| Message ID | 20190103085005.32053-8-nborisov@suse.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
| Series | More misc fixes | expand |
On Thu, Jan 03, 2019 at 10:50:05AM +0200, Nikolay Borisov wrote: > Add a couple of comments regarding the logic flow in shrink_delalloc. > Then, cease using max_reclaim as a temporary variable when calculating > nr_pages. Finally give max_reclaim a more becoming name, which > uneqivocally shows at what this variable really holds. No functional > changes. > > Signed-off-by: Nikolay Borisov <nborisov@suse.com> > --- > fs/btrfs/extent-tree.c | 33 ++++++++++++++++++++++----------- > 1 file changed, 22 insertions(+), 11 deletions(-) > > diff --git a/fs/btrfs/extent-tree.c b/fs/btrfs/extent-tree.c > index b15afeae16df..615441f4d458 100644 > --- a/fs/btrfs/extent-tree.c > +++ b/fs/btrfs/extent-tree.c > @@ -4743,7 +4743,7 @@ static void shrink_delalloc(struct btrfs_fs_info *fs_info, u64 to_reclaim, > struct btrfs_space_info *space_info; > struct btrfs_trans_handle *trans; > u64 delalloc_bytes; > - u64 max_reclaim; > + u64 async_pages; This could be int, as it's compared to an atomic_t, otherwise ok. > u64 items; > long time_left; > unsigned long nr_pages; > @@ -4768,25 +4768,36 @@ static void shrink_delalloc(struct btrfs_fs_info *fs_info, u64 to_reclaim, > > loops = 0; > while (delalloc_bytes && loops < 3) { > - max_reclaim = min(delalloc_bytes, to_reclaim); > - nr_pages = max_reclaim >> PAGE_SHIFT; > + nr_pages = min(delalloc_bytes, to_reclaim) >> PAGE_SHIFT; > + > + /* > + * Triggers inode writeback for up to nr_pages. This will invoke > + * ->writepages callback and trigger delalloc filling > + * (btrfs_run_delalloc_range()). > + */ > btrfs_writeback_inodes_sb_nr(fs_info, nr_pages, items); > + > /* > - * We need to wait for the async pages to actually start before > - * we do anything. > + * We need to wait for the compressed pages to start before > + * we continue. > */ > - max_reclaim = atomic_read(&fs_info->async_delalloc_pages); > - if (!max_reclaim) > + async_pages = atomic_read(&fs_info->async_delalloc_pages); > + if (!async_pages) > goto skip_async; > > - if (max_reclaim <= nr_pages) > - max_reclaim = 0; > + /* > + * Calculate how many compressed pages we want to be written > + * before we continue. I.e if there are more async pages than we > + * require wait_event will wait until nr_pages are written. > + */ > + if (async_pages <= nr_pages) > + async_pages = 0; > else > - max_reclaim -= nr_pages; > + async_pages -= nr_pages; > > wait_event(fs_info->async_submit_wait, > atomic_read(&fs_info->async_delalloc_pages) <= > - (int)max_reclaim); > + (int)async_pages); > skip_async: > spin_lock(&space_info->lock); > if (list_empty(&space_info->tickets) && > -- > 2.17.1
On Fri, Jan 04, 2019 at 04:35:43PM +0100, David Sterba wrote: > > @@ -4743,7 +4743,7 @@ static void shrink_delalloc(struct btrfs_fs_info *fs_info, u64 to_reclaim, > > struct btrfs_space_info *space_info; > > struct btrfs_trans_handle *trans; > > u64 delalloc_bytes; > > - u64 max_reclaim; > > + u64 async_pages; > > This could be int, as it's compared to an atomic_t, otherwise ok. As this is a different change, please send a separate patch, I've applied the rename to misc-next.
diff --git a/fs/btrfs/extent-tree.c b/fs/btrfs/extent-tree.c index b15afeae16df..615441f4d458 100644 --- a/fs/btrfs/extent-tree.c +++ b/fs/btrfs/extent-tree.c @@ -4743,7 +4743,7 @@ static void shrink_delalloc(struct btrfs_fs_info *fs_info, u64 to_reclaim, struct btrfs_space_info *space_info; struct btrfs_trans_handle *trans; u64 delalloc_bytes; - u64 max_reclaim; + u64 async_pages; u64 items; long time_left; unsigned long nr_pages; @@ -4768,25 +4768,36 @@ static void shrink_delalloc(struct btrfs_fs_info *fs_info, u64 to_reclaim, loops = 0; while (delalloc_bytes && loops < 3) { - max_reclaim = min(delalloc_bytes, to_reclaim); - nr_pages = max_reclaim >> PAGE_SHIFT; + nr_pages = min(delalloc_bytes, to_reclaim) >> PAGE_SHIFT; + + /* + * Triggers inode writeback for up to nr_pages. This will invoke + * ->writepages callback and trigger delalloc filling + * (btrfs_run_delalloc_range()). + */ btrfs_writeback_inodes_sb_nr(fs_info, nr_pages, items); + /* - * We need to wait for the async pages to actually start before - * we do anything. + * We need to wait for the compressed pages to start before + * we continue. */ - max_reclaim = atomic_read(&fs_info->async_delalloc_pages); - if (!max_reclaim) + async_pages = atomic_read(&fs_info->async_delalloc_pages); + if (!async_pages) goto skip_async; - if (max_reclaim <= nr_pages) - max_reclaim = 0; + /* + * Calculate how many compressed pages we want to be written + * before we continue. I.e if there are more async pages than we + * require wait_event will wait until nr_pages are written. + */ + if (async_pages <= nr_pages) + async_pages = 0; else - max_reclaim -= nr_pages; + async_pages -= nr_pages; wait_event(fs_info->async_submit_wait, atomic_read(&fs_info->async_delalloc_pages) <= - (int)max_reclaim); + (int)async_pages); skip_async: spin_lock(&space_info->lock); if (list_empty(&space_info->tickets) &&
Add a couple of comments regarding the logic flow in shrink_delalloc. Then, cease using max_reclaim as a temporary variable when calculating nr_pages. Finally give max_reclaim a more becoming name, which uneqivocally shows at what this variable really holds. No functional changes. Signed-off-by: Nikolay Borisov <nborisov@suse.com> --- fs/btrfs/extent-tree.c | 33 ++++++++++++++++++++++----------- 1 file changed, 22 insertions(+), 11 deletions(-)