| Message ID | 20201213154534.54826-5-songmuchun@bytedance.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
| Series | Free some vmemmap pages of HugeTLB page | expand |
On 12/13/20 7:45 AM, Muchun Song wrote: > In the subsequent patch, we will allocate the vmemmap pages when free > HugeTLB pages. But update_and_free_page() is called from a non-task > context(and hold hugetlb_lock), so we can defer the actual freeing in > a workqueue to prevent use GFP_ATOMIC to allocate the vmemmap pages. > > Signed-off-by: Muchun Song <songmuchun@bytedance.com> It is unfortunate we need to add this complexitty, but I can not think of another way. One small comment (no required change) below. Reviewed-by: Mike Kravetz <mike.kravetz@oracle.com> > --- > mm/hugetlb.c | 77 ++++++++++++++++++++++++++++++++++++++++++++++++---- > mm/hugetlb_vmemmap.c | 12 -------- > mm/hugetlb_vmemmap.h | 17 ++++++++++++ > 3 files changed, 88 insertions(+), 18 deletions(-) > > diff --git a/mm/hugetlb.c b/mm/hugetlb.c > index 140135fc8113..0ff9b90e524f 100644 > --- a/mm/hugetlb.c > +++ b/mm/hugetlb.c > @@ -1292,15 +1292,76 @@ static inline void destroy_compound_gigantic_page(struct page *page, > unsigned int order) { } > #endif > > -static void update_and_free_page(struct hstate *h, struct page *page) > +static void __free_hugepage(struct hstate *h, struct page *page); > + > +/* > + * As update_and_free_page() is be called from a non-task context(and hold > + * hugetlb_lock), we can defer the actual freeing in a workqueue to prevent > + * use GFP_ATOMIC to allocate a lot of vmemmap pages. > + * > + * update_hpage_vmemmap_workfn() locklessly retrieves the linked list of > + * pages to be freed and frees them one-by-one. As the page->mapping pointer > + * is going to be cleared in update_hpage_vmemmap_workfn() anyway, it is > + * reused as the llist_node structure of a lockless linked list of huge > + * pages to be freed. > + */ > +static LLIST_HEAD(hpage_update_freelist); > + > +static void update_hpage_vmemmap_workfn(struct work_struct *work) > { > - int i; > + struct llist_node *node; > + struct page *page; > + > + node = llist_del_all(&hpage_update_freelist); > > + while (node) { > + page = container_of((struct address_space **)node, > + struct page, mapping); > + node = node->next; > + page->mapping = NULL; > + __free_hugepage(page_hstate(page), page); > + > + cond_resched(); > + } > +} > +static DECLARE_WORK(hpage_update_work, update_hpage_vmemmap_workfn); > + > +static inline void __update_and_free_page(struct hstate *h, struct page *page) > +{ > + /* No need to allocate vmemmap pages */ > + if (!free_vmemmap_pages_per_hpage(h)) { > + __free_hugepage(h, page); > + return; > + } > + > + /* > + * Defer freeing to avoid using GFP_ATOMIC to allocate vmemmap > + * pages. > + * > + * Only call schedule_work() if hpage_update_freelist is previously > + * empty. Otherwise, schedule_work() had been called but the workfn > + * hasn't retrieved the list yet. > + */ > + if (llist_add((struct llist_node *)&page->mapping, > + &hpage_update_freelist)) > + schedule_work(&hpage_update_work); > +} > + > +static void update_and_free_page(struct hstate *h, struct page *page) > +{ > if (hstate_is_gigantic(h) && !gigantic_page_runtime_supported()) > return; > > h->nr_huge_pages--; > h->nr_huge_pages_node[page_to_nid(page)]--; > + > + __update_and_free_page(h, page); > +} > + > +static void __free_hugepage(struct hstate *h, struct page *page) > +{ > + int i; > + Can we add a comment here saying that this is where the call to allocate vmemmmap pages will be inserted in a later patch. Such a comment would help a bit to understand the restructuring of the code.
On Thu, Dec 17, 2020 at 7:48 AM Mike Kravetz <mike.kravetz@oracle.com> wrote: > > On 12/13/20 7:45 AM, Muchun Song wrote: > > In the subsequent patch, we will allocate the vmemmap pages when free > > HugeTLB pages. But update_and_free_page() is called from a non-task > > context(and hold hugetlb_lock), so we can defer the actual freeing in > > a workqueue to prevent use GFP_ATOMIC to allocate the vmemmap pages. > > > > Signed-off-by: Muchun Song <songmuchun@bytedance.com> > > It is unfortunate we need to add this complexitty, but I can not think > of another way. One small comment (no required change) below. > > Reviewed-by: Mike Kravetz <mike.kravetz@oracle.com> Thank you. > > > --- > > mm/hugetlb.c | 77 ++++++++++++++++++++++++++++++++++++++++++++++++---- > > mm/hugetlb_vmemmap.c | 12 -------- > > mm/hugetlb_vmemmap.h | 17 ++++++++++++ > > 3 files changed, 88 insertions(+), 18 deletions(-) > > > > diff --git a/mm/hugetlb.c b/mm/hugetlb.c > > index 140135fc8113..0ff9b90e524f 100644 > > --- a/mm/hugetlb.c > > +++ b/mm/hugetlb.c > > @@ -1292,15 +1292,76 @@ static inline void destroy_compound_gigantic_page(struct page *page, > > unsigned int order) { } > > #endif > > > > -static void update_and_free_page(struct hstate *h, struct page *page) > > +static void __free_hugepage(struct hstate *h, struct page *page); > > + > > +/* > > + * As update_and_free_page() is be called from a non-task context(and hold > > + * hugetlb_lock), we can defer the actual freeing in a workqueue to prevent > > + * use GFP_ATOMIC to allocate a lot of vmemmap pages. > > + * > > + * update_hpage_vmemmap_workfn() locklessly retrieves the linked list of > > + * pages to be freed and frees them one-by-one. As the page->mapping pointer > > + * is going to be cleared in update_hpage_vmemmap_workfn() anyway, it is > > + * reused as the llist_node structure of a lockless linked list of huge > > + * pages to be freed. > > + */ > > +static LLIST_HEAD(hpage_update_freelist); > > + > > +static void update_hpage_vmemmap_workfn(struct work_struct *work) > > { > > - int i; > > + struct llist_node *node; > > + struct page *page; > > + > > + node = llist_del_all(&hpage_update_freelist); > > > > + while (node) { > > + page = container_of((struct address_space **)node, > > + struct page, mapping); > > + node = node->next; > > + page->mapping = NULL; > > + __free_hugepage(page_hstate(page), page); > > + > > + cond_resched(); > > + } > > +} > > +static DECLARE_WORK(hpage_update_work, update_hpage_vmemmap_workfn); > > + > > +static inline void __update_and_free_page(struct hstate *h, struct page *page) > > +{ > > + /* No need to allocate vmemmap pages */ > > + if (!free_vmemmap_pages_per_hpage(h)) { > > + __free_hugepage(h, page); > > + return; > > + } > > + > > + /* > > + * Defer freeing to avoid using GFP_ATOMIC to allocate vmemmap > > + * pages. > > + * > > + * Only call schedule_work() if hpage_update_freelist is previously > > + * empty. Otherwise, schedule_work() had been called but the workfn > > + * hasn't retrieved the list yet. > > + */ > > + if (llist_add((struct llist_node *)&page->mapping, > > + &hpage_update_freelist)) > > + schedule_work(&hpage_update_work); > > +} > > + > > +static void update_and_free_page(struct hstate *h, struct page *page) > > +{ > > if (hstate_is_gigantic(h) && !gigantic_page_runtime_supported()) > > return; > > > > h->nr_huge_pages--; > > h->nr_huge_pages_node[page_to_nid(page)]--; > > + > > + __update_and_free_page(h, page); > > +} > > + > > +static void __free_hugepage(struct hstate *h, struct page *page) > > +{ > > + int i; > > + > > Can we add a comment here saying that this is where the call to allocate > vmemmmap pages will be inserted in a later patch. Such a comment would > help a bit to understand the restructuring of the code. OK. Will do. Thanks. > > -- > Mike Kravetz > > > for (i = 0; i < pages_per_huge_page(h); i++) { > > page[i].flags &= ~(1 << PG_locked | 1 << PG_error | > > 1 << PG_referenced | 1 << PG_dirty | > > @@ -1313,13 +1374,17 @@ static void update_and_free_page(struct hstate *h, struct page *page) > > set_page_refcounted(page); > > if (hstate_is_gigantic(h)) { > > /* > > - * Temporarily drop the hugetlb_lock, because > > - * we might block in free_gigantic_page(). > > + * Temporarily drop the hugetlb_lock only when this type of > > + * HugeTLB page does not support vmemmap optimization (which > > + * contex do not hold the hugetlb_lock), because we might block > > + * in free_gigantic_page(). > > */ > > - spin_unlock(&hugetlb_lock); > > + if (!free_vmemmap_pages_per_hpage(h)) > > + spin_unlock(&hugetlb_lock); > > destroy_compound_gigantic_page(page, huge_page_order(h)); > > free_gigantic_page(page, huge_page_order(h)); > > - spin_lock(&hugetlb_lock); > > + if (!free_vmemmap_pages_per_hpage(h)) > > + spin_lock(&hugetlb_lock); > > } else { > > __free_pages(page, huge_page_order(h)); > > }
diff --git a/mm/hugetlb.c b/mm/hugetlb.c index 140135fc8113..0ff9b90e524f 100644 --- a/mm/hugetlb.c +++ b/mm/hugetlb.c @@ -1292,15 +1292,76 @@ static inline void destroy_compound_gigantic_page(struct page *page, unsigned int order) { } #endif -static void update_and_free_page(struct hstate *h, struct page *page) +static void __free_hugepage(struct hstate *h, struct page *page); + +/* + * As update_and_free_page() is be called from a non-task context(and hold + * hugetlb_lock), we can defer the actual freeing in a workqueue to prevent + * use GFP_ATOMIC to allocate a lot of vmemmap pages. + * + * update_hpage_vmemmap_workfn() locklessly retrieves the linked list of + * pages to be freed and frees them one-by-one. As the page->mapping pointer + * is going to be cleared in update_hpage_vmemmap_workfn() anyway, it is + * reused as the llist_node structure of a lockless linked list of huge + * pages to be freed. + */ +static LLIST_HEAD(hpage_update_freelist); + +static void update_hpage_vmemmap_workfn(struct work_struct *work) { - int i; + struct llist_node *node; + struct page *page; + + node = llist_del_all(&hpage_update_freelist); + while (node) { + page = container_of((struct address_space **)node, + struct page, mapping); + node = node->next; + page->mapping = NULL; + __free_hugepage(page_hstate(page), page); + + cond_resched(); + } +} +static DECLARE_WORK(hpage_update_work, update_hpage_vmemmap_workfn); + +static inline void __update_and_free_page(struct hstate *h, struct page *page) +{ + /* No need to allocate vmemmap pages */ + if (!free_vmemmap_pages_per_hpage(h)) { + __free_hugepage(h, page); + return; + } + + /* + * Defer freeing to avoid using GFP_ATOMIC to allocate vmemmap + * pages. + * + * Only call schedule_work() if hpage_update_freelist is previously + * empty. Otherwise, schedule_work() had been called but the workfn + * hasn't retrieved the list yet. + */ + if (llist_add((struct llist_node *)&page->mapping, + &hpage_update_freelist)) + schedule_work(&hpage_update_work); +} + +static void update_and_free_page(struct hstate *h, struct page *page) +{ if (hstate_is_gigantic(h) && !gigantic_page_runtime_supported()) return; h->nr_huge_pages--; h->nr_huge_pages_node[page_to_nid(page)]--; + + __update_and_free_page(h, page); +} + +static void __free_hugepage(struct hstate *h, struct page *page) +{ + int i; + for (i = 0; i < pages_per_huge_page(h); i++) { page[i].flags &= ~(1 << PG_locked | 1 << PG_error | 1 << PG_referenced | 1 << PG_dirty | @@ -1313,13 +1374,17 @@ static void update_and_free_page(struct hstate *h, struct page *page) set_page_refcounted(page); if (hstate_is_gigantic(h)) { /* - * Temporarily drop the hugetlb_lock, because - * we might block in free_gigantic_page(). + * Temporarily drop the hugetlb_lock only when this type of + * HugeTLB page does not support vmemmap optimization (which + * contex do not hold the hugetlb_lock), because we might block + * in free_gigantic_page(). */ - spin_unlock(&hugetlb_lock); + if (!free_vmemmap_pages_per_hpage(h)) + spin_unlock(&hugetlb_lock); destroy_compound_gigantic_page(page, huge_page_order(h)); free_gigantic_page(page, huge_page_order(h)); - spin_lock(&hugetlb_lock); + if (!free_vmemmap_pages_per_hpage(h)) + spin_lock(&hugetlb_lock); } else { __free_pages(page, huge_page_order(h)); } diff --git a/mm/hugetlb_vmemmap.c b/mm/hugetlb_vmemmap.c index 5a714bd60d6b..6d4e77a2b6c7 100644 --- a/mm/hugetlb_vmemmap.c +++ b/mm/hugetlb_vmemmap.c @@ -180,18 +180,6 @@ #define RESERVE_VMEMMAP_NR 2U #define RESERVE_VMEMMAP_SIZE (RESERVE_VMEMMAP_NR << PAGE_SHIFT) -/* - * How many vmemmap pages associated with a HugeTLB page that can be freed - * to the buddy allocator. - * - * Todo: Returns zero for now, which means the feature is disabled. We will - * enable it once all the infrastructure is there. - */ -static inline unsigned int free_vmemmap_pages_per_hpage(struct hstate *h) -{ - return 0; -} - static inline unsigned long free_vmemmap_pages_size_per_hpage(struct hstate *h) { return (unsigned long)free_vmemmap_pages_per_hpage(h) << PAGE_SHIFT; diff --git a/mm/hugetlb_vmemmap.h b/mm/hugetlb_vmemmap.h index 6923f03534d5..01f8637adbe0 100644 --- a/mm/hugetlb_vmemmap.h +++ b/mm/hugetlb_vmemmap.h @@ -12,9 +12,26 @@ #ifdef CONFIG_HUGETLB_PAGE_FREE_VMEMMAP void free_huge_page_vmemmap(struct hstate *h, struct page *head); + +/* + * How many vmemmap pages associated with a HugeTLB page that can be freed + * to the buddy allocator. + * + * Todo: Returns zero for now, which means the feature is disabled. We will + * enable it once all the infrastructure is there. + */ +static inline unsigned int free_vmemmap_pages_per_hpage(struct hstate *h) +{ + return 0; +} #else static inline void free_huge_page_vmemmap(struct hstate *h, struct page *head) { } + +static inline unsigned int free_vmemmap_pages_per_hpage(struct hstate *h) +{ + return 0; +} #endif /* CONFIG_HUGETLB_PAGE_FREE_VMEMMAP */ #endif /* _LINUX_HUGETLB_VMEMMAP_H */
In the subsequent patch, we will allocate the vmemmap pages when free HugeTLB pages. But update_and_free_page() is called from a non-task context(and hold hugetlb_lock), so we can defer the actual freeing in a workqueue to prevent use GFP_ATOMIC to allocate the vmemmap pages. Signed-off-by: Muchun Song <songmuchun@bytedance.com> --- mm/hugetlb.c | 77 ++++++++++++++++++++++++++++++++++++++++++++++++---- mm/hugetlb_vmemmap.c | 12 -------- mm/hugetlb_vmemmap.h | 17 ++++++++++++ 3 files changed, 88 insertions(+), 18 deletions(-)