X-Git-Url: https://git.sesse.net/?a=blobdiff_plain;f=libbcachefs%2Futil.c;h=539735033947b91f4bd26768d51ca3049512dd45;hb=e773e86495534ede8efc0f3059cf0939c374a925;hp=fea80e248667c5efc3786aeb1676c89ad7cca218;hpb=35fca2f044d375b1590f499cfd34bef38ca0f8f1;p=bcachefs-tools-debian

diff --git a/libbcachefs/util.c b/libbcachefs/util.c
index fea80e2..5397350 100644
--- a/libbcachefs/util.c
+++ b/libbcachefs/util.c
@@ -1,3 +1,4 @@
+// SPDX-License-Identifier: GPL-2.0
 /*
  * random utiility code, for bcache but in theory not specific to bcache
  *
@@ -7,8 +8,10 @@
 
 #include <linux/bio.h>
 #include <linux/blkdev.h>
+#include <linux/console.h>
 #include <linux/ctype.h>
 #include <linux/debugfs.h>
+#include <linux/eytzinger.h>
 #include <linux/freezer.h>
 #include <linux/kthread.h>
 #include <linux/log2.h>
@@ -20,22 +23,25 @@
 #include <linux/string.h>
 #include <linux/types.h>
 #include <linux/sched/clock.h>
+#include <linux/mean_and_variance.h>
 
-#include "eytzinger.h"
 #include "util.h"
 
 static const char si_units[] = "?kMGTPEZY";
 
-static int __bch2_strtoh(const char *cp, u64 *res,
-			 u64 t_max, bool t_signed)
+/* string_get_size units: */
+static const char *const units_2[] = {
+	"B", "KiB", "MiB", "GiB", "TiB", "PiB", "EiB", "ZiB", "YiB"
+};
+static const char *const units_10[] = {
+	"B", "kB", "MB", "GB", "TB", "PB", "EB", "ZB", "YB"
+};
+
+static int parse_u64(const char *cp, u64 *res)
 {
-	bool positive = *cp != '-';
-	unsigned u;
+	const char *start = cp;
 	u64 v = 0;
 
-	if (*cp == '+' || *cp == '-')
-		cp++;
-
 	if (!isdigit(*cp))
 		return -EINVAL;
 
@@ -49,22 +55,122 @@ static int __bch2_strtoh(const char *cp, u64 *res,
 		cp++;
 	} while (isdigit(*cp));
 
+	*res = v;
+	return cp - start;
+}
+
+static int bch2_pow(u64 n, u64 p, u64 *res)
+{
+	*res = 1;
+
+	while (p--) {
+		if (*res > div_u64(U64_MAX, n))
+			return -ERANGE;
+		*res *= n;
+	}
+	return 0;
+}
+
+static int parse_unit_suffix(const char *cp, u64 *res)
+{
+	const char *start = cp;
+	u64 base = 1024;
+	unsigned u;
+	int ret;
+
+	if (*cp == ' ')
+		cp++;
+
 	for (u = 1; u < strlen(si_units); u++)
 		if (*cp == si_units[u]) {
 			cp++;
 			goto got_unit;
 		}
-	u = 0;
+
+	for (u = 0; u < ARRAY_SIZE(units_2); u++)
+		if (!strncmp(cp, units_2[u], strlen(units_2[u]))) {
+			cp += strlen(units_2[u]);
+			goto got_unit;
+		}
+
+	for (u = 0; u < ARRAY_SIZE(units_10); u++)
+		if (!strncmp(cp, units_10[u], strlen(units_10[u]))) {
+			cp += strlen(units_10[u]);
+			base = 1000;
+			goto got_unit;
+		}
+
+	*res = 1;
+	return 0;
 got_unit:
-	if (*cp == '\n')
+	ret = bch2_pow(base, u, res);
+	if (ret)
+		return ret;
+
+	return cp - start;
+}
+
+#define parse_or_ret(cp, _f)			\
+do {						\
+	int _ret = _f;				\
+	if (_ret < 0)				\
+		return _ret;			\
+	cp += _ret;				\
+} while (0)
+
+static int __bch2_strtou64_h(const char *cp, u64 *res)
+{
+	const char *start = cp;
+	u64 v = 0, b, f_n = 0, f_d = 1;
+	int ret;
+
+	parse_or_ret(cp, parse_u64(cp, &v));
+
+	if (*cp == '.') {
 		cp++;
-	if (*cp)
-		return -EINVAL;
+		ret = parse_u64(cp, &f_n);
+		if (ret < 0)
+			return ret;
+		cp += ret;
+
+		ret = bch2_pow(10, ret, &f_d);
+		if (ret)
+			return ret;
+	}
 
-	if (fls64(v) + u * 10 > 64)
+	parse_or_ret(cp, parse_unit_suffix(cp, &b));
+
+	if (v > div_u64(U64_MAX, b))
+		return -ERANGE;
+	v *= b;
+
+	if (f_n > div_u64(U64_MAX, b))
 		return -ERANGE;
 
-	v <<= u * 10;
+	f_n = div_u64(f_n * b, f_d);
+	if (v + f_n < v)
+		return -ERANGE;
+	v += f_n;
+
+	*res = v;
+	return cp - start;
+}
+
+static int __bch2_strtoh(const char *cp, u64 *res,
+			 u64 t_max, bool t_signed)
+{
+	bool positive = *cp != '-';
+	u64 v = 0;
+
+	if (*cp == '+' || *cp == '-')
+		cp++;
+
+	parse_or_ret(cp, __bch2_strtou64_h(cp, &v));
+
+	if (*cp == '\n')
+		cp++;
+	if (*cp)
+		return -EINVAL;
 
 	if (positive) {
 		if (v > t_max)
@@ -85,7 +191,7 @@ got_unit:
 #define STRTO_H(name, type)					\
 int bch2_ ## name ## _h(const char *cp, type *res)		\
 {								\
-	u64 v;							\
+	u64 v = 0;						\
 	int ret = __bch2_strtoh(cp, &v, ANYSINT_MAX(type),	\
 			ANYSINT_MAX(type) != ((type) ~0ULL));	\
 	*res = v;						\
@@ -98,62 +204,10 @@ STRTO_H(strtoll, long long)
 STRTO_H(strtoull, unsigned long long)
 STRTO_H(strtou64, u64)
 
-void bch2_hprint(struct printbuf *buf, s64 v)
-{
-	int u, t = 0;
-
-	for (u = 0; v >= 1024 || v <= -1024; u++) {
-		t = v & ~(~0U << 10);
-		v >>= 10;
-	}
-
-	pr_buf(buf, "%lli", v);
-
-	/*
-	 * 103 is magic: t is in the range [-1023, 1023] and we want
-	 * to turn it into [-9, 9]
-	 */
-	if (u && v < 100 && v > -100)
-		pr_buf(buf, ".%i", t / 103);
-	if (u)
-		pr_buf(buf, "%c", si_units[u]);
-}
-
-void bch2_string_opt_to_text(struct printbuf *out,
-			     const char * const list[],
-			     size_t selected)
-{
-	size_t i;
-
-	for (i = 0; list[i]; i++)
-		pr_buf(out, i == selected ? "[%s] " : "%s ", list[i]);
-}
-
-void bch2_flags_to_text(struct printbuf *out,
-			const char * const list[], u64 flags)
-{
-	unsigned bit, nr = 0;
-	bool first = true;
-
-	if (out->pos != out->end)
-		*out->pos = '\0';
-
-	while (list[nr])
-		nr++;
-
-	while (flags && (bit = __ffs(flags)) < nr) {
-		pr_buf(out, "%s", list[bit]);
-		if (!first)
-			pr_buf(out, ",");
-		first = false;
-		flags ^= 1 << bit;
-	}
-}
-
 u64 bch2_read_flag_list(char *opt, const char * const list[])
 {
 	u64 ret = 0;
-	char *p, *s, *d = kstrndup(opt, PAGE_SIZE - 1, GFP_KERNEL);
+	char *p, *s, *d = kstrdup(opt, GFP_KERNEL);
 
 	if (!d)
 		return -ENOMEM;
@@ -162,6 +216,7 @@ u64 bch2_read_flag_list(char *opt, const char * const list[])
 
 	while ((p = strsep(&s, ","))) {
 		int flag = match_string(list, -1, p);
+
 		if (flag < 0) {
 			ret = -1;
 			break;
@@ -186,196 +241,257 @@ bool bch2_is_zero(const void *_p, size_t n)
 	return true;
 }
 
-static void bch2_quantiles_update(struct quantiles *q, u64 v)
+void bch2_prt_u64_base2_nbits(struct printbuf *out, u64 v, unsigned nr_bits)
 {
-	unsigned i = 0;
+	while (nr_bits)
+		prt_char(out, '0' + ((v >> --nr_bits) & 1));
+}
 
-	while (i < ARRAY_SIZE(q->entries)) {
-		struct quantile_entry *e = q->entries + i;
+void bch2_prt_u64_base2(struct printbuf *out, u64 v)
+{
+	bch2_prt_u64_base2_nbits(out, v, fls64(v) ?: 1);
+}
 
-		if (unlikely(!e->step)) {
-			e->m = v;
-			e->step = max_t(unsigned, v / 2, 1024);
-		} else if (e->m > v) {
-			e->m = e->m >= e->step
-				? e->m - e->step
-				: 0;
-		} else if (e->m < v) {
-			e->m = e->m + e->step > e->m
-				? e->m + e->step
-				: U32_MAX;
-		}
+void bch2_print_string_as_lines(const char *prefix, const char *lines)
+{
+	const char *p;
 
-		if ((e->m > v ? e->m - v : v - e->m) < e->step)
-			e->step = max_t(unsigned, e->step / 2, 1);
+	if (!lines) {
+		printk("%s (null)\n", prefix);
+		return;
+	}
 
-		if (v >= e->m)
+	console_lock();
+	while (1) {
+		p = strchrnul(lines, '\n');
+		printk("%s%.*s\n", prefix, (int) (p - lines), lines);
+		if (!*p)
 			break;
-
-		i = eytzinger0_child(i, v > e->m);
+		lines = p + 1;
 	}
+	console_unlock();
 }
 
-/* time stats: */
-
-static void bch2_time_stats_update_one(struct time_stats *stats,
-				       u64 start, u64 end)
+int bch2_save_backtrace(bch_stacktrace *stack, struct task_struct *task, unsigned skipnr,
+			gfp_t gfp)
 {
-	u64 duration, freq;
-
-	duration	= time_after64(end, start)
-		? end - start : 0;
-	freq		= time_after64(end, stats->last_event)
-		? end - stats->last_event : 0;
+#ifdef CONFIG_STACKTRACE
+	unsigned nr_entries = 0;
 
-	stats->count++;
+	stack->nr = 0;
+	int ret = darray_make_room_gfp(stack, 32, gfp);
+	if (ret)
+		return ret;
 
-	stats->average_duration = stats->average_duration
-		? ewma_add(stats->average_duration, duration, 6)
-		: duration;
+	if (!down_read_trylock(&task->signal->exec_update_lock))
+		return -1;
 
-	stats->average_frequency = stats->average_frequency
-		? ewma_add(stats->average_frequency, freq, 6)
-		: freq;
-
-	stats->max_duration = max(stats->max_duration, duration);
+	do {
+		nr_entries = stack_trace_save_tsk(task, stack->data, stack->size, skipnr + 1);
+	} while (nr_entries == stack->size &&
+		 !(ret = darray_make_room(stack, stack->size * 2)));
 
-	stats->last_event = end;
+	stack->nr = nr_entries;
+	up_read(&task->signal->exec_update_lock);
 
-	bch2_quantiles_update(&stats->quantiles, duration);
+	return ret;
+#else
+	return 0;
+#endif
 }
 
-void __bch2_time_stats_update(struct time_stats *stats, u64 start, u64 end)
+void bch2_prt_backtrace(struct printbuf *out, bch_stacktrace *stack)
 {
-	unsigned long flags;
-
-	if (!stats->buffer) {
-		spin_lock_irqsave(&stats->lock, flags);
-		bch2_time_stats_update_one(stats, start, end);
-
-		if (stats->average_frequency < 32 &&
-		    stats->count > 1024)
-			stats->buffer =
-				alloc_percpu_gfp(struct time_stat_buffer,
-						 GFP_ATOMIC);
-		spin_unlock_irqrestore(&stats->lock, flags);
-	} else {
-		struct time_stat_buffer_entry *i;
-		struct time_stat_buffer *b;
-
-		preempt_disable();
-		b = this_cpu_ptr(stats->buffer);
-
-		BUG_ON(b->nr >= ARRAY_SIZE(b->entries));
-		b->entries[b->nr++] = (struct time_stat_buffer_entry) {
-			.start = start,
-			.end = end
-		};
-
-		if (b->nr == ARRAY_SIZE(b->entries)) {
-			spin_lock_irqsave(&stats->lock, flags);
-			for (i = b->entries;
-			     i < b->entries + ARRAY_SIZE(b->entries);
-			     i++)
-				bch2_time_stats_update_one(stats, i->start, i->end);
-			spin_unlock_irqrestore(&stats->lock, flags);
-
-			b->nr = 0;
-		}
-
-		preempt_enable();
+	darray_for_each(*stack, i) {
+		prt_printf(out, "[<0>] %pB", (void *) *i);
+		prt_newline(out);
 	}
 }
 
-static const struct time_unit {
-	const char	*name;
-	u32		nsecs;
-} time_units[] = {
-	{ "ns",		1		},
-	{ "us",		NSEC_PER_USEC	},
-	{ "ms",		NSEC_PER_MSEC	},
-	{ "sec",	NSEC_PER_SEC	},
-};
-
-static const struct time_unit *pick_time_units(u64 ns)
+int bch2_prt_task_backtrace(struct printbuf *out, struct task_struct *task, unsigned skipnr, gfp_t gfp)
 {
-	const struct time_unit *u;
+	bch_stacktrace stack = { 0 };
+	int ret = bch2_save_backtrace(&stack, task, skipnr + 1, gfp);
 
-	for (u = time_units;
-	     u + 1 < time_units + ARRAY_SIZE(time_units) &&
-	     ns >= u[1].nsecs << 1;
-	     u++)
-		;
+	bch2_prt_backtrace(out, &stack);
+	darray_exit(&stack);
+	return ret;
+}
 
-	return u;
+#ifndef __KERNEL__
+#include <time.h>
+void bch2_prt_datetime(struct printbuf *out, time64_t sec)
+{
+	time_t t = sec;
+	char buf[64];
+	ctime_r(&t, buf);
+	strim(buf);
+	prt_str(out, buf);
+}
+#else
+void bch2_prt_datetime(struct printbuf *out, time64_t sec)
+{
+	char buf[64];
+	snprintf(buf, sizeof(buf), "%ptT", &sec);
+	prt_u64(out, sec);
 }
+#endif
 
-static void pr_time_units(struct printbuf *out, u64 ns)
+void bch2_pr_time_units(struct printbuf *out, u64 ns)
 {
 	const struct time_unit *u = pick_time_units(ns);
 
-	pr_buf(out, "%llu %s", div_u64(ns, u->nsecs), u->name);
+	prt_printf(out, "%llu %s", div_u64(ns, u->nsecs), u->name);
 }
 
-size_t bch2_time_stats_print(struct time_stats *stats, char *buf, size_t len)
+static void bch2_pr_time_units_aligned(struct printbuf *out, u64 ns)
 {
-	struct printbuf out = _PBUF(buf, len);
-	const struct time_unit *u;
-	u64 freq = READ_ONCE(stats->average_frequency);
-	u64 q, last_q = 0;
-	int i;
-
-	pr_buf(&out, "count:\t\t%llu\n",
-			 stats->count);
-	pr_buf(&out, "rate:\t\t%llu/sec\n",
-	       freq ?  div64_u64(NSEC_PER_SEC, freq) : 0);
+	const struct time_unit *u = pick_time_units(ns);
 
-	pr_buf(&out, "frequency:\t");
-	pr_time_units(&out, freq);
+	prt_printf(out, "%llu ", div64_u64(ns, u->nsecs));
+	prt_tab_rjust(out);
+	prt_printf(out, "%s", u->name);
+}
 
-	pr_buf(&out, "\navg duration:\t");
-	pr_time_units(&out, stats->average_duration);
+static inline void pr_name_and_units(struct printbuf *out, const char *name, u64 ns)
+{
+	prt_str(out, name);
+	prt_tab(out);
+	bch2_pr_time_units_aligned(out, ns);
+	prt_newline(out);
+}
 
-	pr_buf(&out, "\nmax duration:\t");
-	pr_time_units(&out, stats->max_duration);
+#define TABSTOP_SIZE 12
 
-	i = eytzinger0_first(NR_QUANTILES);
-	u = pick_time_units(stats->quantiles.entries[i].m);
+void bch2_time_stats_to_text(struct printbuf *out, struct time_stats *stats)
+{
+	struct quantiles *quantiles = time_stats_to_quantiles(stats);
+	s64 f_mean = 0, d_mean = 0;
+	u64 f_stddev = 0, d_stddev = 0;
 
-	pr_buf(&out, "\nquantiles (%s):\t", u->name);
-	eytzinger0_for_each(i, NR_QUANTILES) {
-		bool is_last = eytzinger0_next(i, NR_QUANTILES) == -1;
+	if (stats->buffer) {
+		int cpu;
 
-		q = max(stats->quantiles.entries[i].m, last_q);
-		pr_buf(&out, "%llu%s",
-		       div_u64(q, u->nsecs),
-		       is_last ? "\n" : " ");
-		last_q = q;
+		spin_lock_irq(&stats->lock);
+		for_each_possible_cpu(cpu)
+			__time_stats_clear_buffer(stats, per_cpu_ptr(stats->buffer, cpu));
+		spin_unlock_irq(&stats->lock);
 	}
 
-	return out.pos - buf;
-}
-
-void bch2_time_stats_exit(struct time_stats *stats)
-{
-	free_percpu(stats->buffer);
-}
+	/*
+	 * 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);
+	}
 
-void bch2_time_stats_init(struct time_stats *stats)
-{
-	memset(stats, 0, sizeof(*stats));
-	spin_lock_init(&stats->lock);
+	printbuf_tabstop_push(out, out->indent + TABSTOP_SIZE);
+	prt_printf(out, "count:");
+	prt_tab(out);
+	prt_printf(out, "%llu ",
+			 stats->duration_stats.n);
+	printbuf_tabstop_pop(out);
+	prt_newline(out);
+
+	printbuf_tabstops_reset(out);
+
+	printbuf_tabstop_push(out, out->indent + 20);
+	printbuf_tabstop_push(out, TABSTOP_SIZE + 2);
+	printbuf_tabstop_push(out, 0);
+	printbuf_tabstop_push(out, TABSTOP_SIZE + 2);
+
+	prt_tab(out);
+	prt_printf(out, "since mount");
+	prt_tab_rjust(out);
+	prt_tab(out);
+	prt_printf(out, "recent");
+	prt_tab_rjust(out);
+	prt_newline(out);
+
+	printbuf_tabstops_reset(out);
+	printbuf_tabstop_push(out, out->indent + 20);
+	printbuf_tabstop_push(out, TABSTOP_SIZE);
+	printbuf_tabstop_push(out, 2);
+	printbuf_tabstop_push(out, TABSTOP_SIZE);
+
+	prt_printf(out, "duration of events");
+	prt_newline(out);
+	printbuf_indent_add(out, 2);
+
+	pr_name_and_units(out, "min:", stats->min_duration);
+	pr_name_and_units(out, "max:", stats->max_duration);
+	pr_name_and_units(out, "total:", stats->total_duration);
+
+	prt_printf(out, "mean:");
+	prt_tab(out);
+	bch2_pr_time_units_aligned(out, d_mean);
+	prt_tab(out);
+	bch2_pr_time_units_aligned(out, mean_and_variance_weighted_get_mean(stats->duration_stats_weighted, TIME_STATS_MV_WEIGHT));
+	prt_newline(out);
+
+	prt_printf(out, "stddev:");
+	prt_tab(out);
+	bch2_pr_time_units_aligned(out, d_stddev);
+	prt_tab(out);
+	bch2_pr_time_units_aligned(out, mean_and_variance_weighted_get_stddev(stats->duration_stats_weighted, TIME_STATS_MV_WEIGHT));
+
+	printbuf_indent_sub(out, 2);
+	prt_newline(out);
+
+	prt_printf(out, "time between events");
+	prt_newline(out);
+	printbuf_indent_add(out, 2);
+
+	pr_name_and_units(out, "min:", stats->min_freq);
+	pr_name_and_units(out, "max:", stats->max_freq);
+
+	prt_printf(out, "mean:");
+	prt_tab(out);
+	bch2_pr_time_units_aligned(out, f_mean);
+	prt_tab(out);
+	bch2_pr_time_units_aligned(out, mean_and_variance_weighted_get_mean(stats->freq_stats_weighted, TIME_STATS_MV_WEIGHT));
+	prt_newline(out);
+
+	prt_printf(out, "stddev:");
+	prt_tab(out);
+	bch2_pr_time_units_aligned(out, f_stddev);
+	prt_tab(out);
+	bch2_pr_time_units_aligned(out, mean_and_variance_weighted_get_stddev(stats->freq_stats_weighted, TIME_STATS_MV_WEIGHT));
+
+	printbuf_indent_sub(out, 2);
+	prt_newline(out);
+
+	printbuf_tabstops_reset(out);
+
+	if (quantiles) {
+		int i = eytzinger0_first(NR_QUANTILES);
+		const struct time_unit *u =
+			pick_time_units(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(quantiles->entries[i].m, last_q);
+			prt_printf(out, "%llu ", div_u64(q, u->nsecs));
+			if (is_last)
+				prt_newline(out);
+			last_q = q;
+		}
+	}
 }
 
 /* ratelimit: */
 
 /**
  * bch2_ratelimit_delay() - return how long to delay until the next time to do
- * some work
- *
- * @d - the struct bch_ratelimit to update
- *
- * Returns the amount of time to delay by, in jiffies
+ *		some work
+ * @d:		the struct bch_ratelimit to update
+ * Returns:	the amount of time to delay by, in jiffies
  */
 u64 bch2_ratelimit_delay(struct bch_ratelimit *d)
 {
@@ -388,9 +504,8 @@ u64 bch2_ratelimit_delay(struct bch_ratelimit *d)
 
 /**
  * bch2_ratelimit_increment() - increment @d by the amount of work done
- *
- * @d - the struct bch_ratelimit to update
- * @done - the amount of work done, in arbitrary units
+ * @d:		the struct bch_ratelimit to update
+ * @done:	the amount of work done, in arbitrary units
  */
 void bch2_ratelimit_increment(struct bch_ratelimit *d, u64 done)
 {
@@ -469,82 +584,79 @@ void bch2_pd_controller_init(struct bch_pd_controller *pd)
 	pd->backpressure	= 1;
 }
 
-size_t bch2_pd_controller_print_debug(struct bch_pd_controller *pd, char *buf)
+void bch2_pd_controller_debug_to_text(struct printbuf *out, struct bch_pd_controller *pd)
 {
-	/* 2^64 - 1 is 20 digits, plus null byte */
-	char rate[21];
-	char actual[21];
-	char target[21];
-	char proportional[21];
-	char derivative[21];
-	char change[21];
-	s64 next_io;
+	if (!out->nr_tabstops)
+		printbuf_tabstop_push(out, 20);
 
-	bch2_hprint(&PBUF(rate),	pd->rate.rate);
-	bch2_hprint(&PBUF(actual),	pd->last_actual);
-	bch2_hprint(&PBUF(target),	pd->last_target);
-	bch2_hprint(&PBUF(proportional), pd->last_proportional);
-	bch2_hprint(&PBUF(derivative),	pd->last_derivative);
-	bch2_hprint(&PBUF(change),	pd->last_change);
+	prt_printf(out, "rate:");
+	prt_tab(out);
+	prt_human_readable_s64(out, pd->rate.rate);
+	prt_newline(out);
 
-	next_io = div64_s64(pd->rate.next - local_clock(), NSEC_PER_MSEC);
+	prt_printf(out, "target:");
+	prt_tab(out);
+	prt_human_readable_u64(out, pd->last_target);
+	prt_newline(out);
 
-	return sprintf(buf,
-		       "rate:\t\t%s/sec\n"
-		       "target:\t\t%s\n"
-		       "actual:\t\t%s\n"
-		       "proportional:\t%s\n"
-		       "derivative:\t%s\n"
-		       "change:\t\t%s/sec\n"
-		       "next io:\t%llims\n",
-		       rate, target, actual, proportional,
-		       derivative, change, next_io);
-}
+	prt_printf(out, "actual:");
+	prt_tab(out);
+	prt_human_readable_u64(out, pd->last_actual);
+	prt_newline(out);
 
-/* misc: */
+	prt_printf(out, "proportional:");
+	prt_tab(out);
+	prt_human_readable_s64(out, pd->last_proportional);
+	prt_newline(out);
 
-void bch2_bio_map(struct bio *bio, void *base)
-{
-	size_t size = bio->bi_iter.bi_size;
-	struct bio_vec *bv = bio->bi_io_vec;
+	prt_printf(out, "derivative:");
+	prt_tab(out);
+	prt_human_readable_s64(out, pd->last_derivative);
+	prt_newline(out);
 
-	BUG_ON(!bio->bi_iter.bi_size);
-	BUG_ON(bio->bi_vcnt);
-	BUG_ON(!bio->bi_max_vecs);
+	prt_printf(out, "change:");
+	prt_tab(out);
+	prt_human_readable_s64(out, pd->last_change);
+	prt_newline(out);
 
-	bv->bv_offset = base ? offset_in_page(base) : 0;
-	goto start;
+	prt_printf(out, "next io:");
+	prt_tab(out);
+	prt_printf(out, "%llims", div64_s64(pd->rate.next - local_clock(), NSEC_PER_MSEC));
+	prt_newline(out);
+}
 
-	for (; size; bio->bi_vcnt++, bv++) {
-		BUG_ON(bio->bi_vcnt >= bio->bi_max_vecs);
+/* misc: */
 
-		bv->bv_offset	= 0;
-start:		bv->bv_len	= min_t(size_t, PAGE_SIZE - bv->bv_offset,
-					size);
-		if (base) {
-			bv->bv_page = is_vmalloc_addr(base)
+void bch2_bio_map(struct bio *bio, void *base, size_t size)
+{
+	while (size) {
+		struct page *page = is_vmalloc_addr(base)
 				? vmalloc_to_page(base)
 				: virt_to_page(base);
+		unsigned offset = offset_in_page(base);
+		unsigned len = min_t(size_t, PAGE_SIZE - offset, size);
 
-			base += bv->bv_len;
-		}
-
-		size -= bv->bv_len;
+		BUG_ON(!bio_add_page(bio, page, len, offset));
+		size -= len;
+		base += len;
 	}
 }
 
-int bch2_bio_alloc_pages(struct bio *bio, gfp_t gfp_mask)
+int bch2_bio_alloc_pages(struct bio *bio, size_t size, gfp_t gfp_mask)
 {
-	int i;
-	struct bio_vec *bv;
+	while (size) {
+		struct page *page = alloc_pages(gfp_mask, 0);
+		unsigned len = min_t(size_t, PAGE_SIZE, size);
 
-	bio_for_each_segment_all(bv, bio, i) {
-		bv->bv_page = alloc_page(gfp_mask);
-		if (!bv->bv_page) {
-			while (--bv >= bio->bi_io_vec)
-				__free_page(bv->bv_page);
+		if (!page)
 			return -ENOMEM;
+
+		if (unlikely(!bio_add_page(bio, page, len, 0))) {
+			__free_page(page);
+			break;
 		}
+
+		size -= len;
 	}
 
 	return 0;
@@ -565,15 +677,16 @@ size_t bch2_rand_range(size_t max)
 	return rand;
 }
 
-void memcpy_to_bio(struct bio *dst, struct bvec_iter dst_iter, void *src)
+void memcpy_to_bio(struct bio *dst, struct bvec_iter dst_iter, const void *src)
 {
 	struct bio_vec bv;
 	struct bvec_iter iter;
 
 	__bio_for_each_segment(bv, dst, iter, dst_iter) {
-		void *dstp = kmap_atomic(bv.bv_page);
+		void *dstp = kmap_local_page(bv.bv_page);
+
 		memcpy(dstp + bv.bv_offset, src, bv.bv_len);
-		kunmap_atomic(dstp);
+		kunmap_local(dstp);
 
 		src += bv.bv_len;
 	}
@@ -585,194 +698,15 @@ void memcpy_from_bio(void *dst, struct bio *src, struct bvec_iter src_iter)
 	struct bvec_iter iter;
 
 	__bio_for_each_segment(bv, src, iter, src_iter) {
-		void *srcp = kmap_atomic(bv.bv_page);
+		void *srcp = kmap_local_page(bv.bv_page);
+
 		memcpy(dst, srcp + bv.bv_offset, bv.bv_len);
-		kunmap_atomic(srcp);
+		kunmap_local(srcp);
 
 		dst += bv.bv_len;
 	}
 }
 
-void bch_scnmemcpy(struct printbuf *out,
-		   const char *src, size_t len)
-{
-	size_t n = printbuf_remaining(out);
-
-	if (n) {
-		n = min(n - 1, len);
-		memcpy(out->pos, src, n);
-		out->pos += n;
-		*out->pos = '\0';
-	}
-}
-
-#include "eytzinger.h"
-
-static int alignment_ok(const void *base, size_t align)
-{
-	return IS_ENABLED(CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS) ||
-		((unsigned long)base & (align - 1)) == 0;
-}
-
-static void u32_swap(void *a, void *b, size_t size)
-{
-	u32 t = *(u32 *)a;
-	*(u32 *)a = *(u32 *)b;
-	*(u32 *)b = t;
-}
-
-static void u64_swap(void *a, void *b, size_t size)
-{
-	u64 t = *(u64 *)a;
-	*(u64 *)a = *(u64 *)b;
-	*(u64 *)b = t;
-}
-
-static void generic_swap(void *a, void *b, size_t size)
-{
-	char t;
-
-	do {
-		t = *(char *)a;
-		*(char *)a++ = *(char *)b;
-		*(char *)b++ = t;
-	} while (--size > 0);
-}
-
-static inline int do_cmp(void *base, size_t n, size_t size,
-			 int (*cmp_func)(const void *, const void *, size_t),
-			 size_t l, size_t r)
-{
-	return cmp_func(base + inorder_to_eytzinger0(l, n) * size,
-			base + inorder_to_eytzinger0(r, n) * size,
-			size);
-}
-
-static inline void do_swap(void *base, size_t n, size_t size,
-			   void (*swap_func)(void *, void *, size_t),
-			   size_t l, size_t r)
-{
-	swap_func(base + inorder_to_eytzinger0(l, n) * size,
-		  base + inorder_to_eytzinger0(r, n) * size,
-		  size);
-}
-
-void eytzinger0_sort(void *base, size_t n, size_t size,
-		     int (*cmp_func)(const void *, const void *, size_t),
-		     void (*swap_func)(void *, void *, size_t))
-{
-	int i, c, r;
-
-	if (!swap_func) {
-		if (size == 4 && alignment_ok(base, 4))
-			swap_func = u32_swap;
-		else if (size == 8 && alignment_ok(base, 8))
-			swap_func = u64_swap;
-		else
-			swap_func = generic_swap;
-	}
-
-	/* heapify */
-	for (i = n / 2 - 1; i >= 0; --i) {
-		for (r = i; r * 2 + 1 < n; r = c) {
-			c = r * 2 + 1;
-
-			if (c + 1 < n &&
-			    do_cmp(base, n, size, cmp_func, c, c + 1) < 0)
-				c++;
-
-			if (do_cmp(base, n, size, cmp_func, r, c) >= 0)
-				break;
-
-			do_swap(base, n, size, swap_func, r, c);
-		}
-	}
-
-	/* sort */
-	for (i = n - 1; i > 0; --i) {
-		do_swap(base, n, size, swap_func, 0, i);
-
-		for (r = 0; r * 2 + 1 < i; r = c) {
-			c = r * 2 + 1;
-
-			if (c + 1 < i &&
-			    do_cmp(base, n, size, cmp_func, c, c + 1) < 0)
-				c++;
-
-			if (do_cmp(base, n, size, cmp_func, r, c) >= 0)
-				break;
-
-			do_swap(base, n, size, swap_func, r, c);
-		}
-	}
-}
-
-void sort_cmp_size(void *base, size_t num, size_t size,
-	  int (*cmp_func)(const void *, const void *, size_t),
-	  void (*swap_func)(void *, void *, size_t size))
-{
-	/* pre-scale counters for performance */
-	int i = (num/2 - 1) * size, n = num * size, c, r;
-
-	if (!swap_func) {
-		if (size == 4 && alignment_ok(base, 4))
-			swap_func = u32_swap;
-		else if (size == 8 && alignment_ok(base, 8))
-			swap_func = u64_swap;
-		else
-			swap_func = generic_swap;
-	}
-
-	/* heapify */
-	for ( ; i >= 0; i -= size) {
-		for (r = i; r * 2 + size < n; r  = c) {
-			c = r * 2 + size;
-			if (c < n - size &&
-			    cmp_func(base + c, base + c + size, size) < 0)
-				c += size;
-			if (cmp_func(base + r, base + c, size) >= 0)
-				break;
-			swap_func(base + r, base + c, size);
-		}
-	}
-
-	/* sort */
-	for (i = n - size; i > 0; i -= size) {
-		swap_func(base, base + i, size);
-		for (r = 0; r * 2 + size < i; r = c) {
-			c = r * 2 + size;
-			if (c < i - size &&
-			    cmp_func(base + c, base + c + size, size) < 0)
-				c += size;
-			if (cmp_func(base + r, base + c, size) >= 0)
-				break;
-			swap_func(base + r, base + c, size);
-		}
-	}
-}
-
-static void mempool_free_vp(void *element, void *pool_data)
-{
-	size_t size = (size_t) pool_data;
-
-	vpfree(element, size);
-}
-
-static void *mempool_alloc_vp(gfp_t gfp_mask, void *pool_data)
-{
-	size_t size = (size_t) pool_data;
-
-	return vpmalloc(size, gfp_mask);
-}
-
-int mempool_init_kvpmalloc_pool(mempool_t *pool, int min_nr, size_t size)
-{
-	return size < PAGE_SIZE
-		? mempool_init_kmalloc_pool(pool, min_nr, size)
-		: mempool_init(pool, min_nr, mempool_alloc_vp,
-			       mempool_free_vp, (void *) size);
-}
-
 #if 0
 void eytzinger1_test(void)
 {
@@ -908,6 +842,7 @@ u64 *bch2_acc_percpu_u64s(u64 __percpu *p, unsigned nr)
 	u64 *ret;
 	int cpu;
 
+	/* access to pcpu vars has to be blocked by other locking */
 	preempt_disable();
 	ret = this_cpu_ptr(p);
 	preempt_enable();
@@ -923,3 +858,39 @@ u64 *bch2_acc_percpu_u64s(u64 __percpu *p, unsigned nr)
 
 	return ret;
 }
+
+void bch2_darray_str_exit(darray_str *d)
+{
+	darray_for_each(*d, i)
+		kfree(*i);
+	darray_exit(d);
+}
+
+int bch2_split_devs(const char *_dev_name, darray_str *ret)
+{
+	darray_init(ret);
+
+	char *dev_name, *s, *orig;
+
+	dev_name = orig = kstrdup(_dev_name, GFP_KERNEL);
+	if (!dev_name)
+		return -ENOMEM;
+
+	while ((s = strsep(&dev_name, ":"))) {
+		char *p = kstrdup(s, GFP_KERNEL);
+		if (!p)
+			goto err;
+
+		if (darray_push(ret, p)) {
+			kfree(p);
+			goto err;
+		}
+	}
+
+	kfree(orig);
+	return 0;
+err:
+	bch2_darray_str_exit(ret);
+	kfree(orig);
+	return -ENOMEM;
+}