Add support to bcache hinting functions and sysfs to hint by the ioprio of
'current' which can be configured with `ionice`.
Cache hinting is configurable by writing 'class,level' pairs to sysfs.
These are the defaults:
echo 2,7 > /sys/block/bcache0/bcache/ioprio_bypass
echo 2,0 > /sys/block/bcache0/bcache/ioprio_writeback
(-p) IO Class (-n) Class level Action
-----------------------------------------------------
(1) Realtime 0-7 Writeback
(2) Best-effort 0 Writeback
(2) Best-effort 1-6 Original bcache logic
(2) Best-effort 7 Bypass cache
(3) Idle n/a Bypass cache
See `man ionice` for more ioprio detail.
Signed-off-by: Eric Wheeler <bcache@xxxxxxxxxxxxxxxxxx>
---
drivers/md/bcache/bcache.h | 3 +++
drivers/md/bcache/request.c | 13 +++++++++++++
drivers/md/bcache/sysfs.c | 45 +++++++++++++++++++++++++++++++++++++++++++
drivers/md/bcache/writeback.c | 3 +++
drivers/md/bcache/writeback.h | 13 +++++++++++++
5 files changed, 77 insertions(+)
diff --git a/drivers/md/bcache/bcache.h b/drivers/md/bcache/bcache.h
index 6b420a5..bf945fa 100644
--- a/drivers/md/bcache/bcache.h
+++ b/drivers/md/bcache/bcache.h
@@ -367,6 +367,9 @@ struct cached_dev {
unsigned writeback_rate_update_seconds;
unsigned writeback_rate_d_term;
unsigned writeback_rate_p_term_inverse;
+
+ unsigned short ioprio_writeback;
+ unsigned short ioprio_bypass;
};
enum alloc_reserve {
diff --git a/drivers/md/bcache/request.c b/drivers/md/bcache/request.c
index 4b177fe..0c0a2f6 100644
--- a/drivers/md/bcache/request.c
+++ b/drivers/md/bcache/request.c
@@ -375,6 +375,7 @@ static bool check_should_bypass(struct cached_dev *dc, struct bio *bio)
unsigned sectors, congested = bch_get_congested(c);
struct task_struct *task = current;
struct io *i;
+ struct io_context *ioc;
if (test_bit(BCACHE_DEV_DETACHING, &dc->disk.flags) ||
c->gc_stats.in_use > CUTOFF_CACHE_ADD ||
@@ -386,6 +387,18 @@ static bool check_should_bypass(struct cached_dev *dc, struct bio *bio)
op_is_write(bio_op(bio))))
goto skip;
+ /* If process ioprio is lower-or-equal to dc->ioprio_bypass, then
+ * hint for bypass. Note that a lower-priority IO class+value
+ * has a greater numeric value. */
+ ioc = get_task_io_context(current, GFP_ATOMIC, NUMA_NO_NODE);
+ if (ioc
+ && ioprio_valid(ioc->ioprio)
+ && ioprio_valid(dc->ioprio_writeback)
+ && ioc->ioprio >= dc->ioprio_bypass) {
+ put_io_context(ioc);
+ return true;
+ }
+
if (bio->bi_iter.bi_sector & (c->sb.block_size - 1) ||
bio_sectors(bio) & (c->sb.block_size - 1)) {
pr_debug("skipping unaligned io");
diff --git a/drivers/md/bcache/sysfs.c b/drivers/md/bcache/sysfs.c
index b3ff57d..83ac010 100644
--- a/drivers/md/bcache/sysfs.c
+++ b/drivers/md/bcache/sysfs.c
@@ -106,6 +106,9 @@ rw_attribute(btree_shrinker_disabled);
rw_attribute(copy_gc_enabled);
rw_attribute(size);
+rw_attribute(ioprio_writeback);
+rw_attribute(ioprio_bypass);
+
SHOW(__bch_cached_dev)
{
struct cached_dev *dc = container_of(kobj, struct cached_dev,
@@ -182,6 +185,17 @@ SHOW(__bch_cached_dev)
return strlen(buf);
}
+ if (attr == &sysfs_ioprio_bypass)
+ return snprintf(buf, PAGE_SIZE-1, "%d,%ld\n",
+ IOPRIO_PRIO_CLASS(dc->ioprio_bypass),
+ IOPRIO_PRIO_DATA(dc->ioprio_bypass));
+
+ if (attr == &sysfs_ioprio_writeback)
+ return snprintf(buf, PAGE_SIZE-1, "%d,%ld\n",
+ IOPRIO_PRIO_CLASS(dc->ioprio_writeback),
+ IOPRIO_PRIO_DATA(dc->ioprio_writeback));
+
+
#undef var
return 0;
}
@@ -194,6 +208,8 @@ STORE(__cached_dev)
unsigned v = size;
struct cache_set *c;
struct kobj_uevent_env *env;
+ unsigned ioprio_class = 0; /* invalid initial ioprio values */
+ unsigned ioprio_value = IOPRIO_BE_NR+1;
#define d_strtoul(var) sysfs_strtoul(var, dc->var)
#define d_strtoul_nonzero(var) sysfs_strtoul_clamp(var, dc->var, 1, INT_MAX)
@@ -282,6 +298,33 @@ STORE(__cached_dev)
if (attr == &sysfs_stop)
bcache_device_stop(&dc->disk);
+ if (attr == &sysfs_ioprio_writeback) {
+ if (sscanf(buf, "%u,%u", &ioprio_class, &ioprio_value) != 2
+ || ioprio_class > IOPRIO_CLASS_IDLE
+ || ioprio_class < IOPRIO_CLASS_RT
+ || ioprio_value >= IOPRIO_BE_NR)
+ pr_err("bcache: ioprio_writeback invalid ioprio class: %u data value: %u\n",
+ ioprio_class, ioprio_value);
+ else if (ioprio_class == IOPRIO_CLASS_IDLE)
+ dc->ioprio_writeback = IOPRIO_PRIO_VALUE(ioprio_class, 0);
+ else
+ dc->ioprio_writeback = IOPRIO_PRIO_VALUE(ioprio_class, ioprio_value);
+ }
+
+ if (attr == &sysfs_ioprio_bypass)
+ {
+ if (sscanf(buf, "%u,%u", &ioprio_class, &ioprio_value) != 2
+ || ioprio_class > IOPRIO_CLASS_IDLE
+ || ioprio_class < IOPRIO_CLASS_RT
+ || ioprio_value >= IOPRIO_BE_NR)
+ pr_err("bcache: ioprio_bypass invalid ioprio class: %u data value: %u\n",
+ ioprio_class, ioprio_value);
+ else if (ioprio_class == IOPRIO_CLASS_IDLE)
+ dc->ioprio_writeback = IOPRIO_PRIO_VALUE(ioprio_class, 0);
+ else
+ dc->ioprio_writeback = IOPRIO_PRIO_VALUE(ioprio_class, ioprio_value);
+ }
+
return size;
}
@@ -334,6 +377,8 @@ static struct attribute *bch_cached_dev_files[] = {
&sysfs_verify,
&sysfs_bypass_torture_test,
#endif
+ &sysfs_ioprio_bypass,
+ &sysfs_ioprio_writeback,
NULL
};
KTYPE(bch_cached_dev);
diff --git a/drivers/md/bcache/writeback.c b/drivers/md/bcache/writeback.c
index d9fd2a6..9772988 100644
--- a/drivers/md/bcache/writeback.c
+++ b/drivers/md/bcache/writeback.c
@@ -514,6 +514,9 @@ void bch_cached_dev_writeback_init(struct cached_dev *dc)
dc->writeback_rate_d_term = 30;
dc->writeback_rate_p_term_inverse = 6000;
+ dc->ioprio_writeback = IOPRIO_PRIO_VALUE(IOPRIO_CLASS_BE, 0);
+ dc->ioprio_bypass = IOPRIO_PRIO_VALUE(IOPRIO_CLASS_BE, IOPRIO_BE_NR-1);
+
INIT_DELAYED_WORK(&dc->writeback_rate_update, update_writeback_rate);
}
diff --git a/drivers/md/bcache/writeback.h b/drivers/md/bcache/writeback.h
index 301eaf5..a21fd87 100644
--- a/drivers/md/bcache/writeback.h
+++ b/drivers/md/bcache/writeback.h
@@ -43,6 +43,7 @@ static inline bool should_writeback(struct cached_dev *dc, struct bio *bio,
unsigned cache_mode, bool would_skip)
{
unsigned in_use = dc->disk.c->gc_stats.in_use;
+ struct io_context *ioc;
if (cache_mode != CACHE_MODE_WRITEBACK ||
test_bit(BCACHE_DEV_DETACHING, &dc->disk.flags) ||
@@ -57,6 +58,18 @@ static inline bool should_writeback(struct cached_dev *dc, struct bio *bio,
if (would_skip)
return false;
+ /* If process ioprio is higher-or-equal to dc->ioprio_writeback, then
+ * hint for writeback. Note that a higher-priority IO class+value
+ * has a lesser numeric value. */
+ ioc = get_task_io_context(current, GFP_ATOMIC, NUMA_NO_NODE);
+ if (ioc
+ && ioprio_valid(ioc->ioprio)
+ && ioprio_valid(dc->ioprio_writeback)
+ && ioc->ioprio <= dc->ioprio_writeback) {
+ put_io_context(ioc);
+ return true;
+ }
+
return bio->bi_opf & REQ_SYNC ||
in_use <= CUTOFF_WRITEBACK;
}
--
1.8.3.1
--
To unsubscribe from this list: send the line "unsubscribe linux-bcache" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
