* the journal that are being staged or in flight.
*/
struct journal_buf {
+ struct closure io;
struct jset *data;
__BKEY_PADDED(key, BCH_REPLICAS_MAX);
unsigned disk_sectors; /* maximum size entry could have been, if
buf_size was bigger */
unsigned u64s_reserved;
- bool noflush; /* write has already been kicked off, and was noflush */
- bool must_flush; /* something wants a flush */
- bool separate_flush;
+ bool noflush:1; /* write has already been kicked off, and was noflush */
+ bool must_flush:1; /* something wants a flush */
+ bool separate_flush:1;
+ bool need_flush_to_write_buffer:1;
+ bool write_started:1;
+ bool write_allocated:1;
+ bool write_done:1;
+ u8 idx;
};
/*
u64 seq;
};
-/*
- * For reserving space in the journal prior to getting a reservation on a
- * particular journal entry:
- */
-struct journal_preres {
- unsigned u64s;
-};
-
union journal_res_state {
struct {
atomic64_t counter;
};
};
-union journal_preres_state {
- struct {
- atomic64_t counter;
- };
-
- struct {
- u64 v;
- };
-
- struct {
- u64 waiting:1,
- reserved:31,
- remaining:32;
- };
-};
-
/* bytes: */
#define JOURNAL_ENTRY_SIZE_MIN (64U << 10) /* 64k */
#define JOURNAL_ENTRY_SIZE_MAX (4U << 20) /* 4M */
JOURNAL_NEED_FLUSH_WRITE,
};
-#define JOURNAL_WATERMARKS() \
- x(any) \
- x(copygc) \
- x(reserved)
-
-enum journal_watermark {
-#define x(n) JOURNAL_WATERMARK_##n,
- JOURNAL_WATERMARKS()
-#undef x
-};
-
-#define JOURNAL_WATERMARK_MASK 3
-
/* Reasons we may fail to get a journal reservation: */
#define JOURNAL_ERRORS() \
x(ok) \
+ x(retry) \
x(blocked) \
x(max_in_flight) \
x(journal_full) \
typedef DARRAY(u64) darray_u64;
+struct journal_bio {
+ struct bch_dev *ca;
+ unsigned buf_idx;
+
+ struct bio bio;
+};
+
/* Embedded in struct bch_fs */
struct journal {
/* Fastpath stuff up front: */
struct {
union journal_res_state reservations;
- enum journal_watermark watermark;
-
- union journal_preres_state prereserved;
+ enum bch_watermark watermark;
} __aligned(SMP_CACHE_BYTES);
*/
darray_u64 early_journal_entries;
+ /*
+ * Protects journal_buf->data, when accessing without a jorunal
+ * reservation: for synchronization between the btree write buffer code
+ * and the journal write path:
+ */
+ struct mutex buf_lock;
/*
* Two journal entries -- one is currently open for new entries, the
* other is possibly being written out.
/* Used when waiting because the journal was full */
wait_queue_head_t wait;
struct closure_waitlist async_wait;
- struct closure_waitlist preres_wait;
- struct closure io;
struct delayed_work write_work;
+ struct workqueue_struct *wq;
/* Sequence number of most recent journal entry (last entry in @pin) */
atomic64_t seq;
unsigned long last_flush_write;
- u64 res_get_blocked_start;
u64 write_start_time;
u64 nr_flush_writes;
u64 nr_noflush_writes;
+ u64 entry_bytes_written;
- struct bch2_time_stats *flush_write_time;
- struct bch2_time_stats *noflush_write_time;
- struct bch2_time_stats *blocked_time;
- struct bch2_time_stats *flush_seq_time;
+ struct time_stats *flush_write_time;
+ struct time_stats *noflush_write_time;
+ struct time_stats *flush_seq_time;
#ifdef CONFIG_DEBUG_LOCK_ALLOC
struct lockdep_map res_map;
u64 *buckets;
/* Bio for journal reads/writes to this device */
- struct bio *bio;
+ struct journal_bio *bio[JOURNAL_BUF_NR];
/* for bch_journal_read_device */
struct closure read;