[Intel-gfx] [RFC PATCH 2/3] drm/i915/perf: allow for CS OA configs to be created lazily

Chris Wilson chris at chris-wilson.co.uk
Mon Oct 8 15:34:40 UTC 2018


Quoting Lionel Landwerlin (2018-10-08 16:18:21)
> Here we introduce a mechanism by which the execbuf part of the i915
> driver will be able to request that a batch buffer containing the
> programming for a particular OA config be created.
> 
> We'll execute these OA configuration buffers right before executing a
> set of userspace commands so that a particular user batchbuffer be
> executed with a given OA configuration.
> 
> This mechanism essentially allows the userspace driver to go through
> several OA configuration without having to open/close the i915/perf
> stream.
> 
> Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin at intel.com>
> ---
>  drivers/gpu/drm/i915/i915_drv.h           |  22 ++-
>  drivers/gpu/drm/i915/i915_perf.c          | 195 ++++++++++++++++++----
>  drivers/gpu/drm/i915/intel_gpu_commands.h |   1 +
>  3 files changed, 187 insertions(+), 31 deletions(-)
> 
> diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h
> index 2264b30ce51a..a35715cd7608 100644
> --- a/drivers/gpu/drm/i915/i915_drv.h
> +++ b/drivers/gpu/drm/i915/i915_drv.h
> @@ -1378,6 +1378,10 @@ struct i915_oa_config {
>         struct attribute *attrs[2];
>         struct device_attribute sysfs_metric_id;
>  
> +       struct i915_vma *vma;
> +
> +       struct list_head vma_link;
> +
>         atomic_t ref_count;
>  };
>  
> @@ -1979,11 +1983,21 @@ struct drm_i915_private {
>                 struct mutex metrics_lock;
>  
>                 /*
> -                * List of dynamic configurations, you need to hold
> -                * dev_priv->perf.metrics_lock to access it.
> +                * List of dynamic configurations (struct i915_oa_config), you
> +                * need to hold dev_priv->perf.metrics_lock to access it.
>                  */
>                 struct idr metrics_idr;
>  
> +               /*
> +                * List of dynamic configurations (struct i915_oa_config)
> +                * which have an allocated buffer in GGTT for reconfiguration,
> +                * you need to hold dev_priv->perf.metrics_lock to access it.
> +                * Elements are added to the list lazilly on execbuf (when a
> +                * particular configuration is requested). The list is freed
> +                * upon closing the perf stream.
> +                */
> +               struct list_head metrics_buffers;
> +
>                 /*
>                  * Lock associated with anything below within this structure
>                  * except exclusive_stream.
> @@ -3315,6 +3329,10 @@ int i915_perf_remove_config_ioctl(struct drm_device *dev, void *data,
>  void i915_oa_init_reg_state(struct intel_engine_cs *engine,
>                             struct i915_gem_context *ctx,
>                             uint32_t *reg_state);
> +int i915_perf_get_oa_config(struct drm_i915_private *i915,
> +                           int metrics_set,
> +                           struct i915_oa_config **out_config,
> +                           struct i915_vma **out_vma);
>  
>  /* i915_gem_evict.c */
>  int __must_check i915_gem_evict_something(struct i915_address_space *vm,
> diff --git a/drivers/gpu/drm/i915/i915_perf.c b/drivers/gpu/drm/i915/i915_perf.c
> index e2a96b6844fe..39c5b44862d4 100644
> --- a/drivers/gpu/drm/i915/i915_perf.c
> +++ b/drivers/gpu/drm/i915/i915_perf.c
> @@ -364,9 +364,16 @@ struct perf_open_properties {
>         int oa_period_exponent;
>  };
>  
> -static void free_oa_config(struct drm_i915_private *dev_priv,
> -                          struct i915_oa_config *oa_config)
> +static void put_oa_config(struct i915_oa_config *oa_config)
>  {
> +       if (!atomic_dec_and_test(&oa_config->ref_count))
> +               return;
> +
> +       if (oa_config->vma) {
> +               list_del(&oa_config->vma_link);
> +               i915_vma_put(oa_config->vma);
> +       }
> +
>         if (!PTR_ERR(oa_config->flex_regs))
>                 kfree(oa_config->flex_regs);
>         if (!PTR_ERR(oa_config->b_counter_regs))
> @@ -376,38 +383,152 @@ static void free_oa_config(struct drm_i915_private *dev_priv,
>         kfree(oa_config);
>  }
>  
> -static void put_oa_config(struct drm_i915_private *dev_priv,
> -                         struct i915_oa_config *oa_config)
> +static u32 *write_cs_mi_lri(u32 *cs, const struct i915_oa_reg *reg_data, u32 n_regs)
>  {
> -       if (!atomic_dec_and_test(&oa_config->ref_count))
> -               return;
> +       u32 i;
>  
> -       free_oa_config(dev_priv, oa_config);
> +       for (i = 0; i < n_regs; i++) {
> +               if ((i % MI_LOAD_REGISTER_IMM_MAX_REGS) == 0) {
> +                       u32 n_lri = min(n_regs - i,
> +                                       (u32) MI_LOAD_REGISTER_IMM_MAX_REGS);
> +
> +                       *cs++ = MI_LOAD_REGISTER_IMM(n_lri);
> +               }
> +               *cs++ = i915_mmio_reg_offset(reg_data[i].addr);
> +               *cs++ = reg_data[i].value;
> +       }
> +
> +       return cs;
>  }
>  
> -static int get_oa_config(struct drm_i915_private *dev_priv,
> -                        int metrics_set,
> -                        struct i915_oa_config **out_config)
> +static int alloc_oa_config_buffer(struct drm_i915_private *i915,
> +                                 struct i915_oa_config *oa_config)
>  {
> +       struct drm_i915_gem_object *bo;
> +       size_t config_length = 0;
>         int ret;
> +       u32 *cs;
>  
> -       if (metrics_set == 1) {
> -               *out_config = &dev_priv->perf.oa.test_config;
> -               atomic_inc(&dev_priv->perf.oa.test_config.ref_count);
> -               return 0;
> +       if (oa_config->mux_regs_len > 0) {
> +               config_length += DIV_ROUND_UP(oa_config->mux_regs_len,
> +                                             MI_LOAD_REGISTER_IMM_MAX_REGS) * 4;
> +               config_length += oa_config->mux_regs_len * 8;
>         }
> +       if (oa_config->b_counter_regs_len > 0) {
> +               config_length += DIV_ROUND_UP(oa_config->b_counter_regs_len,
> +                                             MI_LOAD_REGISTER_IMM_MAX_REGS) * 4;
> +               config_length += oa_config->b_counter_regs_len * 8;
> +       }
> +       if (oa_config->flex_regs_len > 0) {
> +               config_length += DIV_ROUND_UP(oa_config->flex_regs_len,
> +                                             MI_LOAD_REGISTER_IMM_MAX_REGS) * 4;
> +               config_length += oa_config->flex_regs_len * 8;
> +       }
> +       config_length += 4; /* MI_BATCH_BUFFER_END */
> +       config_length = ALIGN(config_length, I915_GTT_PAGE_SIZE);
>  
> -       ret = mutex_lock_interruptible(&dev_priv->perf.metrics_lock);
> +       ret = i915_mutex_lock_interruptible(&i915->drm);
>         if (ret)
>                 return ret;
>  
> -       *out_config = idr_find(&dev_priv->perf.metrics_idr, metrics_set);
> -       if (!*out_config)
> -               ret = -EINVAL;
> -       else
> -               atomic_inc(&(*out_config)->ref_count);
> +       bo = i915_gem_object_create(i915, config_length);
> +       if (IS_ERR(bo)) {
> +               ret = PTR_ERR(bo);
> +               goto unlock;
> +       }
> +
> +       ret = i915_gem_object_set_cache_level(bo, I915_CACHE_LLC);

Don't enable snoop on a batchbuffer.

> +       if (ret)
> +               goto err_unref;
>  
> -       mutex_unlock(&dev_priv->perf.metrics_lock);
> +       oa_config->vma = i915_gem_object_ggtt_pin(bo, NULL, 0, config_length, 0);

Why have you pinned it?

> +       if (IS_ERR(oa_config->vma)) {
> +               ret = PTR_ERR(oa_config->vma);
> +               oa_config->vma = NULL;
> +               goto err_unref;
> +       }
> +
> +       cs = i915_gem_object_pin_map(bo, I915_MAP_WB);
> +       if (IS_ERR(cs)) {
> +               ret = PTR_ERR(cs);
> +               goto err_unpin;
> +       }
> +
> +       memset(cs, 0, config_length);

Already zero. Or use create_internal to avoid shmemfs overhead. And
since you write all bytes, you can just ignore it.

> +       cs = write_cs_mi_lri(cs, oa_config->mux_regs, oa_config->mux_regs_len);
> +       cs = write_cs_mi_lri(cs, oa_config->b_counter_regs, oa_config->b_counter_regs_len);
> +       cs = write_cs_mi_lri(cs, oa_config->flex_regs, oa_config->flex_regs_len);
> +
> +       *cs++ = MI_BATCH_BUFFER_END;
> +
> +       i915_gem_object_unpin_map(bo);
> +
> +       goto unlock;
> +
> +err_unpin:
> +       __i915_vma_unpin(oa_config->vma);
> +
> +err_unref:
> +       oa_config->vma = NULL;
> +       i915_gem_object_put(bo);
> +
> +unlock:
> +       mutex_unlock(&i915->drm.struct_mutex);
> +       return ret;
> +}
> +
> +int i915_perf_get_oa_config(struct drm_i915_private *i915,
> +                           int metrics_set,
> +                           struct i915_oa_config **out_config,
> +                           struct i915_vma **out_vma)
> +{
> +       int ret = 0;
> +       struct i915_oa_config *oa_config;
> +
> +       if (!i915->perf.initialized)
> +               return -ENODEV;
> +
> +       ret = mutex_lock_interruptible(&i915->perf.metrics_lock);
> +       if (ret)
> +               return ret;
> +
> +       if (metrics_set == 1) {
> +               oa_config = &i915->perf.oa.test_config;
> +       } else {
> +               oa_config = idr_find(&i915->perf.metrics_idr, metrics_set);
> +               if (!oa_config) {
> +                       ret = -EINVAL;
> +                       goto unlock;
> +               }
> +       }
> +
> +       if (out_config) {
> +               atomic_inc(&oa_config->ref_count);
> +               *out_config = oa_config;
> +       }
> +
> +       if (out_vma) {
> +               if (oa_config->vma) {
> +                       *out_vma = i915_vma_get(oa_config->vma);
> +               } else {
> +                       ret = alloc_oa_config_buffer(i915, oa_config);
> +                       if (ret) {
> +                               goto err_buf_alloc;
> +                       } else {
> +                               list_add(&oa_config->vma_link,
> +                                        &i915->perf.metrics_buffers);
> +                               *out_vma = i915_vma_get(oa_config->vma);
> +                       }
> +               }

Where is out_vma used so we can check if the litetime tracking is ok as
so far you are releasing it before we know it is idle.
-Chris


More information about the Intel-gfx mailing list