Re: [PATCH v5 11/11] mm: memory_hotplug: memory hotremove supports thp migration
From: Anshuman Khandual <hidden>
Date: 2017-05-19 13:57:39
Also in:
lkml
On 04/21/2017 02:17 AM, Zi Yan wrote:
quoted hunk ↗ jump to hunk
From: Naoya Horiguchi <redacted> This patch enables thp migration for memory hotremove. Signed-off-by: Naoya Horiguchi <redacted> --- ChangeLog v1->v2: - base code switched from alloc_migrate_target to new_node_page() --- include/linux/huge_mm.h | 8 ++++++++ mm/memory_hotplug.c | 17 ++++++++++++++--- 2 files changed, 22 insertions(+), 3 deletions(-)diff --git a/include/linux/huge_mm.h b/include/linux/huge_mm.h index 6f44a2352597..92c2161704c3 100644 --- a/include/linux/huge_mm.h +++ b/include/linux/huge_mm.h@@ -189,6 +189,13 @@ static inline int hpage_nr_pages(struct page *page) return 1; } +static inline int hpage_order(struct page *page) +{ + if (unlikely(PageTransHuge(page))) + return HPAGE_PMD_ORDER; + return 0; +} +
This function seems to be redundant.
quoted hunk ↗ jump to hunk
struct page *follow_devmap_pmd(struct vm_area_struct *vma, unsigned long addr, pmd_t *pmd, int flags); struct page *follow_devmap_pud(struct vm_area_struct *vma, unsigned long addr,@@ -233,6 +240,7 @@ static inline bool thp_migration_supported(void) #define HPAGE_PUD_SIZE ({ BUILD_BUG(); 0; }) #define hpage_nr_pages(x) 1 +#define hpage_order(x) 0 #define transparent_hugepage_enabled(__vma) 0diff --git a/mm/memory_hotplug.c b/mm/memory_hotplug.c index 257166ebdff0..ecae0852994f 100644 --- a/mm/memory_hotplug.c +++ b/mm/memory_hotplug.c@@ -1574,6 +1574,7 @@ static struct page *new_node_page(struct page *page, unsigned long private, int nid = page_to_nid(page); nodemask_t nmask = node_states[N_MEMORY]; struct page *new_page = NULL; + unsigned int order = 0; /* * TODO: allocate a destination hugepage from a nearest neighbor node,@@ -1584,6 +1585,11 @@ static struct page *new_node_page(struct page *page, unsigned long private, return alloc_huge_page_node(page_hstate(compound_head(page)), next_node_in(nid, nmask)); + if (thp_migration_supported() && PageTransHuge(page)) { + order = hpage_order(page);
We have already tested the page as THP, we can just use HPAGE_PMD_ORDER.
quoted hunk ↗ jump to hunk
+ gfp_mask |= GFP_TRANSHUGE; + } + node_clear(nid, nmask); if (PageHighMem(page)@@ -1591,12 +1597,15 @@ static struct page *new_node_page(struct page *page, unsigned long private, gfp_mask |= __GFP_HIGHMEM; if (!nodes_empty(nmask)) - new_page = __alloc_pages_nodemask(gfp_mask, 0, + new_page = __alloc_pages_nodemask(gfp_mask, order, node_zonelist(nid, gfp_mask), &nmask); if (!new_page) - new_page = __alloc_pages(gfp_mask, 0, + new_page = __alloc_pages(gfp_mask, order, node_zonelist(nid, gfp_mask)); + if (new_page && order == hpage_order(page)) + prep_transhuge_page(new_page); +
new_page has been allocated with 'order' already. I guess just checking for PageTransHuge(page) on the old THP 'page' should be sufficient as that has not been changed in any way. -- To unsubscribe, send a message with 'unsubscribe linux-mm' in the body to majordomo@kvack.org. For more info on Linux MM, see: http://www.linux-mm.org/ . Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>