Skip to contentSkip to conversation

drm/i915: make fbdev initialization asynchronous v2

PATCH Applied Jesse Barnes 3 participants 5 messages May 28, 2014 – Aug 29, 2014

Subsystems (2)
lore
Share & tools
lore
#0 lore raw

PATCH 1/1 drm/i915: make fbdev initialization asynchronous v2

This gets us out of our init code and out to userspace quite a bit faster, but does open us up to some bugs given the state of our init time locking.

v2: switch to async_schedule (Chris)
check with lockdep, seems happy (Jesse)
move hotplug enable flag set to fbdev_initial_config (Jesse)

---
 drivers/gpu/drm/i915/i915_dma.c    | 6 ++----
 drivers/gpu/drm/i915/intel_drv.h   | 9 +++++++--
 drivers/gpu/drm/i915/intel_fbdev.c | 9 +++++++--
 3 files changed, 16 insertions(+), 8 deletions(-)
drivers/gpu/drm/i915/i915_dma.c +2 -4
index eedb023..8bbc289 100644
@@ -28,6 +28,7 @@
 
 #define pr_fmt(fmt) KBUILD_MODNAME ": " fmt
 
+#include <linux/async.h>
 #include <drm/drmP.h>
 #include <drm/drm_crtc_helper.h>
 #include <drm/drm_fb_helper.h>
@@ -1367,10 +1368,7 @@ static int i915_load_modeset_init(struct drm_device *dev)
 	 * scanning against hotplug events. Hence do this first and ignore the
 	 * tiny window where we will loose hotplug notifactions.
 	 */
-	intel_fbdev_initial_config(dev);
-
-	/* Only enable hotplug handling once the fbdev is fully set up. */
-	dev_priv->enable_hotplug_processing = true;
+	async_schedule(intel_fbdev_initial_config, dev_priv);
 
 	drm_kms_helper_poll_init(dev);
 
drivers/gpu/drm/i915/intel_drv.h +7 -2
index 328b1a7..a563a83 100644
@@ -25,6 +25,7 @@
 #ifndef __INTEL_DRV_H__
 #define __INTEL_DRV_H__
 
+#include <linux/async.h>
 #include <linux/i2c.h>
 #include <linux/hdmi.h>
 #include <drm/i915_drm.h>
@@ -787,7 +788,7 @@ void intel_dvo_init(struct drm_device *dev);
 /* legacy fbdev emulation in intel_fbdev.c */
 #ifdef CONFIG_DRM_I915_FBDEV
 extern int intel_fbdev_init(struct drm_device *dev);
-extern void intel_fbdev_initial_config(struct drm_device *dev);
+extern void intel_fbdev_initial_config(void *data, async_cookie_t cookie);
 extern void intel_fbdev_fini(struct drm_device *dev);
 extern void intel_fbdev_set_suspend(struct drm_device *dev, int state);
 extern void intel_fbdev_output_poll_changed(struct drm_device *dev);
@@ -798,8 +799,12 @@ static inline int intel_fbdev_init(struct drm_device *dev)
 	return 0;
 }
 
-static inline void intel_fbdev_initial_config(struct drm_device *dev)
+static inline void intel_fbdev_initial_config(void *data, async_cookie_t cookie)
 {
+	struct drm_i915_private *dev_priv = data;
+
+	/* Only enable hotplug handling once the fbdev is fully set up. */
+	dev_priv->enable_hotplug_processing = true;
 }
 
 static inline void intel_fbdev_fini(struct drm_device *dev)
drivers/gpu/drm/i915/intel_fbdev.c +7 -2
index f73ba5e..9e7e45b 100644
@@ -24,6 +24,7 @@
  *     David Airlie
  */
 
+#include <linux/async.h>
 #include <linux/module.h>
 #include <linux/kernel.h>
 #include <linux/errno.h>
@@ -636,13 +637,16 @@ int intel_fbdev_init(struct drm_device *dev)
 	return 0;
 }
 
-void intel_fbdev_initial_config(struct drm_device *dev)
+void intel_fbdev_initial_config(void *data, async_cookie_t cookie)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = data;
 	struct intel_fbdev *ifbdev = dev_priv->fbdev;
 
 	/* Due to peculiar init order wrt to hpd handling this is separate. */
 	drm_fb_helper_initial_config(&ifbdev->helper, ifbdev->preferred_bpp);
+
+	/* Only enable hotplug handling once the fbdev is fully set up. */
+	dev_priv->enable_hotplug_processing = true;
 }
 
 void intel_fbdev_fini(struct drm_device *dev)
@@ -651,6 +655,7 @@ void intel_fbdev_fini(struct drm_device *dev)
 	if (!dev_priv->fbdev)
 		return;
 
+	async_synchronize_full();
 	intel_fbdev_destroy(dev, dev_priv->fbdev);
 	kfree(dev_priv->fbdev);
 	dev_priv->fbdev = NULL;
Signature · 1 line
1.8.4.2
#1 lore raw
Quoted 9 lines · Jesse Barnes

This gets us out of our init code and out to userspace quite a bit faster, but does open us up to some bugs given the state of our init time locking.

v2: switch to async_schedule (Chris)
check with lockdep, seems happy (Jesse)
move hotplug enable flag set to fbdev_initial_config (Jesse)

FBDEV=n and let userspace deal with probing?
-Daniel

Quoted 111 lines
---
 drivers/gpu/drm/i915/i915_dma.c    | 6 ++----
 drivers/gpu/drm/i915/intel_drv.h   | 9 +++++++--
 drivers/gpu/drm/i915/intel_fbdev.c | 9 +++++++--
 3 files changed, 16 insertions(+), 8 deletions(-)
drivers/gpu/drm/i915/i915_dma.c +2 -4
index eedb023..8bbc289 100644
@@ -28,6 +28,7 @@
 
 #define pr_fmt(fmt) KBUILD_MODNAME ": " fmt
 
+#include <linux/async.h>
 #include <drm/drmP.h>
 #include <drm/drm_crtc_helper.h>
 #include <drm/drm_fb_helper.h>
@@ -1367,10 +1368,7 @@ static int i915_load_modeset_init(struct drm_device *dev)
 	 * scanning against hotplug events. Hence do this first and ignore the
 	 * tiny window where we will loose hotplug notifactions.
 	 */
-	intel_fbdev_initial_config(dev);
-
-	/* Only enable hotplug handling once the fbdev is fully set up. */
-	dev_priv->enable_hotplug_processing = true;
+	async_schedule(intel_fbdev_initial_config, dev_priv);
 
 	drm_kms_helper_poll_init(dev);
 
drivers/gpu/drm/i915/intel_drv.h +7 -2
index 328b1a7..a563a83 100644
@@ -25,6 +25,7 @@
 #ifndef __INTEL_DRV_H__
 #define __INTEL_DRV_H__
 
+#include <linux/async.h>
 #include <linux/i2c.h>
 #include <linux/hdmi.h>
 #include <drm/i915_drm.h>
@@ -787,7 +788,7 @@ void intel_dvo_init(struct drm_device *dev);
 /* legacy fbdev emulation in intel_fbdev.c */
 #ifdef CONFIG_DRM_I915_FBDEV
 extern int intel_fbdev_init(struct drm_device *dev);
-extern void intel_fbdev_initial_config(struct drm_device *dev);
+extern void intel_fbdev_initial_config(void *data, async_cookie_t cookie);
 extern void intel_fbdev_fini(struct drm_device *dev);
 extern void intel_fbdev_set_suspend(struct drm_device *dev, int state);
 extern void intel_fbdev_output_poll_changed(struct drm_device *dev);
@@ -798,8 +799,12 @@ static inline int intel_fbdev_init(struct drm_device *dev)
 	return 0;
 }
 
-static inline void intel_fbdev_initial_config(struct drm_device *dev)
+static inline void intel_fbdev_initial_config(void *data, async_cookie_t cookie)
 {
+	struct drm_i915_private *dev_priv = data;
+
+	/* Only enable hotplug handling once the fbdev is fully set up. */
+	dev_priv->enable_hotplug_processing = true;
 }
 
 static inline void intel_fbdev_fini(struct drm_device *dev)
drivers/gpu/drm/i915/intel_fbdev.c +7 -2
index f73ba5e..9e7e45b 100644
@@ -24,6 +24,7 @@
  *     David Airlie
  */
 
+#include <linux/async.h>
 #include <linux/module.h>
 #include <linux/kernel.h>
 #include <linux/errno.h>
@@ -636,13 +637,16 @@ int intel_fbdev_init(struct drm_device *dev)
 	return 0;
 }
 
-void intel_fbdev_initial_config(struct drm_device *dev)
+void intel_fbdev_initial_config(void *data, async_cookie_t cookie)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = data;
 	struct intel_fbdev *ifbdev = dev_priv->fbdev;
 
 	/* Due to peculiar init order wrt to hpd handling this is separate. */
 	drm_fb_helper_initial_config(&ifbdev->helper, ifbdev->preferred_bpp);
+
+	/* Only enable hotplug handling once the fbdev is fully set up. */
+	dev_priv->enable_hotplug_processing = true;
 }
 
 void intel_fbdev_fini(struct drm_device *dev)
@@ -651,6 +655,7 @@ void intel_fbdev_fini(struct drm_device *dev)
 	if (!dev_priv->fbdev)
 		return;
 
+	async_synchronize_full();
 	intel_fbdev_destroy(dev, dev_priv->fbdev);
 	kfree(dev_priv->fbdev);
 	dev_priv->fbdev = NULL;
Signature · 6 lines
1.8.4.2

_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx
Signature · 3 lines
Daniel Vetter
Software Engineer, Intel Corporation
+41 (0) 79 365 57 48 - http://blog.ffwll.ch
#2 lore raw
Quoted 5 lines · On Wed, 28 May 2014 23:40:47 +0200 Daniel Vetter <[email protected]> wrote:
Quoted 4 lines · Jesse Barnes

This gets us out of our init code and out to userspace quite a bit faster, but does open us up to some bugs given the state of our init time locking.

Quoted 7 lines
Quoted 5 lines

v2: switch to async_schedule (Chris)
check with lockdep, seems happy (Jesse)
move hotplug enable flag set to fbdev_initial_config (Jesse)

FBDEV=n and let userspace deal with probing?

That's obviously my preferred solution. :)

Or were you suggesting I make a change to this patch too?

Signature · 1 line
Jesse Barnes, Intel Open Source Technology Center
Reviewed-by on v1 1/1
#3 lore raw
In reply to Jesse Barnes
Quoted 9 lines · Jesse Barnes

This gets us out of our init code and out to userspace quite a bit faster, but does open us up to some bugs given the state of our init time locking.

v2: switch to async_schedule (Chris)
check with lockdep, seems happy (Jesse)
move hotplug enable flag set to fbdev_initial_config (Jesse)

-Chris

Signature · 1 line
Chris Wilson, Intel Open Source Technology Centre
#4 lore raw
Quoted 11 lines · Chris Wilson
Quoted 9 lines · Jesse Barnes

This gets us out of our init code and out to userspace quite a bit faster, but does open us up to some bugs given the state of our init time locking.

v2: switch to async_schedule (Chris)
check with lockdep, seems happy (Jesse)
move hotplug enable flag set to fbdev_initial_config (Jesse)

Queued for -next, thanks for the patch. I've fixed up the conflicts due to the removal of enable_hotplug_processing while applying. -Daniel

Signature · 3 lines
Daniel Vetter
Software Engineer, Intel Corporation
+41 (0) 79 365 57 48 - http://blog.ffwll.ch