diff options
| author | Kent Overstreet <kent.overstreet@gmail.com> | 2021-01-21 23:28:59 +0300 |
|---|---|---|
| committer | Kent Overstreet <kent.overstreet@linux.dev> | 2023-10-23 00:08:52 +0300 |
| commit | 2abe542087d9cb1bc7bb8ac7ae262afccbdb7aa6 (patch) | |
| tree | c33f5733a5cf3f091e59e5084be29ef5819a9f71 /fs/bcachefs/rebalance.c | |
| parent | 7f4e1d5d0faff0d72e9f6708bf98488d76533846 (diff) | |
| download | linux-2abe542087d9cb1bc7bb8ac7ae262afccbdb7aa6.tar.xz | |
bcachefs: Persist 64 bit io clocks
Originally, bcachefs - going back to bcache - stored, for each bucket, a
16 bit counter corresponding to how long it had been since the bucket
was read from. But, this required periodically rescaling counters on
every bucket to avoid wraparound. That wasn't an issue in bcache, where
we'd perodically rewrite the per bucket metadata all at once, but in
bcachefs we're trying to avoid having to walk every single bucket.
This patch switches to persisting 64 bit io clocks, corresponding to the
64 bit bucket timestaps introduced in the previous patch with
KEY_TYPE_alloc_v2.
Signed-off-by: Kent Overstreet <kent.overstreet@gmail.com>
Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
Diffstat (limited to 'fs/bcachefs/rebalance.c')
| -rw-r--r-- | fs/bcachefs/rebalance.c | 10 |
1 files changed, 5 insertions, 5 deletions
diff --git a/fs/bcachefs/rebalance.c b/fs/bcachefs/rebalance.c index f9a12dd797a5..2263ee41c444 100644 --- a/fs/bcachefs/rebalance.c +++ b/fs/bcachefs/rebalance.c @@ -169,12 +169,12 @@ static int bch2_rebalance_thread(void *arg) unsigned long start, prev_start; unsigned long prev_run_time, prev_run_cputime; unsigned long cputime, prev_cputime; - unsigned long io_start; + u64 io_start; long throttle; set_freezable(); - io_start = atomic_long_read(&clock->now); + io_start = atomic64_read(&clock->now); p = rebalance_work(c); prev_start = jiffies; prev_cputime = curr_cputime(); @@ -210,7 +210,7 @@ static int bch2_rebalance_thread(void *arg) (20 - w.dev_most_full_percent), 50); - if (atomic_long_read(&clock->now) + clock->max_slop < + if (atomic64_read(&clock->now) + clock->max_slop < r->throttled_until_iotime) { r->throttled_until_cputime = start + throttle; r->state = REBALANCE_THROTTLED; @@ -229,7 +229,7 @@ static int bch2_rebalance_thread(void *arg) max(p.dev_most_full_percent, 1U) / max(w.dev_most_full_percent, 1U)); - io_start = atomic_long_read(&clock->now); + io_start = atomic64_read(&clock->now); p = w; prev_start = start; prev_cputime = cputime; @@ -274,7 +274,7 @@ void bch2_rebalance_work_to_text(struct printbuf *out, struct bch_fs *c) case REBALANCE_THROTTLED: bch2_hprint(&PBUF(h1), (r->throttled_until_iotime - - atomic_long_read(&c->io_clock[WRITE].now)) << 9); + atomic64_read(&c->io_clock[WRITE].now)) << 9); pr_buf(out, "throttled for %lu sec or %s io\n", (r->throttled_until_cputime - jiffies) / HZ, h1); |