Message ID | alpine.DEB.2.21.2003061421240.7412@chino.kir.corp.google.com (mailing list archive) |
---|---|
State | New, archived |
Headers | show |
Series | [1/2] mm, shmem: add vmstat for hugepage fallback | expand |
On Fri, Mar 6, 2020 at 2:22 PM David Rientjes <rientjes@google.com> wrote: > > The existing thp_fault_fallback indicates when thp attempts to allocate a > hugepage but fails, or if the hugepage cannot be charged to the mem cgroup > hierarchy. > > Extend this to shmem as well. Adds a new thp_file_fallback to complement > thp_file_alloc that gets incremented when a hugepage is attempted to be > allocated but fails, or if it cannot be charged to the mem cgroup > hierarchy. > > Additionally, remove the check for CONFIG_TRANSPARENT_HUGE_PAGECACHE from > shmem_alloc_hugepage() since it is only called with this configuration > option. Looks good to me. Thanks for taking this suggestion. Reviewed-by: Yang Shi <yang.shi@linux.alibaba.com> > > Signed-off-by: David Rientjes <rientjes@google.com> > --- > Documentation/admin-guide/mm/transhuge.rst | 4 ++++ > include/linux/vm_event_item.h | 2 ++ > mm/shmem.c | 10 ++++++---- > mm/vmstat.c | 1 + > 4 files changed, 13 insertions(+), 4 deletions(-) > > diff --git a/Documentation/admin-guide/mm/transhuge.rst b/Documentation/admin-guide/mm/transhuge.rst > --- a/Documentation/admin-guide/mm/transhuge.rst > +++ b/Documentation/admin-guide/mm/transhuge.rst > @@ -319,6 +319,10 @@ thp_file_alloc > is incremented every time a file huge page is successfully > allocated. > > +thp_file_fallback > + is incremented if a file huge page is attempted to be allocated > + but fails and instead falls back to using small pages. > + > thp_file_mapped > is incremented every time a file huge page is mapped into > user address space. > diff --git a/include/linux/vm_event_item.h b/include/linux/vm_event_item.h > --- a/include/linux/vm_event_item.h > +++ b/include/linux/vm_event_item.h > @@ -76,6 +76,7 @@ enum vm_event_item { PGPGIN, PGPGOUT, PSWPIN, PSWPOUT, > THP_COLLAPSE_ALLOC, > THP_COLLAPSE_ALLOC_FAILED, > THP_FILE_ALLOC, > + THP_FILE_FALLBACK, > THP_FILE_MAPPED, > THP_SPLIT_PAGE, > THP_SPLIT_PAGE_FAILED, > @@ -115,6 +116,7 @@ enum vm_event_item { PGPGIN, PGPGOUT, PSWPIN, PSWPOUT, > > #ifndef CONFIG_TRANSPARENT_HUGEPAGE > #define THP_FILE_ALLOC ({ BUILD_BUG(); 0; }) > +#define THP_FILE_FALLBACK ({ BUILD_BUG(); 0; }) > #define THP_FILE_MAPPED ({ BUILD_BUG(); 0; }) > #endif > > diff --git a/mm/shmem.c b/mm/shmem.c > --- a/mm/shmem.c > +++ b/mm/shmem.c > @@ -1472,9 +1472,6 @@ static struct page *shmem_alloc_hugepage(gfp_t gfp, > pgoff_t hindex; > struct page *page; > > - if (!IS_ENABLED(CONFIG_TRANSPARENT_HUGE_PAGECACHE)) > - return NULL; > - > hindex = round_down(index, HPAGE_PMD_NR); > if (xa_find(&mapping->i_pages, &hindex, hindex + HPAGE_PMD_NR - 1, > XA_PRESENT)) > @@ -1486,6 +1483,8 @@ static struct page *shmem_alloc_hugepage(gfp_t gfp, > shmem_pseudo_vma_destroy(&pvma); > if (page) > prep_transhuge_page(page); > + else > + count_vm_event(THP_FILE_FALLBACK); > return page; > } > > @@ -1871,8 +1870,11 @@ static int shmem_getpage_gfp(struct inode *inode, pgoff_t index, > > error = mem_cgroup_try_charge_delay(page, charge_mm, gfp, &memcg, > PageTransHuge(page)); > - if (error) > + if (error) { > + if (PageTransHuge(page)) > + count_vm_event(THP_FILE_FALLBACK); > goto unacct; > + } > error = shmem_add_to_page_cache(page, mapping, hindex, > NULL, gfp & GFP_RECLAIM_MASK); > if (error) { > diff --git a/mm/vmstat.c b/mm/vmstat.c > --- a/mm/vmstat.c > +++ b/mm/vmstat.c > @@ -1257,6 +1257,7 @@ const char * const vmstat_text[] = { > "thp_collapse_alloc", > "thp_collapse_alloc_failed", > "thp_file_alloc", > + "thp_file_fallback", > "thp_file_mapped", > "thp_split_page", > "thp_split_page_failed",
On Fri, Mar 06, 2020 at 02:22:32PM -0800, David Rientjes wrote: > The existing thp_fault_fallback indicates when thp attempts to allocate a > hugepage but fails, or if the hugepage cannot be charged to the mem cgroup > hierarchy. > > Extend this to shmem as well. Adds a new thp_file_fallback to complement > thp_file_alloc that gets incremented when a hugepage is attempted to be > allocated but fails, or if it cannot be charged to the mem cgroup > hierarchy. > > Additionally, remove the check for CONFIG_TRANSPARENT_HUGE_PAGECACHE from > shmem_alloc_hugepage() since it is only called with this configuration > option. > > Signed-off-by: David Rientjes <rientjes@google.com> Acked-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
diff --git a/Documentation/admin-guide/mm/transhuge.rst b/Documentation/admin-guide/mm/transhuge.rst --- a/Documentation/admin-guide/mm/transhuge.rst +++ b/Documentation/admin-guide/mm/transhuge.rst @@ -319,6 +319,10 @@ thp_file_alloc is incremented every time a file huge page is successfully allocated. +thp_file_fallback + is incremented if a file huge page is attempted to be allocated + but fails and instead falls back to using small pages. + thp_file_mapped is incremented every time a file huge page is mapped into user address space. diff --git a/include/linux/vm_event_item.h b/include/linux/vm_event_item.h --- a/include/linux/vm_event_item.h +++ b/include/linux/vm_event_item.h @@ -76,6 +76,7 @@ enum vm_event_item { PGPGIN, PGPGOUT, PSWPIN, PSWPOUT, THP_COLLAPSE_ALLOC, THP_COLLAPSE_ALLOC_FAILED, THP_FILE_ALLOC, + THP_FILE_FALLBACK, THP_FILE_MAPPED, THP_SPLIT_PAGE, THP_SPLIT_PAGE_FAILED, @@ -115,6 +116,7 @@ enum vm_event_item { PGPGIN, PGPGOUT, PSWPIN, PSWPOUT, #ifndef CONFIG_TRANSPARENT_HUGEPAGE #define THP_FILE_ALLOC ({ BUILD_BUG(); 0; }) +#define THP_FILE_FALLBACK ({ BUILD_BUG(); 0; }) #define THP_FILE_MAPPED ({ BUILD_BUG(); 0; }) #endif diff --git a/mm/shmem.c b/mm/shmem.c --- a/mm/shmem.c +++ b/mm/shmem.c @@ -1472,9 +1472,6 @@ static struct page *shmem_alloc_hugepage(gfp_t gfp, pgoff_t hindex; struct page *page; - if (!IS_ENABLED(CONFIG_TRANSPARENT_HUGE_PAGECACHE)) - return NULL; - hindex = round_down(index, HPAGE_PMD_NR); if (xa_find(&mapping->i_pages, &hindex, hindex + HPAGE_PMD_NR - 1, XA_PRESENT)) @@ -1486,6 +1483,8 @@ static struct page *shmem_alloc_hugepage(gfp_t gfp, shmem_pseudo_vma_destroy(&pvma); if (page) prep_transhuge_page(page); + else + count_vm_event(THP_FILE_FALLBACK); return page; } @@ -1871,8 +1870,11 @@ static int shmem_getpage_gfp(struct inode *inode, pgoff_t index, error = mem_cgroup_try_charge_delay(page, charge_mm, gfp, &memcg, PageTransHuge(page)); - if (error) + if (error) { + if (PageTransHuge(page)) + count_vm_event(THP_FILE_FALLBACK); goto unacct; + } error = shmem_add_to_page_cache(page, mapping, hindex, NULL, gfp & GFP_RECLAIM_MASK); if (error) { diff --git a/mm/vmstat.c b/mm/vmstat.c --- a/mm/vmstat.c +++ b/mm/vmstat.c @@ -1257,6 +1257,7 @@ const char * const vmstat_text[] = { "thp_collapse_alloc", "thp_collapse_alloc_failed", "thp_file_alloc", + "thp_file_fallback", "thp_file_mapped", "thp_split_page", "thp_split_page_failed",
The existing thp_fault_fallback indicates when thp attempts to allocate a hugepage but fails, or if the hugepage cannot be charged to the mem cgroup hierarchy. Extend this to shmem as well. Adds a new thp_file_fallback to complement thp_file_alloc that gets incremented when a hugepage is attempted to be allocated but fails, or if it cannot be charged to the mem cgroup hierarchy. Additionally, remove the check for CONFIG_TRANSPARENT_HUGE_PAGECACHE from shmem_alloc_hugepage() since it is only called with this configuration option. Signed-off-by: David Rientjes <rientjes@google.com> --- Documentation/admin-guide/mm/transhuge.rst | 4 ++++ include/linux/vm_event_item.h | 2 ++ mm/shmem.c | 10 ++++++---- mm/vmstat.c | 1 + 4 files changed, 13 insertions(+), 4 deletions(-)