void i915_gem_context_fini(struct drm_device *dev)
{
struct drm_i915_private *dev_priv = dev->dev_private;
struct i915_hw_context *dctx = dev_priv->ring[RCS].default_context;
if (!HAS_HW_CONTEXTS(dev))
return;
/* The only known way to stop the gpu from accessing the hw context is
* to reset it. Do this as the very last operation to avoid confusing
* other code, leading to spurious errors. */
intel_gpu_reset(dev);
/* When default context is created and switched to, base object refcount
* will be 2 (+1 from object creation and +1 from do_switch()).
* i915_gem_context_fini() will be called after gpu_idle() has switched
* to default context. So we need to unreference the base object once
* to offset the do_switch part, so that i915_gem_context_unreference()
* can then free the base object correctly. */
WARN_ON(!dev_priv->ring[RCS].last_context);
if (dev_priv->ring[RCS].last_context == dctx) {
/* Fake switch to NULL context */
WARN_ON(dctx->obj->active);
i915_gem_object_unpin(dctx->obj);
i915_gem_context_unreference(dctx);
}
i915_gem_object_unpin(dctx->obj);
i915_gem_context_unreference(dctx);
dev_priv->ring[RCS].default_context = NULL;
dev_priv->ring[RCS].last_context = NULL;
}
void i915_gem_context_fini(struct drm_device *dev)
{
struct drm_i915_private *dev_priv = dev->dev_private;
if (dev_priv->hw_contexts_disabled)
return;
/* The only known way to stop the gpu from accessing the hw context is
* to reset it. Do this as the very last operation to avoid confusing
* other code, leading to spurious errors. */
intel_gpu_reset(dev);
i915_gem_object_unpin(dev_priv->ring[RCS].default_context->obj);
do_destroy(dev_priv->ring[RCS].default_context);
}
void i915_gem_context_fini(struct drm_device *dev)
{
struct drm_i915_private *dev_priv = dev->dev_private;
struct intel_context *dctx = dev_priv->ring[RCS].default_context;
int i;
if (dctx->legacy_hw_ctx.rcs_state) {
/* The only known way to stop the gpu from accessing the hw context is
* to reset it. Do this as the very last operation to avoid confusing
* other code, leading to spurious errors. */
intel_gpu_reset(dev);
/* When default context is created and switched to, base object refcount
* will be 2 (+1 from object creation and +1 from do_switch()).
* i915_gem_context_fini() will be called after gpu_idle() has switched
* to default context. So we need to unreference the base object once
* to offset the do_switch part, so that i915_gem_context_unreference()
* can then free the base object correctly. */
WARN_ON(!dev_priv->ring[RCS].last_context);
if (dev_priv->ring[RCS].last_context == dctx) {
/* Fake switch to NULL context */
WARN_ON(dctx->legacy_hw_ctx.rcs_state->active);
i915_gem_object_ggtt_unpin(dctx->legacy_hw_ctx.rcs_state);
i915_gem_context_unreference(dctx);
dev_priv->ring[RCS].last_context = NULL;
}
i915_gem_object_ggtt_unpin(dctx->legacy_hw_ctx.rcs_state);
}
for (i = 0; i < I915_NUM_RINGS; i++) {
struct intel_engine_cs *ring = &dev_priv->ring[i];
if (ring->last_context)
i915_gem_context_unreference(ring->last_context);
ring->default_context = NULL;
ring->last_context = NULL;
}
i915_gem_context_unreference(dctx);
}
static int igt_atomic_reset(void *arg)
{
struct drm_i915_private *i915 = arg;
const typeof(*igt_atomic_phases) *p;
int err = 0;
/* Check that the resets are usable from atomic context */
igt_global_reset_lock(i915);
mutex_lock(&i915->drm.struct_mutex);
/* Flush any requests before we get started and check basics */
if (!igt_force_reset(i915))
goto unlock;
for (p = igt_atomic_phases; p->name; p++) {
GEM_TRACE("intel_gpu_reset under %s\n", p->name);
p->critical_section_begin();
reset_prepare(i915);
err = intel_gpu_reset(i915, ALL_ENGINES);
reset_finish(i915);
p->critical_section_end();
if (err) {
pr_err("intel_gpu_reset failed under %s\n", p->name);
break;
}
}
/* As we poke around the guts, do a full reset before continuing. */
igt_force_reset(i915);
unlock:
mutex_unlock(&i915->drm.struct_mutex);
igt_global_reset_unlock(i915);
return err;
}
/**
* i915_reset - reset chip after a hang
* @dev: drm device to reset
*
* Reset the chip. Useful if a hang is detected. Returns zero on successful
* reset or otherwise an error code.
*
* Procedure is fairly simple:
* - reset the chip using the reset reg
* - re-init context state
* - re-init hardware status page
* - re-init ring buffer
* - re-init interrupt state
* - re-init display
*/
int i915_reset(struct drm_device *dev)
{
struct drm_i915_private *dev_priv = dev->dev_private;
bool simulated;
int ret;
intel_reset_gt_powersave(dev);
mutex_lock(&dev->struct_mutex);
i915_gem_reset(dev);
simulated = dev_priv->gpu_error.stop_rings != 0;
ret = intel_gpu_reset(dev, ALL_ENGINES);
/* Also reset the gpu hangman. */
if (simulated) {
DRM_INFO("Simulated gpu hang, resetting stop_rings\n");
dev_priv->gpu_error.stop_rings = 0;
if (ret == -ENODEV) {
DRM_INFO("Reset not implemented, but ignoring "
"error for simulated gpu hangs\n");
ret = 0;
}
}
if (i915_stop_ring_allow_warn(dev_priv))
pr_notice("drm/i915: Resetting chip after gpu hang\n");
if (ret) {
DRM_ERROR("Failed to reset chip: %i\n", ret);
mutex_unlock(&dev->struct_mutex);
return ret;
}
intel_overlay_reset(dev_priv);
/* Ok, now get things going again... */
/*
* Everything depends on having the GTT running, so we need to start
* there. Fortunately we don't need to do this unless we reset the
* chip at a PCI level.
*
* Next we need to restore the context, but we don't use those
* yet either...
*
* Ring buffer needs to be re-initialized in the KMS case, or if X
* was running at the time of the reset (i.e. we weren't VT
* switched away).
*/
/* Used to prevent gem_check_wedged returning -EAGAIN during gpu reset */
dev_priv->gpu_error.reload_in_reset = true;
ret = i915_gem_init_hw(dev);
dev_priv->gpu_error.reload_in_reset = false;
mutex_unlock(&dev->struct_mutex);
if (ret) {
DRM_ERROR("Failed hw init on reset %d\n", ret);
return ret;
}
/*
* rps/rc6 re-init is necessary to restore state lost after the
* reset and the re-install of gt irqs. Skip for ironlake per
* previous concerns that it doesn't respond well to some forms
* of re-init after reset.
*/
if (INTEL_INFO(dev)->gen > 5)
intel_enable_gt_powersave(dev);
return 0;
}
/**
* i915_reset - reset chip after a hang
* @dev: drm device to reset
*
* Reset the chip. Useful if a hang is detected. Returns zero on successful
* reset or otherwise an error code.
*
* Procedure is fairly simple:
* - reset the chip using the reset reg
* - re-init context state
* - re-init hardware status page
* - re-init ring buffer
* - re-init interrupt state
* - re-init display
*/
int i915_reset(struct drm_device *dev)
{
drm_i915_private_t *dev_priv = dev->dev_private;
int ret;
if (!i915_try_reset)
return 0;
mutex_lock(&dev->struct_mutex);
i915_gem_reset(dev);
ret = -ENODEV;
if (get_seconds() - dev_priv->last_gpu_reset < 5)
DRM_ERROR("GPU hanging too fast, declaring wedged!\n");
else
ret = intel_gpu_reset(dev);
dev_priv->last_gpu_reset = get_seconds();
if (ret) {
DRM_ERROR("Failed to reset chip.\n");
mutex_unlock(&dev->struct_mutex);
return ret;
}
/* Ok, now get things going again... */
/*
* Everything depends on having the GTT running, so we need to start
* there. Fortunately we don't need to do this unless we reset the
* chip at a PCI level.
*
* Next we need to restore the context, but we don't use those
* yet either...
*
* Ring buffer needs to be re-initialized in the KMS case, or if X
* was running at the time of the reset (i.e. we weren't VT
* switched away).
*/
if (drm_core_check_feature(dev, DRIVER_MODESET) ||
!dev_priv->mm.suspended) {
struct intel_ring_buffer *ring;
int i;
dev_priv->mm.suspended = 0;
i915_gem_init_swizzling(dev);
for_each_ring(ring, dev_priv, i)
ring->init(ring);
i915_gem_context_init(dev);
i915_gem_init_ppgtt(dev);
/*
* It would make sense to re-init all the other hw state, at
* least the rps/rc6/emon init done within modeset_init_hw. For
* some unknown reason, this blows up my ilk, so don't.
*/
mutex_unlock(&dev->struct_mutex);
drm_irq_uninstall(dev);
drm_irq_install(dev);
} else {
mutex_unlock(&dev->struct_mutex);
}
return 0;
}
/**
* i915_reset - reset chip after a hang
* @dev: drm device to reset
*
* Reset the chip. Useful if a hang is detected. Returns zero on successful
* reset or otherwise an error code.
*
* Procedure is fairly simple:
* - reset the chip using the reset reg
* - re-init context state
* - re-init hardware status page
* - re-init ring buffer
* - re-init interrupt state
* - re-init display
*/
int i915_reset(struct drm_device *dev)
{
drm_i915_private_t *dev_priv = dev->dev_private;
bool simulated;
int ret;
if (!i915_try_reset)
return 0;
mutex_lock(&dev->struct_mutex);
i915_gem_reset(dev);
simulated = dev_priv->gpu_error.stop_rings != 0;
ret = intel_gpu_reset(dev);
/* Also reset the gpu hangman. */
if (simulated) {
DRM_INFO("Simulated gpu hang, resetting stop_rings\n");
dev_priv->gpu_error.stop_rings = 0;
if (ret == -ENODEV) {
DRM_INFO("Reset not implemented, but ignoring "
"error for simulated gpu hangs\n");
ret = 0;
}
}
if (ret) {
DRM_ERROR("Failed to reset chip: %i\n", ret);
mutex_unlock(&dev->struct_mutex);
return ret;
}
/* Ok, now get things going again... */
/*
* Everything depends on having the GTT running, so we need to start
* there. Fortunately we don't need to do this unless we reset the
* chip at a PCI level.
*
* Next we need to restore the context, but we don't use those
* yet either...
*
* Ring buffer needs to be re-initialized in the KMS case, or if X
* was running at the time of the reset (i.e. we weren't VT
* switched away).
*/
if (drm_core_check_feature(dev, DRIVER_MODESET) ||
!dev_priv->ums.mm_suspended) {
dev_priv->ums.mm_suspended = 0;
ret = i915_gem_init_hw(dev);
mutex_unlock(&dev->struct_mutex);
if (ret) {
DRM_ERROR("Failed hw init on reset %d\n", ret);
return ret;
}
drm_irq_uninstall(dev);
drm_irq_install(dev);
intel_hpd_init(dev);
} else {
mutex_unlock(&dev->struct_mutex);
}
return 0;
}
int
i915_reset(struct drm_device *dev)
{
drm_i915_private_t *dev_priv = dev->dev_private;
/*
* We really should only reset the display subsystem if we actually
* need to
*/
bool need_display = true;
int ret;
if (!i915_try_reset)
return (0);
if (!sx_try_xlock(&dev->dev_struct_lock))
return (-EBUSY);
i915_gem_reset(dev);
ret = -ENODEV;
if (time_second - dev_priv->last_gpu_reset < 5) {
DRM_ERROR("GPU hanging too fast, declaring wedged!\n");
} else
ret = intel_gpu_reset(dev);
dev_priv->last_gpu_reset = time_second;
if (ret) {
DRM_ERROR("Failed to reset chip.\n");
DRM_UNLOCK(dev);
return (ret);
}
if (drm_core_check_feature(dev, DRIVER_MODESET) ||
!dev_priv->mm.suspended) {
dev_priv->mm.suspended = 0;
i915_gem_init_swizzling(dev);
dev_priv->rings[RCS].init(&dev_priv->rings[RCS]);
if (HAS_BSD(dev))
dev_priv->rings[VCS].init(&dev_priv->rings[VCS]);
if (HAS_BLT(dev))
dev_priv->rings[BCS].init(&dev_priv->rings[BCS]);
i915_gem_context_init(dev);
i915_gem_init_ppgtt(dev);
drm_irq_uninstall(dev);
drm_mode_config_reset(dev);
DRM_UNLOCK(dev);
drm_irq_install(dev);
DRM_LOCK(dev);
}
DRM_UNLOCK(dev);
if (need_display) {
sx_xlock(&dev->mode_config.mutex);
drm_helper_resume_force_mode(dev);
sx_xunlock(&dev->mode_config.mutex);
}
return (0);
}
/**
* i915_reset - reset chip after a hang
* @dev: drm device to reset
*
* Reset the chip. Useful if a hang is detected. Returns zero on successful
* reset or otherwise an error code.
*
* Procedure is fairly simple:
* - reset the chip using the reset reg
* - re-init context state
* - re-init hardware status page
* - re-init ring buffer
* - re-init interrupt state
* - re-init display
*/
int i915_reset(struct drm_device *dev)
{
struct drm_i915_private *dev_priv = dev->dev_private;
bool simulated;
int ret;
if (!i915.reset)
return 0;
mutex_lock(&dev->struct_mutex);
i915_gem_reset(dev);
simulated = dev_priv->gpu_error.stop_rings != 0;
ret = intel_gpu_reset(dev);
/* Also reset the gpu hangman. */
if (simulated) {
DRM_INFO("Simulated gpu hang, resetting stop_rings\n");
dev_priv->gpu_error.stop_rings = 0;
if (ret == -ENODEV) {
DRM_INFO("Reset not implemented, but ignoring "
"error for simulated gpu hangs\n");
ret = 0;
}
}
if (ret) {
DRM_ERROR("Failed to reset chip: %i\n", ret);
mutex_unlock(&dev->struct_mutex);
return ret;
}
/* Ok, now get things going again... */
/*
* Everything depends on having the GTT running, so we need to start
* there. Fortunately we don't need to do this unless we reset the
* chip at a PCI level.
*
* Next we need to restore the context, but we don't use those
* yet either...
*
* Ring buffer needs to be re-initialized in the KMS case, or if X
* was running at the time of the reset (i.e. we weren't VT
* switched away).
*/
if (drm_core_check_feature(dev, DRIVER_MODESET) ||
!dev_priv->ums.mm_suspended) {
dev_priv->ums.mm_suspended = 0;
ret = i915_gem_init_hw(dev);
mutex_unlock(&dev->struct_mutex);
if (ret) {
DRM_ERROR("Failed hw init on reset %d\n", ret);
return ret;
}
/*
* FIXME: This races pretty badly against concurrent holders of
* ring interrupts. This is possible since we've started to drop
* dev->struct_mutex in select places when waiting for the gpu.
*/
/*
* rps/rc6 re-init is necessary to restore state lost after the
* reset and the re-install of gt irqs. Skip for ironlake per
* previous concerns that it doesn't respond well to some forms
* of re-init after reset.
*/
if (INTEL_INFO(dev)->gen > 5)
intel_reset_gt_powersave(dev);
intel_hpd_init(dev);
} else {
mutex_unlock(&dev->struct_mutex);
}
return 0;
}
