public inbox for intel-gfx@lists.freedesktop.org
 help / color / mirror / Atom feed
From: Ben Widawsky <ben@bwidawsk.net>
To: Daniel Vetter <daniel@ffwll.ch>
Cc: Dave Airlie <airlied@redhat.com>, intel-gfx@lists.freedesktop.org
Subject: Re: [PATCH 11/24] drm: Introduce an iterator over holes in the drm_mm range manager
Date: Thu, 25 Oct 2012 13:54:42 -0700	[thread overview]
Message-ID: <20121025135442.4d1b0850@bwidawsk.net> (raw)
In-Reply-To: <20121025192225.GE5691@phenom.ffwll.local>

On Thu, 25 Oct 2012 21:22:25 +0200
Daniel Vetter <daniel@ffwll.ch> wrote:

> On Thu, Aug 30, 2012 at 04:31:06PM +0100, Chris Wilson wrote:
> > This will be used i915 in forthcoming patches in order to measure the
> > largest contiguous chunk of memory available for enabling chipset
> > features.
> > 
> > Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
> > Cc: Dave Airlie <airlied@redhat.com>
> 
> While discussing the iterator in private a bit with Ben I've noticed that
> you've ignore my bikeshed! tsk ... ;-)
> 
> http://lists.freedesktop.org/archives/intel-gfx/2012-September/020436.html
> 
> -Daniel

While the macro is undeniably horendous to read, I think Daniel's
suggestion makes it slightly better simply for reading.

> 
> > ---
> >  drivers/gpu/drm/drm_mm.c |   55 +++++++++++++++-------------------------------
> >  include/drm/drm_mm.h     |   26 ++++++++++++++++++++++
> >  2 files changed, 44 insertions(+), 37 deletions(-)
> > 
> > diff --git a/drivers/gpu/drm/drm_mm.c b/drivers/gpu/drm/drm_mm.c
> > index 5db8c20..c3d11ec 100644
> > --- a/drivers/gpu/drm/drm_mm.c
> > +++ b/drivers/gpu/drm/drm_mm.c
> > @@ -102,20 +102,6 @@ int drm_mm_pre_get(struct drm_mm *mm)
> >  }
> >  EXPORT_SYMBOL(drm_mm_pre_get);
> >  
> > -static inline unsigned long drm_mm_hole_node_start(struct drm_mm_node *hole_node)
> > -{
> > -	return hole_node->start + hole_node->size;
> > -}
> > -
> > -static inline unsigned long drm_mm_hole_node_end(struct drm_mm_node *hole_node)
> > -{
> > -	struct drm_mm_node *next_node =
> > -		list_entry(hole_node->node_list.next, struct drm_mm_node,
> > -			   node_list);
> > -
> > -	return next_node->start;
> > -}
> > -
> >  static void drm_mm_insert_helper(struct drm_mm_node *hole_node,
> >  				 struct drm_mm_node *node,
> >  				 unsigned long size, unsigned alignment,
> > @@ -127,7 +113,7 @@ static void drm_mm_insert_helper(struct drm_mm_node *hole_node,
> >  	unsigned long adj_start = hole_start;
> >  	unsigned long adj_end = hole_end;
> >  
> > -	BUG_ON(!hole_node->hole_follows || node->allocated);
> > +	BUG_ON(node->allocated);
> >  
> >  	if (mm->color_adjust)
> >  		mm->color_adjust(hole_node, color, &adj_start, &adj_end);
> > @@ -168,15 +154,10 @@ struct drm_mm_node *drm_mm_create_block(struct drm_mm *mm,
> >  {
> >  	struct drm_mm_node *hole, *node;
> >  	unsigned long end = start + size;
> > +	unsigned long hole_start;
> > +	unsigned long hole_end;
> >  
> > -	list_for_each_entry(hole, &mm->hole_stack, hole_stack) {
> > -		unsigned long hole_start;
> > -		unsigned long hole_end;
> > -
> > -		BUG_ON(!hole->hole_follows);
> > -		hole_start = drm_mm_hole_node_start(hole);
> > -		hole_end = drm_mm_hole_node_end(hole);
> > -
> > +	drm_mm_for_each_hole(hole, mm, hole_start, hole_end) {
> >  		if (hole_start > start || hole_end < end)
> >  			continue;
> >  
> > @@ -361,8 +342,10 @@ void drm_mm_remove_node(struct drm_mm_node *node)
> >  				== drm_mm_hole_node_end(node));
> >  		list_del(&node->hole_stack);
> >  	} else
> > -		BUG_ON(drm_mm_hole_node_start(node)
> > -				!= drm_mm_hole_node_end(node));
> > +		BUG_ON(node->start + node->size !=
> > +		       list_entry(node->node_list.next,
> > +				  struct drm_mm_node, node_list)->start);
> > +
> >  
> >  	if (!prev_node->hole_follows) {
> >  		prev_node->hole_follows = 1;
> > @@ -420,6 +403,8 @@ struct drm_mm_node *drm_mm_search_free_generic(const struct drm_mm *mm,
> >  {
> >  	struct drm_mm_node *entry;
> >  	struct drm_mm_node *best;
> > +	unsigned long adj_start;
> > +	unsigned long adj_end;
> >  	unsigned long best_size;
> >  
> >  	BUG_ON(mm->scanned_blocks);
> > @@ -427,17 +412,13 @@ struct drm_mm_node *drm_mm_search_free_generic(const struct drm_mm *mm,
> >  	best = NULL;
> >  	best_size = ~0UL;
> >  
> > -	list_for_each_entry(entry, &mm->hole_stack, hole_stack) {
> > -		unsigned long adj_start = drm_mm_hole_node_start(entry);
> > -		unsigned long adj_end = drm_mm_hole_node_end(entry);
> > -
> > +	drm_mm_for_each_hole(entry, mm, adj_start, adj_end) {
> >  		if (mm->color_adjust) {
> >  			mm->color_adjust(entry, color, &adj_start, &adj_end);
> >  			if (adj_end <= adj_start)
> >  				continue;
> >  		}
> >  
> > -		BUG_ON(!entry->hole_follows);
> >  		if (!check_free_hole(adj_start, adj_end, size, alignment))
> >  			continue;
> >  
> > @@ -464,6 +445,8 @@ struct drm_mm_node *drm_mm_search_free_in_range_generic(const struct drm_mm *mm,
> >  {
> >  	struct drm_mm_node *entry;
> >  	struct drm_mm_node *best;
> > +	unsigned long adj_start;
> > +	unsigned long adj_end;
> >  	unsigned long best_size;
> >  
> >  	BUG_ON(mm->scanned_blocks);
> > @@ -471,13 +454,11 @@ struct drm_mm_node *drm_mm_search_free_in_range_generic(const struct drm_mm *mm,
> >  	best = NULL;
> >  	best_size = ~0UL;
> >  
> > -	list_for_each_entry(entry, &mm->hole_stack, hole_stack) {
> > -		unsigned long adj_start = drm_mm_hole_node_start(entry) < start ?
> > -			start : drm_mm_hole_node_start(entry);
> > -		unsigned long adj_end = drm_mm_hole_node_end(entry) > end ?
> > -			end : drm_mm_hole_node_end(entry);
> > -
> > -		BUG_ON(!entry->hole_follows);
> > +	drm_mm_for_each_hole(entry, mm, adj_start, adj_end) {
> > +		if (adj_start < start)
> > +			adj_start = start;
> > +		if (adj_end > end)
> > +			adj_end = end;
> >  
> >  		if (mm->color_adjust) {
> >  			mm->color_adjust(entry, color, &adj_start, &adj_end);
> > diff --git a/include/drm/drm_mm.h b/include/drm/drm_mm.h
> > index 4020f96..d710a10 100644
> > --- a/include/drm/drm_mm.h
> > +++ b/include/drm/drm_mm.h
> > @@ -89,6 +89,19 @@ static inline bool drm_mm_initialized(struct drm_mm *mm)
> >  {
> >  	return mm->hole_stack.next;
> >  }
> > +
> > +static inline unsigned long drm_mm_hole_node_start(struct drm_mm_node *hole_node)
> > +{
> > +	BUG_ON(!hole_node->hole_follows);
> > +	return hole_node->start + hole_node->size;
> > +}
> > +
> > +static inline unsigned long drm_mm_hole_node_end(struct drm_mm_node *hole_node)
> > +{
> > +	return list_entry(hole_node->node_list.next,
> > +			  struct drm_mm_node, node_list)->start;
> > +}
> > +
> >  #define drm_mm_for_each_node(entry, mm) list_for_each_entry(entry, \
> >  						&(mm)->head_node.node_list, \
> >  						node_list)
> > @@ -99,6 +112,19 @@ static inline bool drm_mm_initialized(struct drm_mm *mm)
> >  	     entry != NULL; entry = next, \
> >  		next = entry ? list_entry(entry->node_list.next, \
> >  			struct drm_mm_node, node_list) : NULL) \
> > +
> > +/* Note that we need to unroll list_for_each_entry in order to inline
> > + * setting hole_start and hole_end on each iteration and keep the
> > + * macro sane.
> > + */
> > +#define drm_mm_for_each_hole(entry, mm, hole_start, hole_end) \
> > +	for (entry = list_entry((mm)->hole_stack.next, typeof(struct drm_mm_node), hole_stack); \
> > +	     &entry->hole_stack != &(mm)->hole_stack ? \
> > +	     hole_start = drm_mm_hole_node_start(entry), \
> > +	     hole_end = drm_mm_hole_node_end(entry) : \
> > +	     0; \
> > +	     entry = list_entry(entry->hole_stack.next, typeof(struct drm_mm_node), hole_stack))
> > +
> >  /*
> >   * Basic range manager support (drm_mm.c)
> >   */
> > -- 
> > 1.7.10.4
> > 
> > _______________________________________________
> > Intel-gfx mailing list
> > Intel-gfx@lists.freedesktop.org
> > http://lists.freedesktop.org/mailman/listinfo/intel-gfx
> 



-- 
Ben Widawsky, Intel Open Source Technology Center

  reply	other threads:[~2012-10-25 20:54 UTC|newest]

Thread overview: 32+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2012-08-30 15:30 Next iteration of stolen support Chris Wilson
2012-08-30 15:30 ` [PATCH 01/24] drm/i915: Introduce drm_i915_gem_object_ops Chris Wilson
2012-08-30 15:30 ` [PATCH 02/24] drm/i915: Pin backing pages whilst exporting through a dmabuf vmap Chris Wilson
2012-08-30 15:30 ` [PATCH 03/24] drm/i915: Pin backing pages for pwrite Chris Wilson
2012-08-30 15:30 ` [PATCH 04/24] drm/i915: Pin backing pages for pread Chris Wilson
2012-08-30 15:31 ` [PATCH 05/24] drm/i915: Replace the array of pages with a scatterlist Chris Wilson
2012-08-30 15:31 ` [PATCH 06/24] drm/i915: Convert the dmabuf object to use the new i915_gem_object_ops Chris Wilson
2012-08-30 15:31 ` [PATCH 07/24] drm: Introduce drm_mm_create_block() Chris Wilson
2012-08-30 15:31 ` [PATCH 08/24] drm/i915: Fix detection of stolen base for gen2 Chris Wilson
2012-08-30 15:31 ` [PATCH 09/24] drm/i915: Fix location of stolen memory register for SandyBridge+ Chris Wilson
2012-08-30 15:31 ` [PATCH 10/24] drm/i915: Avoid clearing preallocated regions from the GTT Chris Wilson
2012-08-30 15:31 ` [PATCH 11/24] drm: Introduce an iterator over holes in the drm_mm range manager Chris Wilson
2012-10-25 19:22   ` Daniel Vetter
2012-10-25 20:54     ` Ben Widawsky [this message]
2012-10-26 10:55       ` Chris Wilson
2012-08-30 15:31 ` [PATCH 12/24] drm/i915: Delay allocation of stolen space for FBC Chris Wilson
2012-08-30 15:31 ` [PATCH 13/24] drm/i915: Defer allocation of stolen memory for FBC until first use Chris Wilson
2012-08-30 15:31 ` [PATCH 14/24] drm/i915: Allow objects to be created with no backing pages, but stolen space Chris Wilson
2012-08-30 15:31 ` [PATCH 15/24] drm/i915: Differentiate between prime and stolen objects Chris Wilson
2012-08-30 15:31 ` [PATCH 16/24] drm/i915: Support readback of stolen objects upon error Chris Wilson
2012-08-30 15:31 ` [PATCH 17/24] drm/i915: Handle stolen objects in pwrite Chris Wilson
2012-08-30 15:31 ` [PATCH 18/24] drm/i915: Handle stolen objects for pread Chris Wilson
2012-08-30 15:31 ` [PATCH 19/24] drm/i915: Introduce i915_gem_object_create_stolen() Chris Wilson
2012-08-30 15:31 ` [PATCH 20/24] drm/i915: Allocate fbcon from stolen memory Chris Wilson
2012-08-30 15:31 ` [PATCH 21/24] drm/i915: Allocate ringbuffers " Chris Wilson
2012-08-30 15:31 ` [PATCH 22/24] drm/i915: Allocate overlay registers " Chris Wilson
2012-08-30 15:31 ` [PATCH 23/24] drm/i915: Use a slab for object allocation Chris Wilson
2012-08-30 18:50   ` Ben Widawsky
2012-08-30 15:31 ` [PATCH 24/24] drm/i915: Introduce mapping of user pages into video memory (userptr) ioctl Chris Wilson
2012-08-30 16:22 ` Next iteration of stolen support Chris Wilson
  -- strict thread matches above, loose matches on Subject: below --
2012-09-04 20:02 Stolen memory, again Chris Wilson
2012-09-04 20:03 ` [PATCH 11/24] drm: Introduce an iterator over holes in the drm_mm range manager Chris Wilson
2012-09-12 13:54   ` Daniel Vetter

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20121025135442.4d1b0850@bwidawsk.net \
    --to=ben@bwidawsk.net \
    --cc=airlied@redhat.com \
    --cc=daniel@ffwll.ch \
    --cc=intel-gfx@lists.freedesktop.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox