LinuxLists.cc - [PATCH 1/3] madvise: Convert madvise_cold_or_pageout_pte

2022-12-07 00:34:42

Subject: [PATCH 1/3] madvise: Convert madvise_cold_or_pageout_pte_range() to use folios

This change removes a number of calls to compound_head(), and saves 1319
bytes of kernel text.

Signed-off-by: Vishal Moola (Oracle) <[email protected]>
---
mm/madvise.c | 88 +++++++++++++++++++++++++++-------------------------
1 file changed, 45 insertions(+), 43 deletions(-)

diff --git a/mm/madvise.c b/mm/madvise.c
index 2baa93ca2310..59bfc6c9c548 100644
--- a/mm/madvise.c
+++ b/mm/madvise.c
@@ -332,8 +332,9 @@ static int madvise_cold_or_pageout_pte_range(pmd_t *pmd,
struct vm_area_struct *vma = walk->vma;
pte_t *orig_pte, *pte, ptent;
spinlock_t *ptl;
+ struct folio *folio = NULL;
struct page *page = NULL;
- LIST_HEAD(page_list);
+ LIST_HEAD(folio_list);

if (fatal_signal_pending(current))
return -EINTR;
@@ -358,23 +359,23 @@ static int madvise_cold_or_pageout_pte_range(pmd_t *pmd,
goto huge_unlock;
}

- page = pmd_page(orig_pmd);
+ folio = pfn_folio(pmd_pfn(orig_pmd));

- /* Do not interfere with other mappings of this page */
- if (page_mapcount(page) != 1)
+ /* Do not interfere with other mappings of this folio */
+ if (folio_mapcount(folio) != 1)
goto huge_unlock;

if (next - addr != HPAGE_PMD_SIZE) {
int err;

- get_page(page);
+ folio_get(folio);
spin_unlock(ptl);
- lock_page(page);
- err = split_huge_page(page);
- unlock_page(page);
- put_page(page);
+ folio_lock(folio);
+ err = split_folio(folio);
+ folio_unlock(folio);
+ folio_put(folio);
if (!err)
- goto regular_page;
+ goto regular_folio;
return 0;
}

@@ -386,25 +387,25 @@ static int madvise_cold_or_pageout_pte_range(pmd_t *pmd,
tlb_remove_pmd_tlb_entry(tlb, pmd, addr);
}

- ClearPageReferenced(page);
- test_and_clear_page_young(page);
+ folio_clear_referenced(folio);
+ folio_test_clear_young(folio);
if (pageout) {
- if (!isolate_lru_page(page)) {
- if (PageUnevictable(page))
- putback_lru_page(page);
+ if (!folio_isolate_lru(folio)) {
+ if (folio_test_unevictable(folio))
+ folio_putback_lru(folio);
else
- list_add(&page->lru, &page_list);
+ list_add(&folio->lru, &folio_list);
}
} else
- deactivate_page(page);
+ deactivate_page(&folio->page);
huge_unlock:
spin_unlock(ptl);
if (pageout)
- reclaim_pages(&page_list);
+ reclaim_pages(&folio_list);
return 0;
}

-regular_page:
+regular_folio:
if (pmd_trans_unstable(pmd))
return 0;
#endif
@@ -424,28 +425,29 @@ static int madvise_cold_or_pageout_pte_range(pmd_t *pmd,
page = vm_normal_page(vma, addr, ptent);
if (!page || is_zone_device_page(page))
continue;
+ folio = page_folio(page);

/*
* Creating a THP page is expensive so split it only if we
* are sure it's worth. Split it if we are only owner.
*/
- if (PageTransCompound(page)) {
- if (page_mapcount(page) != 1)
+ if (folio_test_large(folio)) {
+ if (folio_mapcount(folio) != 1)
break;
- get_page(page);
- if (!trylock_page(page)) {
- put_page(page);
+ folio_get(folio);
+ if (!folio_trylock(folio)) {
+ folio_put(folio);
break;
}
pte_unmap_unlock(orig_pte, ptl);
- if (split_huge_page(page)) {
- unlock_page(page);
- put_page(page);
+ if (split_folio(folio)) {
+ folio_unlock(folio);
+ folio_put(folio);
orig_pte = pte_offset_map_lock(mm, pmd, addr, &ptl);
break;
}
- unlock_page(page);
- put_page(page);
+ folio_unlock(folio);
+ folio_put(folio);
orig_pte = pte = pte_offset_map_lock(mm, pmd, addr, &ptl);
pte--;
addr -= PAGE_SIZE;
@@ -453,13 +455,13 @@ static int madvise_cold_or_pageout_pte_range(pmd_t *pmd,
}

/*
- * Do not interfere with other mappings of this page and
- * non-LRU page.
+ * Do not interfere with other mappings of this folio and
+ * non-LRU folio.
*/
- if (!PageLRU(page) || page_mapcount(page) != 1)
+ if (!folio_test_lru(folio))
continue;

- VM_BUG_ON_PAGE(PageTransCompound(page), page);
+ VM_BUG_ON_PAGE(folio_test_large(folio), folio);

if (pte_young(ptent)) {
ptent = ptep_get_and_clear_full(mm, addr, pte,
@@ -470,28 +472,28 @@ static int madvise_cold_or_pageout_pte_range(pmd_t *pmd,
}

/*
- * We are deactivating a page for accelerating reclaiming.
- * VM couldn't reclaim the page unless we clear PG_young.
+ * We are deactivating a folio for accelerating reclaiming.
+ * VM couldn't reclaim the folio unless we clear PG_young.
* As a side effect, it makes confuse idle-page tracking
* because they will miss recent referenced history.
*/
- ClearPageReferenced(page);
- test_and_clear_page_young(page);
+ folio_clear_referenced(folio);
+ folio_test_clear_young(folio);
if (pageout) {
- if (!isolate_lru_page(page)) {
- if (PageUnevictable(page))
- putback_lru_page(page);
+ if (!folio_isolate_lru(folio)) {
+ if (folio_test_unevictable(folio))
+ folio_putback_lru(folio);
else
- list_add(&page->lru, &page_list);
+ list_add(&folio->lru, &folio_list);
}
} else
- deactivate_page(page);
+ deactivate_page(&folio->page);
}

arch_leave_lazy_mmu_mode();
pte_unmap_unlock(orig_pte, ptl);
if (pageout)
- reclaim_pages(&page_list);
+ reclaim_pages(&folio_list);
cond_resched();

return 0;
--
2.38.1

2022-12-07 10:25:51

by kernel test robot

[permalink] [raw]

Subject: Re: [PATCH 1/3] madvise: Convert madvise_cold_or_pageout_pte_range() to use folios

Hi Vishal,

Thank you for the patch! Yet something to improve:

[auto build test ERROR on linus/master]
[also build test ERROR on v6.1-rc8]
[cannot apply to akpm-mm/mm-everything next-20221207]
[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/Vishal-Moola-Oracle/Convert-deactivate_page-to-deactivate_folio/20221207-082339
patch link: https://lore.kernel.org/r/20221207002158.418789-2-vishal.moola%40gmail.com
patch subject: [PATCH 1/3] madvise: Convert madvise_cold_or_pageout_pte_range() to use folios
config: hexagon-randconfig-r041-20221206
compiler: clang version 16.0.0 (https://github.com/llvm/llvm-project 6e4cea55f0d1104408b26ac574566a0e4de48036)
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/670bf0d6d5222d0c92e72b7869d25fa64f582082
git remote add linux-review https://github.com/intel-lab-lkp/linux
git fetch --no-tags linux-review Vishal-Moola-Oracle/Convert-deactivate_page-to-deactivate_folio/20221207-082339
git checkout 670bf0d6d5222d0c92e72b7869d25fa64f582082
# 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=hexagon SHELL=/bin/bash

If you fix the issue, kindly add following tag where applicable
| Reported-by: kernel test robot <[email protected]>

All errors (new ones prefixed by >>):

In file included from mm/madvise.c:10:
In file included from include/linux/pagemap.h:11:
In file included from include/linux/highmem.h:12:
In file included from include/linux/hardirq.h:11:
In file included from ./arch/hexagon/include/generated/asm/hardirq.h:1:
In file included from include/asm-generic/hardirq.h:17:
In file included from include/linux/irq.h:20:
In file included from include/linux/io.h:13:
In file included from arch/hexagon/include/asm/io.h:334:
include/asm-generic/io.h:547:31: warning: performing pointer arithmetic on a null pointer has undefined behavior [-Wnull-pointer-arithmetic]
val = __raw_readb(PCI_IOBASE + addr);
~~~~~~~~~~ ^
include/asm-generic/io.h:560:61: warning: performing pointer arithmetic on a null pointer has undefined behavior [-Wnull-pointer-arithmetic]
val = __le16_to_cpu((__le16 __force)__raw_readw(PCI_IOBASE + addr));
~~~~~~~~~~ ^
include/uapi/linux/byteorder/little_endian.h:37:51: note: expanded from macro '__le16_to_cpu'
#define __le16_to_cpu(x) ((__force __u16)(__le16)(x))
^
In file included from mm/madvise.c:10:
In file included from include/linux/pagemap.h:11:
In file included from include/linux/highmem.h:12:
In file included from include/linux/hardirq.h:11:
In file included from ./arch/hexagon/include/generated/asm/hardirq.h:1:
In file included from include/asm-generic/hardirq.h:17:
In file included from include/linux/irq.h:20:
In file included from include/linux/io.h:13:
In file included from arch/hexagon/include/asm/io.h:334:
include/asm-generic/io.h:573:61: warning: performing pointer arithmetic on a null pointer has undefined behavior [-Wnull-pointer-arithmetic]
val = __le32_to_cpu((__le32 __force)__raw_readl(PCI_IOBASE + addr));
~~~~~~~~~~ ^
include/uapi/linux/byteorder/little_endian.h:35:51: note: expanded from macro '__le32_to_cpu'
#define __le32_to_cpu(x) ((__force __u32)(__le32)(x))
^
In file included from mm/madvise.c:10:
In file included from include/linux/pagemap.h:11:
In file included from include/linux/highmem.h:12:
In file included from include/linux/hardirq.h:11:
In file included from ./arch/hexagon/include/generated/asm/hardirq.h:1:
In file included from include/asm-generic/hardirq.h:17:
In file included from include/linux/irq.h:20:
In file included from include/linux/io.h:13:
In file included from arch/hexagon/include/asm/io.h:334:
include/asm-generic/io.h:584:33: warning: performing pointer arithmetic on a null pointer has undefined behavior [-Wnull-pointer-arithmetic]
__raw_writeb(value, PCI_IOBASE + addr);
~~~~~~~~~~ ^
include/asm-generic/io.h:594:59: warning: performing pointer arithmetic on a null pointer has undefined behavior [-Wnull-pointer-arithmetic]
__raw_writew((u16 __force)cpu_to_le16(value), PCI_IOBASE + addr);
~~~~~~~~~~ ^
include/asm-generic/io.h:604:59: warning: performing pointer arithmetic on a null pointer has undefined behavior [-Wnull-pointer-arithmetic]
__raw_writel((u32 __force)cpu_to_le32(value), PCI_IOBASE + addr);
~~~~~~~~~~ ^
>> mm/madvise.c:464:43: error: incompatible pointer types passing 'struct folio *' to parameter of type 'struct page *' [-Werror,-Wincompatible-pointer-types]
VM_BUG_ON_PAGE(folio_test_large(folio), folio);
^~~~~
include/linux/mmdebug.h:21:14: note: expanded from macro 'VM_BUG_ON_PAGE'
dump_page(page, "VM_BUG_ON_PAGE(" __stringify(cond)")");\
^~~~
include/linux/mmdebug.h:12:29: note: passing argument to parameter 'page' here
void dump_page(struct page *page, const char *reason);
^
6 warnings and 1 error generated.

vim +464 mm/madvise.c

323
324 static int madvise_cold_or_pageout_pte_range(pmd_t *pmd,
325 unsigned long addr, unsigned long end,
326 struct mm_walk *walk)
327 {
328 struct madvise_walk_private *private = walk->private;
329 struct mmu_gather *tlb = private->tlb;
330 bool pageout = private->pageout;
331 struct mm_struct *mm = tlb->mm;
332 struct vm_area_struct *vma = walk->vma;
333 pte_t *orig_pte, *pte, ptent;
334 spinlock_t *ptl;
335 struct folio *folio = NULL;
336 struct page *page = NULL;
337 LIST_HEAD(folio_list);
338
339 if (fatal_signal_pending(current))
340 return -EINTR;
341
342 #ifdef CONFIG_TRANSPARENT_HUGEPAGE
343 if (pmd_trans_huge(*pmd)) {
344 pmd_t orig_pmd;
345 unsigned long next = pmd_addr_end(addr, end);
346
347 tlb_change_page_size(tlb, HPAGE_PMD_SIZE);
348 ptl = pmd_trans_huge_lock(pmd, vma);
349 if (!ptl)
350 return 0;
351
352 orig_pmd = *pmd;
353 if (is_huge_zero_pmd(orig_pmd))
354 goto huge_unlock;
355
356 if (unlikely(!pmd_present(orig_pmd))) {
357 VM_BUG_ON(thp_migration_supported() &&
358 !is_pmd_migration_entry(orig_pmd));
359 goto huge_unlock;
360 }
361
362 folio = pfn_folio(pmd_pfn(orig_pmd));
363
364 /* Do not interfere with other mappings of this folio */
365 if (folio_mapcount(folio) != 1)
366 goto huge_unlock;
367
368 if (next - addr != HPAGE_PMD_SIZE) {
369 int err;
370
371 folio_get(folio);
372 spin_unlock(ptl);
373 folio_lock(folio);
374 err = split_folio(folio);
375 folio_unlock(folio);
376 folio_put(folio);
377 if (!err)
378 goto regular_folio;
379 return 0;
380 }
381
382 if (pmd_young(orig_pmd)) {
383 pmdp_invalidate(vma, addr, pmd);
384 orig_pmd = pmd_mkold(orig_pmd);
385
386 set_pmd_at(mm, addr, pmd, orig_pmd);
387 tlb_remove_pmd_tlb_entry(tlb, pmd, addr);
388 }
389
390 folio_clear_referenced(folio);
391 folio_test_clear_young(folio);
392 if (pageout) {
393 if (!folio_isolate_lru(folio)) {
394 if (folio_test_unevictable(folio))
395 folio_putback_lru(folio);
396 else
397 list_add(&folio->lru, &folio_list);
398 }
399 } else
400 deactivate_page(&folio->page);
401 huge_unlock:
402 spin_unlock(ptl);
403 if (pageout)
404 reclaim_pages(&folio_list);
405 return 0;
406 }
407
408 regular_folio:
409 if (pmd_trans_unstable(pmd))
410 return 0;
411 #endif
412 tlb_change_page_size(tlb, PAGE_SIZE);
413 orig_pte = pte = pte_offset_map_lock(vma->vm_mm, pmd, addr, &ptl);
414 flush_tlb_batched_pending(mm);
415 arch_enter_lazy_mmu_mode();
416 for (; addr < end; pte++, addr += PAGE_SIZE) {
417 ptent = *pte;
418
419 if (pte_none(ptent))
420 continue;
421
422 if (!pte_present(ptent))
423 continue;
424
425 page = vm_normal_page(vma, addr, ptent);
426 if (!page || is_zone_device_page(page))
427 continue;
428 folio = page_folio(page);
429
430 /*
431 * Creating a THP page is expensive so split it only if we
432 * are sure it's worth. Split it if we are only owner.
433 */
434 if (folio_test_large(folio)) {
435 if (folio_mapcount(folio) != 1)
436 break;
437 folio_get(folio);
438 if (!folio_trylock(folio)) {
439 folio_put(folio);
440 break;
441 }
442 pte_unmap_unlock(orig_pte, ptl);
443 if (split_folio(folio)) {
444 folio_unlock(folio);
445 folio_put(folio);
446 orig_pte = pte_offset_map_lock(mm, pmd, addr, &ptl);
447 break;
448 }
449 folio_unlock(folio);
450 folio_put(folio);
451 orig_pte = pte = pte_offset_map_lock(mm, pmd, addr, &ptl);
452 pte--;
453 addr -= PAGE_SIZE;
454 continue;
455 }
456
457 /*
458 * Do not interfere with other mappings of this folio and
459 * non-LRU folio.
460 */
461 if (!folio_test_lru(folio))
462 continue;
463
> 464 VM_BUG_ON_PAGE(folio_test_large(folio), folio);
465
466 if (pte_young(ptent)) {
467 ptent = ptep_get_and_clear_full(mm, addr, pte,
468 tlb->fullmm);
469 ptent = pte_mkold(ptent);
470 set_pte_at(mm, addr, pte, ptent);
471 tlb_remove_tlb_entry(tlb, pte, addr);
472 }
473
474 /*
475 * We are deactivating a folio for accelerating reclaiming.
476 * VM couldn't reclaim the folio unless we clear PG_young.
477 * As a side effect, it makes confuse idle-page tracking
478 * because they will miss recent referenced history.
479 */
480 folio_clear_referenced(folio);
481 folio_test_clear_young(folio);
482 if (pageout) {
483 if (!folio_isolate_lru(folio)) {
484 if (folio_test_unevictable(folio))
485 folio_putback_lru(folio);
486 else
487 list_add(&folio->lru, &folio_list);
488 }
489 } else
490 deactivate_page(&folio->page);
491 }
492
493 arch_leave_lazy_mmu_mode();
494 pte_unmap_unlock(orig_pte, ptl);
495 if (pageout)
496 reclaim_pages(&folio_list);
497 cond_resched();
498
499 return 0;
500 }
501

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

Attachments:

(No filename) (11.51 kB)
config (100.05 kB)
Download all attachments

2022-12-08 01:11:06

by kernel test robot

[permalink] [raw]

Subject: Re: [PATCH 1/3] madvise: Convert madvise_cold_or_pageout_pte_range() to use folios

Attachments:

(No filename) (8.15 kB)
config (161.04 kB)
Download all attachments