* Re: [PATCH v4 1/5] rmap: move hugetlb try_to_unmap to dedicated function
[not found] <20230313124526.1207490-2-fengwei.yin@intel.com>
@ 2023-03-13 16:55 ` kernel test robot
2023-03-13 17:37 ` kernel test robot
1 sibling, 0 replies; 2+ messages in thread
From: kernel test robot @ 2023-03-13 16:55 UTC (permalink / raw)
To: Yin Fengwei; +Cc: llvm, oe-kbuild-all
Hi Yin,
Thank you for the patch! Yet something to improve:
[auto build test ERROR on akpm-mm/mm-everything]
[also build test ERROR on linus/master v6.3-rc2 next-20230310]
[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/Yin-Fengwei/rmap-move-hugetlb-try_to_unmap-to-dedicated-function/20230313-204656
base: https://git.kernel.org/pub/scm/linux/kernel/git/akpm/mm.git mm-everything
patch link: https://lore.kernel.org/r/20230313124526.1207490-2-fengwei.yin%40intel.com
patch subject: [PATCH v4 1/5] rmap: move hugetlb try_to_unmap to dedicated function
config: arm-buildonly-randconfig-r001-20230312 (https://download.01.org/0day-ci/archive/20230314/202303140016.q8dVJMP2-lkp@intel.com/config)
compiler: clang version 17.0.0 (https://github.com/llvm/llvm-project 67409911353323ca5edf2049ef0df54132fa1ca7)
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
# install arm cross compiling tool for clang build
# apt-get install binutils-arm-linux-gnueabi
# https://github.com/intel-lab-lkp/linux/commit/c43d03caaff9b0b9a869fd2a9b8beed7bc833f2a
git remote add linux-review https://github.com/intel-lab-lkp/linux
git fetch --no-tags linux-review Yin-Fengwei/rmap-move-hugetlb-try_to_unmap-to-dedicated-function/20230313-204656
git checkout c43d03caaff9b0b9a869fd2a9b8beed7bc833f2a
# save the config file
mkdir build_dir && cp config build_dir/.config
COMPILER_INSTALL_PATH=$HOME/0day COMPILER=clang make.cross W=1 O=build_dir ARCH=arm olddefconfig
COMPILER_INSTALL_PATH=$HOME/0day COMPILER=clang make.cross W=1 O=build_dir ARCH=arm SHELL=/bin/bash
If you fix the issue, kindly add following tag where applicable
| Reported-by: kernel test robot <lkp@intel.com>
| Link: https://lore.kernel.org/oe-kbuild-all/202303140016.q8dVJMP2-lkp@intel.com/
All errors (new ones prefixed by >>):
>> mm/rmap.c:1513:6: error: call to undeclared function 'huge_pte_dirty'; ISO C99 and later do not support implicit function declarations [-Wimplicit-function-declaration]
if (huge_pte_dirty(pteval))
^
1 error generated.
vim +/huge_pte_dirty +1513 mm/rmap.c
1443
1444 static bool try_to_unmap_one_hugetlb(struct folio *folio,
1445 struct vm_area_struct *vma, struct mmu_notifier_range range,
1446 struct page_vma_mapped_walk pvmw, unsigned long address,
1447 enum ttu_flags flags)
1448 {
1449 struct mm_struct *mm = vma->vm_mm;
1450 pte_t pteval;
1451 bool ret = true, anon = folio_test_anon(folio);
1452
1453 /*
1454 * The try_to_unmap() is only passed a hugetlb page
1455 * in the case where the hugetlb page is poisoned.
1456 */
1457 VM_BUG_ON_FOLIO(!folio_test_hwpoison(folio), folio);
1458 /*
1459 * huge_pmd_unshare may unmap an entire PMD page.
1460 * There is no way of knowing exactly which PMDs may
1461 * be cached for this mm, so we must flush them all.
1462 * start/end were already adjusted in caller
1463 * (try_to_unmap_one) to cover this range.
1464 */
1465 flush_cache_range(vma, range.start, range.end);
1466
1467 /*
1468 * To call huge_pmd_unshare, i_mmap_rwsem must be
1469 * held in write mode. Caller needs to explicitly
1470 * do this outside rmap routines.
1471 *
1472 * We also must hold hugetlb vma_lock in write mode.
1473 * Lock order dictates acquiring vma_lock BEFORE
1474 * i_mmap_rwsem. We can only try lock here and fail
1475 * if unsuccessful.
1476 */
1477 if (!anon) {
1478 VM_BUG_ON(!(flags & TTU_RMAP_LOCKED));
1479 if (!hugetlb_vma_trylock_write(vma)) {
1480 ret = false;
1481 goto out;
1482 }
1483 if (huge_pmd_unshare(mm, vma, address, pvmw.pte)) {
1484 hugetlb_vma_unlock_write(vma);
1485 flush_tlb_range(vma,
1486 range.start, range.end);
1487 mmu_notifier_invalidate_range(mm,
1488 range.start, range.end);
1489 /*
1490 * The ref count of the PMD page was
1491 * dropped which is part of the way map
1492 * counting is done for shared PMDs.
1493 * Return 'true' here. When there is
1494 * no other sharing, huge_pmd_unshare
1495 * returns false and we will unmap the
1496 * actual page and drop map count
1497 * to zero.
1498 */
1499 goto out;
1500 }
1501 hugetlb_vma_unlock_write(vma);
1502 }
1503 pteval = huge_ptep_clear_flush(vma, address, pvmw.pte);
1504
1505 /*
1506 * Now the pte is cleared. If this pte was uffd-wp armed,
1507 * we may want to replace a none pte with a marker pte if
1508 * it's file-backed, so we don't lose the tracking info.
1509 */
1510 pte_install_uffd_wp_if_needed(vma, address, pvmw.pte, pteval);
1511
1512 /* Set the dirty flag on the folio now the pte is gone. */
> 1513 if (huge_pte_dirty(pteval))
1514 folio_mark_dirty(folio);
1515
1516 /* Update high watermark before we lower rss */
1517 update_hiwater_rss(mm);
1518
1519 /* Poisoned hugetlb folio with TTU_HWPOISON always cleared in flags */
1520 pteval = swp_entry_to_pte(make_hwpoison_entry(&folio->page));
1521 set_huge_pte_at(mm, address, pvmw.pte, pteval);
1522 hugetlb_count_sub(folio_nr_pages(folio), mm);
1523
1524 /*
1525 * No need to call mmu_notifier_invalidate_range() it has be
1526 * done above for all cases requiring it to happen under page
1527 * table lock before mmu_notifier_invalidate_range_end()
1528 *
1529 * See Documentation/mm/mmu_notifier.rst
1530 */
1531 page_remove_rmap(&folio->page, vma, true);
1532 /* No VM_LOCKED set in vma->vm_flags for hugetlb. So not
1533 * necessary to call mlock_drain_local().
1534 */
1535 folio_put(folio);
1536
1537 out:
1538 return ret;
1539 }
1540
--
0-DAY CI Kernel Test Service
https://github.com/intel/lkp-tests
^ permalink raw reply [flat|nested] 2+ messages in thread* Re: [PATCH v4 1/5] rmap: move hugetlb try_to_unmap to dedicated function
[not found] <20230313124526.1207490-2-fengwei.yin@intel.com>
2023-03-13 16:55 ` [PATCH v4 1/5] rmap: move hugetlb try_to_unmap to dedicated function kernel test robot
@ 2023-03-13 17:37 ` kernel test robot
1 sibling, 0 replies; 2+ messages in thread
From: kernel test robot @ 2023-03-13 17:37 UTC (permalink / raw)
To: Yin Fengwei; +Cc: llvm, oe-kbuild-all
Hi Yin,
Thank you for the patch! Yet something to improve:
[auto build test ERROR on akpm-mm/mm-everything]
[also build test ERROR on linus/master v6.3-rc2 next-20230310]
[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/Yin-Fengwei/rmap-move-hugetlb-try_to_unmap-to-dedicated-function/20230313-204656
base: https://git.kernel.org/pub/scm/linux/kernel/git/akpm/mm.git mm-everything
patch link: https://lore.kernel.org/r/20230313124526.1207490-2-fengwei.yin%40intel.com
patch subject: [PATCH v4 1/5] rmap: move hugetlb try_to_unmap to dedicated function
config: x86_64-randconfig-a012-20230313 (https://download.01.org/0day-ci/archive/20230314/202303140156.ZvGUC2dc-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/c43d03caaff9b0b9a869fd2a9b8beed7bc833f2a
git remote add linux-review https://github.com/intel-lab-lkp/linux
git fetch --no-tags linux-review Yin-Fengwei/rmap-move-hugetlb-try_to_unmap-to-dedicated-function/20230313-204656
git checkout c43d03caaff9b0b9a869fd2a9b8beed7bc833f2a
# save the config file
mkdir build_dir && cp config build_dir/.config
COMPILER_INSTALL_PATH=$HOME/0day COMPILER=clang make.cross W=1 O=build_dir ARCH=x86_64 olddefconfig
COMPILER_INSTALL_PATH=$HOME/0day COMPILER=clang make.cross W=1 O=build_dir ARCH=x86_64 SHELL=/bin/bash
If you fix the issue, kindly add following tag where applicable
| Reported-by: kernel test robot <lkp@intel.com>
| Link: https://lore.kernel.org/oe-kbuild-all/202303140156.ZvGUC2dc-lkp@intel.com/
All errors (new ones prefixed by >>):
>> mm/rmap.c:1513:6: error: implicit declaration of function 'huge_pte_dirty' is invalid in C99 [-Werror,-Wimplicit-function-declaration]
if (huge_pte_dirty(pteval))
^
1 error generated.
vim +/huge_pte_dirty +1513 mm/rmap.c
1443
1444 static bool try_to_unmap_one_hugetlb(struct folio *folio,
1445 struct vm_area_struct *vma, struct mmu_notifier_range range,
1446 struct page_vma_mapped_walk pvmw, unsigned long address,
1447 enum ttu_flags flags)
1448 {
1449 struct mm_struct *mm = vma->vm_mm;
1450 pte_t pteval;
1451 bool ret = true, anon = folio_test_anon(folio);
1452
1453 /*
1454 * The try_to_unmap() is only passed a hugetlb page
1455 * in the case where the hugetlb page is poisoned.
1456 */
1457 VM_BUG_ON_FOLIO(!folio_test_hwpoison(folio), folio);
1458 /*
1459 * huge_pmd_unshare may unmap an entire PMD page.
1460 * There is no way of knowing exactly which PMDs may
1461 * be cached for this mm, so we must flush them all.
1462 * start/end were already adjusted in caller
1463 * (try_to_unmap_one) to cover this range.
1464 */
1465 flush_cache_range(vma, range.start, range.end);
1466
1467 /*
1468 * To call huge_pmd_unshare, i_mmap_rwsem must be
1469 * held in write mode. Caller needs to explicitly
1470 * do this outside rmap routines.
1471 *
1472 * We also must hold hugetlb vma_lock in write mode.
1473 * Lock order dictates acquiring vma_lock BEFORE
1474 * i_mmap_rwsem. We can only try lock here and fail
1475 * if unsuccessful.
1476 */
1477 if (!anon) {
1478 VM_BUG_ON(!(flags & TTU_RMAP_LOCKED));
1479 if (!hugetlb_vma_trylock_write(vma)) {
1480 ret = false;
1481 goto out;
1482 }
1483 if (huge_pmd_unshare(mm, vma, address, pvmw.pte)) {
1484 hugetlb_vma_unlock_write(vma);
1485 flush_tlb_range(vma,
1486 range.start, range.end);
1487 mmu_notifier_invalidate_range(mm,
1488 range.start, range.end);
1489 /*
1490 * The ref count of the PMD page was
1491 * dropped which is part of the way map
1492 * counting is done for shared PMDs.
1493 * Return 'true' here. When there is
1494 * no other sharing, huge_pmd_unshare
1495 * returns false and we will unmap the
1496 * actual page and drop map count
1497 * to zero.
1498 */
1499 goto out;
1500 }
1501 hugetlb_vma_unlock_write(vma);
1502 }
1503 pteval = huge_ptep_clear_flush(vma, address, pvmw.pte);
1504
1505 /*
1506 * Now the pte is cleared. If this pte was uffd-wp armed,
1507 * we may want to replace a none pte with a marker pte if
1508 * it's file-backed, so we don't lose the tracking info.
1509 */
1510 pte_install_uffd_wp_if_needed(vma, address, pvmw.pte, pteval);
1511
1512 /* Set the dirty flag on the folio now the pte is gone. */
> 1513 if (huge_pte_dirty(pteval))
1514 folio_mark_dirty(folio);
1515
1516 /* Update high watermark before we lower rss */
1517 update_hiwater_rss(mm);
1518
1519 /* Poisoned hugetlb folio with TTU_HWPOISON always cleared in flags */
1520 pteval = swp_entry_to_pte(make_hwpoison_entry(&folio->page));
1521 set_huge_pte_at(mm, address, pvmw.pte, pteval);
1522 hugetlb_count_sub(folio_nr_pages(folio), mm);
1523
1524 /*
1525 * No need to call mmu_notifier_invalidate_range() it has be
1526 * done above for all cases requiring it to happen under page
1527 * table lock before mmu_notifier_invalidate_range_end()
1528 *
1529 * See Documentation/mm/mmu_notifier.rst
1530 */
1531 page_remove_rmap(&folio->page, vma, true);
1532 /* No VM_LOCKED set in vma->vm_flags for hugetlb. So not
1533 * necessary to call mlock_drain_local().
1534 */
1535 folio_put(folio);
1536
1537 out:
1538 return ret;
1539 }
1540
--
0-DAY CI Kernel Test Service
https://github.com/intel/lkp-tests
^ permalink raw reply [flat|nested] 2+ messages in thread