From: Kent Overstreet <kent.overstreet@linux.dev>
To: linux-bcachefs@vger.kernel.org
Cc: Kent Overstreet <kent.overstreet@linux.dev>
Subject: [PATCH 1/6] bcachefs: move.c exports, refactoring
Date: Tue, 24 Oct 2023 15:14:06 -0400 [thread overview]
Message-ID: <20231024191414.2157874-2-kent.overstreet@linux.dev> (raw)
In-Reply-To: <20231024191414.2157874-1-kent.overstreet@linux.dev>
Prep work for the new rebalance code - we need a few helpers exported.
Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
---
fs/bcachefs/move.c | 119 +++++++++++++++++++++-------------------
fs/bcachefs/move.h | 22 +++++++-
fs/bcachefs/rebalance.c | 3 +-
3 files changed, 85 insertions(+), 59 deletions(-)
diff --git a/fs/bcachefs/move.c b/fs/bcachefs/move.c
index 38b076ff1906..12167791e34c 100644
--- a/fs/bcachefs/move.c
+++ b/fs/bcachefs/move.c
@@ -171,8 +171,8 @@ void bch2_moving_ctxt_do_pending_writes(struct moving_context *ctxt,
}
}
-static void bch2_move_ctxt_wait_for_io(struct moving_context *ctxt,
- struct btree_trans *trans)
+void bch2_move_ctxt_wait_for_io(struct moving_context *ctxt,
+ struct btree_trans *trans)
{
unsigned sectors_pending = atomic_read(&ctxt->write_sectors);
@@ -287,14 +287,13 @@ static int bch2_extent_drop_ptrs(struct btree_trans *trans,
bch2_trans_commit(trans, NULL, NULL, BTREE_INSERT_NOFAIL);
}
-static int bch2_move_extent(struct btree_trans *trans,
- struct btree_iter *iter,
- struct moving_context *ctxt,
- struct move_bucket_in_flight *bucket_in_flight,
- struct bch_io_opts io_opts,
- enum btree_id btree_id,
- struct bkey_s_c k,
- struct data_update_opts data_opts)
+int bch2_move_extent(struct btree_trans *trans,
+ struct btree_iter *iter,
+ struct moving_context *ctxt,
+ struct move_bucket_in_flight *bucket_in_flight,
+ struct bch_io_opts io_opts,
+ struct bkey_s_c k,
+ struct data_update_opts data_opts)
{
struct bch_fs *c = trans->c;
struct bkey_ptrs_c ptrs = bch2_bkey_ptrs_c(k);
@@ -356,7 +355,7 @@ static int bch2_move_extent(struct btree_trans *trans,
io->rbio.bio.bi_end_io = move_read_endio;
ret = bch2_data_update_init(trans, ctxt, &io->write, ctxt->wp,
- io_opts, data_opts, btree_id, k);
+ io_opts, data_opts, iter->btree_id, k);
if (ret && ret != -BCH_ERR_unwritten_extent_update)
goto err_free_pages;
@@ -371,6 +370,9 @@ static int bch2_move_extent(struct btree_trans *trans,
io->write.ctxt = ctxt;
io->write.op.end_io = move_write_done;
+ if (ctxt->rate)
+ bch2_ratelimit_increment(ctxt->rate, k.k->size);
+
if (ctxt->stats) {
atomic64_inc(&ctxt->stats->keys_moved);
atomic64_add(k.k->size, &ctxt->stats->sectors_moved);
@@ -400,7 +402,7 @@ static int bch2_move_extent(struct btree_trans *trans,
closure_get(&ctxt->cl);
bch2_read_extent(trans, &io->rbio,
bkey_start_pos(k.k),
- btree_id, k, 0,
+ iter->btree_id, k, 0,
BCH_READ_NODECODE|
BCH_READ_LAST_FRAGMENT);
return 0;
@@ -464,9 +466,9 @@ struct bch_io_opts *bch2_move_get_io_opts(struct btree_trans *trans,
return &io_opts->fs_io_opts;
}
-static int bch2_move_get_io_opts_one(struct btree_trans *trans,
- struct bch_io_opts *io_opts,
- struct bkey_s_c extent_k)
+int bch2_move_get_io_opts_one(struct btree_trans *trans,
+ struct bch_io_opts *io_opts,
+ struct bkey_s_c extent_k)
{
struct btree_iter iter;
struct bkey_s_c k;
@@ -497,8 +499,8 @@ static int bch2_move_get_io_opts_one(struct btree_trans *trans,
return 0;
}
-static int move_ratelimit(struct btree_trans *trans,
- struct moving_context *ctxt)
+int bch2_move_ratelimit(struct btree_trans *trans,
+ struct moving_context *ctxt)
{
struct bch_fs *c = trans->c;
u64 delay;
@@ -545,7 +547,8 @@ static int move_ratelimit(struct btree_trans *trans,
return 0;
}
-static int __bch2_move_data(struct moving_context *ctxt,
+static int bch2_move_data_btree(struct btree_trans *trans,
+ struct moving_context *ctxt,
struct bpos start,
struct bpos end,
move_pred_fn pred, void *arg,
@@ -555,7 +558,6 @@ static int __bch2_move_data(struct moving_context *ctxt,
struct per_snapshot_io_opts snapshot_io_opts;
struct bch_io_opts *io_opts;
struct bkey_buf sk;
- struct btree_trans *trans = bch2_trans_get(c);
struct btree_iter iter;
struct bkey_s_c k;
struct data_update_opts data_opts;
@@ -577,7 +579,7 @@ static int __bch2_move_data(struct moving_context *ctxt,
if (ctxt->rate)
bch2_ratelimit_reset(ctxt->rate);
- while (!move_ratelimit(trans, ctxt)) {
+ while (!bch2_move_ratelimit(trans, ctxt)) {
bch2_trans_begin(trans);
k = bch2_btree_iter_peek(&iter);
@@ -616,7 +618,7 @@ static int __bch2_move_data(struct moving_context *ctxt,
k = bkey_i_to_s_c(sk.k);
ret2 = bch2_move_extent(trans, &iter, ctxt, NULL,
- *io_opts, btree_id, k, data_opts);
+ *io_opts, k, data_opts);
if (ret2) {
if (bch2_err_matches(ret2, BCH_ERR_transaction_restart))
continue;
@@ -630,9 +632,6 @@ static int __bch2_move_data(struct moving_context *ctxt,
/* XXX signal failure */
goto next;
}
-
- if (ctxt->rate)
- bch2_ratelimit_increment(ctxt->rate, k.k->size);
next:
if (ctxt->stats)
atomic64_add(k.k->size, &ctxt->stats->sectors_seen);
@@ -641,48 +640,60 @@ static int __bch2_move_data(struct moving_context *ctxt,
}
bch2_trans_iter_exit(trans, &iter);
- bch2_trans_put(trans);
bch2_bkey_buf_exit(&sk, c);
per_snapshot_io_opts_exit(&snapshot_io_opts);
return ret;
}
-int bch2_move_data(struct bch_fs *c,
- enum btree_id start_btree_id, struct bpos start_pos,
- enum btree_id end_btree_id, struct bpos end_pos,
- struct bch_ratelimit *rate,
- struct bch_move_stats *stats,
- struct write_point_specifier wp,
- bool wait_on_copygc,
- move_pred_fn pred, void *arg)
+int __bch2_move_data(struct btree_trans *trans,
+ struct moving_context *ctxt,
+ struct bbpos start,
+ struct bbpos end,
+ move_pred_fn pred, void *arg)
{
- struct moving_context ctxt;
+ struct bch_fs *c = trans->c;
enum btree_id id;
int ret = 0;
- bch2_moving_ctxt_init(&ctxt, c, rate, stats, wp, wait_on_copygc);
-
- for (id = start_btree_id;
- id <= min_t(unsigned, end_btree_id, btree_id_nr_alive(c) - 1);
+ for (id = start.btree;
+ id <= min_t(unsigned, end.btree, btree_id_nr_alive(c) - 1);
id++) {
- stats->btree_id = id;
+ ctxt->stats->btree_id = id;
- if (id != BTREE_ID_extents &&
- id != BTREE_ID_reflink)
+ if (!btree_type_has_ptrs(id) ||
+ !bch2_btree_id_root(c, id)->b)
continue;
- if (!bch2_btree_id_root(c, id)->b)
- continue;
-
- ret = __bch2_move_data(&ctxt,
- id == start_btree_id ? start_pos : POS_MIN,
- id == end_btree_id ? end_pos : POS_MAX,
+ ret = bch2_move_data_btree(trans, ctxt,
+ id == start.btree ? start.pos : POS_MIN,
+ id == end.btree ? end.pos : POS_MAX,
pred, arg, id);
if (ret)
break;
}
+ return ret;
+}
+
+int bch2_move_data(struct bch_fs *c,
+ struct bbpos start,
+ struct bbpos end,
+ struct bch_ratelimit *rate,
+ struct bch_move_stats *stats,
+ struct write_point_specifier wp,
+ bool wait_on_copygc,
+ move_pred_fn pred, void *arg)
+{
+
+ struct btree_trans *trans;
+ struct moving_context ctxt;
+ int ret;
+
+ bch2_moving_ctxt_init(&ctxt, c, rate, stats, wp, wait_on_copygc);
+ trans = bch2_trans_get(c);
+ ret = __bch2_move_data(trans, &ctxt, start, end, pred, arg);
+ bch2_trans_put(trans);
bch2_moving_ctxt_exit(&ctxt);
return ret;
@@ -739,7 +750,7 @@ int __bch2_evacuate_bucket(struct btree_trans *trans,
goto err;
}
- while (!(ret = move_ratelimit(trans, ctxt))) {
+ while (!(ret = bch2_move_ratelimit(trans, ctxt))) {
bch2_trans_begin(trans);
ret = bch2_get_next_backpointer(trans, bucket, gen,
@@ -791,7 +802,7 @@ int __bch2_evacuate_bucket(struct btree_trans *trans,
ret = bch2_move_extent(trans, &iter, ctxt,
bucket_in_flight,
- io_opts, bp.btree_id, k, data_opts);
+ io_opts, k, data_opts);
bch2_trans_iter_exit(trans, &iter);
if (bch2_err_matches(ret, BCH_ERR_transaction_restart))
@@ -804,8 +815,6 @@ int __bch2_evacuate_bucket(struct btree_trans *trans,
if (ret)
goto err;
- if (ctxt->rate)
- bch2_ratelimit_increment(ctxt->rate, k.k->size);
if (ctxt->stats)
atomic64_add(k.k->size, &ctxt->stats->sectors_seen);
} else {
@@ -1087,8 +1096,8 @@ int bch2_data_job(struct bch_fs *c,
ret = bch2_replicas_gc2(c) ?: ret;
ret = bch2_move_data(c,
- op.start_btree, op.start_pos,
- op.end_btree, op.end_pos,
+ (struct bbpos) { op.start_btree, op.start_pos },
+ (struct bbpos) { op.end_btree, op.end_pos },
NULL,
stats,
writepoint_hashed((unsigned long) current),
@@ -1111,8 +1120,8 @@ int bch2_data_job(struct bch_fs *c,
ret = bch2_replicas_gc2(c) ?: ret;
ret = bch2_move_data(c,
- op.start_btree, op.start_pos,
- op.end_btree, op.end_pos,
+ (struct bbpos) { op.start_btree, op.start_pos },
+ (struct bbpos) { op.end_btree, op.end_pos },
NULL,
stats,
writepoint_hashed((unsigned long) current),
diff --git a/fs/bcachefs/move.h b/fs/bcachefs/move.h
index aa4b65c4f960..67ca13f7e772 100644
--- a/fs/bcachefs/move.h
+++ b/fs/bcachefs/move.h
@@ -2,6 +2,7 @@
#ifndef _BCACHEFS_MOVE_H
#define _BCACHEFS_MOVE_H
+#include "bbpos.h"
#include "bcachefs_ioctl.h"
#include "btree_iter.h"
#include "buckets.h"
@@ -61,6 +62,9 @@ void bch2_moving_ctxt_init(struct moving_context *, struct bch_fs *,
struct moving_io *bch2_moving_ctxt_next_pending_write(struct moving_context *);
void bch2_moving_ctxt_do_pending_writes(struct moving_context *,
struct btree_trans *);
+void bch2_move_ctxt_wait_for_io(struct moving_context *,
+ struct btree_trans *);
+int bch2_move_ratelimit(struct btree_trans *, struct moving_context *);
/* Inodes in different snapshots may have different IO options: */
struct snapshot_io_opts_entry {
@@ -87,12 +91,26 @@ static inline void per_snapshot_io_opts_exit(struct per_snapshot_io_opts *io_opt
struct bch_io_opts *bch2_move_get_io_opts(struct btree_trans *,
struct per_snapshot_io_opts *, struct bkey_s_c);
+int bch2_move_get_io_opts_one(struct btree_trans *, struct bch_io_opts *, struct bkey_s_c);
int bch2_scan_old_btree_nodes(struct bch_fs *, struct bch_move_stats *);
+int bch2_move_extent(struct btree_trans *,
+ struct btree_iter *,
+ struct moving_context *,
+ struct move_bucket_in_flight *,
+ struct bch_io_opts,
+ struct bkey_s_c,
+ struct data_update_opts);
+
+int __bch2_move_data(struct btree_trans *,
+ struct moving_context *,
+ struct bbpos,
+ struct bbpos,
+ move_pred_fn, void *);
int bch2_move_data(struct bch_fs *,
- enum btree_id, struct bpos,
- enum btree_id, struct bpos,
+ struct bbpos start,
+ struct bbpos end,
struct bch_ratelimit *,
struct bch_move_stats *,
struct write_point_specifier,
diff --git a/fs/bcachefs/rebalance.c b/fs/bcachefs/rebalance.c
index 568f1e8e7507..92403fa79f1f 100644
--- a/fs/bcachefs/rebalance.c
+++ b/fs/bcachefs/rebalance.c
@@ -254,8 +254,7 @@ static int bch2_rebalance_thread(void *arg)
rebalance_work_reset(c);
bch2_move_data(c,
- 0, POS_MIN,
- BTREE_ID_NR, POS_MAX,
+ BBPOS_MIN, BBPOS_MAX,
/* ratelimiting disabled for now */
NULL, /* &r->pd.rate, */
&move_stats,
--
2.42.0
next prev parent reply other threads:[~2023-10-24 19:14 UTC|newest]
Thread overview: 10+ messages / expand[flat|nested] mbox.gz Atom feed top
2023-10-24 19:14 [PATCH 0/6] rebalance_work btree Kent Overstreet
2023-10-24 19:14 ` Kent Overstreet [this message]
2023-10-24 19:14 ` [PATCH 2/6] bcachefs: moving_context now owns a btree_trans Kent Overstreet
2023-10-24 19:14 ` [PATCH 3/6] bcachefs: move: convert to bbpos Kent Overstreet
2023-10-24 19:14 ` [PATCH 4/6] bcachefs: move: move_stats refactoring Kent Overstreet
2023-10-24 19:14 ` [PATCH 5/6] bcachefs: bch2_inum_opts_get() Kent Overstreet
2023-10-24 19:14 ` [PATCH 6/6] bcachefs: rebalance_work Kent Overstreet
2023-11-01 17:02 ` Nathan Chancellor
2023-11-01 17:07 ` Nathan Chancellor
2023-11-02 1:11 ` 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=20231024191414.2157874-2-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