From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from userp2130.oracle.com ([156.151.31.86]:42152 "EHLO userp2130.oracle.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1753948AbdLVTEl (ORCPT ); Fri, 22 Dec 2017 14:04:41 -0500 Date: Fri, 22 Dec 2017 10:58:28 -0800 From: Liu Bo To: Nikolay Borisov Cc: linux-btrfs@vger.kernel.org Subject: Re: [PATCH 04/10] Btrfs: extent map selftest: buffered write vs dio read Message-ID: <20171222185828.GC21643@lim.localdomain> Reply-To: bo.li.liu@oracle.com References: <20171221224256.18196-1-bo.li.liu@oracle.com> <20171221224256.18196-5-bo.li.liu@oracle.com> <703352d3-8b80-1df1-e2ad-192e7e738a8d@suse.com> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii In-Reply-To: <703352d3-8b80-1df1-e2ad-192e7e738a8d@suse.com> Sender: linux-btrfs-owner@vger.kernel.org List-ID: On Fri, Dec 22, 2017 at 09:51:24AM +0200, Nikolay Borisov wrote: > > > On 22.12.2017 00:42, Liu Bo wrote: > > This test case simulates the racy situation of buffered write vs dio > > read, and see if btrfs_get_extent() would return -EEXIST. > > Isn't mixing dio/buffered IO on the same file (range?) considered > dangerous in any case? They are, but it is sometimes the way how applications work. thanks, -liubo > > > > > Signed-off-by: Liu Bo > > --- > > fs/btrfs/tests/extent-map-tests.c | 73 +++++++++++++++++++++++++++++++++++++++ > > 1 file changed, 73 insertions(+) > > > > diff --git a/fs/btrfs/tests/extent-map-tests.c b/fs/btrfs/tests/extent-map-tests.c > > index 0407396..2adf55f 100644 > > --- a/fs/btrfs/tests/extent-map-tests.c > > +++ b/fs/btrfs/tests/extent-map-tests.c > > @@ -181,6 +181,78 @@ static void test_case_2(struct extent_map_tree *em_tree) > > free_extent_map_tree(em_tree); > > } > > > > +static void __test_case_3(struct extent_map_tree *em_tree, u64 start) > > +{ > > + struct extent_map *em; > > + u64 len = SZ_4K; > > + int ret; > > + > > + em = alloc_extent_map(); > > + if (!em) > > + /* Skip this test on error. */ > > + return; > > + > > + /* Add [4K, 8K) */ > > + em->start = SZ_4K; > > + em->len = SZ_4K; > > + em->block_start = SZ_4K; > > + em->block_len = SZ_4K; > > + ret = add_extent_mapping(em_tree, em, 0); > > + ASSERT(ret == 0); > > + free_extent_map(em); > > + > > + em = alloc_extent_map(); > > + if (!em) > > + goto out; > > + > > + /* Add [0, 16K) */ > > + em->start = 0; > > + em->len = SZ_16K; > > + em->block_start = 0; > > + em->block_len = SZ_16K; > > + ret = btrfs_add_extent_mapping(em_tree, &em, start, len); > > + if (ret) > > + test_msg("case3 [0x%llx 0x%llx): ret %d\n", > > + start, start + len, ret); > > + /* > > + * Since bytes within em are contiguous, em->block_start is identical to > > + * em->start. > > + */ > > + if (em && > > + (start < em->start || start + len > extent_map_end(em) || > > + em->start != em->block_start || em->len != em->block_len)) > > + test_msg("case3 [0x%llx 0x%llx): ret %d em (start 0x%llx len 0x%llx block_start 0x%llx block_len 0x%llx)\n", > > + start, start + len, ret, em->start, em->len, > > + em->block_start, em->block_len); > > + free_extent_map(em); > > +out: > > + /* free memory */ > > + free_extent_map_tree(em_tree); > > +} > > + > > +/* > > + * Test scenario: > > + * > > + * Suppose that no extent map has been loaded into memory yet. > > + * There is a file extent [0, 16K), two jobs are running concurrently > > + * against it, t1 is buffered writing to [4K, 8K) and t2 is doing dio > > + * read from [0, 4K) or [8K, 12K) or [12K, 16K). > > + * > > + * t1 goes ahead of t2 and adds em [4K, 8K) into tree. > > + * > > + * t1 t2 > > + * cow_file_range() btrfs_get_extent() > > + * -> lookup_extent_mapping() > > + * -> add_extent_mapping() > > + * -> add_extent_mapping() > > + */ > > +static void test_case_3(struct extent_map_tree *em_tree) > > +{ > > + __test_case_3(em_tree, 0); > > + __test_case_3(em_tree, SZ_8K); > > + __test_case_3(em_tree, (12 * 1024ULL)); > > +} > > + > > int btrfs_test_extent_map() > > { > > struct extent_map_tree *em_tree; > > @@ -196,6 +268,7 @@ int btrfs_test_extent_map() > > > > test_case_1(em_tree); > > test_case_2(em_tree); > > + test_case_3(em_tree); > > > > kfree(em_tree); > > return 0; > >