[v4] drm/i915/lmem: debugfs for LMEM details
diff mbox series

Message ID 20191224084300.5435-1-ramalingam.c@intel.com
State New
Headers show
Series
  • [v4] drm/i915/lmem: debugfs for LMEM details
Related show

Commit Message

Ramalingam C Dec. 24, 2019, 8:43 a.m. UTC
From: Lukasz Fiedorowicz <lukasz.fiedorowicz@intel.com>

Debugfs i915_gem_object is extended to enable the IGTs to
detect the LMEM's availability and the total size of LMEM.

v2: READ_ONCE is used [Chris]
v3: %pa is used for printing the resource [Chris]
v4: All regions' details added to debugfs [Chris]

Signed-off-by: Lukasz Fiedorowicz <lukasz.fiedorowicz@intel.com>
Signed-off-by: Matthew Auld <matthew.auld@intel.com>
Signed-off-by: Stuart Summers <stuart.summers@intel.com>
Signed-off-by: Ramalingam C <ramalingam.c@intel.com>
Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
---
 drivers/gpu/drm/i915/i915_debugfs.c        | 12 +++++++++++-
 drivers/gpu/drm/i915/intel_memory_region.c | 19 ++++++++++++++++++-
 drivers/gpu/drm/i915/intel_memory_region.h |  4 ++++
 3 files changed, 33 insertions(+), 2 deletions(-)

Comments

Chris Wilson Dec. 27, 2019, 11 a.m. UTC | #1
Quoting Ramalingam C (2019-12-24 08:43:00)
> From: Lukasz Fiedorowicz <lukasz.fiedorowicz@intel.com>
> 
> Debugfs i915_gem_object is extended to enable the IGTs to
> detect the LMEM's availability and the total size of LMEM.
> 
> v2: READ_ONCE is used [Chris]
> v3: %pa is used for printing the resource [Chris]
> v4: All regions' details added to debugfs [Chris]
> 
> Signed-off-by: Lukasz Fiedorowicz <lukasz.fiedorowicz@intel.com>
> Signed-off-by: Matthew Auld <matthew.auld@intel.com>
> Signed-off-by: Stuart Summers <stuart.summers@intel.com>
> Signed-off-by: Ramalingam C <ramalingam.c@intel.com>
> Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
> ---
>  drivers/gpu/drm/i915/i915_debugfs.c        | 12 +++++++++++-
>  drivers/gpu/drm/i915/intel_memory_region.c | 19 ++++++++++++++++++-
>  drivers/gpu/drm/i915/intel_memory_region.h |  4 ++++
>  3 files changed, 33 insertions(+), 2 deletions(-)
> 
> diff --git a/drivers/gpu/drm/i915/i915_debugfs.c b/drivers/gpu/drm/i915/i915_debugfs.c
> index d28468eaed57..35f8db7690a5 100644
> --- a/drivers/gpu/drm/i915/i915_debugfs.c
> +++ b/drivers/gpu/drm/i915/i915_debugfs.c
> @@ -367,13 +367,23 @@ static void print_context_stats(struct seq_file *m,
>  static int i915_gem_object_info(struct seq_file *m, void *data)
>  {
>         struct drm_i915_private *i915 = node_to_i915(m->private);
> +       struct intel_memory_region *mem;
> +       enum intel_region_id id;
>  
>         seq_printf(m, "%u shrinkable [%u free] objects, %llu bytes\n",
>                    i915->mm.shrink_count,
>                    atomic_read(&i915->mm.free_count),
>                    i915->mm.shrink_memory);
>  
> -       seq_putc(m, '\n');

Keep the newline before the switch to context info and after the general
memory info.

> +       for (id = INTEL_REGION_SMEM; id < INTEL_REGION_UNKNOWN; id++) {

We don't have a for_each_memregion() yet?

#define for_each_memory_region(mr, i915, id) \
	for (id = 0; id < ARRAY_SIZE((i915)->mm.regions); id++) \
		for_each_if(mr = i915->mm.regions[id])

> +               mem = i915->mm.regions[id];
> +               if (!mem)
> +                       continue;
> +
> +               seq_printf(m, "%s total: %pa bytes, available %pa bytes\n",
> +                          intel_get_mem_region_name(mem), &mem->total,
> +                          &mem->avail);
> +       }
>  
>         print_context_stats(m, i915);
>  
> diff --git a/drivers/gpu/drm/i915/intel_memory_region.c b/drivers/gpu/drm/i915/intel_memory_region.c
> index e24c280e5930..824aa13e993d 100644
> --- a/drivers/gpu/drm/i915/intel_memory_region.c
> +++ b/drivers/gpu/drm/i915/intel_memory_region.c
> @@ -10,6 +10,20 @@
>  #define REGION_MAP(type, inst) \
>         BIT((type) + INTEL_MEMORY_TYPE_SHIFT) | BIT(inst)
>  
> +char *intel_get_mem_region_name(struct intel_memory_region *mem)

const char. But as this does not scale, just put a char name[8] into
intel_memory_region and populate it on init.

> +{
> +       switch (MEMORY_TYPE_FROM_REGION(mem->id)) {
> +       case INTEL_REGION_SMEM:
> +               return "SYSTEM MEM";

WHY ALL THE SHOUTING OF DIFFERING NAMES?

Just system, local, stolen would do.

> +       case INTEL_REGION_LMEM:
> +               return "LOCAL MEM";
> +       case INTEL_REGION_STOLEN:
> +               return "STOLEN MEM";
> +       default:
> +               return NULL;
> +       }
> +}

Patch
diff mbox series

diff --git a/drivers/gpu/drm/i915/i915_debugfs.c b/drivers/gpu/drm/i915/i915_debugfs.c
index d28468eaed57..35f8db7690a5 100644
--- a/drivers/gpu/drm/i915/i915_debugfs.c
+++ b/drivers/gpu/drm/i915/i915_debugfs.c
@@ -367,13 +367,23 @@  static void print_context_stats(struct seq_file *m,
 static int i915_gem_object_info(struct seq_file *m, void *data)
 {
 	struct drm_i915_private *i915 = node_to_i915(m->private);
+	struct intel_memory_region *mem;
+	enum intel_region_id id;
 
 	seq_printf(m, "%u shrinkable [%u free] objects, %llu bytes\n",
 		   i915->mm.shrink_count,
 		   atomic_read(&i915->mm.free_count),
 		   i915->mm.shrink_memory);
 
-	seq_putc(m, '\n');
+	for (id = INTEL_REGION_SMEM; id < INTEL_REGION_UNKNOWN; id++) {
+		mem = i915->mm.regions[id];
+		if (!mem)
+			continue;
+
+		seq_printf(m, "%s total: %pa bytes, available %pa bytes\n",
+			   intel_get_mem_region_name(mem), &mem->total,
+			   &mem->avail);
+	}
 
 	print_context_stats(m, i915);
 
diff --git a/drivers/gpu/drm/i915/intel_memory_region.c b/drivers/gpu/drm/i915/intel_memory_region.c
index e24c280e5930..824aa13e993d 100644
--- a/drivers/gpu/drm/i915/intel_memory_region.c
+++ b/drivers/gpu/drm/i915/intel_memory_region.c
@@ -10,6 +10,20 @@ 
 #define REGION_MAP(type, inst) \
 	BIT((type) + INTEL_MEMORY_TYPE_SHIFT) | BIT(inst)
 
+char *intel_get_mem_region_name(struct intel_memory_region *mem)
+{
+	switch (MEMORY_TYPE_FROM_REGION(mem->id)) {
+	case INTEL_REGION_SMEM:
+		return "SYSTEM MEM";
+	case INTEL_REGION_LMEM:
+		return "LOCAL MEM";
+	case INTEL_REGION_STOLEN:
+		return "STOLEN MEM";
+	default:
+		return NULL;
+	}
+}
+
 const u32 intel_region_map[] = {
 	[INTEL_REGION_SMEM] = REGION_MAP(INTEL_MEMORY_SYSTEM, 0),
 	[INTEL_REGION_LMEM] = REGION_MAP(INTEL_MEMORY_LOCAL, 0),
@@ -37,7 +51,7 @@  __intel_memory_region_put_pages_buddy(struct intel_memory_region *mem,
 				      struct list_head *blocks)
 {
 	mutex_lock(&mem->mm_lock);
-	intel_memory_region_free_pages(mem, blocks);
+	mem->avail += intel_memory_region_free_pages(mem, blocks);
 	mutex_unlock(&mem->mm_lock);
 }
 
@@ -106,6 +120,7 @@  __intel_memory_region_get_pages_buddy(struct intel_memory_region *mem,
 			break;
 	} while (1);
 
+	mem->avail -= size;
 	mutex_unlock(&mem->mm_lock);
 	return 0;
 
@@ -164,6 +179,8 @@  intel_memory_region_create(struct drm_i915_private *i915,
 	mem->io_start = io_start;
 	mem->min_page_size = min_page_size;
 	mem->ops = ops;
+	mem->total = size;
+	mem->avail = mem->total;
 
 	mutex_init(&mem->objects.lock);
 	INIT_LIST_HEAD(&mem->objects.list);
diff --git a/drivers/gpu/drm/i915/intel_memory_region.h b/drivers/gpu/drm/i915/intel_memory_region.h
index 238722009677..c312f93fc94d 100644
--- a/drivers/gpu/drm/i915/intel_memory_region.h
+++ b/drivers/gpu/drm/i915/intel_memory_region.h
@@ -82,6 +82,8 @@  struct intel_memory_region {
 
 	resource_size_t io_start;
 	resource_size_t min_page_size;
+	resource_size_t total;
+	resource_size_t avail;
 
 	unsigned int type;
 	unsigned int instance;
@@ -96,6 +98,8 @@  struct intel_memory_region {
 	} objects;
 };
 
+char *intel_get_mem_region_name(struct intel_memory_region *mem);
+
 int intel_memory_region_init_buddy(struct intel_memory_region *mem);
 void intel_memory_region_release_buddy(struct intel_memory_region *mem);