Re: [RFC v4+ hot_track 03/19] vfs: add I/O frequency update function
On Thu, Nov 8, 2012 at 2:49 AM, Darrick J. Wong wrote:
> On Wed, Nov 07, 2012 at 04:27:05PM +0800, Zhi Yong Wu wrote:
>> On Wed, Nov 7, 2012 at 6:45 AM, Darrick J. Wong
>> wrote:
>> > On Mon, Oct 29, 2012 at 12:30:45PM +0800, zwu.ker...@gmail.com wrote:
>> >> From: Zhi Yong Wu
>> >>
>> >> Add some util helpers to update access frequencies
>> >> for one file or its range.
>> >>
>> >> Signed-off-by: Zhi Yong Wu
>> >> ---
>> >> fs/hot_tracking.c| 179
>> >> ++
>> >> fs/hot_tracking.h|7 ++
>> >> include/linux/hot_tracking.h |2 +
>> >> 3 files changed, 188 insertions(+), 0 deletions(-)
>> >>
>> >> diff --git a/fs/hot_tracking.c b/fs/hot_tracking.c
>> >> index 68591f0..0a7d9a3 100644
>> >> --- a/fs/hot_tracking.c
>> >> +++ b/fs/hot_tracking.c
>> >> @@ -172,6 +172,137 @@ static void hot_inode_tree_exit(struct hot_info
>> >> *root)
>> >> }
>> >> }
>> >>
>> >> +struct hot_inode_item
>> >> +*hot_inode_item_find(struct hot_info *root, u64 ino)
>> >> +{
>> >> + struct hot_inode_item *he;
>> >> + int ret;
>> >> +
>> >> +again:
>> >> + spin_lock(>lock);
>> >> + he = radix_tree_lookup(>hot_inode_tree, ino);
>> >> + if (he) {
>> >> + kref_get(>hot_inode.refs);
>> >> + spin_unlock(>lock);
>> >> + return he;
>> >> + }
>> >> + spin_unlock(>lock);
>> >> +
>> >> + he = kmem_cache_zalloc(hot_inode_item_cachep,
>> >> + GFP_KERNEL | GFP_NOFS);
>> >> + if (!he)
>> >> + return ERR_PTR(-ENOMEM);
>> >> +
>> >> + hot_inode_item_init(he, ino, >hot_inode_tree);
>> >> +
>> >> + ret = radix_tree_preload(GFP_NOFS & ~__GFP_HIGHMEM);
>> >> + if (ret) {
>> >> + kmem_cache_free(hot_inode_item_cachep, he);
>> >> + return ERR_PTR(ret);
>> >> + }
>> >> +
>> >> + spin_lock(>lock);
>> >> + ret = radix_tree_insert(>hot_inode_tree, ino, he);
>> >> + if (ret == -EEXIST) {
>> >> + kmem_cache_free(hot_inode_item_cachep, he);
>> >> + spin_unlock(>lock);
>> >> + radix_tree_preload_end();
>> >> + goto again;
>> >> + }
>> >> + spin_unlock(>lock);
>> >> + radix_tree_preload_end();
>> >> +
>> >> + kref_get(>hot_inode.refs);
>> >> + return he;
>> >> +}
>> >> +EXPORT_SYMBOL_GPL(hot_inode_item_find);
>> >> +
>> >> +static struct hot_range_item
>> >> +*hot_range_item_find(struct hot_inode_item *he,
>> >> + u32 start)
>> >> +{
>> >> + struct hot_range_item *hr;
>> >> + int ret;
>> >> +
>> >> +again:
>> >> + spin_lock(>lock);
>> >> + hr = radix_tree_lookup(>hot_range_tree, start);
>> >> + if (hr) {
>> >> + kref_get(>hot_range.refs);
>> >> + spin_unlock(>lock);
>> >> + return hr;
>> >> + }
>> >> + spin_unlock(>lock);
>> >> +
>> >> + hr = kmem_cache_zalloc(hot_range_item_cachep,
>> >> + GFP_KERNEL | GFP_NOFS);
>> >> + if (!hr)
>> >> + return ERR_PTR(-ENOMEM);
>> >> +
>> >> + hot_range_item_init(hr, start, he);
>> >> +
>> >> + ret = radix_tree_preload(GFP_NOFS & ~__GFP_HIGHMEM);
>> >> + if (ret) {
>> >> + kmem_cache_free(hot_range_item_cachep, hr);
>> >> + return ERR_PTR(ret);
>> >> + }
>> >> +
>> >> + spin_lock(>lock);
>> >> + ret = radix_tree_insert(>hot_range_tree, start, hr);
>> >> + if (ret == -EEXIST) {
>> >> + kmem_cache_free(hot_range_item_cachep, hr);
>> >> + spin_unlock(>lock);
>> >> + radix_tree_preload_end();
>> >> + goto again;
>> >> + }
>> >> + spin_unlock(>lock);
>> >> + radix_tree_preload_end();
>> >> +
>> >> + kref_get(>hot_range.refs);
>> >> + return hr;
>> >> +}
>> >> +
>> >> +/*
>> >> + * This function does the actual work of updating
>> >> + * the frequency numbers, whatever they turn out to be.
>> >> + */
>> >> +static u64 hot_average_update(struct timespec old_atime,
>> >> + struct timespec cur_time, u64 old_avg)
>> >> +{
>> >> + struct timespec delta_ts;
>> >> + u64 new_avg;
>> >> + u64 new_delta;
>> >> +
>> >> + delta_ts = timespec_sub(cur_time, old_atime);
>> >> + new_delta = timespec_to_ns(_ts) >> FREQ_POWER;
>> >> +
>> >> + new_avg = (old_avg << FREQ_POWER) - old_avg + new_delta;
>> >> + new_avg = new_avg >> FREQ_POWER;
>> >> +
>> >> + return new_avg;
>> >> +}
>> >> +
>> >> +static void hot_freq_data_update(struct hot_freq_data *freq_data, bool
>> >> write)
>> >> +{
>> >> + struct timespec cur_time = current_kernel_time();
>> >> +
>> >> + if (write) {
>> >> + freq_data->nr_writes += 1;
>> >> + freq_data->avg_delta_writes = hot_average_update(
>> >> + freq_data->last_write_time,
>> >> + cur_time,
>> >> +
Re: [RFC v4+ hot_track 03/19] vfs: add I/O frequency update function
On Wed, Nov 07, 2012 at 04:27:05PM +0800, Zhi Yong Wu wrote:
> On Wed, Nov 7, 2012 at 6:45 AM, Darrick J. Wong
> wrote:
> > On Mon, Oct 29, 2012 at 12:30:45PM +0800, zwu.ker...@gmail.com wrote:
> >> From: Zhi Yong Wu
> >>
> >> Add some util helpers to update access frequencies
> >> for one file or its range.
> >>
> >> Signed-off-by: Zhi Yong Wu
> >> ---
> >> fs/hot_tracking.c| 179
> >> ++
> >> fs/hot_tracking.h|7 ++
> >> include/linux/hot_tracking.h |2 +
> >> 3 files changed, 188 insertions(+), 0 deletions(-)
> >>
> >> diff --git a/fs/hot_tracking.c b/fs/hot_tracking.c
> >> index 68591f0..0a7d9a3 100644
> >> --- a/fs/hot_tracking.c
> >> +++ b/fs/hot_tracking.c
> >> @@ -172,6 +172,137 @@ static void hot_inode_tree_exit(struct hot_info
> >> *root)
> >> }
> >> }
> >>
> >> +struct hot_inode_item
> >> +*hot_inode_item_find(struct hot_info *root, u64 ino)
> >> +{
> >> + struct hot_inode_item *he;
> >> + int ret;
> >> +
> >> +again:
> >> + spin_lock(>lock);
> >> + he = radix_tree_lookup(>hot_inode_tree, ino);
> >> + if (he) {
> >> + kref_get(>hot_inode.refs);
> >> + spin_unlock(>lock);
> >> + return he;
> >> + }
> >> + spin_unlock(>lock);
> >> +
> >> + he = kmem_cache_zalloc(hot_inode_item_cachep,
> >> + GFP_KERNEL | GFP_NOFS);
> >> + if (!he)
> >> + return ERR_PTR(-ENOMEM);
> >> +
> >> + hot_inode_item_init(he, ino, >hot_inode_tree);
> >> +
> >> + ret = radix_tree_preload(GFP_NOFS & ~__GFP_HIGHMEM);
> >> + if (ret) {
> >> + kmem_cache_free(hot_inode_item_cachep, he);
> >> + return ERR_PTR(ret);
> >> + }
> >> +
> >> + spin_lock(>lock);
> >> + ret = radix_tree_insert(>hot_inode_tree, ino, he);
> >> + if (ret == -EEXIST) {
> >> + kmem_cache_free(hot_inode_item_cachep, he);
> >> + spin_unlock(>lock);
> >> + radix_tree_preload_end();
> >> + goto again;
> >> + }
> >> + spin_unlock(>lock);
> >> + radix_tree_preload_end();
> >> +
> >> + kref_get(>hot_inode.refs);
> >> + return he;
> >> +}
> >> +EXPORT_SYMBOL_GPL(hot_inode_item_find);
> >> +
> >> +static struct hot_range_item
> >> +*hot_range_item_find(struct hot_inode_item *he,
> >> + u32 start)
> >> +{
> >> + struct hot_range_item *hr;
> >> + int ret;
> >> +
> >> +again:
> >> + spin_lock(>lock);
> >> + hr = radix_tree_lookup(>hot_range_tree, start);
> >> + if (hr) {
> >> + kref_get(>hot_range.refs);
> >> + spin_unlock(>lock);
> >> + return hr;
> >> + }
> >> + spin_unlock(>lock);
> >> +
> >> + hr = kmem_cache_zalloc(hot_range_item_cachep,
> >> + GFP_KERNEL | GFP_NOFS);
> >> + if (!hr)
> >> + return ERR_PTR(-ENOMEM);
> >> +
> >> + hot_range_item_init(hr, start, he);
> >> +
> >> + ret = radix_tree_preload(GFP_NOFS & ~__GFP_HIGHMEM);
> >> + if (ret) {
> >> + kmem_cache_free(hot_range_item_cachep, hr);
> >> + return ERR_PTR(ret);
> >> + }
> >> +
> >> + spin_lock(>lock);
> >> + ret = radix_tree_insert(>hot_range_tree, start, hr);
> >> + if (ret == -EEXIST) {
> >> + kmem_cache_free(hot_range_item_cachep, hr);
> >> + spin_unlock(>lock);
> >> + radix_tree_preload_end();
> >> + goto again;
> >> + }
> >> + spin_unlock(>lock);
> >> + radix_tree_preload_end();
> >> +
> >> + kref_get(>hot_range.refs);
> >> + return hr;
> >> +}
> >> +
> >> +/*
> >> + * This function does the actual work of updating
> >> + * the frequency numbers, whatever they turn out to be.
> >> + */
> >> +static u64 hot_average_update(struct timespec old_atime,
> >> + struct timespec cur_time, u64 old_avg)
> >> +{
> >> + struct timespec delta_ts;
> >> + u64 new_avg;
> >> + u64 new_delta;
> >> +
> >> + delta_ts = timespec_sub(cur_time, old_atime);
> >> + new_delta = timespec_to_ns(_ts) >> FREQ_POWER;
> >> +
> >> + new_avg = (old_avg << FREQ_POWER) - old_avg + new_delta;
> >> + new_avg = new_avg >> FREQ_POWER;
> >> +
> >> + return new_avg;
> >> +}
> >> +
> >> +static void hot_freq_data_update(struct hot_freq_data *freq_data, bool
> >> write)
> >> +{
> >> + struct timespec cur_time = current_kernel_time();
> >> +
> >> + if (write) {
> >> + freq_data->nr_writes += 1;
> >> + freq_data->avg_delta_writes = hot_average_update(
> >> + freq_data->last_write_time,
> >> + cur_time,
> >> + freq_data->avg_delta_writes);
> >> + freq_data->last_write_time = cur_time;
> >> + } else {
> >> + freq_data->nr_reads += 1;
> >> + freq_data->avg_delta_reads =
Re: [RFC v4+ hot_track 03/19] vfs: add I/O frequency update function
On Wed, Nov 7, 2012 at 6:45 AM, Darrick J. Wong wrote:
> On Mon, Oct 29, 2012 at 12:30:45PM +0800, zwu.ker...@gmail.com wrote:
>> From: Zhi Yong Wu
>>
>> Add some util helpers to update access frequencies
>> for one file or its range.
>>
>> Signed-off-by: Zhi Yong Wu
>> ---
>> fs/hot_tracking.c| 179
>> ++
>> fs/hot_tracking.h|7 ++
>> include/linux/hot_tracking.h |2 +
>> 3 files changed, 188 insertions(+), 0 deletions(-)
>>
>> diff --git a/fs/hot_tracking.c b/fs/hot_tracking.c
>> index 68591f0..0a7d9a3 100644
>> --- a/fs/hot_tracking.c
>> +++ b/fs/hot_tracking.c
>> @@ -172,6 +172,137 @@ static void hot_inode_tree_exit(struct hot_info *root)
>> }
>> }
>>
>> +struct hot_inode_item
>> +*hot_inode_item_find(struct hot_info *root, u64 ino)
>> +{
>> + struct hot_inode_item *he;
>> + int ret;
>> +
>> +again:
>> + spin_lock(>lock);
>> + he = radix_tree_lookup(>hot_inode_tree, ino);
>> + if (he) {
>> + kref_get(>hot_inode.refs);
>> + spin_unlock(>lock);
>> + return he;
>> + }
>> + spin_unlock(>lock);
>> +
>> + he = kmem_cache_zalloc(hot_inode_item_cachep,
>> + GFP_KERNEL | GFP_NOFS);
>> + if (!he)
>> + return ERR_PTR(-ENOMEM);
>> +
>> + hot_inode_item_init(he, ino, >hot_inode_tree);
>> +
>> + ret = radix_tree_preload(GFP_NOFS & ~__GFP_HIGHMEM);
>> + if (ret) {
>> + kmem_cache_free(hot_inode_item_cachep, he);
>> + return ERR_PTR(ret);
>> + }
>> +
>> + spin_lock(>lock);
>> + ret = radix_tree_insert(>hot_inode_tree, ino, he);
>> + if (ret == -EEXIST) {
>> + kmem_cache_free(hot_inode_item_cachep, he);
>> + spin_unlock(>lock);
>> + radix_tree_preload_end();
>> + goto again;
>> + }
>> + spin_unlock(>lock);
>> + radix_tree_preload_end();
>> +
>> + kref_get(>hot_inode.refs);
>> + return he;
>> +}
>> +EXPORT_SYMBOL_GPL(hot_inode_item_find);
>> +
>> +static struct hot_range_item
>> +*hot_range_item_find(struct hot_inode_item *he,
>> + u32 start)
>> +{
>> + struct hot_range_item *hr;
>> + int ret;
>> +
>> +again:
>> + spin_lock(>lock);
>> + hr = radix_tree_lookup(>hot_range_tree, start);
>> + if (hr) {
>> + kref_get(>hot_range.refs);
>> + spin_unlock(>lock);
>> + return hr;
>> + }
>> + spin_unlock(>lock);
>> +
>> + hr = kmem_cache_zalloc(hot_range_item_cachep,
>> + GFP_KERNEL | GFP_NOFS);
>> + if (!hr)
>> + return ERR_PTR(-ENOMEM);
>> +
>> + hot_range_item_init(hr, start, he);
>> +
>> + ret = radix_tree_preload(GFP_NOFS & ~__GFP_HIGHMEM);
>> + if (ret) {
>> + kmem_cache_free(hot_range_item_cachep, hr);
>> + return ERR_PTR(ret);
>> + }
>> +
>> + spin_lock(>lock);
>> + ret = radix_tree_insert(>hot_range_tree, start, hr);
>> + if (ret == -EEXIST) {
>> + kmem_cache_free(hot_range_item_cachep, hr);
>> + spin_unlock(>lock);
>> + radix_tree_preload_end();
>> + goto again;
>> + }
>> + spin_unlock(>lock);
>> + radix_tree_preload_end();
>> +
>> + kref_get(>hot_range.refs);
>> + return hr;
>> +}
>> +
>> +/*
>> + * This function does the actual work of updating
>> + * the frequency numbers, whatever they turn out to be.
>> + */
>> +static u64 hot_average_update(struct timespec old_atime,
>> + struct timespec cur_time, u64 old_avg)
>> +{
>> + struct timespec delta_ts;
>> + u64 new_avg;
>> + u64 new_delta;
>> +
>> + delta_ts = timespec_sub(cur_time, old_atime);
>> + new_delta = timespec_to_ns(_ts) >> FREQ_POWER;
>> +
>> + new_avg = (old_avg << FREQ_POWER) - old_avg + new_delta;
>> + new_avg = new_avg >> FREQ_POWER;
>> +
>> + return new_avg;
>> +}
>> +
>> +static void hot_freq_data_update(struct hot_freq_data *freq_data, bool
>> write)
>> +{
>> + struct timespec cur_time = current_kernel_time();
>> +
>> + if (write) {
>> + freq_data->nr_writes += 1;
>> + freq_data->avg_delta_writes = hot_average_update(
>> + freq_data->last_write_time,
>> + cur_time,
>> + freq_data->avg_delta_writes);
>> + freq_data->last_write_time = cur_time;
>> + } else {
>> + freq_data->nr_reads += 1;
>> + freq_data->avg_delta_reads = hot_average_update(
>> + freq_data->last_read_time,
>> + cur_time,
>> + freq_data->avg_delta_reads);
>
> I think you could just pass in a pointer to
> freq_data->avg_delta_{writes,reads} here...
why?
>
>> + freq_data->last_read_time = cur_time;
>> + }
>> +}
>> +
>> /*
>> *
Re: [RFC v4+ hot_track 03/19] vfs: add I/O frequency update function
On Wed, Nov 7, 2012 at 6:45 AM, Darrick J. Wong darrick.w...@oracle.com wrote:
On Mon, Oct 29, 2012 at 12:30:45PM +0800, zwu.ker...@gmail.com wrote:
From: Zhi Yong Wu wu...@linux.vnet.ibm.com
Add some util helpers to update access frequencies
for one file or its range.
Signed-off-by: Zhi Yong Wu wu...@linux.vnet.ibm.com
---
fs/hot_tracking.c| 179
++
fs/hot_tracking.h|7 ++
include/linux/hot_tracking.h |2 +
3 files changed, 188 insertions(+), 0 deletions(-)
diff --git a/fs/hot_tracking.c b/fs/hot_tracking.c
index 68591f0..0a7d9a3 100644
--- a/fs/hot_tracking.c
+++ b/fs/hot_tracking.c
@@ -172,6 +172,137 @@ static void hot_inode_tree_exit(struct hot_info *root)
}
}
+struct hot_inode_item
+*hot_inode_item_find(struct hot_info *root, u64 ino)
+{
+ struct hot_inode_item *he;
+ int ret;
+
+again:
+ spin_lock(root-lock);
+ he = radix_tree_lookup(root-hot_inode_tree, ino);
+ if (he) {
+ kref_get(he-hot_inode.refs);
+ spin_unlock(root-lock);
+ return he;
+ }
+ spin_unlock(root-lock);
+
+ he = kmem_cache_zalloc(hot_inode_item_cachep,
+ GFP_KERNEL | GFP_NOFS);
+ if (!he)
+ return ERR_PTR(-ENOMEM);
+
+ hot_inode_item_init(he, ino, root-hot_inode_tree);
+
+ ret = radix_tree_preload(GFP_NOFS ~__GFP_HIGHMEM);
+ if (ret) {
+ kmem_cache_free(hot_inode_item_cachep, he);
+ return ERR_PTR(ret);
+ }
+
+ spin_lock(root-lock);
+ ret = radix_tree_insert(root-hot_inode_tree, ino, he);
+ if (ret == -EEXIST) {
+ kmem_cache_free(hot_inode_item_cachep, he);
+ spin_unlock(root-lock);
+ radix_tree_preload_end();
+ goto again;
+ }
+ spin_unlock(root-lock);
+ radix_tree_preload_end();
+
+ kref_get(he-hot_inode.refs);
+ return he;
+}
+EXPORT_SYMBOL_GPL(hot_inode_item_find);
+
+static struct hot_range_item
+*hot_range_item_find(struct hot_inode_item *he,
+ u32 start)
+{
+ struct hot_range_item *hr;
+ int ret;
+
+again:
+ spin_lock(he-lock);
+ hr = radix_tree_lookup(he-hot_range_tree, start);
+ if (hr) {
+ kref_get(hr-hot_range.refs);
+ spin_unlock(he-lock);
+ return hr;
+ }
+ spin_unlock(he-lock);
+
+ hr = kmem_cache_zalloc(hot_range_item_cachep,
+ GFP_KERNEL | GFP_NOFS);
+ if (!hr)
+ return ERR_PTR(-ENOMEM);
+
+ hot_range_item_init(hr, start, he);
+
+ ret = radix_tree_preload(GFP_NOFS ~__GFP_HIGHMEM);
+ if (ret) {
+ kmem_cache_free(hot_range_item_cachep, hr);
+ return ERR_PTR(ret);
+ }
+
+ spin_lock(he-lock);
+ ret = radix_tree_insert(he-hot_range_tree, start, hr);
+ if (ret == -EEXIST) {
+ kmem_cache_free(hot_range_item_cachep, hr);
+ spin_unlock(he-lock);
+ radix_tree_preload_end();
+ goto again;
+ }
+ spin_unlock(he-lock);
+ radix_tree_preload_end();
+
+ kref_get(hr-hot_range.refs);
+ return hr;
+}
+
+/*
+ * This function does the actual work of updating
+ * the frequency numbers, whatever they turn out to be.
+ */
+static u64 hot_average_update(struct timespec old_atime,
+ struct timespec cur_time, u64 old_avg)
+{
+ struct timespec delta_ts;
+ u64 new_avg;
+ u64 new_delta;
+
+ delta_ts = timespec_sub(cur_time, old_atime);
+ new_delta = timespec_to_ns(delta_ts) FREQ_POWER;
+
+ new_avg = (old_avg FREQ_POWER) - old_avg + new_delta;
+ new_avg = new_avg FREQ_POWER;
+
+ return new_avg;
+}
+
+static void hot_freq_data_update(struct hot_freq_data *freq_data, bool
write)
+{
+ struct timespec cur_time = current_kernel_time();
+
+ if (write) {
+ freq_data-nr_writes += 1;
+ freq_data-avg_delta_writes = hot_average_update(
+ freq_data-last_write_time,
+ cur_time,
+ freq_data-avg_delta_writes);
+ freq_data-last_write_time = cur_time;
+ } else {
+ freq_data-nr_reads += 1;
+ freq_data-avg_delta_reads = hot_average_update(
+ freq_data-last_read_time,
+ cur_time,
+ freq_data-avg_delta_reads);
I think you could just pass in a pointer to
freq_data-avg_delta_{writes,reads} here...
why?
+ freq_data-last_read_time = cur_time;
+ }
+}
+
/*
* Initialize kmem cache for hot_inode_item and hot_range_item.
*/
@@ -199,6 +330,54 @@ err:
EXPORT_SYMBOL_GPL(hot_cache_init);
/*
+ * Main function to update access frequency from
Re: [RFC v4+ hot_track 03/19] vfs: add I/O frequency update function
On Wed, Nov 07, 2012 at 04:27:05PM +0800, Zhi Yong Wu wrote:
On Wed, Nov 7, 2012 at 6:45 AM, Darrick J. Wong darrick.w...@oracle.com
wrote:
On Mon, Oct 29, 2012 at 12:30:45PM +0800, zwu.ker...@gmail.com wrote:
From: Zhi Yong Wu wu...@linux.vnet.ibm.com
Add some util helpers to update access frequencies
for one file or its range.
Signed-off-by: Zhi Yong Wu wu...@linux.vnet.ibm.com
---
fs/hot_tracking.c| 179
++
fs/hot_tracking.h|7 ++
include/linux/hot_tracking.h |2 +
3 files changed, 188 insertions(+), 0 deletions(-)
diff --git a/fs/hot_tracking.c b/fs/hot_tracking.c
index 68591f0..0a7d9a3 100644
--- a/fs/hot_tracking.c
+++ b/fs/hot_tracking.c
@@ -172,6 +172,137 @@ static void hot_inode_tree_exit(struct hot_info
*root)
}
}
+struct hot_inode_item
+*hot_inode_item_find(struct hot_info *root, u64 ino)
+{
+ struct hot_inode_item *he;
+ int ret;
+
+again:
+ spin_lock(root-lock);
+ he = radix_tree_lookup(root-hot_inode_tree, ino);
+ if (he) {
+ kref_get(he-hot_inode.refs);
+ spin_unlock(root-lock);
+ return he;
+ }
+ spin_unlock(root-lock);
+
+ he = kmem_cache_zalloc(hot_inode_item_cachep,
+ GFP_KERNEL | GFP_NOFS);
+ if (!he)
+ return ERR_PTR(-ENOMEM);
+
+ hot_inode_item_init(he, ino, root-hot_inode_tree);
+
+ ret = radix_tree_preload(GFP_NOFS ~__GFP_HIGHMEM);
+ if (ret) {
+ kmem_cache_free(hot_inode_item_cachep, he);
+ return ERR_PTR(ret);
+ }
+
+ spin_lock(root-lock);
+ ret = radix_tree_insert(root-hot_inode_tree, ino, he);
+ if (ret == -EEXIST) {
+ kmem_cache_free(hot_inode_item_cachep, he);
+ spin_unlock(root-lock);
+ radix_tree_preload_end();
+ goto again;
+ }
+ spin_unlock(root-lock);
+ radix_tree_preload_end();
+
+ kref_get(he-hot_inode.refs);
+ return he;
+}
+EXPORT_SYMBOL_GPL(hot_inode_item_find);
+
+static struct hot_range_item
+*hot_range_item_find(struct hot_inode_item *he,
+ u32 start)
+{
+ struct hot_range_item *hr;
+ int ret;
+
+again:
+ spin_lock(he-lock);
+ hr = radix_tree_lookup(he-hot_range_tree, start);
+ if (hr) {
+ kref_get(hr-hot_range.refs);
+ spin_unlock(he-lock);
+ return hr;
+ }
+ spin_unlock(he-lock);
+
+ hr = kmem_cache_zalloc(hot_range_item_cachep,
+ GFP_KERNEL | GFP_NOFS);
+ if (!hr)
+ return ERR_PTR(-ENOMEM);
+
+ hot_range_item_init(hr, start, he);
+
+ ret = radix_tree_preload(GFP_NOFS ~__GFP_HIGHMEM);
+ if (ret) {
+ kmem_cache_free(hot_range_item_cachep, hr);
+ return ERR_PTR(ret);
+ }
+
+ spin_lock(he-lock);
+ ret = radix_tree_insert(he-hot_range_tree, start, hr);
+ if (ret == -EEXIST) {
+ kmem_cache_free(hot_range_item_cachep, hr);
+ spin_unlock(he-lock);
+ radix_tree_preload_end();
+ goto again;
+ }
+ spin_unlock(he-lock);
+ radix_tree_preload_end();
+
+ kref_get(hr-hot_range.refs);
+ return hr;
+}
+
+/*
+ * This function does the actual work of updating
+ * the frequency numbers, whatever they turn out to be.
+ */
+static u64 hot_average_update(struct timespec old_atime,
+ struct timespec cur_time, u64 old_avg)
+{
+ struct timespec delta_ts;
+ u64 new_avg;
+ u64 new_delta;
+
+ delta_ts = timespec_sub(cur_time, old_atime);
+ new_delta = timespec_to_ns(delta_ts) FREQ_POWER;
+
+ new_avg = (old_avg FREQ_POWER) - old_avg + new_delta;
+ new_avg = new_avg FREQ_POWER;
+
+ return new_avg;
+}
+
+static void hot_freq_data_update(struct hot_freq_data *freq_data, bool
write)
+{
+ struct timespec cur_time = current_kernel_time();
+
+ if (write) {
+ freq_data-nr_writes += 1;
+ freq_data-avg_delta_writes = hot_average_update(
+ freq_data-last_write_time,
+ cur_time,
+ freq_data-avg_delta_writes);
+ freq_data-last_write_time = cur_time;
+ } else {
+ freq_data-nr_reads += 1;
+ freq_data-avg_delta_reads = hot_average_update(
+ freq_data-last_read_time,
+ cur_time,
+ freq_data-avg_delta_reads);
I think you could just pass in a pointer to
freq_data-avg_delta_{writes,reads} here...
why?
freq_data-avg_delta_{reads,writes} seems to be an in/out
Re: [RFC v4+ hot_track 03/19] vfs: add I/O frequency update function
On Thu, Nov 8, 2012 at 2:49 AM, Darrick J. Wong darrick.w...@oracle.com wrote:
On Wed, Nov 07, 2012 at 04:27:05PM +0800, Zhi Yong Wu wrote:
On Wed, Nov 7, 2012 at 6:45 AM, Darrick J. Wong darrick.w...@oracle.com
wrote:
On Mon, Oct 29, 2012 at 12:30:45PM +0800, zwu.ker...@gmail.com wrote:
From: Zhi Yong Wu wu...@linux.vnet.ibm.com
Add some util helpers to update access frequencies
for one file or its range.
Signed-off-by: Zhi Yong Wu wu...@linux.vnet.ibm.com
---
fs/hot_tracking.c| 179
++
fs/hot_tracking.h|7 ++
include/linux/hot_tracking.h |2 +
3 files changed, 188 insertions(+), 0 deletions(-)
diff --git a/fs/hot_tracking.c b/fs/hot_tracking.c
index 68591f0..0a7d9a3 100644
--- a/fs/hot_tracking.c
+++ b/fs/hot_tracking.c
@@ -172,6 +172,137 @@ static void hot_inode_tree_exit(struct hot_info
*root)
}
}
+struct hot_inode_item
+*hot_inode_item_find(struct hot_info *root, u64 ino)
+{
+ struct hot_inode_item *he;
+ int ret;
+
+again:
+ spin_lock(root-lock);
+ he = radix_tree_lookup(root-hot_inode_tree, ino);
+ if (he) {
+ kref_get(he-hot_inode.refs);
+ spin_unlock(root-lock);
+ return he;
+ }
+ spin_unlock(root-lock);
+
+ he = kmem_cache_zalloc(hot_inode_item_cachep,
+ GFP_KERNEL | GFP_NOFS);
+ if (!he)
+ return ERR_PTR(-ENOMEM);
+
+ hot_inode_item_init(he, ino, root-hot_inode_tree);
+
+ ret = radix_tree_preload(GFP_NOFS ~__GFP_HIGHMEM);
+ if (ret) {
+ kmem_cache_free(hot_inode_item_cachep, he);
+ return ERR_PTR(ret);
+ }
+
+ spin_lock(root-lock);
+ ret = radix_tree_insert(root-hot_inode_tree, ino, he);
+ if (ret == -EEXIST) {
+ kmem_cache_free(hot_inode_item_cachep, he);
+ spin_unlock(root-lock);
+ radix_tree_preload_end();
+ goto again;
+ }
+ spin_unlock(root-lock);
+ radix_tree_preload_end();
+
+ kref_get(he-hot_inode.refs);
+ return he;
+}
+EXPORT_SYMBOL_GPL(hot_inode_item_find);
+
+static struct hot_range_item
+*hot_range_item_find(struct hot_inode_item *he,
+ u32 start)
+{
+ struct hot_range_item *hr;
+ int ret;
+
+again:
+ spin_lock(he-lock);
+ hr = radix_tree_lookup(he-hot_range_tree, start);
+ if (hr) {
+ kref_get(hr-hot_range.refs);
+ spin_unlock(he-lock);
+ return hr;
+ }
+ spin_unlock(he-lock);
+
+ hr = kmem_cache_zalloc(hot_range_item_cachep,
+ GFP_KERNEL | GFP_NOFS);
+ if (!hr)
+ return ERR_PTR(-ENOMEM);
+
+ hot_range_item_init(hr, start, he);
+
+ ret = radix_tree_preload(GFP_NOFS ~__GFP_HIGHMEM);
+ if (ret) {
+ kmem_cache_free(hot_range_item_cachep, hr);
+ return ERR_PTR(ret);
+ }
+
+ spin_lock(he-lock);
+ ret = radix_tree_insert(he-hot_range_tree, start, hr);
+ if (ret == -EEXIST) {
+ kmem_cache_free(hot_range_item_cachep, hr);
+ spin_unlock(he-lock);
+ radix_tree_preload_end();
+ goto again;
+ }
+ spin_unlock(he-lock);
+ radix_tree_preload_end();
+
+ kref_get(hr-hot_range.refs);
+ return hr;
+}
+
+/*
+ * This function does the actual work of updating
+ * the frequency numbers, whatever they turn out to be.
+ */
+static u64 hot_average_update(struct timespec old_atime,
+ struct timespec cur_time, u64 old_avg)
+{
+ struct timespec delta_ts;
+ u64 new_avg;
+ u64 new_delta;
+
+ delta_ts = timespec_sub(cur_time, old_atime);
+ new_delta = timespec_to_ns(delta_ts) FREQ_POWER;
+
+ new_avg = (old_avg FREQ_POWER) - old_avg + new_delta;
+ new_avg = new_avg FREQ_POWER;
+
+ return new_avg;
+}
+
+static void hot_freq_data_update(struct hot_freq_data *freq_data, bool
write)
+{
+ struct timespec cur_time = current_kernel_time();
+
+ if (write) {
+ freq_data-nr_writes += 1;
+ freq_data-avg_delta_writes = hot_average_update(
+ freq_data-last_write_time,
+ cur_time,
+ freq_data-avg_delta_writes);
+ freq_data-last_write_time = cur_time;
+ } else {
+ freq_data-nr_reads += 1;
+ freq_data-avg_delta_reads = hot_average_update(
+ freq_data-last_read_time,
+ cur_time,
+ freq_data-avg_delta_reads);
I think you could just pass in a pointer to
Re: [RFC v4+ hot_track 03/19] vfs: add I/O frequency update function
On Wed, Nov 7, 2012 at 6:37 AM, David Sterba wrote:
> On Mon, Oct 29, 2012 at 12:30:45PM +0800, zwu.ker...@gmail.com wrote:
>> --- a/fs/hot_tracking.c
>> +++ b/fs/hot_tracking.c
>> +struct hot_inode_item
>> +*hot_inode_item_find(struct hot_info *root, u64 ino)
>> +{
>> + struct hot_inode_item *he;
>> + int ret;
>> +
>> +again:
>> + spin_lock(>lock);
>> + he = radix_tree_lookup(>hot_inode_tree, ino);
>> + if (he) {
>> + kref_get(>hot_inode.refs);
>> + spin_unlock(>lock);
>> + return he;
>> + }
>> + spin_unlock(>lock);
>> +
>> + he = kmem_cache_zalloc(hot_inode_item_cachep,
>> + GFP_KERNEL | GFP_NOFS);
>> + if (!he)
>> + return ERR_PTR(-ENOMEM);
>> +
>> + hot_inode_item_init(he, ino, >hot_inode_tree);
>> +
>> + ret = radix_tree_preload(GFP_NOFS & ~__GFP_HIGHMEM);
>> + if (ret) {
>> + kmem_cache_free(hot_inode_item_cachep, he);
>
> radix_tree_preload_end()
>
>> + return ERR_PTR(ret);
>> + }
>> +
>> + spin_lock(>lock);
>> + ret = radix_tree_insert(>hot_inode_tree, ino, he);
>> + if (ret == -EEXIST) {
>> + kmem_cache_free(hot_inode_item_cachep, he);
>> + spin_unlock(>lock);
>> + radix_tree_preload_end();
>> + goto again;
>> + }
>> + spin_unlock(>lock);
>> + radix_tree_preload_end();
>> +
>> + kref_get(>hot_inode.refs);
>> + return he;
>> +}
>> +EXPORT_SYMBOL_GPL(hot_inode_item_find);
>> +
>> +static struct hot_range_item
>> +*hot_range_item_find(struct hot_inode_item *he,
>> + u32 start)
>> +{
>> + struct hot_range_item *hr;
>> + int ret;
>> +
>> +again:
>> + spin_lock(>lock);
>> + hr = radix_tree_lookup(>hot_range_tree, start);
>> + if (hr) {
>> + kref_get(>hot_range.refs);
>> + spin_unlock(>lock);
>> + return hr;
>> + }
>> + spin_unlock(>lock);
>> +
>> + hr = kmem_cache_zalloc(hot_range_item_cachep,
>> + GFP_KERNEL | GFP_NOFS);
>> + if (!hr)
>> + return ERR_PTR(-ENOMEM);
>> +
>> + hot_range_item_init(hr, start, he);
>> +
>> + ret = radix_tree_preload(GFP_NOFS & ~__GFP_HIGHMEM);
>> + if (ret) {
>> + kmem_cache_free(hot_range_item_cachep, hr);
>
> radix_tree_preload_end()
I checked some kernel existing cases about the usage of
radix_tree_preload(), it seems that when radix_tree_preload() fail,
its error handling doesn't need call radix_tree_preload_end() any
more.
>
>> + return ERR_PTR(ret);
>> + }
>> +
>> + spin_lock(>lock);
>> + ret = radix_tree_insert(>hot_range_tree, start, hr);
>> + if (ret == -EEXIST) {
>> + kmem_cache_free(hot_range_item_cachep, hr);
>> + spin_unlock(>lock);
>> + radix_tree_preload_end();
ditto.
>> + goto again;
>> + }
>> + spin_unlock(>lock);
>> + radix_tree_preload_end();
>> +
>> + kref_get(>hot_range.refs);
>> + return hr;
>> +}
>
> david
--
Regards,
Zhi Yong Wu
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majord...@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
Re: [RFC v4+ hot_track 03/19] vfs: add I/O frequency update function
On Mon, Oct 29, 2012 at 12:30:45PM +0800, zwu.ker...@gmail.com wrote:
> From: Zhi Yong Wu
>
> Add some util helpers to update access frequencies
> for one file or its range.
>
> Signed-off-by: Zhi Yong Wu
> ---
> fs/hot_tracking.c| 179
> ++
> fs/hot_tracking.h|7 ++
> include/linux/hot_tracking.h |2 +
> 3 files changed, 188 insertions(+), 0 deletions(-)
>
> diff --git a/fs/hot_tracking.c b/fs/hot_tracking.c
> index 68591f0..0a7d9a3 100644
> --- a/fs/hot_tracking.c
> +++ b/fs/hot_tracking.c
> @@ -172,6 +172,137 @@ static void hot_inode_tree_exit(struct hot_info *root)
> }
> }
>
> +struct hot_inode_item
> +*hot_inode_item_find(struct hot_info *root, u64 ino)
> +{
> + struct hot_inode_item *he;
> + int ret;
> +
> +again:
> + spin_lock(>lock);
> + he = radix_tree_lookup(>hot_inode_tree, ino);
> + if (he) {
> + kref_get(>hot_inode.refs);
> + spin_unlock(>lock);
> + return he;
> + }
> + spin_unlock(>lock);
> +
> + he = kmem_cache_zalloc(hot_inode_item_cachep,
> + GFP_KERNEL | GFP_NOFS);
> + if (!he)
> + return ERR_PTR(-ENOMEM);
> +
> + hot_inode_item_init(he, ino, >hot_inode_tree);
> +
> + ret = radix_tree_preload(GFP_NOFS & ~__GFP_HIGHMEM);
> + if (ret) {
> + kmem_cache_free(hot_inode_item_cachep, he);
> + return ERR_PTR(ret);
> + }
> +
> + spin_lock(>lock);
> + ret = radix_tree_insert(>hot_inode_tree, ino, he);
> + if (ret == -EEXIST) {
> + kmem_cache_free(hot_inode_item_cachep, he);
> + spin_unlock(>lock);
> + radix_tree_preload_end();
> + goto again;
> + }
> + spin_unlock(>lock);
> + radix_tree_preload_end();
> +
> + kref_get(>hot_inode.refs);
> + return he;
> +}
> +EXPORT_SYMBOL_GPL(hot_inode_item_find);
> +
> +static struct hot_range_item
> +*hot_range_item_find(struct hot_inode_item *he,
> + u32 start)
> +{
> + struct hot_range_item *hr;
> + int ret;
> +
> +again:
> + spin_lock(>lock);
> + hr = radix_tree_lookup(>hot_range_tree, start);
> + if (hr) {
> + kref_get(>hot_range.refs);
> + spin_unlock(>lock);
> + return hr;
> + }
> + spin_unlock(>lock);
> +
> + hr = kmem_cache_zalloc(hot_range_item_cachep,
> + GFP_KERNEL | GFP_NOFS);
> + if (!hr)
> + return ERR_PTR(-ENOMEM);
> +
> + hot_range_item_init(hr, start, he);
> +
> + ret = radix_tree_preload(GFP_NOFS & ~__GFP_HIGHMEM);
> + if (ret) {
> + kmem_cache_free(hot_range_item_cachep, hr);
> + return ERR_PTR(ret);
> + }
> +
> + spin_lock(>lock);
> + ret = radix_tree_insert(>hot_range_tree, start, hr);
> + if (ret == -EEXIST) {
> + kmem_cache_free(hot_range_item_cachep, hr);
> + spin_unlock(>lock);
> + radix_tree_preload_end();
> + goto again;
> + }
> + spin_unlock(>lock);
> + radix_tree_preload_end();
> +
> + kref_get(>hot_range.refs);
> + return hr;
> +}
> +
> +/*
> + * This function does the actual work of updating
> + * the frequency numbers, whatever they turn out to be.
> + */
> +static u64 hot_average_update(struct timespec old_atime,
> + struct timespec cur_time, u64 old_avg)
> +{
> + struct timespec delta_ts;
> + u64 new_avg;
> + u64 new_delta;
> +
> + delta_ts = timespec_sub(cur_time, old_atime);
> + new_delta = timespec_to_ns(_ts) >> FREQ_POWER;
> +
> + new_avg = (old_avg << FREQ_POWER) - old_avg + new_delta;
> + new_avg = new_avg >> FREQ_POWER;
> +
> + return new_avg;
> +}
> +
> +static void hot_freq_data_update(struct hot_freq_data *freq_data, bool write)
> +{
> + struct timespec cur_time = current_kernel_time();
> +
> + if (write) {
> + freq_data->nr_writes += 1;
> + freq_data->avg_delta_writes = hot_average_update(
> + freq_data->last_write_time,
> + cur_time,
> + freq_data->avg_delta_writes);
> + freq_data->last_write_time = cur_time;
> + } else {
> + freq_data->nr_reads += 1;
> + freq_data->avg_delta_reads = hot_average_update(
> + freq_data->last_read_time,
> + cur_time,
> + freq_data->avg_delta_reads);
I think you could just pass in a pointer to
freq_data->avg_delta_{writes,reads} here...
> + freq_data->last_read_time = cur_time;
> + }
> +}
> +
> /*
> * Initialize kmem cache for hot_inode_item and hot_range_item.
> */
> @@ -199,6 +330,54 @@ err:
> EXPORT_SYMBOL_GPL(hot_cache_init);
>
> /*
> + * Main function to update access frequency from read/writepage(s) hooks
> + */
Re: [RFC v4+ hot_track 03/19] vfs: add I/O frequency update function
On Mon, Oct 29, 2012 at 12:30:45PM +0800, zwu.ker...@gmail.com wrote:
> --- a/fs/hot_tracking.c
> +++ b/fs/hot_tracking.c
> +struct hot_inode_item
> +*hot_inode_item_find(struct hot_info *root, u64 ino)
> +{
> + struct hot_inode_item *he;
> + int ret;
> +
> +again:
> + spin_lock(>lock);
> + he = radix_tree_lookup(>hot_inode_tree, ino);
> + if (he) {
> + kref_get(>hot_inode.refs);
> + spin_unlock(>lock);
> + return he;
> + }
> + spin_unlock(>lock);
> +
> + he = kmem_cache_zalloc(hot_inode_item_cachep,
> + GFP_KERNEL | GFP_NOFS);
> + if (!he)
> + return ERR_PTR(-ENOMEM);
> +
> + hot_inode_item_init(he, ino, >hot_inode_tree);
> +
> + ret = radix_tree_preload(GFP_NOFS & ~__GFP_HIGHMEM);
> + if (ret) {
> + kmem_cache_free(hot_inode_item_cachep, he);
radix_tree_preload_end()
> + return ERR_PTR(ret);
> + }
> +
> + spin_lock(>lock);
> + ret = radix_tree_insert(>hot_inode_tree, ino, he);
> + if (ret == -EEXIST) {
> + kmem_cache_free(hot_inode_item_cachep, he);
> + spin_unlock(>lock);
> + radix_tree_preload_end();
> + goto again;
> + }
> + spin_unlock(>lock);
> + radix_tree_preload_end();
> +
> + kref_get(>hot_inode.refs);
> + return he;
> +}
> +EXPORT_SYMBOL_GPL(hot_inode_item_find);
> +
> +static struct hot_range_item
> +*hot_range_item_find(struct hot_inode_item *he,
> + u32 start)
> +{
> + struct hot_range_item *hr;
> + int ret;
> +
> +again:
> + spin_lock(>lock);
> + hr = radix_tree_lookup(>hot_range_tree, start);
> + if (hr) {
> + kref_get(>hot_range.refs);
> + spin_unlock(>lock);
> + return hr;
> + }
> + spin_unlock(>lock);
> +
> + hr = kmem_cache_zalloc(hot_range_item_cachep,
> + GFP_KERNEL | GFP_NOFS);
> + if (!hr)
> + return ERR_PTR(-ENOMEM);
> +
> + hot_range_item_init(hr, start, he);
> +
> + ret = radix_tree_preload(GFP_NOFS & ~__GFP_HIGHMEM);
> + if (ret) {
> + kmem_cache_free(hot_range_item_cachep, hr);
radix_tree_preload_end()
> + return ERR_PTR(ret);
> + }
> +
> + spin_lock(>lock);
> + ret = radix_tree_insert(>hot_range_tree, start, hr);
> + if (ret == -EEXIST) {
> + kmem_cache_free(hot_range_item_cachep, hr);
> + spin_unlock(>lock);
> + radix_tree_preload_end();
> + goto again;
> + }
> + spin_unlock(>lock);
> + radix_tree_preload_end();
> +
> + kref_get(>hot_range.refs);
> + return hr;
> +}
david
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majord...@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
Re: [RFC v4+ hot_track 03/19] vfs: add I/O frequency update function
On Mon, Oct 29, 2012 at 12:30:45PM +0800, zwu.ker...@gmail.com wrote:
--- a/fs/hot_tracking.c
+++ b/fs/hot_tracking.c
+struct hot_inode_item
+*hot_inode_item_find(struct hot_info *root, u64 ino)
+{
+ struct hot_inode_item *he;
+ int ret;
+
+again:
+ spin_lock(root-lock);
+ he = radix_tree_lookup(root-hot_inode_tree, ino);
+ if (he) {
+ kref_get(he-hot_inode.refs);
+ spin_unlock(root-lock);
+ return he;
+ }
+ spin_unlock(root-lock);
+
+ he = kmem_cache_zalloc(hot_inode_item_cachep,
+ GFP_KERNEL | GFP_NOFS);
+ if (!he)
+ return ERR_PTR(-ENOMEM);
+
+ hot_inode_item_init(he, ino, root-hot_inode_tree);
+
+ ret = radix_tree_preload(GFP_NOFS ~__GFP_HIGHMEM);
+ if (ret) {
+ kmem_cache_free(hot_inode_item_cachep, he);
radix_tree_preload_end()
+ return ERR_PTR(ret);
+ }
+
+ spin_lock(root-lock);
+ ret = radix_tree_insert(root-hot_inode_tree, ino, he);
+ if (ret == -EEXIST) {
+ kmem_cache_free(hot_inode_item_cachep, he);
+ spin_unlock(root-lock);
+ radix_tree_preload_end();
+ goto again;
+ }
+ spin_unlock(root-lock);
+ radix_tree_preload_end();
+
+ kref_get(he-hot_inode.refs);
+ return he;
+}
+EXPORT_SYMBOL_GPL(hot_inode_item_find);
+
+static struct hot_range_item
+*hot_range_item_find(struct hot_inode_item *he,
+ u32 start)
+{
+ struct hot_range_item *hr;
+ int ret;
+
+again:
+ spin_lock(he-lock);
+ hr = radix_tree_lookup(he-hot_range_tree, start);
+ if (hr) {
+ kref_get(hr-hot_range.refs);
+ spin_unlock(he-lock);
+ return hr;
+ }
+ spin_unlock(he-lock);
+
+ hr = kmem_cache_zalloc(hot_range_item_cachep,
+ GFP_KERNEL | GFP_NOFS);
+ if (!hr)
+ return ERR_PTR(-ENOMEM);
+
+ hot_range_item_init(hr, start, he);
+
+ ret = radix_tree_preload(GFP_NOFS ~__GFP_HIGHMEM);
+ if (ret) {
+ kmem_cache_free(hot_range_item_cachep, hr);
radix_tree_preload_end()
+ return ERR_PTR(ret);
+ }
+
+ spin_lock(he-lock);
+ ret = radix_tree_insert(he-hot_range_tree, start, hr);
+ if (ret == -EEXIST) {
+ kmem_cache_free(hot_range_item_cachep, hr);
+ spin_unlock(he-lock);
+ radix_tree_preload_end();
+ goto again;
+ }
+ spin_unlock(he-lock);
+ radix_tree_preload_end();
+
+ kref_get(hr-hot_range.refs);
+ return hr;
+}
david
--
To unsubscribe from this list: send the line unsubscribe linux-kernel in
the body of a message to majord...@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
Re: [RFC v4+ hot_track 03/19] vfs: add I/O frequency update function
On Mon, Oct 29, 2012 at 12:30:45PM +0800, zwu.ker...@gmail.com wrote:
From: Zhi Yong Wu wu...@linux.vnet.ibm.com
Add some util helpers to update access frequencies
for one file or its range.
Signed-off-by: Zhi Yong Wu wu...@linux.vnet.ibm.com
---
fs/hot_tracking.c| 179
++
fs/hot_tracking.h|7 ++
include/linux/hot_tracking.h |2 +
3 files changed, 188 insertions(+), 0 deletions(-)
diff --git a/fs/hot_tracking.c b/fs/hot_tracking.c
index 68591f0..0a7d9a3 100644
--- a/fs/hot_tracking.c
+++ b/fs/hot_tracking.c
@@ -172,6 +172,137 @@ static void hot_inode_tree_exit(struct hot_info *root)
}
}
+struct hot_inode_item
+*hot_inode_item_find(struct hot_info *root, u64 ino)
+{
+ struct hot_inode_item *he;
+ int ret;
+
+again:
+ spin_lock(root-lock);
+ he = radix_tree_lookup(root-hot_inode_tree, ino);
+ if (he) {
+ kref_get(he-hot_inode.refs);
+ spin_unlock(root-lock);
+ return he;
+ }
+ spin_unlock(root-lock);
+
+ he = kmem_cache_zalloc(hot_inode_item_cachep,
+ GFP_KERNEL | GFP_NOFS);
+ if (!he)
+ return ERR_PTR(-ENOMEM);
+
+ hot_inode_item_init(he, ino, root-hot_inode_tree);
+
+ ret = radix_tree_preload(GFP_NOFS ~__GFP_HIGHMEM);
+ if (ret) {
+ kmem_cache_free(hot_inode_item_cachep, he);
+ return ERR_PTR(ret);
+ }
+
+ spin_lock(root-lock);
+ ret = radix_tree_insert(root-hot_inode_tree, ino, he);
+ if (ret == -EEXIST) {
+ kmem_cache_free(hot_inode_item_cachep, he);
+ spin_unlock(root-lock);
+ radix_tree_preload_end();
+ goto again;
+ }
+ spin_unlock(root-lock);
+ radix_tree_preload_end();
+
+ kref_get(he-hot_inode.refs);
+ return he;
+}
+EXPORT_SYMBOL_GPL(hot_inode_item_find);
+
+static struct hot_range_item
+*hot_range_item_find(struct hot_inode_item *he,
+ u32 start)
+{
+ struct hot_range_item *hr;
+ int ret;
+
+again:
+ spin_lock(he-lock);
+ hr = radix_tree_lookup(he-hot_range_tree, start);
+ if (hr) {
+ kref_get(hr-hot_range.refs);
+ spin_unlock(he-lock);
+ return hr;
+ }
+ spin_unlock(he-lock);
+
+ hr = kmem_cache_zalloc(hot_range_item_cachep,
+ GFP_KERNEL | GFP_NOFS);
+ if (!hr)
+ return ERR_PTR(-ENOMEM);
+
+ hot_range_item_init(hr, start, he);
+
+ ret = radix_tree_preload(GFP_NOFS ~__GFP_HIGHMEM);
+ if (ret) {
+ kmem_cache_free(hot_range_item_cachep, hr);
+ return ERR_PTR(ret);
+ }
+
+ spin_lock(he-lock);
+ ret = radix_tree_insert(he-hot_range_tree, start, hr);
+ if (ret == -EEXIST) {
+ kmem_cache_free(hot_range_item_cachep, hr);
+ spin_unlock(he-lock);
+ radix_tree_preload_end();
+ goto again;
+ }
+ spin_unlock(he-lock);
+ radix_tree_preload_end();
+
+ kref_get(hr-hot_range.refs);
+ return hr;
+}
+
+/*
+ * This function does the actual work of updating
+ * the frequency numbers, whatever they turn out to be.
+ */
+static u64 hot_average_update(struct timespec old_atime,
+ struct timespec cur_time, u64 old_avg)
+{
+ struct timespec delta_ts;
+ u64 new_avg;
+ u64 new_delta;
+
+ delta_ts = timespec_sub(cur_time, old_atime);
+ new_delta = timespec_to_ns(delta_ts) FREQ_POWER;
+
+ new_avg = (old_avg FREQ_POWER) - old_avg + new_delta;
+ new_avg = new_avg FREQ_POWER;
+
+ return new_avg;
+}
+
+static void hot_freq_data_update(struct hot_freq_data *freq_data, bool write)
+{
+ struct timespec cur_time = current_kernel_time();
+
+ if (write) {
+ freq_data-nr_writes += 1;
+ freq_data-avg_delta_writes = hot_average_update(
+ freq_data-last_write_time,
+ cur_time,
+ freq_data-avg_delta_writes);
+ freq_data-last_write_time = cur_time;
+ } else {
+ freq_data-nr_reads += 1;
+ freq_data-avg_delta_reads = hot_average_update(
+ freq_data-last_read_time,
+ cur_time,
+ freq_data-avg_delta_reads);
I think you could just pass in a pointer to
freq_data-avg_delta_{writes,reads} here...
+ freq_data-last_read_time = cur_time;
+ }
+}
+
/*
* Initialize kmem cache for hot_inode_item and hot_range_item.
*/
@@ -199,6 +330,54 @@ err:
EXPORT_SYMBOL_GPL(hot_cache_init);
/*
+ * Main function to update access frequency from read/writepage(s) hooks
+ */
+void hot_update_freqs(struct inode *inode, u64 start,
+
Re: [RFC v4+ hot_track 03/19] vfs: add I/O frequency update function
On Wed, Nov 7, 2012 at 6:37 AM, David Sterba d...@jikos.cz wrote:
On Mon, Oct 29, 2012 at 12:30:45PM +0800, zwu.ker...@gmail.com wrote:
--- a/fs/hot_tracking.c
+++ b/fs/hot_tracking.c
+struct hot_inode_item
+*hot_inode_item_find(struct hot_info *root, u64 ino)
+{
+ struct hot_inode_item *he;
+ int ret;
+
+again:
+ spin_lock(root-lock);
+ he = radix_tree_lookup(root-hot_inode_tree, ino);
+ if (he) {
+ kref_get(he-hot_inode.refs);
+ spin_unlock(root-lock);
+ return he;
+ }
+ spin_unlock(root-lock);
+
+ he = kmem_cache_zalloc(hot_inode_item_cachep,
+ GFP_KERNEL | GFP_NOFS);
+ if (!he)
+ return ERR_PTR(-ENOMEM);
+
+ hot_inode_item_init(he, ino, root-hot_inode_tree);
+
+ ret = radix_tree_preload(GFP_NOFS ~__GFP_HIGHMEM);
+ if (ret) {
+ kmem_cache_free(hot_inode_item_cachep, he);
radix_tree_preload_end()
+ return ERR_PTR(ret);
+ }
+
+ spin_lock(root-lock);
+ ret = radix_tree_insert(root-hot_inode_tree, ino, he);
+ if (ret == -EEXIST) {
+ kmem_cache_free(hot_inode_item_cachep, he);
+ spin_unlock(root-lock);
+ radix_tree_preload_end();
+ goto again;
+ }
+ spin_unlock(root-lock);
+ radix_tree_preload_end();
+
+ kref_get(he-hot_inode.refs);
+ return he;
+}
+EXPORT_SYMBOL_GPL(hot_inode_item_find);
+
+static struct hot_range_item
+*hot_range_item_find(struct hot_inode_item *he,
+ u32 start)
+{
+ struct hot_range_item *hr;
+ int ret;
+
+again:
+ spin_lock(he-lock);
+ hr = radix_tree_lookup(he-hot_range_tree, start);
+ if (hr) {
+ kref_get(hr-hot_range.refs);
+ spin_unlock(he-lock);
+ return hr;
+ }
+ spin_unlock(he-lock);
+
+ hr = kmem_cache_zalloc(hot_range_item_cachep,
+ GFP_KERNEL | GFP_NOFS);
+ if (!hr)
+ return ERR_PTR(-ENOMEM);
+
+ hot_range_item_init(hr, start, he);
+
+ ret = radix_tree_preload(GFP_NOFS ~__GFP_HIGHMEM);
+ if (ret) {
+ kmem_cache_free(hot_range_item_cachep, hr);
radix_tree_preload_end()
I checked some kernel existing cases about the usage of
radix_tree_preload(), it seems that when radix_tree_preload() fail,
its error handling doesn't need call radix_tree_preload_end() any
more.
+ return ERR_PTR(ret);
+ }
+
+ spin_lock(he-lock);
+ ret = radix_tree_insert(he-hot_range_tree, start, hr);
+ if (ret == -EEXIST) {
+ kmem_cache_free(hot_range_item_cachep, hr);
+ spin_unlock(he-lock);
+ radix_tree_preload_end();
ditto.
+ goto again;
+ }
+ spin_unlock(he-lock);
+ radix_tree_preload_end();
+
+ kref_get(hr-hot_range.refs);
+ return hr;
+}
david
--
Regards,
Zhi Yong Wu
--
To unsubscribe from this list: send the line unsubscribe linux-kernel in
the body of a message to majord...@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
Re: [RFC v4+ hot_track 03/19] vfs: add I/O frequency update function
On Mon, Nov 5, 2012 at 7:07 PM, Steven Whitehouse wrote:
> Hi,
>
> On Mon, 2012-10-29 at 12:30 +0800, zwu.ker...@gmail.com wrote:
>> From: Zhi Yong Wu
>>
>> Add some util helpers to update access frequencies
>> for one file or its range.
>>
>> Signed-off-by: Zhi Yong Wu
>> ---
>> fs/hot_tracking.c| 179
>> ++
>> fs/hot_tracking.h|7 ++
>> include/linux/hot_tracking.h |2 +
>> 3 files changed, 188 insertions(+), 0 deletions(-)
>>
>> diff --git a/fs/hot_tracking.c b/fs/hot_tracking.c
>> index 68591f0..0a7d9a3 100644
>> --- a/fs/hot_tracking.c
>> +++ b/fs/hot_tracking.c
>> @@ -172,6 +172,137 @@ static void hot_inode_tree_exit(struct hot_info *root)
>> }
>> }
>>
>> +struct hot_inode_item
>> +*hot_inode_item_find(struct hot_info *root, u64 ino)
>> +{
>> + struct hot_inode_item *he;
>> + int ret;
>> +
>> +again:
>> + spin_lock(>lock);
>> + he = radix_tree_lookup(>hot_inode_tree, ino);
>> + if (he) {
>> + kref_get(>hot_inode.refs);
>> + spin_unlock(>lock);
>> + return he;
>> + }
>> + spin_unlock(>lock);
>> +
>> + he = kmem_cache_zalloc(hot_inode_item_cachep,
>> + GFP_KERNEL | GFP_NOFS);
> This doesn't look quite right... which of these two did you mean? I
> assume probably just GFP_NOFS
Yes, good catch, thanks.
>
>> + if (!he)
>> + return ERR_PTR(-ENOMEM);
>> +
>> + hot_inode_item_init(he, ino, >hot_inode_tree);
>> +
>> + ret = radix_tree_preload(GFP_NOFS & ~__GFP_HIGHMEM);
>> + if (ret) {
>> + kmem_cache_free(hot_inode_item_cachep, he);
>> + return ERR_PTR(ret);
>> + }
>> +
>> + spin_lock(>lock);
>> + ret = radix_tree_insert(>hot_inode_tree, ino, he);
>> + if (ret == -EEXIST) {
>> + kmem_cache_free(hot_inode_item_cachep, he);
>> + spin_unlock(>lock);
>> + radix_tree_preload_end();
>> + goto again;
>> + }
>> + spin_unlock(>lock);
>> + radix_tree_preload_end();
>> +
>> + kref_get(>hot_inode.refs);
>> + return he;
>> +}
>> +EXPORT_SYMBOL_GPL(hot_inode_item_find);
>> +
>> +static struct hot_range_item
>> +*hot_range_item_find(struct hot_inode_item *he,
>> + u32 start)
>> +{
>> + struct hot_range_item *hr;
>> + int ret;
>> +
>> +again:
>> + spin_lock(>lock);
>> + hr = radix_tree_lookup(>hot_range_tree, start);
>> + if (hr) {
>> + kref_get(>hot_range.refs);
>> + spin_unlock(>lock);
>> + return hr;
>> + }
>> + spin_unlock(>lock);
>> +
>> + hr = kmem_cache_zalloc(hot_range_item_cachep,
>> + GFP_KERNEL | GFP_NOFS);
> Likewise, here too.
ditto
>
> Steve.
>
>
>
--
Regards,
Zhi Yong Wu
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majord...@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
Re: [RFC v4+ hot_track 03/19] vfs: add I/O frequency update function
Hi,
On Mon, 2012-10-29 at 12:30 +0800, zwu.ker...@gmail.com wrote:
> From: Zhi Yong Wu
>
> Add some util helpers to update access frequencies
> for one file or its range.
>
> Signed-off-by: Zhi Yong Wu
> ---
> fs/hot_tracking.c| 179
> ++
> fs/hot_tracking.h|7 ++
> include/linux/hot_tracking.h |2 +
> 3 files changed, 188 insertions(+), 0 deletions(-)
>
> diff --git a/fs/hot_tracking.c b/fs/hot_tracking.c
> index 68591f0..0a7d9a3 100644
> --- a/fs/hot_tracking.c
> +++ b/fs/hot_tracking.c
> @@ -172,6 +172,137 @@ static void hot_inode_tree_exit(struct hot_info *root)
> }
> }
>
> +struct hot_inode_item
> +*hot_inode_item_find(struct hot_info *root, u64 ino)
> +{
> + struct hot_inode_item *he;
> + int ret;
> +
> +again:
> + spin_lock(>lock);
> + he = radix_tree_lookup(>hot_inode_tree, ino);
> + if (he) {
> + kref_get(>hot_inode.refs);
> + spin_unlock(>lock);
> + return he;
> + }
> + spin_unlock(>lock);
> +
> + he = kmem_cache_zalloc(hot_inode_item_cachep,
> + GFP_KERNEL | GFP_NOFS);
This doesn't look quite right... which of these two did you mean? I
assume probably just GFP_NOFS
> + if (!he)
> + return ERR_PTR(-ENOMEM);
> +
> + hot_inode_item_init(he, ino, >hot_inode_tree);
> +
> + ret = radix_tree_preload(GFP_NOFS & ~__GFP_HIGHMEM);
> + if (ret) {
> + kmem_cache_free(hot_inode_item_cachep, he);
> + return ERR_PTR(ret);
> + }
> +
> + spin_lock(>lock);
> + ret = radix_tree_insert(>hot_inode_tree, ino, he);
> + if (ret == -EEXIST) {
> + kmem_cache_free(hot_inode_item_cachep, he);
> + spin_unlock(>lock);
> + radix_tree_preload_end();
> + goto again;
> + }
> + spin_unlock(>lock);
> + radix_tree_preload_end();
> +
> + kref_get(>hot_inode.refs);
> + return he;
> +}
> +EXPORT_SYMBOL_GPL(hot_inode_item_find);
> +
> +static struct hot_range_item
> +*hot_range_item_find(struct hot_inode_item *he,
> + u32 start)
> +{
> + struct hot_range_item *hr;
> + int ret;
> +
> +again:
> + spin_lock(>lock);
> + hr = radix_tree_lookup(>hot_range_tree, start);
> + if (hr) {
> + kref_get(>hot_range.refs);
> + spin_unlock(>lock);
> + return hr;
> + }
> + spin_unlock(>lock);
> +
> + hr = kmem_cache_zalloc(hot_range_item_cachep,
> + GFP_KERNEL | GFP_NOFS);
Likewise, here too.
Steve.
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majord...@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
Re: [RFC v4+ hot_track 03/19] vfs: add I/O frequency update function
Hi,
On Mon, 2012-10-29 at 12:30 +0800, zwu.ker...@gmail.com wrote:
From: Zhi Yong Wu wu...@linux.vnet.ibm.com
Add some util helpers to update access frequencies
for one file or its range.
Signed-off-by: Zhi Yong Wu wu...@linux.vnet.ibm.com
---
fs/hot_tracking.c| 179
++
fs/hot_tracking.h|7 ++
include/linux/hot_tracking.h |2 +
3 files changed, 188 insertions(+), 0 deletions(-)
diff --git a/fs/hot_tracking.c b/fs/hot_tracking.c
index 68591f0..0a7d9a3 100644
--- a/fs/hot_tracking.c
+++ b/fs/hot_tracking.c
@@ -172,6 +172,137 @@ static void hot_inode_tree_exit(struct hot_info *root)
}
}
+struct hot_inode_item
+*hot_inode_item_find(struct hot_info *root, u64 ino)
+{
+ struct hot_inode_item *he;
+ int ret;
+
+again:
+ spin_lock(root-lock);
+ he = radix_tree_lookup(root-hot_inode_tree, ino);
+ if (he) {
+ kref_get(he-hot_inode.refs);
+ spin_unlock(root-lock);
+ return he;
+ }
+ spin_unlock(root-lock);
+
+ he = kmem_cache_zalloc(hot_inode_item_cachep,
+ GFP_KERNEL | GFP_NOFS);
This doesn't look quite right... which of these two did you mean? I
assume probably just GFP_NOFS
+ if (!he)
+ return ERR_PTR(-ENOMEM);
+
+ hot_inode_item_init(he, ino, root-hot_inode_tree);
+
+ ret = radix_tree_preload(GFP_NOFS ~__GFP_HIGHMEM);
+ if (ret) {
+ kmem_cache_free(hot_inode_item_cachep, he);
+ return ERR_PTR(ret);
+ }
+
+ spin_lock(root-lock);
+ ret = radix_tree_insert(root-hot_inode_tree, ino, he);
+ if (ret == -EEXIST) {
+ kmem_cache_free(hot_inode_item_cachep, he);
+ spin_unlock(root-lock);
+ radix_tree_preload_end();
+ goto again;
+ }
+ spin_unlock(root-lock);
+ radix_tree_preload_end();
+
+ kref_get(he-hot_inode.refs);
+ return he;
+}
+EXPORT_SYMBOL_GPL(hot_inode_item_find);
+
+static struct hot_range_item
+*hot_range_item_find(struct hot_inode_item *he,
+ u32 start)
+{
+ struct hot_range_item *hr;
+ int ret;
+
+again:
+ spin_lock(he-lock);
+ hr = radix_tree_lookup(he-hot_range_tree, start);
+ if (hr) {
+ kref_get(hr-hot_range.refs);
+ spin_unlock(he-lock);
+ return hr;
+ }
+ spin_unlock(he-lock);
+
+ hr = kmem_cache_zalloc(hot_range_item_cachep,
+ GFP_KERNEL | GFP_NOFS);
Likewise, here too.
Steve.
--
To unsubscribe from this list: send the line unsubscribe linux-kernel in
the body of a message to majord...@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
Re: [RFC v4+ hot_track 03/19] vfs: add I/O frequency update function
On Mon, Nov 5, 2012 at 7:07 PM, Steven Whitehouse swhit...@redhat.com wrote:
Hi,
On Mon, 2012-10-29 at 12:30 +0800, zwu.ker...@gmail.com wrote:
From: Zhi Yong Wu wu...@linux.vnet.ibm.com
Add some util helpers to update access frequencies
for one file or its range.
Signed-off-by: Zhi Yong Wu wu...@linux.vnet.ibm.com
---
fs/hot_tracking.c| 179
++
fs/hot_tracking.h|7 ++
include/linux/hot_tracking.h |2 +
3 files changed, 188 insertions(+), 0 deletions(-)
diff --git a/fs/hot_tracking.c b/fs/hot_tracking.c
index 68591f0..0a7d9a3 100644
--- a/fs/hot_tracking.c
+++ b/fs/hot_tracking.c
@@ -172,6 +172,137 @@ static void hot_inode_tree_exit(struct hot_info *root)
}
}
+struct hot_inode_item
+*hot_inode_item_find(struct hot_info *root, u64 ino)
+{
+ struct hot_inode_item *he;
+ int ret;
+
+again:
+ spin_lock(root-lock);
+ he = radix_tree_lookup(root-hot_inode_tree, ino);
+ if (he) {
+ kref_get(he-hot_inode.refs);
+ spin_unlock(root-lock);
+ return he;
+ }
+ spin_unlock(root-lock);
+
+ he = kmem_cache_zalloc(hot_inode_item_cachep,
+ GFP_KERNEL | GFP_NOFS);
This doesn't look quite right... which of these two did you mean? I
assume probably just GFP_NOFS
Yes, good catch, thanks.
+ if (!he)
+ return ERR_PTR(-ENOMEM);
+
+ hot_inode_item_init(he, ino, root-hot_inode_tree);
+
+ ret = radix_tree_preload(GFP_NOFS ~__GFP_HIGHMEM);
+ if (ret) {
+ kmem_cache_free(hot_inode_item_cachep, he);
+ return ERR_PTR(ret);
+ }
+
+ spin_lock(root-lock);
+ ret = radix_tree_insert(root-hot_inode_tree, ino, he);
+ if (ret == -EEXIST) {
+ kmem_cache_free(hot_inode_item_cachep, he);
+ spin_unlock(root-lock);
+ radix_tree_preload_end();
+ goto again;
+ }
+ spin_unlock(root-lock);
+ radix_tree_preload_end();
+
+ kref_get(he-hot_inode.refs);
+ return he;
+}
+EXPORT_SYMBOL_GPL(hot_inode_item_find);
+
+static struct hot_range_item
+*hot_range_item_find(struct hot_inode_item *he,
+ u32 start)
+{
+ struct hot_range_item *hr;
+ int ret;
+
+again:
+ spin_lock(he-lock);
+ hr = radix_tree_lookup(he-hot_range_tree, start);
+ if (hr) {
+ kref_get(hr-hot_range.refs);
+ spin_unlock(he-lock);
+ return hr;
+ }
+ spin_unlock(he-lock);
+
+ hr = kmem_cache_zalloc(hot_range_item_cachep,
+ GFP_KERNEL | GFP_NOFS);
Likewise, here too.
ditto
Steve.
--
Regards,
Zhi Yong Wu
--
To unsubscribe from this list: send the line unsubscribe linux-kernel in
the body of a message to majord...@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
[RFC v4+ hot_track 03/19] vfs: add I/O frequency update function
From: Zhi Yong Wu
Add some util helpers to update access frequencies
for one file or its range.
Signed-off-by: Zhi Yong Wu
---
fs/hot_tracking.c| 179 ++
fs/hot_tracking.h|7 ++
include/linux/hot_tracking.h |2 +
3 files changed, 188 insertions(+), 0 deletions(-)
diff --git a/fs/hot_tracking.c b/fs/hot_tracking.c
index 68591f0..0a7d9a3 100644
--- a/fs/hot_tracking.c
+++ b/fs/hot_tracking.c
@@ -172,6 +172,137 @@ static void hot_inode_tree_exit(struct hot_info *root)
}
}
+struct hot_inode_item
+*hot_inode_item_find(struct hot_info *root, u64 ino)
+{
+ struct hot_inode_item *he;
+ int ret;
+
+again:
+ spin_lock(>lock);
+ he = radix_tree_lookup(>hot_inode_tree, ino);
+ if (he) {
+ kref_get(>hot_inode.refs);
+ spin_unlock(>lock);
+ return he;
+ }
+ spin_unlock(>lock);
+
+ he = kmem_cache_zalloc(hot_inode_item_cachep,
+ GFP_KERNEL | GFP_NOFS);
+ if (!he)
+ return ERR_PTR(-ENOMEM);
+
+ hot_inode_item_init(he, ino, >hot_inode_tree);
+
+ ret = radix_tree_preload(GFP_NOFS & ~__GFP_HIGHMEM);
+ if (ret) {
+ kmem_cache_free(hot_inode_item_cachep, he);
+ return ERR_PTR(ret);
+ }
+
+ spin_lock(>lock);
+ ret = radix_tree_insert(>hot_inode_tree, ino, he);
+ if (ret == -EEXIST) {
+ kmem_cache_free(hot_inode_item_cachep, he);
+ spin_unlock(>lock);
+ radix_tree_preload_end();
+ goto again;
+ }
+ spin_unlock(>lock);
+ radix_tree_preload_end();
+
+ kref_get(>hot_inode.refs);
+ return he;
+}
+EXPORT_SYMBOL_GPL(hot_inode_item_find);
+
+static struct hot_range_item
+*hot_range_item_find(struct hot_inode_item *he,
+ u32 start)
+{
+ struct hot_range_item *hr;
+ int ret;
+
+again:
+ spin_lock(>lock);
+ hr = radix_tree_lookup(>hot_range_tree, start);
+ if (hr) {
+ kref_get(>hot_range.refs);
+ spin_unlock(>lock);
+ return hr;
+ }
+ spin_unlock(>lock);
+
+ hr = kmem_cache_zalloc(hot_range_item_cachep,
+ GFP_KERNEL | GFP_NOFS);
+ if (!hr)
+ return ERR_PTR(-ENOMEM);
+
+ hot_range_item_init(hr, start, he);
+
+ ret = radix_tree_preload(GFP_NOFS & ~__GFP_HIGHMEM);
+ if (ret) {
+ kmem_cache_free(hot_range_item_cachep, hr);
+ return ERR_PTR(ret);
+ }
+
+ spin_lock(>lock);
+ ret = radix_tree_insert(>hot_range_tree, start, hr);
+ if (ret == -EEXIST) {
+ kmem_cache_free(hot_range_item_cachep, hr);
+ spin_unlock(>lock);
+ radix_tree_preload_end();
+ goto again;
+ }
+ spin_unlock(>lock);
+ radix_tree_preload_end();
+
+ kref_get(>hot_range.refs);
+ return hr;
+}
+
+/*
+ * This function does the actual work of updating
+ * the frequency numbers, whatever they turn out to be.
+ */
+static u64 hot_average_update(struct timespec old_atime,
+ struct timespec cur_time, u64 old_avg)
+{
+ struct timespec delta_ts;
+ u64 new_avg;
+ u64 new_delta;
+
+ delta_ts = timespec_sub(cur_time, old_atime);
+ new_delta = timespec_to_ns(_ts) >> FREQ_POWER;
+
+ new_avg = (old_avg << FREQ_POWER) - old_avg + new_delta;
+ new_avg = new_avg >> FREQ_POWER;
+
+ return new_avg;
+}
+
+static void hot_freq_data_update(struct hot_freq_data *freq_data, bool write)
+{
+ struct timespec cur_time = current_kernel_time();
+
+ if (write) {
+ freq_data->nr_writes += 1;
+ freq_data->avg_delta_writes = hot_average_update(
+ freq_data->last_write_time,
+ cur_time,
+ freq_data->avg_delta_writes);
+ freq_data->last_write_time = cur_time;
+ } else {
+ freq_data->nr_reads += 1;
+ freq_data->avg_delta_reads = hot_average_update(
+ freq_data->last_read_time,
+ cur_time,
+ freq_data->avg_delta_reads);
+ freq_data->last_read_time = cur_time;
+ }
+}
+
/*
* Initialize kmem cache for hot_inode_item and hot_range_item.
*/
@@ -199,6 +330,54 @@ err:
EXPORT_SYMBOL_GPL(hot_cache_init);
/*
+ * Main function to update access frequency from read/writepage(s) hooks
+ */
+void hot_update_freqs(struct inode *inode, u64 start,
+ u64 len, int rw)
+{
+ struct hot_info *root = inode->i_sb->s_hot_root;
+ struct hot_inode_item *he;
+ struct hot_range_item *hr;
+ u32 cur, end;
+
+ if (!root || (len == 0))
+
[RFC v4+ hot_track 03/19] vfs: add I/O frequency update function
From: Zhi Yong Wu wu...@linux.vnet.ibm.com
Add some util helpers to update access frequencies
for one file or its range.
Signed-off-by: Zhi Yong Wu wu...@linux.vnet.ibm.com
---
fs/hot_tracking.c| 179 ++
fs/hot_tracking.h|7 ++
include/linux/hot_tracking.h |2 +
3 files changed, 188 insertions(+), 0 deletions(-)
diff --git a/fs/hot_tracking.c b/fs/hot_tracking.c
index 68591f0..0a7d9a3 100644
--- a/fs/hot_tracking.c
+++ b/fs/hot_tracking.c
@@ -172,6 +172,137 @@ static void hot_inode_tree_exit(struct hot_info *root)
}
}
+struct hot_inode_item
+*hot_inode_item_find(struct hot_info *root, u64 ino)
+{
+ struct hot_inode_item *he;
+ int ret;
+
+again:
+ spin_lock(root-lock);
+ he = radix_tree_lookup(root-hot_inode_tree, ino);
+ if (he) {
+ kref_get(he-hot_inode.refs);
+ spin_unlock(root-lock);
+ return he;
+ }
+ spin_unlock(root-lock);
+
+ he = kmem_cache_zalloc(hot_inode_item_cachep,
+ GFP_KERNEL | GFP_NOFS);
+ if (!he)
+ return ERR_PTR(-ENOMEM);
+
+ hot_inode_item_init(he, ino, root-hot_inode_tree);
+
+ ret = radix_tree_preload(GFP_NOFS ~__GFP_HIGHMEM);
+ if (ret) {
+ kmem_cache_free(hot_inode_item_cachep, he);
+ return ERR_PTR(ret);
+ }
+
+ spin_lock(root-lock);
+ ret = radix_tree_insert(root-hot_inode_tree, ino, he);
+ if (ret == -EEXIST) {
+ kmem_cache_free(hot_inode_item_cachep, he);
+ spin_unlock(root-lock);
+ radix_tree_preload_end();
+ goto again;
+ }
+ spin_unlock(root-lock);
+ radix_tree_preload_end();
+
+ kref_get(he-hot_inode.refs);
+ return he;
+}
+EXPORT_SYMBOL_GPL(hot_inode_item_find);
+
+static struct hot_range_item
+*hot_range_item_find(struct hot_inode_item *he,
+ u32 start)
+{
+ struct hot_range_item *hr;
+ int ret;
+
+again:
+ spin_lock(he-lock);
+ hr = radix_tree_lookup(he-hot_range_tree, start);
+ if (hr) {
+ kref_get(hr-hot_range.refs);
+ spin_unlock(he-lock);
+ return hr;
+ }
+ spin_unlock(he-lock);
+
+ hr = kmem_cache_zalloc(hot_range_item_cachep,
+ GFP_KERNEL | GFP_NOFS);
+ if (!hr)
+ return ERR_PTR(-ENOMEM);
+
+ hot_range_item_init(hr, start, he);
+
+ ret = radix_tree_preload(GFP_NOFS ~__GFP_HIGHMEM);
+ if (ret) {
+ kmem_cache_free(hot_range_item_cachep, hr);
+ return ERR_PTR(ret);
+ }
+
+ spin_lock(he-lock);
+ ret = radix_tree_insert(he-hot_range_tree, start, hr);
+ if (ret == -EEXIST) {
+ kmem_cache_free(hot_range_item_cachep, hr);
+ spin_unlock(he-lock);
+ radix_tree_preload_end();
+ goto again;
+ }
+ spin_unlock(he-lock);
+ radix_tree_preload_end();
+
+ kref_get(hr-hot_range.refs);
+ return hr;
+}
+
+/*
+ * This function does the actual work of updating
+ * the frequency numbers, whatever they turn out to be.
+ */
+static u64 hot_average_update(struct timespec old_atime,
+ struct timespec cur_time, u64 old_avg)
+{
+ struct timespec delta_ts;
+ u64 new_avg;
+ u64 new_delta;
+
+ delta_ts = timespec_sub(cur_time, old_atime);
+ new_delta = timespec_to_ns(delta_ts) FREQ_POWER;
+
+ new_avg = (old_avg FREQ_POWER) - old_avg + new_delta;
+ new_avg = new_avg FREQ_POWER;
+
+ return new_avg;
+}
+
+static void hot_freq_data_update(struct hot_freq_data *freq_data, bool write)
+{
+ struct timespec cur_time = current_kernel_time();
+
+ if (write) {
+ freq_data-nr_writes += 1;
+ freq_data-avg_delta_writes = hot_average_update(
+ freq_data-last_write_time,
+ cur_time,
+ freq_data-avg_delta_writes);
+ freq_data-last_write_time = cur_time;
+ } else {
+ freq_data-nr_reads += 1;
+ freq_data-avg_delta_reads = hot_average_update(
+ freq_data-last_read_time,
+ cur_time,
+ freq_data-avg_delta_reads);
+ freq_data-last_read_time = cur_time;
+ }
+}
+
/*
* Initialize kmem cache for hot_inode_item and hot_range_item.
*/
@@ -199,6 +330,54 @@ err:
EXPORT_SYMBOL_GPL(hot_cache_init);
/*
+ * Main function to update access frequency from read/writepage(s) hooks
+ */
+void hot_update_freqs(struct inode *inode, u64 start,
+ u64 len, int rw)
+{
+ struct hot_info *root = inode-i_sb-s_hot_root;
+ struct hot_inode_item *he;
+ struct
