Andrew has two typo fix, should I resend all patch with fix-patch squashed?
+ mm-memory-failure-add-unmap_posioned_folio-fix.patch added to
mm-unstable branch
+
mm-memory_hotplug-check-hwpoisoned-page-firstly-in-do_migrate_range-fix.patch
added to mm-unstable branch
On 2024/8/26 0:48, kernel test robot wrote:
tree: https://git.kernel.org/pub/scm/linux/kernel/git/next/linux-next.git master
head: c79c85875f1af04040fe4492ed94ce37ad729c4d
commit: fb3b45758493a2451a9fdcf7f8c034d614d85939 [6060/6439] mm: memory_hotplug: check hwpoisoned page firstly in do_migrate_range()
config: x86_64-rhel-8.3-ltp (https://download.01.org/0day-ci/archive/20240826/202408260054.ka3zWBLH-lkp@xxxxxxxxx/config)
compiler: gcc-12 (Debian 12.2.0-14) 12.2.0
reproduce (this is a W=1 build): (https://download.01.org/0day-ci/archive/20240826/202408260054.ka3zWBLH-lkp@xxxxxxxxx/reproduce)
If you fix the issue in a separate patch/commit (i.e. not just a new version of
the same patch/commit), kindly add following tags
| Reported-by: kernel test robot <lkp@xxxxxxxxx>
| Closes: https://lore.kernel.org/oe-kbuild-all/202408260054.ka3zWBLH-lkp@xxxxxxxxx/
Note: the linux-next/master HEAD c79c85875f1af04040fe4492ed94ce37ad729c4d builds fine.
It may have been fixed somewhere.
All errors (new ones prefixed by >>):
mm/memory_hotplug.c: In function 'do_migrate_range':
mm/memory_hotplug.c:1811:33: error: implicit declaration of function 'unmap_posioned_folio'; did you mean 'unmap_poisoned_folio'? [-Werror=implicit-function-declaration]
1811 | unmap_posioned_folio(folio, TTU_IGNORE_MLOCK);
| ^~~~~~~~~~~~~~~~~~~~
| unmap_poisoned_folio
cc1: some warnings being treated as errors
vim +1811 mm/memory_hotplug.c
1772
1773 static void do_migrate_range(unsigned long start_pfn, unsigned long end_pfn)
1774 {
1775 unsigned long pfn;
1776 struct page *page;
1777 LIST_HEAD(source);
1778 static DEFINE_RATELIMIT_STATE(migrate_rs, DEFAULT_RATELIMIT_INTERVAL,
1779 DEFAULT_RATELIMIT_BURST);
1780
1781 for (pfn = start_pfn; pfn < end_pfn; pfn++) {
1782 struct folio *folio;
1783 bool isolated;
1784
1785 if (!pfn_valid(pfn))
1786 continue;
1787 page = pfn_to_page(pfn);
1788 folio = page_folio(page);
1789
1790 /*
1791 * No reference or lock is held on the folio, so it might
1792 * be modified concurrently (e.g. split). As such,
1793 * folio_nr_pages() may read garbage. This is fine as the outer
1794 * loop will revisit the split folio later.
1795 */
1796 if (folio_test_large(folio))
1797 pfn = folio_pfn(folio) + folio_nr_pages(folio) - 1;
1798
1799 /*
1800 * HWPoison pages have elevated reference counts so the migration would
1801 * fail on them. It also doesn't make any sense to migrate them in the
1802 * first place. Still try to unmap such a page in case it is still mapped
1803 * (e.g. current hwpoison implementation doesn't unmap KSM pages but keep
1804 * the unmap as the catch all safety net).
1805 */
1806 if (folio_test_hwpoison(folio) ||
1807 (folio_test_large(folio) && folio_test_has_hwpoisoned(folio))) {
1808 if (WARN_ON(folio_test_lru(folio)))
1809 folio_isolate_lru(folio);
1810 if (folio_mapped(folio))
1811 unmap_posioned_folio(folio, TTU_IGNORE_MLOCK);
1812 continue;
1813 }
1814
1815 if (folio_test_hugetlb(folio)) {
1816 isolate_hugetlb(folio, &source);
1817 continue;
1818 }
1819
1820 if (!get_page_unless_zero(page))
1821 continue;
1822 /*
1823 * We can skip free pages. And we can deal with pages on
1824 * LRU and non-lru movable pages.
1825 */
1826 if (PageLRU(page))
1827 isolated = isolate_lru_page(page);
1828 else
1829 isolated = isolate_movable_page(page, ISOLATE_UNEVICTABLE);
1830 if (isolated) {
1831 list_add_tail(&page->lru, &source);
1832 if (!__PageMovable(page))
1833 inc_node_page_state(page, NR_ISOLATED_ANON +
1834 page_is_file_lru(page));
1835
1836 } else {
1837 if (__ratelimit(&migrate_rs)) {
1838 pr_warn("failed to isolate pfn %lx\n", pfn);
1839 dump_page(page, "isolation failed");
1840 }
1841 }
1842 put_page(page);
1843 }
1844 if (!list_empty(&source)) {
1845 nodemask_t nmask = node_states[N_MEMORY];
1846 struct migration_target_control mtc = {
1847 .nmask = &nmask,
1848 .gfp_mask = GFP_USER | __GFP_MOVABLE | __GFP_RETRY_MAYFAIL,
1849 .reason = MR_MEMORY_HOTPLUG,
1850 };
1851 int ret;
1852
1853 /*
1854 * We have checked that migration range is on a single zone so
1855 * we can use the nid of the first page to all the others.
1856 */
1857 mtc.nid = page_to_nid(list_first_entry(&source, struct page, lru));
1858
1859 /*
1860 * try to allocate from a different node but reuse this node
1861 * if there are no other online nodes to be used (e.g. we are
1862 * offlining a part of the only existing node)
1863 */
1864 node_clear(mtc.nid, nmask);
1865 if (nodes_empty(nmask))
1866 node_set(mtc.nid, nmask);
1867 ret = migrate_pages(&source, alloc_migration_target, NULL,
1868 (unsigned long)&mtc, MIGRATE_SYNC, MR_MEMORY_HOTPLUG, NULL);
1869 if (ret) {
1870 list_for_each_entry(page, &source, lru) {
1871 if (__ratelimit(&migrate_rs)) {
1872 pr_warn("migrating pfn %lx failed ret:%d\n",
1873 page_to_pfn(page), ret);
1874 dump_page(page, "migration failure");
1875 }
1876 }
1877 putback_movable_pages(&source);
1878 }
1879 }
1880 }
1881
