2.44.0

[PATCH 07/11] drm/xe: Convert xe_gem_fault to use direct xe_pm_runtime calls

[Rodrigo Vivi]

The gem page fault is one of the outer bound protections where
we want to ensure that the hardware is in D0 before proceeding
with memory access. Let's convert it towards the xe_pm_runtime
functions directly so we can then convert the mem_access to be
inner protection only and then Kill it for good.

Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Reviewed-by: Matthew Auld <matthew.auld@intel.com>
---
 drivers/gpu/drm/xe/xe_bo.c | 5 +++--
 1 file changed, 3 insertions(+), 2 deletions(-)

diff --git a/drivers/gpu/drm/xe/xe_bo.c b/drivers/gpu/drm/xe/xe_bo.c
index d7e0deb1b869..7ee4406da127 100644
--- a/drivers/gpu/drm/xe/xe_bo.c
+++ b/drivers/gpu/drm/xe/xe_bo.c
@@ -22,6 +22,7 @@
 #include "xe_gt.h"
 #include "xe_map.h"
 #include "xe_migrate.h"
+#include "xe_pm.h"
 #include "xe_preempt_fence.h"
 #include "xe_res_cursor.h"
 #include "xe_trace.h"
@@ -1107,7 +1108,7 @@ static vm_fault_t xe_gem_fault(struct vm_fault *vmf)
 	int idx;
 
 	if (needs_rpm)
-		xe_device_mem_access_get(xe);
+		xe_pm_runtime_get(xe);
 
 	ret = ttm_bo_vm_reserve(tbo, vmf);
 	if (ret)
@@ -1138,7 +1139,7 @@ static vm_fault_t xe_gem_fault(struct vm_fault *vmf)
 	dma_resv_unlock(tbo->base.resv);
 out:
 	if (needs_rpm)
-		xe_device_mem_access_put(xe);
+		xe_pm_runtime_put(xe);
 
 	return ret;
 }
-- 
2.44.0

[PATCH 08/11] drm/xe: Removing extra mem_access protection from runtime pm

[Rodrigo Vivi]

This is not needed any longer, now that we have all the protection
in place with the runtime pm itself.

Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Reviewed-by: Matthew Auld <matthew.auld@intel.com>
---
 drivers/gpu/drm/xe/xe_device.c | 8 --------
 drivers/gpu/drm/xe/xe_device.h | 1 -
 drivers/gpu/drm/xe/xe_pm.c     | 3 ---
 3 files changed, 12 deletions(-)

diff --git a/drivers/gpu/drm/xe/xe_device.c b/drivers/gpu/drm/xe/xe_device.c
index 82f686595b16..4754da58c112 100644
--- a/drivers/gpu/drm/xe/xe_device.c
+++ b/drivers/gpu/drm/xe/xe_device.c
@@ -639,14 +639,6 @@ u32 xe_device_ccs_bytes(struct xe_device *xe, u64 size)
 		DIV_ROUND_UP_ULL(size, NUM_BYTES_PER_CCS_BYTE(xe)) : 0;
 }
 
-bool xe_device_mem_access_ongoing(struct xe_device *xe)
-{
-	if (xe_pm_read_callback_task(xe) != NULL)
-		return true;
-
-	return atomic_read(&xe->mem_access.ref);
-}
-
 /**
  * xe_device_assert_mem_access - Inspect the current runtime_pm state.
  * @xe: xe device instance
diff --git a/drivers/gpu/drm/xe/xe_device.h b/drivers/gpu/drm/xe/xe_device.h
index 2653f53bee4e..2327b6c0ae6a 100644
--- a/drivers/gpu/drm/xe/xe_device.h
+++ b/drivers/gpu/drm/xe/xe_device.h
@@ -138,7 +138,6 @@ bool xe_device_mem_access_get_if_ongoing(struct xe_device *xe);
 void xe_device_mem_access_put(struct xe_device *xe);
 
 void xe_device_assert_mem_access(struct xe_device *xe);
-bool xe_device_mem_access_ongoing(struct xe_device *xe);
 
 static inline bool xe_device_in_fault_mode(struct xe_device *xe)
 {
diff --git a/drivers/gpu/drm/xe/xe_pm.c b/drivers/gpu/drm/xe/xe_pm.c
index 9d87a68ba6eb..92e159a8d52e 100644
--- a/drivers/gpu/drm/xe/xe_pm.c
+++ b/drivers/gpu/drm/xe/xe_pm.c
@@ -296,9 +296,6 @@ int xe_pm_runtime_suspend(struct xe_device *xe)
 	u8 id;
 	int err = 0;
 
-	if (xe->d3cold.allowed && xe_device_mem_access_ongoing(xe))
-		return -EBUSY;
-
 	/* Disable access_ongoing asserts and prevent recursive pm calls */
 	xe_pm_write_callback_task(xe, current);
 
-- 
2.44.0

[PATCH 09/11] drm/xe: Convert mem_access_if_ongoing to direct xe_pm_runtime_get_if_active

[Rodrigo Vivi]

Now that assert_mem_access is relying directly on the pm_runtime state
instead of the counters, there's no reason why we cannot use
the pm_runtime functions directly.

Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Reviewed-by: Matthew Auld <matthew.auld@intel.com>
---
 drivers/gpu/drm/xe/xe_device.c | 17 -----------------
 drivers/gpu/drm/xe/xe_device.h |  1 -
 drivers/gpu/drm/xe/xe_guc_ct.c |  8 ++++----
 3 files changed, 4 insertions(+), 22 deletions(-)

diff --git a/drivers/gpu/drm/xe/xe_device.c b/drivers/gpu/drm/xe/xe_device.c
index 4754da58c112..0a51eddd264b 100644
--- a/drivers/gpu/drm/xe/xe_device.c
+++ b/drivers/gpu/drm/xe/xe_device.c
@@ -655,23 +655,6 @@ void xe_device_assert_mem_access(struct xe_device *xe)
 	xe_assert(xe, !xe_pm_runtime_suspended(xe));
 }
 
-bool xe_device_mem_access_get_if_ongoing(struct xe_device *xe)
-{
-	bool active;
-
-	if (xe_pm_read_callback_task(xe) == current)
-		return true;
-
-	active = xe_pm_runtime_get_if_active(xe);
-	if (active) {
-		int ref = atomic_inc_return(&xe->mem_access.ref);
-
-		xe_assert(xe, ref != S32_MAX);
-	}
-
-	return active;
-}
-
 void xe_device_mem_access_get(struct xe_device *xe)
 {
 	int ref;
diff --git a/drivers/gpu/drm/xe/xe_device.h b/drivers/gpu/drm/xe/xe_device.h
index 2327b6c0ae6a..b45592b0bf19 100644
--- a/drivers/gpu/drm/xe/xe_device.h
+++ b/drivers/gpu/drm/xe/xe_device.h
@@ -134,7 +134,6 @@ static inline struct xe_force_wake *gt_to_fw(struct xe_gt *gt)
 }
 
 void xe_device_mem_access_get(struct xe_device *xe);
-bool xe_device_mem_access_get_if_ongoing(struct xe_device *xe);
 void xe_device_mem_access_put(struct xe_device *xe);
 
 void xe_device_assert_mem_access(struct xe_device *xe);
diff --git a/drivers/gpu/drm/xe/xe_guc_ct.c b/drivers/gpu/drm/xe/xe_guc_ct.c
index 355edd4d758a..8d7b1b42b2e6 100644
--- a/drivers/gpu/drm/xe/xe_guc_ct.c
+++ b/drivers/gpu/drm/xe/xe_guc_ct.c
@@ -1203,7 +1203,7 @@ void xe_guc_ct_fast_path(struct xe_guc_ct *ct)
 	bool ongoing;
 	int len;
 
-	ongoing = xe_device_mem_access_get_if_ongoing(ct_to_xe(ct));
+	ongoing = xe_pm_runtime_get_if_active(ct_to_xe(ct));
 	if (!ongoing && xe_pm_read_callback_task(ct_to_xe(ct)) == NULL)
 		return;
 
@@ -1216,7 +1216,7 @@ void xe_guc_ct_fast_path(struct xe_guc_ct *ct)
 	spin_unlock(&ct->fast_lock);
 
 	if (ongoing)
-		xe_device_mem_access_put(xe);
+		xe_pm_runtime_put(xe);
 }
 
 /* Returns less than zero on error, 0 on done, 1 on more available */
@@ -1273,7 +1273,7 @@ static void g2h_worker_func(struct work_struct *w)
 	 * responses, if the worker here is blocked on those callbacks
 	 * completing, creating a deadlock.
 	 */
-	ongoing = xe_device_mem_access_get_if_ongoing(ct_to_xe(ct));
+	ongoing = xe_pm_runtime_get_if_active(ct_to_xe(ct));
 	if (!ongoing && xe_pm_read_callback_task(ct_to_xe(ct)) == NULL)
 		return;
 
@@ -1292,7 +1292,7 @@ static void g2h_worker_func(struct work_struct *w)
 	} while (ret == 1);
 
 	if (ongoing)
-		xe_device_mem_access_put(ct_to_xe(ct));
+		xe_pm_runtime_put(ct_to_xe(ct));
 }
 
 static void guc_ctb_snapshot_capture(struct xe_device *xe, struct guc_ctb *ctb,
-- 
2.44.0

[PATCH 10/11] drm/xe: Ensure all the inner access are using the _noresume variant

[Rodrigo Vivi]

At this point mem_access references should be only used as inner
points of the execution and a get with synchronous resume previously
called at an outer point.

So, before killing mem_acces in favor of direct accsess, let's
ensure that we first convert them towards the new _noresume
variant that will WARN us if no inner caller happened.

Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Reviewed-by: Matthew Auld <matthew.auld@intel.com>
---
 drivers/gpu/drm/xe/xe_device.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/gpu/drm/xe/xe_device.c b/drivers/gpu/drm/xe/xe_device.c
index 0a51eddd264b..6ab402153242 100644
--- a/drivers/gpu/drm/xe/xe_device.c
+++ b/drivers/gpu/drm/xe/xe_device.c
@@ -671,7 +671,7 @@ void xe_device_mem_access_get(struct xe_device *xe)
 	if (xe_pm_read_callback_task(xe) == current)
 		return;
 
-	xe_pm_runtime_get(xe);
+	xe_pm_runtime_get_noresume(xe);
 	ref = atomic_inc_return(&xe->mem_access.ref);
 
 	xe_assert(xe, ref != S32_MAX);
-- 
2.44.0

[PATCH 11/11] drm/xe: Kill xe_device_mem_access_{get*,put}

[Rodrigo Vivi]

Let's simply convert all the current callers towards direct
xe_pm_runtime access and remove this extra layer of indirection.

v2: Convert all the current callers instead of a big refactor
at once.

Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Reviewed-by: Matthew Auld <matthew.auld@intel.com>
---
 drivers/gpu/drm/xe/display/xe_fb_pin.c |  7 ++---
 drivers/gpu/drm/xe/xe_bo.c             |  8 +++---
 drivers/gpu/drm/xe/xe_device.c         | 36 --------------------------
 drivers/gpu/drm/xe/xe_device.h         |  3 ---
 drivers/gpu/drm/xe/xe_device_types.h   |  3 ---
 drivers/gpu/drm/xe/xe_exec_queue.c     |  6 ++---
 drivers/gpu/drm/xe/xe_ggtt.c           |  9 ++++---
 drivers/gpu/drm/xe/xe_sched_job.c      |  5 ++--
 drivers/gpu/drm/xe/xe_vm.c             |  6 ++---
 9 files changed, 22 insertions(+), 61 deletions(-)

diff --git a/drivers/gpu/drm/xe/display/xe_fb_pin.c b/drivers/gpu/drm/xe/display/xe_fb_pin.c
index 722c84a56607..403ed2d42f6b 100644
--- a/drivers/gpu/drm/xe/display/xe_fb_pin.c
+++ b/drivers/gpu/drm/xe/display/xe_fb_pin.c
@@ -10,6 +10,7 @@
 #include "intel_fb_pin.h"
 #include "xe_ggtt.h"
 #include "xe_gt.h"
+#include "xe_pm.h"
 
 #include <drm/ttm/ttm_bo.h>
 
@@ -190,7 +191,7 @@ static int __xe_pin_fb_vma_ggtt(struct intel_framebuffer *fb,
 	/* TODO: Consider sharing framebuffer mapping?
 	 * embed i915_vma inside intel_framebuffer
 	 */
-	xe_device_mem_access_get(tile_to_xe(ggtt->tile));
+	xe_pm_runtime_get_noresume(tile_to_xe(ggtt->tile));
 	ret = mutex_lock_interruptible(&ggtt->lock);
 	if (ret)
 		goto out;
@@ -242,7 +243,7 @@ static int __xe_pin_fb_vma_ggtt(struct intel_framebuffer *fb,
 out_unlock:
 	mutex_unlock(&ggtt->lock);
 out:
-	xe_device_mem_access_put(tile_to_xe(ggtt->tile));
+	xe_pm_runtime_put(tile_to_xe(ggtt->tile));
 	return ret;
 }
 
@@ -381,4 +382,4 @@ struct i915_address_space *intel_dpt_create(struct intel_framebuffer *fb)
 void intel_dpt_destroy(struct i915_address_space *vm)
 {
 	return;
-}
\ No newline at end of file
+}
diff --git a/drivers/gpu/drm/xe/xe_bo.c b/drivers/gpu/drm/xe/xe_bo.c
index 7ee4406da127..4c9a8ab4e840 100644
--- a/drivers/gpu/drm/xe/xe_bo.c
+++ b/drivers/gpu/drm/xe/xe_bo.c
@@ -716,7 +716,7 @@ static int xe_bo_move(struct ttm_buffer_object *ttm_bo, bool evict,
 
 	xe_assert(xe, migrate);
 	trace_xe_bo_move(bo, new_mem->mem_type, old_mem_type, move_lacks_source);
-	xe_device_mem_access_get(xe);
+	xe_pm_runtime_get_noresume(xe);
 
 	if (xe_bo_is_pinned(bo) && !xe_bo_is_user(bo)) {
 		/*
@@ -740,7 +740,7 @@ static int xe_bo_move(struct ttm_buffer_object *ttm_bo, bool evict,
 
 				if (XE_WARN_ON(new_mem->start == XE_BO_INVALID_OFFSET)) {
 					ret = -EINVAL;
-					xe_device_mem_access_put(xe);
+					xe_pm_runtime_put(xe);
 					goto out;
 				}
 
@@ -758,7 +758,7 @@ static int xe_bo_move(struct ttm_buffer_object *ttm_bo, bool evict,
 						new_mem, handle_system_ccs);
 		if (IS_ERR(fence)) {
 			ret = PTR_ERR(fence);
-			xe_device_mem_access_put(xe);
+			xe_pm_runtime_put(xe);
 			goto out;
 		}
 		if (!move_lacks_source) {
@@ -783,7 +783,7 @@ static int xe_bo_move(struct ttm_buffer_object *ttm_bo, bool evict,
 		dma_fence_put(fence);
 	}
 
-	xe_device_mem_access_put(xe);
+	xe_pm_runtime_put(xe);
 
 out:
 	return ret;
diff --git a/drivers/gpu/drm/xe/xe_device.c b/drivers/gpu/drm/xe/xe_device.c
index 6ab402153242..31182ebe0b89 100644
--- a/drivers/gpu/drm/xe/xe_device.c
+++ b/drivers/gpu/drm/xe/xe_device.c
@@ -655,42 +655,6 @@ void xe_device_assert_mem_access(struct xe_device *xe)
 	xe_assert(xe, !xe_pm_runtime_suspended(xe));
 }
 
-void xe_device_mem_access_get(struct xe_device *xe)
-{
-	int ref;
-
-	/*
-	 * This looks racy, but should be fine since the pm_callback_task only
-	 * transitions from NULL -> current (and back to NULL again), during the
-	 * runtime_resume() or runtime_suspend() callbacks, for which there can
-	 * only be a single one running for our device. We only need to prevent
-	 * recursively calling the runtime_get or runtime_put from those
-	 * callbacks, as well as preventing triggering any access_ongoing
-	 * asserts.
-	 */
-	if (xe_pm_read_callback_task(xe) == current)
-		return;
-
-	xe_pm_runtime_get_noresume(xe);
-	ref = atomic_inc_return(&xe->mem_access.ref);
-
-	xe_assert(xe, ref != S32_MAX);
-
-}
-
-void xe_device_mem_access_put(struct xe_device *xe)
-{
-	int ref;
-
-	if (xe_pm_read_callback_task(xe) == current)
-		return;
-
-	ref = atomic_dec_return(&xe->mem_access.ref);
-	xe_pm_runtime_put(xe);
-
-	xe_assert(xe, ref >= 0);
-}
-
 void xe_device_snapshot_print(struct xe_device *xe, struct drm_printer *p)
 {
 	struct xe_gt *gt;
diff --git a/drivers/gpu/drm/xe/xe_device.h b/drivers/gpu/drm/xe/xe_device.h
index b45592b0bf19..355bddfc8274 100644
--- a/drivers/gpu/drm/xe/xe_device.h
+++ b/drivers/gpu/drm/xe/xe_device.h
@@ -133,9 +133,6 @@ static inline struct xe_force_wake *gt_to_fw(struct xe_gt *gt)
 	return &gt->mmio.fw;
 }
 
-void xe_device_mem_access_get(struct xe_device *xe);
-void xe_device_mem_access_put(struct xe_device *xe);
-
 void xe_device_assert_mem_access(struct xe_device *xe);
 
 static inline bool xe_device_in_fault_mode(struct xe_device *xe)
diff --git a/drivers/gpu/drm/xe/xe_device_types.h b/drivers/gpu/drm/xe/xe_device_types.h
index 9785eef2e5a4..4bec8b90a37c 100644
--- a/drivers/gpu/drm/xe/xe_device_types.h
+++ b/drivers/gpu/drm/xe/xe_device_types.h
@@ -377,9 +377,6 @@ struct xe_device {
 	 * triggering additional actions when they occur.
 	 */
 	struct {
-		/** @mem_access.ref: ref count of memory accesses */
-		atomic_t ref;
-
 		/**
 		 * @mem_access.vram_userfault: Encapsulate vram_userfault
 		 * related stuff
diff --git a/drivers/gpu/drm/xe/xe_exec_queue.c b/drivers/gpu/drm/xe/xe_exec_queue.c
index f69a9c99329c..42ddf857e9e8 100644
--- a/drivers/gpu/drm/xe/xe_exec_queue.c
+++ b/drivers/gpu/drm/xe/xe_exec_queue.c
@@ -217,7 +217,7 @@ void xe_exec_queue_fini(struct xe_exec_queue *q)
 	for (i = 0; i < q->width; ++i)
 		xe_lrc_finish(q->lrc + i);
 	if (!(q->flags & EXEC_QUEUE_FLAG_PERMANENT) && (q->flags & EXEC_QUEUE_FLAG_VM || !q->vm))
-		xe_device_mem_access_put(gt_to_xe(q->gt));
+		xe_pm_runtime_put(gt_to_xe(q->gt));
 	__xe_exec_queue_free(q);
 }
 
@@ -589,7 +589,7 @@ int xe_exec_queue_create_ioctl(struct drm_device *dev, void *data,
 				return -EINVAL;
 
 			/* The migration vm doesn't hold rpm ref */
-			xe_device_mem_access_get(xe);
+			xe_pm_runtime_get_noresume(xe);
 
 			flags = EXEC_QUEUE_FLAG_VM | (id ? EXEC_QUEUE_FLAG_BIND_ENGINE_CHILD : 0);
 
@@ -598,7 +598,7 @@ int xe_exec_queue_create_ioctl(struct drm_device *dev, void *data,
 						   args->width, hwe, flags,
 						   args->extensions);
 
-			xe_device_mem_access_put(xe); /* now held by engine */
+			xe_pm_runtime_put(xe); /* now held by engine */
 
 			xe_vm_put(migrate_vm);
 			if (IS_ERR(new)) {
diff --git a/drivers/gpu/drm/xe/xe_ggtt.c b/drivers/gpu/drm/xe/xe_ggtt.c
index 0f96b7db6dcc..d73056aa2362 100644
--- a/drivers/gpu/drm/xe/xe_ggtt.c
+++ b/drivers/gpu/drm/xe/xe_ggtt.c
@@ -20,6 +20,7 @@
 #include "xe_gt_printk.h"
 #include "xe_gt_tlb_invalidation.h"
 #include "xe_map.h"
+#include "xe_pm.h"
 #include "xe_sriov.h"
 #include "xe_wopcm.h"
 
@@ -408,14 +409,14 @@ static int __xe_ggtt_insert_bo_at(struct xe_ggtt *ggtt, struct xe_bo *bo,
 	if (err)
 		return err;
 
-	xe_device_mem_access_get(tile_to_xe(ggtt->tile));
+	xe_pm_runtime_get_noresume(tile_to_xe(ggtt->tile));
 	mutex_lock(&ggtt->lock);
 	err = drm_mm_insert_node_in_range(&ggtt->mm, &bo->ggtt_node, bo->size,
 					  alignment, 0, start, end, 0);
 	if (!err)
 		xe_ggtt_map_bo(ggtt, bo);
 	mutex_unlock(&ggtt->lock);
-	xe_device_mem_access_put(tile_to_xe(ggtt->tile));
+	xe_pm_runtime_put(tile_to_xe(ggtt->tile));
 
 	return err;
 }
@@ -433,7 +434,7 @@ int xe_ggtt_insert_bo(struct xe_ggtt *ggtt, struct xe_bo *bo)
 
 void xe_ggtt_remove_node(struct xe_ggtt *ggtt, struct drm_mm_node *node)
 {
-	xe_device_mem_access_get(tile_to_xe(ggtt->tile));
+	xe_pm_runtime_get_noresume(tile_to_xe(ggtt->tile));
 	mutex_lock(&ggtt->lock);
 
 	xe_ggtt_clear(ggtt, node->start, node->size);
@@ -443,7 +444,7 @@ void xe_ggtt_remove_node(struct xe_ggtt *ggtt, struct drm_mm_node *node)
 	xe_ggtt_invalidate(ggtt);
 
 	mutex_unlock(&ggtt->lock);
-	xe_device_mem_access_put(tile_to_xe(ggtt->tile));
+	xe_pm_runtime_put(tile_to_xe(ggtt->tile));
 }
 
 void xe_ggtt_remove_bo(struct xe_ggtt *ggtt, struct xe_bo *bo)
diff --git a/drivers/gpu/drm/xe/xe_sched_job.c b/drivers/gpu/drm/xe/xe_sched_job.c
index 8151ddafb940..8ba35b1698ed 100644
--- a/drivers/gpu/drm/xe/xe_sched_job.c
+++ b/drivers/gpu/drm/xe/xe_sched_job.c
@@ -15,6 +15,7 @@
 #include "xe_hw_fence.h"
 #include "xe_lrc.h"
 #include "xe_macros.h"
+#include "xe_pm.h"
 #include "xe_trace.h"
 #include "xe_vm.h"
 
@@ -157,7 +158,7 @@ struct xe_sched_job *xe_sched_job_create(struct xe_exec_queue *q,
 
 	/* All other jobs require a VM to be open which has a ref */
 	if (unlikely(q->flags & EXEC_QUEUE_FLAG_KERNEL))
-		xe_device_mem_access_get(job_to_xe(job));
+		xe_pm_runtime_get_noresume(job_to_xe(job));
 	xe_device_assert_mem_access(job_to_xe(job));
 
 	trace_xe_sched_job_create(job);
@@ -190,7 +191,7 @@ void xe_sched_job_destroy(struct kref *ref)
 		container_of(ref, struct xe_sched_job, refcount);
 
 	if (unlikely(job->q->flags & EXEC_QUEUE_FLAG_KERNEL))
-		xe_device_mem_access_put(job_to_xe(job));
+		xe_pm_runtime_put(job_to_xe(job));
 	xe_exec_queue_put(job->q);
 	dma_fence_put(job->fence);
 	drm_sched_job_cleanup(&job->drm);
diff --git a/drivers/gpu/drm/xe/xe_vm.c b/drivers/gpu/drm/xe/xe_vm.c
index cbb9b8935c90..8263ae98c268 100644
--- a/drivers/gpu/drm/xe/xe_vm.c
+++ b/drivers/gpu/drm/xe/xe_vm.c
@@ -1341,7 +1341,7 @@ struct xe_vm *xe_vm_create(struct xe_device *xe, u32 flags)
 	vm->pt_ops = &xelp_pt_ops;
 
 	if (!(flags & XE_VM_FLAG_MIGRATION))
-		xe_device_mem_access_get(xe);
+		xe_pm_runtime_get_noresume(xe);
 
 	vm_resv_obj = drm_gpuvm_resv_object_alloc(&xe->drm);
 	if (!vm_resv_obj) {
@@ -1452,7 +1452,7 @@ struct xe_vm *xe_vm_create(struct xe_device *xe, u32 flags)
 		xe_range_fence_tree_fini(&vm->rftree[id]);
 	kfree(vm);
 	if (!(flags & XE_VM_FLAG_MIGRATION))
-		xe_device_mem_access_put(xe);
+		xe_pm_runtime_put(xe);
 	return ERR_PTR(err);
 }
 
@@ -1575,7 +1575,7 @@ static void vm_destroy_work_func(struct work_struct *w)
 	mutex_destroy(&vm->snap_mutex);
 
 	if (!(vm->flags & XE_VM_FLAG_MIGRATION)) {
-		xe_device_mem_access_put(xe);
+		xe_pm_runtime_put(xe);
 
 		if (xe->info.has_asid && vm->usm.asid) {
 			mutex_lock(&xe->usm.lock);
-- 
2.44.0

Re: [PATCH 02/11] drm/xe: Introduce intel_runtime_pm_get_noresume at compat-i915-headers for display

[Francois Dugast]

Hi,

On Thu, Mar 14, 2024 at 10:10:12AM -0400, Rodrigo Vivi wrote:
> The i915-display will start using the intel_runtime_pm_noresume.
> So we need to add the compat header before it.

Or "So we need to add it to the compat header before"?

> 
> Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
> ---
>  drivers/gpu/drm/xe/compat-i915-headers/i915_drv.h | 8 ++++++++
>  1 file changed, 8 insertions(+)
> 
> diff --git a/drivers/gpu/drm/xe/compat-i915-headers/i915_drv.h b/drivers/gpu/drm/xe/compat-i915-headers/i915_drv.h
> index fef969112b1d..ecaaef3df4bf 100644
> --- a/drivers/gpu/drm/xe/compat-i915-headers/i915_drv.h
> +++ b/drivers/gpu/drm/xe/compat-i915-headers/i915_drv.h
> @@ -176,6 +176,14 @@ static inline intel_wakeref_t intel_runtime_pm_get_if_in_use(struct xe_runtime_p
>  	return xe_pm_runtime_get_if_in_use(xe);
>  }
>  
> +static inline intel_wakeref_t intel_runtime_pm_get_noresume(struct xe_runtime_pm *pm)
> +{
> +	struct xe_device *xe = container_of(pm, struct xe_device, runtime_pm);
> +
> +	xe_pm_runtime_get_noresume(xe);
> +	return true;
> +}
> +

LGTM but wondering if this and the next patch in the series should be
combined in order to have at least one use of this new definition.

Either way:
Reviewed-by: Francois Dugast <francois.dugast@intel.com>

Francois

>  static inline void intel_runtime_pm_put_unchecked(struct xe_runtime_pm *pm)
>  {
>  	struct xe_device *xe = container_of(pm, struct xe_device, runtime_pm);
> -- 
> 2.44.0
> 

Re: [PATCH 03/11] drm/i915/display: convert inner wakeref get towards get_if_in_use

[Francois Dugast]

On Thu, Mar 14, 2024 at 10:10:13AM -0400, Rodrigo Vivi wrote:
> This patch brings no functional change. Since at this point of
> the code we are already asserting a wakeref was held, it means
> that we are with runtime_pm 'in_use' and in practical terms we
> are only bumping the pm_runtime usage counter and moving on.
> 
> However, xe driver has a lockdep annotation that warned us that
> if a sync resume was actually called at this point, we could have
> a deadlock because we are inside the power_domains->lock locked
> area and the resume would call the irq_reset, which would also
> try to get the power_domains->lock.
> 
> For this reason, let's convert this call to a safer option and
> calm lockdep on.
> 
> v2: use _noresume variant instead of get_in_use (Ville, Imre)
> 
> Cc: Ville Syrjälä <ville.syrjala@linux.intel.com>
> Acked-by: Imre Deak <imre.deak@intel.com>
> Cc: Matthew Auld <matthew.auld@intel.com>
> Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>

Reviewed-by: Francois Dugast <francois.dugast@intel.com>

> ---
>  drivers/gpu/drm/i915/display/intel_display_power.c | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/drivers/gpu/drm/i915/display/intel_display_power.c b/drivers/gpu/drm/i915/display/intel_display_power.c
> index 6fd4fa52253a..048943d0a881 100644
> --- a/drivers/gpu/drm/i915/display/intel_display_power.c
> +++ b/drivers/gpu/drm/i915/display/intel_display_power.c
> @@ -646,7 +646,7 @@ release_async_put_domains(struct i915_power_domains *power_domains,
>  	 * power well disabling.
>  	 */
>  	assert_rpm_raw_wakeref_held(rpm);
> -	wakeref = intel_runtime_pm_get(rpm);
> +	wakeref = intel_runtime_pm_get_noresume(rpm);
>  
>  	for_each_power_domain(domain, mask) {
>  		/* Clear before put, so put's sanity check is happy. */
> -- 
> 2.44.0
> 

✗ Fi.CI.SPARSE: warning for series starting with [01/11] drm/xe: Introduce xe_pm_runtime_get_noresume for inner callers

[Patchwork]

== Series Details ==

Series: series starting with [01/11] drm/xe: Introduce xe_pm_runtime_get_noresume for inner callers
URL   : https://patchwork.freedesktop.org/series/131141/
State : warning

== Summary ==

Error: dim sparse failed
Sparse version: v0.6.2
Fast mode used, each commit won't be checked separately.
-
+./arch/x86/include/asm/bitops.h:116:1: warning: unreplaced symbol 'return'
+./arch/x86/include/asm/bitops.h:147:1: warning: unreplaced symbol 'return'
+./arch/x86/include/asm/bitops.h:149:9: warning: unreplaced symbol 'oldbit'
+./arch/x86/include/asm/bitops.h:153:26: warning: unreplaced symbol 'oldbit'
+./arch/x86/include/asm/bitops.h:155:16: warning: unreplaced symbol 'oldbit'
+./arch/x86/include/asm/bitops.h:155:9: warning: unreplaced symbol 'return'
+./arch/x86/include/asm/bitops.h:173:1: warning: unreplaced symbol 'return'
+./arch/x86/include/asm/bitops.h:175:9: warning: unreplaced symbol 'oldbit'
+./arch/x86/include/asm/bitops.h:179:35: warning: unreplaced symbol 'oldbit'
+./arch/x86/include/asm/bitops.h:181:16: warning: unreplaced symbol 'oldbit'
+./arch/x86/include/asm/bitops.h:181:9: warning: unreplaced symbol 'return'
+./arch/x86/include/asm/bitops.h:185:1: warning: unreplaced symbol 'return'
+./arch/x86/include/asm/bitops.h:187:9: warning: unreplaced symbol 'oldbit'
+./arch/x86/include/asm/bitops.h:191:35: warning: unreplaced symbol 'oldbit'
+./arch/x86/include/asm/bitops.h:194:16: warning: unreplaced symbol 'oldbit'
+./arch/x86/include/asm/bitops.h:194:9: warning: unreplaced symbol 'return'
+./arch/x86/include/asm/bitops.h:236:1: warning: unreplaced symbol 'return'
+./arch/x86/include/asm/bitops.h:238:9: warning: unreplaced symbol 'return'
+./arch/x86/include/asm/bitops.h:66:1: warning: unreplaced symbol 'return'
+./arch/x86/include/asm/bitops.h:92:1: warning: unreplaced symbol 'return'
+./include/asm-generic/bitops/generic-non-atomic.h:100:17: warning: unreplaced symbol 'old'
+./include/asm-generic/bitops/generic-non-atomic.h:100:23: warning: unreplaced symbol 'mask'
+./include/asm-generic/bitops/generic-non-atomic.h:100:9: warning: unreplaced symbol 'return'
+./include/asm-generic/bitops/generic-non-atomic.h:105:1: warning: unreplaced symbol 'return'
+./include/asm-generic/bitops/generic-non-atomic.h:107:9: warning: unreplaced symbol 'mask'
+./include/asm-generic/bitops/generic-non-atomic.h:108:9: warning: unreplaced symbol 'p'
+./include/asm-generic/bitops/generic-non-atomic.h:109:9: warning: unreplaced symbol 'old'
+./include/asm-generic/bitops/generic-non-atomic.h:111:10: warning: unreplaced symbol 'p'
+./include/asm-generic/bitops/generic-non-atomic.h:111:14: warning: unreplaced symbol 'old'
+./include/asm-generic/bitops/generic-non-atomic.h:111:20: warning: unreplaced symbol 'mask'
+./include/asm-generic/bitops/generic-non-atomic.h:112:17: warning: unreplaced symbol 'old'
+./include/asm-generic/bitops/generic-non-atomic.h:112:23: warning: unreplaced symbol 'mask'
+./include/asm-generic/bitops/generic-non-atomic.h:112:9: warning: unreplaced symbol 'return'
+./include/asm-generic/bitops/generic-non-atomic.h:121:1: warning: unreplaced symbol 'return'
+./include/asm-generic/bitops/generic-non-atomic.h:128:9: warning: unreplaced symbol 'return'
+./include/asm-generic/bitops/generic-non-atomic.h:166:1: warning: unreplaced symbol 'return'
+./include/asm-generic/bitops/generic-non-atomic.h:168:9: warning: unreplaced symbol 'p'
+./include/asm-generic/bitops/generic-non-atomic.h:169:9: warning: unreplaced symbol 'mask'
+./include/asm-generic/bitops/generic-non-atomic.h:170:9: warning: unreplaced symbol 'val'
+./include/asm-generic/bitops/generic-non-atomic.h:172:19: warning: unreplaced symbol 'val'
+./include/asm-generic/bitops/generic-non-atomic.h:172:25: warning: unreplaced symbol 'mask'
+./include/asm-generic/bitops/generic-non-atomic.h:172:9: warning: unreplaced symbol 'return'
+./include/asm-generic/bitops/generic-non-atomic.h:28:1: warning: unreplaced symbol 'return'
+./include/asm-generic/bitops/generic-non-atomic.h:30:9: warning: unreplaced symbol 'mask'
+./include/asm-generic/bitops/generic-non-atomic.h:31:9: warning: unreplaced symbol 'p'
+./include/asm-generic/bitops/generic-non-atomic.h:33:10: warning: unreplaced symbol 'p'
+./include/asm-generic/bitops/generic-non-atomic.h:33:16: warning: unreplaced symbol 'mask'
+./include/asm-generic/bitops/generic-non-atomic.h:37:1: warning: unreplaced symbol 'return'
+./include/asm-generic/bitops/generic-non-atomic.h:39:9: warning: unreplaced symbol 'mask'
+./include/asm-generic/bitops/generic-non-atomic.h:40:9: warning: unreplaced symbol 'p'
+./include/asm-generic/bitops/generic-non-atomic.h:42:10: warning: unreplaced symbol 'p'
+./include/asm-generic/bitops/generic-non-atomic.h:42:16: warning: unreplaced symbol 'mask'
+./include/asm-generic/bitops/generic-non-atomic.h:55:1: warning: unreplaced symbol 'return'
+./include/asm-generic/bitops/generic-non-atomic.h:57:9: warning: unreplaced symbol 'mask'
+./include/asm-generic/bitops/generic-non-atomic.h:58:9: warning: unreplaced symbol 'p'
+./include/asm-generic/bitops/generic-non-atomic.h:60:10: warning: unreplaced symbol 'p'
+./include/asm-generic/bitops/generic-non-atomic.h:60:15: warning: unreplaced symbol 'mask'
+./include/asm-generic/bitops/generic-non-atomic.h:73:1: warning: unreplaced symbol 'return'
+./include/asm-generic/bitops/generic-non-atomic.h:75:9: warning: unreplaced symbol 'mask'
+./include/asm-generic/bitops/generic-non-atomic.h:76:9: warning: unreplaced symbol 'p'
+./include/asm-generic/bitops/generic-non-atomic.h:77:9: warning: unreplaced symbol 'old'
+./include/asm-generic/bitops/generic-non-atomic.h:79:10: warning: unreplaced symbol 'p'
+./include/asm-generic/bitops/generic-non-atomic.h:79:14: warning: unreplaced symbol 'old'
+./include/asm-generic/bitops/generic-non-atomic.h:79:20: warning: unreplaced symbol 'mask'
+./include/asm-generic/bitops/generic-non-atomic.h:80:17: warning: unreplaced symbol 'old'
+./include/asm-generic/bitops/generic-non-atomic.h:80:23: warning: unreplaced symbol 'mask'
+./include/asm-generic/bitops/generic-non-atomic.h:80:9: warning: unreplaced symbol 'return'
+./include/asm-generic/bitops/generic-non-atomic.h:93:1: warning: unreplaced symbol 'return'
+./include/asm-generic/bitops/generic-non-atomic.h:95:9: warning: unreplaced symbol 'mask'
+./include/asm-generic/bitops/generic-non-atomic.h:96:9: warning: unreplaced symbol 'p'
+./include/asm-generic/bitops/generic-non-atomic.h:97:9: warning: unreplaced symbol 'old'
+./include/asm-generic/bitops/generic-non-atomic.h:99:10: warning: unreplaced symbol 'p'
+./include/asm-generic/bitops/generic-non-atomic.h:99:14: warning: unreplaced symbol 'old'
+./include/asm-generic/bitops/generic-non-atomic.h:99:21: warning: unreplaced symbol 'mask'
+./include/asm-generic/bitops/instrumented-non-atomic.h:100:9: warning: unreplaced symbol 'return'
+./include/asm-generic/bitops/instrumented-non-atomic.h:112:1: warning: unreplaced symbol 'return'
+./include/asm-generic/bitops/instrumented-non-atomic.h:115:9: warning: unreplaced symbol 'return'
+./include/asm-generic/bitops/instrumented-non-atomic.h:127:1: warning: unreplaced symbol 'return'
+./include/asm-generic/bitops/instrumented-non-atomic.h:130:9: warning: unreplaced symbol 'return'
+./include/asm-generic/bitops/instrumented-non-atomic.h:139:1: warning: unreplaced symbol 'return'
+./include/asm-generic/bitops/instrumented-non-atomic.h:142:9: warning: unreplaced symbol 'return'
+./include/asm-generic/bitops/instrumented-non-atomic.h:26:1: warning: unreplaced symbol 'return'
+./include/asm-generic/bitops/instrumented-non-atomic.h:42:1: warning: unreplaced symbol 'return'
+./include/asm-generic/bitops/instrumented-non-atomic.h:58:1: warning: unreplaced symbol 'return'
+./include/asm-generic/bitops/instrumented-non-atomic.h:97:1: warning: unreplaced symbol 'return'
+./include/linux/find.h:203:45: warning: shift count is negative (-12)
+./include/linux/find.h:203:45: warning: shift count is negative (-12)

✗ Fi.CI.BAT: failure for series starting with [01/11] drm/xe: Introduce xe_pm_runtime_get_noresume for inner callers

[Patchwork]

[-- Attachment #1: Type: text/plain, Size: 11462 bytes --]

== Series Details ==

Series: series starting with [01/11] drm/xe: Introduce xe_pm_runtime_get_noresume for inner callers
URL   : https://patchwork.freedesktop.org/series/131141/
State : failure

== Summary ==

CI Bug Log - changes from CI_DRM_14433 -> Patchwork_131141v1
====================================================

Summary
-------

  **FAILURE**

  Serious unknown changes coming with Patchwork_131141v1 absolutely need to be
  verified manually.
  
  If you think the reported changes have nothing to do with the changes
  introduced in Patchwork_131141v1, please notify your bug team (I915-ci-infra@lists.freedesktop.org) to allow them
  to document this new failure mode, which will reduce false positives in CI.

  External URL: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_131141v1/index.html

Participating hosts (37 -> 36)
------------------------------

  Additional (1): fi-kbl-8809g 
  Missing    (2): bat-dg1-7 fi-snb-2520m 

Possible new issues
-------------------

  Here are the unknown changes that may have been introduced in Patchwork_131141v1:

### IGT changes ###

#### Possible regressions ####

  * igt@gem_busy@busy@all-engines:
    - fi-kbl-8809g:       NOTRUN -> [ABORT][1]
   [1]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_131141v1/fi-kbl-8809g/igt@gem_busy@busy@all-engines.html

  * igt@gem_close_race@basic-process:
    - bat-dg2-8:          [PASS][2] -> [INCOMPLETE][3]
   [2]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_14433/bat-dg2-8/igt@gem_close_race@basic-process.html
   [3]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_131141v1/bat-dg2-8/igt@gem_close_race@basic-process.html
    - bat-mtlp-6:         [PASS][4] -> [DMESG-WARN][5]
   [4]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_14433/bat-mtlp-6/igt@gem_close_race@basic-process.html
   [5]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_131141v1/bat-mtlp-6/igt@gem_close_race@basic-process.html

  * igt@i915_module_load@load:
    - bat-atsm-1:         [PASS][6] -> [INCOMPLETE][7]
   [6]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_14433/bat-atsm-1/igt@i915_module_load@load.html
   [7]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_131141v1/bat-atsm-1/igt@i915_module_load@load.html
    - bat-dg2-9:          [PASS][8] -> [INCOMPLETE][9]
   [8]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_14433/bat-dg2-9/igt@i915_module_load@load.html
   [9]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_131141v1/bat-dg2-9/igt@i915_module_load@load.html
    - fi-kbl-guc:         [PASS][10] -> [ABORT][11]
   [10]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_14433/fi-kbl-guc/igt@i915_module_load@load.html
   [11]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_131141v1/fi-kbl-guc/igt@i915_module_load@load.html

  * igt@kms_busy@basic@flip:
    - fi-pnv-d510:        [PASS][12] -> [ABORT][13]
   [12]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_14433/fi-pnv-d510/igt@kms_busy@basic@flip.html
   [13]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_131141v1/fi-pnv-d510/igt@kms_busy@basic@flip.html

  * igt@kms_busy@basic@modeset:
    - bat-adln-1:         [PASS][14] -> [INCOMPLETE][15]
   [14]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_14433/bat-adln-1/igt@kms_busy@basic@modeset.html
   [15]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_131141v1/bat-adln-1/igt@kms_busy@basic@modeset.html
    - bat-rplp-1:         [PASS][16] -> [INCOMPLETE][17]
   [16]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_14433/bat-rplp-1/igt@kms_busy@basic@modeset.html
   [17]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_131141v1/bat-rplp-1/igt@kms_busy@basic@modeset.html
    - fi-ivb-3770:        [PASS][18] -> [INCOMPLETE][19]
   [18]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_14433/fi-ivb-3770/igt@kms_busy@basic@modeset.html
   [19]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_131141v1/fi-ivb-3770/igt@kms_busy@basic@modeset.html
    - bat-adls-6:         [PASS][20] -> [ABORT][21]
   [20]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_14433/bat-adls-6/igt@kms_busy@basic@modeset.html
   [21]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_131141v1/bat-adls-6/igt@kms_busy@basic@modeset.html
    - fi-ilk-650:         [PASS][22] -> [INCOMPLETE][23]
   [22]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_14433/fi-ilk-650/igt@kms_busy@basic@modeset.html
   [23]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_131141v1/fi-ilk-650/igt@kms_busy@basic@modeset.html
    - bat-arls-1:         [PASS][24] -> [ABORT][25]
   [24]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_14433/bat-arls-1/igt@kms_busy@basic@modeset.html
   [25]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_131141v1/bat-arls-1/igt@kms_busy@basic@modeset.html

  * igt@kms_flip@basic-flip-vs-dpms@b-dp2:
    - bat-rpls-3:         [PASS][26] -> [ABORT][27]
   [26]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_14433/bat-rpls-3/igt@kms_flip@basic-flip-vs-dpms@b-dp2.html
   [27]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_131141v1/bat-rpls-3/igt@kms_flip@basic-flip-vs-dpms@b-dp2.html

  * igt@kms_flip@basic-flip-vs-dpms@b-edp1:
    - bat-jsl-3:          [PASS][28] -> [INCOMPLETE][29]
   [28]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_14433/bat-jsl-3/igt@kms_flip@basic-flip-vs-dpms@b-edp1.html
   [29]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_131141v1/bat-jsl-3/igt@kms_flip@basic-flip-vs-dpms@b-edp1.html

  * igt@kms_flip@basic-flip-vs-dpms@b-hdmi-a1:
    - fi-cfl-guc:         [PASS][30] -> [INCOMPLETE][31]
   [30]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_14433/fi-cfl-guc/igt@kms_flip@basic-flip-vs-dpms@b-hdmi-a1.html
   [31]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_131141v1/fi-cfl-guc/igt@kms_flip@basic-flip-vs-dpms@b-hdmi-a1.html
    - fi-tgl-1115g4:      [PASS][32] -> [INCOMPLETE][33]
   [32]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_14433/fi-tgl-1115g4/igt@kms_flip@basic-flip-vs-dpms@b-hdmi-a1.html
   [33]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_131141v1/fi-tgl-1115g4/igt@kms_flip@basic-flip-vs-dpms@b-hdmi-a1.html

  * igt@kms_flip@basic-flip-vs-dpms@b-hdmi-a2:
    - fi-glk-j4005:       [PASS][34] -> [INCOMPLETE][35]
   [34]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_14433/fi-glk-j4005/igt@kms_flip@basic-flip-vs-dpms@b-hdmi-a2.html
   [35]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_131141v1/fi-glk-j4005/igt@kms_flip@basic-flip-vs-dpms@b-hdmi-a2.html

  
#### Suppressed ####

  The following results come from untrusted machines, tests, or statuses.
  They do not affect the overall result.

  * igt@kms_busy@basic@modeset:
    - {bat-mtlp-9}:       [PASS][36] -> [ABORT][37]
   [36]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_14433/bat-mtlp-9/igt@kms_busy@basic@modeset.html
   [37]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_131141v1/bat-mtlp-9/igt@kms_busy@basic@modeset.html
    - {bat-arls-4}:       [PASS][38] -> [ABORT][39]
   [38]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_14433/bat-arls-4/igt@kms_busy@basic@modeset.html
   [39]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_131141v1/bat-arls-4/igt@kms_busy@basic@modeset.html

  
Known issues
------------

  Here are the changes found in Patchwork_131141v1 that come from known issues:

### CI changes ###

#### Issues hit ####

  * boot:
    - bat-jsl-1:          [PASS][40] -> [FAIL][41] ([i915#8293])
   [40]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_14433/bat-jsl-1/boot.html
   [41]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_131141v1/bat-jsl-1/boot.html
    - fi-cfl-8109u:       [PASS][42] -> [FAIL][43] ([i915#8293])
   [42]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_14433/fi-cfl-8109u/boot.html
   [43]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_131141v1/fi-cfl-8109u/boot.html

  

### IGT changes ###

#### Issues hit ####

  * igt@fbdev@eof:
    - fi-kbl-8809g:       NOTRUN -> [SKIP][44] +4 other tests skip
   [44]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_131141v1/fi-kbl-8809g/igt@fbdev@eof.html

  * igt@fbdev@info:
    - fi-kbl-8809g:       NOTRUN -> [SKIP][45] ([i915#1849])
   [45]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_131141v1/fi-kbl-8809g/igt@fbdev@info.html

  * igt@kms_busy@basic@modeset:
    - fi-apl-guc:         [PASS][46] -> [INCOMPLETE][47] ([i915#10056])
   [46]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_14433/fi-apl-guc/igt@kms_busy@basic@modeset.html
   [47]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_131141v1/fi-apl-guc/igt@kms_busy@basic@modeset.html
    - fi-kbl-7567u:       [PASS][48] -> [INCOMPLETE][49] ([i915#10056])
   [48]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_14433/fi-kbl-7567u/igt@kms_busy@basic@modeset.html
   [49]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_131141v1/fi-kbl-7567u/igt@kms_busy@basic@modeset.html
    - bat-dg2-14:         [PASS][50] -> [INCOMPLETE][51] ([i915#10056])
   [50]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_14433/bat-dg2-14/igt@kms_busy@basic@modeset.html
   [51]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_131141v1/bat-dg2-14/igt@kms_busy@basic@modeset.html
    - bat-mtlp-8:         [PASS][52] -> [INCOMPLETE][53] ([i915#10056])
   [52]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_14433/bat-mtlp-8/igt@kms_busy@basic@modeset.html
   [53]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_131141v1/bat-mtlp-8/igt@kms_busy@basic@modeset.html
    - bat-adlp-6:         [PASS][54] -> [INCOMPLETE][55] ([i915#10056])
   [54]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_14433/bat-adlp-6/igt@kms_busy@basic@modeset.html
   [55]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_131141v1/bat-adlp-6/igt@kms_busy@basic@modeset.html

  
  {name}: This element is suppressed. This means it is ignored when computing
          the status of the difference (SUCCESS, WARNING, or FAILURE).

  [i915#10056]: https://gitlab.freedesktop.org/drm/intel/issues/10056
  [i915#1849]: https://gitlab.freedesktop.org/drm/intel/issues/1849
  [i915#8293]: https://gitlab.freedesktop.org/drm/intel/issues/8293


Build changes
-------------

  * Linux: CI_DRM_14433 -> Patchwork_131141v1

  CI-20190529: 20190529
  CI_DRM_14433: f5246a1cf97dd1b9bf955148b0ca2cbf0b18d31e @ git://anongit.freedesktop.org/gfx-ci/linux
  IGT_7765: 67615901aedcc79d8d1523c6484c67c11ba0e316 @ https://gitlab.freedesktop.org/drm/igt-gpu-tools.git
  Patchwork_131141v1: f5246a1cf97dd1b9bf955148b0ca2cbf0b18d31e @ git://anongit.freedesktop.org/gfx-ci/linux


### Linux commits

e69cf4d9a53b drm/xe: Kill xe_device_mem_access_{get*,put}
61ef856ddb5c drm/xe: Ensure all the inner access are using the _noresume variant
647d7a5b64aa drm/xe: Convert mem_access_if_ongoing to direct xe_pm_runtime_get_if_active
3db70fccb52d drm/xe: Removing extra mem_access protection from runtime pm
d99a6c1af946 drm/xe: Convert xe_gem_fault to use direct xe_pm_runtime calls
ff92126792b8 drm/xe: Remove useless mem_access during probe
1fb4ebb91fee drm/xe: Convert GSC HDCP from mem_access to direct xe_pm_runtime calls
91f51eecc7a4 drm/xe: Move lockdep protection from mem_access to xe_pm_runtime
98db7e114c42 drm/i915/display: convert inner wakeref get towards get_if_in_use
d4f1b5b53f14 drm/xe: Introduce intel_runtime_pm_get_noresume at compat-i915-headers for display
ca7a1aaa4dab drm/xe: Introduce xe_pm_runtime_get_noresume for inner callers

== Logs ==

For more details see: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_131141v1/index.html

[-- Attachment #2: Type: text/html, Size: 12926 bytes --]

Re: [PATCH 02/11] drm/xe: Introduce intel_runtime_pm_get_noresume at compat-i915-headers for display

[Rodrigo Vivi]

On Thu, Mar 14, 2024 at 07:04:02PM +0100, Francois Dugast wrote:
> Hi,
> 
> On Thu, Mar 14, 2024 at 10:10:12AM -0400, Rodrigo Vivi wrote:
> > The i915-display will start using the intel_runtime_pm_noresume.
> > So we need to add the compat header before it.
> 
> Or "So we need to add it to the compat header before"?
> 
> > 
> > Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
> > ---
> >  drivers/gpu/drm/xe/compat-i915-headers/i915_drv.h | 8 ++++++++
> >  1 file changed, 8 insertions(+)
> > 
> > diff --git a/drivers/gpu/drm/xe/compat-i915-headers/i915_drv.h b/drivers/gpu/drm/xe/compat-i915-headers/i915_drv.h
> > index fef969112b1d..ecaaef3df4bf 100644
> > --- a/drivers/gpu/drm/xe/compat-i915-headers/i915_drv.h
> > +++ b/drivers/gpu/drm/xe/compat-i915-headers/i915_drv.h
> > @@ -176,6 +176,14 @@ static inline intel_wakeref_t intel_runtime_pm_get_if_in_use(struct xe_runtime_p
> >  	return xe_pm_runtime_get_if_in_use(xe);
> >  }
> >  
> > +static inline intel_wakeref_t intel_runtime_pm_get_noresume(struct xe_runtime_pm *pm)
> > +{
> > +	struct xe_device *xe = container_of(pm, struct xe_device, runtime_pm);
> > +
> > +	xe_pm_runtime_get_noresume(xe);
> > +	return true;
> > +}
> > +
> 
> LGTM but wondering if this and the next patch in the series should be
> combined in order to have at least one use of this new definition.

I preferred to keep separate because they are in different drivers and
because the compilation doesn't break without the user here.
(I just double checked that).

> 
> Either way:
> Reviewed-by: Francois Dugast <francois.dugast@intel.com>
> 
> Francois
> 
> >  static inline void intel_runtime_pm_put_unchecked(struct xe_runtime_pm *pm)
> >  {
> >  	struct xe_device *xe = container_of(pm, struct xe_device, runtime_pm);
> > -- 
> > 2.44.0
> >