From: Zhi Yong Wu <wuzhy@xxxxxxxxxxxxxxxxxx>
Signed-off-by: Zhi Yong Wu <wuzhy@xxxxxxxxxxxxxxxxxx>
---
fs/hot_tracking.c | 164 +++++++++++++++++++++++++++++++++++++++++++++++++++++
fs/hot_tracking.h | 54 +++++++++++++++++
2 files changed, 218 insertions(+), 0 deletions(-)
diff --git a/fs/hot_tracking.c b/fs/hot_tracking.c
index b5568bc..05624ad 100644
--- a/fs/hot_tracking.c
+++ b/fs/hot_tracking.c
@@ -331,6 +331,170 @@ static void hot_freq_data_update(struct hot_freq_data *freq_data, bool write)
}
}
+static u64 hot_raw_shift(u64 counter, u32 bits, bool dir)
+{
+ if (dir)
+ return counter << bits;
+ else
+ return counter >> bits;
+}
+
+/*
+ * hot_temp_calc() is responsible for distilling the six heat
+ * criteria, which are described in detail in hot_tracking.h) down into a single
+ * temperature value for the data, which is an integer between 0
+ * and HEAT_MAX_VALUE.
+ *
+ * To accomplish this, the raw values from the hot_freq_data structure
+ * are shifted various ways in order to make the temperature calculation more
+ * or less sensitive to each value.
+ *
+ * Once this calibration has happened, we do some additional normalization and
+ * make sure that everything fits nicely in a u32. From there, we take a very
+ * rudimentary kind of "average" of each of the values, where the *_COEFF_POWER
+ * values act as weights for the average.
+ *
+ * Finally, we use the HEAT_HASH_BITS value, which determines the size of the
+ * heat list array, to normalize the temperature to the proper granularity.
+ */
+u32 hot_temp_calc(struct hot_freq_data *freq_data)
+{
+ u32 result = 0;
+
+ struct timespec ckt = current_kernel_time();
+ u64 cur_time = timespec_to_ns(&ckt);
+
+ u32 nrr_heat = (u32)hot_raw_shift((u64)freq_data->nr_reads,
+ NRR_MULTIPLIER_POWER, true);
+ u32 nrw_heat = (u32)hot_raw_shift((u64)freq_data->nr_writes,
+ NRW_MULTIPLIER_POWER, true);
+
+ u64 ltr_heat =
+ hot_raw_shift((cur_time - timespec_to_ns(&freq_data->last_read_time)),
+ LTR_DIVIDER_POWER, false);
+ u64 ltw_heat =
+ hot_raw_shift((cur_time - timespec_to_ns(&freq_data->last_write_time)),
+ LTW_DIVIDER_POWER, false);
+
+ u64 avr_heat =
+ hot_raw_shift((((u64) -1) - freq_data->avg_delta_reads),
+ AVR_DIVIDER_POWER, false);
+ u64 avw_heat =
+ hot_raw_shift((((u64) -1) - freq_data->avg_delta_writes),
+ AVW_DIVIDER_POWER, false);
+
+ /* ltr_heat is now guaranteed to be u32 safe */
+ if (ltr_heat >= hot_raw_shift((u64) 1, 32, true))
+ ltr_heat = 0;
+ else
+ ltr_heat = hot_raw_shift((u64) 1, 32, true) - ltr_heat;
+
+ /* ltw_heat is now guaranteed to be u32 safe */
+ if (ltw_heat >= hot_raw_shift((u64) 1, 32, true))
+ ltw_heat = 0;
+ else
+ ltw_heat = hot_raw_shift((u64) 1, 32, true) - ltw_heat;
+
+ /* avr_heat is now guaranteed to be u32 safe */
+ if (avr_heat >= hot_raw_shift((u64) 1, 32, true))
+ avr_heat = (u32) -1;
+
+ /* avw_heat is now guaranteed to be u32 safe */
+ if (avw_heat >= hot_raw_shift((u64) 1, 32, true))
+ avw_heat = (u32) -1;
+
+ nrr_heat = (u32)hot_raw_shift((u64)nrr_heat,
+ (3 - NRR_COEFF_POWER), false);
+ nrw_heat = (u32)hot_raw_shift((u64)nrw_heat,
+ (3 - NRW_COEFF_POWER), false);
+ ltr_heat = hot_raw_shift(ltr_heat, (3 - LTR_COEFF_POWER), false);
+ ltw_heat = hot_raw_shift(ltw_heat, (3 - LTW_COEFF_POWER), false);
+ avr_heat = hot_raw_shift(avr_heat, (3 - AVR_COEFF_POWER), false);
+ avw_heat = hot_raw_shift(avw_heat, (3 - AVW_COEFF_POWER), false);
+
+ result = nrr_heat + nrw_heat + (u32) ltr_heat +
+ (u32) ltw_heat + (u32) avr_heat + (u32) avw_heat;
+
+ return result;
+}
+
+/*
+ * Calculate a new temperature and, if necessary,
+ * move the list_head corresponding to this inode or range
+ * to the proper list with the new temperature
+ */
+static void hot_map_array_update(struct hot_freq_data *freq_data,
+ struct hot_info *root)
+{
+ struct hot_map_head *buckets, *cur_bucket;
+ struct hot_comm_item *comm_item;
+ struct hot_inode_item *he;
+ struct hot_range_item *hr;
+ u8 a_temp, b_temp;
+ u32 temp = 0;
+
+ comm_item = container_of(freq_data,
+ struct hot_comm_item, hot_freq_data);
+
+ if (freq_data->flags & FREQ_DATA_TYPE_INODE) {
+ he = container_of(comm_item,
+ struct hot_inode_item, hot_inode);
+ buckets = root->heat_inode_map;
+
+ spin_lock(&he->hot_inode.lock);
+ temp = hot_temp_calc(freq_data);
+ spin_unlock(&he->hot_inode.lock);
+
+ if (he == NULL)
+ return;
+
+ spin_lock(&he->hot_inode.lock);
+ a_temp = temp >> (32 - HEAT_MAP_BITS);
+ b_temp = freq_data->last_temp >> (32 - HEAT_MAP_BITS);
+ if (list_empty(&he->hot_inode.n_list) || (a_temp != b_temp)) {
+ if (!list_empty(&he->hot_inode.n_list)) {
+ list_del_init(&he->hot_inode.n_list);
+ root->hot_map_nr--;
+ }
+
+ cur_bucket = buckets + a_temp;
+ list_add_tail(&he->hot_inode.n_list,
+ &cur_bucket->node_list);
+ root->hot_map_nr++;
+ freq_data->last_temp = temp;
+ }
+ spin_unlock(&he->hot_inode.lock);
+ } else if (freq_data->flags & FREQ_DATA_TYPE_RANGE) {
+ hr = container_of(comm_item,
+ struct hot_range_item, hot_range);
+ buckets = root->heat_range_map;
+
+ spin_lock(&hr->hot_range.lock);
+ temp = hot_temp_calc(freq_data);
+ spin_unlock(&hr->hot_range.lock);
+
+ if (hr == NULL)
+ return;
+
+ spin_lock(&hr->hot_range.lock);
+ a_temp = temp >> (32 - HEAT_MAP_BITS);
+ b_temp = freq_data->last_temp >> (32 - HEAT_MAP_BITS);
+ if (list_empty(&hr->hot_range.n_list) || (a_temp != b_temp)) {
+ if (!list_empty(&hr->hot_range.n_list)) {
+ list_del_init(&hr->hot_range.n_list);
+ root->hot_map_nr--;
+ }
+
+ cur_bucket = buckets + a_temp;
+ list_add_tail(&hr->hot_range.n_list,
+ &cur_bucket->node_list);
+ root->hot_map_nr++;
+ freq_data->last_temp = temp;
+ }
+ spin_unlock(&hr->hot_range.lock);
+ }
+}
+
/*
* Initialize inode and range map arrays.
*/
diff --git a/fs/hot_tracking.h b/fs/hot_tracking.h
index 3e5f5d0..be2365c 100644
--- a/fs/hot_tracking.h
+++ b/fs/hot_tracking.h
@@ -25,8 +25,62 @@
#define FREQ_POWER 4
+/*
+ * The following comments explain what exactly comprises a unit of heat.
+ *
+ * Each of six values of heat are calculated and combined in order to form an
+ * overall temperature for the data:
+ *
+ * NRR - number of reads since mount
+ * NRW - number of writes since mount
+ * LTR - time elapsed since last read (ns)
+ * LTW - time elapsed since last write (ns)
+ * AVR - average delta between recent reads (ns)
+ * AVW - average delta between recent writes (ns)
+ *
+ * These values are divided (right-shifted) according to the *_DIVIDER_POWER
+ * values defined below to bring the numbers into a reasonable range. You can
+ * modify these values to fit your needs. However, each heat unit is a u32 and
+ * thus maxes out at 2^32 - 1. Therefore, you must choose your dividers quite
+ * carefully or else they could max out or be stuck at zero quite easily.
+ *
+ * (E.g., if you chose AVR_DIVIDER_POWER = 0, nothing less than 4s of atime
+ * delta would bring the temperature above zero, ever.)
+ *
+ * Finally, each value is added to the overall temperature between 0 and 8
+ * times, depending on its *_COEFF_POWER value. Note that the coefficients are
+ * also actually implemented with shifts, so take care to treat these values
+ * as powers of 2. (I.e., 0 means we'll add it to the temp once; 1 = 2x, etc.)
+ */
+
+/* NRR/NRW heat unit = 2^X accesses */
+#define NRR_MULTIPLIER_POWER 20
+#define NRR_COEFF_POWER 0
+#define NRW_MULTIPLIER_POWER 20
+#define NRW_COEFF_POWER 0
+
+/* LTR/LTW heat unit = 2^X ns of age */
+#define LTR_DIVIDER_POWER 30
+#define LTR_COEFF_POWER 1
+#define LTW_DIVIDER_POWER 30
+#define LTW_COEFF_POWER 1
+
+/*
+ * AVR/AVW cold unit = 2^X ns of average delta
+ * AVR/AVW heat unit = HEAT_MAX_VALUE - cold unit
+ *
+ * E.g., data with an average delta between 0 and 2^X ns
+ * will have a cold value of 0, which means a heat value
+ * equal to HEAT_MAX_VALUE.
+ */
+#define AVR_DIVIDER_POWER 40
+#define AVR_COEFF_POWER 0
+#define AVW_DIVIDER_POWER 40
+#define AVW_COEFF_POWER 0
+
struct hot_inode_item
*hot_inode_item_find(struct hot_info *root, u64 ino);
void hot_inode_item_put(struct hot_inode_item *he);
+u32 hot_temp_calc(struct hot_freq_data *freq_data);
#endif /* __HOT_TRACKING__ */
--
1.7.6.5
--
To unsubscribe from this list: send the line "unsubscribe linux-ext4" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
