From 3167760f83899ccda312b9ad9306ec9e5dda06d4 Mon Sep 17 00:00:00 2001 From: Dan Magenheimer Date: Wed, 21 Sep 2011 11:56:28 -0400 Subject: mm: cleancache: s/flush/invalidate/ Per akpm suggestions alter the use of the term flush to be invalidate. The next patch will do this across all MM. This change is completely cosmetic. [v9: akpm@linux-foundation.org: change "flush" to "invalidate", part 3] Signed-off-by: Dan Magenheimer Cc: Kamezawa Hiroyuki Cc: Jan Beulich Reviewed-by: Seth Jennings Cc: Jeremy Fitzhardinge Cc: Hugh Dickins Cc: Johannes Weiner Cc: Nitin Gupta Cc: Matthew Wilcox Cc: Chris Mason Cc: Rik Riel Cc: Andrew Morton [v10: Fixed fs: move code out of buffer.c conflict change] Signed-off-by: Konrad Rzeszutek Wilk --- mm/filemap.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'mm/filemap.c') diff --git a/mm/filemap.c b/mm/filemap.c index a0701e6eec10..0aa3faa48219 100644 --- a/mm/filemap.c +++ b/mm/filemap.c @@ -123,7 +123,7 @@ void __delete_from_page_cache(struct page *page) if (PageUptodate(page) && PageMappedToDisk(page)) cleancache_put_page(page); else - cleancache_flush_page(mapping, page); + cleancache_invalidate_page(mapping, page); radix_tree_delete(&mapping->page_tree, page->index); page->mapping = NULL; -- cgit v1.2.3 From 3deaa7190a8da38453c4fabd9dec7f66d17fff67 Mon Sep 17 00:00:00 2001 From: Shaohua Li Date: Fri, 3 Feb 2012 15:37:17 -0800 Subject: readahead: fix pipeline break caused by block plug Herbert Poetzl reported a performance regression since 2.6.39. The test is a simple dd read, but with big block size. The reason is: T1: ra (A, A+128k), (A+128k, A+256k) T2: lock_page for page A, submit the 256k T3: hit page A+128K, ra (A+256k, A+384). the range isn't submitted because of plug and there isn't any lock_page till we hit page A+256k because all pages from A to A+256k is in memory T4: hit page A+256k, ra (A+384, A+ 512). Because of plug, the range isn't submitted again. T5: lock_page A+256k, so (A+256k, A+512k) will be submitted. The task is waitting for (A+256k, A+512k) finish. There is no request to disk in T3 and T4, so readahead pipeline breaks. We really don't need block plug for generic_file_aio_read() for buffered I/O. The readahead already has plug and has fine grained control when I/O should be submitted. Deleting plug for buffered I/O fixes the regression. One side effect is plug makes the request size 256k, the size is 128k without it. This is because default ra size is 128k and not a reason we need plug here. Vivek said: : We submit some readahead IO to device request queue but because of nested : plug, queue never gets unplugged. When read logic reaches a page which is : not in page cache, it waits for page to be read from the disk : (lock_page_killable()) and that time we flush the plug list. : : So effectively read ahead logic is kind of broken in parts because of : nested plugging. Removing top level plug (generic_file_aio_read()) for : buffered reads, will allow unplugging queue earlier for readahead. Signed-off-by: Shaohua Li Signed-off-by: Wu Fengguang Reported-by: Herbert Poetzl Tested-by: Eric Dumazet Cc: Christoph Hellwig Cc: Jens Axboe Cc: Vivek Goyal Cc: Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- mm/filemap.c | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-) (limited to 'mm/filemap.c') diff --git a/mm/filemap.c b/mm/filemap.c index 97f49ed35bd2..b66275757c28 100644 --- a/mm/filemap.c +++ b/mm/filemap.c @@ -1400,15 +1400,12 @@ generic_file_aio_read(struct kiocb *iocb, const struct iovec *iov, unsigned long seg = 0; size_t count; loff_t *ppos = &iocb->ki_pos; - struct blk_plug plug; count = 0; retval = generic_segment_checks(iov, &nr_segs, &count, VERIFY_WRITE); if (retval) return retval; - blk_start_plug(&plug); - /* coalesce the iovecs and go direct-to-BIO for O_DIRECT */ if (filp->f_flags & O_DIRECT) { loff_t size; @@ -1424,8 +1421,12 @@ generic_file_aio_read(struct kiocb *iocb, const struct iovec *iov, retval = filemap_write_and_wait_range(mapping, pos, pos + iov_length(iov, nr_segs) - 1); if (!retval) { + struct blk_plug plug; + + blk_start_plug(&plug); retval = mapping->a_ops->direct_IO(READ, iocb, iov, pos, nr_segs); + blk_finish_plug(&plug); } if (retval > 0) { *ppos = pos + retval; @@ -1481,7 +1482,6 @@ generic_file_aio_read(struct kiocb *iocb, const struct iovec *iov, break; } out: - blk_finish_plug(&plug); return retval; } EXPORT_SYMBOL(generic_file_aio_read); -- cgit v1.2.3 From 9b04c5fec43c0da610a2c37f70c5b013101a6ad7 Mon Sep 17 00:00:00 2001 From: Cong Wang Date: Fri, 25 Nov 2011 23:14:39 +0800 Subject: mm: remove the second argument of k[un]map_atomic() Signed-off-by: Cong Wang --- mm/bounce.c | 4 ++-- mm/filemap.c | 8 ++++---- mm/ksm.c | 12 ++++++------ mm/memory.c | 4 ++-- mm/shmem.c | 4 ++-- mm/swapfile.c | 30 +++++++++++++++--------------- mm/vmalloc.c | 8 ++++---- 7 files changed, 35 insertions(+), 35 deletions(-) (limited to 'mm/filemap.c') diff --git a/mm/bounce.c b/mm/bounce.c index 4e9ae722af83..d1be02ca1889 100644 --- a/mm/bounce.c +++ b/mm/bounce.c @@ -50,9 +50,9 @@ static void bounce_copy_vec(struct bio_vec *to, unsigned char *vfrom) unsigned char *vto; local_irq_save(flags); - vto = kmap_atomic(to->bv_page, KM_BOUNCE_READ); + vto = kmap_atomic(to->bv_page); memcpy(vto + to->bv_offset, vfrom, to->bv_len); - kunmap_atomic(vto, KM_BOUNCE_READ); + kunmap_atomic(vto); local_irq_restore(flags); } diff --git a/mm/filemap.c b/mm/filemap.c index b66275757c28..2f8165075a5a 100644 --- a/mm/filemap.c +++ b/mm/filemap.c @@ -1318,10 +1318,10 @@ int file_read_actor(read_descriptor_t *desc, struct page *page, * taking the kmap. */ if (!fault_in_pages_writeable(desc->arg.buf, size)) { - kaddr = kmap_atomic(page, KM_USER0); + kaddr = kmap_atomic(page); left = __copy_to_user_inatomic(desc->arg.buf, kaddr + offset, size); - kunmap_atomic(kaddr, KM_USER0); + kunmap_atomic(kaddr); if (left == 0) goto success; } @@ -2045,7 +2045,7 @@ size_t iov_iter_copy_from_user_atomic(struct page *page, size_t copied; BUG_ON(!in_atomic()); - kaddr = kmap_atomic(page, KM_USER0); + kaddr = kmap_atomic(page); if (likely(i->nr_segs == 1)) { int left; char __user *buf = i->iov->iov_base + i->iov_offset; @@ -2055,7 +2055,7 @@ size_t iov_iter_copy_from_user_atomic(struct page *page, copied = __iovec_copy_from_user_inatomic(kaddr + offset, i->iov, i->iov_offset, bytes); } - kunmap_atomic(kaddr, KM_USER0); + kunmap_atomic(kaddr); return copied; } diff --git a/mm/ksm.c b/mm/ksm.c index 310544a379ae..a6d3fb7e6c10 100644 --- a/mm/ksm.c +++ b/mm/ksm.c @@ -672,9 +672,9 @@ error: static u32 calc_checksum(struct page *page) { u32 checksum; - void *addr = kmap_atomic(page, KM_USER0); + void *addr = kmap_atomic(page); checksum = jhash2(addr, PAGE_SIZE / 4, 17); - kunmap_atomic(addr, KM_USER0); + kunmap_atomic(addr); return checksum; } @@ -683,11 +683,11 @@ static int memcmp_pages(struct page *page1, struct page *page2) char *addr1, *addr2; int ret; - addr1 = kmap_atomic(page1, KM_USER0); - addr2 = kmap_atomic(page2, KM_USER1); + addr1 = kmap_atomic(page1); + addr2 = kmap_atomic(page2); ret = memcmp(addr1, addr2, PAGE_SIZE); - kunmap_atomic(addr2, KM_USER1); - kunmap_atomic(addr1, KM_USER0); + kunmap_atomic(addr2); + kunmap_atomic(addr1); return ret; } diff --git a/mm/memory.c b/mm/memory.c index fa2f04e0337c..347e5fad1cfa 100644 --- a/mm/memory.c +++ b/mm/memory.c @@ -2447,7 +2447,7 @@ static inline void cow_user_page(struct page *dst, struct page *src, unsigned lo * fails, we just zero-fill it. Live with it. */ if (unlikely(!src)) { - void *kaddr = kmap_atomic(dst, KM_USER0); + void *kaddr = kmap_atomic(dst); void __user *uaddr = (void __user *)(va & PAGE_MASK); /* @@ -2458,7 +2458,7 @@ static inline void cow_user_page(struct page *dst, struct page *src, unsigned lo */ if (__copy_from_user_inatomic(kaddr, uaddr, PAGE_SIZE)) clear_page(kaddr); - kunmap_atomic(kaddr, KM_USER0); + kunmap_atomic(kaddr); flush_dcache_page(dst); } else copy_user_highpage(dst, src, va, vma); diff --git a/mm/shmem.c b/mm/shmem.c index 269d049294ab..b7e195571862 100644 --- a/mm/shmem.c +++ b/mm/shmem.c @@ -1656,9 +1656,9 @@ static int shmem_symlink(struct inode *dir, struct dentry *dentry, const char *s } inode->i_mapping->a_ops = &shmem_aops; inode->i_op = &shmem_symlink_inode_operations; - kaddr = kmap_atomic(page, KM_USER0); + kaddr = kmap_atomic(page); memcpy(kaddr, symname, len); - kunmap_atomic(kaddr, KM_USER0); + kunmap_atomic(kaddr); set_page_dirty(page); unlock_page(page); page_cache_release(page); diff --git a/mm/swapfile.c b/mm/swapfile.c index d999f090dfda..00a962caab1a 100644 --- a/mm/swapfile.c +++ b/mm/swapfile.c @@ -2427,9 +2427,9 @@ int add_swap_count_continuation(swp_entry_t entry, gfp_t gfp_mask) if (!(count & COUNT_CONTINUED)) goto out; - map = kmap_atomic(list_page, KM_USER0) + offset; + map = kmap_atomic(list_page) + offset; count = *map; - kunmap_atomic(map, KM_USER0); + kunmap_atomic(map); /* * If this continuation count now has some space in it, @@ -2472,7 +2472,7 @@ static bool swap_count_continued(struct swap_info_struct *si, offset &= ~PAGE_MASK; page = list_entry(head->lru.next, struct page, lru); - map = kmap_atomic(page, KM_USER0) + offset; + map = kmap_atomic(page) + offset; if (count == SWAP_MAP_MAX) /* initial increment from swap_map */ goto init_map; /* jump over SWAP_CONT_MAX checks */ @@ -2482,26 +2482,26 @@ static bool swap_count_continued(struct swap_info_struct *si, * Think of how you add 1 to 999 */ while (*map == (SWAP_CONT_MAX | COUNT_CONTINUED)) { - kunmap_atomic(map, KM_USER0); + kunmap_atomic(map); page = list_entry(page->lru.next, struct page, lru); BUG_ON(page == head); - map = kmap_atomic(page, KM_USER0) + offset; + map = kmap_atomic(page) + offset; } if (*map == SWAP_CONT_MAX) { - kunmap_atomic(map, KM_USER0); + kunmap_atomic(map); page = list_entry(page->lru.next, struct page, lru); if (page == head) return false; /* add count continuation */ - map = kmap_atomic(page, KM_USER0) + offset; + map = kmap_atomic(page) + offset; init_map: *map = 0; /* we didn't zero the page */ } *map += 1; - kunmap_atomic(map, KM_USER0); + kunmap_atomic(map); page = list_entry(page->lru.prev, struct page, lru); while (page != head) { - map = kmap_atomic(page, KM_USER0) + offset; + map = kmap_atomic(page) + offset; *map = COUNT_CONTINUED; - kunmap_atomic(map, KM_USER0); + kunmap_atomic(map); page = list_entry(page->lru.prev, struct page, lru); } return true; /* incremented */ @@ -2512,22 +2512,22 @@ init_map: *map = 0; /* we didn't zero the page */ */ BUG_ON(count != COUNT_CONTINUED); while (*map == COUNT_CONTINUED) { - kunmap_atomic(map, KM_USER0); + kunmap_atomic(map); page = list_entry(page->lru.next, struct page, lru); BUG_ON(page == head); - map = kmap_atomic(page, KM_USER0) + offset; + map = kmap_atomic(page) + offset; } BUG_ON(*map == 0); *map -= 1; if (*map == 0) count = 0; - kunmap_atomic(map, KM_USER0); + kunmap_atomic(map); page = list_entry(page->lru.prev, struct page, lru); while (page != head) { - map = kmap_atomic(page, KM_USER0) + offset; + map = kmap_atomic(page) + offset; *map = SWAP_CONT_MAX | count; count = COUNT_CONTINUED; - kunmap_atomic(map, KM_USER0); + kunmap_atomic(map); page = list_entry(page->lru.prev, struct page, lru); } return count == COUNT_CONTINUED; diff --git a/mm/vmalloc.c b/mm/vmalloc.c index 86ce9a526c17..94dff883b449 100644 --- a/mm/vmalloc.c +++ b/mm/vmalloc.c @@ -1906,9 +1906,9 @@ static int aligned_vread(char *buf, char *addr, unsigned long count) * we can expect USER0 is not used (see vread/vwrite's * function description) */ - void *map = kmap_atomic(p, KM_USER0); + void *map = kmap_atomic(p); memcpy(buf, map + offset, length); - kunmap_atomic(map, KM_USER0); + kunmap_atomic(map); } else memset(buf, 0, length); @@ -1945,9 +1945,9 @@ static int aligned_vwrite(char *buf, char *addr, unsigned long count) * we can expect USER0 is not used (see vread/vwrite's * function description) */ - void *map = kmap_atomic(p, KM_USER0); + void *map = kmap_atomic(p); memcpy(map + offset, buf, length); - kunmap_atomic(map, KM_USER0); + kunmap_atomic(map); } addr += length; buf += length; -- cgit v1.2.3 From 1010bb1b80edb0713415dfe1f97114d320f58c4f Mon Sep 17 00:00:00 2001 From: Fengguang Wu Date: Wed, 21 Mar 2012 16:34:08 -0700 Subject: mm: don't set __GFP_WRITE on ramfs/sysfs writes There is not much point in skipping zones during allocation based on the dirty usage which they'll never contribute to. And we'd like to avoid page reclaim waits when writing to ramfs/sysfs etc. Signed-off-by: Fengguang Wu Acked-by: Johannes Weiner Cc: Jan Kara Cc: Greg Thelen Cc: Ying Han Cc: KAMEZAWA Hiroyuki Cc: Rik van Riel Cc: Mel Gorman Acked-by: Minchan Kim Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- mm/filemap.c | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) (limited to 'mm/filemap.c') diff --git a/mm/filemap.c b/mm/filemap.c index 2f8165075a5a..e8cf8aefd6f6 100644 --- a/mm/filemap.c +++ b/mm/filemap.c @@ -2341,7 +2341,9 @@ struct page *grab_cache_page_write_begin(struct address_space *mapping, struct page *page; gfp_t gfp_notmask = 0; - gfp_mask = mapping_gfp_mask(mapping) | __GFP_WRITE; + gfp_mask = mapping_gfp_mask(mapping); + if (mapping_cap_account_dirty(mapping)) + gfp_mask |= __GFP_WRITE; if (flags & AOP_FLAG_NOFS) gfp_notmask = __GFP_FS; repeat: -- cgit v1.2.3 From 9a3c531df9462df6cc2b060f749651723ffc180c Mon Sep 17 00:00:00 2001 From: Andi Kleen Date: Wed, 21 Mar 2012 16:34:09 -0700 Subject: mm: update stale lock ordering comment for memory-failure.c When i_mmap_lock changed to a mutex the locking order in memory failure was changed to take the sleeping lock first. But the big fat mm lock ordering comment (BFMLO) wasn't updated. Do this here. Pointed out by Andrew. Signed-off-by: Andi Kleen Cc: Peter Zijlstra Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- mm/filemap.c | 5 ++--- 1 file changed, 2 insertions(+), 3 deletions(-) (limited to 'mm/filemap.c') diff --git a/mm/filemap.c b/mm/filemap.c index e8cf8aefd6f6..f3230604006c 100644 --- a/mm/filemap.c +++ b/mm/filemap.c @@ -101,9 +101,8 @@ * ->inode->i_lock (zap_pte_range->set_page_dirty) * ->private_lock (zap_pte_range->__set_page_dirty_buffers) * - * (code doesn't rely on that order, so you could switch it around) - * ->tasklist_lock (memory_failure, collect_procs_ao) - * ->i_mmap_mutex + * ->i_mmap_mutex + * ->tasklist_lock (memory_failure, collect_procs_ao) */ /* -- cgit v1.2.3 From cc9a6c8776615f9c194ccf0b63a0aa5628235545 Mon Sep 17 00:00:00 2001 From: Mel Gorman Date: Wed, 21 Mar 2012 16:34:11 -0700 Subject: cpuset: mm: reduce large amounts of memory barrier related damage v3 Commit c0ff7453bb5c ("cpuset,mm: fix no node to alloc memory when changing cpuset's mems") wins a super prize for the largest number of memory barriers entered into fast paths for one commit. [get|put]_mems_allowed is incredibly heavy with pairs of full memory barriers inserted into a number of hot paths. This was detected while investigating at large page allocator slowdown introduced some time after 2.6.32. The largest portion of this overhead was shown by oprofile to be at an mfence introduced by this commit into the page allocator hot path. For extra style points, the commit introduced the use of yield() in an implementation of what looks like a spinning mutex. This patch replaces the full memory barriers on both read and write sides with a sequence counter with just read barriers on the fast path side. This is much cheaper on some architectures, including x86. The main bulk of the patch is the retry logic if the nodemask changes in a manner that can cause a false failure. While updating the nodemask, a check is made to see if a false failure is a risk. If it is, the sequence number gets bumped and parallel allocators will briefly stall while the nodemask update takes place. In a page fault test microbenchmark, oprofile samples from __alloc_pages_nodemask went from 4.53% of all samples to 1.15%. The actual results were 3.3.0-rc3 3.3.0-rc3 rc3-vanilla nobarrier-v2r1 Clients 1 UserTime 0.07 ( 0.00%) 0.08 (-14.19%) Clients 2 UserTime 0.07 ( 0.00%) 0.07 ( 2.72%) Clients 4 UserTime 0.08 ( 0.00%) 0.07 ( 3.29%) Clients 1 SysTime 0.70 ( 0.00%) 0.65 ( 6.65%) Clients 2 SysTime 0.85 ( 0.00%) 0.82 ( 3.65%) Clients 4 SysTime 1.41 ( 0.00%) 1.41 ( 0.32%) Clients 1 WallTime 0.77 ( 0.00%) 0.74 ( 4.19%) Clients 2 WallTime 0.47 ( 0.00%) 0.45 ( 3.73%) Clients 4 WallTime 0.38 ( 0.00%) 0.37 ( 1.58%) Clients 1 Flt/sec/cpu 497620.28 ( 0.00%) 520294.53 ( 4.56%) Clients 2 Flt/sec/cpu 414639.05 ( 0.00%) 429882.01 ( 3.68%) Clients 4 Flt/sec/cpu 257959.16 ( 0.00%) 258761.48 ( 0.31%) Clients 1 Flt/sec 495161.39 ( 0.00%) 517292.87 ( 4.47%) Clients 2 Flt/sec 820325.95 ( 0.00%) 850289.77 ( 3.65%) Clients 4 Flt/sec 1020068.93 ( 0.00%) 1022674.06 ( 0.26%) MMTests Statistics: duration Sys Time Running Test (seconds) 135.68 132.17 User+Sys Time Running Test (seconds) 164.2 160.13 Total Elapsed Time (seconds) 123.46 120.87 The overall improvement is small but the System CPU time is much improved and roughly in correlation to what oprofile reported (these performance figures are without profiling so skew is expected). The actual number of page faults is noticeably improved. For benchmarks like kernel builds, the overall benefit is marginal but the system CPU time is slightly reduced. To test the actual bug the commit fixed I opened two terminals. The first ran within a cpuset and continually ran a small program that faulted 100M of anonymous data. In a second window, the nodemask of the cpuset was continually randomised in a loop. Without the commit, the program would fail every so often (usually within 10 seconds) and obviously with the commit everything worked fine. With this patch applied, it also worked fine so the fix should be functionally equivalent. Signed-off-by: Mel Gorman Cc: Miao Xie Cc: David Rientjes Cc: Peter Zijlstra Cc: Christoph Lameter Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- include/linux/cpuset.h | 47 ++++++++++++++++++++--------------------------- include/linux/init_task.h | 8 ++++++++ include/linux/sched.h | 2 +- kernel/cpuset.c | 43 ++++++++----------------------------------- kernel/fork.c | 1 + mm/filemap.c | 11 +++++++---- mm/hugetlb.c | 15 +++++++++++---- mm/mempolicy.c | 28 +++++++++++++++++++++------- mm/page_alloc.c | 33 +++++++++++++++++++++++---------- mm/slab.c | 13 ++++++++----- mm/slub.c | 40 +++++++++++++++++++++++++--------------- mm/vmscan.c | 2 -- 12 files changed, 133 insertions(+), 110 deletions(-) (limited to 'mm/filemap.c') diff --git a/include/linux/cpuset.h b/include/linux/cpuset.h index e9eaec522655..7a7e5fd2a277 100644 --- a/include/linux/cpuset.h +++ b/include/linux/cpuset.h @@ -89,42 +89,33 @@ extern void rebuild_sched_domains(void); extern void cpuset_print_task_mems_allowed(struct task_struct *p); /* - * reading current mems_allowed and mempolicy in the fastpath must protected - * by get_mems_allowed() + * get_mems_allowed is required when making decisions involving mems_allowed + * such as during page allocation. mems_allowed can be updated in parallel + * and depending on the new value an operation can fail potentially causing + * process failure. A retry loop with get_mems_allowed and put_mems_allowed + * prevents these artificial failures. */ -static inline void get_mems_allowed(void) +static inline unsigned int get_mems_allowed(void) { - current->mems_allowed_change_disable++; - - /* - * ensure that reading mems_allowed and mempolicy happens after the - * update of ->mems_allowed_change_disable. - * - * the write-side task finds ->mems_allowed_change_disable is not 0, - * and knows the read-side task is reading mems_allowed or mempolicy, - * so it will clear old bits lazily. - */ - smp_mb(); + return read_seqcount_begin(¤t->mems_allowed_seq); } -static inline void put_mems_allowed(void) +/* + * If this returns false, the operation that took place after get_mems_allowed + * may have failed. It is up to the caller to retry the operation if + * appropriate. + */ +static inline bool put_mems_allowed(unsigned int seq) { - /* - * ensure that reading mems_allowed and mempolicy before reducing - * mems_allowed_change_disable. - * - * the write-side task will know that the read-side task is still - * reading mems_allowed or mempolicy, don't clears old bits in the - * nodemask. - */ - smp_mb(); - --ACCESS_ONCE(current->mems_allowed_change_disable); + return !read_seqcount_retry(¤t->mems_allowed_seq, seq); } static inline void set_mems_allowed(nodemask_t nodemask) { task_lock(current); + write_seqcount_begin(¤t->mems_allowed_seq); current->mems_allowed = nodemask; + write_seqcount_end(¤t->mems_allowed_seq); task_unlock(current); } @@ -234,12 +225,14 @@ static inline void set_mems_allowed(nodemask_t nodemask) { } -static inline void get_mems_allowed(void) +static inline unsigned int get_mems_allowed(void) { + return 0; } -static inline void put_mems_allowed(void) +static inline bool put_mems_allowed(unsigned int seq) { + return true; } #endif /* !CONFIG_CPUSETS */ diff --git a/include/linux/init_task.h b/include/linux/init_task.h index f994d51f70f2..e4baff5f7ff4 100644 --- a/include/linux/init_task.h +++ b/include/linux/init_task.h @@ -29,6 +29,13 @@ extern struct fs_struct init_fs; #define INIT_GROUP_RWSEM(sig) #endif +#ifdef CONFIG_CPUSETS +#define INIT_CPUSET_SEQ \ + .mems_allowed_seq = SEQCNT_ZERO, +#else +#define INIT_CPUSET_SEQ +#endif + #define INIT_SIGNALS(sig) { \ .nr_threads = 1, \ .wait_chldexit = __WAIT_QUEUE_HEAD_INITIALIZER(sig.wait_chldexit),\ @@ -192,6 +199,7 @@ extern struct cred init_cred; INIT_FTRACE_GRAPH \ INIT_TRACE_RECURSION \ INIT_TASK_RCU_PREEMPT(tsk) \ + INIT_CPUSET_SEQ \ } diff --git a/include/linux/sched.h b/include/linux/sched.h index e074e1e54f85..0c147a4260a5 100644 --- a/include/linux/sched.h +++ b/include/linux/sched.h @@ -1514,7 +1514,7 @@ struct task_struct { #endif #ifdef CONFIG_CPUSETS nodemask_t mems_allowed; /* Protected by alloc_lock */ - int mems_allowed_change_disable; + seqcount_t mems_allowed_seq; /* Seqence no to catch updates */ int cpuset_mem_spread_rotor; int cpuset_slab_spread_rotor; #endif diff --git a/kernel/cpuset.c b/kernel/cpuset.c index 5d575836dba6..1010cc61931f 100644 --- a/kernel/cpuset.c +++ b/kernel/cpuset.c @@ -964,7 +964,6 @@ static void cpuset_change_task_nodemask(struct task_struct *tsk, { bool need_loop; -repeat: /* * Allow tasks that have access to memory reserves because they have * been OOM killed to get memory anywhere. @@ -983,45 +982,19 @@ repeat: */ need_loop = task_has_mempolicy(tsk) || !nodes_intersects(*newmems, tsk->mems_allowed); - nodes_or(tsk->mems_allowed, tsk->mems_allowed, *newmems); - mpol_rebind_task(tsk, newmems, MPOL_REBIND_STEP1); - /* - * ensure checking ->mems_allowed_change_disable after setting all new - * allowed nodes. - * - * the read-side task can see an nodemask with new allowed nodes and - * old allowed nodes. and if it allocates page when cpuset clears newly - * disallowed ones continuous, it can see the new allowed bits. - * - * And if setting all new allowed nodes is after the checking, setting - * all new allowed nodes and clearing newly disallowed ones will be done - * continuous, and the read-side task may find no node to alloc page. - */ - smp_mb(); + if (need_loop) + write_seqcount_begin(&tsk->mems_allowed_seq); - /* - * Allocation of memory is very fast, we needn't sleep when waiting - * for the read-side. - */ - while (need_loop && ACCESS_ONCE(tsk->mems_allowed_change_disable)) { - task_unlock(tsk); - if (!task_curr(tsk)) - yield(); - goto repeat; - } - - /* - * ensure checking ->mems_allowed_change_disable before clearing all new - * disallowed nodes. - * - * if clearing newly disallowed bits before the checking, the read-side - * task may find no node to alloc page. - */ - smp_mb(); + nodes_or(tsk->mems_allowed, tsk->mems_allowed, *newmems); + mpol_rebind_task(tsk, newmems, MPOL_REBIND_STEP1); mpol_rebind_task(tsk, newmems, MPOL_REBIND_STEP2); tsk->mems_allowed = *newmems; + + if (need_loop) + write_seqcount_end(&tsk->mems_allowed_seq); + task_unlock(tsk); } diff --git a/kernel/fork.c b/kernel/fork.c index a9e99f3c18e0..9cc227d54102 100644 --- a/kernel/fork.c +++ b/kernel/fork.c @@ -1237,6 +1237,7 @@ static struct task_struct *copy_process(unsigned long clone_flags, #ifdef CONFIG_CPUSETS p->cpuset_mem_spread_rotor = NUMA_NO_NODE; p->cpuset_slab_spread_rotor = NUMA_NO_NODE; + seqcount_init(&p->mems_allowed_seq); #endif #ifdef CONFIG_TRACE_IRQFLAGS p->irq_events = 0; diff --git a/mm/filemap.c b/mm/filemap.c index f3230604006c..843042045dc9 100644 --- a/mm/filemap.c +++ b/mm/filemap.c @@ -499,10 +499,13 @@ struct page *__page_cache_alloc(gfp_t gfp) struct page *page; if (cpuset_do_page_mem_spread()) { - get_mems_allowed(); - n = cpuset_mem_spread_node(); - page = alloc_pages_exact_node(n, gfp, 0); - put_mems_allowed(); + unsigned int cpuset_mems_cookie; + do { + cpuset_mems_cookie = get_mems_allowed(); + n = cpuset_mem_spread_node(); + page = alloc_pages_exact_node(n, gfp, 0); + } while (!put_mems_allowed(cpuset_mems_cookie) && !page); + return page; } return alloc_pages(gfp, 0); diff --git a/mm/hugetlb.c b/mm/hugetlb.c index 62f9fada4d6d..b1c314877334 100644 --- a/mm/hugetlb.c +++ b/mm/hugetlb.c @@ -454,14 +454,16 @@ static struct page *dequeue_huge_page_vma(struct hstate *h, struct vm_area_struct *vma, unsigned long address, int avoid_reserve) { - struct page *page = NULL; + struct page *page; struct mempolicy *mpol; nodemask_t *nodemask; struct zonelist *zonelist; struct zone *zone; struct zoneref *z; + unsigned int cpuset_mems_cookie; - get_mems_allowed(); +retry_cpuset: + cpuset_mems_cookie = get_mems_allowed(); zonelist = huge_zonelist(vma, address, htlb_alloc_mask, &mpol, &nodemask); /* @@ -488,10 +490,15 @@ static struct page *dequeue_huge_page_vma(struct hstate *h, } } } -err: + mpol_cond_put(mpol); - put_mems_allowed(); + if (unlikely(!put_mems_allowed(cpuset_mems_cookie) && !page)) + goto retry_cpuset; return page; + +err: + mpol_cond_put(mpol); + return NULL; } static void update_and_free_page(struct hstate *h, struct page *page) diff --git a/mm/mempolicy.c b/mm/mempolicy.c index 71e1a523e209..cfb6c8678754 100644 --- a/mm/mempolicy.c +++ b/mm/mempolicy.c @@ -1850,18 +1850,24 @@ struct page * alloc_pages_vma(gfp_t gfp, int order, struct vm_area_struct *vma, unsigned long addr, int node) { - struct mempolicy *pol = get_vma_policy(current, vma, addr); + struct mempolicy *pol; struct zonelist *zl; struct page *page; + unsigned int cpuset_mems_cookie; + +retry_cpuset: + pol = get_vma_policy(current, vma, addr); + cpuset_mems_cookie = get_mems_allowed(); - get_mems_allowed(); if (unlikely(pol->mode == MPOL_INTERLEAVE)) { unsigned nid; nid = interleave_nid(pol, vma, addr, PAGE_SHIFT + order); mpol_cond_put(pol); page = alloc_page_interleave(gfp, order, nid); - put_mems_allowed(); + if (unlikely(!put_mems_allowed(cpuset_mems_cookie) && !page)) + goto retry_cpuset; + return page; } zl = policy_zonelist(gfp, pol, node); @@ -1872,7 +1878,8 @@ alloc_pages_vma(gfp_t gfp, int order, struct vm_area_struct *vma, struct page *page = __alloc_pages_nodemask(gfp, order, zl, policy_nodemask(gfp, pol)); __mpol_put(pol); - put_mems_allowed(); + if (unlikely(!put_mems_allowed(cpuset_mems_cookie) && !page)) + goto retry_cpuset; return page; } /* @@ -1880,7 +1887,8 @@ alloc_pages_vma(gfp_t gfp, int order, struct vm_area_struct *vma, */ page = __alloc_pages_nodemask(gfp, order, zl, policy_nodemask(gfp, pol)); - put_mems_allowed(); + if (unlikely(!put_mems_allowed(cpuset_mems_cookie) && !page)) + goto retry_cpuset; return page; } @@ -1907,11 +1915,14 @@ struct page *alloc_pages_current(gfp_t gfp, unsigned order) { struct mempolicy *pol = current->mempolicy; struct page *page; + unsigned int cpuset_mems_cookie; if (!pol || in_interrupt() || (gfp & __GFP_THISNODE)) pol = &default_policy; - get_mems_allowed(); +retry_cpuset: + cpuset_mems_cookie = get_mems_allowed(); + /* * No reference counting needed for current->mempolicy * nor system default_policy @@ -1922,7 +1933,10 @@ struct page *alloc_pages_current(gfp_t gfp, unsigned order) page = __alloc_pages_nodemask(gfp, order, policy_zonelist(gfp, pol, numa_node_id()), policy_nodemask(gfp, pol)); - put_mems_allowed(); + + if (unlikely(!put_mems_allowed(cpuset_mems_cookie) && !page)) + goto retry_cpuset; + return page; } EXPORT_SYMBOL(alloc_pages_current); diff --git a/mm/page_alloc.c b/mm/page_alloc.c index 673596ad9c80..40de6854b980 100644 --- a/mm/page_alloc.c +++ b/mm/page_alloc.c @@ -2380,8 +2380,9 @@ __alloc_pages_nodemask(gfp_t gfp_mask, unsigned int order, { enum zone_type high_zoneidx = gfp_zone(gfp_mask); struct zone *preferred_zone; - struct page *page; + struct page *page = NULL; int migratetype = allocflags_to_migratetype(gfp_mask); + unsigned int cpuset_mems_cookie; gfp_mask &= gfp_allowed_mask; @@ -2400,15 +2401,15 @@ __alloc_pages_nodemask(gfp_t gfp_mask, unsigned int order, if (unlikely(!zonelist->_zonerefs->zone)) return NULL; - get_mems_allowed(); +retry_cpuset: + cpuset_mems_cookie = get_mems_allowed(); + /* The preferred zone is used for statistics later */ first_zones_zonelist(zonelist, high_zoneidx, nodemask ? : &cpuset_current_mems_allowed, &preferred_zone); - if (!preferred_zone) { - put_mems_allowed(); - return NULL; - } + if (!preferred_zone) + goto out; /* First allocation attempt */ page = get_page_from_freelist(gfp_mask|__GFP_HARDWALL, nodemask, order, @@ -2418,9 +2419,19 @@ __alloc_pages_nodemask(gfp_t gfp_mask, unsigned int order, page = __alloc_pages_slowpath(gfp_mask, order, zonelist, high_zoneidx, nodemask, preferred_zone, migratetype); - put_mems_allowed(); trace_mm_page_alloc(page, order, gfp_mask, migratetype); + +out: + /* + * When updating a task's mems_allowed, it is possible to race with + * parallel threads in such a way that an allocation can fail while + * the mask is being updated. If a page allocation is about to fail, + * check if the cpuset changed during allocation and if so, retry. + */ + if (unlikely(!put_mems_allowed(cpuset_mems_cookie) && !page)) + goto retry_cpuset; + return page; } EXPORT_SYMBOL(__alloc_pages_nodemask); @@ -2634,13 +2645,15 @@ void si_meminfo_node(struct sysinfo *val, int nid) bool skip_free_areas_node(unsigned int flags, int nid) { bool ret = false; + unsigned int cpuset_mems_cookie; if (!(flags & SHOW_MEM_FILTER_NODES)) goto out; - get_mems_allowed(); - ret = !node_isset(nid, cpuset_current_mems_allowed); - put_mems_allowed(); + do { + cpuset_mems_cookie = get_mems_allowed(); + ret = !node_isset(nid, cpuset_current_mems_allowed); + } while (!put_mems_allowed(cpuset_mems_cookie)); out: return ret; } diff --git a/mm/slab.c b/mm/slab.c index f0bd7857ab3b..29c8716eb7a9 100644 --- a/mm/slab.c +++ b/mm/slab.c @@ -3284,12 +3284,10 @@ static void *alternate_node_alloc(struct kmem_cache *cachep, gfp_t flags) if (in_interrupt() || (flags & __GFP_THISNODE)) return NULL; nid_alloc = nid_here = numa_mem_id(); - get_mems_allowed(); if (cpuset_do_slab_mem_spread() && (cachep->flags & SLAB_MEM_SPREAD)) nid_alloc = cpuset_slab_spread_node(); else if (current->mempolicy) nid_alloc = slab_node(current->mempolicy); - put_mems_allowed(); if (nid_alloc != nid_here) return ____cache_alloc_node(cachep, flags, nid_alloc); return NULL; @@ -3312,14 +3310,17 @@ static void *fallback_alloc(struct kmem_cache *cache, gfp_t flags) enum zone_type high_zoneidx = gfp_zone(flags); void *obj = NULL; int nid; + unsigned int cpuset_mems_cookie; if (flags & __GFP_THISNODE) return NULL; - get_mems_allowed(); - zonelist = node_zonelist(slab_node(current->mempolicy), flags); local_flags = flags & (GFP_CONSTRAINT_MASK|GFP_RECLAIM_MASK); +retry_cpuset: + cpuset_mems_cookie = get_mems_allowed(); + zonelist = node_zonelist(slab_node(current->mempolicy), flags); + retry: /* * Look through allowed nodes for objects available @@ -3372,7 +3373,9 @@ retry: } } } - put_mems_allowed(); + + if (unlikely(!put_mems_allowed(cpuset_mems_cookie) && !obj)) + goto retry_cpuset; return obj; } diff --git a/mm/slub.c b/mm/slub.c index 4907563ef7ff..f4a6229848fd 100644 --- a/mm/slub.c +++ b/mm/slub.c @@ -1581,6 +1581,7 @@ static struct page *get_any_partial(struct kmem_cache *s, gfp_t flags, struct zone *zone; enum zone_type high_zoneidx = gfp_zone(flags); void *object; + unsigned int cpuset_mems_cookie; /* * The defrag ratio allows a configuration of the tradeoffs between @@ -1604,23 +1605,32 @@ static struct page *get_any_partial(struct kmem_cache *s, gfp_t flags, get_cycles() % 1024 > s->remote_node_defrag_ratio) return NULL; - get_mems_allowed(); - zonelist = node_zonelist(slab_node(current->mempolicy), flags); - for_each_zone_zonelist(zone, z, zonelist, high_zoneidx) { - struct kmem_cache_node *n; - - n = get_node(s, zone_to_nid(zone)); - - if (n && cpuset_zone_allowed_hardwall(zone, flags) && - n->nr_partial > s->min_partial) { - object = get_partial_node(s, n, c); - if (object) { - put_mems_allowed(); - return object; + do { + cpuset_mems_cookie = get_mems_allowed(); + zonelist = node_zonelist(slab_node(current->mempolicy), flags); + for_each_zone_zonelist(zone, z, zonelist, high_zoneidx) { + struct kmem_cache_node *n; + + n = get_node(s, zone_to_nid(zone)); + + if (n && cpuset_zone_allowed_hardwall(zone, flags) && + n->nr_partial > s->min_partial) { + object = get_partial_node(s, n, c); + if (object) { + /* + * Return the object even if + * put_mems_allowed indicated that + * the cpuset mems_allowed was + * updated in parallel. It's a + * harmless race between the alloc + * and the cpuset update. + */ + put_mems_allowed(cpuset_mems_cookie); + return object; + } } } - } - put_mems_allowed(); + } while (!put_mems_allowed(cpuset_mems_cookie)); #endif return NULL; } diff --git a/mm/vmscan.c b/mm/vmscan.c index 440af1d899b9..55d86c9506f3 100644 --- a/mm/vmscan.c +++ b/mm/vmscan.c @@ -2343,7 +2343,6 @@ static unsigned long do_try_to_free_pages(struct zonelist *zonelist, unsigned long writeback_threshold; bool aborted_reclaim; - get_mems_allowed(); delayacct_freepages_start(); if (global_reclaim(sc)) @@ -2407,7 +2406,6 @@ static unsigned long do_try_to_free_pages(struct zonelist *zonelist, out: delayacct_freepages_end(); - put_mems_allowed(); if (sc->nr_reclaimed) return sc->nr_reclaimed; -- cgit v1.2.3 From 0fc9d1040313047edf6a39fd4d7c7defdca97c62 Mon Sep 17 00:00:00 2001 From: Konstantin Khlebnikov Date: Wed, 28 Mar 2012 14:42:54 -0700 Subject: radix-tree: use iterators in find_get_pages* functions Replace radix_tree_gang_lookup_slot() and radix_tree_gang_lookup_tag_slot() in page-cache lookup functions with brand-new radix-tree direct iterating. This avoids the double-scanning and pointer copying. Iterator don't stop after nr_pages page-get fails in a row, it continue lookup till the radix-tree end. Thus we can safely remove these restart conditions. Unfortunately, old implementation didn't forbid nr_pages == 0, this corner case does not fit into new code, so the patch adds an extra check at the beginning. Signed-off-by: Konstantin Khlebnikov Tested-by: Hugh Dickins Cc: Christoph Hellwig Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- mm/filemap.c | 86 +++++++++++++++++++++++++++--------------------------------- 1 file changed, 38 insertions(+), 48 deletions(-) (limited to 'mm/filemap.c') diff --git a/mm/filemap.c b/mm/filemap.c index c3811bc6b9e3..79c4b2b0b14e 100644 --- a/mm/filemap.c +++ b/mm/filemap.c @@ -813,20 +813,19 @@ EXPORT_SYMBOL(find_or_create_page); unsigned find_get_pages(struct address_space *mapping, pgoff_t start, unsigned int nr_pages, struct page **pages) { - unsigned int i; - unsigned int ret; - unsigned int nr_found, nr_skip; + struct radix_tree_iter iter; + void **slot; + unsigned ret = 0; + + if (unlikely(!nr_pages)) + return 0; rcu_read_lock(); restart: - nr_found = radix_tree_gang_lookup_slot(&mapping->page_tree, - (void ***)pages, NULL, start, nr_pages); - ret = 0; - nr_skip = 0; - for (i = 0; i < nr_found; i++) { + radix_tree_for_each_slot(slot, &mapping->page_tree, &iter, start) { struct page *page; repeat: - page = radix_tree_deref_slot((void **)pages[i]); + page = radix_tree_deref_slot(slot); if (unlikely(!page)) continue; @@ -837,7 +836,7 @@ repeat: * when entry at index 0 moves out of or back * to root: none yet gotten, safe to restart. */ - WARN_ON(start | i); + WARN_ON(iter.index); goto restart; } /* @@ -845,7 +844,6 @@ repeat: * here as an exceptional entry: so skip over it - * we only reach this from invalidate_mapping_pages(). */ - nr_skip++; continue; } @@ -853,21 +851,16 @@ repeat: goto repeat; /* Has the page moved? */ - if (unlikely(page != *((void **)pages[i]))) { + if (unlikely(page != *slot)) { page_cache_release(page); goto repeat; } pages[ret] = page; - ret++; + if (++ret == nr_pages) + break; } - /* - * If all entries were removed before we could secure them, - * try again, because callers stop trying once 0 is returned. - */ - if (unlikely(!ret && nr_found > nr_skip)) - goto restart; rcu_read_unlock(); return ret; } @@ -887,21 +880,22 @@ repeat: unsigned find_get_pages_contig(struct address_space *mapping, pgoff_t index, unsigned int nr_pages, struct page **pages) { - unsigned int i; - unsigned int ret; - unsigned int nr_found; + struct radix_tree_iter iter; + void **slot; + unsigned int ret = 0; + + if (unlikely(!nr_pages)) + return 0; rcu_read_lock(); restart: - nr_found = radix_tree_gang_lookup_slot(&mapping->page_tree, - (void ***)pages, NULL, index, nr_pages); - ret = 0; - for (i = 0; i < nr_found; i++) { + radix_tree_for_each_contig(slot, &mapping->page_tree, &iter, index) { struct page *page; repeat: - page = radix_tree_deref_slot((void **)pages[i]); + page = radix_tree_deref_slot(slot); + /* The hole, there no reason to continue */ if (unlikely(!page)) - continue; + break; if (radix_tree_exception(page)) { if (radix_tree_deref_retry(page)) { @@ -924,7 +918,7 @@ repeat: goto repeat; /* Has the page moved? */ - if (unlikely(page != *((void **)pages[i]))) { + if (unlikely(page != *slot)) { page_cache_release(page); goto repeat; } @@ -934,14 +928,14 @@ repeat: * otherwise we can get both false positives and false * negatives, which is just confusing to the caller. */ - if (page->mapping == NULL || page->index != index) { + if (page->mapping == NULL || page->index != iter.index) { page_cache_release(page); break; } pages[ret] = page; - ret++; - index++; + if (++ret == nr_pages) + break; } rcu_read_unlock(); return ret; @@ -962,19 +956,20 @@ EXPORT_SYMBOL(find_get_pages_contig); unsigned find_get_pages_tag(struct address_space *mapping, pgoff_t *index, int tag, unsigned int nr_pages, struct page **pages) { - unsigned int i; - unsigned int ret; - unsigned int nr_found; + struct radix_tree_iter iter; + void **slot; + unsigned ret = 0; + + if (unlikely(!nr_pages)) + return 0; rcu_read_lock(); restart: - nr_found = radix_tree_gang_lookup_tag_slot(&mapping->page_tree, - (void ***)pages, *index, nr_pages, tag); - ret = 0; - for (i = 0; i < nr_found; i++) { + radix_tree_for_each_tagged(slot, &mapping->page_tree, + &iter, *index, tag) { struct page *page; repeat: - page = radix_tree_deref_slot((void **)pages[i]); + page = radix_tree_deref_slot(slot); if (unlikely(!page)) continue; @@ -998,21 +993,16 @@ repeat: goto repeat; /* Has the page moved? */ - if (unlikely(page != *((void **)pages[i]))) { + if (unlikely(page != *slot)) { page_cache_release(page); goto repeat; } pages[ret] = page; - ret++; + if (++ret == nr_pages) + break; } - /* - * If all entries were removed before we could secure them, - * try again, because callers stop trying once 0 is returned. - */ - if (unlikely(!ret && nr_found)) - goto restart; rcu_read_unlock(); if (ret) -- cgit v1.2.3