From: Jan Kara <jack@suse.cz>
To: Tejun Heo <tj@kernel.org>
Cc: Jan Kara <jack@suse.cz>,
axboe@kernel.dk, linux-kernel@vger.kernel.org, hch@infradead.org,
hannes@cmpxchg.org, linux-fsdevel@vger.kernel.org,
vgoyal@redhat.com, lizefan@huawei.com, cgroups@vger.kernel.org,
linux-mm@kvack.org, mhocko@suse.cz, clm@fb.com,
fengguang.wu@intel.com, david@fromorbit.com, gthelen@google.com,
khlebnikov@yandex-team.ru
Subject: Re: [PATCH block/for-4.3] writeback: remove wb_writeback_work->single_wait/done
Date: Wed, 8 Jul 2015 10:24:53 +0200 [thread overview]
Message-ID: <20150708082453.GC725@quack.suse.cz> (raw)
In-Reply-To: <20150703221223.GH5273@mtj.duckdns.org>
On Fri 03-07-15 18:12:23, Tejun Heo wrote:
> Hello, Jan.
>
> So, something like the following. It depends on other changes so
> won't apply as-is. I'll repost it as part of a patchset once -rc1
> drops.
>
> Thanks!
>
> ------ 8< ------
> wb_writeback_work->single_wait/done are used for the wait mechanism
> for synchronous wb_work (wb_writeback_work) items which are issued
> when bdi_split_work_to_wbs() fails to allocate memory for asynchronous
> wb_work items; however, there's no reason to use a separate wait
> mechanism for this. bdi_split_work_to_wbs() can simply use on-stack
> fallback wb_work item and separate wb_completion to wait for it.
>
> This patch removes wb_work->single_wait/done and the related code and
> make bdi_split_work_to_wbs() use on-stack fallback wb_work and
> wb_completion instead.
>
> Signed-off-by: Tejun Heo <tj@kernel.org>
> Suggested-by: Jan Kara <jack@suse.cz>
Thanks! The patch looks good. You can add:
Reviewed-by: Jan Kara <jack@suse.com>
Honza
> ---
> fs/fs-writeback.c | 116 +++++++++++++-----------------------------------------
> 1 file changed, 30 insertions(+), 86 deletions(-)
>
> --- a/fs/fs-writeback.c
> +++ b/fs/fs-writeback.c
> @@ -53,8 +53,6 @@ struct wb_writeback_work {
> unsigned int for_background:1;
> unsigned int for_sync:1; /* sync(2) WB_SYNC_ALL writeback */
> unsigned int auto_free:1; /* free on completion */
> - unsigned int single_wait:1;
> - unsigned int single_done:1;
> enum wb_reason reason; /* why was writeback initiated? */
>
> struct list_head list; /* pending work list */
> @@ -181,11 +179,8 @@ static void wb_queue_work(struct bdi_wri
> trace_writeback_queue(wb->bdi, work);
>
> spin_lock_bh(&wb->work_lock);
> - if (!test_bit(WB_registered, &wb->state)) {
> - if (work->single_wait)
> - work->single_done = 1;
> + if (!test_bit(WB_registered, &wb->state))
> goto out_unlock;
> - }
> if (work->done)
> atomic_inc(&work->done->cnt);
> list_add_tail(&work->list, &wb->work_list);
> @@ -737,32 +732,6 @@ int inode_congested(struct inode *inode,
> EXPORT_SYMBOL_GPL(inode_congested);
>
> /**
> - * wb_wait_for_single_work - wait for completion of a single bdi_writeback_work
> - * @bdi: bdi the work item was issued to
> - * @work: work item to wait for
> - *
> - * Wait for the completion of @work which was issued to one of @bdi's
> - * bdi_writeback's. The caller must have set @work->single_wait before
> - * issuing it. This wait operates independently fo
> - * wb_wait_for_completion() and also disables automatic freeing of @work.
> - */
> -static void wb_wait_for_single_work(struct backing_dev_info *bdi,
> - struct wb_writeback_work *work)
> -{
> - if (WARN_ON_ONCE(!work->single_wait))
> - return;
> -
> - wait_event(bdi->wb_waitq, work->single_done);
> -
> - /*
> - * Paired with smp_wmb() in wb_do_writeback() and ensures that all
> - * modifications to @work prior to assertion of ->single_done is
> - * visible to the caller once this function returns.
> - */
> - smp_rmb();
> -}
> -
> -/**
> * wb_split_bdi_pages - split nr_pages to write according to bandwidth
> * @wb: target bdi_writeback to split @nr_pages to
> * @nr_pages: number of pages to write for the whole bdi
> @@ -791,38 +760,6 @@ static long wb_split_bdi_pages(struct bd
> }
>
> /**
> - * wb_clone_and_queue_work - clone a wb_writeback_work and issue it to a wb
> - * @wb: target bdi_writeback
> - * @base_work: source wb_writeback_work
> - *
> - * Try to make a clone of @base_work and issue it to @wb. If cloning
> - * succeeds, %true is returned; otherwise, @base_work is issued directly
> - * and %false is returned. In the latter case, the caller is required to
> - * wait for @base_work's completion using wb_wait_for_single_work().
> - *
> - * A clone is auto-freed on completion. @base_work never is.
> - */
> -static bool wb_clone_and_queue_work(struct bdi_writeback *wb,
> - struct wb_writeback_work *base_work)
> -{
> - struct wb_writeback_work *work;
> -
> - work = kmalloc(sizeof(*work), GFP_ATOMIC);
> - if (work) {
> - *work = *base_work;
> - work->auto_free = 1;
> - work->single_wait = 0;
> - } else {
> - work = base_work;
> - work->auto_free = 0;
> - work->single_wait = 1;
> - }
> - work->single_done = 0;
> - wb_queue_work(wb, work);
> - return work != base_work;
> -}
> -
> -/**
> * bdi_split_work_to_wbs - split a wb_writeback_work to all wb's of a bdi
> * @bdi: target backing_dev_info
> * @base_work: wb_writeback_work to issue
> @@ -837,7 +774,6 @@ static void bdi_split_work_to_wbs(struct
> struct wb_writeback_work *base_work,
> bool skip_if_busy)
> {
> - long nr_pages = base_work->nr_pages;
> int next_memcg_id = 0;
> struct bdi_writeback *wb;
> struct wb_iter iter;
> @@ -849,17 +785,39 @@ static void bdi_split_work_to_wbs(struct
> restart:
> rcu_read_lock();
> bdi_for_each_wb(wb, bdi, &iter, next_memcg_id) {
> + DEFINE_WB_COMPLETION_ONSTACK(fallback_work_done);
> + struct wb_writeback_work fallback_work;
> + struct wb_writeback_work *work;
> + long nr_pages;
> +
> if (!wb_has_dirty_io(wb) ||
> (skip_if_busy && writeback_in_progress(wb)))
> continue;
>
> - base_work->nr_pages = wb_split_bdi_pages(wb, nr_pages);
> - if (!wb_clone_and_queue_work(wb, base_work)) {
> - next_memcg_id = wb->memcg_css->id + 1;
> - rcu_read_unlock();
> - wb_wait_for_single_work(bdi, base_work);
> - goto restart;
> + nr_pages = wb_split_bdi_pages(wb, base_work->nr_pages);
> +
> + work = kmalloc(sizeof(*work), GFP_ATOMIC);
> + if (work) {
> + *work = *base_work;
> + work->nr_pages = nr_pages;
> + work->auto_free = 1;
> + wb_queue_work(wb, work);
> + continue;
> }
> +
> + /* alloc failed, execute synchronously using on-stack fallback */
> + work = &fallback_work;
> + *work = *base_work;
> + work->nr_pages = nr_pages;
> + work->auto_free = 0;
> + work->done = &fallback_work_done;
> +
> + wb_queue_work(wb, work);
> +
> + next_memcg_id = wb->memcg_css->id + 1;
> + rcu_read_unlock();
> + wb_wait_for_completion(bdi, &fallback_work_done);
> + goto restart;
> }
> rcu_read_unlock();
> }
> @@ -901,8 +859,6 @@ static void bdi_split_work_to_wbs(struct
> if (bdi_has_dirty_io(bdi) &&
> (!skip_if_busy || !writeback_in_progress(&bdi->wb))) {
> base_work->auto_free = 0;
> - base_work->single_wait = 0;
> - base_work->single_done = 0;
> wb_queue_work(&bdi->wb, base_work);
> }
> }
> @@ -1793,26 +1749,14 @@ static long wb_do_writeback(struct bdi_w
> set_bit(WB_writeback_running, &wb->state);
> while ((work = get_next_work_item(wb)) != NULL) {
> struct wb_completion *done = work->done;
> - bool need_wake_up = false;
>
> trace_writeback_exec(wb->bdi, work);
>
> wrote += wb_writeback(wb, work);
>
> - if (work->single_wait) {
> - WARN_ON_ONCE(work->auto_free);
> - /* paired w/ rmb in wb_wait_for_single_work() */
> - smp_wmb();
> - work->single_done = 1;
> - need_wake_up = true;
> - } else if (work->auto_free) {
> + if (work->auto_free)
> kfree(work);
> - }
> -
> if (done && atomic_dec_and_test(&done->cnt))
> - need_wake_up = true;
> -
> - if (need_wake_up)
> wake_up_all(&wb->bdi->wb_waitq);
> }
>
--
Jan Kara <jack@suse.cz>
SUSE Labs, CR
next prev parent reply other threads:[~2015-07-08 8:25 UTC|newest]
Thread overview: 131+ messages / expand[flat|nested] mbox.gz Atom feed top
2015-05-22 21:13 [PATCHSET 1/3 v4 block/for-4.2/core] writeback: cgroup writeback support Tejun Heo
2015-05-22 21:13 ` [PATCH 01/51] page_writeback: revive cancel_dirty_page() in a restricted form Tejun Heo
2015-05-22 21:13 ` [PATCH 02/51] memcg: add per cgroup dirty page accounting Tejun Heo
2015-05-22 21:13 ` [PATCH 03/51] blkcg: move block/blk-cgroup.h to include/linux/blk-cgroup.h Tejun Heo
2015-05-22 21:13 ` [PATCH 04/51] update !CONFIG_BLK_CGROUP dummies in include/linux/blk-cgroup.h Tejun Heo
2015-05-22 21:13 ` [PATCH 05/51] blkcg: always create the blkcg_gq for the root blkcg Tejun Heo
2015-05-22 21:13 ` [PATCH 06/51] memcg: add mem_cgroup_root_css Tejun Heo
2015-06-17 14:56 ` Michal Hocko
2015-06-17 18:25 ` Tejun Heo
2015-06-18 11:12 ` Michal Hocko
2015-06-18 17:49 ` Tejun Heo
2015-06-19 9:18 ` Michal Hocko
2015-06-19 15:17 ` Tejun Heo
2015-05-22 21:13 ` [PATCH 07/51] blkcg: add blkcg_root_css Tejun Heo
2015-05-22 21:13 ` [PATCH 08/51] cgroup, block: implement task_get_css() and use it in bio_associate_current() Tejun Heo
2015-05-22 21:13 ` [PATCH 09/51] blkcg: implement task_get_blkcg_css() Tejun Heo
2015-05-22 21:13 ` [PATCH 10/51] blkcg: implement bio_associate_blkcg() Tejun Heo
2015-05-22 21:13 ` [PATCH 11/51] memcg: implement mem_cgroup_css_from_page() Tejun Heo
2015-05-22 23:28 ` Johannes Weiner
2015-05-24 21:24 ` Tejun Heo
2015-05-27 12:58 ` Johannes Weiner
2015-05-27 16:13 ` [PATCH v2 " Tejun Heo
2015-05-27 17:09 ` Johannes Weiner
2015-05-27 17:48 ` Tejun Heo
2015-05-27 17:57 ` [PATCH v3 " Tejun Heo
2015-05-28 0:00 ` [PATCH v4 " Tejun Heo
2015-05-22 21:13 ` [PATCH 12/51] writeback: move backing_dev_info->state into bdi_writeback Tejun Heo
2015-05-22 21:13 ` [PATCH 13/51] writeback: move backing_dev_info->bdi_stat[] " Tejun Heo
2015-05-22 21:13 ` [PATCH 14/51] writeback: move bandwidth related fields from backing_dev_info " Tejun Heo
2015-05-22 21:13 ` [PATCH 15/51] writeback: s/bdi/wb/ in mm/page-writeback.c Tejun Heo
2015-05-22 21:13 ` [PATCH 16/51] writeback: move backing_dev_info->wb_lock and ->worklist into bdi_writeback Tejun Heo
2015-06-07 0:49 ` Sasha Levin
2015-06-08 5:57 ` [PATCH block/for-4.2-writeback] v9fs: fix error handling in v9fs_session_init() Tejun Heo
2015-06-08 15:10 ` Jens Axboe
2015-05-22 21:13 ` [PATCH 17/51] writeback: reorganize mm/backing-dev.c Tejun Heo
2015-05-22 21:13 ` [PATCH 18/51] writeback: separate out include/linux/backing-dev-defs.h Tejun Heo
2015-05-22 21:13 ` [PATCH 19/51] bdi: make inode_to_bdi() inline Tejun Heo
2015-06-30 6:47 ` Jan Kara
2015-05-22 21:13 ` [PATCH 20/51] writeback: add @gfp to wb_init() Tejun Heo
2015-05-22 21:13 ` [PATCH 21/51] bdi: separate out congested state into a separate struct Tejun Heo
2015-06-30 9:21 ` Jan Kara
2015-05-22 21:13 ` [PATCH 22/51] writeback: add {CONFIG|BDI_CAP|FS}_CGROUP_WRITEBACK Tejun Heo
2015-06-30 9:37 ` Jan Kara
2015-07-02 1:10 ` Tejun Heo
2015-07-03 10:49 ` Jan Kara
2015-07-03 17:14 ` Tejun Heo
2015-05-22 21:13 ` [PATCH 23/51] writeback: make backing_dev_info host cgroup-specific bdi_writebacks Tejun Heo
2015-06-30 10:14 ` Jan Kara
2015-05-22 21:13 ` [PATCH 24/51] writeback, blkcg: associate each blkcg_gq with the corresponding bdi_writeback_congested Tejun Heo
2015-06-30 9:08 ` Jan Kara
2015-05-22 21:13 ` [PATCH 25/51] writeback: attribute stats to the matching per-cgroup bdi_writeback Tejun Heo
2015-06-30 14:17 ` Jan Kara
2015-05-22 21:13 ` [PATCH 26/51] writeback: let balance_dirty_pages() work on the matching cgroup bdi_writeback Tejun Heo
2015-06-30 14:31 ` Jan Kara
2015-07-02 1:26 ` Tejun Heo
2015-05-22 21:13 ` [PATCH 27/51] writeback: make congestion functions per bdi_writeback Tejun Heo
2015-06-30 14:50 ` Jan Kara
2015-05-22 21:13 ` [PATCH 28/51] writeback, blkcg: restructure blk_{set|clear}_queue_congested() Tejun Heo
2015-06-30 15:02 ` Jan Kara
2015-07-02 1:38 ` Tejun Heo
2015-07-03 12:16 ` Jan Kara
2015-05-22 21:13 ` [PATCH 29/51] writeback, blkcg: propagate non-root blkcg congestion state Tejun Heo
2015-06-30 15:03 ` Jan Kara
2015-05-22 21:13 ` [PATCH 30/51] writeback: implement and use inode_congested() Tejun Heo
2015-06-30 15:21 ` Jan Kara
2015-07-02 1:46 ` Tejun Heo
2015-07-03 12:17 ` Jan Kara
2015-07-03 17:07 ` Tejun Heo
2015-07-04 15:12 ` [PATCH block/for-4.3] writeback: explain why @inode is allowed to be NULL for inode_congested() Tejun Heo
2015-07-08 8:12 ` Jan Kara
2015-05-22 21:13 ` [PATCH 31/51] writeback: implement WB_has_dirty_io wb_state flag Tejun Heo
2015-06-30 15:42 ` Jan Kara
2015-05-22 21:13 ` [PATCH 32/51] writeback: implement backing_dev_info->tot_write_bandwidth Tejun Heo
2015-06-30 16:14 ` Jan Kara
2015-06-30 16:42 ` Jan Kara
2015-05-22 21:13 ` [PATCH 33/51] writeback: make bdi_has_dirty_io() take multiple bdi_writeback's into account Tejun Heo
2015-06-30 16:48 ` Jan Kara
2015-07-02 2:01 ` Tejun Heo
2015-05-22 21:13 ` [PATCH 34/51] writeback: don't issue wb_writeback_work if clean Tejun Heo
2015-06-30 16:18 ` Jan Kara
2015-05-22 21:13 ` [PATCH 35/51] writeback: make bdi->min/max_ratio handling cgroup writeback aware Tejun Heo
2015-07-01 7:00 ` Jan Kara
2015-05-22 21:13 ` [PATCH 36/51] writeback: implement bdi_for_each_wb() Tejun Heo
2015-07-01 7:27 ` Jan Kara
2015-07-02 2:22 ` Tejun Heo
2015-07-03 12:26 ` Jan Kara
2015-07-03 17:06 ` Tejun Heo
2015-05-22 21:13 ` [PATCH 37/51] writeback: remove bdi_start_writeback() Tejun Heo
2015-07-01 7:30 ` Jan Kara
2015-05-22 21:13 ` [PATCH 38/51] writeback: make laptop_mode_timer_fn() handle multiple bdi_writeback's Tejun Heo
2015-07-01 7:32 ` Jan Kara
2015-05-22 21:13 ` [PATCH 39/51] writeback: make writeback_in_progress() take bdi_writeback instead of backing_dev_info Tejun Heo
2015-07-01 7:47 ` Jan Kara
2015-07-02 2:28 ` Tejun Heo
2015-05-22 21:13 ` [PATCH 40/51] writeback: make bdi_start_background_writeback() " Tejun Heo
2015-07-01 7:50 ` Jan Kara
2015-07-02 2:29 ` Tejun Heo
2015-07-06 19:36 ` [PATCH block/for-4.3] writeback: update writeback tracepoints to report cgroup Tejun Heo
2015-07-08 8:17 ` Jan Kara
2015-05-22 21:13 ` [PATCH 41/51] writeback: make wakeup_flusher_threads() handle multiple bdi_writeback's Tejun Heo
2015-07-01 8:15 ` Jan Kara
2015-07-02 2:37 ` Tejun Heo
2015-07-03 13:02 ` Jan Kara
2015-07-03 16:33 ` Tejun Heo
2015-05-22 21:13 ` [PATCH 42/51] writeback: make wakeup_dirtytime_writeback() " Tejun Heo
2015-07-01 8:20 ` Jan Kara
2015-05-22 21:13 ` [PATCH 43/51] writeback: add wb_writeback_work->auto_free Tejun Heo
2015-05-22 21:13 ` [PATCH 44/51] writeback: implement bdi_wait_for_completion() Tejun Heo
2015-07-01 16:04 ` Jan Kara
2015-07-02 3:06 ` Tejun Heo
2015-07-03 12:36 ` Jan Kara
2015-07-03 17:02 ` Tejun Heo
2015-07-01 16:09 ` Jan Kara
2015-07-02 3:01 ` Tejun Heo
2015-05-22 21:13 ` [PATCH 45/51] writeback: implement wb_wait_for_single_work() Tejun Heo
2015-07-01 19:07 ` Jan Kara
2015-07-02 3:07 ` Tejun Heo
2015-07-03 22:12 ` [PATCH block/for-4.3] writeback: remove wb_writeback_work->single_wait/done Tejun Heo
2015-07-08 8:24 ` Jan Kara [this message]
2015-05-22 21:14 ` [PATCH 46/51] writeback: restructure try_writeback_inodes_sb[_nr]() Tejun Heo
2015-05-22 21:14 ` [PATCH 47/51] writeback: make writeback initiation functions handle multiple bdi_writeback's Tejun Heo
2015-05-22 21:14 ` [PATCH 48/51] writeback: dirty inodes against their matching cgroup bdi_writeback's Tejun Heo
2015-07-01 19:16 ` Jan Kara
2015-05-22 21:14 ` [PATCH 49/51] buffer, writeback: make __block_write_full_page() honor cgroup writeback Tejun Heo
2015-07-01 19:21 ` Jan Kara
2015-07-01 19:28 ` Jan Kara
2015-05-22 21:14 ` [PATCH 50/51] mpage: make __mpage_writepage() " Tejun Heo
2015-07-01 19:26 ` Jan Kara
2015-05-22 21:14 ` [PATCH 51/51] ext2: enable cgroup writeback support Tejun Heo
2015-07-01 19:29 ` Jan Kara
2015-07-02 3:08 ` Tejun Heo
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20150708082453.GC725@quack.suse.cz \
--to=jack@suse.cz \
--cc=axboe@kernel.dk \
--cc=cgroups@vger.kernel.org \
--cc=clm@fb.com \
--cc=david@fromorbit.com \
--cc=fengguang.wu@intel.com \
--cc=gthelen@google.com \
--cc=hannes@cmpxchg.org \
--cc=hch@infradead.org \
--cc=khlebnikov@yandex-team.ru \
--cc=linux-fsdevel@vger.kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-mm@kvack.org \
--cc=lizefan@huawei.com \
--cc=mhocko@suse.cz \
--cc=tj@kernel.org \
--cc=vgoyal@redhat.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).