linux-fsdevel.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: "Darrick J. Wong" <djwong@kernel.org>
To: Kent Overstreet <kent.overstreet@linux.dev>
Cc: Joshua Ashton <joshua@froggi.es>,
	linux-kernel@vger.kernel.org, linux-fsdevel@vger.kernel.org,
	linux-bcache@vger.kernel.org, linux-bcachefs@vger.kernel.org
Subject: Re: [PATCH 3/5] bcachefs: bch2_time_stats_to_seq_buf()
Date: Sat, 27 Jan 2024 09:45:43 -0800	[thread overview]
Message-ID: <20240127174543.GD6184@frogsfrogsfrogs> (raw)
In-Reply-To: <5efhxgruvtlrjiqe52kgmnduguw3tmg2jjugfzi7dhbywljrwi@zu2ndkvzunzs>

On Sat, Jan 27, 2024 at 10:20:48AM -0500, Kent Overstreet wrote:
> On Sat, Jan 27, 2024 at 03:49:29AM +0000, Joshua Ashton wrote:
> > Kernel patches need descriptions.
> 
> I think this one was pretty clear from the name and type signature :)

I was mildly confused by this patch until I read the /next/ patch and
realized "Oh, patch 3 ports the timestats rendering code to seq_buf so
that patch 4 can lift it to lib/."

But I also thought the purpose of all this was kinda obvious from the
cover letter, which was why I didn't say anything yesterday afternoon.

I really don't care about these sorts of things.  IMO the obviously
interesting meat of the patchset is the patch that hoisted this to lib/
and the critical review is brainstorming how I'd use this new piece of
functionality, deciding if the api was ergonomic enough to start
prototyping upon, and sending a conditional RVB if I demonstrated enough
understanding of the problem being solved.

As for rearranging deck chairs inside fs/bcachefs, that's Kent's thing.

--D

> > 
> > On January 26, 2024 10:06:53 PM GMT, Kent Overstreet <kent.overstreet@linux.dev> wrote:
> > >Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
> > >---
> > > fs/bcachefs/super.c |   2 +
> > > fs/bcachefs/util.c  | 129 +++++++++++++++++++++++++++++++++++++++-----
> > > fs/bcachefs/util.h  |   4 ++
> > > 3 files changed, 121 insertions(+), 14 deletions(-)
> > >
> > >diff --git a/fs/bcachefs/super.c b/fs/bcachefs/super.c
> > >index da8697c79a97..e74534096cb5 100644
> > >--- a/fs/bcachefs/super.c
> > >+++ b/fs/bcachefs/super.c
> > >@@ -1262,6 +1262,8 @@ static struct bch_dev *__bch2_dev_alloc(struct bch_fs *c,
> > > 
> > > 	bch2_time_stats_init(&ca->io_latency[READ]);
> > > 	bch2_time_stats_init(&ca->io_latency[WRITE]);
> > >+	ca->io_latency[READ].quantiles_enabled = true;
> > >+	ca->io_latency[WRITE].quantiles_enabled = true;
> > > 
> > > 	ca->mi = bch2_mi_to_cpu(member);
> > > 
> > >diff --git a/fs/bcachefs/util.c b/fs/bcachefs/util.c
> > >index c7cf9c6fcf9a..f2c8550c1331 100644
> > >--- a/fs/bcachefs/util.c
> > >+++ b/fs/bcachefs/util.c
> > >@@ -505,10 +505,8 @@ static inline void pr_name_and_units(struct printbuf *out, const char *name, u64
> > > 
> > > void bch2_time_stats_to_text(struct printbuf *out, struct bch2_time_stats *stats)
> > > {
> > >-	const struct time_unit *u;
> > > 	s64 f_mean = 0, d_mean = 0;
> > >-	u64 q, last_q = 0, f_stddev = 0, d_stddev = 0;
> > >-	int i;
> > >+	u64 f_stddev = 0, d_stddev = 0;
> > > 
> > > 	if (stats->buffer) {
> > > 		int cpu;
> > >@@ -607,19 +605,122 @@ void bch2_time_stats_to_text(struct printbuf *out, struct bch2_time_stats *stats
> > > 
> > > 	printbuf_tabstops_reset(out);
> > > 
> > >-	i = eytzinger0_first(NR_QUANTILES);
> > >-	u = pick_time_units(stats->quantiles.entries[i].m);
> > >+	if (stats->quantiles_enabled) {
> > >+		int i = eytzinger0_first(NR_QUANTILES);
> > >+		const struct time_unit *u =
> > >+			pick_time_units(stats->quantiles.entries[i].m);
> > >+		u64 last_q = 0;
> > >+
> > >+		prt_printf(out, "quantiles (%s):\t", u->name);
> > >+		eytzinger0_for_each(i, NR_QUANTILES) {
> > >+			bool is_last = eytzinger0_next(i, NR_QUANTILES) == -1;
> > >+
> > >+			u64 q = max(stats->quantiles.entries[i].m, last_q);
> > >+			prt_printf(out, "%llu ", div_u64(q, u->nsecs));
> > >+			if (is_last)
> > >+				prt_newline(out);
> > >+			last_q = q;
> > >+		}
> > >+	}
> > >+}
> > >+
> > >+#include <linux/seq_buf.h>
> > >+
> > >+static void seq_buf_time_units_aligned(struct seq_buf *out, u64 ns)
> > >+{
> > >+	const struct time_unit *u = pick_time_units(ns);
> > >+
> > >+	seq_buf_printf(out, "%8llu %s", div64_u64(ns, u->nsecs), u->name);
> > >+}
> > >+
> > >+void bch2_time_stats_to_seq_buf(struct seq_buf *out, struct bch2_time_stats *stats)
> > >+{
> > >+	s64 f_mean = 0, d_mean = 0;
> > >+	u64 f_stddev = 0, d_stddev = 0;
> > >+
> > >+	if (stats->buffer) {
> > >+		int cpu;
> > > 
> > >-	prt_printf(out, "quantiles (%s):\t", u->name);
> > >-	eytzinger0_for_each(i, NR_QUANTILES) {
> > >-		bool is_last = eytzinger0_next(i, NR_QUANTILES) == -1;
> > >+		spin_lock_irq(&stats->lock);
> > >+		for_each_possible_cpu(cpu)
> > >+			__bch2_time_stats_clear_buffer(stats, per_cpu_ptr(stats->buffer, cpu));
> > >+		spin_unlock_irq(&stats->lock);
> > >+	}
> > > 
> > >-		q = max(stats->quantiles.entries[i].m, last_q);
> > >-		prt_printf(out, "%llu ",
> > >-		       div_u64(q, u->nsecs));
> > >-		if (is_last)
> > >-			prt_newline(out);
> > >-		last_q = q;
> > >+	/*
> > >+	 * avoid divide by zero
> > >+	 */
> > >+	if (stats->freq_stats.n) {
> > >+		f_mean = mean_and_variance_get_mean(stats->freq_stats);
> > >+		f_stddev = mean_and_variance_get_stddev(stats->freq_stats);
> > >+		d_mean = mean_and_variance_get_mean(stats->duration_stats);
> > >+		d_stddev = mean_and_variance_get_stddev(stats->duration_stats);
> > >+	}
> > >+
> > >+	seq_buf_printf(out, "count: %llu\n", stats->duration_stats.n);
> > >+
> > >+	seq_buf_printf(out, "                       since mount        recent\n");
> > >+
> > >+	seq_buf_printf(out, "duration of events\n");
> > >+
> > >+	seq_buf_printf(out, "  min:                     ");
> > >+	seq_buf_time_units_aligned(out, stats->min_duration);
> > >+	seq_buf_printf(out, "\n");
> > >+
> > >+	seq_buf_printf(out, "  max:                     ");
> > >+	seq_buf_time_units_aligned(out, stats->max_duration);
> > >+	seq_buf_printf(out, "\n");
> > >+
> > >+	seq_buf_printf(out, "  total:                   ");
> > >+	seq_buf_time_units_aligned(out, stats->total_duration);
> > >+	seq_buf_printf(out, "\n");
> > >+
> > >+	seq_buf_printf(out, "  mean:                    ");
> > >+	seq_buf_time_units_aligned(out, d_mean);
> > >+	seq_buf_time_units_aligned(out, mean_and_variance_weighted_get_mean(stats->duration_stats_weighted));
> > >+	seq_buf_printf(out, "\n");
> > >+
> > >+	seq_buf_printf(out, "  stddev:                  ");
> > >+	seq_buf_time_units_aligned(out, d_stddev);
> > >+	seq_buf_time_units_aligned(out, mean_and_variance_weighted_get_stddev(stats->duration_stats_weighted));
> > >+	seq_buf_printf(out, "\n");
> > >+
> > >+	seq_buf_printf(out, "time between events\n");
> > >+
> > >+	seq_buf_printf(out, "  min:                     ");
> > >+	seq_buf_time_units_aligned(out, stats->min_freq);
> > >+	seq_buf_printf(out, "\n");
> > >+
> > >+	seq_buf_printf(out, "  max:                     ");
> > >+	seq_buf_time_units_aligned(out, stats->max_freq);
> > >+	seq_buf_printf(out, "\n");
> > >+
> > >+	seq_buf_printf(out, "  mean:                    ");
> > >+	seq_buf_time_units_aligned(out, f_mean);
> > >+	seq_buf_time_units_aligned(out, mean_and_variance_weighted_get_mean(stats->freq_stats_weighted));
> > >+	seq_buf_printf(out, "\n");
> > >+
> > >+	seq_buf_printf(out, "  stddev:                  ");
> > >+	seq_buf_time_units_aligned(out, f_stddev);
> > >+	seq_buf_time_units_aligned(out, mean_and_variance_weighted_get_stddev(stats->freq_stats_weighted));
> > >+	seq_buf_printf(out, "\n");
> > >+
> > >+	if (stats->quantiles_enabled) {
> > >+		int i = eytzinger0_first(NR_QUANTILES);
> > >+		const struct time_unit *u =
> > >+			pick_time_units(stats->quantiles.entries[i].m);
> > >+		u64 last_q = 0;
> > >+
> > >+		prt_printf(out, "quantiles (%s):\t", u->name);
> > >+		eytzinger0_for_each(i, NR_QUANTILES) {
> > >+			bool is_last = eytzinger0_next(i, NR_QUANTILES) == -1;
> > >+
> > >+			u64 q = max(stats->quantiles.entries[i].m, last_q);
> > >+			seq_buf_printf(out, "%llu ", div_u64(q, u->nsecs));
> > >+			if (is_last)
> > >+				seq_buf_printf(out, "\n");
> > >+			last_q = q;
> > >+		}
> > > 	}
> > > }
> > > #else
> > >diff --git a/fs/bcachefs/util.h b/fs/bcachefs/util.h
> > >index c3b11c3d24ea..7ff2d4fe26f6 100644
> > >--- a/fs/bcachefs/util.h
> > >+++ b/fs/bcachefs/util.h
> > >@@ -382,6 +382,7 @@ struct bch2_time_stat_buffer {
> > > 
> > > struct bch2_time_stats {
> > > 	spinlock_t	lock;
> > >+	bool		quantiles_enabled;
> > > 	/* all fields are in nanoseconds */
> > > 	u64             min_duration;
> > > 	u64		max_duration;
> > >@@ -435,6 +436,9 @@ static inline bool track_event_change(struct bch2_time_stats *stats,
> > > 
> > > void bch2_time_stats_to_text(struct printbuf *, struct bch2_time_stats *);
> > > 
> > >+struct seq_buf;
> > >+void bch2_time_stats_to_seq_buf(struct seq_buf *, struct bch2_time_stats *);
> > >+
> > > void bch2_time_stats_exit(struct bch2_time_stats *);
> > > void bch2_time_stats_init(struct bch2_time_stats *);
> > > 
> > >-- 
> > >2.43.0
> > >
> > >
> > 
> > - Joshie 🐸✨
> 

  reply	other threads:[~2024-01-27 17:45 UTC|newest]

Thread overview: 13+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2024-01-26 22:06 [PATCH 1/5] mean and variance: Promote to lib/math Kent Overstreet
2024-01-26 22:06 ` [PATCH 2/5] eytzinger: Promote to include/linux/ Kent Overstreet
2024-01-27  2:17   ` Darrick J. Wong
2024-01-26 22:06 ` [PATCH 3/5] bcachefs: bch2_time_stats_to_seq_buf() Kent Overstreet
     [not found]   ` <DA69BC9C-D7FF-48F7-9F6D-EF96DC4C5C7B@froggi.es>
2024-01-27 15:20     ` Kent Overstreet
2024-01-27 17:45       ` Darrick J. Wong [this message]
2024-01-26 22:06 ` [PATCH 4/5] time_stats: Promote to lib/ Kent Overstreet
2024-01-27  1:53   ` Darrick J. Wong
2024-02-01 19:57     ` Darrick J. Wong
2024-01-26 22:06 ` [PATCH 5/5] bcache: Convert to lib/time_stats Kent Overstreet
2024-01-27  7:09   ` Coly Li
2024-02-02  7:30 ` [PATCH 1/5] mean and variance: Promote to lib/math Darrick J. Wong
2024-02-02  9:56   ` 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=20240127174543.GD6184@frogsfrogsfrogs \
    --to=djwong@kernel.org \
    --cc=joshua@froggi.es \
    --cc=kent.overstreet@linux.dev \
    --cc=linux-bcache@vger.kernel.org \
    --cc=linux-bcachefs@vger.kernel.org \
    --cc=linux-fsdevel@vger.kernel.org \
    --cc=linux-kernel@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).