Re: [PATCH 18/21] mm/cma: support CMA in Designated Movable Blocks

Re: [PATCH 18/21] mm/cma: support CMA in Designated Movable Blocks

[kernel test robot]

Hi Doug,

I love your patch! Perhaps something to improve:

[auto build test WARNING on robh/for-next]
[also build test WARNING on linus/master v6.0-rc5]
[cannot apply to akpm-mm/mm-everything next-20220914]
[If your patch is applied to the wrong git tree, kindly drop us a note.
And when submitting patch, we suggest to use '--base' as documented in
https://git-scm.com/docs/git-format-patch#_base_tree_information]

url:    https://github.com/intel-lab-lkp/linux/commits/Doug-Berger/mm-introduce-Designated-Movable-Blocks/20220914-040216
base:   https://git.kernel.org/pub/scm/linux/kernel/git/robh/linux.git for-next
config: i386-randconfig-a002 (https://download.01.org/0day-ci/archive/20220915/202209150009.PoWlLoNu-lkp@intel.com/config)
compiler: clang version 14.0.6 (https://github.com/llvm/llvm-project f28c006a5895fc0e329fe15fead81e37457cb1d1)
reproduce (this is a W=1 build):
        wget https://raw.githubusercontent.com/intel/lkp-tests/master/sbin/make.cross -O ~/bin/make.cross
        chmod +x ~/bin/make.cross
        # https://github.com/intel-lab-lkp/linux/commit/635e919c92ca242c4b900bdfc7e21529e76f2f8e
        git remote add linux-review https://github.com/intel-lab-lkp/linux
        git fetch --no-tags linux-review Doug-Berger/mm-introduce-Designated-Movable-Blocks/20220914-040216
        git checkout 635e919c92ca242c4b900bdfc7e21529e76f2f8e
        # save the config file
        mkdir build_dir && cp config build_dir/.config
        COMPILER_INSTALL_PATH=$HOME/0day COMPILER=clang make.cross W=1 O=build_dir ARCH=i386 SHELL=/bin/bash

If you fix the issue, kindly add following tag where applicable
Reported-by: kernel test robot <lkp@intel.com>

All warnings (new ones prefixed by >>):

>> mm/page_alloc.c:9236:5: warning: no previous prototype for function '_alloc_contig_range' [-Wmissing-prototypes]
   int _alloc_contig_range(unsigned long start, unsigned long end,
       ^
   mm/page_alloc.c:9236:1: note: declare 'static' if the function is not intended to be used outside of this translation unit
   int _alloc_contig_range(unsigned long start, unsigned long end,
   ^
   static 
   1 warning generated.


vim +/_alloc_contig_range +9236 mm/page_alloc.c

  9235	
> 9236	int _alloc_contig_range(unsigned long start, unsigned long end,
  9237				unsigned int migratetype, gfp_t gfp_mask)
  9238	{
  9239		unsigned long outer_start, outer_end;
  9240		int order;
  9241		int ret = 0;
  9242	
  9243		struct compact_control cc = {
  9244			.nr_migratepages = 0,
  9245			.order = -1,
  9246			.zone = page_zone(pfn_to_page(start)),
  9247			.mode = MIGRATE_SYNC,
  9248			.ignore_skip_hint = true,
  9249			.no_set_skip_hint = true,
  9250			.gfp_mask = current_gfp_context(gfp_mask),
  9251			.alloc_contig = true,
  9252		};
  9253		INIT_LIST_HEAD(&cc.migratepages);
  9254	
  9255		/*
  9256		 * What we do here is we mark all pageblocks in range as
  9257		 * MIGRATE_ISOLATE.  Because pageblock and max order pages may
  9258		 * have different sizes, and due to the way page allocator
  9259		 * work, start_isolate_page_range() has special handlings for this.
  9260		 *
  9261		 * Once the pageblocks are marked as MIGRATE_ISOLATE, we
  9262		 * migrate the pages from an unaligned range (ie. pages that
  9263		 * we are interested in). This will put all the pages in
  9264		 * range back to page allocator as MIGRATE_ISOLATE.
  9265		 *
  9266		 * When this is done, we take the pages in range from page
  9267		 * allocator removing them from the buddy system.  This way
  9268		 * page allocator will never consider using them.
  9269		 *
  9270		 * This lets us mark the pageblocks back as
  9271		 * MIGRATE_CMA/MIGRATE_MOVABLE so that free pages in the
  9272		 * aligned range but not in the unaligned, original range are
  9273		 * put back to page allocator so that buddy can use them.
  9274		 */
  9275	
  9276		ret = start_isolate_page_range(start, end, migratetype, 0, gfp_mask);
  9277		if (ret)
  9278			goto done;
  9279	
  9280		drain_all_pages(cc.zone);
  9281	
  9282		/*
  9283		 * In case of -EBUSY, we'd like to know which page causes problem.
  9284		 * So, just fall through. test_pages_isolated() has a tracepoint
  9285		 * which will report the busy page.
  9286		 *
  9287		 * It is possible that busy pages could become available before
  9288		 * the call to test_pages_isolated, and the range will actually be
  9289		 * allocated.  So, if we fall through be sure to clear ret so that
  9290		 * -EBUSY is not accidentally used or returned to caller.
  9291		 */
  9292		ret = __alloc_contig_migrate_range(&cc, start, end);
  9293		if (ret && ret != -EBUSY)
  9294			goto done;
  9295		ret = 0;
  9296		sync_hugetlb_dissolve();
  9297	
  9298		/*
  9299		 * Pages from [start, end) are within a pageblock_nr_pages
  9300		 * aligned blocks that are marked as MIGRATE_ISOLATE.  What's
  9301		 * more, all pages in [start, end) are free in page allocator.
  9302		 * What we are going to do is to allocate all pages from
  9303		 * [start, end) (that is remove them from page allocator).
  9304		 *
  9305		 * The only problem is that pages at the beginning and at the
  9306		 * end of interesting range may be not aligned with pages that
  9307		 * page allocator holds, ie. they can be part of higher order
  9308		 * pages.  Because of this, we reserve the bigger range and
  9309		 * once this is done free the pages we are not interested in.
  9310		 *
  9311		 * We don't have to hold zone->lock here because the pages are
  9312		 * isolated thus they won't get removed from buddy.
  9313		 */
  9314	
  9315		order = 0;
  9316		outer_start = start;
  9317		while (!PageBuddy(pfn_to_page(outer_start))) {
  9318			if (++order >= MAX_ORDER) {
  9319				outer_start = start;
  9320				break;
  9321			}
  9322			outer_start &= ~0UL << order;
  9323		}
  9324	
  9325		if (outer_start != start) {
  9326			order = buddy_order(pfn_to_page(outer_start));
  9327	
  9328			/*
  9329			 * outer_start page could be small order buddy page and
  9330			 * it doesn't include start page. Adjust outer_start
  9331			 * in this case to report failed page properly
  9332			 * on tracepoint in test_pages_isolated()
  9333			 */
  9334			if (outer_start + (1UL << order) <= start)
  9335				outer_start = start;
  9336		}
  9337	
  9338		/* Make sure the range is really isolated. */
  9339		if (test_pages_isolated(outer_start, end, 0)) {
  9340			ret = -EBUSY;
  9341			goto done;
  9342		}
  9343	
  9344		/* Grab isolated pages from freelists. */
  9345		outer_end = isolate_freepages_range(&cc, outer_start, end);
  9346		if (!outer_end) {
  9347			ret = -EBUSY;
  9348			goto done;
  9349		}
  9350	
  9351		/* Free head and tail (if any) */
  9352		if (start != outer_start)
  9353			free_contig_range(outer_start, start - outer_start);
  9354		if (end != outer_end)
  9355			free_contig_range(end, outer_end - end);
  9356	
  9357	done:
  9358		undo_isolate_page_range(start, end, migratetype);
  9359		return ret;
  9360	}
  9361	

-- 
0-DAY CI Kernel Test Service
https://01.org/lkp