From: Kent Overstreet <kent.overstreet@linux.dev>
To: linux-bcachefs@vger.kernel.org
Cc: Kent Overstreet <kent.overstreet@linux.dev>
Subject: [PATCH 04/17] bcachefs: BTREE_INSERT_JOURNAL_REPLAY now "don't init trans->journal_res"
Date: Fri, 10 Nov 2023 11:31:41 -0500 [thread overview]
Message-ID: <20231110163157.2736111-5-kent.overstreet@linux.dev> (raw)
In-Reply-To: <20231110163157.2736111-1-kent.overstreet@linux.dev>
This slightly changes how trans->journal_res works, in preparation for
changing the btree write buffer flush path to use it.
Now, BTREE_INSERT_JOURNAL_REPLAY means "don't take a journal
reservation; trans->journal_res.seq already refers to the journal
sequence number to pin".
Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
---
fs/bcachefs/btree_trans_commit.c | 19 +++++++++++++++----
fs/bcachefs/recovery.c | 2 ++
2 files changed, 17 insertions(+), 4 deletions(-)
diff --git a/fs/bcachefs/btree_trans_commit.c b/fs/bcachefs/btree_trans_commit.c
index 3152f6abf73f..ec90a06a6cf9 100644
--- a/fs/bcachefs/btree_trans_commit.c
+++ b/fs/bcachefs/btree_trans_commit.c
@@ -677,8 +677,6 @@ bch2_trans_commit_write_locked(struct btree_trans *trans, unsigned flags,
if (unlikely(trans->journal_transaction_names))
journal_transaction_name(trans);
- } else {
- trans->journal_res.seq = c->journal.replay_journal_seq;
}
/*
@@ -897,7 +895,8 @@ static inline int do_bch2_trans_commit(struct btree_trans *trans, unsigned flags
* Drop journal reservation after dropping write locks, since dropping
* the journal reservation may kick off a journal write:
*/
- bch2_journal_res_put(&c->journal, &trans->journal_res);
+ if (likely(!(flags & BTREE_INSERT_JOURNAL_REPLAY)))
+ bch2_journal_res_put(&c->journal, &trans->journal_res);
return ret;
}
@@ -1140,7 +1139,8 @@ int __bch2_trans_commit(struct btree_trans *trans, unsigned flags)
}
retry:
bch2_trans_verify_not_in_restart(trans);
- memset(&trans->journal_res, 0, sizeof(trans->journal_res));
+ if (likely(!(flags & BTREE_INSERT_JOURNAL_REPLAY)))
+ memset(&trans->journal_res, 0, sizeof(trans->journal_res));
ret = do_bch2_trans_commit(trans, flags, &i, _RET_IP_);
@@ -1165,5 +1165,16 @@ int __bch2_trans_commit(struct btree_trans *trans, unsigned flags)
if (ret)
goto out;
+ /*
+ * We might have done another transaction commit in the error path -
+ * i.e. btree write buffer flush - which will have made use of
+ * trans->journal_res, but with BTREE_INSERT_JOURNAL_REPLAY that is how
+ * the journal sequence number to pin is passed in - so we must restart:
+ */
+ if (flags & BTREE_INSERT_JOURNAL_REPLAY) {
+ ret = -BCH_ERR_transaction_restart_nested;
+ goto out;
+ }
+
goto retry;
}
diff --git a/fs/bcachefs/recovery.c b/fs/bcachefs/recovery.c
index 9c30500ce920..8fe39a91beed 100644
--- a/fs/bcachefs/recovery.c
+++ b/fs/bcachefs/recovery.c
@@ -98,6 +98,8 @@ static int bch2_journal_replay_key(struct btree_trans *trans,
unsigned update_flags = BTREE_TRIGGER_NORUN;
int ret;
+ trans->journal_res.seq = k->journal_seq;
+
/*
* BTREE_UPDATE_KEY_CACHE_RECLAIM disables key cache lookup/update to
* keep the key cache coherent with the underlying btree. Nothing
--
2.42.0
next prev parent reply other threads:[~2023-11-10 16:32 UTC|newest]
Thread overview: 29+ messages / expand[flat|nested] mbox.gz Atom feed top
2023-11-10 16:31 [PATCH 00/17] btree write buffer & journal optimizations Kent Overstreet
2023-11-10 16:31 ` [PATCH 01/17] bcachefs: Kill journal pre-reservations Kent Overstreet
2023-11-10 16:31 ` [PATCH 02/17] bcachefs: track_event_change() Kent Overstreet
2023-11-10 16:31 ` [PATCH 03/17] bcachefs: Journal pins must always have a flush_fn Kent Overstreet
2023-11-13 15:22 ` Brian Foster
2023-11-13 16:36 ` Kent Overstreet
2023-11-13 17:08 ` Brian Foster
2023-11-10 16:31 ` Kent Overstreet [this message]
2023-11-10 16:31 ` [PATCH 05/17] bcachefs: Kill BTREE_UPDATE_PREJOURNAL Kent Overstreet
2023-11-13 15:29 ` Brian Foster
2023-11-13 16:49 ` Kent Overstreet
2023-11-14 13:17 ` Brian Foster
2023-11-10 16:31 ` [PATCH 06/17] bcachefs: Go rw before journal replay Kent Overstreet
2023-11-10 16:31 ` [PATCH 07/17] bcachefs: Make journal replay more efficient Kent Overstreet
2023-11-14 13:19 ` Brian Foster
2023-11-15 1:50 ` Kent Overstreet
2023-11-10 16:31 ` [PATCH 08/17] bcachefs: Don't flush journal after replay Kent Overstreet
2023-11-10 16:31 ` [PATCH 09/17] bcachefs: Unwritten journal buffers are always dirty Kent Overstreet
2023-11-10 16:31 ` [PATCH 10/17] bcachefs: journal->buf_lock Kent Overstreet
2023-11-10 16:31 ` [PATCH 11/17] bcachefs: bch2_journal_block_reservations() Kent Overstreet
2023-11-10 16:31 ` [PATCH 12/17] bcachefs: Clean up btree write buffer write ref handling Kent Overstreet
2023-11-10 16:31 ` [PATCH 13/17] bcachefs: bch2_btree_write_buffer_flush_locked() Kent Overstreet
2023-11-10 16:31 ` [PATCH 14/17] bcachefs: bch2_btree_write_buffer_flush() -> bch2_btree_write_buffer_tryflush() Kent Overstreet
2023-11-10 16:31 ` [PATCH 15/17] bcachefs: Improve btree write buffer tracepoints Kent Overstreet
2023-11-10 16:31 ` [PATCH 16/17] bcachefs: btree write buffer now slurps keys from journal Kent Overstreet
2023-11-21 10:56 ` Geert Uytterhoeven
2023-11-21 16:52 ` Kent Overstreet
2023-11-10 16:31 ` [PATCH 17/17] bcachefs: Inline btree write buffer sort Kent Overstreet
2023-11-10 16:42 ` [PATCH 00/17] btree write buffer & journal optimizations Kent Overstreet
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=20231110163157.2736111-5-kent.overstreet@linux.dev \
--to=kent.overstreet@linux.dev \
--cc=linux-bcachefs@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