[Intel-gfx] [PATCH v2] drm/i915: Fix races on fbdev

[Lukas Wunner]

The ->lastclose callback invokes intel_fbdev_restore_mode() and has
been witnessed to run before intel_fbdev_initial_config_async()
has finished.

We might likewise receive hotplug events before we've had a chance to
fully set up the fbdev.

Fix by waiting for the asynchronous thread to finish.

v2:
An async_synchronize_full() was also added to intel_fbdev_set_suspend()
in v1 which turned out to be entirely gratuitous. It caused a deadlock
on suspend (discovered by CI, thanks to Damien Lespiau and Tomi Sarvela
for CI support) and was unnecessary since a device is never suspended
until its ->probe callback (and all asynchronous tasks it scheduled)
have finished. See dpm_prepare(), which calls wait_for_device_probe(),
which calls async_synchronize_full().

Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=93580
Reported-by: Gustav Fägerlind <gustav.fagerlind at gmail.com>
Reported-by: "Li, Weinan Z" <weinan.z.li at intel.com>
Cc: Chris Wilson <chris at chris-wilson.co.uk>
Cc: stable at vger.kernel.org
Signed-off-by: Lukas Wunner <lukas at wunner.de>
---
 drivers/gpu/drm/i915/i915_dma.c    | 8 +++-----
 drivers/gpu/drm/i915/intel_fbdev.c | 3 +++
 2 files changed, 6 insertions(+), 5 deletions(-)

diff --git a/drivers/gpu/drm/i915/i915_dma.c b/drivers/gpu/drm/i915/i915_dma.c
index 4aa3db6..9d76dfb 100644
--- a/drivers/gpu/drm/i915/i915_dma.c
+++ b/drivers/gpu/drm/i915/i915_dma.c
@@ -430,11 +430,9 @@ static int i915_load_modeset_init(struct drm_device *dev)
 	 * Some ports require correctly set-up hpd registers for detection to
 	 * work properly (leading to ghost connected connector status), e.g. VGA
 	 * on gm45.  Hence we can only set up the initial fbdev config after hpd
-	 * irqs are fully enabled. Now we should scan for the initial config
-	 * only once hotplug handling is enabled, but due to screwed-up locking
-	 * around kms/fbdev init we can't protect the fdbev initial config
-	 * scanning against hotplug events. Hence do this first and ignore the
-	 * tiny window where we will loose hotplug notifactions.
+	 * irqs are fully enabled. We protect the fbdev initial config scanning
+	 * against hotplug events by waiting in intel_fbdev_output_poll_changed
+	 * until the asynchronous thread has finished.
 	 */
 	intel_fbdev_initial_config_async(dev);
 
diff --git a/drivers/gpu/drm/i915/intel_fbdev.c b/drivers/gpu/drm/i915/intel_fbdev.c
index ae9cf6f..7e73acc 100644
--- a/drivers/gpu/drm/i915/intel_fbdev.c
+++ b/drivers/gpu/drm/i915/intel_fbdev.c
@@ -800,6 +800,8 @@ void intel_fbdev_set_suspend(struct drm_device *dev, int state, bool synchronous
 void intel_fbdev_output_poll_changed(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
+
+	async_synchronize_full();
 	if (dev_priv->fbdev)
 		drm_fb_helper_hotplug_event(&dev_priv->fbdev->helper);
 }
@@ -811,6 +813,7 @@ void intel_fbdev_restore_mode(struct drm_device *dev)
 	struct intel_fbdev *ifbdev = dev_priv->fbdev;
 	struct drm_fb_helper *fb_helper;
 
+	async_synchronize_full();
 	if (!ifbdev)
 		return;
 
-- 
1.8.5.2 (Apple Git-48)