* [PATCH] Btrfs: do not use extent commit root for sending
@ 2014-01-12 13:38 Wang Shilong
2014-01-12 13:38 ` [PATCH 1/3] Btrfs: fix missing skinny metadata check in scrub_stripe() Wang Shilong
` (2 more replies)
0 siblings, 3 replies; 6+ messages in thread
From: Wang Shilong @ 2014-01-12 13:38 UTC (permalink / raw)
To: linux-btrfs
From: Wang Shilong <wangsl.fnst@cn.fujitsu.com>
Now we have kicked off transaction from btrfs send, it is not safe
that we use extent commit root to search.
I happended to catch this problem when running sending and snapshot
in my desktop.
Signed-off-by: Wang Shilong <wangsl.fnst@cn.fujitsu.com>
---
fs/btrfs/send.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/fs/btrfs/send.c b/fs/btrfs/send.c
index 591063d..e159df1 100644
--- a/fs/btrfs/send.c
+++ b/fs/btrfs/send.c
@@ -1225,7 +1225,7 @@ static int find_extent_clone(struct send_ctx *sctx,
extent_item_pos = logical - found_key.objectid;
ret = iterate_extent_inodes(sctx->send_root->fs_info,
- found_key.objectid, extent_item_pos, 1,
+ found_key.objectid, extent_item_pos, 0,
__iterate_backrefs, backref_ctx);
if (ret < 0)
--
1.8.4
^ permalink raw reply related [flat|nested] 6+ messages in thread* [PATCH 1/3] Btrfs: fix missing skinny metadata check in scrub_stripe() 2014-01-12 13:38 [PATCH] Btrfs: do not use extent commit root for sending Wang Shilong @ 2014-01-12 13:38 ` Wang Shilong 2014-01-12 13:38 ` [PATCH 2/3] Btrfs: fix to search previous metadata extent item since skinny metadata Wang Shilong 2014-01-12 13:38 ` [PATCH 3/3] Btrfs: fix missing inline refs when walking backrefs Wang Shilong 2 siblings, 0 replies; 6+ messages in thread From: Wang Shilong @ 2014-01-12 13:38 UTC (permalink / raw) To: linux-btrfs From: Wang Shilong <wangsl.fnst@cn.fujitsu.com> Check if we support skinny metadata firstly and fix to use right type to search. Signed-off-by: Wang Shilong <wangsl.fnst@cn.fujitsu.com> --- fs/btrfs/scrub.c | 5 ++++- 1 file changed, 4 insertions(+), 1 deletion(-) diff --git a/fs/btrfs/scrub.c b/fs/btrfs/scrub.c index 7806e2c..e0677e4 100644 --- a/fs/btrfs/scrub.c +++ b/fs/btrfs/scrub.c @@ -2373,8 +2373,11 @@ static noinline_for_stack int scrub_stripe(struct scrub_ctx *sctx, scrub_blocked_if_needed(fs_info); } + if (btrfs_fs_incompat(fs_info, SKINNY_METADATA)) + key.type = BTRFS_METADATA_ITEM_KEY; + else + key.type = BTRFS_EXTENT_ITEM_KEY; key.objectid = logical; - key.type = BTRFS_EXTENT_ITEM_KEY; key.offset = (u64)-1; ret = btrfs_search_slot(NULL, root, &key, path, 0, 0); -- 1.8.4 ^ permalink raw reply related [flat|nested] 6+ messages in thread
* [PATCH 2/3] Btrfs: fix to search previous metadata extent item since skinny metadata 2014-01-12 13:38 [PATCH] Btrfs: do not use extent commit root for sending Wang Shilong 2014-01-12 13:38 ` [PATCH 1/3] Btrfs: fix missing skinny metadata check in scrub_stripe() Wang Shilong @ 2014-01-12 13:38 ` Wang Shilong 2014-01-12 13:38 ` [PATCH 3/3] Btrfs: fix missing inline refs when walking backrefs Wang Shilong 2 siblings, 0 replies; 6+ messages in thread From: Wang Shilong @ 2014-01-12 13:38 UTC (permalink / raw) To: linux-btrfs From: Wang Shilong <wangsl.fnst@cn.fujitsu.com> There is a bug that using btrfs_previous_item() to search metadata extent item. This is because in btrfs_previous_item(), we need type match, however, since skinny metada was introduced by josef, we may mix this two types. So just use btrfs_previous_item() is not working right. To keep btrfs_previous_item() like normal tree search, i introduce another function btrfs_previous_extent_item(). Signed-off-by: Wang Shilong <wangsl.fnst@cn.fujitsu.com> --- fs/btrfs/backref.c | 3 +-- fs/btrfs/ctree.c | 43 +++++++++++++++++++++++++++++++++++++++++++ fs/btrfs/ctree.h | 2 ++ fs/btrfs/scrub.c | 3 +-- 4 files changed, 47 insertions(+), 4 deletions(-) diff --git a/fs/btrfs/backref.c b/fs/btrfs/backref.c index 835b6c9..964679c 100644 --- a/fs/btrfs/backref.c +++ b/fs/btrfs/backref.c @@ -1303,8 +1303,7 @@ int extent_from_logical(struct btrfs_fs_info *fs_info, u64 logical, ret = btrfs_search_slot(NULL, fs_info->extent_root, &key, path, 0, 0); if (ret < 0) return ret; - ret = btrfs_previous_item(fs_info->extent_root, path, - 0, BTRFS_EXTENT_ITEM_KEY); + ret = btrfs_previous_extent_item(fs_info->extent_root, path, 0); if (ret < 0) return ret; diff --git a/fs/btrfs/ctree.c b/fs/btrfs/ctree.c index 9e9de68..30f5b11 100644 --- a/fs/btrfs/ctree.c +++ b/fs/btrfs/ctree.c @@ -5955,3 +5955,46 @@ int btrfs_previous_item(struct btrfs_root *root, } return 1; } + +/* + * search in extent tree to find a previous Metadata/Data extent item with + * min objecitd. + * + * returns 0 if something is found, 1 if nothing was found and < 0 on error + */ +int btrfs_previous_extent_item(struct btrfs_root *root, + struct btrfs_path *path, u64 min_objectid) +{ + struct btrfs_key found_key; + struct extent_buffer *leaf; + u32 nritems; + int ret; + + while (1) { + if (path->slots[0] == 0) { + btrfs_set_path_blocking(path); + ret = btrfs_prev_leaf(root, path); + if (ret != 0) + return ret; + } else { + path->slots[0]--; + } + leaf = path->nodes[0]; + nritems = btrfs_header_nritems(leaf); + if (nritems == 0) + return 1; + if (path->slots[0] == nritems) + path->slots[0]--; + + btrfs_item_key_to_cpu(leaf, &found_key, path->slots[0]); + if (found_key.objectid < min_objectid) + break; + if (found_key.type == BTRFS_EXTENT_ITEM_KEY || + found_key.type == BTRFS_METADATA_ITEM_KEY) + return 0; + if (found_key.objectid == min_objectid && + found_key.type < BTRFS_EXTENT_ITEM_KEY) + break; + } + return 1; +} diff --git a/fs/btrfs/ctree.h b/fs/btrfs/ctree.h index 3cebb4a..81f8190 100644 --- a/fs/btrfs/ctree.h +++ b/fs/btrfs/ctree.h @@ -3331,6 +3331,8 @@ int btrfs_comp_cpu_keys(struct btrfs_key *k1, struct btrfs_key *k2); int btrfs_previous_item(struct btrfs_root *root, struct btrfs_path *path, u64 min_objectid, int type); +int btrfs_previous_extent_item(struct btrfs_root *root, + struct btrfs_path *path, u64 min_objectid); void btrfs_set_item_key_safe(struct btrfs_root *root, struct btrfs_path *path, struct btrfs_key *new_key); struct extent_buffer *btrfs_root_node(struct btrfs_root *root); diff --git a/fs/btrfs/scrub.c b/fs/btrfs/scrub.c index e0677e4..51c342b 100644 --- a/fs/btrfs/scrub.c +++ b/fs/btrfs/scrub.c @@ -2385,8 +2385,7 @@ static noinline_for_stack int scrub_stripe(struct scrub_ctx *sctx, goto out; if (ret > 0) { - ret = btrfs_previous_item(root, path, 0, - BTRFS_EXTENT_ITEM_KEY); + ret = btrfs_previous_extent_item(root, path, 0); if (ret < 0) goto out; if (ret > 0) { -- 1.8.4 ^ permalink raw reply related [flat|nested] 6+ messages in thread
* [PATCH 3/3] Btrfs: fix missing inline refs when walking backrefs 2014-01-12 13:38 [PATCH] Btrfs: do not use extent commit root for sending Wang Shilong 2014-01-12 13:38 ` [PATCH 1/3] Btrfs: fix missing skinny metadata check in scrub_stripe() Wang Shilong 2014-01-12 13:38 ` [PATCH 2/3] Btrfs: fix to search previous metadata extent item since skinny metadata Wang Shilong @ 2014-01-12 13:38 ` Wang Shilong 2014-01-12 15:36 ` Filipe David Manana 2 siblings, 1 reply; 6+ messages in thread From: Wang Shilong @ 2014-01-12 13:38 UTC (permalink / raw) To: linux-btrfs From: Wang Shilong <wangsl.fnst@cn.fujitsu.com> If @slot=0, we may have an expected item in the previous leaf, So we should handle that case, otherwise, we will miss inline refs ,fix it. Signed-off-by: Wang Shilong <wangsl.fnst@cn.fujitsu.com> --- fs/btrfs/backref.c | 13 ++++++------- 1 file changed, 6 insertions(+), 7 deletions(-) diff --git a/fs/btrfs/backref.c b/fs/btrfs/backref.c index 964679c..4ccd726 100644 --- a/fs/btrfs/backref.c +++ b/fs/btrfs/backref.c @@ -893,14 +893,13 @@ again: spin_unlock(&delayed_refs->lock); } - if (path->slots[0]) { - struct extent_buffer *leaf; - int slot; + ret = btrfs_previous_extent_item(fs_info->extent_root, path, + key.objectid); + if (ret < 0) + goto out; - path->slots[0]--; - leaf = path->nodes[0]; - slot = path->slots[0]; - btrfs_item_key_to_cpu(leaf, &key, slot); + if (ret == 0) { + btrfs_item_key_to_cpu(path->nodes[0], &key, path->slots[0]); if (key.objectid == bytenr && (key.type == BTRFS_EXTENT_ITEM_KEY || key.type == BTRFS_METADATA_ITEM_KEY)) { -- 1.8.4 ^ permalink raw reply related [flat|nested] 6+ messages in thread
* Re: [PATCH 3/3] Btrfs: fix missing inline refs when walking backrefs 2014-01-12 13:38 ` [PATCH 3/3] Btrfs: fix missing inline refs when walking backrefs Wang Shilong @ 2014-01-12 15:36 ` Filipe David Manana 2014-01-13 1:27 ` Wang Shilong 0 siblings, 1 reply; 6+ messages in thread From: Filipe David Manana @ 2014-01-12 15:36 UTC (permalink / raw) To: Wang Shilong; +Cc: linux-btrfs@vger.kernel.org On Sun, Jan 12, 2014 at 1:38 PM, Wang Shilong <wangshilong1991@gmail.com> wrote: > From: Wang Shilong <wangsl.fnst@cn.fujitsu.com> > > If @slot=0, we may have an expected item in the previous leaf, > So we should handle that case, otherwise, we will miss inline refs > ,fix it. Hi Shilong. How can this happen exactly? So the search key, regardless of having an BTRFS_EXTENT_ITEM_KEY or a BTRFS_METADATA_ITEM_KEY type, always has an offset set to (u64)-1. This means the btrfs_search_slot call will always return 1 (not found, as expected) or an error. And btrfs_search_slot, when it doesn't find a key and if that key should be the first item in a leaf, it makes the path point to the previous leaf with a "path->slots[0] == btrfs_header_nritems(path->nodes[0])" (see [1]), otherwise to some value > 0. The only cases I can imagine getting "path->slots[0] == 0" is if we hit the first, i.e. left most, leaf in the btree or if the btree consists only of 1 leaf (no nodes) - in these cases there's isn't a previous leaf. Did I miss something here? thanks [1] - https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/tree/fs/btrfs/ctree.c?id=refs/tags/v3.13-rc8#n2629 > > Signed-off-by: Wang Shilong <wangsl.fnst@cn.fujitsu.com> > --- > fs/btrfs/backref.c | 13 ++++++------- > 1 file changed, 6 insertions(+), 7 deletions(-) > > diff --git a/fs/btrfs/backref.c b/fs/btrfs/backref.c > index 964679c..4ccd726 100644 > --- a/fs/btrfs/backref.c > +++ b/fs/btrfs/backref.c > @@ -893,14 +893,13 @@ again: > spin_unlock(&delayed_refs->lock); > } > > - if (path->slots[0]) { > - struct extent_buffer *leaf; > - int slot; > + ret = btrfs_previous_extent_item(fs_info->extent_root, path, > + key.objectid); > + if (ret < 0) > + goto out; > > - path->slots[0]--; > - leaf = path->nodes[0]; > - slot = path->slots[0]; > - btrfs_item_key_to_cpu(leaf, &key, slot); > + if (ret == 0) { > + btrfs_item_key_to_cpu(path->nodes[0], &key, path->slots[0]); > if (key.objectid == bytenr && > (key.type == BTRFS_EXTENT_ITEM_KEY || > key.type == BTRFS_METADATA_ITEM_KEY)) { > -- > 1.8.4 > > -- > To unsubscribe from this list: send the line "unsubscribe linux-btrfs" in > the body of a message to majordomo@vger.kernel.org > More majordomo info at http://vger.kernel.org/majordomo-info.html -- Filipe David Manana, "Reasonable men adapt themselves to the world. Unreasonable men adapt the world to themselves. That's why all progress depends on unreasonable men." ^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: [PATCH 3/3] Btrfs: fix missing inline refs when walking backrefs 2014-01-12 15:36 ` Filipe David Manana @ 2014-01-13 1:27 ` Wang Shilong 0 siblings, 0 replies; 6+ messages in thread From: Wang Shilong @ 2014-01-13 1:27 UTC (permalink / raw) To: fdmanana; +Cc: Wang Shilong, linux-btrfs@vger.kernel.org Hi Filipe, On 01/12/2014 11:36 PM, Filipe David Manana wrote: > On Sun, Jan 12, 2014 at 1:38 PM, Wang Shilong <wangshilong1991@gmail.com> wrote: >> From: Wang Shilong <wangsl.fnst@cn.fujitsu.com> >> >> If @slot=0, we may have an expected item in the previous leaf, >> So we should handle that case, otherwise, we will miss inline refs >> ,fix it. > Hi Shilong. > > How can this happen exactly? > > So the search key, regardless of having an BTRFS_EXTENT_ITEM_KEY or a > BTRFS_METADATA_ITEM_KEY type, always has an offset set to (u64)-1. > This means the btrfs_search_slot call will always return 1 (not found, > as expected) or an error. > > And btrfs_search_slot, when it doesn't find a key and if that key > should be the first item in a leaf, it makes the path point to the > previous leaf with a "path->slots[0] == > btrfs_header_nritems(path->nodes[0])" (see [1]), otherwise to some > value > 0. The only cases I can imagine getting "path->slots[0] == 0" > is if we hit the first, i.e. left most, leaf in the btree or if the > btree consists only of 1 leaf (no nodes) - in these cases there's > isn't a previous leaf. > > Did I miss something here? > > thanks > > [1] - https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/tree/fs/btrfs/ctree.c?id=refs/tags/v3.13-rc8#n2629 You are right, i was missing something when i made this patch. Really thanks for you reviewing and correcting me!:-) Thanks, Wang > >> Signed-off-by: Wang Shilong <wangsl.fnst@cn.fujitsu.com> >> --- >> fs/btrfs/backref.c | 13 ++++++------- >> 1 file changed, 6 insertions(+), 7 deletions(-) >> >> diff --git a/fs/btrfs/backref.c b/fs/btrfs/backref.c >> index 964679c..4ccd726 100644 >> --- a/fs/btrfs/backref.c >> +++ b/fs/btrfs/backref.c >> @@ -893,14 +893,13 @@ again: >> spin_unlock(&delayed_refs->lock); >> } >> >> - if (path->slots[0]) { >> - struct extent_buffer *leaf; >> - int slot; >> + ret = btrfs_previous_extent_item(fs_info->extent_root, path, >> + key.objectid); >> + if (ret < 0) >> + goto out; >> >> - path->slots[0]--; >> - leaf = path->nodes[0]; >> - slot = path->slots[0]; >> - btrfs_item_key_to_cpu(leaf, &key, slot); >> + if (ret == 0) { >> + btrfs_item_key_to_cpu(path->nodes[0], &key, path->slots[0]); >> if (key.objectid == bytenr && >> (key.type == BTRFS_EXTENT_ITEM_KEY || >> key.type == BTRFS_METADATA_ITEM_KEY)) { >> -- >> 1.8.4 >> >> -- >> To unsubscribe from this list: send the line "unsubscribe linux-btrfs" in >> the body of a message to majordomo@vger.kernel.org >> More majordomo info at http://vger.kernel.org/majordomo-info.html > > ^ permalink raw reply [flat|nested] 6+ messages in thread
end of thread, other threads:[~2014-01-13 1:28 UTC | newest] Thread overview: 6+ messages (download: mbox.gz follow: Atom feed -- links below jump to the message on this page -- 2014-01-12 13:38 [PATCH] Btrfs: do not use extent commit root for sending Wang Shilong 2014-01-12 13:38 ` [PATCH 1/3] Btrfs: fix missing skinny metadata check in scrub_stripe() Wang Shilong 2014-01-12 13:38 ` [PATCH 2/3] Btrfs: fix to search previous metadata extent item since skinny metadata Wang Shilong 2014-01-12 13:38 ` [PATCH 3/3] Btrfs: fix missing inline refs when walking backrefs Wang Shilong 2014-01-12 15:36 ` Filipe David Manana 2014-01-13 1:27 ` Wang Shilong
This is an external index of several public inboxes, see mirroring instructions on how to clone and mirror all data and code used by this external index.