Re: [PATCH 04/12] drm/ttm: add common accounting to the resource mgr

[Daniel Vetter <daniel@ffwll.ch>]

On Tue, Aug 31, 2021 at 02:57:05PM +0200, Christian König wrote:
> Am 31.08.21 um 14:52 schrieb Daniel Vetter:
> > On Mon, Aug 30, 2021 at 10:56:59AM +0200, Christian König wrote:
> > > It makes sense to have this in the common manager for debugging and
> > > accounting of how much resources are used.
> > > 
> > > Signed-off-by: Christian König <christian.koenig@amd.com>
> > > ---
> > >   drivers/gpu/drm/ttm/ttm_resource.c |  8 ++++++++
> > >   include/drm/ttm/ttm_resource.h     | 18 ++++++++++++++++++
> > >   2 files changed, 26 insertions(+)
> > > 
> > > diff --git a/drivers/gpu/drm/ttm/ttm_resource.c b/drivers/gpu/drm/ttm/ttm_resource.c
> > > index a4c495da0040..426e6841fc89 100644
> > > --- a/drivers/gpu/drm/ttm/ttm_resource.c
> > > +++ b/drivers/gpu/drm/ttm/ttm_resource.c
> > > @@ -33,6 +33,8 @@ void ttm_resource_init(struct ttm_buffer_object *bo,
> > >                          const struct ttm_place *place,
> > >                          struct ttm_resource *res)
> > >   {
> > > +	struct ttm_resource_manager *man;
> > > +
> > >   	res->start = 0;
> > >   	res->num_pages = PFN_UP(bo->base.size);
> > >   	res->mem_type = place->mem_type;
> > > @@ -42,12 +44,16 @@ void ttm_resource_init(struct ttm_buffer_object *bo,
> > >   	res->bus.is_iomem = false;
> > >   	res->bus.caching = ttm_cached;
> > >   	res->bo = bo;
> > > +
> > > +	man = ttm_manager_type(bo->bdev, place->mem_type);
> > > +	atomic64_add(bo->base.size, &man->usage);
> > >   }
> > >   EXPORT_SYMBOL(ttm_resource_init);
> > >   void ttm_resource_fini(struct ttm_resource_manager *man,
> > >   		       struct ttm_resource *res)
> > >   {
> > > +	atomic64_sub(res->bo->base.size, &man->usage);
> > >   }
> > >   EXPORT_SYMBOL(ttm_resource_fini);
> > > @@ -100,6 +106,7 @@ void ttm_resource_manager_init(struct ttm_resource_manager *man,
> > >   	spin_lock_init(&man->move_lock);
> > >   	man->bdev = bdev;
> > >   	man->size = p_size;
> > > +	atomic64_set(&man->usage, 0);
> > >   	for (i = 0; i < TTM_MAX_BO_PRIORITY; ++i)
> > >   		INIT_LIST_HEAD(&man->lru[i]);
> > > @@ -172,6 +179,7 @@ void ttm_resource_manager_debug(struct ttm_resource_manager *man,
> > >   	drm_printf(p, "  use_type: %d\n", man->use_type);
> > >   	drm_printf(p, "  use_tt: %d\n", man->use_tt);
> > >   	drm_printf(p, "  size: %llu\n", man->size);
> > > +	drm_printf(p, "  usage: %llu\n", atomic64_read(&man->usage));
> > >   	if (man->func->debug)
> > >   		man->func->debug(man, p);
> > >   }
> > > diff --git a/include/drm/ttm/ttm_resource.h b/include/drm/ttm/ttm_resource.h
> > > index e8080192cae4..526fe359c603 100644
> > > --- a/include/drm/ttm/ttm_resource.h
> > > +++ b/include/drm/ttm/ttm_resource.h
> > > @@ -27,6 +27,7 @@
> > >   #include <linux/types.h>
> > >   #include <linux/mutex.h>
> > > +#include <linux/atomic.h>
> > >   #include <linux/dma-buf-map.h>
> > >   #include <linux/dma-fence.h>
> > >   #include <drm/drm_print.h>
> > > @@ -110,6 +111,7 @@ struct ttm_resource_manager_func {
> > >    * static information. bdev::driver::io_mem_free is never used.
> > >    * @lru: The lru list for this memory type.
> > >    * @move: The fence of the last pipelined move operation.
> > > + * @usage: How much of the region is used.
> > >    *
> > >    * This structure is used to identify and manage memory types for a device.
> > >    */
> > > @@ -134,6 +136,9 @@ struct ttm_resource_manager {
> > >   	 * Protected by @move_lock.
> > >   	 */
> > >   	struct dma_fence *move;
> > > +
> > > +	/* Own protection */

Please document struct members with the inline kerneldoc style, that way
the comment here shows up in the kerneldoc too.

Would be good to just convert the entire struct I think.

> > > +	atomic64_t usage;
> > Shouldn't we keep track of this together with the lru updates, under the
> > same spinlock?
> 
> Mhm, what should that be good for? As far as I know we use it for two use
> cases:
> 1. Early abort when size-usage < requested.

But doesn't have all kinds of potential temporary leaks again? Except this
time around we can't even fix them eventually, because these allocations
aren't on the lru list at all.

> 2. Statistics for debugging.
> 
> Especially the first use case is rather important under memory pressure to
> avoid costly acquiring of a contended lock.

Or maybe I'm just not understanding where this matters? Don't you need to
grab the lru lock anyway under memory pressure?

> > Otherwise this usage here just becomes kinda meaningless I think, and just
> > for some debugging. I really don't like sprinkling random atomic_t around
> > (mostly because i915-gem code has gone totally overboard with them, with
> > complete disregard to complexity of the result).
> 
> Well this here just replaces what drivers did anyway and the cost of inc/dec
> an atomic is pretty much negligible.

Complexity in understanding the locking. The cpu has not problem with this
stuff ofc.
-Daniel

> 
> Christian.
> 
> > -Daniel
> > 
> > >   };
> > >   /**
> > > @@ -260,6 +265,19 @@ ttm_resource_manager_cleanup(struct ttm_resource_manager *man)
> > >   	man->move = NULL;
> > >   }
> > > +/**
> > > + * ttm_resource_manager_usage
> > > + *
> > > + * @man: A memory manager object.
> > > + *
> > > + * Return how many resources are currently used.
> > > + */
> > > +static inline uint64_t
> > > +ttm_resource_manager_usage(struct ttm_resource_manager *man)
> > > +{
> > > +	return atomic64_read(&man->usage);
> > > +}
> > > +
> > >   void ttm_resource_init(struct ttm_buffer_object *bo,
> > >                          const struct ttm_place *place,
> > >                          struct ttm_resource *res);
> > > -- 
> > > 2.25.1
> > > 
> 

-- 
Daniel Vetter
Software Engineer, Intel Corporation
http://blog.ffwll.ch