From: "Theodore Ts'o" <tytso@mit.edu>
To: Ext4 Developers List <linux-ext4@vger.kernel.org>,
Linux Kernel Developers List <linux-kernel@vger.kernel.org>
Cc: Mingming Cao <cmm@us.ibm.com>, "Theodore Ts'o" <tytso@mit.edu>
Subject: [PATCH 43/52] percpu_counter: new function percpu_counter_sum_and_set
Date: Sat, 5 Jul 2008 13:36:09 -0400 [thread overview]
Message-ID: <1215279378-30504-44-git-send-email-tytso@mit.edu> (raw)
In-Reply-To: <1215279378-30504-43-git-send-email-tytso@mit.edu>
From: Mingming Cao <cmm@us.ibm.com>
Delayed allocation need to check free blocks at every write time.
percpu_counter_read_positive() is not quit accurate. delayed
allocation need a more accurate accounting, but using
percpu_counter_sum_positive() is frequently is quite expensive.
This patch added a new function to update center counter when sum
per-cpu counter, to increase the accurate rate for next
percpu_counter_read() and require less calling expensive
percpu_counter_sum().
Signed-off-by: Mingming Cao <cmm@us.ibm.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
---
fs/ext4/balloc.c | 2 +-
include/linux/percpu_counter.h | 12 +++++++++---
lib/percpu_counter.c | 7 ++++++-
3 files changed, 16 insertions(+), 5 deletions(-)
diff --git a/fs/ext4/balloc.c b/fs/ext4/balloc.c
index 25f63d8..6369bac 100644
--- a/fs/ext4/balloc.c
+++ b/fs/ext4/balloc.c
@@ -1621,7 +1621,7 @@ ext4_fsblk_t ext4_has_free_blocks(struct ext4_sb_info *sbi,
#ifdef CONFIG_SMP
if (free_blocks - root_blocks < FBC_BATCH)
free_blocks =
- percpu_counter_sum_positive(&sbi->s_freeblocks_counter);
+ percpu_counter_sum_and_set(&sbi->s_freeblocks_counter);
#endif
if (free_blocks - root_blocks < nblocks)
return free_blocks - root_blocks;
diff --git a/include/linux/percpu_counter.h b/include/linux/percpu_counter.h
index 9007ccd..2083888 100644
--- a/include/linux/percpu_counter.h
+++ b/include/linux/percpu_counter.h
@@ -35,7 +35,7 @@ int percpu_counter_init_irq(struct percpu_counter *fbc, s64 amount);
void percpu_counter_destroy(struct percpu_counter *fbc);
void percpu_counter_set(struct percpu_counter *fbc, s64 amount);
void __percpu_counter_add(struct percpu_counter *fbc, s64 amount, s32 batch);
-s64 __percpu_counter_sum(struct percpu_counter *fbc);
+s64 __percpu_counter_sum(struct percpu_counter *fbc, int set);
static inline void percpu_counter_add(struct percpu_counter *fbc, s64 amount)
{
@@ -44,13 +44,19 @@ static inline void percpu_counter_add(struct percpu_counter *fbc, s64 amount)
static inline s64 percpu_counter_sum_positive(struct percpu_counter *fbc)
{
- s64 ret = __percpu_counter_sum(fbc);
+ s64 ret = __percpu_counter_sum(fbc, 0);
return ret < 0 ? 0 : ret;
}
+static inline s64 percpu_counter_sum_and_set(struct percpu_counter *fbc)
+{
+ return __percpu_counter_sum(fbc, 1);
+}
+
+
static inline s64 percpu_counter_sum(struct percpu_counter *fbc)
{
- return __percpu_counter_sum(fbc);
+ return __percpu_counter_sum(fbc, 0);
}
static inline s64 percpu_counter_read(struct percpu_counter *fbc)
diff --git a/lib/percpu_counter.c b/lib/percpu_counter.c
index 1191744..4a8ba4b 100644
--- a/lib/percpu_counter.c
+++ b/lib/percpu_counter.c
@@ -52,7 +52,7 @@ EXPORT_SYMBOL(__percpu_counter_add);
* Add up all the per-cpu counts, return the result. This is a more accurate
* but much slower version of percpu_counter_read_positive()
*/
-s64 __percpu_counter_sum(struct percpu_counter *fbc)
+s64 __percpu_counter_sum(struct percpu_counter *fbc, int set)
{
s64 ret;
int cpu;
@@ -62,7 +62,12 @@ s64 __percpu_counter_sum(struct percpu_counter *fbc)
for_each_online_cpu(cpu) {
s32 *pcount = per_cpu_ptr(fbc->counters, cpu);
ret += *pcount;
+ if (set)
+ *pcount = 0;
}
+ if (set)
+ fbc->count = ret;
+
spin_unlock(&fbc->lock);
return ret;
}
--
1.5.6.rc3.1.g36b7.dirty
next prev parent reply other threads:[~2008-07-05 17:41 UTC|newest]
Thread overview: 53+ messages / expand[flat|nested] mbox.gz Atom feed top
2008-07-05 17:35 Ext4 patches for the next merge window Theodore Ts'o
2008-07-05 17:35 ` [PATCH 01/52] ext4: fix comments to say "ext4" Theodore Ts'o
2008-07-05 17:35 ` [PATCH 02/52] ext4: start searching for the right extent from the goal group Theodore Ts'o
2008-07-05 17:35 ` [PATCH 03/52] ext4: Use BUG_ON() instead of BUG() Theodore Ts'o
2008-07-05 17:35 ` [PATCH 04/52] ext4: switch to seq_files Theodore Ts'o
2008-07-05 17:35 ` [PATCH 05/52] ext4: improve some code in rb tree part of dir.c Theodore Ts'o
2008-07-05 17:35 ` [PATCH 06/52] ext4: Fix ext4_mb_init_cache return error Theodore Ts'o
2008-07-05 17:35 ` [PATCH 07/52] ext4: add error processing when calling ext4_mb_init_cache in mballoc Theodore Ts'o
2008-07-05 17:35 ` [PATCH 08/52] ext4: miscellaneous error checks and coding cleanups for mballoc Theodore Ts'o
2008-07-05 17:35 ` [PATCH 09/52] ext4: remove double definitions of xattr macros Theodore Ts'o
2008-07-05 17:35 ` [PATCH 10/52] ext4: Rename read_block_bitmap() to ext4_read_block_bitmap() Theodore Ts'o
2008-07-05 17:35 ` [PATCH 11/52] ext4: Remove unused variable from ext4_show_options Theodore Ts'o
2008-07-05 17:35 ` [PATCH 12/52] jbd2: Add commit time into the commit block Theodore Ts'o
2008-07-05 17:35 ` [PATCH 13/52] ext4: New inode allocation for FLEX_BG meta-data groups Theodore Ts'o
2008-07-05 17:35 ` [PATCH 14/52] jbd2: fix race between jbd2_journal_try_to_free_buffers() and jbd2 commit transaction Theodore Ts'o
2008-07-05 17:35 ` [PATCH 15/52] ext4: remove redundant code in ext4_fill_super() Theodore Ts'o
2008-07-05 17:35 ` [PATCH 16/52] ext4: remove quota allocation when ext4_mb_new_blocks fails Theodore Ts'o
2008-07-05 17:35 ` [PATCH 17/52] ext4: Update i_disksize properly when allocating from fallocate area Theodore Ts'o
2008-07-05 17:35 ` [PATCH 18/52] ext4: return error when calling ext4_ext_split failed Theodore Ts'o
2008-07-05 17:35 ` [PATCH 19/52] ext4: Make ext4_ext_find_extent fills ext_path completely Theodore Ts'o
2008-07-05 17:35 ` [PATCH 20/52] ext4: Fix ext4_ext_journal_restart() to reflect errors up to the caller Theodore Ts'o
2008-07-05 17:35 ` [PATCH 21/52] ext4: cleanup never-used magic numbers from htree code Theodore Ts'o
2008-07-05 17:35 ` [PATCH 22/52] ext4: Fix sparse warning Theodore Ts'o
2008-07-05 17:35 ` [PATCH 23/52] ext4: fix ext4_init_block_bitmap() for metablock block group Theodore Ts'o
2008-07-05 17:35 ` [PATCH 24/52] ext4: Use inode preallocation with -o noextents Theodore Ts'o
2008-07-05 17:35 ` [PATCH 25/52] ext4: cleanup block allocator Theodore Ts'o
2008-07-05 17:35 ` [PATCH 26/52] ext4: call blkdev_issue_flush on fsync Theodore Ts'o
2008-07-05 17:35 ` [PATCH 27/52] ext4: mballoc avoid use root reserved blocks for non root allocation Theodore Ts'o
2008-07-05 17:35 ` [PATCH 28/52] ext4: Set journal pointer to NULL when journal is released Theodore Ts'o
2008-07-05 17:35 ` [PATCH 29/52] ext4: use atomic functions to set bh_state Theodore Ts'o
2008-07-05 17:35 ` [PATCH 30/52] ext4: Add missing unlock to an error path in ext4_quota_write() Theodore Ts'o
2008-07-05 17:35 ` [PATCH 31/52] ext4: fix online resize with mballoc Theodore Ts'o
2008-07-05 17:35 ` [PATCH 32/52] ext4: Documentation updates Theodore Ts'o
2008-07-05 17:35 ` [PATCH 33/52] ext4: Use page_mkwrite vma_operations to get mmap write notification Theodore Ts'o
2008-07-05 17:36 ` [PATCH 34/52] vfs: Move mark_inode_dirty() from under page lock in generic_write_end() Theodore Ts'o
2008-07-05 17:36 ` [PATCH 35/52] ext4: Invert the locking order of page_lock and transaction start Theodore Ts'o
2008-07-05 17:36 ` [PATCH 36/52] ext4: Fix lock inversion in ext4_ext_truncate() Theodore Ts'o
2008-07-05 17:36 ` [PATCH 37/52] vfs: export filemap_fdatawrite_range() Theodore Ts'o
2008-07-05 17:36 ` [PATCH 38/52] jbd2: Implement data=ordered mode handling via inodes Theodore Ts'o
2008-07-05 17:36 ` [PATCH 39/52] ext4: Use new framework for data=ordered mode in JBD2 Theodore Ts'o
2008-07-05 17:36 ` [PATCH 40/52] jbd2: Remove data=ordered mode support using jbd buffer heads Theodore Ts'o
2008-07-05 17:36 ` [PATCH 41/52] vfs: add basic delayed allocation support Theodore Ts'o
2008-07-05 17:36 ` [PATCH 42/52] ext4: Add " Theodore Ts'o
2008-07-05 17:36 ` Theodore Ts'o [this message]
2008-07-05 17:36 ` [PATCH 44/52] ext4: delayed allocation ENOSPC handling Theodore Ts'o
2008-07-05 17:36 ` [PATCH 45/52] mm: Add range_cont mode for writeback Theodore Ts'o
2008-07-05 17:36 ` [PATCH 46/52] ext4: Invert lock ordering of page_lock and transaction start in delalloc Theodore Ts'o
2008-07-05 17:36 ` [PATCH 47/52] ext4: Add ordered mode support for delalloc Theodore Ts'o
2008-07-05 17:36 ` [PATCH 48/52] ext4: Handle page without buffers in ext4_*_writepage() Theodore Ts'o
2008-07-05 17:36 ` [PATCH 49/52] ext4: fix delalloc i_disksize early update issue Theodore Ts'o
2008-07-05 17:36 ` [PATCH 50/52] ext4: Enable delalloc by default Theodore Ts'o
2008-07-05 17:36 ` [PATCH 51/52] ext4: Don't allow nonextenst mount option for large filesystem Theodore Ts'o
2008-07-05 17:36 ` [PATCH 52/52] ext4: Documention update for new ordered mode and delayed allocation Theodore Ts'o
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=1215279378-30504-44-git-send-email-tytso@mit.edu \
--to=tytso@mit.edu \
--cc=cmm@us.ibm.com \
--cc=linux-ext4@vger.kernel.org \
--cc=linux-kernel@vger.kernel.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox
all inboxes | Powered by JetHome®