From 2cb4b05e7647891b46b91c07c9a60304803d1688 Mon Sep 17 00:00:00 2001 From: Changli Gao Date: Tue, 29 Jun 2010 13:09:18 +0200 Subject: splice: direct_splice_actor() should not use pos in sd direct_splice_actor() shouldn't use sd->pos, as sd->pos is for file reading, file->f_pos should be used instead. Signed-off-by: Changli Gao Signed-off-by: Miklos Szeredi ---- fs/splice.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) Signed-off-by: Jens Axboe --- fs/splice.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/fs/splice.c b/fs/splice.c index 740e6b9faf7a..41900496d3bb 100644 --- a/fs/splice.c +++ b/fs/splice.c @@ -1282,7 +1282,8 @@ static int direct_splice_actor(struct pipe_inode_info *pipe, { struct file *file = sd->u.file; - return do_splice_from(pipe, file, &sd->pos, sd->total_len, sd->flags); + return do_splice_from(pipe, file, &file->f_pos, sd->total_len, + sd->flags); } /** -- cgit v1.2.3 From 19c9a49b432f245c6293508d164a4350f1f2c601 Mon Sep 17 00:00:00 2001 From: Changli Gao Date: Tue, 29 Jun 2010 13:10:36 +0200 Subject: splice: check f_mode for seekable file check f_mode for seekable file As a seekable file is allowed without a llseek function, so the old way isn't work any more. Signed-off-by: Changli Gao Signed-off-by: Miklos Szeredi ---- fs/splice.c | 6 ++---- 1 file changed, 2 insertions(+), 4 deletions(-) Signed-off-by: Jens Axboe --- fs/splice.c | 6 ++---- 1 file changed, 2 insertions(+), 4 deletions(-) diff --git a/fs/splice.c b/fs/splice.c index 41900496d3bb..efdbfece9932 100644 --- a/fs/splice.c +++ b/fs/splice.c @@ -1372,8 +1372,7 @@ static long do_splice(struct file *in, loff_t __user *off_in, if (off_in) return -ESPIPE; if (off_out) { - if (!out->f_op || !out->f_op->llseek || - out->f_op->llseek == no_llseek) + if (!(out->f_mode & FMODE_PWRITE)) return -EINVAL; if (copy_from_user(&offset, off_out, sizeof(loff_t))) return -EFAULT; @@ -1393,8 +1392,7 @@ static long do_splice(struct file *in, loff_t __user *off_in, if (off_out) return -ESPIPE; if (off_in) { - if (!in->f_op || !in->f_op->llseek || - in->f_op->llseek == no_llseek) + if (!(in->f_mode & FMODE_PREAD)) return -EINVAL; if (copy_from_user(&offset, off_in, sizeof(loff_t))) return -EFAULT; -- cgit v1.2.3 From 06d738fa9155ff16dba3d7e501ba4581d01a98cb Mon Sep 17 00:00:00 2001 From: Randy Dunlap Date: Thu, 1 Jul 2010 08:26:34 +0200 Subject: fs-writeback: fix kernel-doc warnings Fix kernel-doc to match the function's changed args. Warning(fs/fs-writeback.c:190): No description found for parameter 'args' Warning(fs/fs-writeback.c:190): Excess function parameter 'sb' description in 'bdi_queue_work_onstack' Signed-off-by: Randy Dunlap Cc: Christoph Hellwig Signed-off-by: Jens Axboe --- fs/fs-writeback.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/fs/fs-writeback.c b/fs/fs-writeback.c index 0609607d3955..6981e4b7c148 100644 --- a/fs/fs-writeback.c +++ b/fs/fs-writeback.c @@ -179,7 +179,7 @@ static void bdi_alloc_queue_work(struct backing_dev_info *bdi, /** * bdi_queue_work_onstack - start and wait for writeback - * @sb: write inodes from this super_block + * @args: parameters to control the work queue writeback * * Description: * This function initiates writeback and waits for the operation to -- cgit v1.2.3 From 9c3a8ee8a1d72c5c0d7fbdf426d80e270ddfa54c Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Thu, 10 Jun 2010 12:07:27 +0200 Subject: writeback: remove writeback_inodes_wbc This was just an odd wrapper around writeback_inodes_wb. Removing this also allows to get rid of the bdi member of struct writeback_control which was rather out of place there. Signed-off-by: Christoph Hellwig Signed-off-by: Jens Axboe --- fs/afs/write.c | 1 - fs/btrfs/extent_io.c | 2 -- fs/fs-writeback.c | 12 ++---------- include/linux/writeback.h | 5 ++--- mm/backing-dev.c | 3 +-- mm/page-writeback.c | 3 +-- 6 files changed, 6 insertions(+), 20 deletions(-) diff --git a/fs/afs/write.c b/fs/afs/write.c index 3dab9e9948d0..722743b152d8 100644 --- a/fs/afs/write.c +++ b/fs/afs/write.c @@ -680,7 +680,6 @@ int afs_writeback_all(struct afs_vnode *vnode) { struct address_space *mapping = vnode->vfs_inode.i_mapping; struct writeback_control wbc = { - .bdi = mapping->backing_dev_info, .sync_mode = WB_SYNC_ALL, .nr_to_write = LONG_MAX, .range_cyclic = 1, diff --git a/fs/btrfs/extent_io.c b/fs/btrfs/extent_io.c index a4080c21ec55..d74e6af9b53a 100644 --- a/fs/btrfs/extent_io.c +++ b/fs/btrfs/extent_io.c @@ -2594,7 +2594,6 @@ int extent_write_full_page(struct extent_io_tree *tree, struct page *page, .sync_io = wbc->sync_mode == WB_SYNC_ALL, }; struct writeback_control wbc_writepages = { - .bdi = wbc->bdi, .sync_mode = wbc->sync_mode, .older_than_this = NULL, .nr_to_write = 64, @@ -2628,7 +2627,6 @@ int extent_write_locked_range(struct extent_io_tree *tree, struct inode *inode, .sync_io = mode == WB_SYNC_ALL, }; struct writeback_control wbc_writepages = { - .bdi = inode->i_mapping->backing_dev_info, .sync_mode = mode, .older_than_this = NULL, .nr_to_write = nr_pages * 2, diff --git a/fs/fs-writeback.c b/fs/fs-writeback.c index 6981e4b7c148..94a602e98bb5 100644 --- a/fs/fs-writeback.c +++ b/fs/fs-writeback.c @@ -614,8 +614,8 @@ static int writeback_sb_inodes(struct super_block *sb, return 1; } -static void writeback_inodes_wb(struct bdi_writeback *wb, - struct writeback_control *wbc) +void writeback_inodes_wb(struct bdi_writeback *wb, + struct writeback_control *wbc) { int ret = 0; @@ -660,13 +660,6 @@ static void writeback_inodes_wb(struct bdi_writeback *wb, /* Leave any unwritten inodes on b_io */ } -void writeback_inodes_wbc(struct writeback_control *wbc) -{ - struct backing_dev_info *bdi = wbc->bdi; - - writeback_inodes_wb(&bdi->wb, wbc); -} - /* * The maximum number of pages to writeout in a single bdi flush/kupdate * operation. We do this so we don't hold I_SYNC against an inode for @@ -705,7 +698,6 @@ static long wb_writeback(struct bdi_writeback *wb, struct wb_writeback_args *args) { struct writeback_control wbc = { - .bdi = wb->bdi, .sb = args->sb, .sync_mode = args->sync_mode, .older_than_this = NULL, diff --git a/include/linux/writeback.h b/include/linux/writeback.h index d63ef8f9609f..f6756f6a610c 100644 --- a/include/linux/writeback.h +++ b/include/linux/writeback.h @@ -27,8 +27,6 @@ enum writeback_sync_modes { * in a manner such that unspecified fields are set to zero. */ struct writeback_control { - struct backing_dev_info *bdi; /* If !NULL, only write back this - queue */ struct super_block *sb; /* if !NULL, only write inodes from this super_block */ enum writeback_sync_modes sync_mode; @@ -66,7 +64,8 @@ int inode_wait(void *); void writeback_inodes_sb(struct super_block *); int writeback_inodes_sb_if_idle(struct super_block *); void sync_inodes_sb(struct super_block *); -void writeback_inodes_wbc(struct writeback_control *wbc); +void writeback_inodes_wb(struct bdi_writeback *wb, + struct writeback_control *wbc); long wb_do_writeback(struct bdi_writeback *wb, int force_wait); void wakeup_flusher_threads(long nr_pages); diff --git a/mm/backing-dev.c b/mm/backing-dev.c index 660a87a22511..6e0b09a1ec2c 100644 --- a/mm/backing-dev.c +++ b/mm/backing-dev.c @@ -340,14 +340,13 @@ int bdi_has_dirty_io(struct backing_dev_info *bdi) static void bdi_flush_io(struct backing_dev_info *bdi) { struct writeback_control wbc = { - .bdi = bdi, .sync_mode = WB_SYNC_NONE, .older_than_this = NULL, .range_cyclic = 1, .nr_to_write = 1024, }; - writeback_inodes_wbc(&wbc); + writeback_inodes_wb(&bdi->wb, &wbc); } /* diff --git a/mm/page-writeback.c b/mm/page-writeback.c index 54f28bd493d3..37498ef61548 100644 --- a/mm/page-writeback.c +++ b/mm/page-writeback.c @@ -495,7 +495,6 @@ static void balance_dirty_pages(struct address_space *mapping, for (;;) { struct writeback_control wbc = { - .bdi = bdi, .sync_mode = WB_SYNC_NONE, .older_than_this = NULL, .nr_to_write = write_chunk, @@ -537,7 +536,7 @@ static void balance_dirty_pages(struct address_space *mapping, * up. */ if (bdi_nr_reclaimable > bdi_thresh) { - writeback_inodes_wbc(&wbc); + writeback_inodes_wb(&bdi->wb, &wbc); pages_written += write_chunk - wbc.nr_to_write; get_dirty_limits(&background_thresh, &dirty_thresh, &bdi_thresh, bdi); -- cgit v1.2.3 From edadfb10ba35da7253541e4155aa92eff758ebe6 Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Thu, 10 Jun 2010 12:07:54 +0200 Subject: writeback: split writeback_inodes_wb The case where we have a superblock doesn't require a loop here as we scan over all inodes in writeback_sb_inodes. Split it out into a separate helper to make the code simpler. This also allows to get rid of the sb member in struct writeback_control, which was rather out of place there. Also update the comments in writeback_sb_inodes that explain the handling of inodes from wrong superblocks. Signed-off-by: Christoph Hellwig Signed-off-by: Jens Axboe --- fs/fs-writeback.c | 82 ++++++++++++++++++++++++++--------------------- include/linux/writeback.h | 2 -- 2 files changed, 46 insertions(+), 38 deletions(-) diff --git a/fs/fs-writeback.c b/fs/fs-writeback.c index 94a602e98bb5..8cc06d5432b5 100644 --- a/fs/fs-writeback.c +++ b/fs/fs-writeback.c @@ -554,29 +554,41 @@ static bool pin_sb_for_writeback(struct super_block *sb) /* * Write a portion of b_io inodes which belong to @sb. - * If @wbc->sb != NULL, then find and write all such + * + * If @only_this_sb is true, then find and write all such * inodes. Otherwise write only ones which go sequentially * in reverse order. + * * Return 1, if the caller writeback routine should be * interrupted. Otherwise return 0. */ -static int writeback_sb_inodes(struct super_block *sb, - struct bdi_writeback *wb, - struct writeback_control *wbc) +static int writeback_sb_inodes(struct super_block *sb, struct bdi_writeback *wb, + struct writeback_control *wbc, bool only_this_sb) { while (!list_empty(&wb->b_io)) { long pages_skipped; struct inode *inode = list_entry(wb->b_io.prev, struct inode, i_list); - if (wbc->sb && sb != inode->i_sb) { - /* super block given and doesn't - match, skip this inode */ - redirty_tail(inode); - continue; - } - if (sb != inode->i_sb) - /* finish with this superblock */ + + if (inode->i_sb != sb) { + if (only_this_sb) { + /* + * We only want to write back data for this + * superblock, move all inodes not belonging + * to it back onto the dirty list. + */ + redirty_tail(inode); + continue; + } + + /* + * The inode belongs to a different superblock. + * Bounce back to the caller to unpin this and + * pin the next superblock. + */ return 0; + } + if (inode->i_state & (I_NEW | I_WILL_FREE)) { requeue_io(inode); continue; @@ -629,29 +641,12 @@ void writeback_inodes_wb(struct bdi_writeback *wb, struct inode, i_list); struct super_block *sb = inode->i_sb; - if (wbc->sb) { - /* - * We are requested to write out inodes for a specific - * superblock. This means we already have s_umount - * taken by the caller which also waits for us to - * complete the writeout. - */ - if (sb != wbc->sb) { - redirty_tail(inode); - continue; - } - - WARN_ON(!rwsem_is_locked(&sb->s_umount)); - - ret = writeback_sb_inodes(sb, wb, wbc); - } else { - if (!pin_sb_for_writeback(sb)) { - requeue_io(inode); - continue; - } - ret = writeback_sb_inodes(sb, wb, wbc); - drop_super(sb); + if (!pin_sb_for_writeback(sb)) { + requeue_io(inode); + continue; } + ret = writeback_sb_inodes(sb, wb, wbc, false); + drop_super(sb); if (ret) break; @@ -660,6 +655,19 @@ void writeback_inodes_wb(struct bdi_writeback *wb, /* Leave any unwritten inodes on b_io */ } +static void __writeback_inodes_sb(struct super_block *sb, + struct bdi_writeback *wb, struct writeback_control *wbc) +{ + WARN_ON(!rwsem_is_locked(&sb->s_umount)); + + wbc->wb_start = jiffies; /* livelock avoidance */ + spin_lock(&inode_lock); + if (!wbc->for_kupdate || list_empty(&wb->b_io)) + queue_io(wb, wbc->older_than_this); + writeback_sb_inodes(sb, wb, wbc, true); + spin_unlock(&inode_lock); +} + /* * The maximum number of pages to writeout in a single bdi flush/kupdate * operation. We do this so we don't hold I_SYNC against an inode for @@ -698,7 +706,6 @@ static long wb_writeback(struct bdi_writeback *wb, struct wb_writeback_args *args) { struct writeback_control wbc = { - .sb = args->sb, .sync_mode = args->sync_mode, .older_than_this = NULL, .for_kupdate = args->for_kupdate, @@ -736,7 +743,10 @@ static long wb_writeback(struct bdi_writeback *wb, wbc.more_io = 0; wbc.nr_to_write = MAX_WRITEBACK_PAGES; wbc.pages_skipped = 0; - writeback_inodes_wb(wb, &wbc); + if (args->sb) + __writeback_inodes_sb(args->sb, wb, &wbc); + else + writeback_inodes_wb(wb, &wbc); args->nr_pages -= MAX_WRITEBACK_PAGES - wbc.nr_to_write; wrote += MAX_WRITEBACK_PAGES - wbc.nr_to_write; diff --git a/include/linux/writeback.h b/include/linux/writeback.h index f6756f6a610c..c24eca71e80c 100644 --- a/include/linux/writeback.h +++ b/include/linux/writeback.h @@ -27,8 +27,6 @@ enum writeback_sync_modes { * in a manner such that unspecified fields are set to zero. */ struct writeback_control { - struct super_block *sb; /* if !NULL, only write inodes from - this super_block */ enum writeback_sync_modes sync_mode; unsigned long *older_than_this; /* If !NULL, only write back inodes older than this */ -- cgit v1.2.3 From 83ba7b071f30f7c01f72518ad72d5cd203c27502 Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Tue, 6 Jul 2010 08:59:53 +0200 Subject: writeback: simplify the write back thread queue First remove items from work_list as soon as we start working on them. This means we don't have to track any pending or visited state and can get rid of all the RCU magic freeing the work items - we can simply free them once the operation has finished. Second use a real completion for tracking synchronous requests - if the caller sets the completion pointer we complete it, otherwise use it as a boolean indicator that we can free the work item directly. Third unify struct wb_writeback_args and struct bdi_work into a single data structure, wb_writeback_work. Previous we set all parameters into a struct wb_writeback_args, copied it into struct bdi_work, copied it again on the stack to use it there. Instead of just allocate one structure dynamically or on the stack and use it all the way through the stack. Signed-off-by: Christoph Hellwig Signed-off-by: Jens Axboe --- fs/fs-writeback.c | 253 ++++++++++++-------------------------------- include/linux/backing-dev.h | 2 - mm/backing-dev.c | 14 +-- 3 files changed, 72 insertions(+), 197 deletions(-) diff --git a/fs/fs-writeback.c b/fs/fs-writeback.c index 8cc06d5432b5..d5be1693ac93 100644 --- a/fs/fs-writeback.c +++ b/fs/fs-writeback.c @@ -38,43 +38,18 @@ int nr_pdflush_threads; /* * Passed into wb_writeback(), essentially a subset of writeback_control */ -struct wb_writeback_args { +struct wb_writeback_work { long nr_pages; struct super_block *sb; enum writeback_sync_modes sync_mode; unsigned int for_kupdate:1; unsigned int range_cyclic:1; unsigned int for_background:1; -}; -/* - * Work items for the bdi_writeback threads - */ -struct bdi_work { struct list_head list; /* pending work list */ - struct rcu_head rcu_head; /* for RCU free/clear of work */ - - unsigned long seen; /* threads that have seen this work */ - atomic_t pending; /* number of threads still to do work */ - - struct wb_writeback_args args; /* writeback arguments */ - - unsigned long state; /* flag bits, see WS_* */ -}; - -enum { - WS_INPROGRESS = 0, - WS_ONSTACK, + struct completion *done; /* set if the caller waits */ }; -static inline void bdi_work_init(struct bdi_work *work, - struct wb_writeback_args *args) -{ - INIT_RCU_HEAD(&work->rcu_head); - work->args = *args; - __set_bit(WS_INPROGRESS, &work->state); -} - /** * writeback_in_progress - determine whether there is writeback in progress * @bdi: the device's backing_dev_info structure. @@ -87,49 +62,11 @@ int writeback_in_progress(struct backing_dev_info *bdi) return !list_empty(&bdi->work_list); } -static void bdi_work_free(struct rcu_head *head) -{ - struct bdi_work *work = container_of(head, struct bdi_work, rcu_head); - - clear_bit(WS_INPROGRESS, &work->state); - smp_mb__after_clear_bit(); - wake_up_bit(&work->state, WS_INPROGRESS); - - if (!test_bit(WS_ONSTACK, &work->state)) - kfree(work); -} - -static void wb_clear_pending(struct bdi_writeback *wb, struct bdi_work *work) -{ - /* - * The caller has retrieved the work arguments from this work, - * drop our reference. If this is the last ref, delete and free it - */ - if (atomic_dec_and_test(&work->pending)) { - struct backing_dev_info *bdi = wb->bdi; - - spin_lock(&bdi->wb_lock); - list_del_rcu(&work->list); - spin_unlock(&bdi->wb_lock); - - call_rcu(&work->rcu_head, bdi_work_free); - } -} - -static void bdi_queue_work(struct backing_dev_info *bdi, struct bdi_work *work) +static void bdi_queue_work(struct backing_dev_info *bdi, + struct wb_writeback_work *work) { - work->seen = bdi->wb_mask; - BUG_ON(!work->seen); - atomic_set(&work->pending, bdi->wb_cnt); - BUG_ON(!bdi->wb_cnt); - - /* - * list_add_tail_rcu() contains the necessary barriers to - * make sure the above stores are seen before the item is - * noticed on the list - */ spin_lock(&bdi->wb_lock); - list_add_tail_rcu(&work->list, &bdi->work_list); + list_add_tail(&work->list, &bdi->work_list); spin_unlock(&bdi->wb_lock); /* @@ -146,55 +83,29 @@ static void bdi_queue_work(struct backing_dev_info *bdi, struct bdi_work *work) } } -/* - * Used for on-stack allocated work items. The caller needs to wait until - * the wb threads have acked the work before it's safe to continue. - */ -static void bdi_wait_on_work_done(struct bdi_work *work) -{ - wait_on_bit(&work->state, WS_INPROGRESS, bdi_sched_wait, - TASK_UNINTERRUPTIBLE); -} - -static void bdi_alloc_queue_work(struct backing_dev_info *bdi, - struct wb_writeback_args *args) +static void +__bdi_start_writeback(struct backing_dev_info *bdi, long nr_pages, + bool range_cyclic, bool for_background) { - struct bdi_work *work; + struct wb_writeback_work *work; /* * This is WB_SYNC_NONE writeback, so if allocation fails just * wakeup the thread for old dirty data writeback */ - work = kmalloc(sizeof(*work), GFP_ATOMIC); - if (work) { - bdi_work_init(work, args); - bdi_queue_work(bdi, work); - } else { - struct bdi_writeback *wb = &bdi->wb; - - if (wb->task) - wake_up_process(wb->task); + work = kzalloc(sizeof(*work), GFP_ATOMIC); + if (!work) { + if (bdi->wb.task) + wake_up_process(bdi->wb.task); + return; } -} -/** - * bdi_queue_work_onstack - start and wait for writeback - * @args: parameters to control the work queue writeback - * - * Description: - * This function initiates writeback and waits for the operation to - * complete. Callers must hold the sb s_umount semaphore for - * reading, to avoid having the super disappear before we are done. - */ -static void bdi_queue_work_onstack(struct wb_writeback_args *args) -{ - struct bdi_work work; + work->sync_mode = WB_SYNC_NONE; + work->nr_pages = nr_pages; + work->range_cyclic = range_cyclic; + work->for_background = for_background; - bdi_work_init(&work, args); - __set_bit(WS_ONSTACK, &work.state); - - bdi_queue_work(args->sb->s_bdi, &work); - bdi_wait_on_work_done(&work); + bdi_queue_work(bdi, work); } /** @@ -210,13 +121,7 @@ static void bdi_queue_work_onstack(struct wb_writeback_args *args) */ void bdi_start_writeback(struct backing_dev_info *bdi, long nr_pages) { - struct wb_writeback_args args = { - .sync_mode = WB_SYNC_NONE, - .nr_pages = nr_pages, - .range_cyclic = 1, - }; - - bdi_alloc_queue_work(bdi, &args); + __bdi_start_writeback(bdi, nr_pages, true, false); } /** @@ -230,13 +135,7 @@ void bdi_start_writeback(struct backing_dev_info *bdi, long nr_pages) */ void bdi_start_background_writeback(struct backing_dev_info *bdi) { - struct wb_writeback_args args = { - .sync_mode = WB_SYNC_NONE, - .nr_pages = LONG_MAX, - .for_background = 1, - .range_cyclic = 1, - }; - bdi_alloc_queue_work(bdi, &args); + __bdi_start_writeback(bdi, LONG_MAX, true, true); } /* @@ -703,14 +602,14 @@ static inline bool over_bground_thresh(void) * all dirty pages if they are all attached to "old" mappings. */ static long wb_writeback(struct bdi_writeback *wb, - struct wb_writeback_args *args) + struct wb_writeback_work *work) { struct writeback_control wbc = { - .sync_mode = args->sync_mode, + .sync_mode = work->sync_mode, .older_than_this = NULL, - .for_kupdate = args->for_kupdate, - .for_background = args->for_background, - .range_cyclic = args->range_cyclic, + .for_kupdate = work->for_kupdate, + .for_background = work->for_background, + .range_cyclic = work->range_cyclic, }; unsigned long oldest_jif; long wrote = 0; @@ -730,24 +629,24 @@ static long wb_writeback(struct bdi_writeback *wb, /* * Stop writeback when nr_pages has been consumed */ - if (args->nr_pages <= 0) + if (work->nr_pages <= 0) break; /* * For background writeout, stop when we are below the * background dirty threshold */ - if (args->for_background && !over_bground_thresh()) + if (work->for_background && !over_bground_thresh()) break; wbc.more_io = 0; wbc.nr_to_write = MAX_WRITEBACK_PAGES; wbc.pages_skipped = 0; - if (args->sb) - __writeback_inodes_sb(args->sb, wb, &wbc); + if (work->sb) + __writeback_inodes_sb(work->sb, wb, &wbc); else writeback_inodes_wb(wb, &wbc); - args->nr_pages -= MAX_WRITEBACK_PAGES - wbc.nr_to_write; + work->nr_pages -= MAX_WRITEBACK_PAGES - wbc.nr_to_write; wrote += MAX_WRITEBACK_PAGES - wbc.nr_to_write; /* @@ -783,31 +682,21 @@ static long wb_writeback(struct bdi_writeback *wb, } /* - * Return the next bdi_work struct that hasn't been processed by this - * wb thread yet. ->seen is initially set for each thread that exists - * for this device, when a thread first notices a piece of work it - * clears its bit. Depending on writeback type, the thread will notify - * completion on either receiving the work (WB_SYNC_NONE) or after - * it is done (WB_SYNC_ALL). + * Return the next wb_writeback_work struct that hasn't been processed yet. */ -static struct bdi_work *get_next_work_item(struct backing_dev_info *bdi, - struct bdi_writeback *wb) +static struct wb_writeback_work * +get_next_work_item(struct backing_dev_info *bdi, struct bdi_writeback *wb) { - struct bdi_work *work, *ret = NULL; - - rcu_read_lock(); - - list_for_each_entry_rcu(work, &bdi->work_list, list) { - if (!test_bit(wb->nr, &work->seen)) - continue; - clear_bit(wb->nr, &work->seen); + struct wb_writeback_work *work = NULL; - ret = work; - break; + spin_lock(&bdi->wb_lock); + if (!list_empty(&bdi->work_list)) { + work = list_entry(bdi->work_list.next, + struct wb_writeback_work, list); + list_del_init(&work->list); } - - rcu_read_unlock(); - return ret; + spin_unlock(&bdi->wb_lock); + return work; } static long wb_check_old_data_flush(struct bdi_writeback *wb) @@ -832,14 +721,14 @@ static long wb_check_old_data_flush(struct bdi_writeback *wb) (inodes_stat.nr_inodes - inodes_stat.nr_unused); if (nr_pages) { - struct wb_writeback_args args = { + struct wb_writeback_work work = { .nr_pages = nr_pages, .sync_mode = WB_SYNC_NONE, .for_kupdate = 1, .range_cyclic = 1, }; - return wb_writeback(wb, &args); + return wb_writeback(wb, &work); } return 0; @@ -851,33 +740,27 @@ static long wb_check_old_data_flush(struct bdi_writeback *wb) long wb_do_writeback(struct bdi_writeback *wb, int force_wait) { struct backing_dev_info *bdi = wb->bdi; - struct bdi_work *work; + struct wb_writeback_work *work; long wrote = 0; while ((work = get_next_work_item(bdi, wb)) != NULL) { - struct wb_writeback_args args = work->args; - /* * Override sync mode, in case we must wait for completion + * because this thread is exiting now. */ if (force_wait) - work->args.sync_mode = args.sync_mode = WB_SYNC_ALL; + work->sync_mode = WB_SYNC_ALL; - /* - * If this isn't a data integrity operation, just notify - * that we have seen this work and we are now starting it. - */ - if (!test_bit(WS_ONSTACK, &work->state)) - wb_clear_pending(wb, work); - - wrote += wb_writeback(wb, &args); + wrote += wb_writeback(wb, work); /* - * This is a data integrity writeback, so only do the - * notification when we have completed the work. + * Notify the caller of completion if this is a synchronous + * work item, otherwise just free it. */ - if (test_bit(WS_ONSTACK, &work->state)) - wb_clear_pending(wb, work); + if (work->done) + complete(work->done); + else + kfree(work); } /* @@ -940,14 +823,9 @@ int bdi_writeback_task(struct bdi_writeback *wb) void wakeup_flusher_threads(long nr_pages) { struct backing_dev_info *bdi; - struct wb_writeback_args args = { - .sync_mode = WB_SYNC_NONE, - }; - if (nr_pages) { - args.nr_pages = nr_pages; - } else { - args.nr_pages = global_page_state(NR_FILE_DIRTY) + + if (!nr_pages) { + nr_pages = global_page_state(NR_FILE_DIRTY) + global_page_state(NR_UNSTABLE_NFS); } @@ -955,7 +833,7 @@ void wakeup_flusher_threads(long nr_pages) list_for_each_entry_rcu(bdi, &bdi_list, bdi_list) { if (!bdi_has_dirty_io(bdi)) continue; - bdi_alloc_queue_work(bdi, &args); + __bdi_start_writeback(bdi, nr_pages, false, false); } rcu_read_unlock(); } @@ -1164,17 +1042,20 @@ void writeback_inodes_sb(struct super_block *sb) { unsigned long nr_dirty = global_page_state(NR_FILE_DIRTY); unsigned long nr_unstable = global_page_state(NR_UNSTABLE_NFS); - struct wb_writeback_args args = { + DECLARE_COMPLETION_ONSTACK(done); + struct wb_writeback_work work = { .sb = sb, .sync_mode = WB_SYNC_NONE, + .done = &done, }; WARN_ON(!rwsem_is_locked(&sb->s_umount)); - args.nr_pages = nr_dirty + nr_unstable + + work.nr_pages = nr_dirty + nr_unstable + (inodes_stat.nr_inodes - inodes_stat.nr_unused); - bdi_queue_work_onstack(&args); + bdi_queue_work(sb->s_bdi, &work); + wait_for_completion(&done); } EXPORT_SYMBOL(writeback_inodes_sb); @@ -1206,16 +1087,20 @@ EXPORT_SYMBOL(writeback_inodes_sb_if_idle); */ void sync_inodes_sb(struct super_block *sb) { - struct wb_writeback_args args = { + DECLARE_COMPLETION_ONSTACK(done); + struct wb_writeback_work work = { .sb = sb, .sync_mode = WB_SYNC_ALL, .nr_pages = LONG_MAX, .range_cyclic = 0, + .done = &done, }; WARN_ON(!rwsem_is_locked(&sb->s_umount)); - bdi_queue_work_onstack(&args); + bdi_queue_work(sb->s_bdi, &work); + wait_for_completion(&done); + wait_sb_inodes(sb); } EXPORT_SYMBOL(sync_inodes_sb); diff --git a/include/linux/backing-dev.h b/include/linux/backing-dev.h index 9ae2889096b6..e9aec0d099df 100644 --- a/include/linux/backing-dev.h +++ b/include/linux/backing-dev.h @@ -82,8 +82,6 @@ struct backing_dev_info { struct bdi_writeback wb; /* default writeback info for this bdi */ spinlock_t wb_lock; /* protects update side of wb_list */ struct list_head wb_list; /* the flusher threads hanging off this bdi */ - unsigned long wb_mask; /* bitmask of registered tasks */ - unsigned int wb_cnt; /* number of registered tasks */ struct list_head work_list; diff --git a/mm/backing-dev.c b/mm/backing-dev.c index 6e0b09a1ec2c..123bcef13e51 100644 --- a/mm/backing-dev.c +++ b/mm/backing-dev.c @@ -104,15 +104,13 @@ static int bdi_debug_stats_show(struct seq_file *m, void *v) "b_more_io: %8lu\n" "bdi_list: %8u\n" "state: %8lx\n" - "wb_mask: %8lx\n" - "wb_list: %8u\n" - "wb_cnt: %8u\n", + "wb_list: %8u\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), nr_wb, nr_dirty, nr_io, nr_more_io, - !list_empty(&bdi->bdi_list), bdi->state, bdi->wb_mask, - !list_empty(&bdi->wb_list), bdi->wb_cnt); + !list_empty(&bdi->bdi_list), bdi->state, + !list_empty(&bdi->wb_list)); #undef K return 0; @@ -674,12 +672,6 @@ int bdi_init(struct backing_dev_info *bdi) bdi_wb_init(&bdi->wb, bdi); - /* - * Just one thread support for now, hard code mask and count - */ - bdi->wb_mask = 1; - bdi->wb_cnt = 1; - for (i = 0; i < NR_BDI_STAT_ITEMS; i++) { err = percpu_counter_init(&bdi->bdi_stat[i], 0); if (err) -- cgit v1.2.3