swap: add a simple detector for inappropriate swapin readahead

The swapin readahead does a blind readahead whether or not the swapin is
sequential.  This is ok for harddisk because large reads have relatively
small costs and if the readahead pages are unneeded they can be reclaimed
easily.  But for SSD devices large reads are more expensive than small
one.  If readahead pages are unneeded, reading them in caused significant
overhead

This patch addes a simple random read detection similar to file mmap
readahead.  If a random read is detected, swapin readahead will be
skipped.  This improves a lot for a swap workload with random IO in a fast
SSD.

I run anonymous mmap write micro benchmark, which will triger swapin/swapout.

			runtime changes with patch
randwrite harddisk	-38.7%
seqwrite harddisk	-1.1%
randwrite SSD		-46.9%
seqwrite SSD		+0.3%

For both harddisk and SSD, the randwrite swap workload run time is reduced
significantly.  Sequential write swap workload hasn't chanage.

Interestingly, the randwrite harddisk test is improved too.  This might be
because swapin readahead needs to allocate extra memory, which further
tights memory pressure, so more swapout/swapin.

Signed-off-by: Shaohua Li <shli@fusionio.com>
Acked-by: Rik van Riel <riel@redhat.com>
Cc: Konstantin Khlebnikov <khlebnikov@openvz.org>
Cc: Wu Fengguang <fengguang.wu@intel.com>
Cc: Hugh Dickins <hughd@google.com>
Cc: Minchan Kim <minchan@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>

diff --git a/include/linux/rmap.h b/include/linux/rmap.h
index b2cce644ffc712faf6630e23c9d5fdbc10c3ae3d..1767045c287294836f274c030cda9dabdbc3f5bb 100644 (file)
--- a/include/linux/rmap.h
+++ b/include/linux/rmap.h
@@ -35,6 +35,9 @@ struct anon_vma {
         * anon_vma if they are the last user on release
         */
        atomic_t refcount;
+#ifdef CONFIG_SWAP
+       atomic_t swapra_miss;
+#endif
 
        /*
         * NOTE: the LSB of the rb_root.rb_node is set by

diff --git a/mm/internal.h b/mm/internal.h
index 8312d4fadf59dd32f21733b66ab14737e64da7b7..d09f59366de09692eb9c0bf73106f123bad5b77c 100644 (file)
--- a/mm/internal.h
+++ b/mm/internal.h
@@ -12,6 +12,7 @@
 #define __MM_INTERNAL_H
 
 #include <linux/mm.h>
+#include <linux/rmap.h>
 
 void free_pgtables(struct mmu_gather *tlb, struct vm_area_struct *start_vma,
                unsigned long floor, unsigned long ceiling);
@@ -358,4 +359,55 @@ extern unsigned long vm_mmap_pgoff(struct file *, unsigned long,
 extern void set_pageblock_order(void);
 unsigned long reclaim_clean_pages_from_list(struct zone *zone,
                                            struct list_head *page_list);
+
+/*
+ * Unnecessary readahead harms performance. 1. for SSD, big size read is more
+ * expensive than small size read, so extra unnecessary read only has overhead.
+ * For harddisk, this overhead doesn't exist. 2. unnecessary readahead will
+ * allocate extra memroy, which further tights memory pressure, so more
+ * swapout/swapin.
+ * These adds a simple swap random access detection. In swap page fault, if
+ * page is found in swap cache, decrease an account of vma, otherwise we need
+ * do sync swapin and the account is increased. Optionally swapin will do
+ * readahead if the counter is below a threshold.
+ */
+#ifdef CONFIG_SWAP
+#define SWAPRA_MISS_THRESHOLD  (100)
+#define SWAPRA_MAX_MISS ((SWAPRA_MISS_THRESHOLD) * 10)
+static inline void swap_cache_hit(struct vm_area_struct *vma)
+{
+       if (vma && vma->anon_vma)
+               atomic_dec_if_positive(&vma->anon_vma->swapra_miss);
+}
+
+static inline void swap_cache_miss(struct vm_area_struct *vma)
+{
+       if (!vma || !vma->anon_vma)
+               return;
+       if (atomic_read(&vma->anon_vma->swapra_miss) < SWAPRA_MAX_MISS)
+               atomic_inc(&vma->anon_vma->swapra_miss);
+}
+
+static inline int swap_cache_skip_readahead(struct vm_area_struct *vma)
+{
+       if (!vma || !vma->anon_vma)
+               return 0;
+       return atomic_read(&vma->anon_vma->swapra_miss) >
+               SWAPRA_MISS_THRESHOLD;
+}
+#else
+static inline void swap_cache_hit(struct vm_area_struct *vma)
+{
+}
+
+static inline void swap_cache_miss(struct vm_area_struct *vma)
+{
+}
+
+static inline int swap_cache_skip_readahead(struct vm_area_struct *vma)
+{
+       return 0;
+}
+#endif /* CONFIG_SWAP */
+
 #endif /* __MM_INTERNAL_H */

diff --git a/mm/memory.c b/mm/memory.c
index be4a2e6c487af29197cdd6c15fd208243e2b1cad..1a92d87575d74133e0dd62259ff07f8d1033d199 100644 (file)
--- a/mm/memory.c
+++ b/mm/memory.c
@@ -2967,7 +2967,8 @@ static int do_swap_page(struct mm_struct *mm, struct vm_area_struct *vma,
                ret = VM_FAULT_HWPOISON;
                delayacct_clear_flag(DELAYACCT_PF_SWAPIN);
                goto out_release;
-       }
+       } else if (!(flags & FAULT_FLAG_TRIED))
+               swap_cache_hit(vma);
 
        locked = lock_page_or_retry(page, mm, flags);
 

diff --git a/mm/rmap.c b/mm/rmap.c
index 11b25a1394bd38f54696e091a9094ab3e2e00df7..4d9c7acba5d9d2e48fd3d9dc2cbc5a3b5c87ed68 100644 (file)
--- a/mm/rmap.c
+++ b/mm/rmap.c
@@ -366,6 +366,9 @@ static void anon_vma_ctor(void *data)
 
        mutex_init(&anon_vma->mutex);
        atomic_set(&anon_vma->refcount, 0);
+#ifdef CONFIG_SWAP
+       atomic_set(&anon_vma->swapra_miss, 0);
+#endif
        anon_vma->rb_root = RB_ROOT;
 }
 

diff --git a/mm/shmem.c b/mm/shmem.c
index cc12072f8787d5adacc6cd49a78849ca5c2005a6..217aa9cd59df904e5031cef207052e65aec34900 100644 (file)
--- a/mm/shmem.c
+++ b/mm/shmem.c
@@ -922,6 +922,7 @@ static struct page *shmem_swapin(swp_entry_t swap, gfp_t gfp,
        pvma.vm_pgoff = index + info->vfs_inode.i_ino;
        pvma.vm_ops = NULL;
        pvma.vm_policy = spol;
+       pvma.anon_vma = NULL;
        return swapin_readahead(swap, gfp, &pvma, 0);
 }
 

diff --git a/mm/swap_state.c b/mm/swap_state.c
index 0cb36fb1f61cc539baa143319c40da30ada3d04e..d1f6c2df820e995ffb067597e779c9a491d8bfe7 100644 (file)
--- a/mm/swap_state.c
+++ b/mm/swap_state.c
@@ -20,6 +20,7 @@
 #include <linux/page_cgroup.h>
 
 #include <asm/pgtable.h>
+#include "internal.h"
 
 /*
  * swapper_space is a fiction, retained to simplify the path through
@@ -379,6 +380,10 @@ struct page *swapin_readahead(swp_entry_t entry, gfp_t gfp_mask,
        unsigned long mask = (1UL << page_cluster) - 1;
        struct blk_plug plug;
 
+       swap_cache_miss(vma);
+       if (swap_cache_skip_readahead(vma))
+               goto skip;
+
        /* Read a page_cluster sized and aligned cluster around offset. */
        start_offset = offset & ~mask;
        end_offset = offset | mask;
@@ -397,5 +402,6 @@ struct page *swapin_readahead(swp_entry_t entry, gfp_t gfp_mask,
        blk_finish_plug(&plug);
 
        lru_add_drain();        /* Push any new pages onto the LRU now */
+skip:
        return read_swap_cache_async(entry, gfp_mask, vma, addr);
 }