* mm/kvm/vfio/ppc64: Migrate compound pages out of CMA region
Subsystems (4)
v4 cover * mm/kvm/vfio/ppc64: Migrate compound pages out of CMA region
ppc64 use CMA area for the allocation of guest page table (hash page table). We won't be able to start guest if we fail to allocate hash page table. We have observed hash table allocation failure because we failed to migrate pages out of CMA region because they were pinned. This happen when we are using VFIO. VFIO on ppc64 pins the entire guest RAM. If the guest RAM pages get allocated out of CMA region, we won't be able to migrate those pages. The pages are also pinned for the lifetime of the guest.
Currently we support migration of non-compound pages. With THP and with the addition of
hugetlb migration we can end up allocating compound pages from CMA region. This
patch series add support for migrating compound pages. The first path adds the helper get_user_pages_cma_migrate() which pin the page making sure we migrate them out of CMA region before incrementing the reference count.
Changes from V3:
* Move the hugetlb check before transhuge check
* Use compound head page when isolating hugetlb page
Aneesh Kumar K.V (3):
mm: Add get_user_pages_cma_migrate
powerpc/mm/iommu: Allow migration of cma allocated pages during
mm_iommu_get
powerpc/mm/iommu: Allow large IOMMU page size only for hugetlb backing
arch/powerpc/mm/mmu_context_iommu.c | 140 ++++++++-------------------- include/linux/hugetlb.h | 2 + include/linux/migrate.h | 3 + mm/hugetlb.c | 4 +- mm/migrate.c | 132 ++++++++++++++++++++++++++ 5 files changed, 179 insertions(+), 102 deletions(-)
Signature · 1 line
2.17.2
PATCH v4 1/3 mm: Add get_user_pages_cma_migrate
This helper does a get_user_pages_fast and if it find pages in the CMA area it will try to migrate them before taking page reference. This makes sure that we don't keep non-movable pages (due to page reference count) in the CMA area. Not able to move pages out of CMA area result in CMA allocation failures.
--- include/linux/hugetlb.h | 2 + include/linux/migrate.h | 3 + mm/hugetlb.c | 4 +- mm/migrate.c | 132 ++++++++++++++++++++++++++++++++++++++++ 4 files changed, 139 insertions(+), 2 deletions(-)
include/linux/hugetlb.h +2 -0
@@ -371,6 +371,8 @@ struct page *alloc_huge_page_nodemask(struct hstate *h, int preferred_nid, nodemask_t *nmask);struct page *alloc_huge_page_vma (struct hstate *h,struct vm_area_struct *vma,unsigned long address); +struct page *alloc_migrate_huge_page (struct hstate *h,gfp_t gfp_mask, +int nid,nodemask_t *nmask);int huge_add_to_page_cache (struct page *page,struct address_space *mapping,pgoff_t idx);
include/linux/migrate.h +3 -0
@@ -286,6 +286,9 @@ static inline int migrate_vma(const struct migrate_vma_ops *ops, }#endif /* IS_ENABLED(CONFIG_MIGRATE_VMA_HELPER) */ +extern int get_user_pages_cma_migrate (unsigned long start,int nr_pages,int write, +struct page **pages); +#endif /* CONFIG_MIGRATION */ #endif /* _LINUX_MIGRATE_H */
mm/hugetlb.c +2 -2
@@ -1585,8 +1585,8 @@ static struct page *alloc_surplus_huge_page(struct hstate *h, gfp_t gfp_mask,return page; } -static struct page *alloc_migrate_huge_page (struct hstate *h,gfp_t gfp_mask, -int nid,nodemask_t *nmask) +struct page *alloc_migrate_huge_page (struct hstate *h,gfp_t gfp_mask, +int nid,nodemask_t *nmask) {struct page *page;
mm/migrate.c +132 -0
135 lines not on this page · raw message
Signature · 1 line
2.17.2
PATCH v4 2/3 powerpc/mm/iommu: Allow migration of cma allocated pages during mm_iommu_get
In reply to Aneesh Kumar K.VCurrent code doesn't do page migration if the page allocated is a compound page. With HugeTLB migration support, we can end up allocating hugetlb pages from CMA region. Also THP pages can be allocated from CMA region. This patch updates the code to handle compound pages correctly.
This use the new helper get_user_pages_cma_migrate. It does one get_user_pages with right count, instead of doing one get_user_pages per page. That avoids reading page table multiple times.
The patch also convert the hpas member of mm_iommu_table_group_mem_t to a union. We use the same storage location to store pointers to struct page. We cannot update alll the code path use struct page *, because we access hpas in real mode and we can't do that struct page * to pfn conversion in real mode.
--- arch/powerpc/mm/mmu_context_iommu.c | 120 ++++++++-------------------- 1 file changed, 35 insertions(+), 85 deletions(-)
arch/powerpc/mm/mmu_context_iommu.c +35 -85
@@ -21,6 +21,7 @@#include <linux/sizes.h> #include <asm/mmu_context.h> #include <asm/pte-walk.h> +#include <linux/mm_inline.h> static DEFINE_MUTEX (mem_list_mutex);
@@ -34,8 +35,18 @@ struct mm_iommu_table_group_mem_t {atomic64_t mapped;unsigned int pageshift;u64 ua;/* userspace address */ -u64 entries;/* number of entries in hpas[] */ -u64 *hpas;/* vmalloc'ed */ +u64 entries;/* number of entries in hpages[] */ +/* +* in mm_iommu_get we temporarily use this to store +* struct page address. +* +* We need to convert ua to hpa in real mode. Make it +* simpler by storing physicall address. +*/ +union { +struct page **hpages;/* vmalloc'ed */ +phys_addr_t *hpas; + }; }; static long mm_iommu_adjust_locked_vm(struct mm_struct *mm,
@@ -78,63 +89,14 @@ bool mm_iommu_preregistered(struct mm_struct *mm) }EXPORT_SYMBOL_GPL (mm_iommu_preregistered); -/* -* Taken from alloc_migrate_target with changes to remove CMA allocations -*/ -struct page *new_iommu_non_cma_page (struct page *page,unsigned long private) -{ -gfp_t gfp_mask =GFP_USER ; -struct page *new_page; - -if (PageCompound (page)) -return NULL ; - -if (PageHighMem (page)) - gfp_mask |= __GFP_HIGHMEM; - -/* -* We don't want the allocation to force an OOM if possibe -*/ - new_page =alloc_page (gfp_mask | __GFP_NORETRY | __GFP_NOWARN); -return new_page; -} - -static int mm_iommu_move_page_from_cma (struct page *page) -{ -int ret =0 ; -LIST_HEAD (cma_migrate_pages); - -/* Ignore huge pages for now */ -if (PageCompound (page)) -return -EBUSY ; - -lru_add_drain (); - ret =isolate_lru_page (page); -if (ret) -return ret; - -list_add (&page->lru, &cma_migrate_pages); -put_page (page);/* Drop the gup reference */ - - ret =migrate_pages (&cma_migrate_pages, new_iommu_non_cma_page, -NULL ,0 ,MIGRATE_SYNC ,MR_CONTIG_RANGE ); -if (ret) { -if (!list_empty (&cma_migrate_pages)) -putback_movable_pages (&cma_migrate_pages); - } - -return 0 ; -} -long mm_iommu_get (struct mm_struct *mm,unsigned long ua,unsigned long entries,struct mm_iommu_table_group_mem_t **pmem) {struct mm_iommu_table_group_mem_t *mem; -long i, j, ret =0 , locked_entries =0 ; +long i, ret =0 , locked_entries =0 ;unsigned int pageshift;unsigned long flags;unsigned long cur_ua; -struct page *page =NULL ;mutex_lock (&mem_list_mutex);
@@ -181,41 +143,24 @@ long mm_iommu_get(struct mm_struct *mm, unsigned long ua, unsigned long entries, goto unlock_exit; } + ret =get_user_pages_cma_migrate (ua, entries,1 , mem->hpages); +if (ret != entries) { +/* free the reference taken */ +for (i =0 ; i < ret; i++) +put_page (mem->hpages[i]); + +vfree (mem->hpas); +kfree (mem); + ret = -EFAULT ; + goto unlock_exit; + }else + ret =0 ; + + pageshift =PAGE_SHIFT ;for (i =0 ; i < entries; ++i) { +struct page *page = mem->hpages[i]; cur_ua = ua + (i <<PAGE_SHIFT ); -if (1 !=get_user_pages_fast (cur_ua, -1 /* pages */ ,1 /* iswrite */ , &page)) { - ret = -EFAULT ; -for (j =0 ; j < i; ++j) -put_page (pfn_to_page (mem->hpas[j] >> -PAGE_SHIFT )); -vfree (mem->hpas); -kfree (mem); - goto unlock_exit; - } -/* -* If we get a page from the CMA zone, since we are going to -* be pinning these entries, we might as well move them out -* of the CMA zone if possible. NOTE: faulting in + migration -* can be expensive. Batching can be considered later -*/ -if (is_migrate_cma_page (page)) { -if (mm_iommu_move_page_from_cma (page)) - goto populate; -if (1 !=get_user_pages_fast (cur_ua, -1 /* pages */ ,1 /* iswrite */ , - &page)) { - ret = -EFAULT ; -for (j =0 ; j < i; ++j) -put_page (pfn_to_page (mem->hpas[j] >> -PAGE_SHIFT )); -vfree (mem->hpas); -kfree (mem); - goto unlock_exit; - } - } -populate: - pageshift =PAGE_SHIFT ; +if (mem->pageshift >PAGE_SHIFT &&PageCompound (page)) {pte_t *pte;struct page *head =compound_head (page);
@@ -233,7 +178,12 @@ long mm_iommu_get(struct mm_struct *mm, unsigned long ua, unsigned long entries,local_irq_restore (flags); } mem->pageshift =min (mem->pageshift, pageshift); +/* +* We don't need struct page reference any more, switch +* physicall address. +*/ mem->hpas[i] =page_to_pfn (page) <<PAGE_SHIFT ; + }atomic64_set (&mem->mapped,1 );
Signature · 1 line
2.17.2
PATCH v4 3/3 powerpc/mm/iommu: Allow large IOMMU page size only for hugetlb backing
In reply to Aneesh Kumar K.VTHP pages can get split during different code paths. An incremented reference count do imply we will not split the compound page. But the pmd entry can be converted to level 4 pte entries. Keep the code simpler by allowing large IOMMU page size only if the guest ram is backed by hugetlb pages.
--- arch/powerpc/mm/mmu_context_iommu.c | 24 +++++++----------------- 1 file changed, 7 insertions(+), 17 deletions(-)
arch/powerpc/mm/mmu_context_iommu.c +7 -17
@@ -95,8 +95,6 @@ long mm_iommu_get(struct mm_struct *mm, unsigned long ua, unsigned long entries,struct mm_iommu_table_group_mem_t *mem;long i, ret =0 , locked_entries =0 ;unsigned int pageshift; -unsigned long flags; -unsigned long cur_ua;mutex_lock (&mem_list_mutex);
@@ -159,23 +157,15 @@ long mm_iommu_get(struct mm_struct *mm, unsigned long ua, unsigned long entries, pageshift =PAGE_SHIFT ;for (i =0 ; i < entries; ++i) {struct page *page = mem->hpages[i]; - cur_ua = ua + (i <<PAGE_SHIFT ); -if (mem->pageshift >PAGE_SHIFT &&PageCompound (page)) { -pte_t *pte; +/* +* Allow to use larger than 64k IOMMU pages. Only do that +* if we are backed by hugetlb. +*/ +if ((mem->pageshift >PAGE_SHIFT ) &&PageHuge (page)) {struct page *head =compound_head (page); -unsigned int compshift =compound_order (head); -unsigned int pteshift; - -local_irq_save (flags);/* disables as well */ - pte =find_linux_pte (mm->pgd, cur_ua,NULL , &pteshift); - -/* Double check it is still the same pinned page */ -if (pte &&pte_page (*pte) == head && - pteshift == compshift +PAGE_SHIFT ) - pageshift = max_t(unsigned int, pteshift, -PAGE_SHIFT ); -local_irq_restore (flags); + + pageshift =compound_order (head) +PAGE_SHIFT ; } mem->pageshift =min (mem->pageshift, pageshift);/*
Signature · 1 line
2.17.2
Andrew Morton
maintainerQuoted 1 line · On Wed, 21 Nov 2018 14:52:56 +0530 "Aneesh Kumar K.V"…
Subject: [PATCH V4 0/3] *
mm/kvm/vfio/ppc64: Migrate compound pages out of CMA region
Asterisk in title is strange?
Quoted 8 lines
ppc64 use CMA area for the allocation of guest page table (hash page table). We won't be able to start guest if we fail to allocate hash page table. We have observed hash table allocation failure because we failed to migrate pages out of CMA region because they were pinned. This happen when we are using VFIO. VFIO on ppc64 pins the entire guest RAM. If the guest RAM pages get allocated out of CMA region, we won't be able to migrate those pages. The pages are also pinned for the lifetime of the guest.
Quoted 5 lines
Currently we support migration of non-compound pages. With THP and with the addition of
hugetlb migration we can end up allocating compound pages from CMA region. This
patch series add support for migrating compound pages. The first path adds the helperget_user_pages_cma_migrate()which pin the page making sure we migrate them out of CMA region before incrementing the reference count.
Very little review activity. Perhaps Andrey and/or Michal can find the time..
Quoted 1 line
mm/migrate.c| 108 ++++++++++++++++++++++++++++++++++++++++
can we make this code disappear when CONFIG_CMA=n?
Quoted 5 lines · Andrew Morton
Quoted 1 line · On Wed, 21 Nov 2018 14:52:56 +0530 "Aneesh Kumar K.V"…
Subject: [PATCH V4 0/3] *
mm/kvm/vfio/ppc64: Migrate compound pages out of CMA regionAsterisk in title is strange?
My mistake while editing git-format-patch cover-letter.
Quoted 15 lines
Quoted 9 lines
ppc64 use CMA area for the allocation of guest page table (hash page table). We won't be able to start guest if we fail to allocate hash page table. We have observed hash table allocation failure because we failed to migrate pages out of CMA region because they were pinned. This happen when we are using VFIO. VFIO on ppc64 pins the entire guest RAM. If the guest RAM pages get allocated out of CMA region, we won't be able to migrate those pages. The pages are also pinned for the lifetime of the guest.
Quoted 5 lines
Currently we support migration of non-compound pages. With THP and with the addition of
hugetlb migration we can end up allocating compound pages from CMA region. This
patch series add support for migrating compound pages. The first path adds the helperget_user_pages_cma_migrate()which pin the page making sure we migrate them out of CMA region before incrementing the reference count.
Quoted 7 lines
Very little review activity. Perhaps Andrey and/or Michal can find the time..
Quoted 1 line
mm/migrate.c| 108 ++++++++++++++++++++++++++++++++++++++++can we make this code disappear when
CONFIG_CMA=n?
We can definitely do
static inline int get_user_pages_cma_migrate(unsigned long start, int nr_pages, int write, struct page **pages)
{
return get_user_pages_fast(start, nr_pages, write, pages);
}with #ifdef CONFIG_CMA around but that is unnecessary #ifdef in the code. If CMA config is disabled, we will not be doing any migrate. Hence wondering whether we need an alternative definition for CONFIG_CMA=n
-aneesh
Michael Ellerman
reviewerRe: [PATCH V4 1/3] mm: Add get_user_pages_cma_migrate
In reply to Aneesh Kumar K.VQuoted 5 lines · Aneesh Kumar K.V
This helper does a
get_user_pages_fastand if it find pages in the CMA area it will try to migrate them before taking page reference. This makes sure that we don't keep non-movable pages (due to page reference count) in the CMA area. Not able to move pages out of CMA area result in CMA allocation failures.
Quoted 7 lines
Signed-off-by Aneesh Kumar K.V <[email protected]>--- include/linux/hugetlb.h | 2 + include/linux/migrate.h | 3 + mm/hugetlb.c | 4 +- mm/migrate.c | 132 ++++++++++++++++++++++++++++++++++++++++ 4 files changed, 139 insertions(+), 2 deletions(-)
I'd rather not merge this much mm/ code via the powerpc tree without acks.
Anyone?
cheers
Quoted 184 lines
include/linux/hugetlb.h +2
-0@@ -371,6 +371,8 @@ struct page *alloc_huge_page_nodemask(struct hstate *h, int preferred_nid, nodemask_t *nmask);struct page *alloc_huge_page_vma (struct hstate *h,struct vm_area_struct *vma,unsigned long address); +struct page *alloc_migrate_huge_page (struct hstate *h,gfp_t gfp_mask, +int nid,nodemask_t *nmask);int huge_add_to_page_cache (struct page *page,struct address_space *mapping,pgoff_t idx);include/linux/migrate.h +3
-0@@ -286,6 +286,9 @@ static inline int migrate_vma(const struct migrate_vma_ops *ops, }#endif /* IS_ENABLED(CONFIG_MIGRATE_VMA_HELPER) */ +extern int get_user_pages_cma_migrate (unsigned long start,int nr_pages,int write, +struct page **pages); +#endif /* CONFIG_MIGRATION */ #endif /* _LINUX_MIGRATE_H */ mm/hugetlb.c +2
-2@@ -1585,8 +1585,8 @@ static struct page *alloc_surplus_huge_page(struct hstate *h, gfp_t gfp_mask,return page; } -static struct page *alloc_migrate_huge_page (struct hstate *h,gfp_t gfp_mask, -int nid,nodemask_t *nmask) +struct page *alloc_migrate_huge_page (struct hstate *h,gfp_t gfp_mask, +int nid,nodemask_t *nmask) {struct page *page;mm/migrate.c +132
-0@@ -2946,3 +2946,135 @@ int migrate_vma(const struct migrate_vma_ops *ops, }EXPORT_SYMBOL (migrate_vma);#endif /* defined(MIGRATE_VMA_HELPER) */ + +static struct page *new_non_cma_page (struct page *page,unsigned long private) +{ +/* +* We want to make sure we allocate the new page from the same node +* as the source page. +*/ +int nid =page_to_nid (page); +gfp_t gfp_mask =GFP_USER | __GFP_THISNODE; + +if (PageHighMem (page)) + gfp_mask |= __GFP_HIGHMEM; + +#ifdef CONFIG_HUGETLB_PAGE +if (PageHuge (page)) { +struct hstate *h =page_hstate (page); +/* +* We don't want to dequeue from the pool because pool pages will +* mostly be from the CMA region. +*/ +return alloc_migrate_huge_page (h, gfp_mask, nid,NULL ); + } +#endif +if (PageTransHuge (page)) { +struct page *thp; +gfp_t thp_gfpmask =GFP_TRANSHUGE | __GFP_THISNODE; + +/* +* Remove the movable mask so that we don't allocate from +* CMA area again. +*/ + thp_gfpmask &= ~__GFP_MOVABLE; + thp =__alloc_pages_node (nid, thp_gfpmask,HPAGE_PMD_ORDER ); +if (!thp) +return NULL ; +prep_transhuge_page (thp); +return thp; + } + +return __alloc_pages_node (nid, gfp_mask,0 ); +} + +/** +* get_user_pages_cma_migrate() - pin user pages in memory by migrating pages in CMA region +* @start: starting user address +* @nr_pages: number of pages from start to pin +* @write: whether pages will be written to +* @pages: array that receives pointers to the pages pinned. +* Should be at least nr_pages long. +* +* Attempt to pin user pages in memory without taking mm->mmap_sem. +* If not successful, it will fall back to taking the lock and +* calling get_user_pages(). +* +* If the pinned pages are backed by CMA region, we migrate those pages out, +* allocating new pages from non-CMA region. This helps in avoiding keeping +* pages pinned in the CMA region for a long time thereby resulting in +* CMA allocation failures. +* +* Returns number of pages pinned. This may be fewer than the number +* requested. If nr_pages is 0 or negative, returns 0. If no pages +* were pinned, returns -errno. +*/ + +int get_user_pages_cma_migrate (unsigned long start,int nr_pages,int write, +struct page **pages) +{ +int i, ret; +bool drain_allow = true; +bool migrate_allow = true; +LIST_HEAD (cma_page_list); + +get_user_again: + ret =get_user_pages_fast (start, nr_pages, write, pages); +if (ret <=0 ) +return ret; + +for (i =0 ; i < ret; ++i) { +/* +* If we get a page from the CMA zone, since we are going to +* be pinning these entries, we might as well move them out +* of the CMA zone if possible. +*/ +if (is_migrate_cma_page (pages[i]) && migrate_allow) { + +struct page *head =compound_head (pages[i]); + +if (PageHuge (head)) +isolate_huge_page (head, &cma_page_list); +else { +if (!PageLRU (head) && drain_allow) { +lru_add_drain_all (); + drain_allow = false; + } + +if (!isolate_lru_page (head)) { +list_add_tail (&head->lru, &cma_page_list); +mod_node_page_state (page_pgdat (head), +NR_ISOLATED_ANON + +page_is_file_cache (head), +hpage_nr_pages (head)); + } + } + } + } +if (!list_empty (&cma_page_list)) { +/* +* drop the above get_user_pages reference. +*/ +for (i =0 ; i < ret; ++i) +put_page (pages[i]); + +if (migrate_pages (&cma_page_list, new_non_cma_page, +NULL ,0 ,MIGRATE_SYNC ,MR_CONTIG_RANGE )) { +/* +* some of the pages failed migration. Do get_user_pages +* without migration. +*/ + migrate_allow = false; + +if (!list_empty (&cma_page_list)) +putback_movable_pages (&cma_page_list); + } +/* +* We did migrate all the pages, Try to get the page references again +* migrating any new CMA pages which we failed to isolate earlier. +*/ + drain_allow = true; + goto get_user_again; + } +return ret; +}Signature · 1 line
2.17.2
Quoted 14 lines · Andrew Morton
Quoted 1 line · On Wed, 21 Nov 2018 14:52:56 +0530 "Aneesh Kumar K.V"…
Subject: [PATCH V4 0/3] *
mm/kvm/vfio/ppc64: Migrate compound pages out of CMA regionAsterisk in title is strange?
Quoted 8 lines
ppc64 use CMA area for the allocation of guest page table (hash page table). We won't be able to start guest if we fail to allocate hash page table. We have observed hash table allocation failure because we failed to migrate pages out of CMA region because they were pinned. This happen when we are using VFIO. VFIO on ppc64 pins the entire guest RAM. If the guest RAM pages get allocated out of CMA region, we won't be able to migrate those pages. The pages are also pinned for the lifetime of the guest.
Quoted 8 lines
Quoted 5 lines
Currently we support migration of non-compound pages. With THP and with the addition of
hugetlb migration we can end up allocating compound pages from CMA region. This
patch series add support for migrating compound pages. The first path adds the helperget_user_pages_cma_migrate()which pin the page making sure we migrate them out of CMA region before incrementing the reference count.Very little review activity. Perhaps Andrey and/or Michal can find the time..
I will unlikely find some time before the end of the year. Sorry about that.
Signature · 2 lines
Michal Hocko SUSE Labs