The patch titled
readahead: introduce context readahead algorithm
has been removed from the -mm tree. Its filename was
readahead-introduce-context-readahead-algorithm-update.patch
This patch was dropped because it was folded into readahead-introduce-context-readahead-algorithm.patch
The current -mm tree may be found at http://userweb.kernel.org/~akpm/mmotm/
------------------------------------------------------
Subject: readahead: introduce context readahead algorithm
From: Wu Fengguang <fengguang.wu@xxxxxxxxx>
Introduce page cache context based readahead algorithm.
This is to better support concurrent read streams in general.
RATIONALE
---------
The current readahead algorithm detects interleaved reads in a _passive_ way.
Given a sequence of interleaved streams 1,1001,2,1002,3,4,1003,5,1004,1005,6,...
By checking for (offset == prev_offset + 1), it will discover the sequentialness
between 3,4 and between 1004,1005, and start doing sequential readahead for the
individual streams since page 4 and page 1005.
The context readahead algorithm guarantees to discover the sequentialness no
matter how the streams are interleaved. For the above example, it will start
sequential readahead since page 2 and 1002.
The trick is to poke for page @offset-1 in the page cache when it has no other
clues on the sequentialness of request @offset: if the current requenst belongs
to a sequential stream, that stream must have accessed page @offset-1 recently,
and the page will still be cached now. So if page @offset-1 is there, we can
take request @offset as a sequential access.
BENEFICIARIES
-------------
- strictly interleaved reads i.e. 1,1001,2,1002,3,1003,...
the current readahead will take them as silly random reads;
the context readahead will take them as two sequential streams.
- cooperative IO processes i.e. NFS and SCST
They create a thread pool, farming off (sequential) IO requests to different
threads which will be performing interleaved IO.
It was not easy(or possible) to reliably tell from file->f_ra all those
cooperative processes working on the same sequential stream, since they will
have different file->f_ra instances. And NFSD's file->f_ra is particularly
unusable, since their file objects are dynamically created for each request.
The nfsd does have code trying to restore the f_ra bits, but not satisfactory.
The new scheme is to detect the sequential pattern via looking up the page
cache, which provides one single and consistent view of the pages recently
accessed. That makes sequential detection for cooperative processes possible.
USER REPORT
-----------
Vladislav recommends the addition of context readahead as a result of his SCST
benchmarks. It leads to 6%~40% performance gains in various cases and achieves
equal performance in others. http://lkml.org/lkml/2009/3/19/239
OVERHEADS
---------
In theory, it introduces one extra page cache lookup per random read. However
the below benchmark shows context readahead to be slightly faster, wondering..
Randomly reading 200MB amount of data on a sparse file, repeat 20 times for
each block size. The average throughputs are:
original ra context ra gain
4K random reads: 65.561MB/s 65.648MB/s +0.1%
16K random reads: 124.767MB/s 124.951MB/s +0.1%
64K random reads: 162.123MB/s 162.278MB/s +0.1%
Cc: Jens Axboe <jens.axboe@xxxxxxxxxx>
Cc: Jeff Moyer <jmoyer@xxxxxxxxxx>
Tested-by: Vladislav Bolkhovitin <vst@xxxxxxxx>
Signed-off-by: Wu Fengguang <fengguang.wu@xxxxxxxxx>
Signed-off-by: Andrew Morton <akpm@xxxxxxxxxxxxxxxxxxxx>
---
mm/readahead.c | 10 +++++-----
1 file changed, 5 insertions(+), 5 deletions(-)
diff -puN mm/readahead.c~readahead-introduce-context-readahead-algorithm-update mm/readahead.c
--- a/mm/readahead.c~readahead-introduce-context-readahead-algorithm-update
+++ a/mm/readahead.c
@@ -330,14 +330,14 @@ static unsigned long get_next_ra_size(st
*/
/*
- * Count continuously cached pages from @offset-1 to @offset-@max,
+ * Count contiguously cached pages from @offset-1 to @offset-@max,
* this count is a conservative estimation of
* - length of the sequential read sequence, or
* - thrashing threshold in memory tight systems
*/
-static unsigned long count_history_pages(struct address_space *mapping,
- struct file_ra_state *ra,
- pgoff_t offset, unsigned long max)
+static pgoff_t count_history_pages(struct address_space *mapping,
+ struct file_ra_state *ra,
+ pgoff_t offset, unsigned long max)
{
pgoff_t head;
@@ -357,7 +357,7 @@ static int try_context_readahead(struct
unsigned long req_size,
unsigned long max)
{
- unsigned long size;
+ pgoff_t size;
size = count_history_pages(mapping, ra, offset, max);
_
Patches currently in -mm which might be from fengguang.wu@xxxxxxxxx are
origin.patch
readahead-make-mmap_miss-an-unsigned-int.patch
readahead-move-max_sane_readahead-calls-into-force_page_cache_readahead.patch
readahead-apply-max_sane_readahead-limit-in-ondemand_readahead.patch
readahead-remove-one-unnecessary-radix-tree-lookup.patch
readahead-increase-interleaved-readahead-size.patch
readahead-remove-sync-async-readahead-call-dependency.patch
readahead-clean-up-and-simplify-the-code-for-filemap-page-fault-readahead.patch
readahead-sequential-mmap-readahead.patch
readahead-enforce-full-readahead-size-on-async-mmap-readahead.patch
readahead-record-mmap-read-around-states-in-file_ra_state.patch
radix-tree-add-radix_tree_prev_hole.patch
readahead-move-the-random-read-case-to-bottom.patch
readahead-introduce-context-readahead-algorithm.patch
readahead-introduce-context-readahead-algorithm-update.patch
readahead-remove-redundant-test-in-shrink_readahead_size_eio.patch
readahead-enforce-full-sync-mmap-readahead-size.patch
mm-introduce-pagehuge-for-testing-huge-gigantic-pages.patch
mm-introduce-pagehuge-for-testing-huge-gigantic-pages-update.patch
proc-kpagecount-kpageflags-code-cleanup.patch
proc-export-more-page-flags-in-proc-kpageflags.patch
pagemap-document-clarifications.patch
pagemap-document-9-more-exported-page-flags.patch
pagemap-add-page-types-tool.patch
pagemap-add-page-types-tool-fix.patch
vmscan-cleanup-the-scan-batching-code.patch
vmscan-dont-export-nr_saved_scan-in-proc-zoneinfo.patch
vmscan-zvc-updates-in-shrink_active_list-can-be-done-once.patch
mm-remove-__invalidate_mapping_pages-variant.patch
vmscan-report-vm_flags-in-page_referenced.patch
vmscan-make-mapped-executable-pages-the-first-class-citizen.patch
vmscan-merge-duplicate-code-in-shrink_active_list.patch
writeback-skip-new-or-to-be-freed-inodes.patch
vmscan-properly-account-for-the-number-of-page-cache-pages-zone_reclaim-can-reclaim.patch
vmscan-properly-account-for-the-number-of-page-cache-pages-zone_reclaim-can-reclaim-vmscan-fix-use-of-delta-in-zone_pagecache_reclaimable.patch
vmscan-properly-account-for-the-number-of-page-cache-pages-zone_reclaim-can-reclaim-mm-fix-documentation-of-min_unmapped_ratio.patch
vmscan-do-not-unconditionally-treat-zones-that-fail-zone_reclaim-as-full.patch
vmscan-count-the-number-of-times-zone_reclaim-scans-and-fails.patch
vmscan-count-the-number-of-times-zone_reclaim-scans-and-fails-fix.patch
--
To unsubscribe from this list: send the line "unsubscribe mm-commits" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
