Merge branch 'drm-core-next' of ../linux-2.6 into drm-next

1 files changed, 94 insertions, 8 deletions

diff --git a/drivers/gpu/drm/i915/i915_irq.c b/drivers/gpu/drm/i915/i915_irq.c
index 6c89f2ff2495..c3ceffa46ea0 100644
--- a/drivers/gpu/drm/i915/i915_irq.c
+++ b/drivers/gpu/drm/i915/i915_irq.c
@@ -31,6 +31,7 @@
 #include "drm.h"
 #include "i915_drm.h"
 #include "i915_drv.h"
+#include "i915_trace.h"
 #include "intel_drv.h"
 
 #define MAX_NOPID ((u32)~0)
@@ -279,7 +280,9 @@ irqreturn_t igdng_irq_handler(struct drm_device *dev)
                 }
 
                 if (gt_iir & GT_USER_INTERRUPT) {
-                        dev_priv->mm.irq_gem_seqno = i915_get_gem_seqno(dev);
+                        u32 seqno = i915_get_gem_seqno(dev);
+                        dev_priv->mm.irq_gem_seqno = seqno;
+                        trace_i915_gem_request_complete(dev, seqno);
                         DRM_WAKEUP(&dev_priv->irq_queue);
                 }
 
@@ -302,12 +305,25 @@ static void i915_error_work_func(struct work_struct *work)
         drm_i915_private_t *dev_priv = container_of(work, drm_i915_private_t,
                                                     error_work);
         struct drm_device *dev = dev_priv->dev;
-        char *event_string = "ERROR=1";
-        char *envp[] = { event_string, NULL };
+        char *error_event[] = { "ERROR=1", NULL };
+        char *reset_event[] = { "RESET=1", NULL };
+        char *reset_done_event[] = { "ERROR=0", NULL };
 
         DRM_DEBUG("generating error event\n");
-
-        kobject_uevent_env(&dev->primary->kdev.kobj, KOBJ_CHANGE, envp);
+        kobject_uevent_env(&dev->primary->kdev.kobj, KOBJ_CHANGE, error_event);
+
+        if (atomic_read(&dev_priv->mm.wedged)) {
+                if (IS_I965G(dev)) {
+                        DRM_DEBUG("resetting chip\n");
+                        kobject_uevent_env(&dev->primary->kdev.kobj, KOBJ_CHANGE, reset_event);
+                        if (!i965_reset(dev, GDRST_RENDER)) {
+                                atomic_set(&dev_priv->mm.wedged, 0);
+                                kobject_uevent_env(&dev->primary->kdev.kobj, KOBJ_CHANGE, reset_done_event);
+                        }
+                } else {
+                        printk("reboot required\n");
+                }
+        }
 }
 
 /**
@@ -372,7 +388,7 @@ out:
  * so userspace knows something bad happened (should trigger collection
  * of a ring dump etc.).
  */
-static void i915_handle_error(struct drm_device *dev)
+static void i915_handle_error(struct drm_device *dev, bool wedged)
 {
         struct drm_i915_private *dev_priv = dev->dev_private;
         u32 eir = I915_READ(EIR);
@@ -482,6 +498,16 @@ static void i915_handle_error(struct drm_device *dev)
                 I915_WRITE(IIR, I915_RENDER_COMMAND_PARSER_ERROR_INTERRUPT);
         }
 
+        if (wedged) {
+                atomic_set(&dev_priv->mm.wedged, 1);
+
+                /*
+                 * Wakeup waiting processes so they don't hang
+                 */
+                printk("i915: Waking up sleeping processes\n");
+                DRM_WAKEUP(&dev_priv->irq_queue);
+        }
+
         queue_work(dev_priv->wq, &dev_priv->error_work);
 }
 
@@ -527,7 +553,7 @@ irqreturn_t i915_driver_irq_handler(DRM_IRQ_ARGS)
                 pipeb_stats = I915_READ(PIPEBSTAT);
 
                 if (iir & I915_RENDER_COMMAND_PARSER_ERROR_INTERRUPT)
-                        i915_handle_error(dev);
+                        i915_handle_error(dev, false);
 
                 /*
                  * Clear the PIPE(A|B)STAT regs before the IIR
@@ -599,8 +625,12 @@ irqreturn_t i915_driver_irq_handler(DRM_IRQ_ARGS)
                 }
 
                 if (iir & I915_USER_INTERRUPT) {
-                        dev_priv->mm.irq_gem_seqno = i915_get_gem_seqno(dev);
+                        u32 seqno = i915_get_gem_seqno(dev);
+                        dev_priv->mm.irq_gem_seqno = seqno;
+                        trace_i915_gem_request_complete(dev, seqno);
                         DRM_WAKEUP(&dev_priv->irq_queue);
+                        dev_priv->hangcheck_count = 0;
+                        mod_timer(&dev_priv->hangcheck_timer, jiffies + DRM_I915_HANGCHECK_PERIOD);
                 }
 
                 if (pipea_stats & vblank_status) {
@@ -695,6 +725,16 @@ void i915_user_irq_put(struct drm_device *dev)
         spin_unlock_irqrestore(&dev_priv->user_irq_lock, irqflags);
 }
 
+void i915_trace_irq_get(struct drm_device *dev, u32 seqno)
+{
+        drm_i915_private_t *dev_priv = (drm_i915_private_t *) dev->dev_private;
+
+        if (dev_priv->trace_irq_seqno == 0)
+                i915_user_irq_get(dev);
+
+        dev_priv->trace_irq_seqno = seqno;
+}
+
 static int i915_wait_irq(struct drm_device * dev, int irq_nr)
 {
         drm_i915_private_t *dev_priv = (drm_i915_private_t *) dev->dev_private;
@@ -880,6 +920,52 @@ int i915_vblank_swap(struct drm_device *dev, void *data,
         return -EINVAL;
 }
 
+struct drm_i915_gem_request *i915_get_tail_request(struct drm_device *dev) {
+        drm_i915_private_t *dev_priv = dev->dev_private;
+        return list_entry(dev_priv->mm.request_list.prev, struct drm_i915_gem_request, list);
+}
+
+/**
+ * This is called when the chip hasn't reported back with completed
+ * batchbuffers in a long time. The first time this is called we simply record
+ * ACTHD. If ACTHD hasn't changed by the time the hangcheck timer elapses
+ * again, we assume the chip is wedged and try to fix it.
+ */
+void i915_hangcheck_elapsed(unsigned long data)
+{
+        struct drm_device *dev = (struct drm_device *)data;
+        drm_i915_private_t *dev_priv = dev->dev_private;
+        uint32_t acthd;
+       
+        if (!IS_I965G(dev))
+                acthd = I915_READ(ACTHD);
+        else
+                acthd = I915_READ(ACTHD_I965);
+
+        /* If all work is done then ACTHD clearly hasn't advanced. */
+        if (list_empty(&dev_priv->mm.request_list) ||
+                       i915_seqno_passed(i915_get_gem_seqno(dev), i915_get_tail_request(dev)->seqno)) {
+                dev_priv->hangcheck_count = 0;
+                return;
+        }
+
+        if (dev_priv->last_acthd == acthd && dev_priv->hangcheck_count > 0) {
+                DRM_ERROR("Hangcheck timer elapsed... GPU hung\n");
+                i915_handle_error(dev, true);
+                return;
+        } 
+
+        /* Reset timer case chip hangs without another request being added */
+        mod_timer(&dev_priv->hangcheck_timer, jiffies + DRM_I915_HANGCHECK_PERIOD);
+
+        if (acthd != dev_priv->last_acthd)
+                dev_priv->hangcheck_count = 0;
+        else
+                dev_priv->hangcheck_count++;
+
+        dev_priv->last_acthd = acthd;
+}
+
 /* drm_dma.h hooks
 */
 static void igdng_irq_preinstall(struct drm_device *dev)