From: Jan Schmidt <list.btrfs@jan-o-sch.net>
To: chris.mason@fusionio.com, linux-btrfs@vger.kernel.org
Subject: [PATCH 1/7] Btrfs: avoid waiting for delayed refs when we must not
Date: Wed, 27 Jun 2012 19:24:53 +0200 [thread overview]
Message-ID: <1340817899-16343-2-git-send-email-list.btrfs@jan-o-sch.net> (raw)
In-Reply-To: <1340817899-16343-1-git-send-email-list.btrfs@jan-o-sch.net>
We track two conditions to decide if we should sleep while waiting for more
delayed refs, the number of delayed refs (num_refs) and the first entry in
the list of blockers (first_seq).
When we suspect staleness, we save num_refs and do one more cycle. If
nothing changes, we then save first_seq for later comparison and do
wait_event. We ought to save first_seq the very same moment we're saving
num_refs. Otherwise we cannot be sure that nothing has changed and we might
start waiting when we shouldn't, which could lead to starvation.
Signed-off-by: Jan Schmidt <list.btrfs@jan-o-sch.net>
---
fs/btrfs/extent-tree.c | 11 ++++++-----
1 files changed, 6 insertions(+), 5 deletions(-)
diff --git a/fs/btrfs/extent-tree.c b/fs/btrfs/extent-tree.c
index 4b5a1e1..6e1d367 100644
--- a/fs/btrfs/extent-tree.c
+++ b/fs/btrfs/extent-tree.c
@@ -2347,12 +2347,10 @@ next:
return count;
}
-
static void wait_for_more_refs(struct btrfs_delayed_ref_root *delayed_refs,
- unsigned long num_refs)
+ unsigned long num_refs,
+ struct list_head *first_seq)
{
- struct list_head *first_seq = delayed_refs->seq_head.next;
-
spin_unlock(&delayed_refs->lock);
pr_debug("waiting for more refs (num %ld, first %p)\n",
num_refs, first_seq);
@@ -2381,6 +2379,7 @@ int btrfs_run_delayed_refs(struct btrfs_trans_handle *trans,
struct btrfs_delayed_ref_root *delayed_refs;
struct btrfs_delayed_ref_node *ref;
struct list_head cluster;
+ struct list_head *first_seq = NULL;
int ret;
u64 delayed_start;
int run_all = count == (unsigned long)-1;
@@ -2436,8 +2435,10 @@ again:
*/
consider_waiting = 1;
num_refs = delayed_refs->num_entries;
+ first_seq = root->fs_info->tree_mod_seq_list.next;
} else {
- wait_for_more_refs(delayed_refs, num_refs);
+ wait_for_more_refs(delayed_refs,
+ num_refs, first_seq);
/*
* after waiting, things have changed. we
* dropped the lock and someone else might have
--
1.7.3.4
next prev parent reply other threads:[~2012-06-27 17:25 UTC|newest]
Thread overview: 8+ messages / expand[flat|nested] mbox.gz Atom feed top
2012-06-27 17:24 [PATCH 0/7] Btrfs: fixes for 3.5-rc5 Jan Schmidt
2012-06-27 17:24 ` Jan Schmidt [this message]
2012-06-27 17:24 ` [PATCH 2/7] Btrfs: support root level changes in __resolve_indirect_ref Jan Schmidt
2012-06-27 17:24 ` [PATCH 3/7] Btrfs: fix tree mod log for root replacements at leaf level Jan Schmidt
2012-06-27 17:24 ` [PATCH 4/7] Btrfs: always put insert_ptr modifications into the tree mod log Jan Schmidt
2012-06-27 17:24 ` [PATCH 5/7] Btrfs: leave critical region in btrfs_find_all_roots as soon as possible Jan Schmidt
2012-06-27 17:24 ` [PATCH 6/7] Btrfs: fix tree mod log rewind of ADD operations Jan Schmidt
2012-06-27 17:24 ` [PATCH 7/7] Btrfs: resolve tree mod log locking issue in btrfs_next_leaf Jan Schmidt
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=1340817899-16343-2-git-send-email-list.btrfs@jan-o-sch.net \
--to=list.btrfs@jan-o-sch.net \
--cc=chris.mason@fusionio.com \
--cc=linux-btrfs@vger.kernel.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).