mm/compaction: Break out of loop on !PageBuddy in isolate_freepages_block

We received several reports of bad page state when freeing CMA pages previously allocated with alloc_contig_range:

<1>[ 1258.084111] BUG: Bad page state in process Binder_A  pfn:63202
<1>[ 1258.089763] page:d21130b0 count:0 mapcount:1 mapping:  (null) index:0x7dfbf
<1>[ 1258.096109] page flags: 0x40080068(uptodate|lru|active|swapbacked)

Based on the page state, it looks like the page was still in use. The page flags do not make sense for the use case though. Further debugging showed that despite alloc_contig_range returning success, at least one page in the range still remained in the buddy allocator.

There is an issue with isolate_freepages_block. In strict mode (which CMA uses), if any pages in the range cannot be isolated, isolate_freepages_block should return failure 0. The current check keeps track of the total number of isolated pages and compares against the size of the range:

        if (strict && nr_strict_required > total_isolated)
                total_isolated = 0;

After taking the zone lock, if one of the pages in the range is not in the buddy allocator, we continue through the loop and do not increment total_isolated. If in the last iteration of the loop we isolate more than one page (e.g. last page needed is a higher order page), the check for total_isolated may pass and we fail to detect that a page was skipped. The fix is to bail out if the loop immediately if we are in strict mode. There's no benfit to continuing anyway since we need all pages to be isolated. Additionally, drop the error checking based on nr_strict_required and just check the pfn ranges. This matches with what isolate_freepages_range does.

mm/compaction.c +4 -6

@@ -242,7 +242,6 @@ static unsigned long isolate_freepages_block(struct compact_control *cc,
 {
 	int nr_scanned = 0, total_isolated = 0;
 	struct page *cursor, *valid_page = NULL;
-	unsigned long nr_strict_required = end_pfn - blockpfn;
 	unsigned long flags;
 	bool locked = false;
 	bool checked_pageblock = false;

@@ -256,11 +255,12 @@ static unsigned long isolate_freepages_block(struct compact_control *cc,
 
 		nr_scanned++;
 		if (!pfn_valid_within(blockpfn))
-			continue;
+			goto isolate_fail;
+
 		if (!valid_page)
 			valid_page = page;
 		if (!PageBuddy(page))
-			continue;
+			goto isolate_fail;
 
 		/*
 		 * The zone lock must be held to isolate freepages.

@@ -289,12 +289,10 @@ static unsigned long isolate_freepages_block(struct compact_control *cc,
 
 		/* Recheck this is a buddy page under lock */
 		if (!PageBuddy(page))
-			continue;
+			goto isolate_fail;
 
 		/* Found a free page, break it into order-0 pages */
 		isolated = split_free_page(page);
-		if (!isolated && strict)
-			break;
 		total_isolated += isolated;
 		for (i = 0; i < isolated; i++) {
 			list_add(&page->lru, freelist);

Quoted 72 lines · On Thu, 6 Mar 2014 10:21:32 -0800 Laura Abbott <[email protected]> wrote:

We received several reports of bad page state when freeing CMA pages previously allocated with alloc_contig_range:

<1>[ 1258.084111] BUG: Bad page state in process Binder_A  pfn:63202
<1>[ 1258.089763] page:d21130b0 count:0 mapcount:1 mapping:  (null) index:0x7dfbf
<1>[ 1258.096109] page flags: 0x40080068(uptodate|lru|active|swapbacked)

Based on the page state, it looks like the page was still in use. The page flags do not make sense for the use case though. Further debugging showed that despite alloc_contig_range returning success, at least one page in the range still remained in the buddy allocator.

There is an issue with isolate_freepages_block. In strict mode (which CMA uses), if any pages in the range cannot be isolated, isolate_freepages_block should return failure 0. The current check keeps track of the total number of isolated pages and compares against the size of the range:

         if (strict && nr_strict_required > total_isolated)
                 total_isolated = 0;

After taking the zone lock, if one of the pages in the range is not in the buddy allocator, we continue through the loop and do not increment total_isolated. If in the last iteration of the loop we isolate more than one page (e.g. last page needed is a higher order page), the check for total_isolated may pass and we fail to detect that a page was skipped. The fix is to bail out if the loop immediately if we are in strict mode. There's no benfit to continuing anyway since we need all pages to be isolated. Additionally, drop the error checking based on nr_strict_required and just check the pfn ranges. This matches with what isolate_freepages_range does.

mm/compaction.c +4 -6

@@ -242,7 +242,6 @@ static unsigned long isolate_freepages_block(struct compact_control *cc,
  {
  	int nr_scanned = 0, total_isolated = 0;
  	struct page *cursor, *valid_page = NULL;
-	unsigned long nr_strict_required = end_pfn - blockpfn;
  	unsigned long flags;
  	bool locked = false;
  	bool checked_pageblock = false;

@@ -256,11 +255,12 @@ static unsigned long isolate_freepages_block(struct compact_control *cc,
 
  		nr_scanned++;
  		if (!pfn_valid_within(blockpfn))
-			continue;
+			goto isolate_fail;
+
  		if (!valid_page)
  			valid_page = page;
  		if (!PageBuddy(page))
-			continue;
+			goto isolate_fail;
 
  		/*
  		 * The zone lock must be held to isolate freepages.

@@ -289,12 +289,10 @@ static unsigned long isolate_freepages_block(struct compact_control *cc,
 
  		/* Recheck this is a buddy page under lock */
  		if (!PageBuddy(page))
-			continue;
+			goto isolate_fail;
 
  		/* Found a free page, break it into order-0 pages */
  		isolated = split_free_page(page);
-		if (!isolated && strict)
-			break;
  		total_isolated += isolated;
  		for (i = 0; i < isolated; i++) {
  			list_add(&page->lru, freelist);

mm/compaction.c static unsigned long isolate_freepages_block(struct compact_control *cc, · line 307

Quoted 5 lines

mm/compaction.c +1 -0

@@ -305,7 +303,15 @@ static unsigned long isolate_freepages_block(struct compact_control *cc,
  		if (isolated) {
  			blockpfn += isolated - 1;
  			cursor += isolated - 1;
+			continue;
  		}

We can make the code a little more efficient and (I think) clearer by moving that `if (isolated)' test.

mm/compaction.c static unsigned long isolate_freepages_block(struct compact_control *cc, · line 314

Quoted 7 lines

mm/compaction.c +7 -0

+
+isolate_fail:
+		if (strict)
+			break;
+		else
+			continue;
+

And I don't think this `continue' has any benefit.

mm/compaction.c renamed from mm/compaction.c~mm-compaction-break-out-of-loop-on-pagebuddy-in-isolate_freepages_block-fix +7 -10

@@ -293,14 +293,14 @@ static unsigned long isolate_freepages_b
 
  		/* Found a free page, break it into order-0 pages */
  		isolated = split_free_page(page);
-		total_isolated += isolated;
-		for (i = 0; i < isolated; i++) {
-			list_add(&page->lru, freelist);
-			page++;
-		}
-
-		/* If a page was split, advance to the end of it */
  		if (isolated) {
+			total_isolated += isolated;
+			for (i = 0; i < isolated; i++) {
+				list_add(&page->lru, freelist);
+				page++;
+			}
+
+			/* If a page was split, advance to the end of it */
  			blockpfn += isolated - 1;
  			cursor += isolated - 1;
  			continue;

@@ -309,9 +309,6 @@ static unsigned long isolate_freepages_b
  isolate_fail:
  		if (strict)
  			break;
-		else
-			continue;
-
  	}
 
  	trace_mm_compaction_isolate_freepages(nr_scanned, total_isolated);

Problem is, I can't be bothered testing this.

We received several reports of bad page state when freeing CMA pages previously allocated with alloc_contig_range:

<1>[ 1258.084111] BUG: Bad page state in process Binder_A  pfn:63202
<1>[ 1258.089763] page:d21130b0 count:0 mapcount:1 mapping:  (null) index:0x7dfbf
<1>[ 1258.096109] page flags: 0x40080068(uptodate|lru|active|swapbacked)

Based on the page state, it looks like the page was still in use. The page flags do not make sense for the use case though. Further debugging showed that despite alloc_contig_range returning success, at least one page in the range still remained in the buddy allocator.

There is an issue with isolate_freepages_block. In strict mode (which CMA uses), if any pages in the range cannot be isolated, isolate_freepages_block should return failure 0. The current check keeps track of the total number of isolated pages and compares against the size of the range:

        if (strict && nr_strict_required > total_isolated)
                total_isolated = 0;

After taking the zone lock, if one of the pages in the range is not in the buddy allocator, we continue through the loop and do not increment total_isolated. If in the last iteration of the loop we isolate more than one page (e.g. last page needed is a higher order page), the check for total_isolated may pass and we fail to detect that a page was skipped. The fix is to bail out if the loop immediately if we are in strict mode. There's no benfit to continuing anyway since we need all pages to be isolated. Additionally, drop the error checking based on nr_strict_required and just check the pfn ranges. This matches with what isolate_freepages_range does.

Signed-off-by Laura Abbott <[email protected]>

Quoted 32 lines · Laura Abbott

We received several reports of bad page state when freeing CMA pages previously allocated with alloc_contig_range:

<1>[ 1258.084111] BUG: Bad page state in process Binder_A  pfn:63202
<1>[ 1258.089763] page:d21130b0 count:0 mapcount:1 mapping:  (null) index:0x7dfbf
<1>[ 1258.096109] page flags: 0x40080068(uptodate|lru|active|swapbacked)

Based on the page state, it looks like the page was still in use. The page flags do not make sense for the use case though. Further debugging showed that despite alloc_contig_range returning success, at least one page in the range still remained in the buddy allocator.

There is an issue with isolate_freepages_block. In strict mode (which CMA uses), if any pages in the range cannot be isolated, isolate_freepages_block should return failure 0. The current check keeps track of the total number of isolated pages and compares against the size of the range:

        if (strict && nr_strict_required > total_isolated)
                total_isolated = 0;

After taking the zone lock, if one of the pages in the range is not in the buddy allocator, we continue through the loop and do not increment total_isolated. If in the last iteration of the loop we isolate more than one page (e.g. last page needed is a higher order page), the check for total_isolated may pass and we fail to detect that a page was skipped. The fix is to bail out if the loop immediately if we are in strict mode. There's no benfit to continuing anyway since we need all pages to be isolated. Additionally, drop the error checking based on nr_strict_required and just check the pfn ranges. This matches with what isolate_freepages_range does.

Quoted 1 line

Signed-off-by Laura Abbott <[email protected]>

Nice catch! stable stuff?

We received several reports of bad page state when freeing CMA pages previously allocated with alloc_contig_range:

<1>[ 1258.084111] BUG: Bad page state in process Binder_A  pfn:63202
<1>[ 1258.089763] page:d21130b0 count:0 mapcount:1 mapping:  (null) index:0x7dfbf
<1>[ 1258.096109] page flags: 0x40080068(uptodate|lru|active|swapbacked)

Based on the page state, it looks like the page was still in use. The page flags do not make sense for the use case though. Further debugging showed that despite alloc_contig_range returning success, at least one page in the range still remained in the buddy allocator.

There is an issue with isolate_freepages_block. In strict mode (which CMA uses), if any pages in the range cannot be isolated, isolate_freepages_block should return failure 0. The current check keeps track of the total number of isolated pages and compares against the size of the range:

         if (strict && nr_strict_required > total_isolated)
                 total_isolated = 0;

After taking the zone lock, if one of the pages in the range is not in the buddy allocator, we continue through the loop and do not increment total_isolated. If in the last iteration of the loop we isolate more than one page (e.g. last page needed is a higher order page), the check for total_isolated may pass and we fail to detect that a page was skipped. The fix is to bail out if the loop immediately if we are in strict mode. There's no benfit to continuing anyway since we need all pages to be isolated. Additionally, drop the error checking based on nr_strict_required and just check the pfn ranges. This matches with what isolate_freepages_range does.

Signed-off-by Laura Abbott <[email protected]>

---
v2: Addressed several comments by Vlastimil

  mm/compaction.c |   20 +++++++++++++-------
  1 files changed, 13 insertions(+), 7 deletions(-)

mm/compaction.c +12 -6

index 5142920..054c28b 100644
@@ -242,7 +242,6 @@ static unsigned long isolate_freepages_block(struct compact_control *cc,
  {
  	int nr_scanned = 0, total_isolated = 0;
  	struct page *cursor, *valid_page = NULL;
-	unsigned long nr_strict_required = end_pfn - blockpfn;
  	unsigned long flags;
  	bool locked = false;
  	bool checked_pageblock = false;

@@ -256,11 +255,12 @@ static unsigned long isolate_freepages_block(struct compact_control *cc,
 
  		nr_scanned++;
  		if (!pfn_valid_within(blockpfn))
-			continue;
+			goto isolate_fail;
+
  		if (!valid_page)
  			valid_page = page;
  		if (!PageBuddy(page))
-			continue;
+			goto isolate_fail;
 
  		/*
  		 * The zone lock must be held to isolate freepages.

@@ -289,12 +289,10 @@ static unsigned long isolate_freepages_block(struct compact_control *cc,
 
  		/* Recheck this is a buddy page under lock */
  		if (!PageBuddy(page))
-			continue;
+			goto isolate_fail;
 
  		/* Found a free page, break it into order-0 pages */
  		isolated = split_free_page(page);
-		if (!isolated && strict)
-			break;
  		total_isolated += isolated;
  		for (i = 0; i < isolated; i++) {
  			list_add(&page->lru, freelist);

@@ -305,7 +303,15 @@ static unsigned long isolate_freepages_block(struct compact_control *cc,
  		if (isolated) {
  			blockpfn += isolated - 1;
  			cursor += isolated - 1;
+			continue;
  		}
+
+isolate_fail:
+		if (strict)
+			break;
+		else
+			continue;
+
  	}
 
  	trace_mm_compaction_isolate_freepages(nr_scanned, total_isolated);

We received several reports of bad page state when freeing CMA pages previously allocated with alloc_contig_range:

<1>[ 1258.084111] BUG: Bad page state in process Binder_A  pfn:63202
<1>[ 1258.089763] page:d21130b0 count:0 mapcount:1 mapping:  (null) index:0x7dfbf
<1>[ 1258.096109] page flags: 0x40080068(uptodate|lru|active|swapbacked)

Based on the page state, it looks like the page was still in use. The page flags do not make sense for the use case though. Further debugging showed that despite alloc_contig_range returning success, at least one page in the range still remained in the buddy allocator.

There is an issue with isolate_freepages_block. In strict mode (which CMA uses), if any pages in the range cannot be isolated, isolate_freepages_block should return failure 0. The current check keeps track of the total number of isolated pages and compares against the size of the range:

        if (strict && nr_strict_required > total_isolated)
                total_isolated = 0;

After taking the zone lock, if one of the pages in the range is not in the buddy allocator, we continue through the loop and do not increment total_isolated. If in the last iteration of the loop we isolate more than one page (e.g. last page needed is a higher order page), the check for total_isolated may pass and we fail to detect that a page was skipped. The fix is to bail out if the loop immediately if we are in strict mode. There's no benfit to continuing anyway since we need all pages to be isolated. Additionally, drop the error checking based on nr_strict_required and just check the pfn ranges. This matches with what isolate_freepages_range does.

Signed-off-by Laura Abbott <[email protected]>