kernel-fxtec-pro1x/mm/backing-dev.c

669 lines
17 KiB
C
Raw Normal View History

#include <linux/wait.h>
#include <linux/backing-dev.h>
writeback: switch to per-bdi threads for flushing data This gets rid of pdflush for bdi writeout and kupdated style cleaning. pdflush writeout suffers from lack of locality and also requires more threads to handle the same workload, since it has to work in a non-blocking fashion against each queue. This also introduces lumpy behaviour and potential request starvation, since pdflush can be starved for queue access if others are accessing it. A sample ffsb workload that does random writes to files is about 8% faster here on a simple SATA drive during the benchmark phase. File layout also seems a LOT more smooth in vmstat: r b swpd free buff cache si so bi bo in cs us sy id wa 0 1 0 608848 2652 375372 0 0 0 71024 604 24 1 10 48 42 0 1 0 549644 2712 433736 0 0 0 60692 505 27 1 8 48 44 1 0 0 476928 2784 505192 0 0 4 29540 553 24 0 9 53 37 0 1 0 457972 2808 524008 0 0 0 54876 331 16 0 4 38 58 0 1 0 366128 2928 614284 0 0 4 92168 710 58 0 13 53 34 0 1 0 295092 3000 684140 0 0 0 62924 572 23 0 9 53 37 0 1 0 236592 3064 741704 0 0 4 58256 523 17 0 8 48 44 0 1 0 165608 3132 811464 0 0 0 57460 560 21 0 8 54 38 0 1 0 102952 3200 873164 0 0 4 74748 540 29 1 10 48 41 0 1 0 48604 3252 926472 0 0 0 53248 469 29 0 7 47 45 where vanilla tends to fluctuate a lot in the creation phase: r b swpd free buff cache si so bi bo in cs us sy id wa 1 1 0 678716 5792 303380 0 0 0 74064 565 50 1 11 52 36 1 0 0 662488 5864 319396 0 0 4 352 302 329 0 2 47 51 0 1 0 599312 5924 381468 0 0 0 78164 516 55 0 9 51 40 0 1 0 519952 6008 459516 0 0 4 78156 622 56 1 11 52 37 1 1 0 436640 6092 541632 0 0 0 82244 622 54 0 11 48 41 0 1 0 436640 6092 541660 0 0 0 8 152 39 0 0 51 49 0 1 0 332224 6200 644252 0 0 4 102800 728 46 1 13 49 36 1 0 0 274492 6260 701056 0 0 4 12328 459 49 0 7 50 43 0 1 0 211220 6324 763356 0 0 0 106940 515 37 1 10 51 39 1 0 0 160412 6376 813468 0 0 0 8224 415 43 0 6 49 45 1 1 0 85980 6452 886556 0 0 4 113516 575 39 1 11 54 34 0 2 0 85968 6452 886620 0 0 0 1640 158 211 0 0 46 54 A 10 disk test with btrfs performs 26% faster with per-bdi flushing. A SSD based writeback test on XFS performs over 20% better as well, with the throughput being very stable around 1GB/sec, where pdflush only manages 750MB/sec and fluctuates wildly while doing so. Random buffered writes to many files behave a lot better as well, as does random mmap'ed writes. A separate thread is added to sync the super blocks. In the long term, adding sync_supers_bdi() functionality could get rid of this thread again. Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
2009-09-09 01:08:54 -06:00
#include <linux/kthread.h>
#include <linux/freezer.h>
#include <linux/fs.h>
#include <linux/pagemap.h>
writeback: switch to per-bdi threads for flushing data This gets rid of pdflush for bdi writeout and kupdated style cleaning. pdflush writeout suffers from lack of locality and also requires more threads to handle the same workload, since it has to work in a non-blocking fashion against each queue. This also introduces lumpy behaviour and potential request starvation, since pdflush can be starved for queue access if others are accessing it. A sample ffsb workload that does random writes to files is about 8% faster here on a simple SATA drive during the benchmark phase. File layout also seems a LOT more smooth in vmstat: r b swpd free buff cache si so bi bo in cs us sy id wa 0 1 0 608848 2652 375372 0 0 0 71024 604 24 1 10 48 42 0 1 0 549644 2712 433736 0 0 0 60692 505 27 1 8 48 44 1 0 0 476928 2784 505192 0 0 4 29540 553 24 0 9 53 37 0 1 0 457972 2808 524008 0 0 0 54876 331 16 0 4 38 58 0 1 0 366128 2928 614284 0 0 4 92168 710 58 0 13 53 34 0 1 0 295092 3000 684140 0 0 0 62924 572 23 0 9 53 37 0 1 0 236592 3064 741704 0 0 4 58256 523 17 0 8 48 44 0 1 0 165608 3132 811464 0 0 0 57460 560 21 0 8 54 38 0 1 0 102952 3200 873164 0 0 4 74748 540 29 1 10 48 41 0 1 0 48604 3252 926472 0 0 0 53248 469 29 0 7 47 45 where vanilla tends to fluctuate a lot in the creation phase: r b swpd free buff cache si so bi bo in cs us sy id wa 1 1 0 678716 5792 303380 0 0 0 74064 565 50 1 11 52 36 1 0 0 662488 5864 319396 0 0 4 352 302 329 0 2 47 51 0 1 0 599312 5924 381468 0 0 0 78164 516 55 0 9 51 40 0 1 0 519952 6008 459516 0 0 4 78156 622 56 1 11 52 37 1 1 0 436640 6092 541632 0 0 0 82244 622 54 0 11 48 41 0 1 0 436640 6092 541660 0 0 0 8 152 39 0 0 51 49 0 1 0 332224 6200 644252 0 0 4 102800 728 46 1 13 49 36 1 0 0 274492 6260 701056 0 0 4 12328 459 49 0 7 50 43 0 1 0 211220 6324 763356 0 0 0 106940 515 37 1 10 51 39 1 0 0 160412 6376 813468 0 0 0 8224 415 43 0 6 49 45 1 1 0 85980 6452 886556 0 0 4 113516 575 39 1 11 54 34 0 2 0 85968 6452 886620 0 0 0 1640 158 211 0 0 46 54 A 10 disk test with btrfs performs 26% faster with per-bdi flushing. A SSD based writeback test on XFS performs over 20% better as well, with the throughput being very stable around 1GB/sec, where pdflush only manages 750MB/sec and fluctuates wildly while doing so. Random buffered writes to many files behave a lot better as well, as does random mmap'ed writes. A separate thread is added to sync the super blocks. In the long term, adding sync_supers_bdi() functionality could get rid of this thread again. Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
2009-09-09 01:08:54 -06:00
#include <linux/mm.h>
#include <linux/sched.h>
#include <linux/module.h>
#include <linux/writeback.h>
#include <linux/device.h>
#include <trace/events/writeback.h>
static atomic_long_t bdi_seq = ATOMIC_LONG_INIT(0);
struct backing_dev_info default_backing_dev_info = {
.name = "default",
.ra_pages = VM_MAX_READAHEAD * 1024 / PAGE_CACHE_SIZE,
.state = 0,
.capabilities = BDI_CAP_MAP_COPY,
};
EXPORT_SYMBOL_GPL(default_backing_dev_info);
struct backing_dev_info noop_backing_dev_info = {
.name = "noop",
.capabilities = BDI_CAP_NO_ACCT_AND_WRITEBACK,
};
EXPORT_SYMBOL_GPL(noop_backing_dev_info);
static struct class *bdi_class;
/*
* bdi_lock protects updates to bdi_list. bdi_list has RCU reader side
* locking.
*/
writeback: switch to per-bdi threads for flushing data This gets rid of pdflush for bdi writeout and kupdated style cleaning. pdflush writeout suffers from lack of locality and also requires more threads to handle the same workload, since it has to work in a non-blocking fashion against each queue. This also introduces lumpy behaviour and potential request starvation, since pdflush can be starved for queue access if others are accessing it. A sample ffsb workload that does random writes to files is about 8% faster here on a simple SATA drive during the benchmark phase. File layout also seems a LOT more smooth in vmstat: r b swpd free buff cache si so bi bo in cs us sy id wa 0 1 0 608848 2652 375372 0 0 0 71024 604 24 1 10 48 42 0 1 0 549644 2712 433736 0 0 0 60692 505 27 1 8 48 44 1 0 0 476928 2784 505192 0 0 4 29540 553 24 0 9 53 37 0 1 0 457972 2808 524008 0 0 0 54876 331 16 0 4 38 58 0 1 0 366128 2928 614284 0 0 4 92168 710 58 0 13 53 34 0 1 0 295092 3000 684140 0 0 0 62924 572 23 0 9 53 37 0 1 0 236592 3064 741704 0 0 4 58256 523 17 0 8 48 44 0 1 0 165608 3132 811464 0 0 0 57460 560 21 0 8 54 38 0 1 0 102952 3200 873164 0 0 4 74748 540 29 1 10 48 41 0 1 0 48604 3252 926472 0 0 0 53248 469 29 0 7 47 45 where vanilla tends to fluctuate a lot in the creation phase: r b swpd free buff cache si so bi bo in cs us sy id wa 1 1 0 678716 5792 303380 0 0 0 74064 565 50 1 11 52 36 1 0 0 662488 5864 319396 0 0 4 352 302 329 0 2 47 51 0 1 0 599312 5924 381468 0 0 0 78164 516 55 0 9 51 40 0 1 0 519952 6008 459516 0 0 4 78156 622 56 1 11 52 37 1 1 0 436640 6092 541632 0 0 0 82244 622 54 0 11 48 41 0 1 0 436640 6092 541660 0 0 0 8 152 39 0 0 51 49 0 1 0 332224 6200 644252 0 0 4 102800 728 46 1 13 49 36 1 0 0 274492 6260 701056 0 0 4 12328 459 49 0 7 50 43 0 1 0 211220 6324 763356 0 0 0 106940 515 37 1 10 51 39 1 0 0 160412 6376 813468 0 0 0 8224 415 43 0 6 49 45 1 1 0 85980 6452 886556 0 0 4 113516 575 39 1 11 54 34 0 2 0 85968 6452 886620 0 0 0 1640 158 211 0 0 46 54 A 10 disk test with btrfs performs 26% faster with per-bdi flushing. A SSD based writeback test on XFS performs over 20% better as well, with the throughput being very stable around 1GB/sec, where pdflush only manages 750MB/sec and fluctuates wildly while doing so. Random buffered writes to many files behave a lot better as well, as does random mmap'ed writes. A separate thread is added to sync the super blocks. In the long term, adding sync_supers_bdi() functionality could get rid of this thread again. Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
2009-09-09 01:08:54 -06:00
DEFINE_SPINLOCK(bdi_lock);
LIST_HEAD(bdi_list);
writeback: switch to per-bdi threads for flushing data This gets rid of pdflush for bdi writeout and kupdated style cleaning. pdflush writeout suffers from lack of locality and also requires more threads to handle the same workload, since it has to work in a non-blocking fashion against each queue. This also introduces lumpy behaviour and potential request starvation, since pdflush can be starved for queue access if others are accessing it. A sample ffsb workload that does random writes to files is about 8% faster here on a simple SATA drive during the benchmark phase. File layout also seems a LOT more smooth in vmstat: r b swpd free buff cache si so bi bo in cs us sy id wa 0 1 0 608848 2652 375372 0 0 0 71024 604 24 1 10 48 42 0 1 0 549644 2712 433736 0 0 0 60692 505 27 1 8 48 44 1 0 0 476928 2784 505192 0 0 4 29540 553 24 0 9 53 37 0 1 0 457972 2808 524008 0 0 0 54876 331 16 0 4 38 58 0 1 0 366128 2928 614284 0 0 4 92168 710 58 0 13 53 34 0 1 0 295092 3000 684140 0 0 0 62924 572 23 0 9 53 37 0 1 0 236592 3064 741704 0 0 4 58256 523 17 0 8 48 44 0 1 0 165608 3132 811464 0 0 0 57460 560 21 0 8 54 38 0 1 0 102952 3200 873164 0 0 4 74748 540 29 1 10 48 41 0 1 0 48604 3252 926472 0 0 0 53248 469 29 0 7 47 45 where vanilla tends to fluctuate a lot in the creation phase: r b swpd free buff cache si so bi bo in cs us sy id wa 1 1 0 678716 5792 303380 0 0 0 74064 565 50 1 11 52 36 1 0 0 662488 5864 319396 0 0 4 352 302 329 0 2 47 51 0 1 0 599312 5924 381468 0 0 0 78164 516 55 0 9 51 40 0 1 0 519952 6008 459516 0 0 4 78156 622 56 1 11 52 37 1 1 0 436640 6092 541632 0 0 0 82244 622 54 0 11 48 41 0 1 0 436640 6092 541660 0 0 0 8 152 39 0 0 51 49 0 1 0 332224 6200 644252 0 0 4 102800 728 46 1 13 49 36 1 0 0 274492 6260 701056 0 0 4 12328 459 49 0 7 50 43 0 1 0 211220 6324 763356 0 0 0 106940 515 37 1 10 51 39 1 0 0 160412 6376 813468 0 0 0 8224 415 43 0 6 49 45 1 1 0 85980 6452 886556 0 0 4 113516 575 39 1 11 54 34 0 2 0 85968 6452 886620 0 0 0 1640 158 211 0 0 46 54 A 10 disk test with btrfs performs 26% faster with per-bdi flushing. A SSD based writeback test on XFS performs over 20% better as well, with the throughput being very stable around 1GB/sec, where pdflush only manages 750MB/sec and fluctuates wildly while doing so. Random buffered writes to many files behave a lot better as well, as does random mmap'ed writes. A separate thread is added to sync the super blocks. In the long term, adding sync_supers_bdi() functionality could get rid of this thread again. Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
2009-09-09 01:08:54 -06:00
writeback: replace custom worker pool implementation with unbound workqueue Writeback implements its own worker pool - each bdi can be associated with a worker thread which is created and destroyed dynamically. The worker thread for the default bdi is always present and serves as the "forker" thread which forks off worker threads for other bdis. there's no reason for writeback to implement its own worker pool when using unbound workqueue instead is much simpler and more efficient. This patch replaces custom worker pool implementation in writeback with an unbound workqueue. The conversion isn't too complicated but the followings are worth mentioning. * bdi_writeback->last_active, task and wakeup_timer are removed. delayed_work ->dwork is added instead. Explicit timer handling is no longer necessary. Everything works by either queueing / modding / flushing / canceling the delayed_work item. * bdi_writeback_thread() becomes bdi_writeback_workfn() which runs off bdi_writeback->dwork. On each execution, it processes bdi->work_list and reschedules itself if there are more things to do. The function also handles low-mem condition, which used to be handled by the forker thread. If the function is running off a rescuer thread, it only writes out limited number of pages so that the rescuer can serve other bdis too. This preserves the flusher creation failure behavior of the forker thread. * INIT_LIST_HEAD(&bdi->bdi_list) is used to tell bdi_writeback_workfn() about on-going bdi unregistration so that it always drains work_list even if it's running off the rescuer. Note that the original code was broken in this regard. Under memory pressure, a bdi could finish unregistration with non-empty work_list. * The default bdi is no longer special. It now is treated the same as any other bdi and bdi_cap_flush_forker() is removed. * BDI_pending is no longer used. Removed. * Some tracepoints become non-applicable. The following TPs are removed - writeback_nothread, writeback_wake_thread, writeback_wake_forker_thread, writeback_thread_start, writeback_thread_stop. Everything, including devices coming and going away and rescuer operation under simulated memory pressure, seems to work fine in my test setup. Signed-off-by: Tejun Heo <tj@kernel.org> Reviewed-by: Jan Kara <jack@suse.cz> Cc: Jens Axboe <axboe@kernel.dk> Cc: Fengguang Wu <fengguang.wu@intel.com> Cc: Jeff Moyer <jmoyer@redhat.com>
2013-04-01 20:08:06 -06:00
/* bdi_wq serves all asynchronous writeback tasks */
struct workqueue_struct *bdi_wq;
static void bdi_lock_two(struct bdi_writeback *wb1, struct bdi_writeback *wb2)
writeback: split inode_wb_list_lock into bdi_writeback.list_lock Split the global inode_wb_list_lock into a per-bdi_writeback list_lock, as it's currently the most contended lock in the system for metadata heavy workloads. It won't help for single-filesystem workloads for which we'll need the I/O-less balance_dirty_pages, but at least we can dedicate a cpu to spinning on each bdi now for larger systems. Based on earlier patches from Nick Piggin and Dave Chinner. It reduces lock contentions to 1/4 in this test case: 10 HDD JBOD, 100 dd on each disk, XFS, 6GB ram lock_stat version 0.3 ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- class name con-bounces contentions waittime-min waittime-max waittime-total acq-bounces acquisitions holdtime-min holdtime-max holdtime-total ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- vanilla 2.6.39-rc3: inode_wb_list_lock: 42590 44433 0.12 147.74 144127.35 252274 886792 0.08 121.34 917211.23 ------------------ inode_wb_list_lock 2 [<ffffffff81165da5>] bdev_inode_switch_bdi+0x29/0x85 inode_wb_list_lock 34 [<ffffffff8115bd0b>] inode_wb_list_del+0x22/0x49 inode_wb_list_lock 12893 [<ffffffff8115bb53>] __mark_inode_dirty+0x170/0x1d0 inode_wb_list_lock 10702 [<ffffffff8115afef>] writeback_single_inode+0x16d/0x20a ------------------ inode_wb_list_lock 2 [<ffffffff81165da5>] bdev_inode_switch_bdi+0x29/0x85 inode_wb_list_lock 19 [<ffffffff8115bd0b>] inode_wb_list_del+0x22/0x49 inode_wb_list_lock 5550 [<ffffffff8115bb53>] __mark_inode_dirty+0x170/0x1d0 inode_wb_list_lock 8511 [<ffffffff8115b4ad>] writeback_sb_inodes+0x10f/0x157 2.6.39-rc3 + patch: &(&wb->list_lock)->rlock: 11383 11657 0.14 151.69 40429.51 90825 527918 0.11 145.90 556843.37 ------------------------ &(&wb->list_lock)->rlock 10 [<ffffffff8115b189>] inode_wb_list_del+0x5f/0x86 &(&wb->list_lock)->rlock 1493 [<ffffffff8115b1ed>] writeback_inodes_wb+0x3d/0x150 &(&wb->list_lock)->rlock 3652 [<ffffffff8115a8e9>] writeback_sb_inodes+0x123/0x16f &(&wb->list_lock)->rlock 1412 [<ffffffff8115a38e>] writeback_single_inode+0x17f/0x223 ------------------------ &(&wb->list_lock)->rlock 3 [<ffffffff8110b5af>] bdi_lock_two+0x46/0x4b &(&wb->list_lock)->rlock 6 [<ffffffff8115b189>] inode_wb_list_del+0x5f/0x86 &(&wb->list_lock)->rlock 2061 [<ffffffff8115af97>] __mark_inode_dirty+0x173/0x1cf &(&wb->list_lock)->rlock 2629 [<ffffffff8115a8e9>] writeback_sb_inodes+0x123/0x16f hughd@google.com: fix recursive lock when bdi_lock_two() is called with new the same as old akpm@linux-foundation.org: cleanup bdev_inode_switch_bdi() comment Signed-off-by: Christoph Hellwig <hch@lst.de> Signed-off-by: Hugh Dickins <hughd@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
2011-04-21 18:19:44 -06:00
{
if (wb1 < wb2) {
spin_lock(&wb1->list_lock);
spin_lock_nested(&wb2->list_lock, 1);
} else {
spin_lock(&wb2->list_lock);
spin_lock_nested(&wb1->list_lock, 1);
}
}
#ifdef CONFIG_DEBUG_FS
#include <linux/debugfs.h>
#include <linux/seq_file.h>
static struct dentry *bdi_debug_root;
static void bdi_debug_init(void)
{
bdi_debug_root = debugfs_create_dir("bdi", NULL);
}
static int bdi_debug_stats_show(struct seq_file *m, void *v)
{
struct backing_dev_info *bdi = m->private;
struct bdi_writeback *wb = &bdi->wb;
unsigned long background_thresh;
unsigned long dirty_thresh;
unsigned long bdi_thresh;
unsigned long nr_dirty, nr_io, nr_more_io;
struct inode *inode;
nr_dirty = nr_io = nr_more_io = 0;
writeback: split inode_wb_list_lock into bdi_writeback.list_lock Split the global inode_wb_list_lock into a per-bdi_writeback list_lock, as it's currently the most contended lock in the system for metadata heavy workloads. It won't help for single-filesystem workloads for which we'll need the I/O-less balance_dirty_pages, but at least we can dedicate a cpu to spinning on each bdi now for larger systems. Based on earlier patches from Nick Piggin and Dave Chinner. It reduces lock contentions to 1/4 in this test case: 10 HDD JBOD, 100 dd on each disk, XFS, 6GB ram lock_stat version 0.3 ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- class name con-bounces contentions waittime-min waittime-max waittime-total acq-bounces acquisitions holdtime-min holdtime-max holdtime-total ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- vanilla 2.6.39-rc3: inode_wb_list_lock: 42590 44433 0.12 147.74 144127.35 252274 886792 0.08 121.34 917211.23 ------------------ inode_wb_list_lock 2 [<ffffffff81165da5>] bdev_inode_switch_bdi+0x29/0x85 inode_wb_list_lock 34 [<ffffffff8115bd0b>] inode_wb_list_del+0x22/0x49 inode_wb_list_lock 12893 [<ffffffff8115bb53>] __mark_inode_dirty+0x170/0x1d0 inode_wb_list_lock 10702 [<ffffffff8115afef>] writeback_single_inode+0x16d/0x20a ------------------ inode_wb_list_lock 2 [<ffffffff81165da5>] bdev_inode_switch_bdi+0x29/0x85 inode_wb_list_lock 19 [<ffffffff8115bd0b>] inode_wb_list_del+0x22/0x49 inode_wb_list_lock 5550 [<ffffffff8115bb53>] __mark_inode_dirty+0x170/0x1d0 inode_wb_list_lock 8511 [<ffffffff8115b4ad>] writeback_sb_inodes+0x10f/0x157 2.6.39-rc3 + patch: &(&wb->list_lock)->rlock: 11383 11657 0.14 151.69 40429.51 90825 527918 0.11 145.90 556843.37 ------------------------ &(&wb->list_lock)->rlock 10 [<ffffffff8115b189>] inode_wb_list_del+0x5f/0x86 &(&wb->list_lock)->rlock 1493 [<ffffffff8115b1ed>] writeback_inodes_wb+0x3d/0x150 &(&wb->list_lock)->rlock 3652 [<ffffffff8115a8e9>] writeback_sb_inodes+0x123/0x16f &(&wb->list_lock)->rlock 1412 [<ffffffff8115a38e>] writeback_single_inode+0x17f/0x223 ------------------------ &(&wb->list_lock)->rlock 3 [<ffffffff8110b5af>] bdi_lock_two+0x46/0x4b &(&wb->list_lock)->rlock 6 [<ffffffff8115b189>] inode_wb_list_del+0x5f/0x86 &(&wb->list_lock)->rlock 2061 [<ffffffff8115af97>] __mark_inode_dirty+0x173/0x1cf &(&wb->list_lock)->rlock 2629 [<ffffffff8115a8e9>] writeback_sb_inodes+0x123/0x16f hughd@google.com: fix recursive lock when bdi_lock_two() is called with new the same as old akpm@linux-foundation.org: cleanup bdev_inode_switch_bdi() comment Signed-off-by: Christoph Hellwig <hch@lst.de> Signed-off-by: Hugh Dickins <hughd@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
2011-04-21 18:19:44 -06:00
spin_lock(&wb->list_lock);
list_for_each_entry(inode, &wb->b_dirty, i_wb_list)
nr_dirty++;
list_for_each_entry(inode, &wb->b_io, i_wb_list)
nr_io++;
list_for_each_entry(inode, &wb->b_more_io, i_wb_list)
nr_more_io++;
writeback: split inode_wb_list_lock into bdi_writeback.list_lock Split the global inode_wb_list_lock into a per-bdi_writeback list_lock, as it's currently the most contended lock in the system for metadata heavy workloads. It won't help for single-filesystem workloads for which we'll need the I/O-less balance_dirty_pages, but at least we can dedicate a cpu to spinning on each bdi now for larger systems. Based on earlier patches from Nick Piggin and Dave Chinner. It reduces lock contentions to 1/4 in this test case: 10 HDD JBOD, 100 dd on each disk, XFS, 6GB ram lock_stat version 0.3 ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- class name con-bounces contentions waittime-min waittime-max waittime-total acq-bounces acquisitions holdtime-min holdtime-max holdtime-total ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- vanilla 2.6.39-rc3: inode_wb_list_lock: 42590 44433 0.12 147.74 144127.35 252274 886792 0.08 121.34 917211.23 ------------------ inode_wb_list_lock 2 [<ffffffff81165da5>] bdev_inode_switch_bdi+0x29/0x85 inode_wb_list_lock 34 [<ffffffff8115bd0b>] inode_wb_list_del+0x22/0x49 inode_wb_list_lock 12893 [<ffffffff8115bb53>] __mark_inode_dirty+0x170/0x1d0 inode_wb_list_lock 10702 [<ffffffff8115afef>] writeback_single_inode+0x16d/0x20a ------------------ inode_wb_list_lock 2 [<ffffffff81165da5>] bdev_inode_switch_bdi+0x29/0x85 inode_wb_list_lock 19 [<ffffffff8115bd0b>] inode_wb_list_del+0x22/0x49 inode_wb_list_lock 5550 [<ffffffff8115bb53>] __mark_inode_dirty+0x170/0x1d0 inode_wb_list_lock 8511 [<ffffffff8115b4ad>] writeback_sb_inodes+0x10f/0x157 2.6.39-rc3 + patch: &(&wb->list_lock)->rlock: 11383 11657 0.14 151.69 40429.51 90825 527918 0.11 145.90 556843.37 ------------------------ &(&wb->list_lock)->rlock 10 [<ffffffff8115b189>] inode_wb_list_del+0x5f/0x86 &(&wb->list_lock)->rlock 1493 [<ffffffff8115b1ed>] writeback_inodes_wb+0x3d/0x150 &(&wb->list_lock)->rlock 3652 [<ffffffff8115a8e9>] writeback_sb_inodes+0x123/0x16f &(&wb->list_lock)->rlock 1412 [<ffffffff8115a38e>] writeback_single_inode+0x17f/0x223 ------------------------ &(&wb->list_lock)->rlock 3 [<ffffffff8110b5af>] bdi_lock_two+0x46/0x4b &(&wb->list_lock)->rlock 6 [<ffffffff8115b189>] inode_wb_list_del+0x5f/0x86 &(&wb->list_lock)->rlock 2061 [<ffffffff8115af97>] __mark_inode_dirty+0x173/0x1cf &(&wb->list_lock)->rlock 2629 [<ffffffff8115a8e9>] writeback_sb_inodes+0x123/0x16f hughd@google.com: fix recursive lock when bdi_lock_two() is called with new the same as old akpm@linux-foundation.org: cleanup bdev_inode_switch_bdi() comment Signed-off-by: Christoph Hellwig <hch@lst.de> Signed-off-by: Hugh Dickins <hughd@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
2011-04-21 18:19:44 -06:00
spin_unlock(&wb->list_lock);
global_dirty_limits(&background_thresh, &dirty_thresh);
bdi_thresh = bdi_dirty_limit(bdi, dirty_thresh);
#define K(x) ((x) << (PAGE_SHIFT - 10))
seq_printf(m,
"BdiWriteback: %10lu kB\n"
"BdiReclaimable: %10lu kB\n"
"BdiDirtyThresh: %10lu kB\n"
"DirtyThresh: %10lu kB\n"
"BackgroundThresh: %10lu kB\n"
"BdiDirtied: %10lu kB\n"
"BdiWritten: %10lu kB\n"
"BdiWriteBandwidth: %10lu kBps\n"
"b_dirty: %10lu\n"
"b_io: %10lu\n"
"b_more_io: %10lu\n"
"bdi_list: %10u\n"
"state: %10lx\n",
(unsigned long) K(bdi_stat(bdi, BDI_WRITEBACK)),
(unsigned long) K(bdi_stat(bdi, BDI_RECLAIMABLE)),
K(bdi_thresh),
K(dirty_thresh),
K(background_thresh),
(unsigned long) K(bdi_stat(bdi, BDI_DIRTIED)),
(unsigned long) K(bdi_stat(bdi, BDI_WRITTEN)),
(unsigned long) K(bdi->write_bandwidth),
nr_dirty,
nr_io,
nr_more_io,
!list_empty(&bdi->bdi_list), bdi->state);
#undef K
return 0;
}
static int bdi_debug_stats_open(struct inode *inode, struct file *file)
{
return single_open(file, bdi_debug_stats_show, inode->i_private);
}
static const struct file_operations bdi_debug_stats_fops = {
.open = bdi_debug_stats_open,
.read = seq_read,
.llseek = seq_lseek,
.release = single_release,
};
static void bdi_debug_register(struct backing_dev_info *bdi, const char *name)
{
bdi->debug_dir = debugfs_create_dir(name, bdi_debug_root);
bdi->debug_stats = debugfs_create_file("stats", 0444, bdi->debug_dir,
bdi, &bdi_debug_stats_fops);
}
static void bdi_debug_unregister(struct backing_dev_info *bdi)
{
debugfs_remove(bdi->debug_stats);
debugfs_remove(bdi->debug_dir);
}
#else
static inline void bdi_debug_init(void)
{
}
static inline void bdi_debug_register(struct backing_dev_info *bdi,
const char *name)
{
}
static inline void bdi_debug_unregister(struct backing_dev_info *bdi)
{
}
#endif
static ssize_t read_ahead_kb_store(struct device *dev,
struct device_attribute *attr,
const char *buf, size_t count)
{
struct backing_dev_info *bdi = dev_get_drvdata(dev);
unsigned long read_ahead_kb;
ssize_t ret;
ret = kstrtoul(buf, 10, &read_ahead_kb);
if (ret < 0)
return ret;
bdi->ra_pages = read_ahead_kb >> (PAGE_SHIFT - 10);
return count;
}
#define K(pages) ((pages) << (PAGE_SHIFT - 10))
#define BDI_SHOW(name, expr) \
static ssize_t name##_show(struct device *dev, \
struct device_attribute *attr, char *page) \
{ \
struct backing_dev_info *bdi = dev_get_drvdata(dev); \
\
return snprintf(page, PAGE_SIZE-1, "%lld\n", (long long)expr); \
} \
static DEVICE_ATTR_RW(name);
BDI_SHOW(read_ahead_kb, K(bdi->ra_pages))
static ssize_t min_ratio_store(struct device *dev,
struct device_attribute *attr, const char *buf, size_t count)
{
struct backing_dev_info *bdi = dev_get_drvdata(dev);
unsigned int ratio;
ssize_t ret;
ret = kstrtouint(buf, 10, &ratio);
if (ret < 0)
return ret;
ret = bdi_set_min_ratio(bdi, ratio);
if (!ret)
ret = count;
return ret;
}
BDI_SHOW(min_ratio, bdi->min_ratio)
static ssize_t max_ratio_store(struct device *dev,
struct device_attribute *attr, const char *buf, size_t count)
{
struct backing_dev_info *bdi = dev_get_drvdata(dev);
unsigned int ratio;
ssize_t ret;
ret = kstrtouint(buf, 10, &ratio);
if (ret < 0)
return ret;
ret = bdi_set_max_ratio(bdi, ratio);
if (!ret)
ret = count;
return ret;
}
BDI_SHOW(max_ratio, bdi->max_ratio)
bdi: allow block devices to say that they require stable page writes This patchset ("stable page writes, part 2") makes some key modifications to the original 'stable page writes' patchset. First, it provides creators (devices and filesystems) of a backing_dev_info a flag that declares whether or not it is necessary to ensure that page contents cannot change during writeout. It is no longer assumed that this is true of all devices (which was never true anyway). Second, the flag is used to relaxed the wait_on_page_writeback calls so that wait only occurs if the device needs it. Third, it fixes up the remaining disk-backed filesystems to use this improved conditional-wait logic to provide stable page writes on those filesystems. It is hoped that (for people not using checksumming devices, anyway) this patchset will give back unnecessary performance decreases since the original stable page write patchset went into 3.0. Sorry about not fixing it sooner. Complaints were registered by several people about the long write latencies introduced by the original stable page write patchset. Generally speaking, the kernel ought to allocate as little extra memory as possible to facilitate writeout, but for people who simply cannot wait, a second page stability strategy is (re)introduced: snapshotting page contents. The waiting behavior is still the default strategy; to enable page snapshotting, a superblock flag (MS_SNAP_STABLE) must be set. This flag is used to bandaid^Henable stable page writeback on ext3[1], and is not used anywhere else. Given that there are already a few storage devices and network FSes that have rolled their own page stability wait/page snapshot code, it would be nice to move towards consolidating all of these. It seems possible that iscsi and raid5 may wish to use the new stable page write support to enable zero-copy writeout. Thank you to Jan Kara for helping fix a couple more filesystems. Per Andrew Morton's request, here are the result of using dbench to measure latencies on ext2: 3.8.0-rc3: Operation Count AvgLat MaxLat ---------------------------------------- WriteX 109347 0.028 59.817 ReadX 347180 0.004 3.391 Flush 15514 29.828 287.283 Throughput 57.429 MB/sec 4 clients 4 procs max_latency=287.290 ms 3.8.0-rc3 + patches: WriteX 105556 0.029 4.273 ReadX 335004 0.005 4.112 Flush 14982 30.540 298.634 Throughput 55.4496 MB/sec 4 clients 4 procs max_latency=298.650 ms As you can see, for ext2 the maximum write latency decreases from ~60ms on a laptop hard disk to ~4ms. I'm not sure why the flush latencies increase, though I suspect that being able to dirty pages faster gives the flusher more work to do. On ext4, the average write latency decreases as well as all the maximum latencies: 3.8.0-rc3: WriteX 85624 0.152 33.078 ReadX 272090 0.010 61.210 Flush 12129 36.219 168.260 Throughput 44.8618 MB/sec 4 clients 4 procs max_latency=168.276 ms 3.8.0-rc3 + patches: WriteX 86082 0.141 30.928 ReadX 273358 0.010 36.124 Flush 12214 34.800 165.689 Throughput 44.9941 MB/sec 4 clients 4 procs max_latency=165.722 ms XFS seems to exhibit similar latency improvements as ext2: 3.8.0-rc3: WriteX 125739 0.028 104.343 ReadX 399070 0.005 4.115 Flush 17851 25.004 131.390 Throughput 66.0024 MB/sec 4 clients 4 procs max_latency=131.406 ms 3.8.0-rc3 + patches: WriteX 123529 0.028 6.299 ReadX 392434 0.005 4.287 Flush 17549 25.120 188.687 Throughput 64.9113 MB/sec 4 clients 4 procs max_latency=188.704 ms ...and btrfs, just to round things out, also shows some latency decreases: 3.8.0-rc3: WriteX 67122 0.083 82.355 ReadX 212719 0.005 2.828 Flush 9547 47.561 147.418 Throughput 35.3391 MB/sec 4 clients 4 procs max_latency=147.433 ms 3.8.0-rc3 + patches: WriteX 64898 0.101 71.631 ReadX 206673 0.005 7.123 Flush 9190 47.963 219.034 Throughput 34.0795 MB/sec 4 clients 4 procs max_latency=219.044 ms Before this patchset, all filesystems would block, regardless of whether or not it was necessary. ext3 would wait, but still generate occasional checksum errors. The network filesystems were left to do their own thing, so they'd wait too. After this patchset, all the disk filesystems except ext3 and btrfs will wait only if the hardware requires it. ext3 (if necessary) snapshots pages instead of blocking, and btrfs provides its own bdi so the mm will never wait. Network filesystems haven't been touched, so either they provide their own wait code, or they don't block at all. The blocking behavior is back to what it was before 3.0 if you don't have a disk requiring stable page writes. This patchset has been tested on 3.8.0-rc3 on x64 with ext3, ext4, and xfs. I've spot-checked 3.8.0-rc4 and seem to be getting the same results as -rc3. [1] The alternative fixes to ext3 include fixing the locking order and page bit handling like we did for ext4 (but then why not just use ext4?), or setting PG_writeback so early that ext3 becomes extremely slow. I tried that, but the number of write()s I could initiate dropped by nearly an order of magnitude. That was a bit much even for the author of the stable page series! :) This patch: Creates a per-backing-device flag that tracks whether or not pages must be held immutable during writeout. Eventually it will be used to waive wait_for_page_writeback() if nothing requires stable pages. Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Jan Kara <jack@suse.cz> Cc: Adrian Hunter <adrian.hunter@intel.com> Cc: Andy Lutomirski <luto@amacapital.net> Cc: Artem Bityutskiy <dedekind1@gmail.com> Cc: Joel Becker <jlbec@evilplan.org> Cc: Mark Fasheh <mfasheh@suse.com> Cc: Steven Whitehouse <swhiteho@redhat.com> Cc: Jens Axboe <axboe@kernel.dk> Cc: Eric Van Hensbergen <ericvh@gmail.com> Cc: Ron Minnich <rminnich@sandia.gov> Cc: Latchesar Ionkov <lucho@ionkov.net> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2013-02-21 17:42:48 -07:00
static ssize_t stable_pages_required_show(struct device *dev,
struct device_attribute *attr,
char *page)
{
struct backing_dev_info *bdi = dev_get_drvdata(dev);
return snprintf(page, PAGE_SIZE-1, "%d\n",
bdi_cap_stable_pages_required(bdi) ? 1 : 0);
}
static DEVICE_ATTR_RO(stable_pages_required);
static struct attribute *bdi_dev_attrs[] = {
&dev_attr_read_ahead_kb.attr,
&dev_attr_min_ratio.attr,
&dev_attr_max_ratio.attr,
&dev_attr_stable_pages_required.attr,
NULL,
};
ATTRIBUTE_GROUPS(bdi_dev);
static __init int bdi_class_init(void)
{
bdi_class = class_create(THIS_MODULE, "bdi");
if (IS_ERR(bdi_class))
return PTR_ERR(bdi_class);
bdi_class->dev_groups = bdi_dev_groups;
bdi_debug_init();
return 0;
}
postcore_initcall(bdi_class_init);
static int __init default_bdi_init(void)
{
int err;
writeback: replace custom worker pool implementation with unbound workqueue Writeback implements its own worker pool - each bdi can be associated with a worker thread which is created and destroyed dynamically. The worker thread for the default bdi is always present and serves as the "forker" thread which forks off worker threads for other bdis. there's no reason for writeback to implement its own worker pool when using unbound workqueue instead is much simpler and more efficient. This patch replaces custom worker pool implementation in writeback with an unbound workqueue. The conversion isn't too complicated but the followings are worth mentioning. * bdi_writeback->last_active, task and wakeup_timer are removed. delayed_work ->dwork is added instead. Explicit timer handling is no longer necessary. Everything works by either queueing / modding / flushing / canceling the delayed_work item. * bdi_writeback_thread() becomes bdi_writeback_workfn() which runs off bdi_writeback->dwork. On each execution, it processes bdi->work_list and reschedules itself if there are more things to do. The function also handles low-mem condition, which used to be handled by the forker thread. If the function is running off a rescuer thread, it only writes out limited number of pages so that the rescuer can serve other bdis too. This preserves the flusher creation failure behavior of the forker thread. * INIT_LIST_HEAD(&bdi->bdi_list) is used to tell bdi_writeback_workfn() about on-going bdi unregistration so that it always drains work_list even if it's running off the rescuer. Note that the original code was broken in this regard. Under memory pressure, a bdi could finish unregistration with non-empty work_list. * The default bdi is no longer special. It now is treated the same as any other bdi and bdi_cap_flush_forker() is removed. * BDI_pending is no longer used. Removed. * Some tracepoints become non-applicable. The following TPs are removed - writeback_nothread, writeback_wake_thread, writeback_wake_forker_thread, writeback_thread_start, writeback_thread_stop. Everything, including devices coming and going away and rescuer operation under simulated memory pressure, seems to work fine in my test setup. Signed-off-by: Tejun Heo <tj@kernel.org> Reviewed-by: Jan Kara <jack@suse.cz> Cc: Jens Axboe <axboe@kernel.dk> Cc: Fengguang Wu <fengguang.wu@intel.com> Cc: Jeff Moyer <jmoyer@redhat.com>
2013-04-01 20:08:06 -06:00
bdi_wq = alloc_workqueue("writeback", WQ_MEM_RECLAIM | WQ_FREEZABLE |
WQ_UNBOUND | WQ_SYSFS, 0);
writeback: replace custom worker pool implementation with unbound workqueue Writeback implements its own worker pool - each bdi can be associated with a worker thread which is created and destroyed dynamically. The worker thread for the default bdi is always present and serves as the "forker" thread which forks off worker threads for other bdis. there's no reason for writeback to implement its own worker pool when using unbound workqueue instead is much simpler and more efficient. This patch replaces custom worker pool implementation in writeback with an unbound workqueue. The conversion isn't too complicated but the followings are worth mentioning. * bdi_writeback->last_active, task and wakeup_timer are removed. delayed_work ->dwork is added instead. Explicit timer handling is no longer necessary. Everything works by either queueing / modding / flushing / canceling the delayed_work item. * bdi_writeback_thread() becomes bdi_writeback_workfn() which runs off bdi_writeback->dwork. On each execution, it processes bdi->work_list and reschedules itself if there are more things to do. The function also handles low-mem condition, which used to be handled by the forker thread. If the function is running off a rescuer thread, it only writes out limited number of pages so that the rescuer can serve other bdis too. This preserves the flusher creation failure behavior of the forker thread. * INIT_LIST_HEAD(&bdi->bdi_list) is used to tell bdi_writeback_workfn() about on-going bdi unregistration so that it always drains work_list even if it's running off the rescuer. Note that the original code was broken in this regard. Under memory pressure, a bdi could finish unregistration with non-empty work_list. * The default bdi is no longer special. It now is treated the same as any other bdi and bdi_cap_flush_forker() is removed. * BDI_pending is no longer used. Removed. * Some tracepoints become non-applicable. The following TPs are removed - writeback_nothread, writeback_wake_thread, writeback_wake_forker_thread, writeback_thread_start, writeback_thread_stop. Everything, including devices coming and going away and rescuer operation under simulated memory pressure, seems to work fine in my test setup. Signed-off-by: Tejun Heo <tj@kernel.org> Reviewed-by: Jan Kara <jack@suse.cz> Cc: Jens Axboe <axboe@kernel.dk> Cc: Fengguang Wu <fengguang.wu@intel.com> Cc: Jeff Moyer <jmoyer@redhat.com>
2013-04-01 20:08:06 -06:00
if (!bdi_wq)
return -ENOMEM;
err = bdi_init(&default_backing_dev_info);
if (!err)
bdi_register(&default_backing_dev_info, NULL, "default");
err = bdi_init(&noop_backing_dev_info);
return err;
}
subsys_initcall(default_bdi_init);
writeback: switch to per-bdi threads for flushing data This gets rid of pdflush for bdi writeout and kupdated style cleaning. pdflush writeout suffers from lack of locality and also requires more threads to handle the same workload, since it has to work in a non-blocking fashion against each queue. This also introduces lumpy behaviour and potential request starvation, since pdflush can be starved for queue access if others are accessing it. A sample ffsb workload that does random writes to files is about 8% faster here on a simple SATA drive during the benchmark phase. File layout also seems a LOT more smooth in vmstat: r b swpd free buff cache si so bi bo in cs us sy id wa 0 1 0 608848 2652 375372 0 0 0 71024 604 24 1 10 48 42 0 1 0 549644 2712 433736 0 0 0 60692 505 27 1 8 48 44 1 0 0 476928 2784 505192 0 0 4 29540 553 24 0 9 53 37 0 1 0 457972 2808 524008 0 0 0 54876 331 16 0 4 38 58 0 1 0 366128 2928 614284 0 0 4 92168 710 58 0 13 53 34 0 1 0 295092 3000 684140 0 0 0 62924 572 23 0 9 53 37 0 1 0 236592 3064 741704 0 0 4 58256 523 17 0 8 48 44 0 1 0 165608 3132 811464 0 0 0 57460 560 21 0 8 54 38 0 1 0 102952 3200 873164 0 0 4 74748 540 29 1 10 48 41 0 1 0 48604 3252 926472 0 0 0 53248 469 29 0 7 47 45 where vanilla tends to fluctuate a lot in the creation phase: r b swpd free buff cache si so bi bo in cs us sy id wa 1 1 0 678716 5792 303380 0 0 0 74064 565 50 1 11 52 36 1 0 0 662488 5864 319396 0 0 4 352 302 329 0 2 47 51 0 1 0 599312 5924 381468 0 0 0 78164 516 55 0 9 51 40 0 1 0 519952 6008 459516 0 0 4 78156 622 56 1 11 52 37 1 1 0 436640 6092 541632 0 0 0 82244 622 54 0 11 48 41 0 1 0 436640 6092 541660 0 0 0 8 152 39 0 0 51 49 0 1 0 332224 6200 644252 0 0 4 102800 728 46 1 13 49 36 1 0 0 274492 6260 701056 0 0 4 12328 459 49 0 7 50 43 0 1 0 211220 6324 763356 0 0 0 106940 515 37 1 10 51 39 1 0 0 160412 6376 813468 0 0 0 8224 415 43 0 6 49 45 1 1 0 85980 6452 886556 0 0 4 113516 575 39 1 11 54 34 0 2 0 85968 6452 886620 0 0 0 1640 158 211 0 0 46 54 A 10 disk test with btrfs performs 26% faster with per-bdi flushing. A SSD based writeback test on XFS performs over 20% better as well, with the throughput being very stable around 1GB/sec, where pdflush only manages 750MB/sec and fluctuates wildly while doing so. Random buffered writes to many files behave a lot better as well, as does random mmap'ed writes. A separate thread is added to sync the super blocks. In the long term, adding sync_supers_bdi() functionality could get rid of this thread again. Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
2009-09-09 01:08:54 -06:00
int bdi_has_dirty_io(struct backing_dev_info *bdi)
{
return wb_has_dirty_io(&bdi->wb);
}
/*
* This function is used when the first inode for this bdi is marked dirty. It
* wakes-up the corresponding bdi thread which should then take care of the
* periodic background write-out of dirty inodes. Since the write-out would
* starts only 'dirty_writeback_interval' centisecs from now anyway, we just
* set up a timer which wakes the bdi thread up later.
*
* Note, we wouldn't bother setting up the timer, but this function is on the
* fast-path (used by '__mark_inode_dirty()'), so we save few context switches
* by delaying the wake-up.
backing_dev: fix hung task on sync bdi_wakeup_thread_delayed() used the mod_delayed_work() function to schedule work to writeback dirty inodes. The problem with this is that it can delay work that is scheduled for immediate execution, such as the work from sync_inodes_sb(). This can happen since mod_delayed_work() can now steal work from a work_queue. This fixes the problem by using queue_delayed_work() instead. This is a regression caused by commit 839a8e8660b6 ("writeback: replace custom worker pool implementation with unbound workqueue"). The reason that this causes a problem is that laptop-mode will change the delay, dirty_writeback_centisecs, to 60000 (10 minutes) by default. In the case that bdi_wakeup_thread_delayed() races with sync_inodes_sb(), sync will be stopped for 10 minutes and trigger a hung task. Even if dirty_writeback_centisecs is not long enough to cause a hung task, we still don't want to delay sync for that long. We fix the problem by using queue_delayed_work() when we want to schedule writeback sometime in future. This function doesn't change the timer if it is already armed. For the same reason, we also change bdi_writeback_workfn() to immediately queue the work again in the case that the work_list is not empty. The same problem can happen if the sync work is run on the rescue worker. [jack@suse.cz: update changelog, add comment, use bdi_wakeup_thread_delayed()] Signed-off-by: Derek Basehore <dbasehore@chromium.org> Reviewed-by: Jan Kara <jack@suse.cz> Cc: Alexander Viro <viro@zento.linux.org.uk> Reviewed-by: Tejun Heo <tj@kernel.org> Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org> Cc: "Darrick J. Wong" <darrick.wong@oracle.com> Cc: Derek Basehore <dbasehore@chromium.org> Cc: Kees Cook <keescook@chromium.org> Cc: Benson Leung <bleung@chromium.org> Cc: Sonny Rao <sonnyrao@chromium.org> Cc: Luigi Semenzato <semenzato@chromium.org> Cc: Jens Axboe <axboe@kernel.dk> Cc: Dave Chinner <david@fromorbit.com> Cc: <stable@vger.kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-04-03 15:46:22 -06:00
*
* We have to be careful not to postpone flush work if it is scheduled for
* earlier. Thus we use queue_delayed_work().
*/
void bdi_wakeup_thread_delayed(struct backing_dev_info *bdi)
{
unsigned long timeout;
timeout = msecs_to_jiffies(dirty_writeback_interval * 10);
spin_lock_bh(&bdi->wb_lock);
if (test_bit(BDI_registered, &bdi->state))
queue_delayed_work(bdi_wq, &bdi->wb.dwork, timeout);
spin_unlock_bh(&bdi->wb_lock);
writeback: switch to per-bdi threads for flushing data This gets rid of pdflush for bdi writeout and kupdated style cleaning. pdflush writeout suffers from lack of locality and also requires more threads to handle the same workload, since it has to work in a non-blocking fashion against each queue. This also introduces lumpy behaviour and potential request starvation, since pdflush can be starved for queue access if others are accessing it. A sample ffsb workload that does random writes to files is about 8% faster here on a simple SATA drive during the benchmark phase. File layout also seems a LOT more smooth in vmstat: r b swpd free buff cache si so bi bo in cs us sy id wa 0 1 0 608848 2652 375372 0 0 0 71024 604 24 1 10 48 42 0 1 0 549644 2712 433736 0 0 0 60692 505 27 1 8 48 44 1 0 0 476928 2784 505192 0 0 4 29540 553 24 0 9 53 37 0 1 0 457972 2808 524008 0 0 0 54876 331 16 0 4 38 58 0 1 0 366128 2928 614284 0 0 4 92168 710 58 0 13 53 34 0 1 0 295092 3000 684140 0 0 0 62924 572 23 0 9 53 37 0 1 0 236592 3064 741704 0 0 4 58256 523 17 0 8 48 44 0 1 0 165608 3132 811464 0 0 0 57460 560 21 0 8 54 38 0 1 0 102952 3200 873164 0 0 4 74748 540 29 1 10 48 41 0 1 0 48604 3252 926472 0 0 0 53248 469 29 0 7 47 45 where vanilla tends to fluctuate a lot in the creation phase: r b swpd free buff cache si so bi bo in cs us sy id wa 1 1 0 678716 5792 303380 0 0 0 74064 565 50 1 11 52 36 1 0 0 662488 5864 319396 0 0 4 352 302 329 0 2 47 51 0 1 0 599312 5924 381468 0 0 0 78164 516 55 0 9 51 40 0 1 0 519952 6008 459516 0 0 4 78156 622 56 1 11 52 37 1 1 0 436640 6092 541632 0 0 0 82244 622 54 0 11 48 41 0 1 0 436640 6092 541660 0 0 0 8 152 39 0 0 51 49 0 1 0 332224 6200 644252 0 0 4 102800 728 46 1 13 49 36 1 0 0 274492 6260 701056 0 0 4 12328 459 49 0 7 50 43 0 1 0 211220 6324 763356 0 0 0 106940 515 37 1 10 51 39 1 0 0 160412 6376 813468 0 0 0 8224 415 43 0 6 49 45 1 1 0 85980 6452 886556 0 0 4 113516 575 39 1 11 54 34 0 2 0 85968 6452 886620 0 0 0 1640 158 211 0 0 46 54 A 10 disk test with btrfs performs 26% faster with per-bdi flushing. A SSD based writeback test on XFS performs over 20% better as well, with the throughput being very stable around 1GB/sec, where pdflush only manages 750MB/sec and fluctuates wildly while doing so. Random buffered writes to many files behave a lot better as well, as does random mmap'ed writes. A separate thread is added to sync the super blocks. In the long term, adding sync_supers_bdi() functionality could get rid of this thread again. Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
2009-09-09 01:08:54 -06:00
}
/*
* Remove bdi from bdi_list, and ensure that it is no longer visible
*/
static void bdi_remove_from_list(struct backing_dev_info *bdi)
{
spin_lock_bh(&bdi_lock);
list_del_rcu(&bdi->bdi_list);
spin_unlock_bh(&bdi_lock);
synchronize_rcu_expedited();
}
int bdi_register(struct backing_dev_info *bdi, struct device *parent,
const char *fmt, ...)
{
va_list args;
struct device *dev;
if (bdi->dev) /* The driver needs to use separate queues per device */
return 0;
va_start(args, fmt);
dev = device_create_vargs(bdi_class, parent, MKDEV(0, 0), bdi, fmt, args);
va_end(args);
if (IS_ERR(dev))
return PTR_ERR(dev);
bdi->dev = dev;
writeback: switch to per-bdi threads for flushing data This gets rid of pdflush for bdi writeout and kupdated style cleaning. pdflush writeout suffers from lack of locality and also requires more threads to handle the same workload, since it has to work in a non-blocking fashion against each queue. This also introduces lumpy behaviour and potential request starvation, since pdflush can be starved for queue access if others are accessing it. A sample ffsb workload that does random writes to files is about 8% faster here on a simple SATA drive during the benchmark phase. File layout also seems a LOT more smooth in vmstat: r b swpd free buff cache si so bi bo in cs us sy id wa 0 1 0 608848 2652 375372 0 0 0 71024 604 24 1 10 48 42 0 1 0 549644 2712 433736 0 0 0 60692 505 27 1 8 48 44 1 0 0 476928 2784 505192 0 0 4 29540 553 24 0 9 53 37 0 1 0 457972 2808 524008 0 0 0 54876 331 16 0 4 38 58 0 1 0 366128 2928 614284 0 0 4 92168 710 58 0 13 53 34 0 1 0 295092 3000 684140 0 0 0 62924 572 23 0 9 53 37 0 1 0 236592 3064 741704 0 0 4 58256 523 17 0 8 48 44 0 1 0 165608 3132 811464 0 0 0 57460 560 21 0 8 54 38 0 1 0 102952 3200 873164 0 0 4 74748 540 29 1 10 48 41 0 1 0 48604 3252 926472 0 0 0 53248 469 29 0 7 47 45 where vanilla tends to fluctuate a lot in the creation phase: r b swpd free buff cache si so bi bo in cs us sy id wa 1 1 0 678716 5792 303380 0 0 0 74064 565 50 1 11 52 36 1 0 0 662488 5864 319396 0 0 4 352 302 329 0 2 47 51 0 1 0 599312 5924 381468 0 0 0 78164 516 55 0 9 51 40 0 1 0 519952 6008 459516 0 0 4 78156 622 56 1 11 52 37 1 1 0 436640 6092 541632 0 0 0 82244 622 54 0 11 48 41 0 1 0 436640 6092 541660 0 0 0 8 152 39 0 0 51 49 0 1 0 332224 6200 644252 0 0 4 102800 728 46 1 13 49 36 1 0 0 274492 6260 701056 0 0 4 12328 459 49 0 7 50 43 0 1 0 211220 6324 763356 0 0 0 106940 515 37 1 10 51 39 1 0 0 160412 6376 813468 0 0 0 8224 415 43 0 6 49 45 1 1 0 85980 6452 886556 0 0 4 113516 575 39 1 11 54 34 0 2 0 85968 6452 886620 0 0 0 1640 158 211 0 0 46 54 A 10 disk test with btrfs performs 26% faster with per-bdi flushing. A SSD based writeback test on XFS performs over 20% better as well, with the throughput being very stable around 1GB/sec, where pdflush only manages 750MB/sec and fluctuates wildly while doing so. Random buffered writes to many files behave a lot better as well, as does random mmap'ed writes. A separate thread is added to sync the super blocks. In the long term, adding sync_supers_bdi() functionality could get rid of this thread again. Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
2009-09-09 01:08:54 -06:00
bdi_debug_register(bdi, dev_name(dev));
set_bit(BDI_registered, &bdi->state);
spin_lock_bh(&bdi_lock);
list_add_tail_rcu(&bdi->bdi_list, &bdi_list);
spin_unlock_bh(&bdi_lock);
trace_writeback_bdi_register(bdi);
return 0;
}
EXPORT_SYMBOL(bdi_register);
int bdi_register_dev(struct backing_dev_info *bdi, dev_t dev)
{
return bdi_register(bdi, NULL, "%u:%u", MAJOR(dev), MINOR(dev));
}
EXPORT_SYMBOL(bdi_register_dev);
writeback: switch to per-bdi threads for flushing data This gets rid of pdflush for bdi writeout and kupdated style cleaning. pdflush writeout suffers from lack of locality and also requires more threads to handle the same workload, since it has to work in a non-blocking fashion against each queue. This also introduces lumpy behaviour and potential request starvation, since pdflush can be starved for queue access if others are accessing it. A sample ffsb workload that does random writes to files is about 8% faster here on a simple SATA drive during the benchmark phase. File layout also seems a LOT more smooth in vmstat: r b swpd free buff cache si so bi bo in cs us sy id wa 0 1 0 608848 2652 375372 0 0 0 71024 604 24 1 10 48 42 0 1 0 549644 2712 433736 0 0 0 60692 505 27 1 8 48 44 1 0 0 476928 2784 505192 0 0 4 29540 553 24 0 9 53 37 0 1 0 457972 2808 524008 0 0 0 54876 331 16 0 4 38 58 0 1 0 366128 2928 614284 0 0 4 92168 710 58 0 13 53 34 0 1 0 295092 3000 684140 0 0 0 62924 572 23 0 9 53 37 0 1 0 236592 3064 741704 0 0 4 58256 523 17 0 8 48 44 0 1 0 165608 3132 811464 0 0 0 57460 560 21 0 8 54 38 0 1 0 102952 3200 873164 0 0 4 74748 540 29 1 10 48 41 0 1 0 48604 3252 926472 0 0 0 53248 469 29 0 7 47 45 where vanilla tends to fluctuate a lot in the creation phase: r b swpd free buff cache si so bi bo in cs us sy id wa 1 1 0 678716 5792 303380 0 0 0 74064 565 50 1 11 52 36 1 0 0 662488 5864 319396 0 0 4 352 302 329 0 2 47 51 0 1 0 599312 5924 381468 0 0 0 78164 516 55 0 9 51 40 0 1 0 519952 6008 459516 0 0 4 78156 622 56 1 11 52 37 1 1 0 436640 6092 541632 0 0 0 82244 622 54 0 11 48 41 0 1 0 436640 6092 541660 0 0 0 8 152 39 0 0 51 49 0 1 0 332224 6200 644252 0 0 4 102800 728 46 1 13 49 36 1 0 0 274492 6260 701056 0 0 4 12328 459 49 0 7 50 43 0 1 0 211220 6324 763356 0 0 0 106940 515 37 1 10 51 39 1 0 0 160412 6376 813468 0 0 0 8224 415 43 0 6 49 45 1 1 0 85980 6452 886556 0 0 4 113516 575 39 1 11 54 34 0 2 0 85968 6452 886620 0 0 0 1640 158 211 0 0 46 54 A 10 disk test with btrfs performs 26% faster with per-bdi flushing. A SSD based writeback test on XFS performs over 20% better as well, with the throughput being very stable around 1GB/sec, where pdflush only manages 750MB/sec and fluctuates wildly while doing so. Random buffered writes to many files behave a lot better as well, as does random mmap'ed writes. A separate thread is added to sync the super blocks. In the long term, adding sync_supers_bdi() functionality could get rid of this thread again. Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
2009-09-09 01:08:54 -06:00
/*
* Remove bdi from the global list and shutdown any threads we have running
*/
static void bdi_wb_shutdown(struct backing_dev_info *bdi)
{
writeback: switch to per-bdi threads for flushing data This gets rid of pdflush for bdi writeout and kupdated style cleaning. pdflush writeout suffers from lack of locality and also requires more threads to handle the same workload, since it has to work in a non-blocking fashion against each queue. This also introduces lumpy behaviour and potential request starvation, since pdflush can be starved for queue access if others are accessing it. A sample ffsb workload that does random writes to files is about 8% faster here on a simple SATA drive during the benchmark phase. File layout also seems a LOT more smooth in vmstat: r b swpd free buff cache si so bi bo in cs us sy id wa 0 1 0 608848 2652 375372 0 0 0 71024 604 24 1 10 48 42 0 1 0 549644 2712 433736 0 0 0 60692 505 27 1 8 48 44 1 0 0 476928 2784 505192 0 0 4 29540 553 24 0 9 53 37 0 1 0 457972 2808 524008 0 0 0 54876 331 16 0 4 38 58 0 1 0 366128 2928 614284 0 0 4 92168 710 58 0 13 53 34 0 1 0 295092 3000 684140 0 0 0 62924 572 23 0 9 53 37 0 1 0 236592 3064 741704 0 0 4 58256 523 17 0 8 48 44 0 1 0 165608 3132 811464 0 0 0 57460 560 21 0 8 54 38 0 1 0 102952 3200 873164 0 0 4 74748 540 29 1 10 48 41 0 1 0 48604 3252 926472 0 0 0 53248 469 29 0 7 47 45 where vanilla tends to fluctuate a lot in the creation phase: r b swpd free buff cache si so bi bo in cs us sy id wa 1 1 0 678716 5792 303380 0 0 0 74064 565 50 1 11 52 36 1 0 0 662488 5864 319396 0 0 4 352 302 329 0 2 47 51 0 1 0 599312 5924 381468 0 0 0 78164 516 55 0 9 51 40 0 1 0 519952 6008 459516 0 0 4 78156 622 56 1 11 52 37 1 1 0 436640 6092 541632 0 0 0 82244 622 54 0 11 48 41 0 1 0 436640 6092 541660 0 0 0 8 152 39 0 0 51 49 0 1 0 332224 6200 644252 0 0 4 102800 728 46 1 13 49 36 1 0 0 274492 6260 701056 0 0 4 12328 459 49 0 7 50 43 0 1 0 211220 6324 763356 0 0 0 106940 515 37 1 10 51 39 1 0 0 160412 6376 813468 0 0 0 8224 415 43 0 6 49 45 1 1 0 85980 6452 886556 0 0 4 113516 575 39 1 11 54 34 0 2 0 85968 6452 886620 0 0 0 1640 158 211 0 0 46 54 A 10 disk test with btrfs performs 26% faster with per-bdi flushing. A SSD based writeback test on XFS performs over 20% better as well, with the throughput being very stable around 1GB/sec, where pdflush only manages 750MB/sec and fluctuates wildly while doing so. Random buffered writes to many files behave a lot better as well, as does random mmap'ed writes. A separate thread is added to sync the super blocks. In the long term, adding sync_supers_bdi() functionality could get rid of this thread again. Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
2009-09-09 01:08:54 -06:00
if (!bdi_cap_writeback_dirty(bdi))
return;
/*
writeback: move bdi threads exiting logic to the forker thread Currently, bdi threads can decide to exit if there were no useful activities for 5 minutes. However, this causes nasty races: we can easily oops in the 'bdi_queue_work()' if the bdi thread decides to exit while we are waking it up. And even if we do not oops, but the bdi tread exits immediately after we wake it up, we'd lose the wake-up event and have an unnecessary delay (up to 5 secs) in the bdi work processing. This patch makes the forker thread to be the central place which not only creates bdi threads, but also kills them if they were inactive long enough. This better design-wise. Another reason why this change was done is to prepare for the further changes which will prevent the bdi threads from waking up every 5 sec and wasting power. Indeed, when the task does not wake up periodically anymore, it won't be able to exit either. This patch also moves the the 'wake_up_bit()' call from the bdi thread to the forker thread as well. So now the forker thread sets the BDI_pending bit, then forks the task or kills it, then clears the bit and wakes up the waiting process. The only process which may wain on the bit is 'bdi_wb_shutdown()'. This function was changed as well - now it first removes the bdi from the 'bdi_list', then waits on the 'BDI_pending' bit. Once it wakes up, it is guaranteed that the forker thread won't race with it, because the bdi is not visible. Note, the forker thread sets the 'BDI_pending' bit under the 'bdi->wb_lock' which is essential for proper serialization. And additionally, when we change 'bdi->wb.task', we now take the 'bdi->work_lock', to make sure that we do not lose wake-ups which we otherwise would when raced with, say, 'bdi_queue_work()'. Signed-off-by: Artem Bityutskiy <Artem.Bityutskiy@nokia.com> Reviewed-by: Christoph Hellwig <hch@lst.de> Signed-off-by: Jens Axboe <jaxboe@fusionio.com>
2010-07-25 05:29:20 -06:00
* Make sure nobody finds us on the bdi_list anymore
writeback: switch to per-bdi threads for flushing data This gets rid of pdflush for bdi writeout and kupdated style cleaning. pdflush writeout suffers from lack of locality and also requires more threads to handle the same workload, since it has to work in a non-blocking fashion against each queue. This also introduces lumpy behaviour and potential request starvation, since pdflush can be starved for queue access if others are accessing it. A sample ffsb workload that does random writes to files is about 8% faster here on a simple SATA drive during the benchmark phase. File layout also seems a LOT more smooth in vmstat: r b swpd free buff cache si so bi bo in cs us sy id wa 0 1 0 608848 2652 375372 0 0 0 71024 604 24 1 10 48 42 0 1 0 549644 2712 433736 0 0 0 60692 505 27 1 8 48 44 1 0 0 476928 2784 505192 0 0 4 29540 553 24 0 9 53 37 0 1 0 457972 2808 524008 0 0 0 54876 331 16 0 4 38 58 0 1 0 366128 2928 614284 0 0 4 92168 710 58 0 13 53 34 0 1 0 295092 3000 684140 0 0 0 62924 572 23 0 9 53 37 0 1 0 236592 3064 741704 0 0 4 58256 523 17 0 8 48 44 0 1 0 165608 3132 811464 0 0 0 57460 560 21 0 8 54 38 0 1 0 102952 3200 873164 0 0 4 74748 540 29 1 10 48 41 0 1 0 48604 3252 926472 0 0 0 53248 469 29 0 7 47 45 where vanilla tends to fluctuate a lot in the creation phase: r b swpd free buff cache si so bi bo in cs us sy id wa 1 1 0 678716 5792 303380 0 0 0 74064 565 50 1 11 52 36 1 0 0 662488 5864 319396 0 0 4 352 302 329 0 2 47 51 0 1 0 599312 5924 381468 0 0 0 78164 516 55 0 9 51 40 0 1 0 519952 6008 459516 0 0 4 78156 622 56 1 11 52 37 1 1 0 436640 6092 541632 0 0 0 82244 622 54 0 11 48 41 0 1 0 436640 6092 541660 0 0 0 8 152 39 0 0 51 49 0 1 0 332224 6200 644252 0 0 4 102800 728 46 1 13 49 36 1 0 0 274492 6260 701056 0 0 4 12328 459 49 0 7 50 43 0 1 0 211220 6324 763356 0 0 0 106940 515 37 1 10 51 39 1 0 0 160412 6376 813468 0 0 0 8224 415 43 0 6 49 45 1 1 0 85980 6452 886556 0 0 4 113516 575 39 1 11 54 34 0 2 0 85968 6452 886620 0 0 0 1640 158 211 0 0 46 54 A 10 disk test with btrfs performs 26% faster with per-bdi flushing. A SSD based writeback test on XFS performs over 20% better as well, with the throughput being very stable around 1GB/sec, where pdflush only manages 750MB/sec and fluctuates wildly while doing so. Random buffered writes to many files behave a lot better as well, as does random mmap'ed writes. A separate thread is added to sync the super blocks. In the long term, adding sync_supers_bdi() functionality could get rid of this thread again. Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
2009-09-09 01:08:54 -06:00
*/
writeback: move bdi threads exiting logic to the forker thread Currently, bdi threads can decide to exit if there were no useful activities for 5 minutes. However, this causes nasty races: we can easily oops in the 'bdi_queue_work()' if the bdi thread decides to exit while we are waking it up. And even if we do not oops, but the bdi tread exits immediately after we wake it up, we'd lose the wake-up event and have an unnecessary delay (up to 5 secs) in the bdi work processing. This patch makes the forker thread to be the central place which not only creates bdi threads, but also kills them if they were inactive long enough. This better design-wise. Another reason why this change was done is to prepare for the further changes which will prevent the bdi threads from waking up every 5 sec and wasting power. Indeed, when the task does not wake up periodically anymore, it won't be able to exit either. This patch also moves the the 'wake_up_bit()' call from the bdi thread to the forker thread as well. So now the forker thread sets the BDI_pending bit, then forks the task or kills it, then clears the bit and wakes up the waiting process. The only process which may wain on the bit is 'bdi_wb_shutdown()'. This function was changed as well - now it first removes the bdi from the 'bdi_list', then waits on the 'BDI_pending' bit. Once it wakes up, it is guaranteed that the forker thread won't race with it, because the bdi is not visible. Note, the forker thread sets the 'BDI_pending' bit under the 'bdi->wb_lock' which is essential for proper serialization. And additionally, when we change 'bdi->wb.task', we now take the 'bdi->work_lock', to make sure that we do not lose wake-ups which we otherwise would when raced with, say, 'bdi_queue_work()'. Signed-off-by: Artem Bityutskiy <Artem.Bityutskiy@nokia.com> Reviewed-by: Christoph Hellwig <hch@lst.de> Signed-off-by: Jens Axboe <jaxboe@fusionio.com>
2010-07-25 05:29:20 -06:00
bdi_remove_from_list(bdi);
writeback: switch to per-bdi threads for flushing data This gets rid of pdflush for bdi writeout and kupdated style cleaning. pdflush writeout suffers from lack of locality and also requires more threads to handle the same workload, since it has to work in a non-blocking fashion against each queue. This also introduces lumpy behaviour and potential request starvation, since pdflush can be starved for queue access if others are accessing it. A sample ffsb workload that does random writes to files is about 8% faster here on a simple SATA drive during the benchmark phase. File layout also seems a LOT more smooth in vmstat: r b swpd free buff cache si so bi bo in cs us sy id wa 0 1 0 608848 2652 375372 0 0 0 71024 604 24 1 10 48 42 0 1 0 549644 2712 433736 0 0 0 60692 505 27 1 8 48 44 1 0 0 476928 2784 505192 0 0 4 29540 553 24 0 9 53 37 0 1 0 457972 2808 524008 0 0 0 54876 331 16 0 4 38 58 0 1 0 366128 2928 614284 0 0 4 92168 710 58 0 13 53 34 0 1 0 295092 3000 684140 0 0 0 62924 572 23 0 9 53 37 0 1 0 236592 3064 741704 0 0 4 58256 523 17 0 8 48 44 0 1 0 165608 3132 811464 0 0 0 57460 560 21 0 8 54 38 0 1 0 102952 3200 873164 0 0 4 74748 540 29 1 10 48 41 0 1 0 48604 3252 926472 0 0 0 53248 469 29 0 7 47 45 where vanilla tends to fluctuate a lot in the creation phase: r b swpd free buff cache si so bi bo in cs us sy id wa 1 1 0 678716 5792 303380 0 0 0 74064 565 50 1 11 52 36 1 0 0 662488 5864 319396 0 0 4 352 302 329 0 2 47 51 0 1 0 599312 5924 381468 0 0 0 78164 516 55 0 9 51 40 0 1 0 519952 6008 459516 0 0 4 78156 622 56 1 11 52 37 1 1 0 436640 6092 541632 0 0 0 82244 622 54 0 11 48 41 0 1 0 436640 6092 541660 0 0 0 8 152 39 0 0 51 49 0 1 0 332224 6200 644252 0 0 4 102800 728 46 1 13 49 36 1 0 0 274492 6260 701056 0 0 4 12328 459 49 0 7 50 43 0 1 0 211220 6324 763356 0 0 0 106940 515 37 1 10 51 39 1 0 0 160412 6376 813468 0 0 0 8224 415 43 0 6 49 45 1 1 0 85980 6452 886556 0 0 4 113516 575 39 1 11 54 34 0 2 0 85968 6452 886620 0 0 0 1640 158 211 0 0 46 54 A 10 disk test with btrfs performs 26% faster with per-bdi flushing. A SSD based writeback test on XFS performs over 20% better as well, with the throughput being very stable around 1GB/sec, where pdflush only manages 750MB/sec and fluctuates wildly while doing so. Random buffered writes to many files behave a lot better as well, as does random mmap'ed writes. A separate thread is added to sync the super blocks. In the long term, adding sync_supers_bdi() functionality could get rid of this thread again. Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
2009-09-09 01:08:54 -06:00
/* Make sure nobody queues further work */
spin_lock_bh(&bdi->wb_lock);
clear_bit(BDI_registered, &bdi->state);
spin_unlock_bh(&bdi->wb_lock);
writeback: switch to per-bdi threads for flushing data This gets rid of pdflush for bdi writeout and kupdated style cleaning. pdflush writeout suffers from lack of locality and also requires more threads to handle the same workload, since it has to work in a non-blocking fashion against each queue. This also introduces lumpy behaviour and potential request starvation, since pdflush can be starved for queue access if others are accessing it. A sample ffsb workload that does random writes to files is about 8% faster here on a simple SATA drive during the benchmark phase. File layout also seems a LOT more smooth in vmstat: r b swpd free buff cache si so bi bo in cs us sy id wa 0 1 0 608848 2652 375372 0 0 0 71024 604 24 1 10 48 42 0 1 0 549644 2712 433736 0 0 0 60692 505 27 1 8 48 44 1 0 0 476928 2784 505192 0 0 4 29540 553 24 0 9 53 37 0 1 0 457972 2808 524008 0 0 0 54876 331 16 0 4 38 58 0 1 0 366128 2928 614284 0 0 4 92168 710 58 0 13 53 34 0 1 0 295092 3000 684140 0 0 0 62924 572 23 0 9 53 37 0 1 0 236592 3064 741704 0 0 4 58256 523 17 0 8 48 44 0 1 0 165608 3132 811464 0 0 0 57460 560 21 0 8 54 38 0 1 0 102952 3200 873164 0 0 4 74748 540 29 1 10 48 41 0 1 0 48604 3252 926472 0 0 0 53248 469 29 0 7 47 45 where vanilla tends to fluctuate a lot in the creation phase: r b swpd free buff cache si so bi bo in cs us sy id wa 1 1 0 678716 5792 303380 0 0 0 74064 565 50 1 11 52 36 1 0 0 662488 5864 319396 0 0 4 352 302 329 0 2 47 51 0 1 0 599312 5924 381468 0 0 0 78164 516 55 0 9 51 40 0 1 0 519952 6008 459516 0 0 4 78156 622 56 1 11 52 37 1 1 0 436640 6092 541632 0 0 0 82244 622 54 0 11 48 41 0 1 0 436640 6092 541660 0 0 0 8 152 39 0 0 51 49 0 1 0 332224 6200 644252 0 0 4 102800 728 46 1 13 49 36 1 0 0 274492 6260 701056 0 0 4 12328 459 49 0 7 50 43 0 1 0 211220 6324 763356 0 0 0 106940 515 37 1 10 51 39 1 0 0 160412 6376 813468 0 0 0 8224 415 43 0 6 49 45 1 1 0 85980 6452 886556 0 0 4 113516 575 39 1 11 54 34 0 2 0 85968 6452 886620 0 0 0 1640 158 211 0 0 46 54 A 10 disk test with btrfs performs 26% faster with per-bdi flushing. A SSD based writeback test on XFS performs over 20% better as well, with the throughput being very stable around 1GB/sec, where pdflush only manages 750MB/sec and fluctuates wildly while doing so. Random buffered writes to many files behave a lot better as well, as does random mmap'ed writes. A separate thread is added to sync the super blocks. In the long term, adding sync_supers_bdi() functionality could get rid of this thread again. Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
2009-09-09 01:08:54 -06:00
/*
writeback: replace custom worker pool implementation with unbound workqueue Writeback implements its own worker pool - each bdi can be associated with a worker thread which is created and destroyed dynamically. The worker thread for the default bdi is always present and serves as the "forker" thread which forks off worker threads for other bdis. there's no reason for writeback to implement its own worker pool when using unbound workqueue instead is much simpler and more efficient. This patch replaces custom worker pool implementation in writeback with an unbound workqueue. The conversion isn't too complicated but the followings are worth mentioning. * bdi_writeback->last_active, task and wakeup_timer are removed. delayed_work ->dwork is added instead. Explicit timer handling is no longer necessary. Everything works by either queueing / modding / flushing / canceling the delayed_work item. * bdi_writeback_thread() becomes bdi_writeback_workfn() which runs off bdi_writeback->dwork. On each execution, it processes bdi->work_list and reschedules itself if there are more things to do. The function also handles low-mem condition, which used to be handled by the forker thread. If the function is running off a rescuer thread, it only writes out limited number of pages so that the rescuer can serve other bdis too. This preserves the flusher creation failure behavior of the forker thread. * INIT_LIST_HEAD(&bdi->bdi_list) is used to tell bdi_writeback_workfn() about on-going bdi unregistration so that it always drains work_list even if it's running off the rescuer. Note that the original code was broken in this regard. Under memory pressure, a bdi could finish unregistration with non-empty work_list. * The default bdi is no longer special. It now is treated the same as any other bdi and bdi_cap_flush_forker() is removed. * BDI_pending is no longer used. Removed. * Some tracepoints become non-applicable. The following TPs are removed - writeback_nothread, writeback_wake_thread, writeback_wake_forker_thread, writeback_thread_start, writeback_thread_stop. Everything, including devices coming and going away and rescuer operation under simulated memory pressure, seems to work fine in my test setup. Signed-off-by: Tejun Heo <tj@kernel.org> Reviewed-by: Jan Kara <jack@suse.cz> Cc: Jens Axboe <axboe@kernel.dk> Cc: Fengguang Wu <fengguang.wu@intel.com> Cc: Jeff Moyer <jmoyer@redhat.com>
2013-04-01 20:08:06 -06:00
* Drain work list and shutdown the delayed_work. At this point,
* @bdi->bdi_list is empty telling bdi_Writeback_workfn() that @bdi
* is dying and its work_list needs to be drained no matter what.
writeback: switch to per-bdi threads for flushing data This gets rid of pdflush for bdi writeout and kupdated style cleaning. pdflush writeout suffers from lack of locality and also requires more threads to handle the same workload, since it has to work in a non-blocking fashion against each queue. This also introduces lumpy behaviour and potential request starvation, since pdflush can be starved for queue access if others are accessing it. A sample ffsb workload that does random writes to files is about 8% faster here on a simple SATA drive during the benchmark phase. File layout also seems a LOT more smooth in vmstat: r b swpd free buff cache si so bi bo in cs us sy id wa 0 1 0 608848 2652 375372 0 0 0 71024 604 24 1 10 48 42 0 1 0 549644 2712 433736 0 0 0 60692 505 27 1 8 48 44 1 0 0 476928 2784 505192 0 0 4 29540 553 24 0 9 53 37 0 1 0 457972 2808 524008 0 0 0 54876 331 16 0 4 38 58 0 1 0 366128 2928 614284 0 0 4 92168 710 58 0 13 53 34 0 1 0 295092 3000 684140 0 0 0 62924 572 23 0 9 53 37 0 1 0 236592 3064 741704 0 0 4 58256 523 17 0 8 48 44 0 1 0 165608 3132 811464 0 0 0 57460 560 21 0 8 54 38 0 1 0 102952 3200 873164 0 0 4 74748 540 29 1 10 48 41 0 1 0 48604 3252 926472 0 0 0 53248 469 29 0 7 47 45 where vanilla tends to fluctuate a lot in the creation phase: r b swpd free buff cache si so bi bo in cs us sy id wa 1 1 0 678716 5792 303380 0 0 0 74064 565 50 1 11 52 36 1 0 0 662488 5864 319396 0 0 4 352 302 329 0 2 47 51 0 1 0 599312 5924 381468 0 0 0 78164 516 55 0 9 51 40 0 1 0 519952 6008 459516 0 0 4 78156 622 56 1 11 52 37 1 1 0 436640 6092 541632 0 0 0 82244 622 54 0 11 48 41 0 1 0 436640 6092 541660 0 0 0 8 152 39 0 0 51 49 0 1 0 332224 6200 644252 0 0 4 102800 728 46 1 13 49 36 1 0 0 274492 6260 701056 0 0 4 12328 459 49 0 7 50 43 0 1 0 211220 6324 763356 0 0 0 106940 515 37 1 10 51 39 1 0 0 160412 6376 813468 0 0 0 8224 415 43 0 6 49 45 1 1 0 85980 6452 886556 0 0 4 113516 575 39 1 11 54 34 0 2 0 85968 6452 886620 0 0 0 1640 158 211 0 0 46 54 A 10 disk test with btrfs performs 26% faster with per-bdi flushing. A SSD based writeback test on XFS performs over 20% better as well, with the throughput being very stable around 1GB/sec, where pdflush only manages 750MB/sec and fluctuates wildly while doing so. Random buffered writes to many files behave a lot better as well, as does random mmap'ed writes. A separate thread is added to sync the super blocks. In the long term, adding sync_supers_bdi() functionality could get rid of this thread again. Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
2009-09-09 01:08:54 -06:00
*/
writeback: replace custom worker pool implementation with unbound workqueue Writeback implements its own worker pool - each bdi can be associated with a worker thread which is created and destroyed dynamically. The worker thread for the default bdi is always present and serves as the "forker" thread which forks off worker threads for other bdis. there's no reason for writeback to implement its own worker pool when using unbound workqueue instead is much simpler and more efficient. This patch replaces custom worker pool implementation in writeback with an unbound workqueue. The conversion isn't too complicated but the followings are worth mentioning. * bdi_writeback->last_active, task and wakeup_timer are removed. delayed_work ->dwork is added instead. Explicit timer handling is no longer necessary. Everything works by either queueing / modding / flushing / canceling the delayed_work item. * bdi_writeback_thread() becomes bdi_writeback_workfn() which runs off bdi_writeback->dwork. On each execution, it processes bdi->work_list and reschedules itself if there are more things to do. The function also handles low-mem condition, which used to be handled by the forker thread. If the function is running off a rescuer thread, it only writes out limited number of pages so that the rescuer can serve other bdis too. This preserves the flusher creation failure behavior of the forker thread. * INIT_LIST_HEAD(&bdi->bdi_list) is used to tell bdi_writeback_workfn() about on-going bdi unregistration so that it always drains work_list even if it's running off the rescuer. Note that the original code was broken in this regard. Under memory pressure, a bdi could finish unregistration with non-empty work_list. * The default bdi is no longer special. It now is treated the same as any other bdi and bdi_cap_flush_forker() is removed. * BDI_pending is no longer used. Removed. * Some tracepoints become non-applicable. The following TPs are removed - writeback_nothread, writeback_wake_thread, writeback_wake_forker_thread, writeback_thread_start, writeback_thread_stop. Everything, including devices coming and going away and rescuer operation under simulated memory pressure, seems to work fine in my test setup. Signed-off-by: Tejun Heo <tj@kernel.org> Reviewed-by: Jan Kara <jack@suse.cz> Cc: Jens Axboe <axboe@kernel.dk> Cc: Fengguang Wu <fengguang.wu@intel.com> Cc: Jeff Moyer <jmoyer@redhat.com>
2013-04-01 20:08:06 -06:00
mod_delayed_work(bdi_wq, &bdi->wb.dwork, 0);
flush_delayed_work(&bdi->wb.dwork);
WARN_ON(!list_empty(&bdi->work_list));
WARN_ON(delayed_work_pending(&bdi->wb.dwork));
}
/*
* This bdi is going away now, make sure that no super_blocks point to it
*/
static void bdi_prune_sb(struct backing_dev_info *bdi)
{
struct super_block *sb;
spin_lock(&sb_lock);
list_for_each_entry(sb, &super_blocks, s_list) {
if (sb->s_bdi == bdi)
sb->s_bdi = &default_backing_dev_info;
}
spin_unlock(&sb_lock);
}
void bdi_unregister(struct backing_dev_info *bdi)
{
if (bdi->dev) {
bdi_set_min_ratio(bdi, 0);
trace_writeback_bdi_unregister(bdi);
bdi_prune_sb(bdi);
writeback: replace custom worker pool implementation with unbound workqueue Writeback implements its own worker pool - each bdi can be associated with a worker thread which is created and destroyed dynamically. The worker thread for the default bdi is always present and serves as the "forker" thread which forks off worker threads for other bdis. there's no reason for writeback to implement its own worker pool when using unbound workqueue instead is much simpler and more efficient. This patch replaces custom worker pool implementation in writeback with an unbound workqueue. The conversion isn't too complicated but the followings are worth mentioning. * bdi_writeback->last_active, task and wakeup_timer are removed. delayed_work ->dwork is added instead. Explicit timer handling is no longer necessary. Everything works by either queueing / modding / flushing / canceling the delayed_work item. * bdi_writeback_thread() becomes bdi_writeback_workfn() which runs off bdi_writeback->dwork. On each execution, it processes bdi->work_list and reschedules itself if there are more things to do. The function also handles low-mem condition, which used to be handled by the forker thread. If the function is running off a rescuer thread, it only writes out limited number of pages so that the rescuer can serve other bdis too. This preserves the flusher creation failure behavior of the forker thread. * INIT_LIST_HEAD(&bdi->bdi_list) is used to tell bdi_writeback_workfn() about on-going bdi unregistration so that it always drains work_list even if it's running off the rescuer. Note that the original code was broken in this regard. Under memory pressure, a bdi could finish unregistration with non-empty work_list. * The default bdi is no longer special. It now is treated the same as any other bdi and bdi_cap_flush_forker() is removed. * BDI_pending is no longer used. Removed. * Some tracepoints become non-applicable. The following TPs are removed - writeback_nothread, writeback_wake_thread, writeback_wake_forker_thread, writeback_thread_start, writeback_thread_stop. Everything, including devices coming and going away and rescuer operation under simulated memory pressure, seems to work fine in my test setup. Signed-off-by: Tejun Heo <tj@kernel.org> Reviewed-by: Jan Kara <jack@suse.cz> Cc: Jens Axboe <axboe@kernel.dk> Cc: Fengguang Wu <fengguang.wu@intel.com> Cc: Jeff Moyer <jmoyer@redhat.com>
2013-04-01 20:08:06 -06:00
bdi_wb_shutdown(bdi);
bdi_debug_unregister(bdi);
device_unregister(bdi->dev);
bdi->dev = NULL;
}
}
EXPORT_SYMBOL(bdi_unregister);
static void bdi_wb_init(struct bdi_writeback *wb, struct backing_dev_info *bdi)
{
memset(wb, 0, sizeof(*wb));
wb->bdi = bdi;
wb->last_old_flush = jiffies;
INIT_LIST_HEAD(&wb->b_dirty);
INIT_LIST_HEAD(&wb->b_io);
INIT_LIST_HEAD(&wb->b_more_io);
writeback: split inode_wb_list_lock into bdi_writeback.list_lock Split the global inode_wb_list_lock into a per-bdi_writeback list_lock, as it's currently the most contended lock in the system for metadata heavy workloads. It won't help for single-filesystem workloads for which we'll need the I/O-less balance_dirty_pages, but at least we can dedicate a cpu to spinning on each bdi now for larger systems. Based on earlier patches from Nick Piggin and Dave Chinner. It reduces lock contentions to 1/4 in this test case: 10 HDD JBOD, 100 dd on each disk, XFS, 6GB ram lock_stat version 0.3 ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- class name con-bounces contentions waittime-min waittime-max waittime-total acq-bounces acquisitions holdtime-min holdtime-max holdtime-total ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- vanilla 2.6.39-rc3: inode_wb_list_lock: 42590 44433 0.12 147.74 144127.35 252274 886792 0.08 121.34 917211.23 ------------------ inode_wb_list_lock 2 [<ffffffff81165da5>] bdev_inode_switch_bdi+0x29/0x85 inode_wb_list_lock 34 [<ffffffff8115bd0b>] inode_wb_list_del+0x22/0x49 inode_wb_list_lock 12893 [<ffffffff8115bb53>] __mark_inode_dirty+0x170/0x1d0 inode_wb_list_lock 10702 [<ffffffff8115afef>] writeback_single_inode+0x16d/0x20a ------------------ inode_wb_list_lock 2 [<ffffffff81165da5>] bdev_inode_switch_bdi+0x29/0x85 inode_wb_list_lock 19 [<ffffffff8115bd0b>] inode_wb_list_del+0x22/0x49 inode_wb_list_lock 5550 [<ffffffff8115bb53>] __mark_inode_dirty+0x170/0x1d0 inode_wb_list_lock 8511 [<ffffffff8115b4ad>] writeback_sb_inodes+0x10f/0x157 2.6.39-rc3 + patch: &(&wb->list_lock)->rlock: 11383 11657 0.14 151.69 40429.51 90825 527918 0.11 145.90 556843.37 ------------------------ &(&wb->list_lock)->rlock 10 [<ffffffff8115b189>] inode_wb_list_del+0x5f/0x86 &(&wb->list_lock)->rlock 1493 [<ffffffff8115b1ed>] writeback_inodes_wb+0x3d/0x150 &(&wb->list_lock)->rlock 3652 [<ffffffff8115a8e9>] writeback_sb_inodes+0x123/0x16f &(&wb->list_lock)->rlock 1412 [<ffffffff8115a38e>] writeback_single_inode+0x17f/0x223 ------------------------ &(&wb->list_lock)->rlock 3 [<ffffffff8110b5af>] bdi_lock_two+0x46/0x4b &(&wb->list_lock)->rlock 6 [<ffffffff8115b189>] inode_wb_list_del+0x5f/0x86 &(&wb->list_lock)->rlock 2061 [<ffffffff8115af97>] __mark_inode_dirty+0x173/0x1cf &(&wb->list_lock)->rlock 2629 [<ffffffff8115a8e9>] writeback_sb_inodes+0x123/0x16f hughd@google.com: fix recursive lock when bdi_lock_two() is called with new the same as old akpm@linux-foundation.org: cleanup bdev_inode_switch_bdi() comment Signed-off-by: Christoph Hellwig <hch@lst.de> Signed-off-by: Hugh Dickins <hughd@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
2011-04-21 18:19:44 -06:00
spin_lock_init(&wb->list_lock);
writeback: replace custom worker pool implementation with unbound workqueue Writeback implements its own worker pool - each bdi can be associated with a worker thread which is created and destroyed dynamically. The worker thread for the default bdi is always present and serves as the "forker" thread which forks off worker threads for other bdis. there's no reason for writeback to implement its own worker pool when using unbound workqueue instead is much simpler and more efficient. This patch replaces custom worker pool implementation in writeback with an unbound workqueue. The conversion isn't too complicated but the followings are worth mentioning. * bdi_writeback->last_active, task and wakeup_timer are removed. delayed_work ->dwork is added instead. Explicit timer handling is no longer necessary. Everything works by either queueing / modding / flushing / canceling the delayed_work item. * bdi_writeback_thread() becomes bdi_writeback_workfn() which runs off bdi_writeback->dwork. On each execution, it processes bdi->work_list and reschedules itself if there are more things to do. The function also handles low-mem condition, which used to be handled by the forker thread. If the function is running off a rescuer thread, it only writes out limited number of pages so that the rescuer can serve other bdis too. This preserves the flusher creation failure behavior of the forker thread. * INIT_LIST_HEAD(&bdi->bdi_list) is used to tell bdi_writeback_workfn() about on-going bdi unregistration so that it always drains work_list even if it's running off the rescuer. Note that the original code was broken in this regard. Under memory pressure, a bdi could finish unregistration with non-empty work_list. * The default bdi is no longer special. It now is treated the same as any other bdi and bdi_cap_flush_forker() is removed. * BDI_pending is no longer used. Removed. * Some tracepoints become non-applicable. The following TPs are removed - writeback_nothread, writeback_wake_thread, writeback_wake_forker_thread, writeback_thread_start, writeback_thread_stop. Everything, including devices coming and going away and rescuer operation under simulated memory pressure, seems to work fine in my test setup. Signed-off-by: Tejun Heo <tj@kernel.org> Reviewed-by: Jan Kara <jack@suse.cz> Cc: Jens Axboe <axboe@kernel.dk> Cc: Fengguang Wu <fengguang.wu@intel.com> Cc: Jeff Moyer <jmoyer@redhat.com>
2013-04-01 20:08:06 -06:00
INIT_DELAYED_WORK(&wb->dwork, bdi_writeback_workfn);
}
writeback: bdi write bandwidth estimation The estimation value will start from 100MB/s and adapt to the real bandwidth in seconds. It tries to update the bandwidth only when disk is fully utilized. Any inactive period of more than one second will be skipped. The estimated bandwidth will be reflecting how fast the device can writeout when _fully utilized_, and won't drop to 0 when it goes idle. The value will remain constant at disk idle time. At busy write time, if not considering fluctuations, it will also remain high unless be knocked down by possible concurrent reads that compete for the disk time and bandwidth with async writes. The estimation is not done purely in the flusher because there is no guarantee for write_cache_pages() to return timely to update bandwidth. The bdi->avg_write_bandwidth smoothing is very effective for filtering out sudden spikes, however may be a little biased in long term. The overheads are low because the bdi bandwidth update only occurs at 200ms intervals. The 200ms update interval is suitable, because it's not possible to get the real bandwidth for the instance at all, due to large fluctuations. The NFS commits can be as large as seconds worth of data. One XFS completion may be as large as half second worth of data if we are going to increase the write chunk to half second worth of data. In ext4, fluctuations with time period of around 5 seconds is observed. And there is another pattern of irregular periods of up to 20 seconds on SSD tests. That's why we are not only doing the estimation at 200ms intervals, but also averaging them over a period of 3 seconds and then go further to do another level of smoothing in avg_write_bandwidth. CC: Li Shaohua <shaohua.li@intel.com> CC: Peter Zijlstra <a.p.zijlstra@chello.nl> Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
2010-08-29 11:22:30 -06:00
/*
* Initial write bandwidth: 100 MB/s
*/
#define INIT_BW (100 << (20 - PAGE_SHIFT))
int bdi_init(struct backing_dev_info *bdi)
{
writeback: switch to per-bdi threads for flushing data This gets rid of pdflush for bdi writeout and kupdated style cleaning. pdflush writeout suffers from lack of locality and also requires more threads to handle the same workload, since it has to work in a non-blocking fashion against each queue. This also introduces lumpy behaviour and potential request starvation, since pdflush can be starved for queue access if others are accessing it. A sample ffsb workload that does random writes to files is about 8% faster here on a simple SATA drive during the benchmark phase. File layout also seems a LOT more smooth in vmstat: r b swpd free buff cache si so bi bo in cs us sy id wa 0 1 0 608848 2652 375372 0 0 0 71024 604 24 1 10 48 42 0 1 0 549644 2712 433736 0 0 0 60692 505 27 1 8 48 44 1 0 0 476928 2784 505192 0 0 4 29540 553 24 0 9 53 37 0 1 0 457972 2808 524008 0 0 0 54876 331 16 0 4 38 58 0 1 0 366128 2928 614284 0 0 4 92168 710 58 0 13 53 34 0 1 0 295092 3000 684140 0 0 0 62924 572 23 0 9 53 37 0 1 0 236592 3064 741704 0 0 4 58256 523 17 0 8 48 44 0 1 0 165608 3132 811464 0 0 0 57460 560 21 0 8 54 38 0 1 0 102952 3200 873164 0 0 4 74748 540 29 1 10 48 41 0 1 0 48604 3252 926472 0 0 0 53248 469 29 0 7 47 45 where vanilla tends to fluctuate a lot in the creation phase: r b swpd free buff cache si so bi bo in cs us sy id wa 1 1 0 678716 5792 303380 0 0 0 74064 565 50 1 11 52 36 1 0 0 662488 5864 319396 0 0 4 352 302 329 0 2 47 51 0 1 0 599312 5924 381468 0 0 0 78164 516 55 0 9 51 40 0 1 0 519952 6008 459516 0 0 4 78156 622 56 1 11 52 37 1 1 0 436640 6092 541632 0 0 0 82244 622 54 0 11 48 41 0 1 0 436640 6092 541660 0 0 0 8 152 39 0 0 51 49 0 1 0 332224 6200 644252 0 0 4 102800 728 46 1 13 49 36 1 0 0 274492 6260 701056 0 0 4 12328 459 49 0 7 50 43 0 1 0 211220 6324 763356 0 0 0 106940 515 37 1 10 51 39 1 0 0 160412 6376 813468 0 0 0 8224 415 43 0 6 49 45 1 1 0 85980 6452 886556 0 0 4 113516 575 39 1 11 54 34 0 2 0 85968 6452 886620 0 0 0 1640 158 211 0 0 46 54 A 10 disk test with btrfs performs 26% faster with per-bdi flushing. A SSD based writeback test on XFS performs over 20% better as well, with the throughput being very stable around 1GB/sec, where pdflush only manages 750MB/sec and fluctuates wildly while doing so. Random buffered writes to many files behave a lot better as well, as does random mmap'ed writes. A separate thread is added to sync the super blocks. In the long term, adding sync_supers_bdi() functionality could get rid of this thread again. Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
2009-09-09 01:08:54 -06:00
int i, err;
bdi->dev = NULL;
bdi->min_ratio = 0;
bdi->max_ratio = 100;
bdi->max_prop_frac = FPROP_FRAC_BASE;
writeback: switch to per-bdi threads for flushing data This gets rid of pdflush for bdi writeout and kupdated style cleaning. pdflush writeout suffers from lack of locality and also requires more threads to handle the same workload, since it has to work in a non-blocking fashion against each queue. This also introduces lumpy behaviour and potential request starvation, since pdflush can be starved for queue access if others are accessing it. A sample ffsb workload that does random writes to files is about 8% faster here on a simple SATA drive during the benchmark phase. File layout also seems a LOT more smooth in vmstat: r b swpd free buff cache si so bi bo in cs us sy id wa 0 1 0 608848 2652 375372 0 0 0 71024 604 24 1 10 48 42 0 1 0 549644 2712 433736 0 0 0 60692 505 27 1 8 48 44 1 0 0 476928 2784 505192 0 0 4 29540 553 24 0 9 53 37 0 1 0 457972 2808 524008 0 0 0 54876 331 16 0 4 38 58 0 1 0 366128 2928 614284 0 0 4 92168 710 58 0 13 53 34 0 1 0 295092 3000 684140 0 0 0 62924 572 23 0 9 53 37 0 1 0 236592 3064 741704 0 0 4 58256 523 17 0 8 48 44 0 1 0 165608 3132 811464 0 0 0 57460 560 21 0 8 54 38 0 1 0 102952 3200 873164 0 0 4 74748 540 29 1 10 48 41 0 1 0 48604 3252 926472 0 0 0 53248 469 29 0 7 47 45 where vanilla tends to fluctuate a lot in the creation phase: r b swpd free buff cache si so bi bo in cs us sy id wa 1 1 0 678716 5792 303380 0 0 0 74064 565 50 1 11 52 36 1 0 0 662488 5864 319396 0 0 4 352 302 329 0 2 47 51 0 1 0 599312 5924 381468 0 0 0 78164 516 55 0 9 51 40 0 1 0 519952 6008 459516 0 0 4 78156 622 56 1 11 52 37 1 1 0 436640 6092 541632 0 0 0 82244 622 54 0 11 48 41 0 1 0 436640 6092 541660 0 0 0 8 152 39 0 0 51 49 0 1 0 332224 6200 644252 0 0 4 102800 728 46 1 13 49 36 1 0 0 274492 6260 701056 0 0 4 12328 459 49 0 7 50 43 0 1 0 211220 6324 763356 0 0 0 106940 515 37 1 10 51 39 1 0 0 160412 6376 813468 0 0 0 8224 415 43 0 6 49 45 1 1 0 85980 6452 886556 0 0 4 113516 575 39 1 11 54 34 0 2 0 85968 6452 886620 0 0 0 1640 158 211 0 0 46 54 A 10 disk test with btrfs performs 26% faster with per-bdi flushing. A SSD based writeback test on XFS performs over 20% better as well, with the throughput being very stable around 1GB/sec, where pdflush only manages 750MB/sec and fluctuates wildly while doing so. Random buffered writes to many files behave a lot better as well, as does random mmap'ed writes. A separate thread is added to sync the super blocks. In the long term, adding sync_supers_bdi() functionality could get rid of this thread again. Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
2009-09-09 01:08:54 -06:00
spin_lock_init(&bdi->wb_lock);
INIT_LIST_HEAD(&bdi->bdi_list);
writeback: switch to per-bdi threads for flushing data This gets rid of pdflush for bdi writeout and kupdated style cleaning. pdflush writeout suffers from lack of locality and also requires more threads to handle the same workload, since it has to work in a non-blocking fashion against each queue. This also introduces lumpy behaviour and potential request starvation, since pdflush can be starved for queue access if others are accessing it. A sample ffsb workload that does random writes to files is about 8% faster here on a simple SATA drive during the benchmark phase. File layout also seems a LOT more smooth in vmstat: r b swpd free buff cache si so bi bo in cs us sy id wa 0 1 0 608848 2652 375372 0 0 0 71024 604 24 1 10 48 42 0 1 0 549644 2712 433736 0 0 0 60692 505 27 1 8 48 44 1 0 0 476928 2784 505192 0 0 4 29540 553 24 0 9 53 37 0 1 0 457972 2808 524008 0 0 0 54876 331 16 0 4 38 58 0 1 0 366128 2928 614284 0 0 4 92168 710 58 0 13 53 34 0 1 0 295092 3000 684140 0 0 0 62924 572 23 0 9 53 37 0 1 0 236592 3064 741704 0 0 4 58256 523 17 0 8 48 44 0 1 0 165608 3132 811464 0 0 0 57460 560 21 0 8 54 38 0 1 0 102952 3200 873164 0 0 4 74748 540 29 1 10 48 41 0 1 0 48604 3252 926472 0 0 0 53248 469 29 0 7 47 45 where vanilla tends to fluctuate a lot in the creation phase: r b swpd free buff cache si so bi bo in cs us sy id wa 1 1 0 678716 5792 303380 0 0 0 74064 565 50 1 11 52 36 1 0 0 662488 5864 319396 0 0 4 352 302 329 0 2 47 51 0 1 0 599312 5924 381468 0 0 0 78164 516 55 0 9 51 40 0 1 0 519952 6008 459516 0 0 4 78156 622 56 1 11 52 37 1 1 0 436640 6092 541632 0 0 0 82244 622 54 0 11 48 41 0 1 0 436640 6092 541660 0 0 0 8 152 39 0 0 51 49 0 1 0 332224 6200 644252 0 0 4 102800 728 46 1 13 49 36 1 0 0 274492 6260 701056 0 0 4 12328 459 49 0 7 50 43 0 1 0 211220 6324 763356 0 0 0 106940 515 37 1 10 51 39 1 0 0 160412 6376 813468 0 0 0 8224 415 43 0 6 49 45 1 1 0 85980 6452 886556 0 0 4 113516 575 39 1 11 54 34 0 2 0 85968 6452 886620 0 0 0 1640 158 211 0 0 46 54 A 10 disk test with btrfs performs 26% faster with per-bdi flushing. A SSD based writeback test on XFS performs over 20% better as well, with the throughput being very stable around 1GB/sec, where pdflush only manages 750MB/sec and fluctuates wildly while doing so. Random buffered writes to many files behave a lot better as well, as does random mmap'ed writes. A separate thread is added to sync the super blocks. In the long term, adding sync_supers_bdi() functionality could get rid of this thread again. Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
2009-09-09 01:08:54 -06:00
INIT_LIST_HEAD(&bdi->work_list);
bdi_wb_init(&bdi->wb, bdi);
for (i = 0; i < NR_BDI_STAT_ITEMS; i++) {
err = percpu_counter_init(&bdi->bdi_stat[i], 0, GFP_KERNEL);
mm: per device dirty threshold Scale writeback cache per backing device, proportional to its writeout speed. By decoupling the BDI dirty thresholds a number of problems we currently have will go away, namely: - mutual interference starvation (for any number of BDIs); - deadlocks with stacked BDIs (loop, FUSE and local NFS mounts). It might be that all dirty pages are for a single BDI while other BDIs are idling. By giving each BDI a 'fair' share of the dirty limit, each one can have dirty pages outstanding and make progress. A global threshold also creates a deadlock for stacked BDIs; when A writes to B, and A generates enough dirty pages to get throttled, B will never start writeback until the dirty pages go away. Again, by giving each BDI its own 'independent' dirty limit, this problem is avoided. So the problem is to determine how to distribute the total dirty limit across the BDIs fairly and efficiently. A DBI that has a large dirty limit but does not have any dirty pages outstanding is a waste. What is done is to keep a floating proportion between the DBIs based on writeback completions. This way faster/more active devices get a larger share than slower/idle devices. [akpm@linux-foundation.org: fix warnings] [hugh@veritas.com: Fix occasional hang when a task couldn't get out of balance_dirty_pages] Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Signed-off-by: Hugh Dickins <hugh@veritas.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2007-10-17 00:25:50 -06:00
if (err)
goto err;
}
bdi->dirty_exceeded = 0;
writeback: bdi write bandwidth estimation The estimation value will start from 100MB/s and adapt to the real bandwidth in seconds. It tries to update the bandwidth only when disk is fully utilized. Any inactive period of more than one second will be skipped. The estimated bandwidth will be reflecting how fast the device can writeout when _fully utilized_, and won't drop to 0 when it goes idle. The value will remain constant at disk idle time. At busy write time, if not considering fluctuations, it will also remain high unless be knocked down by possible concurrent reads that compete for the disk time and bandwidth with async writes. The estimation is not done purely in the flusher because there is no guarantee for write_cache_pages() to return timely to update bandwidth. The bdi->avg_write_bandwidth smoothing is very effective for filtering out sudden spikes, however may be a little biased in long term. The overheads are low because the bdi bandwidth update only occurs at 200ms intervals. The 200ms update interval is suitable, because it's not possible to get the real bandwidth for the instance at all, due to large fluctuations. The NFS commits can be as large as seconds worth of data. One XFS completion may be as large as half second worth of data if we are going to increase the write chunk to half second worth of data. In ext4, fluctuations with time period of around 5 seconds is observed. And there is another pattern of irregular periods of up to 20 seconds on SSD tests. That's why we are not only doing the estimation at 200ms intervals, but also averaging them over a period of 3 seconds and then go further to do another level of smoothing in avg_write_bandwidth. CC: Li Shaohua <shaohua.li@intel.com> CC: Peter Zijlstra <a.p.zijlstra@chello.nl> Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
2010-08-29 11:22:30 -06:00
bdi->bw_time_stamp = jiffies;
bdi->written_stamp = 0;
writeback: stabilize bdi->dirty_ratelimit There are some imperfections in balanced_dirty_ratelimit. 1) large fluctuations The dirty_rate used for computing balanced_dirty_ratelimit is merely averaged in the past 200ms (very small comparing to the 3s estimation period for write_bw), which makes rather dispersed distribution of balanced_dirty_ratelimit. It's pretty hard to average out the singular points by increasing the estimation period. Considering that the averaging technique will introduce very undesirable time lags, I give it up totally. (btw, the 3s write_bw averaging time lag is much more acceptable because its impact is one-way and therefore won't lead to oscillations.) The more practical way is filtering -- most singular balanced_dirty_ratelimit points can be filtered out by remembering some prev_balanced_rate and prev_prev_balanced_rate. However the more reliable way is to guard balanced_dirty_ratelimit with task_ratelimit. 2) due to truncates and fs redirties, the (write_bw <=> dirty_rate) match could become unbalanced, which may lead to large systematical errors in balanced_dirty_ratelimit. The truncates, due to its possibly bumpy nature, can hardly be compensated smoothly. So let's face it. When some over-estimated balanced_dirty_ratelimit brings dirty_ratelimit high, dirty pages will go higher than the setpoint. task_ratelimit will in turn become lower than dirty_ratelimit. So if we consider both balanced_dirty_ratelimit and task_ratelimit and update dirty_ratelimit only when they are on the same side of dirty_ratelimit, the systematical errors in balanced_dirty_ratelimit won't be able to bring dirty_ratelimit far away. The balanced_dirty_ratelimit estimation may also be inaccurate near @limit or @freerun, however is less an issue. 3) since we ultimately want to - keep the fluctuations of task ratelimit as small as possible - keep the dirty pages around the setpoint as long time as possible the update policy used for (2) also serves the above goals nicely: if for some reason the dirty pages are high (task_ratelimit < dirty_ratelimit), and dirty_ratelimit is low (dirty_ratelimit < balanced_dirty_ratelimit), there is no point to bring up dirty_ratelimit in a hurry only to hurt both the above two goals. So, we make use of task_ratelimit to limit the update of dirty_ratelimit in two ways: 1) avoid changing dirty rate when it's against the position control target (the adjusted rate will slow down the progress of dirty pages going back to setpoint). 2) limit the step size. task_ratelimit is changing values step by step, leaving a consistent trace comparing to the randomly jumping balanced_dirty_ratelimit. task_ratelimit also has the nice smaller errors in stable state and typically larger errors when there are big errors in rate. So it's a pretty good limiting factor for the step size of dirty_ratelimit. Note that bdi->dirty_ratelimit is always tracking balanced_dirty_ratelimit. task_ratelimit is merely used as a limiting factor. Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
2011-08-26 15:53:24 -06:00
bdi->balanced_dirty_ratelimit = INIT_BW;
writeback: dirty rate control It's all about bdi->dirty_ratelimit, which aims to be (write_bw / N) when there are N dd tasks. On write() syscall, use bdi->dirty_ratelimit ============================================ balance_dirty_pages(pages_dirtied) { task_ratelimit = bdi->dirty_ratelimit * bdi_position_ratio(); pause = pages_dirtied / task_ratelimit; sleep(pause); } On every 200ms, update bdi->dirty_ratelimit =========================================== bdi_update_dirty_ratelimit() { task_ratelimit = bdi->dirty_ratelimit * bdi_position_ratio(); balanced_dirty_ratelimit = task_ratelimit * write_bw / dirty_rate; bdi->dirty_ratelimit = balanced_dirty_ratelimit } Estimation of balanced bdi->dirty_ratelimit =========================================== balanced task_ratelimit ----------------------- balance_dirty_pages() needs to throttle tasks dirtying pages such that the total amount of dirty pages stays below the specified dirty limit in order to avoid memory deadlocks. Furthermore we desire fairness in that tasks get throttled proportionally to the amount of pages they dirty. IOW we want to throttle tasks such that we match the dirty rate to the writeout bandwidth, this yields a stable amount of dirty pages: dirty_rate == write_bw (1) The fairness requirement gives us: task_ratelimit = balanced_dirty_ratelimit == write_bw / N (2) where N is the number of dd tasks. We don't know N beforehand, but still can estimate balanced_dirty_ratelimit within 200ms. Start by throttling each dd task at rate task_ratelimit = task_ratelimit_0 (3) (any non-zero initial value is OK) After 200ms, we measured dirty_rate = # of pages dirtied by all dd's / 200ms write_bw = # of pages written to the disk / 200ms For the aggressive dd dirtiers, the equality holds dirty_rate == N * task_rate == N * task_ratelimit_0 (4) Or task_ratelimit_0 == dirty_rate / N (5) Now we conclude that the balanced task ratelimit can be estimated by write_bw balanced_dirty_ratelimit = task_ratelimit_0 * ---------- (6) dirty_rate Because with (4) and (5) we can get the desired equality (1): write_bw balanced_dirty_ratelimit == (dirty_rate / N) * ---------- dirty_rate == write_bw / N Then using the balanced task ratelimit we can compute task pause times like: task_pause = task->nr_dirtied / task_ratelimit task_ratelimit with position control ------------------------------------ However, while the above gives us means of matching the dirty rate to the writeout bandwidth, it at best provides us with a stable dirty page count (assuming a static system). In order to control the dirty page count such that it is high enough to provide performance, but does not exceed the specified limit we need another control. The dirty position control works by extending (2) to task_ratelimit = balanced_dirty_ratelimit * pos_ratio (7) where pos_ratio is a negative feedback function that subjects to 1) f(setpoint) = 1.0 2) df/dx < 0 That is, if the dirty pages are ABOVE the setpoint, we throttle each task a bit more HEAVY than balanced_dirty_ratelimit, so that the dirty pages are created less fast than they are cleaned, thus DROP to the setpoints (and the reverse). Based on (7) and the assumption that both dirty_ratelimit and pos_ratio remains CONSTANT for the past 200ms, we get task_ratelimit_0 = balanced_dirty_ratelimit * pos_ratio (8) Putting (8) into (6), we get the formula used in bdi_update_dirty_ratelimit(): write_bw balanced_dirty_ratelimit *= pos_ratio * ---------- (9) dirty_rate Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
2011-06-12 10:51:31 -06:00
bdi->dirty_ratelimit = INIT_BW;
writeback: bdi write bandwidth estimation The estimation value will start from 100MB/s and adapt to the real bandwidth in seconds. It tries to update the bandwidth only when disk is fully utilized. Any inactive period of more than one second will be skipped. The estimated bandwidth will be reflecting how fast the device can writeout when _fully utilized_, and won't drop to 0 when it goes idle. The value will remain constant at disk idle time. At busy write time, if not considering fluctuations, it will also remain high unless be knocked down by possible concurrent reads that compete for the disk time and bandwidth with async writes. The estimation is not done purely in the flusher because there is no guarantee for write_cache_pages() to return timely to update bandwidth. The bdi->avg_write_bandwidth smoothing is very effective for filtering out sudden spikes, however may be a little biased in long term. The overheads are low because the bdi bandwidth update only occurs at 200ms intervals. The 200ms update interval is suitable, because it's not possible to get the real bandwidth for the instance at all, due to large fluctuations. The NFS commits can be as large as seconds worth of data. One XFS completion may be as large as half second worth of data if we are going to increase the write chunk to half second worth of data. In ext4, fluctuations with time period of around 5 seconds is observed. And there is another pattern of irregular periods of up to 20 seconds on SSD tests. That's why we are not only doing the estimation at 200ms intervals, but also averaging them over a period of 3 seconds and then go further to do another level of smoothing in avg_write_bandwidth. CC: Li Shaohua <shaohua.li@intel.com> CC: Peter Zijlstra <a.p.zijlstra@chello.nl> Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
2010-08-29 11:22:30 -06:00
bdi->write_bandwidth = INIT_BW;
bdi->avg_write_bandwidth = INIT_BW;
err = fprop_local_init_percpu(&bdi->completions, GFP_KERNEL);
mm: per device dirty threshold Scale writeback cache per backing device, proportional to its writeout speed. By decoupling the BDI dirty thresholds a number of problems we currently have will go away, namely: - mutual interference starvation (for any number of BDIs); - deadlocks with stacked BDIs (loop, FUSE and local NFS mounts). It might be that all dirty pages are for a single BDI while other BDIs are idling. By giving each BDI a 'fair' share of the dirty limit, each one can have dirty pages outstanding and make progress. A global threshold also creates a deadlock for stacked BDIs; when A writes to B, and A generates enough dirty pages to get throttled, B will never start writeback until the dirty pages go away. Again, by giving each BDI its own 'independent' dirty limit, this problem is avoided. So the problem is to determine how to distribute the total dirty limit across the BDIs fairly and efficiently. A DBI that has a large dirty limit but does not have any dirty pages outstanding is a waste. What is done is to keep a floating proportion between the DBIs based on writeback completions. This way faster/more active devices get a larger share than slower/idle devices. [akpm@linux-foundation.org: fix warnings] [hugh@veritas.com: Fix occasional hang when a task couldn't get out of balance_dirty_pages] Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Signed-off-by: Hugh Dickins <hugh@veritas.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2007-10-17 00:25:50 -06:00
if (err) {
err:
while (i--)
mm: per device dirty threshold Scale writeback cache per backing device, proportional to its writeout speed. By decoupling the BDI dirty thresholds a number of problems we currently have will go away, namely: - mutual interference starvation (for any number of BDIs); - deadlocks with stacked BDIs (loop, FUSE and local NFS mounts). It might be that all dirty pages are for a single BDI while other BDIs are idling. By giving each BDI a 'fair' share of the dirty limit, each one can have dirty pages outstanding and make progress. A global threshold also creates a deadlock for stacked BDIs; when A writes to B, and A generates enough dirty pages to get throttled, B will never start writeback until the dirty pages go away. Again, by giving each BDI its own 'independent' dirty limit, this problem is avoided. So the problem is to determine how to distribute the total dirty limit across the BDIs fairly and efficiently. A DBI that has a large dirty limit but does not have any dirty pages outstanding is a waste. What is done is to keep a floating proportion between the DBIs based on writeback completions. This way faster/more active devices get a larger share than slower/idle devices. [akpm@linux-foundation.org: fix warnings] [hugh@veritas.com: Fix occasional hang when a task couldn't get out of balance_dirty_pages] Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Signed-off-by: Hugh Dickins <hugh@veritas.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2007-10-17 00:25:50 -06:00
percpu_counter_destroy(&bdi->bdi_stat[i]);
}
return err;
}
EXPORT_SYMBOL(bdi_init);
void bdi_destroy(struct backing_dev_info *bdi)
{
int i;
/*
* Splice our entries to the default_backing_dev_info. This
* condition shouldn't happen. @wb must be empty at this point and
* dirty inodes on it might cause other issues. This workaround is
* added by ce5f8e779519 ("writeback: splice dirty inode entries to
* default bdi on bdi_destroy()") without root-causing the issue.
*
* http://lkml.kernel.org/g/1253038617-30204-11-git-send-email-jens.axboe@oracle.com
* http://thread.gmane.org/gmane.linux.file-systems/35341/focus=35350
*
* We should probably add WARN_ON() to find out whether it still
* happens and track it down if so.
*/
if (bdi_has_dirty_io(bdi)) {
struct bdi_writeback *dst = &default_backing_dev_info.wb;
writeback: split inode_wb_list_lock into bdi_writeback.list_lock Split the global inode_wb_list_lock into a per-bdi_writeback list_lock, as it's currently the most contended lock in the system for metadata heavy workloads. It won't help for single-filesystem workloads for which we'll need the I/O-less balance_dirty_pages, but at least we can dedicate a cpu to spinning on each bdi now for larger systems. Based on earlier patches from Nick Piggin and Dave Chinner. It reduces lock contentions to 1/4 in this test case: 10 HDD JBOD, 100 dd on each disk, XFS, 6GB ram lock_stat version 0.3 ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- class name con-bounces contentions waittime-min waittime-max waittime-total acq-bounces acquisitions holdtime-min holdtime-max holdtime-total ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- vanilla 2.6.39-rc3: inode_wb_list_lock: 42590 44433 0.12 147.74 144127.35 252274 886792 0.08 121.34 917211.23 ------------------ inode_wb_list_lock 2 [<ffffffff81165da5>] bdev_inode_switch_bdi+0x29/0x85 inode_wb_list_lock 34 [<ffffffff8115bd0b>] inode_wb_list_del+0x22/0x49 inode_wb_list_lock 12893 [<ffffffff8115bb53>] __mark_inode_dirty+0x170/0x1d0 inode_wb_list_lock 10702 [<ffffffff8115afef>] writeback_single_inode+0x16d/0x20a ------------------ inode_wb_list_lock 2 [<ffffffff81165da5>] bdev_inode_switch_bdi+0x29/0x85 inode_wb_list_lock 19 [<ffffffff8115bd0b>] inode_wb_list_del+0x22/0x49 inode_wb_list_lock 5550 [<ffffffff8115bb53>] __mark_inode_dirty+0x170/0x1d0 inode_wb_list_lock 8511 [<ffffffff8115b4ad>] writeback_sb_inodes+0x10f/0x157 2.6.39-rc3 + patch: &(&wb->list_lock)->rlock: 11383 11657 0.14 151.69 40429.51 90825 527918 0.11 145.90 556843.37 ------------------------ &(&wb->list_lock)->rlock 10 [<ffffffff8115b189>] inode_wb_list_del+0x5f/0x86 &(&wb->list_lock)->rlock 1493 [<ffffffff8115b1ed>] writeback_inodes_wb+0x3d/0x150 &(&wb->list_lock)->rlock 3652 [<ffffffff8115a8e9>] writeback_sb_inodes+0x123/0x16f &(&wb->list_lock)->rlock 1412 [<ffffffff8115a38e>] writeback_single_inode+0x17f/0x223 ------------------------ &(&wb->list_lock)->rlock 3 [<ffffffff8110b5af>] bdi_lock_two+0x46/0x4b &(&wb->list_lock)->rlock 6 [<ffffffff8115b189>] inode_wb_list_del+0x5f/0x86 &(&wb->list_lock)->rlock 2061 [<ffffffff8115af97>] __mark_inode_dirty+0x173/0x1cf &(&wb->list_lock)->rlock 2629 [<ffffffff8115a8e9>] writeback_sb_inodes+0x123/0x16f hughd@google.com: fix recursive lock when bdi_lock_two() is called with new the same as old akpm@linux-foundation.org: cleanup bdev_inode_switch_bdi() comment Signed-off-by: Christoph Hellwig <hch@lst.de> Signed-off-by: Hugh Dickins <hughd@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
2011-04-21 18:19:44 -06:00
bdi_lock_two(&bdi->wb, dst);
list_splice(&bdi->wb.b_dirty, &dst->b_dirty);
list_splice(&bdi->wb.b_io, &dst->b_io);
list_splice(&bdi->wb.b_more_io, &dst->b_more_io);
writeback: split inode_wb_list_lock into bdi_writeback.list_lock Split the global inode_wb_list_lock into a per-bdi_writeback list_lock, as it's currently the most contended lock in the system for metadata heavy workloads. It won't help for single-filesystem workloads for which we'll need the I/O-less balance_dirty_pages, but at least we can dedicate a cpu to spinning on each bdi now for larger systems. Based on earlier patches from Nick Piggin and Dave Chinner. It reduces lock contentions to 1/4 in this test case: 10 HDD JBOD, 100 dd on each disk, XFS, 6GB ram lock_stat version 0.3 ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- class name con-bounces contentions waittime-min waittime-max waittime-total acq-bounces acquisitions holdtime-min holdtime-max holdtime-total ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- vanilla 2.6.39-rc3: inode_wb_list_lock: 42590 44433 0.12 147.74 144127.35 252274 886792 0.08 121.34 917211.23 ------------------ inode_wb_list_lock 2 [<ffffffff81165da5>] bdev_inode_switch_bdi+0x29/0x85 inode_wb_list_lock 34 [<ffffffff8115bd0b>] inode_wb_list_del+0x22/0x49 inode_wb_list_lock 12893 [<ffffffff8115bb53>] __mark_inode_dirty+0x170/0x1d0 inode_wb_list_lock 10702 [<ffffffff8115afef>] writeback_single_inode+0x16d/0x20a ------------------ inode_wb_list_lock 2 [<ffffffff81165da5>] bdev_inode_switch_bdi+0x29/0x85 inode_wb_list_lock 19 [<ffffffff8115bd0b>] inode_wb_list_del+0x22/0x49 inode_wb_list_lock 5550 [<ffffffff8115bb53>] __mark_inode_dirty+0x170/0x1d0 inode_wb_list_lock 8511 [<ffffffff8115b4ad>] writeback_sb_inodes+0x10f/0x157 2.6.39-rc3 + patch: &(&wb->list_lock)->rlock: 11383 11657 0.14 151.69 40429.51 90825 527918 0.11 145.90 556843.37 ------------------------ &(&wb->list_lock)->rlock 10 [<ffffffff8115b189>] inode_wb_list_del+0x5f/0x86 &(&wb->list_lock)->rlock 1493 [<ffffffff8115b1ed>] writeback_inodes_wb+0x3d/0x150 &(&wb->list_lock)->rlock 3652 [<ffffffff8115a8e9>] writeback_sb_inodes+0x123/0x16f &(&wb->list_lock)->rlock 1412 [<ffffffff8115a38e>] writeback_single_inode+0x17f/0x223 ------------------------ &(&wb->list_lock)->rlock 3 [<ffffffff8110b5af>] bdi_lock_two+0x46/0x4b &(&wb->list_lock)->rlock 6 [<ffffffff8115b189>] inode_wb_list_del+0x5f/0x86 &(&wb->list_lock)->rlock 2061 [<ffffffff8115af97>] __mark_inode_dirty+0x173/0x1cf &(&wb->list_lock)->rlock 2629 [<ffffffff8115a8e9>] writeback_sb_inodes+0x123/0x16f hughd@google.com: fix recursive lock when bdi_lock_two() is called with new the same as old akpm@linux-foundation.org: cleanup bdev_inode_switch_bdi() comment Signed-off-by: Christoph Hellwig <hch@lst.de> Signed-off-by: Hugh Dickins <hughd@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
2011-04-21 18:19:44 -06:00
spin_unlock(&bdi->wb.list_lock);
spin_unlock(&dst->list_lock);
}
bdi_unregister(bdi);
WARN_ON(delayed_work_pending(&bdi->wb.dwork));
backing-dev: ensure wakeup_timer is deleted bdi_prune_sb() in bdi_unregister() attempts to removes the bdi links from all super_blocks and then del_timer_sync() the writeback timer. However, this can race with __mark_inode_dirty(), leading to bdi_wakeup_thread_delayed() rearming the writeback timer on the bdi we're unregistering, after we've called del_timer_sync(). This can end up with the bdi being freed with an active timer inside it, as in the case of the following dump after the removal of an SD card. Fix this by redoing the del_timer_sync() in bdi_destory(). ------------[ cut here ]------------ WARNING: at /home/rabin/kernel/arm/lib/debugobjects.c:262 debug_print_object+0x9c/0xc8() ODEBUG: free active (active state 0) object type: timer_list hint: wakeup_timer_fn+0x0/0x180 Modules linked in: Backtrace: [<c00109dc>] (dump_backtrace+0x0/0x110) from [<c0236e4c>] (dump_stack+0x18/0x1c) r6:c02bc638 r5:00000106 r4:c79f5d18 r3:00000000 [<c0236e34>] (dump_stack+0x0/0x1c) from [<c0025e6c>] (warn_slowpath_common+0x54/0x6c) [<c0025e18>] (warn_slowpath_common+0x0/0x6c) from [<c0025f28>] (warn_slowpath_fmt+0x38/0x40) r8:20000013 r7:c780c6f0 r6:c031613c r5:c780c6f0 r4:c02b1b29 r3:00000009 [<c0025ef0>] (warn_slowpath_fmt+0x0/0x40) from [<c015eb4c>] (debug_print_object+0x9c/0xc8) r3:c02b1b29 r2:c02bc662 [<c015eab0>] (debug_print_object+0x0/0xc8) from [<c015f574>] (debug_check_no_obj_freed+0xac/0x1dc) r6:c7964000 r5:00000001 r4:c7964000 [<c015f4c8>] (debug_check_no_obj_freed+0x0/0x1dc) from [<c00a9e38>] (kmem_cache_free+0x88/0x1f8) [<c00a9db0>] (kmem_cache_free+0x0/0x1f8) from [<c014286c>] (blk_release_queue+0x70/0x78) [<c01427fc>] (blk_release_queue+0x0/0x78) from [<c015290c>] (kobject_release+0x70/0x84) r5:c79641f0 r4:c796420c [<c015289c>] (kobject_release+0x0/0x84) from [<c0153ce4>] (kref_put+0x68/0x80) r7:00000083 r6:c74083d0 r5:c015289c r4:c796420c [<c0153c7c>] (kref_put+0x0/0x80) from [<c01527d0>] (kobject_put+0x48/0x5c) r5:c79643b4 r4:c79641f0 [<c0152788>] (kobject_put+0x0/0x5c) from [<c013ddd8>] (blk_cleanup_queue+0x68/0x74) r4:c7964000 [<c013dd70>] (blk_cleanup_queue+0x0/0x74) from [<c01a6370>] (mmc_blk_put+0x78/0xe8) r5:00000000 r4:c794c400 [<c01a62f8>] (mmc_blk_put+0x0/0xe8) from [<c01a64b4>] (mmc_blk_release+0x24/0x38) r5:c794c400 r4:c0322824 [<c01a6490>] (mmc_blk_release+0x0/0x38) from [<c00de11c>] (__blkdev_put+0xe8/0x170) r5:c78d5e00 r4:c74083c0 [<c00de034>] (__blkdev_put+0x0/0x170) from [<c00de2c0>] (blkdev_put+0x11c/0x12c) r8:c79f5f70 r7:00000001 r6:c74083d0 r5:00000083 r4:c74083c0 r3:00000000 [<c00de1a4>] (blkdev_put+0x0/0x12c) from [<c00b0724>] (kill_block_super+0x60/0x6c) r7:c7942300 r6:c79f4000 r5:00000083 r4:c74083c0 [<c00b06c4>] (kill_block_super+0x0/0x6c) from [<c00b0a94>] (deactivate_locked_super+0x44/0x70) r6:c79f4000 r5:c031af64 r4:c794dc00 r3:c00b06c4 [<c00b0a50>] (deactivate_locked_super+0x0/0x70) from [<c00b1358>] (deactivate_super+0x6c/0x70) r5:c794dc00 r4:c794dc00 [<c00b12ec>] (deactivate_super+0x0/0x70) from [<c00c88b0>] (mntput_no_expire+0x188/0x194) r5:c794dc00 r4:c7942300 [<c00c8728>] (mntput_no_expire+0x0/0x194) from [<c00c95e0>] (sys_umount+0x2e4/0x310) r6:c7942300 r5:00000000 r4:00000000 r3:00000000 [<c00c92fc>] (sys_umount+0x0/0x310) from [<c000d940>] (ret_fast_syscall+0x0/0x30) ---[ end trace e5c83c92ada51c76 ]--- Cc: stable@kernel.org Signed-off-by: Rabin Vincent <rabin.vincent@stericsson.com> Signed-off-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2011-11-11 05:29:04 -07:00
for (i = 0; i < NR_BDI_STAT_ITEMS; i++)
percpu_counter_destroy(&bdi->bdi_stat[i]);
mm: per device dirty threshold Scale writeback cache per backing device, proportional to its writeout speed. By decoupling the BDI dirty thresholds a number of problems we currently have will go away, namely: - mutual interference starvation (for any number of BDIs); - deadlocks with stacked BDIs (loop, FUSE and local NFS mounts). It might be that all dirty pages are for a single BDI while other BDIs are idling. By giving each BDI a 'fair' share of the dirty limit, each one can have dirty pages outstanding and make progress. A global threshold also creates a deadlock for stacked BDIs; when A writes to B, and A generates enough dirty pages to get throttled, B will never start writeback until the dirty pages go away. Again, by giving each BDI its own 'independent' dirty limit, this problem is avoided. So the problem is to determine how to distribute the total dirty limit across the BDIs fairly and efficiently. A DBI that has a large dirty limit but does not have any dirty pages outstanding is a waste. What is done is to keep a floating proportion between the DBIs based on writeback completions. This way faster/more active devices get a larger share than slower/idle devices. [akpm@linux-foundation.org: fix warnings] [hugh@veritas.com: Fix occasional hang when a task couldn't get out of balance_dirty_pages] Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Signed-off-by: Hugh Dickins <hugh@veritas.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2007-10-17 00:25:50 -06:00
fprop_local_destroy_percpu(&bdi->completions);
}
EXPORT_SYMBOL(bdi_destroy);
/*
* For use from filesystems to quickly init and register a bdi associated
* with dirty writeback
*/
int bdi_setup_and_register(struct backing_dev_info *bdi, char *name,
unsigned int cap)
{
int err;
bdi->name = name;
bdi->capabilities = cap;
err = bdi_init(bdi);
if (err)
return err;
err = bdi_register(bdi, NULL, "%.28s-%ld", name,
atomic_long_inc_return(&bdi_seq));
if (err) {
bdi_destroy(bdi);
return err;
}
return 0;
}
EXPORT_SYMBOL(bdi_setup_and_register);
static wait_queue_head_t congestion_wqh[2] = {
__WAIT_QUEUE_HEAD_INITIALIZER(congestion_wqh[0]),
__WAIT_QUEUE_HEAD_INITIALIZER(congestion_wqh[1])
};
static atomic_t nr_bdi_congested[2];
void clear_bdi_congested(struct backing_dev_info *bdi, int sync)
{
enum bdi_state bit;
wait_queue_head_t *wqh = &congestion_wqh[sync];
bit = sync ? BDI_sync_congested : BDI_async_congested;
if (test_and_clear_bit(bit, &bdi->state))
atomic_dec(&nr_bdi_congested[sync]);
smp_mb__after_atomic();
if (waitqueue_active(wqh))
wake_up(wqh);
}
EXPORT_SYMBOL(clear_bdi_congested);
void set_bdi_congested(struct backing_dev_info *bdi, int sync)
{
enum bdi_state bit;
bit = sync ? BDI_sync_congested : BDI_async_congested;
if (!test_and_set_bit(bit, &bdi->state))
atomic_inc(&nr_bdi_congested[sync]);
}
EXPORT_SYMBOL(set_bdi_congested);
/**
* congestion_wait - wait for a backing_dev to become uncongested
* @sync: SYNC or ASYNC IO
* @timeout: timeout in jiffies
*
* Waits for up to @timeout jiffies for a backing_dev (any backing_dev) to exit
* write congestion. If no backing_devs are congested then just wait for the
* next write to be completed.
*/
long congestion_wait(int sync, long timeout)
{
long ret;
unsigned long start = jiffies;
DEFINE_WAIT(wait);
wait_queue_head_t *wqh = &congestion_wqh[sync];
prepare_to_wait(wqh, &wait, TASK_UNINTERRUPTIBLE);
ret = io_schedule_timeout(timeout);
finish_wait(wqh, &wait);
trace_writeback_congestion_wait(jiffies_to_usecs(timeout),
jiffies_to_usecs(jiffies - start));
return ret;
}
EXPORT_SYMBOL(congestion_wait);
mm: per device dirty threshold Scale writeback cache per backing device, proportional to its writeout speed. By decoupling the BDI dirty thresholds a number of problems we currently have will go away, namely: - mutual interference starvation (for any number of BDIs); - deadlocks with stacked BDIs (loop, FUSE and local NFS mounts). It might be that all dirty pages are for a single BDI while other BDIs are idling. By giving each BDI a 'fair' share of the dirty limit, each one can have dirty pages outstanding and make progress. A global threshold also creates a deadlock for stacked BDIs; when A writes to B, and A generates enough dirty pages to get throttled, B will never start writeback until the dirty pages go away. Again, by giving each BDI its own 'independent' dirty limit, this problem is avoided. So the problem is to determine how to distribute the total dirty limit across the BDIs fairly and efficiently. A DBI that has a large dirty limit but does not have any dirty pages outstanding is a waste. What is done is to keep a floating proportion between the DBIs based on writeback completions. This way faster/more active devices get a larger share than slower/idle devices. [akpm@linux-foundation.org: fix warnings] [hugh@veritas.com: Fix occasional hang when a task couldn't get out of balance_dirty_pages] Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Signed-off-by: Hugh Dickins <hugh@veritas.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2007-10-17 00:25:50 -06:00
/**
* wait_iff_congested - Conditionally wait for a backing_dev to become uncongested or a zone to complete writes
* @zone: A zone to check if it is heavily congested
* @sync: SYNC or ASYNC IO
* @timeout: timeout in jiffies
*
* In the event of a congested backing_dev (any backing_dev) and the given
* @zone has experienced recent congestion, this waits for up to @timeout
* jiffies for either a BDI to exit congestion of the given @sync queue
* or a write to complete.
*
* In the absence of zone congestion, cond_resched() is called to yield
* the processor if necessary but otherwise does not sleep.
*
* The return value is 0 if the sleep is for the full timeout. Otherwise,
* it is the number of jiffies that were still remaining when the function
* returned. return_value == timeout implies the function did not sleep.
*/
long wait_iff_congested(struct zone *zone, int sync, long timeout)
{
long ret;
unsigned long start = jiffies;
DEFINE_WAIT(wait);
wait_queue_head_t *wqh = &congestion_wqh[sync];
/*
* If there is no congestion, or heavy congestion is not being
* encountered in the current zone, yield if necessary instead
* of sleeping on the congestion queue
*/
if (atomic_read(&nr_bdi_congested[sync]) == 0 ||
!test_bit(ZONE_CONGESTED, &zone->flags)) {
cond_resched();
/* In case we scheduled, work out time remaining */
ret = timeout - (jiffies - start);
if (ret < 0)
ret = 0;
goto out;
}
/* Sleep until uncongested or a write happens */
prepare_to_wait(wqh, &wait, TASK_UNINTERRUPTIBLE);
ret = io_schedule_timeout(timeout);
finish_wait(wqh, &wait);
out:
trace_writeback_wait_iff_congested(jiffies_to_usecs(timeout),
jiffies_to_usecs(jiffies - start));
return ret;
}
EXPORT_SYMBOL(wait_iff_congested);
int pdflush_proc_obsolete(struct ctl_table *table, int write,
void __user *buffer, size_t *lenp, loff_t *ppos)
{
char kbuf[] = "0\n";
if (*ppos || *lenp < sizeof(kbuf)) {
*lenp = 0;
return 0;
}
if (copy_to_user(buffer, kbuf, sizeof(kbuf)))
return -EFAULT;
printk_once(KERN_WARNING "%s exported in /proc is scheduled for removal\n",
table->procname);
*lenp = 2;
*ppos += *lenp;
return 2;
}