From eb70d2226347efe3b7de771f8034166c2740418d Mon Sep 17 00:00:00 2001 From: Omar Sandoval Date: Mon, 2 Dec 2019 17:34:20 -0800 Subject: btrfs: remove unnecessary pg_offset assignments in __extent_writepage() We're initializing pg_offset to 0, setting it immediately, then reassigning it to 0 again after. The former became unnecessary in 211c17f51f46 ("Fix corners in writepage and btrfs_truncate_page"). The latter is a leftover that should've been removed in 40f765805f08 ("Btrfs: split up __extent_writepage to lower stack usage"). Remove both. Reviewed-by: Johannes Thumshirn Signed-off-by: Omar Sandoval Reviewed-by: David Sterba Signed-off-by: David Sterba --- fs/btrfs/extent_io.c | 4 +--- 1 file changed, 1 insertion(+), 3 deletions(-) (limited to 'fs/btrfs/extent_io.c') diff --git a/fs/btrfs/extent_io.c b/fs/btrfs/extent_io.c index 2f4802f405a2..b2d0418a4c55 100644 --- a/fs/btrfs/extent_io.c +++ b/fs/btrfs/extent_io.c @@ -3562,7 +3562,7 @@ static int __extent_writepage(struct page *page, struct writeback_control *wbc, u64 page_end = start + PAGE_SIZE - 1; int ret; int nr = 0; - size_t pg_offset = 0; + size_t pg_offset; loff_t i_size = i_size_read(inode); unsigned long end_index = i_size >> PAGE_SHIFT; unsigned long nr_written = 0; @@ -3591,8 +3591,6 @@ static int __extent_writepage(struct page *page, struct writeback_control *wbc, flush_dcache_page(page); } - pg_offset = 0; - set_page_extent_mapped(page); if (!epd->extent_locked) { -- cgit v1.2.3 From 169d2c875ec6461f6d18aaddbc1f30ba2b6b5a29 Mon Sep 17 00:00:00 2001 From: Omar Sandoval Date: Mon, 2 Dec 2019 17:34:21 -0800 Subject: btrfs: remove trivial goto label in __extent_writepage() Since 40f765805f08 ("Btrfs: split up __extent_writepage to lower stack usage"), done_unlocked is simply a return 0. Get rid of it. Mid-statement block returns don seem to make the code less readable here. Reviewed-by: Johannes Thumshirn Signed-off-by: Omar Sandoval Reviewed-by: David Sterba Signed-off-by: David Sterba --- fs/btrfs/extent_io.c | 7 ++----- 1 file changed, 2 insertions(+), 5 deletions(-) (limited to 'fs/btrfs/extent_io.c') diff --git a/fs/btrfs/extent_io.c b/fs/btrfs/extent_io.c index b2d0418a4c55..9475e81dc419 100644 --- a/fs/btrfs/extent_io.c +++ b/fs/btrfs/extent_io.c @@ -3596,7 +3596,7 @@ static int __extent_writepage(struct page *page, struct writeback_control *wbc, if (!epd->extent_locked) { ret = writepage_delalloc(inode, page, wbc, start, &nr_written); if (ret == 1) - goto done_unlocked; + return 0; if (ret) goto done; } @@ -3604,7 +3604,7 @@ static int __extent_writepage(struct page *page, struct writeback_control *wbc, ret = __extent_writepage_io(inode, page, wbc, epd, i_size, nr_written, &nr); if (ret == 1) - goto done_unlocked; + return 0; done: if (nr == 0) { @@ -3619,9 +3619,6 @@ done: unlock_page(page); ASSERT(ret <= 0); return ret; - -done_unlocked: - return 0; } void wait_on_extent_buffer_writeback(struct extent_buffer *eb) -- cgit v1.2.3 From f95d713b540487901871f90512f01f2fb3e6dfaa Mon Sep 17 00:00:00 2001 From: Omar Sandoval Date: Mon, 2 Dec 2019 17:34:22 -0800 Subject: btrfs: remove redundant i_size check in __extent_writepage_io() In __extent_writepage_io(), we check whether i_size <= page_offset(page). Note that if i_size < page_offset(page), then i_size >> PAGE_SHIFT < page->index. If i_size == page_offset(page), then i_size >> PAGE_SHIFT == page->index && offset_in_page(i_size) == 0. __extent_writepage() already has a check for these cases that returns without calling __extent_writepage_io(): end_index = i_size >> PAGE_SHIFT pg_offset = offset_in_page(i_size); if (page->index > end_index || (page->index == end_index && !pg_offset)) { page->mapping->a_ops->invalidatepage(page, 0, PAGE_SIZE); unlock_page(page); return 0; } Get rid of the one in __extent_writepage_io(), which was obsoleted in 211c17f51f46 ("Fix corners in writepage and btrfs_truncate_page"). Signed-off-by: Omar Sandoval Reviewed-by: David Sterba Signed-off-by: David Sterba --- fs/btrfs/extent_io.c | 6 ------ 1 file changed, 6 deletions(-) (limited to 'fs/btrfs/extent_io.c') diff --git a/fs/btrfs/extent_io.c b/fs/btrfs/extent_io.c index 9475e81dc419..00ddefcb54c8 100644 --- a/fs/btrfs/extent_io.c +++ b/fs/btrfs/extent_io.c @@ -3455,11 +3455,6 @@ static noinline_for_stack int __extent_writepage_io(struct inode *inode, update_nr_written(wbc, nr_written + 1); end = page_end; - if (i_size <= start) { - btrfs_writepage_endio_finish_ordered(page, start, page_end, 1); - goto done; - } - blocksize = inode->i_sb->s_blocksize; while (cur <= end) { @@ -3540,7 +3535,6 @@ static noinline_for_stack int __extent_writepage_io(struct inode *inode, pg_offset += iosize; nr++; } -done: *nr_ret = nr; return ret; } -- cgit v1.2.3 From 39b07b5d7072f8e9fd8cc2f840d3749f86699bbb Mon Sep 17 00:00:00 2001 From: Omar Sandoval Date: Mon, 2 Dec 2019 17:34:23 -0800 Subject: btrfs: drop create parameter to btrfs_get_extent() We only pass this as 1 from __extent_writepage_io(). The parameter basically means "pretend I didn't pass in a page". This is silly since we can simply not pass in the page. Get rid of the parameter from btrfs_get_extent(), and since it's used as a get_extent_t callback, remove it from get_extent_t and btree_get_extent(), neither of which need it. While we're here, let's document btrfs_get_extent(). Signed-off-by: Omar Sandoval Reviewed-by: David Sterba Signed-off-by: David Sterba --- fs/btrfs/extent_io.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) (limited to 'fs/btrfs/extent_io.c') diff --git a/fs/btrfs/extent_io.c b/fs/btrfs/extent_io.c index 00ddefcb54c8..bbfb102d65b8 100644 --- a/fs/btrfs/extent_io.c +++ b/fs/btrfs/extent_io.c @@ -3043,7 +3043,7 @@ __get_extent_map(struct inode *inode, struct page *page, size_t pg_offset, *em_cached = NULL; } - em = get_extent(BTRFS_I(inode), page, pg_offset, start, len, 0); + em = get_extent(BTRFS_I(inode), page, pg_offset, start, len); if (em_cached && !IS_ERR_OR_NULL(em)) { BUG_ON(*em_cached); refcount_inc(&em->refs); @@ -3466,8 +3466,8 @@ static noinline_for_stack int __extent_writepage_io(struct inode *inode, page_end, 1); break; } - em = btrfs_get_extent(BTRFS_I(inode), page, pg_offset, cur, - end - cur + 1, 1); + em = btrfs_get_extent(BTRFS_I(inode), NULL, 0, cur, + end - cur + 1); if (IS_ERR_OR_NULL(em)) { SetPageError(page); ret = PTR_ERR_OR_ZERO(em); -- cgit v1.2.3 From c8b04030c5ec006e161969d77c07fb03a8b67112 Mon Sep 17 00:00:00 2001 From: Omar Sandoval Date: Mon, 2 Dec 2019 17:34:24 -0800 Subject: btrfs: simplify compressed/inline check in __extent_writepage_io() Commit 7087a9d8db88 ("btrfs: Remove extent_io_ops::writepage_end_io_hook") left this logic in a confusing state. Simplify it. Reviewed-by: Johannes Thumshirn Signed-off-by: Omar Sandoval Reviewed-by: David Sterba Signed-off-by: David Sterba --- fs/btrfs/extent_io.c | 19 ++++--------------- 1 file changed, 4 insertions(+), 15 deletions(-) (limited to 'fs/btrfs/extent_io.c') diff --git a/fs/btrfs/extent_io.c b/fs/btrfs/extent_io.c index bbfb102d65b8..394beb474a69 100644 --- a/fs/btrfs/extent_io.c +++ b/fs/btrfs/extent_io.c @@ -3492,22 +3492,11 @@ static noinline_for_stack int __extent_writepage_io(struct inode *inode, */ if (compressed || block_start == EXTENT_MAP_HOLE || block_start == EXTENT_MAP_INLINE) { - /* - * end_io notification does not happen here for - * compressed extents - */ - if (!compressed) - btrfs_writepage_endio_finish_ordered(page, cur, - cur + iosize - 1, - 1); - else if (compressed) { - /* we don't want to end_page_writeback on - * a compressed extent. this happens - * elsewhere - */ + if (compressed) nr++; - } - + else + btrfs_writepage_endio_finish_ordered(page, cur, + cur + iosize - 1, 1); cur += iosize; pg_offset += iosize; continue; -- cgit v1.2.3 From 556755a8a99be8ca3cd9fbe36aaf9b3b0339a00d Mon Sep 17 00:00:00 2001 From: Josef Bacik Date: Fri, 3 Jan 2020 10:38:44 -0500 Subject: btrfs: fix improper setting of scanned for range cyclic write cache pages We noticed that we were having regular CG OOM kills in cases where there was still enough dirty pages to avoid OOM'ing. It turned out there's this corner case in btrfs's handling of range_cyclic where files that were being redirtied were not getting fully written out because of how we do range_cyclic writeback. We unconditionally were setting scanned = 1; the first time we found any pages in the inode. This isn't actually what we want, we want it to be set if we've scanned the entire file. For range_cyclic we could be starting in the middle or towards the end of the file, so we could write one page and then not write any of the other dirty pages in the file because we set scanned = 1. Fix this by not setting scanned = 1 if we find pages. The rules for setting scanned should be 1) !range_cyclic. In this case we have a specified range to write out. 2) range_cyclic && index == 0. In this case we've started at the beginning and there is no need to loop around a second time. 3) range_cyclic && we started at index > 0 and we've reached the end of the file without satisfying our nr_to_write. This patch fixes both of our writepages implementations to make sure these rules hold true. This fixed our over zealous CG OOMs in production. Fixes: d1310b2e0cd9 ("Btrfs: Split the extent_map code into two parts") Signed-off-by: Josef Bacik Reviewed-by: David Sterba [ add comment ] Signed-off-by: David Sterba --- fs/btrfs/extent_io.c | 12 ++++++++++-- 1 file changed, 10 insertions(+), 2 deletions(-) (limited to 'fs/btrfs/extent_io.c') diff --git a/fs/btrfs/extent_io.c b/fs/btrfs/extent_io.c index 394beb474a69..e2d30287e2d5 100644 --- a/fs/btrfs/extent_io.c +++ b/fs/btrfs/extent_io.c @@ -3919,6 +3919,11 @@ int btree_write_cache_pages(struct address_space *mapping, if (wbc->range_cyclic) { index = mapping->writeback_index; /* Start from prev offset */ end = -1; + /* + * Start from the beginning does not need to cycle over the + * range, mark it as scanned. + */ + scanned = (index == 0); } else { index = wbc->range_start >> PAGE_SHIFT; end = wbc->range_end >> PAGE_SHIFT; @@ -3936,7 +3941,6 @@ retry: tag))) { unsigned i; - scanned = 1; for (i = 0; i < nr_pages; i++) { struct page *page = pvec.pages[i]; @@ -4065,6 +4069,11 @@ static int extent_write_cache_pages(struct address_space *mapping, if (wbc->range_cyclic) { index = mapping->writeback_index; /* Start from prev offset */ end = -1; + /* + * Start from the beginning does not need to cycle over the + * range, mark it as scanned. + */ + scanned = (index == 0); } else { index = wbc->range_start >> PAGE_SHIFT; end = wbc->range_end >> PAGE_SHIFT; @@ -4098,7 +4107,6 @@ retry: &index, end, tag))) { unsigned i; - scanned = 1; for (i = 0; i < nr_pages; i++) { struct page *page = pvec.pages[i]; -- cgit v1.2.3