forked from Minki/linux
drm/i915: extract hangcheck/reset/error_state state into substruct
This has been sprinkled all over the place in dev_priv. I think it'd be good to also move all the code into a separate file like i915_gem_error.c, but that's for another patch. Reviewed-by: Damien Lespiau <damien.lespiau@intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
This commit is contained in:
parent
4b5aed6212
commit
99584db33b
@ -814,11 +814,11 @@ static int i915_error_state_open(struct inode *inode, struct file *file)
|
||||
|
||||
error_priv->dev = dev;
|
||||
|
||||
spin_lock_irqsave(&dev_priv->error_lock, flags);
|
||||
error_priv->error = dev_priv->first_error;
|
||||
spin_lock_irqsave(&dev_priv->gpu_error.lock, flags);
|
||||
error_priv->error = dev_priv->gpu_error.first_error;
|
||||
if (error_priv->error)
|
||||
kref_get(&error_priv->error->ref);
|
||||
spin_unlock_irqrestore(&dev_priv->error_lock, flags);
|
||||
spin_unlock_irqrestore(&dev_priv->gpu_error.lock, flags);
|
||||
|
||||
return single_open(file, i915_error_state, error_priv);
|
||||
}
|
||||
@ -1727,7 +1727,7 @@ i915_ring_stop_read(struct file *filp,
|
||||
int len;
|
||||
|
||||
len = snprintf(buf, sizeof(buf),
|
||||
"0x%08x\n", dev_priv->stop_rings);
|
||||
"0x%08x\n", dev_priv->gpu_error.stop_rings);
|
||||
|
||||
if (len > sizeof(buf))
|
||||
len = sizeof(buf);
|
||||
@ -1763,7 +1763,7 @@ i915_ring_stop_write(struct file *filp,
|
||||
if (ret)
|
||||
return ret;
|
||||
|
||||
dev_priv->stop_rings = val;
|
||||
dev_priv->gpu_error.stop_rings = val;
|
||||
mutex_unlock(&dev->struct_mutex);
|
||||
|
||||
return cnt;
|
||||
|
@ -1605,7 +1605,7 @@ int i915_driver_load(struct drm_device *dev, unsigned long flags)
|
||||
pci_enable_msi(dev->pdev);
|
||||
|
||||
spin_lock_init(&dev_priv->irq_lock);
|
||||
spin_lock_init(&dev_priv->error_lock);
|
||||
spin_lock_init(&dev_priv->gpu_error.lock);
|
||||
spin_lock_init(&dev_priv->rps.lock);
|
||||
mutex_init(&dev_priv->dpio_lock);
|
||||
|
||||
@ -1725,8 +1725,8 @@ int i915_driver_unload(struct drm_device *dev)
|
||||
}
|
||||
|
||||
/* Free error state after interrupts are fully disabled. */
|
||||
del_timer_sync(&dev_priv->hangcheck_timer);
|
||||
cancel_work_sync(&dev_priv->error_work);
|
||||
del_timer_sync(&dev_priv->gpu_error.hangcheck_timer);
|
||||
cancel_work_sync(&dev_priv->gpu_error.work);
|
||||
i915_destroy_error_state(dev);
|
||||
|
||||
if (dev->pdev->msi_enabled)
|
||||
|
@ -779,9 +779,9 @@ int intel_gpu_reset(struct drm_device *dev)
|
||||
}
|
||||
|
||||
/* Also reset the gpu hangman. */
|
||||
if (dev_priv->stop_rings) {
|
||||
if (dev_priv->gpu_error.stop_rings) {
|
||||
DRM_DEBUG("Simulated gpu hang, resetting stop_rings\n");
|
||||
dev_priv->stop_rings = 0;
|
||||
dev_priv->gpu_error.stop_rings = 0;
|
||||
if (ret == -ENODEV) {
|
||||
DRM_ERROR("Reset not implemented, but ignoring "
|
||||
"error for simulated gpu hangs\n");
|
||||
@ -820,12 +820,12 @@ int i915_reset(struct drm_device *dev)
|
||||
i915_gem_reset(dev);
|
||||
|
||||
ret = -ENODEV;
|
||||
if (get_seconds() - dev_priv->last_gpu_reset < 5)
|
||||
if (get_seconds() - dev_priv->gpu_error.last_reset < 5)
|
||||
DRM_ERROR("GPU hanging too fast, declaring wedged!\n");
|
||||
else
|
||||
ret = intel_gpu_reset(dev);
|
||||
|
||||
dev_priv->last_gpu_reset = get_seconds();
|
||||
dev_priv->gpu_error.last_reset = get_seconds();
|
||||
if (ret) {
|
||||
DRM_ERROR("Failed to reset chip.\n");
|
||||
mutex_unlock(&dev->struct_mutex);
|
||||
|
@ -766,6 +766,28 @@ struct i915_gem_mm {
|
||||
u32 object_count;
|
||||
};
|
||||
|
||||
struct i915_gpu_error {
|
||||
/* For hangcheck timer */
|
||||
#define DRM_I915_HANGCHECK_PERIOD 1500 /* in ms */
|
||||
#define DRM_I915_HANGCHECK_JIFFIES msecs_to_jiffies(DRM_I915_HANGCHECK_PERIOD)
|
||||
struct timer_list hangcheck_timer;
|
||||
int hangcheck_count;
|
||||
uint32_t last_acthd[I915_NUM_RINGS];
|
||||
uint32_t prev_instdone[I915_NUM_INSTDONE_REG];
|
||||
|
||||
/* For reset and error_state handling. */
|
||||
spinlock_t lock;
|
||||
/* Protected by the above dev->gpu_error.lock. */
|
||||
struct drm_i915_error_state *first_error;
|
||||
struct work_struct work;
|
||||
struct completion completion;
|
||||
|
||||
unsigned long last_reset;
|
||||
|
||||
/* For gpu hang simulation. */
|
||||
unsigned int stop_rings;
|
||||
};
|
||||
|
||||
typedef struct drm_i915_private {
|
||||
struct drm_device *dev;
|
||||
struct kmem_cache *slab;
|
||||
@ -829,16 +851,6 @@ typedef struct drm_i915_private {
|
||||
int num_pipe;
|
||||
int num_pch_pll;
|
||||
|
||||
/* For hangcheck timer */
|
||||
#define DRM_I915_HANGCHECK_PERIOD 1500 /* in ms */
|
||||
#define DRM_I915_HANGCHECK_JIFFIES msecs_to_jiffies(DRM_I915_HANGCHECK_PERIOD)
|
||||
struct timer_list hangcheck_timer;
|
||||
int hangcheck_count;
|
||||
uint32_t last_acthd[I915_NUM_RINGS];
|
||||
uint32_t prev_instdone[I915_NUM_INSTDONE_REG];
|
||||
|
||||
unsigned int stop_rings;
|
||||
|
||||
unsigned long cfb_size;
|
||||
unsigned int cfb_fb;
|
||||
enum plane cfb_plane;
|
||||
@ -886,11 +898,6 @@ typedef struct drm_i915_private {
|
||||
|
||||
unsigned int fsb_freq, mem_freq, is_ddr3;
|
||||
|
||||
spinlock_t error_lock;
|
||||
/* Protected by dev->error_lock. */
|
||||
struct drm_i915_error_state *first_error;
|
||||
struct work_struct error_work;
|
||||
struct completion error_completion;
|
||||
struct workqueue_struct *wq;
|
||||
|
||||
/* Display functions */
|
||||
@ -949,7 +956,7 @@ typedef struct drm_i915_private {
|
||||
struct drm_mm_node *compressed_fb;
|
||||
struct drm_mm_node *compressed_llb;
|
||||
|
||||
unsigned long last_gpu_reset;
|
||||
struct i915_gpu_error gpu_error;
|
||||
|
||||
/* list of fbdev register on this device */
|
||||
struct intel_fbdev *fbdev;
|
||||
|
@ -90,7 +90,7 @@ static int
|
||||
i915_gem_wait_for_error(struct drm_device *dev)
|
||||
{
|
||||
struct drm_i915_private *dev_priv = dev->dev_private;
|
||||
struct completion *x = &dev_priv->error_completion;
|
||||
struct completion *x = &dev_priv->gpu_error.completion;
|
||||
unsigned long flags;
|
||||
int ret;
|
||||
|
||||
@ -943,7 +943,7 @@ i915_gem_check_wedge(struct drm_i915_private *dev_priv,
|
||||
bool interruptible)
|
||||
{
|
||||
if (atomic_read(&dev_priv->mm.wedged)) {
|
||||
struct completion *x = &dev_priv->error_completion;
|
||||
struct completion *x = &dev_priv->gpu_error.completion;
|
||||
bool recovery_complete;
|
||||
unsigned long flags;
|
||||
|
||||
@ -2045,7 +2045,7 @@ i915_add_request(struct intel_ring_buffer *ring,
|
||||
|
||||
if (!dev_priv->mm.suspended) {
|
||||
if (i915_enable_hangcheck) {
|
||||
mod_timer(&dev_priv->hangcheck_timer,
|
||||
mod_timer(&dev_priv->gpu_error.hangcheck_timer,
|
||||
round_jiffies_up(jiffies + DRM_I915_HANGCHECK_JIFFIES));
|
||||
}
|
||||
if (was_empty) {
|
||||
@ -3803,7 +3803,7 @@ i915_gem_idle(struct drm_device *dev)
|
||||
* And not confound mm.suspended!
|
||||
*/
|
||||
dev_priv->mm.suspended = 1;
|
||||
del_timer_sync(&dev_priv->hangcheck_timer);
|
||||
del_timer_sync(&dev_priv->gpu_error.hangcheck_timer);
|
||||
|
||||
i915_kernel_lost_context(dev);
|
||||
i915_gem_cleanup_ringbuffer(dev);
|
||||
@ -4064,7 +4064,7 @@ i915_gem_load(struct drm_device *dev)
|
||||
INIT_LIST_HEAD(&dev_priv->fence_regs[i].lru_list);
|
||||
INIT_DELAYED_WORK(&dev_priv->mm.retire_work,
|
||||
i915_gem_retire_work_handler);
|
||||
init_completion(&dev_priv->error_completion);
|
||||
init_completion(&dev_priv->gpu_error.completion);
|
||||
|
||||
/* On GEN3 we really need to make sure the ARB C3 LP bit is set */
|
||||
if (IS_GEN3(dev)) {
|
||||
|
@ -356,8 +356,8 @@ static void notify_ring(struct drm_device *dev,
|
||||
|
||||
wake_up_all(&ring->irq_queue);
|
||||
if (i915_enable_hangcheck) {
|
||||
dev_priv->hangcheck_count = 0;
|
||||
mod_timer(&dev_priv->hangcheck_timer,
|
||||
dev_priv->gpu_error.hangcheck_count = 0;
|
||||
mod_timer(&dev_priv->gpu_error.hangcheck_timer,
|
||||
round_jiffies_up(jiffies + DRM_I915_HANGCHECK_JIFFIES));
|
||||
}
|
||||
}
|
||||
@ -863,7 +863,7 @@ done:
|
||||
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);
|
||||
gpu_error.work);
|
||||
struct drm_device *dev = dev_priv->dev;
|
||||
char *error_event[] = { "ERROR=1", NULL };
|
||||
char *reset_event[] = { "RESET=1", NULL };
|
||||
@ -878,7 +878,7 @@ static void i915_error_work_func(struct work_struct *work)
|
||||
atomic_set(&dev_priv->mm.wedged, 0);
|
||||
kobject_uevent_env(&dev->primary->kdev.kobj, KOBJ_CHANGE, reset_done_event);
|
||||
}
|
||||
complete_all(&dev_priv->error_completion);
|
||||
complete_all(&dev_priv->gpu_error.completion);
|
||||
}
|
||||
}
|
||||
|
||||
@ -1255,9 +1255,9 @@ static void i915_capture_error_state(struct drm_device *dev)
|
||||
unsigned long flags;
|
||||
int i, pipe;
|
||||
|
||||
spin_lock_irqsave(&dev_priv->error_lock, flags);
|
||||
error = dev_priv->first_error;
|
||||
spin_unlock_irqrestore(&dev_priv->error_lock, flags);
|
||||
spin_lock_irqsave(&dev_priv->gpu_error.lock, flags);
|
||||
error = dev_priv->gpu_error.first_error;
|
||||
spin_unlock_irqrestore(&dev_priv->gpu_error.lock, flags);
|
||||
if (error)
|
||||
return;
|
||||
|
||||
@ -1341,12 +1341,12 @@ static void i915_capture_error_state(struct drm_device *dev)
|
||||
error->overlay = intel_overlay_capture_error_state(dev);
|
||||
error->display = intel_display_capture_error_state(dev);
|
||||
|
||||
spin_lock_irqsave(&dev_priv->error_lock, flags);
|
||||
if (dev_priv->first_error == NULL) {
|
||||
dev_priv->first_error = error;
|
||||
spin_lock_irqsave(&dev_priv->gpu_error.lock, flags);
|
||||
if (dev_priv->gpu_error.first_error == NULL) {
|
||||
dev_priv->gpu_error.first_error = error;
|
||||
error = NULL;
|
||||
}
|
||||
spin_unlock_irqrestore(&dev_priv->error_lock, flags);
|
||||
spin_unlock_irqrestore(&dev_priv->gpu_error.lock, flags);
|
||||
|
||||
if (error)
|
||||
i915_error_state_free(&error->ref);
|
||||
@ -1358,10 +1358,10 @@ void i915_destroy_error_state(struct drm_device *dev)
|
||||
struct drm_i915_error_state *error;
|
||||
unsigned long flags;
|
||||
|
||||
spin_lock_irqsave(&dev_priv->error_lock, flags);
|
||||
error = dev_priv->first_error;
|
||||
dev_priv->first_error = NULL;
|
||||
spin_unlock_irqrestore(&dev_priv->error_lock, flags);
|
||||
spin_lock_irqsave(&dev_priv->gpu_error.lock, flags);
|
||||
error = dev_priv->gpu_error.first_error;
|
||||
dev_priv->gpu_error.first_error = NULL;
|
||||
spin_unlock_irqrestore(&dev_priv->gpu_error.lock, flags);
|
||||
|
||||
if (error)
|
||||
kref_put(&error->ref, i915_error_state_free);
|
||||
@ -1482,7 +1482,7 @@ void i915_handle_error(struct drm_device *dev, bool wedged)
|
||||
i915_report_and_clear_eir(dev);
|
||||
|
||||
if (wedged) {
|
||||
INIT_COMPLETION(dev_priv->error_completion);
|
||||
INIT_COMPLETION(dev_priv->gpu_error.completion);
|
||||
atomic_set(&dev_priv->mm.wedged, 1);
|
||||
|
||||
/*
|
||||
@ -1492,7 +1492,7 @@ void i915_handle_error(struct drm_device *dev, bool wedged)
|
||||
wake_up_all(&ring->irq_queue);
|
||||
}
|
||||
|
||||
queue_work(dev_priv->wq, &dev_priv->error_work);
|
||||
queue_work(dev_priv->wq, &dev_priv->gpu_error.work);
|
||||
}
|
||||
|
||||
static void i915_pageflip_stall_check(struct drm_device *dev, int pipe)
|
||||
@ -1723,7 +1723,7 @@ static bool i915_hangcheck_hung(struct drm_device *dev)
|
||||
{
|
||||
drm_i915_private_t *dev_priv = dev->dev_private;
|
||||
|
||||
if (dev_priv->hangcheck_count++ > 1) {
|
||||
if (dev_priv->gpu_error.hangcheck_count++ > 1) {
|
||||
bool hung = true;
|
||||
|
||||
DRM_ERROR("Hangcheck timer elapsed... GPU hung\n");
|
||||
@ -1782,25 +1782,29 @@ void i915_hangcheck_elapsed(unsigned long data)
|
||||
goto repeat;
|
||||
}
|
||||
|
||||
dev_priv->hangcheck_count = 0;
|
||||
dev_priv->gpu_error.hangcheck_count = 0;
|
||||
return;
|
||||
}
|
||||
|
||||
i915_get_extra_instdone(dev, instdone);
|
||||
if (memcmp(dev_priv->last_acthd, acthd, sizeof(acthd)) == 0 &&
|
||||
memcmp(dev_priv->prev_instdone, instdone, sizeof(instdone)) == 0) {
|
||||
if (memcmp(dev_priv->gpu_error.last_acthd, acthd,
|
||||
sizeof(acthd)) == 0 &&
|
||||
memcmp(dev_priv->gpu_error.prev_instdone, instdone,
|
||||
sizeof(instdone)) == 0) {
|
||||
if (i915_hangcheck_hung(dev))
|
||||
return;
|
||||
} else {
|
||||
dev_priv->hangcheck_count = 0;
|
||||
dev_priv->gpu_error.hangcheck_count = 0;
|
||||
|
||||
memcpy(dev_priv->last_acthd, acthd, sizeof(acthd));
|
||||
memcpy(dev_priv->prev_instdone, instdone, sizeof(instdone));
|
||||
memcpy(dev_priv->gpu_error.last_acthd, acthd,
|
||||
sizeof(acthd));
|
||||
memcpy(dev_priv->gpu_error.prev_instdone, instdone,
|
||||
sizeof(instdone));
|
||||
}
|
||||
|
||||
repeat:
|
||||
/* Reset timer case chip hangs without another request being added */
|
||||
mod_timer(&dev_priv->hangcheck_timer,
|
||||
mod_timer(&dev_priv->gpu_error.hangcheck_timer,
|
||||
round_jiffies_up(jiffies + DRM_I915_HANGCHECK_JIFFIES));
|
||||
}
|
||||
|
||||
@ -2769,11 +2773,12 @@ void intel_irq_init(struct drm_device *dev)
|
||||
struct drm_i915_private *dev_priv = dev->dev_private;
|
||||
|
||||
INIT_WORK(&dev_priv->hotplug_work, i915_hotplug_work_func);
|
||||
INIT_WORK(&dev_priv->error_work, i915_error_work_func);
|
||||
INIT_WORK(&dev_priv->gpu_error.work, i915_error_work_func);
|
||||
INIT_WORK(&dev_priv->rps.work, gen6_pm_rps_work);
|
||||
INIT_WORK(&dev_priv->l3_parity.error_work, ivybridge_parity_work);
|
||||
|
||||
setup_timer(&dev_priv->hangcheck_timer, i915_hangcheck_elapsed,
|
||||
setup_timer(&dev_priv->gpu_error.hangcheck_timer,
|
||||
i915_hangcheck_elapsed,
|
||||
(unsigned long) dev);
|
||||
|
||||
pm_qos_add_request(&dev_priv->pm_qos, PM_QOS_CPU_DMA_LATENCY, PM_QOS_DEFAULT_VALUE);
|
||||
|
@ -1491,7 +1491,7 @@ void intel_ring_advance(struct intel_ring_buffer *ring)
|
||||
struct drm_i915_private *dev_priv = ring->dev->dev_private;
|
||||
|
||||
ring->tail &= ring->size - 1;
|
||||
if (dev_priv->stop_rings & intel_ring_flag(ring))
|
||||
if (dev_priv->gpu_error.stop_rings & intel_ring_flag(ring))
|
||||
return;
|
||||
ring->write_tail(ring, ring->tail);
|
||||
}
|
||||
|
Loading…
Reference in New Issue
Block a user