Received: from smtp.kernel.org (aws-us-west-2-korg-mail-alma10-1.taild15c8.ts.net [100.103.45.18]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by smtp.subspace.kernel.org (Postfix) with ESMTPS id E989C2D0C94; Sat, 12 Sep 2026 10:52:20 +0000 (UTC) Authentication-Results: smtp.subspace.kernel.org; arc=none smtp.client-ip=100.103.45.18 ARC-Seal:i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1789210342; cv=none; b=evA13q1wOyW6DY3Ww/4bOT+59P6k5lMRvtwbFgHaLMKCbe+OCIGQg2WG/t+xeyVQ7HaYVdfzMHpKG25yur4VeiPfjF+0gssKLSaIQwV4RaNN/WZuZTUFAO5WycXcI4qk0qHv4EltbJMGjwUR/fBqTlK7c69eeVZx82Mg7/8YsVI= ARC-Message-Signature:i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1789210342; c=relaxed/simple; bh=zUYAU3HduHujaPvMm04elUIOdXhyXPToDYIH+wBJ78A=; h=From:To:Cc:Subject:Date:Message-ID:In-Reply-To:References: MIME-Version; b=UsSAe16wxBliVH6MKTq4jcGgkWOblVGO7Uc3CCx12D7XsMi/CEWOlJZgGQHPtq+Tz5J5DQjTVdaJe7Ivc2EUxmh3MisN5V+O/ZDtTY/c8nAgMOCytCTpTP2rrpmVjZL48/PWLxZuQnocJDsnPQ/BqCEzmGFmulcr7dRHo5dhym4= ARC-Authentication-Results:i=1; smtp.subspace.kernel.org; dkim=pass (1024-bit key) header.d=linuxfoundation.org header.i=@linuxfoundation.org header.b=neyc82aI; arc=none smtp.client-ip=100.103.45.18 Authentication-Results: smtp.subspace.kernel.org; dkim=pass (1024-bit key) header.d=linuxfoundation.org header.i=@linuxfoundation.org header.b="neyc82aI" Received: by smtp.kernel.org (Postfix) with ESMTPSA id A710A1F000FF; Sat, 12 Sep 2026 10:52:19 +0000 (UTC) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=linuxfoundation.org; s=korg; t=1789210340; bh=e5kg/hJ4fI4iQdy32GuSjjxcPlqtKoj2b0ST+1d+/W4=; h=From:To:Cc:Subject:Date:In-Reply-To:References; b=neyc82aIjrHzjswvPV32OxQvu+gG2rsUb0SA3g39kHDQz2OcAQAHaXz35m0ptwiHj StF5okUrUqU28cqLwq5WasVT34Hu4ZpbIpDgFu/FBgfbpOnnZlf/zJ3NLfjtVjM92N gB73rUUSPBzpL6RtI89hHW900s8sqrDmqyDT4D74= From: Greg Kroah-Hartman To: stable@vger.kernel.org Cc: Greg Kroah-Hartman , patches@lists.linux.dev, Boris Burkov , Johannes Thumshirn , David Sterba , Sasha Levin Subject: [PATCH 6.18 1004/1518] btrfs: zoned: dont clobber the extent buffer when zeroing it out Date: Sat, 12 Sep 2026 08:52:53 +0200 Message-ID: <20260912065646.165329220@linuxfoundation.org> X-Mailer: git-send-email 2.55.0 In-Reply-To: <20260912065623.398859879@linuxfoundation.org> References: <20260912065623.398859879@linuxfoundation.org> User-Agent: quilt/0.69 X-stable: review X-Patchwork-Hint: ignore Precedence: bulk X-Mailing-List: patches@lists.linux.dev List-Id: List-Subscribe: List-Unsubscribe: MIME-Version: 1.0 Content-Transfer-Encoding: 8bit 6.18-stable review patch. If anyone has any objections, please let me know. ------------------ From: Johannes Thumshirn [ Upstream commit db4b9eefc8ee0bcaeee4d5e6a7313905f6a2fe7c ] On a zoned filesystem a freed-but-still-dirty tree block is written out as zeros (EXTENT_BUFFER_ZONED_ZEROOUT) only to keep the zone write pointer advancing. btree_csum_one_bio() implemented this by memzeroing the extent buffer's own folios before submission. That destroys the in-memory buffer while it may still be referenced. In particular btrfs_free_tree_block() can run on it afterwards and reads the header to add a delayed reference; once the header has been zeroed it frees bytenr 0 and corrupts the extent tree (the btrfs_header_bytenr(buf) != 0 ASSERT in btrfs_free_tree_block(), or an "unable to find ref" abort). It is flaky and reproduces under fsstress, e.g. generic/461 and generic/013. Write the zeros to disk from the shared zero page instead and leave the extent buffer content untouched, so any later reference - including the delayed reference from btrfs_free_tree_block() - still sees a valid header. end_bbio_meta_write() now clears writeback on the buffer's own folios, as the bio no longer carries them. Fixes: aa6313e6ff2b ("btrfs: zoned: don't clear dirty flag of extent buffer") Assisted-by: LLM (debugging, commit message) Reviewed-by: Boris Burkov Signed-off-by: Johannes Thumshirn Signed-off-by: David Sterba Signed-off-by: Sasha Levin --- fs/btrfs/disk-io.c | 13 +++++++------ fs/btrfs/extent_io.c | 31 ++++++++++++++++++++++++------- 2 files changed, 31 insertions(+), 13 deletions(-) diff --git a/fs/btrfs/disk-io.c b/fs/btrfs/disk-io.c index b956434208fe6..15b6a67d17ddc 100644 --- a/fs/btrfs/disk-io.c +++ b/fs/btrfs/disk-io.c @@ -278,14 +278,15 @@ int btree_csum_one_bio(struct btrfs_bio *bbio) return -EIO; /* - * If an extent_buffer is marked as EXTENT_BUFFER_ZONED_ZEROOUT, don't - * checksum it but zero-out its content. This is done to preserve - * ordering of I/O without unnecessarily writing out data. + * An extent_buffer marked EXTENT_BUFFER_ZONED_ZEROOUT is written out as + * zeros to preserve ordering of I/O without persisting the now + * unnecessary block. The bio is fed from the shared zero page (see + * write_one_eb()), so there is nothing to checksum here. Crucially, the + * buffer's own content is left intact: it may still be referenced, e.g. + * btrfs_free_tree_block() reads its header to add a delayed reference. */ - if (test_bit(EXTENT_BUFFER_ZONED_ZEROOUT, &eb->bflags)) { - memzero_extent_buffer(eb, 0, eb->len); + if (test_bit(EXTENT_BUFFER_ZONED_ZEROOUT, &eb->bflags)) return 0; - } if (WARN_ON_ONCE(found_start != eb->start)) return -EIO; diff --git a/fs/btrfs/extent_io.c b/fs/btrfs/extent_io.c index 80fc2d59bfeb8..90a6b161d7c11 100644 --- a/fs/btrfs/extent_io.c +++ b/fs/btrfs/extent_io.c @@ -2182,14 +2182,17 @@ static struct extent_buffer *find_extent_buffer_nolock( static void end_bbio_meta_write(struct btrfs_bio *bbio) { struct extent_buffer *eb = bbio->private; - struct folio_iter fi; if (bbio->bio.bi_status != BLK_STS_OK) set_btree_ioerr(eb); - bio_for_each_folio_all(fi, &bbio->bio) { - btrfs_meta_folio_clear_writeback(fi.folio, eb); - } + /* + * Clear writeback on the buffer's own folios. The bio may carry the + * shared zero page instead (EXTENT_BUFFER_ZONED_ZEROOUT), so iterate + * the extent buffer folios rather than the bio folios. + */ + for (int i = 0; i < num_extent_folios(eb); i++) + btrfs_meta_folio_clear_writeback(eb->folios[i], eb); buffer_tree_clear_mark(eb, PAGECACHE_TAG_WRITEBACK); clear_and_wake_up_bit(EXTENT_BUFFER_WRITEBACK, &eb->bflags); @@ -2230,7 +2233,8 @@ static noinline_for_stack void write_one_eb(struct extent_buffer *eb, struct btrfs_fs_info *fs_info = eb->fs_info; struct btrfs_bio *bbio; - prepare_eb_write(eb); + if (!test_bit(EXTENT_BUFFER_ZONED_ZEROOUT, &eb->bflags)) + prepare_eb_write(eb); bbio = btrfs_bio_alloc(INLINE_EXTENT_BUFFER_PAGES, REQ_OP_WRITE | REQ_META | wbc_to_write_flags(wbc), @@ -2250,8 +2254,21 @@ static noinline_for_stack void write_one_eb(struct extent_buffer *eb, btrfs_meta_folio_set_writeback(folio, eb); if (!folio_test_dirty(folio)) wbc->nr_to_write -= folio_nr_pages(folio); - bio_add_folio_nofail(&bbio->bio, folio, range_len, - offset_in_folio(folio, range_start)); + if (test_bit(EXTENT_BUFFER_ZONED_ZEROOUT, &eb->bflags)) { + u32 off = 0; + + while (off < range_len) { + u32 add = min_t(u32, PAGE_SIZE, range_len - off); + + bio_add_folio_nofail(&bbio->bio, + page_folio(ZERO_PAGE(0)), + add, 0); + off += add; + } + } else { + bio_add_folio_nofail(&bbio->bio, folio, range_len, + offset_in_folio(folio, range_start)); + } wbc_account_cgroup_owner(wbc, folio, range_len); folio_unlock(folio); } -- 2.53.0