Re: [PATCHv2] mm/gup: speed up check_and_migrate_cma_pages() on huge page
On Mon, Jun 24, 2019 at 01:34:01PM +0800, Pingfan Liu wrote: > On Mon, Jun 24, 2019 at 1:32 PM Pingfan Liu wrote: > > > > On Mon, Jun 24, 2019 at 12:43 PM Ira Weiny wrote: > > > > > > On Mon, Jun 24, 2019 at 12:12:41PM +0800, Pingfan Liu wrote: > > > > Both hugetlb and thp locate on the same migration type of pageblock, > > > > since > > > > they are allocated from a free_list[]. Based on this fact, it is enough > > > > to > > > > check on a single subpage to decide the migration type of the whole huge > > > > page. By this way, it saves (2M/4K - 1) times loop for pmd_huge on x86, > > > > similar on other archs. > > > > > > > > Furthermore, when executing isolate_huge_page(), it avoid taking global > > > > hugetlb_lock many times, and meanless remove/add to the local link list > > > > cma_page_list. > > > > > > > > Signed-off-by: Pingfan Liu > > > > Cc: Andrew Morton > > > > Cc: Ira Weiny > > > > Cc: Mike Rapoport > > > > Cc: "Kirill A. Shutemov" > > > > Cc: Thomas Gleixner > > > > Cc: John Hubbard > > > > Cc: "Aneesh Kumar K.V" > > > > Cc: Christoph Hellwig > > > > Cc: Keith Busch > > > > Cc: Mike Kravetz > > > > Cc: Linux-kernel@vger.kernel.org > > > > --- > > > > mm/gup.c | 19 --- > > > > 1 file changed, 12 insertions(+), 7 deletions(-) > > > > > > > > diff --git a/mm/gup.c b/mm/gup.c > > > > index ddde097..544f5de 100644 > > > > --- a/mm/gup.c > > > > +++ b/mm/gup.c > > > > @@ -1342,19 +1342,22 @@ static long check_and_migrate_cma_pages(struct > > > > task_struct *tsk, > > > > LIST_HEAD(cma_page_list); > > > > > > > > check_again: > > > > - for (i = 0; i < nr_pages; i++) { > > > > + for (i = 0; i < nr_pages;) { > > > > + > > > > + struct page *head = compound_head(pages[i]); > > > > + long step = 1; > > > > + > > > > + if (PageCompound(head)) > > > > + step = compound_order(head) - (pages[i] - head); > > > > > > Sorry if I missed this last time. compound_order() is not correct here. > > For thp, prep_transhuge_page()->prep_compound_page()->set_compound_order(). > > For smaller hugetlb, > > prep_new_huge_page()->prep_compound_page()->set_compound_order(). > > For gigantic page, prep_compound_gigantic_page()->set_compound_order(). > > > > Do I miss anything? > > > Oh, got it. It should be 1< > Thanks, > > Pingfan > > [...]
Re: [PATCHv2] mm/gup: speed up check_and_migrate_cma_pages() on huge page
On Mon, Jun 24, 2019 at 1:32 PM Pingfan Liu wrote: > > On Mon, Jun 24, 2019 at 12:43 PM Ira Weiny wrote: > > > > On Mon, Jun 24, 2019 at 12:12:41PM +0800, Pingfan Liu wrote: > > > Both hugetlb and thp locate on the same migration type of pageblock, since > > > they are allocated from a free_list[]. Based on this fact, it is enough to > > > check on a single subpage to decide the migration type of the whole huge > > > page. By this way, it saves (2M/4K - 1) times loop for pmd_huge on x86, > > > similar on other archs. > > > > > > Furthermore, when executing isolate_huge_page(), it avoid taking global > > > hugetlb_lock many times, and meanless remove/add to the local link list > > > cma_page_list. > > > > > > Signed-off-by: Pingfan Liu > > > Cc: Andrew Morton > > > Cc: Ira Weiny > > > Cc: Mike Rapoport > > > Cc: "Kirill A. Shutemov" > > > Cc: Thomas Gleixner > > > Cc: John Hubbard > > > Cc: "Aneesh Kumar K.V" > > > Cc: Christoph Hellwig > > > Cc: Keith Busch > > > Cc: Mike Kravetz > > > Cc: Linux-kernel@vger.kernel.org > > > --- > > > mm/gup.c | 19 --- > > > 1 file changed, 12 insertions(+), 7 deletions(-) > > > > > > diff --git a/mm/gup.c b/mm/gup.c > > > index ddde097..544f5de 100644 > > > --- a/mm/gup.c > > > +++ b/mm/gup.c > > > @@ -1342,19 +1342,22 @@ static long check_and_migrate_cma_pages(struct > > > task_struct *tsk, > > > LIST_HEAD(cma_page_list); > > > > > > check_again: > > > - for (i = 0; i < nr_pages; i++) { > > > + for (i = 0; i < nr_pages;) { > > > + > > > + struct page *head = compound_head(pages[i]); > > > + long step = 1; > > > + > > > + if (PageCompound(head)) > > > + step = compound_order(head) - (pages[i] - head); > > > > Sorry if I missed this last time. compound_order() is not correct here. > For thp, prep_transhuge_page()->prep_compound_page()->set_compound_order(). > For smaller hugetlb, > prep_new_huge_page()->prep_compound_page()->set_compound_order(). > For gigantic page, prep_compound_gigantic_page()->set_compound_order(). > > Do I miss anything? > Oh, got it. It should be 1< Thanks, > Pingfan > [...]
Re: [PATCHv2] mm/gup: speed up check_and_migrate_cma_pages() on huge page
On Mon, Jun 24, 2019 at 12:43 PM Ira Weiny wrote: > > On Mon, Jun 24, 2019 at 12:12:41PM +0800, Pingfan Liu wrote: > > Both hugetlb and thp locate on the same migration type of pageblock, since > > they are allocated from a free_list[]. Based on this fact, it is enough to > > check on a single subpage to decide the migration type of the whole huge > > page. By this way, it saves (2M/4K - 1) times loop for pmd_huge on x86, > > similar on other archs. > > > > Furthermore, when executing isolate_huge_page(), it avoid taking global > > hugetlb_lock many times, and meanless remove/add to the local link list > > cma_page_list. > > > > Signed-off-by: Pingfan Liu > > Cc: Andrew Morton > > Cc: Ira Weiny > > Cc: Mike Rapoport > > Cc: "Kirill A. Shutemov" > > Cc: Thomas Gleixner > > Cc: John Hubbard > > Cc: "Aneesh Kumar K.V" > > Cc: Christoph Hellwig > > Cc: Keith Busch > > Cc: Mike Kravetz > > Cc: Linux-kernel@vger.kernel.org > > --- > > mm/gup.c | 19 --- > > 1 file changed, 12 insertions(+), 7 deletions(-) > > > > diff --git a/mm/gup.c b/mm/gup.c > > index ddde097..544f5de 100644 > > --- a/mm/gup.c > > +++ b/mm/gup.c > > @@ -1342,19 +1342,22 @@ static long check_and_migrate_cma_pages(struct > > task_struct *tsk, > > LIST_HEAD(cma_page_list); > > > > check_again: > > - for (i = 0; i < nr_pages; i++) { > > + for (i = 0; i < nr_pages;) { > > + > > + struct page *head = compound_head(pages[i]); > > + long step = 1; > > + > > + if (PageCompound(head)) > > + step = compound_order(head) - (pages[i] - head); > > Sorry if I missed this last time. compound_order() is not correct here. For thp, prep_transhuge_page()->prep_compound_page()->set_compound_order(). For smaller hugetlb, prep_new_huge_page()->prep_compound_page()->set_compound_order(). For gigantic page, prep_compound_gigantic_page()->set_compound_order(). Do I miss anything? Thanks, Pingfan [...]
Re: [PATCHv2] mm/gup: speed up check_and_migrate_cma_pages() on huge page
On Mon, Jun 24, 2019 at 12:12:41PM +0800, Pingfan Liu wrote: > Both hugetlb and thp locate on the same migration type of pageblock, since > they are allocated from a free_list[]. Based on this fact, it is enough to > check on a single subpage to decide the migration type of the whole huge > page. By this way, it saves (2M/4K - 1) times loop for pmd_huge on x86, > similar on other archs. > > Furthermore, when executing isolate_huge_page(), it avoid taking global > hugetlb_lock many times, and meanless remove/add to the local link list > cma_page_list. > > Signed-off-by: Pingfan Liu > Cc: Andrew Morton > Cc: Ira Weiny > Cc: Mike Rapoport > Cc: "Kirill A. Shutemov" > Cc: Thomas Gleixner > Cc: John Hubbard > Cc: "Aneesh Kumar K.V" > Cc: Christoph Hellwig > Cc: Keith Busch > Cc: Mike Kravetz > Cc: Linux-kernel@vger.kernel.org > --- > mm/gup.c | 19 --- > 1 file changed, 12 insertions(+), 7 deletions(-) > > diff --git a/mm/gup.c b/mm/gup.c > index ddde097..544f5de 100644 > --- a/mm/gup.c > +++ b/mm/gup.c > @@ -1342,19 +1342,22 @@ static long check_and_migrate_cma_pages(struct > task_struct *tsk, > LIST_HEAD(cma_page_list); > > check_again: > - for (i = 0; i < nr_pages; i++) { > + for (i = 0; i < nr_pages;) { > + > + struct page *head = compound_head(pages[i]); > + long step = 1; > + > + if (PageCompound(head)) > + step = compound_order(head) - (pages[i] - head); Sorry if I missed this last time. compound_order() is not correct here. Ira > /* >* 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])) { > - > - struct page *head = compound_head(pages[i]); > - > - if (PageHuge(head)) { > + if (is_migrate_cma_page(head)) { > + if (PageHuge(head)) > isolate_huge_page(head, _page_list); > - } else { > + else { > if (!PageLRU(head) && drain_allow) { > lru_add_drain_all(); > drain_allow = false; > @@ -1369,6 +1372,8 @@ static long check_and_migrate_cma_pages(struct > task_struct *tsk, > } > } > } > + > + i += step; > } > > if (!list_empty(_page_list)) { > -- > 2.7.5 >