APPLIED: [PATCH][Lucid SRU] drm: mm: fix range restricted allocations

[Tim Gardner]

On 10/14/2011 04:05 PM, Seth Forshee wrote:
> From: Daniel Vetter<daniel.vetter at ffwll.ch>
>
> With the code cleanup in
>
> 7a6b2896f261894dde287d3faefa4b432cddca53 is the first bad commit
> commit 7a6b2896f261894dde287d3faefa4b432cddca53
> Author: Daniel Vetter<daniel.vetter at ffwll.ch>
> Date:   Fri Jul 2 15:02:15 2010 +0100
>
>      drm_mm: extract check_free_mm_node
>
> I've botched up the range-restriction checks. The result is usually
> an X server dying with SIGBUS in libpixman (software fallback rendering).
> Change the code to adjust the start and end for range restricted
> allocations. IMHO this even makes the code a bit clearer.
>
> Fixes regression bug: https://bugs.freedesktop.org/show_bug.cgi?id=29738
>
> Reported-by-Tested-by: Till MAtthiesen<entropy at everymail.net>
> Acked-by: Alex Deucher<alexdeucher at gmail.com>
> Signed-off-by: Daniel Vetter<daniel.vetter at ffwll.ch>
> Signed-off-by: Dave Airlie<airlied at redhat.com>
> (backported from commit 7521473305f1379403b893a30ac09a2132dc1e25 upstream)
>
> BugLink: http://bugs.launchpad.net/bugs/873130
> Signed-off-by: Seth Forshee<seth.forshee at canonical.com>
> ---
>   drivers/gpu/drm/drm_mm.c |   29 +++++++++++++++--------------
>   1 files changed, 15 insertions(+), 14 deletions(-)
>
> diff --git a/drivers/gpu/drm/drm_mm.c b/drivers/gpu/drm/drm_mm.c
> index f1d3314..5a23a9b 100644
> --- a/drivers/gpu/drm/drm_mm.c
> +++ b/drivers/gpu/drm/drm_mm.c
> @@ -330,21 +330,21 @@ void drm_mm_put_block(struct drm_mm_node *cur)
>
>   EXPORT_SYMBOL(drm_mm_put_block);
>
> -static int check_free_mm_node(struct drm_mm_node *entry, unsigned long size,
> -			      unsigned alignment)
> +static int check_free_hole(unsigned long start, unsigned long end,
> +			   unsigned long size, unsigned alignment)
>   {
>   	unsigned wasted = 0;
>
> -	if (entry->size<  size)
> +	if (end - start<  size)
>   		return 0;
>
>   	if (alignment) {
> -		register unsigned tmp = entry->start % alignment;
> +		unsigned tmp = start % alignment;
>   		if (tmp)
>   			wasted = alignment - tmp;
>   	}
>
> -	if (entry->size>= size + wasted) {
> +	if (end>= start + size + wasted) {
>   		return 1;
>   	}
>
> @@ -369,7 +369,8 @@ struct drm_mm_node *drm_mm_search_free(const struct drm_mm *mm,
>   	list_for_each(list, free_stack) {
>   		entry = list_entry(list, struct drm_mm_node, fl_entry);
>
> -		if (!check_free_mm_node(entry, size, alignment))
> +		if (!check_free_hole(entry->start, entry->start + entry->size,
> +				     size, alignment))
>   			continue;
>
>   		if (!best_match)
> @@ -392,7 +393,6 @@ struct drm_mm_node *drm_mm_search_free_in_range(const struct drm_mm *mm,
>   						unsigned long end,
>   						int best_match)
>   {
> -	struct list_head *list;
>   	const struct list_head *free_stack =&mm->fl_entry;
>   	struct drm_mm_node *entry;
>   	struct drm_mm_node *best;
> @@ -403,13 +403,13 @@ struct drm_mm_node *drm_mm_search_free_in_range(const struct drm_mm *mm,
>   	best = NULL;
>   	best_size = ~0UL;
>
> -	list_for_each(list, free_stack) {
> -		entry = list_entry(list, struct drm_mm_node, fl_entry);
> -
> -		if (entry->start>  end || (entry->start+entry->size)<  start)
> -			continue;
> +	list_for_each_entry(entry, free_stack, fl_entry) {
> +		unsigned long adj_start = entry->start<  start ?
> +			start : entry->start;
> +		unsigned long adj_end = entry->start + entry->size>  end ?
> +			end : entry->start + entry->size;
>
> -		if (!check_free_mm_node(entry, size, alignment))
> +		if (!check_free_hole(adj_start, adj_end, size, alignment))
>   			continue;
>
>   		if (!best_match)
> @@ -502,7 +502,8 @@ int drm_mm_scan_add_block(struct drm_mm_node *node)
>   	node->fl_entry.prev = prev_free;
>   	node->fl_entry.next = next_free;
>
> -	if (check_free_mm_node(node, mm->scan_size, mm->scan_alignment)) {
> +	if (check_free_hole(node->start, node->start + node->size,
> +			    mm->scan_size, mm->scan_alignment)) {
>   		mm->scan_hit_start = node->start;
>   		mm->scan_hit_size = node->size;
>


-- 
Tim Gardner tim.gardner at canonical.com