* [Ocfs2-devel] [PATCH 00/34] OCFS2: Add trace event and replace mlog(0).
@ 2010-12-23 7:19 Tao Ma
2010-12-23 7:30 ` [Ocfs2-devel] [PATCH 01/34] ocfs2: Remove unused truncate function from alloc.c Tao Ma
` (38 more replies)
0 siblings, 39 replies; 53+ messages in thread
From: Tao Ma @ 2010-12-23 7:19 UTC (permalink / raw)
To: ocfs2-devel
Hi all,
This patch set is my attempt to replace our old debug(mlog) with new
trace events.
Wengang Wang has did some work in this field about one year ago.
http://oss.oracle.com/pipermail/ocfs2-devel/2009-September/005230.html
But at that time we don't have ocfs2 1.6 introduced, so we delay this
work until now after ocfs2 1.6 and UEK(Unbreakable Enterprise Kernel)
are released.
All patches except one(see below) are only mlog related, so they
shouldn't affect the normal operations in ocfs2 and they should be easy
for review(patch 1-5 were already sent to the mail list around one month
ago).
The only one that changes the codes somehow are patch 15, "ocfs2: Little
refactoring against ocfs2_iget", but it looks straightforward.
Regards,
Tao
^ permalink raw reply [flat|nested] 53+ messages in thread
* [Ocfs2-devel] [PATCH 01/34] ocfs2: Remove unused truncate function from alloc.c
2010-12-23 7:19 [Ocfs2-devel] [PATCH 00/34] OCFS2: Add trace event and replace mlog(0) Tao Ma
@ 2010-12-23 7:30 ` Tao Ma
2010-12-28 22:54 ` Mark Fasheh
2010-12-23 7:30 ` [Ocfs2-devel] [PATCH 02/34] ocfs2: Remove ENTRY from masklog Tao Ma
` (37 subsequent siblings)
38 siblings, 1 reply; 53+ messages in thread
From: Tao Ma @ 2010-12-23 7:30 UTC (permalink / raw)
To: ocfs2-devel
From: Tao Ma <tao.ma@oracle.com>
Tristan Ye has done some refactoring against our truncate
process, so some functions like ocfs2_prepare_truncate and
ocfs2_free_truncate_context are no use and we'd better
remove them.
Signed-off-by: Tao Ma <tao.ma@oracle.com>
---
fs/ocfs2/alloc.c | 74 ------------------------------------------------------
fs/ocfs2/alloc.h | 4 ---
2 files changed, 0 insertions(+), 78 deletions(-)
diff --git a/fs/ocfs2/alloc.c b/fs/ocfs2/alloc.c
index 592fae5..cd9d1bc 100644
--- a/fs/ocfs2/alloc.c
+++ b/fs/ocfs2/alloc.c
@@ -565,7 +565,6 @@ static inline int ocfs2_et_sanity_check(struct ocfs2_extent_tree *et)
return ret;
}
-static void ocfs2_free_truncate_context(struct ocfs2_truncate_context *tc);
static int ocfs2_cache_extent_block_free(struct ocfs2_cached_dealloc_ctxt *ctxt,
struct ocfs2_extent_block *eb);
static void ocfs2_adjust_rightmost_records(handle_t *handle,
@@ -7139,64 +7138,6 @@ bail:
}
/*
- * Expects the inode to already be locked.
- */
-int ocfs2_prepare_truncate(struct ocfs2_super *osb,
- struct inode *inode,
- struct buffer_head *fe_bh,
- struct ocfs2_truncate_context **tc)
-{
- int status;
- unsigned int new_i_clusters;
- struct ocfs2_dinode *fe;
- struct ocfs2_extent_block *eb;
- struct buffer_head *last_eb_bh = NULL;
-
- mlog_entry_void();
-
- *tc = NULL;
-
- new_i_clusters = ocfs2_clusters_for_bytes(osb->sb,
- i_size_read(inode));
- fe = (struct ocfs2_dinode *) fe_bh->b_data;
-
- mlog(0, "fe->i_clusters = %u, new_i_clusters = %u, fe->i_size ="
- "%llu\n", le32_to_cpu(fe->i_clusters), new_i_clusters,
- (unsigned long long)le64_to_cpu(fe->i_size));
-
- *tc = kzalloc(sizeof(struct ocfs2_truncate_context), GFP_KERNEL);
- if (!(*tc)) {
- status = -ENOMEM;
- mlog_errno(status);
- goto bail;
- }
- ocfs2_init_dealloc_ctxt(&(*tc)->tc_dealloc);
-
- if (fe->id2.i_list.l_tree_depth) {
- status = ocfs2_read_extent_block(INODE_CACHE(inode),
- le64_to_cpu(fe->i_last_eb_blk),
- &last_eb_bh);
- if (status < 0) {
- mlog_errno(status);
- goto bail;
- }
- eb = (struct ocfs2_extent_block *) last_eb_bh->b_data;
- }
-
- (*tc)->tc_last_eb_bh = last_eb_bh;
-
- status = 0;
-bail:
- if (status < 0) {
- if (*tc)
- ocfs2_free_truncate_context(*tc);
- *tc = NULL;
- }
- mlog_exit_void();
- return status;
-}
-
-/*
* 'start' is inclusive, 'end' is not.
*/
int ocfs2_truncate_inline(struct inode *inode, struct buffer_head *di_bh,
@@ -7270,18 +7211,3 @@ out_commit:
out:
return ret;
}
-
-static void ocfs2_free_truncate_context(struct ocfs2_truncate_context *tc)
-{
- /*
- * The caller is responsible for completing deallocation
- * before freeing the context.
- */
- if (tc->tc_dealloc.c_first_suballocator != NULL)
- mlog(ML_NOTICE,
- "Truncate completion has non-empty dealloc context\n");
-
- brelse(tc->tc_last_eb_bh);
-
- kfree(tc);
-}
diff --git a/fs/ocfs2/alloc.h b/fs/ocfs2/alloc.h
index 55762b5..3bd08a0 100644
--- a/fs/ocfs2/alloc.h
+++ b/fs/ocfs2/alloc.h
@@ -228,10 +228,6 @@ struct ocfs2_truncate_context {
int ocfs2_zero_range_for_truncate(struct inode *inode, handle_t *handle,
u64 range_start, u64 range_end);
-int ocfs2_prepare_truncate(struct ocfs2_super *osb,
- struct inode *inode,
- struct buffer_head *fe_bh,
- struct ocfs2_truncate_context **tc);
int ocfs2_commit_truncate(struct ocfs2_super *osb,
struct inode *inode,
struct buffer_head *di_bh);
--
1.7.0.4
^ permalink raw reply related [flat|nested] 53+ messages in thread
* [Ocfs2-devel] [PATCH 02/34] ocfs2: Remove ENTRY from masklog.
2010-12-23 7:19 [Ocfs2-devel] [PATCH 00/34] OCFS2: Add trace event and replace mlog(0) Tao Ma
2010-12-23 7:30 ` [Ocfs2-devel] [PATCH 01/34] ocfs2: Remove unused truncate function from alloc.c Tao Ma
@ 2010-12-23 7:30 ` Tao Ma
2010-12-23 7:30 ` [Ocfs2-devel] [PATCH 03/34] ocfs2: Remove EXIT " Tao Ma
` (36 subsequent siblings)
38 siblings, 0 replies; 53+ messages in thread
From: Tao Ma @ 2010-12-23 7:30 UTC (permalink / raw)
To: ocfs2-devel
From: Tao Ma <tao.ma@oracle.com>
ENTRY is used to record the entry of a function.
But because it is added in so many functions, if we enable it,
the system logs get filled up quickly and cause too much I/O.
So actually no one can open it for a production system or even
for a test.
So for mlog_entry_void, we just remove it.
for mlog_entry(...), we replace it with mlog(0,...), and they
will be replace by trace event later.
Signed-off-by: Tao Ma <tao.ma@oracle.com>
---
fs/ocfs2/alloc.c | 28 +--------------
fs/ocfs2/aops.c | 16 ++++-----
fs/ocfs2/buffer_head_io.c | 10 ++---
fs/ocfs2/cluster/heartbeat.c | 2 -
fs/ocfs2/cluster/masklog.c | 1 -
fs/ocfs2/cluster/masklog.h | 13 +------
fs/ocfs2/dcache.c | 4 +-
fs/ocfs2/dir.c | 20 +++--------
fs/ocfs2/dlm/dlmast.c | 19 ++--------
fs/ocfs2/dlm/dlmconvert.c | 6 ++--
fs/ocfs2/dlm/dlmdomain.c | 8 ++--
fs/ocfs2/dlm/dlmlock.c | 10 ++----
fs/ocfs2/dlm/dlmmaster.c | 6 +---
fs/ocfs2/dlm/dlmrecovery.c | 4 --
fs/ocfs2/dlm/dlmthread.c | 8 ++--
fs/ocfs2/dlm/dlmunlock.c | 4 +--
fs/ocfs2/dlmglue.c | 78 ------------------------------------------
fs/ocfs2/export.c | 12 +++---
fs/ocfs2/extent_map.c | 8 ++--
fs/ocfs2/file.c | 73 +++++++++++++++++----------------------
fs/ocfs2/inode.c | 28 ++++++---------
fs/ocfs2/journal.c | 43 ++++++----------------
fs/ocfs2/localalloc.c | 26 +++-----------
fs/ocfs2/mmap.c | 2 +-
fs/ocfs2/namei.c | 74 +++++++++++++++++++---------------------
fs/ocfs2/quota_global.c | 14 +++-----
fs/ocfs2/quota_local.c | 2 +-
fs/ocfs2/resize.c | 8 +---
fs/ocfs2/slot_map.c | 2 -
fs/ocfs2/suballoc.c | 35 +++---------------
fs/ocfs2/super.c | 32 +++--------------
fs/ocfs2/symlink.c | 6 ---
fs/ocfs2/xattr.c | 4 +-
33 files changed, 168 insertions(+), 438 deletions(-)
diff --git a/fs/ocfs2/alloc.c b/fs/ocfs2/alloc.c
index cd9d1bc..c12e9d8 100644
--- a/fs/ocfs2/alloc.c
+++ b/fs/ocfs2/alloc.c
@@ -965,8 +965,6 @@ int ocfs2_num_free_extents(struct ocfs2_super *osb,
struct buffer_head *eb_bh = NULL;
u64 last_eb_blk = 0;
- mlog_entry_void();
-
el = et->et_root_el;
last_eb_blk = ocfs2_et_get_last_eb_blk(et);
@@ -1010,8 +1008,6 @@ static int ocfs2_create_new_meta_bhs(handle_t *handle,
OCFS2_SB(ocfs2_metadata_cache_get_super(et->et_ci));
struct ocfs2_extent_block *eb;
- mlog_entry_void();
-
count = 0;
while (count < wanted) {
status = ocfs2_claim_metadata(handle,
@@ -1173,8 +1169,6 @@ static int ocfs2_add_branch(handle_t *handle,
struct ocfs2_extent_list *el;
u32 new_cpos, root_end;
- mlog_entry_void();
-
BUG_ON(!last_eb_bh || !*last_eb_bh);
if (eb_bh) {
@@ -1353,8 +1347,6 @@ static int ocfs2_shift_tree_depth(handle_t *handle,
struct ocfs2_extent_list *root_el;
struct ocfs2_extent_list *eb_el;
- mlog_entry_void();
-
status = ocfs2_create_new_meta_bhs(handle, et, 1, meta_ac,
&new_eb_bh);
if (status < 0) {
@@ -1446,8 +1438,6 @@ static int ocfs2_find_branch_target(struct ocfs2_extent_tree *et,
struct buffer_head *bh = NULL;
struct buffer_head *lowest_bh = NULL;
- mlog_entry_void();
-
*target_bh = NULL;
el = et->et_root_el;
@@ -5795,8 +5785,8 @@ int ocfs2_truncate_log_append(struct ocfs2_super *osb,
struct ocfs2_dinode *di;
struct ocfs2_truncate_log *tl;
- mlog_entry("start_blk = %llu, num_clusters = %u\n",
- (unsigned long long)start_blk, num_clusters);
+ mlog(0, "start_blk = %llu, num_clusters = %u\n",
+ (unsigned long long)start_blk, num_clusters);
BUG_ON(mutex_trylock(&tl_inode->i_mutex));
@@ -5877,8 +5867,6 @@ static int ocfs2_replay_truncate_records(struct ocfs2_super *osb,
struct inode *tl_inode = osb->osb_tl_inode;
struct buffer_head *tl_bh = osb->osb_tl_bh;
- mlog_entry_void();
-
di = (struct ocfs2_dinode *) tl_bh->b_data;
tl = &di->id2.i_dealloc;
i = le16_to_cpu(tl->tl_used) - 1;
@@ -5946,8 +5934,6 @@ int __ocfs2_flush_truncate_log(struct ocfs2_super *osb)
struct ocfs2_dinode *di;
struct ocfs2_truncate_log *tl;
- mlog_entry_void();
-
BUG_ON(mutex_trylock(&tl_inode->i_mutex));
di = (struct ocfs2_dinode *) tl_bh->b_data;
@@ -6029,8 +6015,6 @@ static void ocfs2_truncate_log_worker(struct work_struct *work)
container_of(work, struct ocfs2_super,
osb_truncate_log_wq.work);
- mlog_entry_void();
-
status = ocfs2_flush_truncate_log(osb);
if (status < 0)
mlog_errno(status);
@@ -6171,8 +6155,6 @@ int ocfs2_complete_truncate_log_recovery(struct ocfs2_super *osb,
struct inode *tl_inode = osb->osb_tl_inode;
struct ocfs2_truncate_log *tl;
- mlog_entry_void();
-
if (OCFS2_I(tl_inode)->ip_blkno == le64_to_cpu(tl_copy->i_blkno)) {
mlog(ML_ERROR, "Asked to recover my own truncate log!\n");
return -EINVAL;
@@ -6225,8 +6207,6 @@ void ocfs2_truncate_log_shutdown(struct ocfs2_super *osb)
int status;
struct inode *tl_inode = osb->osb_tl_inode;
- mlog_entry_void();
-
if (tl_inode) {
cancel_delayed_work(&osb->osb_truncate_log_wq);
flush_workqueue(ocfs2_wq);
@@ -6248,8 +6228,6 @@ int ocfs2_truncate_log_init(struct ocfs2_super *osb)
struct inode *tl_inode = NULL;
struct buffer_head *tl_bh = NULL;
- mlog_entry_void();
-
status = ocfs2_get_truncate_log_info(osb,
osb->slot_num,
&tl_inode,
@@ -7002,8 +6980,6 @@ int ocfs2_commit_truncate(struct ocfs2_super *osb,
struct ocfs2_extent_tree et;
struct ocfs2_cached_dealloc_ctxt dealloc;
- mlog_entry_void();
-
ocfs2_init_dinode_extent_tree(&et, INODE_CACHE(inode), di_bh);
ocfs2_init_dealloc_ctxt(&dealloc);
diff --git a/fs/ocfs2/aops.c b/fs/ocfs2/aops.c
index f1e962c..9bff526 100644
--- a/fs/ocfs2/aops.c
+++ b/fs/ocfs2/aops.c
@@ -59,8 +59,8 @@ static int ocfs2_symlink_get_block(struct inode *inode, sector_t iblock,
struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
void *kaddr;
- mlog_entry("(0x%p, %llu, 0x%p, %d)\n", inode,
- (unsigned long long)iblock, bh_result, create);
+ mlog(0, "(0x%p, %llu, 0x%p, %d)\n", inode,
+ (unsigned long long)iblock, bh_result, create);
BUG_ON(ocfs2_inode_is_fast_symlink(inode));
@@ -136,8 +136,8 @@ int ocfs2_get_block(struct inode *inode, sector_t iblock,
u64 p_blkno, count, past_eof;
struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
- mlog_entry("(0x%p, %llu, 0x%p, %d)\n", inode,
- (unsigned long long)iblock, bh_result, create);
+ mlog(0, "(0x%p, %llu, 0x%p, %d)\n", inode,
+ (unsigned long long)iblock, bh_result, create);
if (OCFS2_I(inode)->ip_flags & OCFS2_INODE_SYSTEM_FILE)
mlog(ML_NOTICE, "get_block on system inode 0x%p (%lu)\n",
@@ -278,7 +278,7 @@ static int ocfs2_readpage(struct file *file, struct page *page)
loff_t start = (loff_t)page->index << PAGE_CACHE_SHIFT;
int ret, unlock = 1;
- mlog_entry("(0x%p, %lu)\n", file, (page ? page->index : 0));
+ mlog(0, "(0x%p, %lu)\n", file, (page ? page->index : 0));
ret = ocfs2_inode_lock_with_page(inode, NULL, 0, page);
if (ret != 0) {
@@ -398,7 +398,7 @@ static int ocfs2_writepage(struct page *page, struct writeback_control *wbc)
{
int ret;
- mlog_entry("(0x%p)\n", page);
+ mlog(0, "(0x%p)\n", page);
ret = block_write_full_page(page, ocfs2_get_block, wbc);
@@ -450,7 +450,7 @@ static sector_t ocfs2_bmap(struct address_space *mapping, sector_t block)
int err = 0;
struct inode *inode = mapping->host;
- mlog_entry("(block = %llu)\n", (unsigned long long)block);
+ mlog(0, "(block = %llu)\n", (unsigned long long)block);
/* We don't need to lock journal system files, since they aren't
* accessed concurrently from multiple nodes.
@@ -615,8 +615,6 @@ static ssize_t ocfs2_direct_IO(int rw,
struct inode *inode = file->f_path.dentry->d_inode->i_mapping->host;
int ret;
- mlog_entry_void();
-
/*
* Fallback to buffered I/O if we see an inode without
* extents.
diff --git a/fs/ocfs2/buffer_head_io.c b/fs/ocfs2/buffer_head_io.c
index f9d5d3f..024bd45 100644
--- a/fs/ocfs2/buffer_head_io.c
+++ b/fs/ocfs2/buffer_head_io.c
@@ -55,8 +55,8 @@ int ocfs2_write_block(struct ocfs2_super *osb, struct buffer_head *bh,
{
int ret = 0;
- mlog_entry("(bh->b_blocknr = %llu, ci=%p)\n",
- (unsigned long long)bh->b_blocknr, ci);
+ mlog(0, "(bh->b_blocknr = %llu, ci=%p)\n",
+ (unsigned long long)bh->b_blocknr, ci);
BUG_ON(bh->b_blocknr < OCFS2_SUPER_BLOCK_BLKNO);
BUG_ON(buffer_jbd(bh));
@@ -186,8 +186,8 @@ int ocfs2_read_blocks(struct ocfs2_caching_info *ci, u64 block, int nr,
struct buffer_head *bh;
struct super_block *sb = ocfs2_metadata_cache_get_super(ci);
- mlog_entry("(ci=%p, block=(%llu), nr=(%d), flags=%d)\n",
- ci, (unsigned long long)block, nr, flags);
+ mlog(0, "(ci=%p, block=(%llu), nr=(%d), flags=%d)\n",
+ ci, (unsigned long long)block, nr, flags);
BUG_ON(!ci);
BUG_ON((flags & OCFS2_BH_READAHEAD) &&
@@ -408,8 +408,6 @@ int ocfs2_write_super_or_backup(struct ocfs2_super *osb,
int ret = 0;
struct ocfs2_dinode *di = (struct ocfs2_dinode *)bh->b_data;
- mlog_entry_void();
-
BUG_ON(buffer_jbd(bh));
ocfs2_check_super_or_backup(osb->sb, bh->b_blocknr);
diff --git a/fs/ocfs2/cluster/heartbeat.c b/fs/ocfs2/cluster/heartbeat.c
index 9f26ac9..4301242 100644
--- a/fs/ocfs2/cluster/heartbeat.c
+++ b/fs/ocfs2/cluster/heartbeat.c
@@ -1603,8 +1603,6 @@ static int o2hb_populate_slot_data(struct o2hb_region *reg)
struct o2hb_disk_slot *slot;
struct o2hb_disk_heartbeat_block *hb_block;
- mlog_entry_void();
-
ret = o2hb_read_slots(reg, reg->hr_blocks);
if (ret) {
mlog_errno(ret);
diff --git a/fs/ocfs2/cluster/masklog.c b/fs/ocfs2/cluster/masklog.c
index c7fba39..b9ff806 100644
--- a/fs/ocfs2/cluster/masklog.c
+++ b/fs/ocfs2/cluster/masklog.c
@@ -80,7 +80,6 @@ struct mlog_attribute {
}
static struct mlog_attribute mlog_attrs[MLOG_MAX_BITS] = {
- define_mask(ENTRY),
define_mask(EXIT),
define_mask(TCP),
define_mask(MSG),
diff --git a/fs/ocfs2/cluster/masklog.h b/fs/ocfs2/cluster/masklog.h
index ea2ed9f..72d3599 100644
--- a/fs/ocfs2/cluster/masklog.h
+++ b/fs/ocfs2/cluster/masklog.h
@@ -82,7 +82,6 @@
/* bits that are frequently given and infrequently matched in the low word */
/* NOTE: If you add a flag, you need to also update mlog.c! */
-#define ML_ENTRY 0x0000000000000001ULL /* func call entry */
#define ML_EXIT 0x0000000000000002ULL /* func call exit */
#define ML_TCP 0x0000000000000004ULL /* net cluster/tcp.c */
#define ML_MSG 0x0000000000000008ULL /* net network messages */
@@ -123,7 +122,7 @@
#define ML_CLUSTER 0x0000001000000000ULL /* cluster stack */
#define MLOG_INITIAL_AND_MASK (ML_ERROR|ML_NOTICE)
-#define MLOG_INITIAL_NOT_MASK (ML_ENTRY|ML_EXIT)
+#define MLOG_INITIAL_NOT_MASK (ML_EXIT)
#ifndef MLOG_MASK_PREFIX
#define MLOG_MASK_PREFIX 0
#endif
@@ -222,14 +221,6 @@ extern struct mlog_bits mlog_and_bits, mlog_not_bits;
} while (0)
#if defined(CONFIG_OCFS2_DEBUG_MASKLOG)
-#define mlog_entry(fmt, args...) do { \
- mlog(ML_ENTRY, "ENTRY:" fmt , ##args); \
-} while (0)
-
-#define mlog_entry_void() do { \
- mlog(ML_ENTRY, "ENTRY:\n"); \
-} while (0)
-
/*
* We disable this for sparse.
*/
@@ -266,8 +257,6 @@ extern struct mlog_bits mlog_and_bits, mlog_not_bits;
mlog(ML_EXIT, "EXIT\n"); \
} while (0)
#else
-#define mlog_entry(...) do { } while (0)
-#define mlog_entry_void(...) do { } while (0)
#define mlog_exit(...) do { } while (0)
#define mlog_exit_ptr(...) do { } while (0)
#define mlog_exit_void(...) do { } while (0)
diff --git a/fs/ocfs2/dcache.c b/fs/ocfs2/dcache.c
index 895532a..351f3d0 100644
--- a/fs/ocfs2/dcache.c
+++ b/fs/ocfs2/dcache.c
@@ -56,8 +56,8 @@ static int ocfs2_dentry_revalidate(struct dentry *dentry,
int ret = 0; /* if all else fails, just return false */
struct ocfs2_super *osb = OCFS2_SB(dentry->d_sb);
- mlog_entry("(0x%p, '%.*s')\n", dentry,
- dentry->d_name.len, dentry->d_name.name);
+ mlog(0, "(0x%p, '%.*s')\n", dentry,
+ dentry->d_name.len, dentry->d_name.name);
/* For a negative dentry -
* check the generation number of the parent and compare with the
diff --git a/fs/ocfs2/dir.c b/fs/ocfs2/dir.c
index c49f6de..7483e9e 100644
--- a/fs/ocfs2/dir.c
+++ b/fs/ocfs2/dir.c
@@ -367,8 +367,6 @@ static int inline ocfs2_search_dirblock(struct buffer_head *bh,
int de_len;
int ret = 0;
- mlog_entry_void();
-
de_buf = first_de;
dlimit = de_buf + bytes;
@@ -706,8 +704,6 @@ static struct buffer_head *ocfs2_find_entry_el(const char *name, int namelen,
int num = 0;
int nblocks, i, err;
- mlog_entry_void();
-
sb = dir->i_sb;
nblocks = i_size_read(dir) >> sb->s_blocksize_bits;
@@ -1166,7 +1162,7 @@ static int __ocfs2_delete_entry(handle_t *handle, struct inode *dir,
int i, status = -ENOENT;
ocfs2_journal_access_func access = ocfs2_journal_access_db;
- mlog_entry("(0x%p, 0x%p, 0x%p, 0x%p)\n", handle, dir, de_del, bh);
+ mlog(0, "(0x%p, 0x%p, 0x%p, 0x%p)\n", handle, dir, de_del, bh);
if (OCFS2_I(dir)->ip_dyn_features & OCFS2_INLINE_DATA_FL)
access = ocfs2_journal_access_di;
@@ -1632,8 +1628,6 @@ int __ocfs2_add_entry(handle_t *handle,
struct buffer_head *insert_bh = lookup->dl_leaf_bh;
char *data_start = insert_bh->b_data;
- mlog_entry_void();
-
if (!namelen)
return -EINVAL;
@@ -2028,8 +2022,8 @@ int ocfs2_readdir(struct file * filp, void * dirent, filldir_t filldir)
struct inode *inode = filp->f_path.dentry->d_inode;
int lock_level = 0;
- mlog_entry("dirino=%llu\n",
- (unsigned long long)OCFS2_I(inode)->ip_blkno);
+ mlog(0, "dirino=%llu\n",
+ (unsigned long long)OCFS2_I(inode)->ip_blkno);
error = ocfs2_inode_lock_atime(inode, filp->f_vfsmnt, &lock_level);
if (lock_level && error >= 0) {
@@ -2114,8 +2108,8 @@ int ocfs2_check_dir_for_entry(struct inode *dir,
int ret;
struct ocfs2_dir_lookup_result lookup = { NULL, };
- mlog_entry("dir %llu, name '%.*s'\n",
- (unsigned long long)OCFS2_I(dir)->ip_blkno, namelen, name);
+ mlog(0, "dir %llu, name '%.*s'\n",
+ (unsigned long long)OCFS2_I(dir)->ip_blkno, namelen, name);
ret = -EEXIST;
if (ocfs2_find_entry(name, namelen, dir, &lookup) == 0)
@@ -2324,8 +2318,6 @@ static int ocfs2_fill_new_dir_el(struct ocfs2_super *osb,
struct buffer_head *new_bh = NULL;
struct ocfs2_dir_entry *de;
- mlog_entry_void();
-
if (ocfs2_new_dir_wants_trailer(inode))
size = ocfs2_dir_trailer_blk_off(parent->i_sb);
@@ -3268,8 +3260,6 @@ static int ocfs2_extend_dir(struct ocfs2_super *osb,
struct ocfs2_extent_tree et;
struct buffer_head *dx_root_bh = lookup->dl_dx_root_bh;
- mlog_entry_void();
-
if (OCFS2_I(dir)->ip_dyn_features & OCFS2_INLINE_DATA_FL) {
/*
* This would be a code error as an inline directory should
diff --git a/fs/ocfs2/dlm/dlmast.c b/fs/ocfs2/dlm/dlmast.c
index f449991..6526292 100644
--- a/fs/ocfs2/dlm/dlmast.c
+++ b/fs/ocfs2/dlm/dlmast.c
@@ -90,8 +90,6 @@ static int dlm_should_cancel_bast(struct dlm_ctxt *dlm, struct dlm_lock *lock)
void __dlm_queue_ast(struct dlm_ctxt *dlm, struct dlm_lock *lock)
{
- mlog_entry_void();
-
BUG_ON(!dlm);
BUG_ON(!lock);
@@ -134,8 +132,6 @@ void __dlm_queue_ast(struct dlm_ctxt *dlm, struct dlm_lock *lock)
void dlm_queue_ast(struct dlm_ctxt *dlm, struct dlm_lock *lock)
{
- mlog_entry_void();
-
BUG_ON(!dlm);
BUG_ON(!lock);
@@ -147,8 +143,6 @@ void dlm_queue_ast(struct dlm_ctxt *dlm, struct dlm_lock *lock)
void __dlm_queue_bast(struct dlm_ctxt *dlm, struct dlm_lock *lock)
{
- mlog_entry_void();
-
BUG_ON(!dlm);
BUG_ON(!lock);
assert_spin_locked(&dlm->ast_lock);
@@ -167,8 +161,6 @@ void __dlm_queue_bast(struct dlm_ctxt *dlm, struct dlm_lock *lock)
void dlm_queue_bast(struct dlm_ctxt *dlm, struct dlm_lock *lock)
{
- mlog_entry_void();
-
BUG_ON(!dlm);
BUG_ON(!lock);
@@ -213,8 +205,6 @@ void dlm_do_local_ast(struct dlm_ctxt *dlm, struct dlm_lock_resource *res,
dlm_astlockfunc_t *fn;
struct dlm_lockstatus *lksb;
- mlog_entry_void();
-
lksb = lock->lksb;
fn = lock->ast;
BUG_ON(lock->ml.node != dlm->node_num);
@@ -231,8 +221,6 @@ int dlm_do_remote_ast(struct dlm_ctxt *dlm, struct dlm_lock_resource *res,
struct dlm_lockstatus *lksb;
int lksbflags;
- mlog_entry_void();
-
lksb = lock->lksb;
BUG_ON(lock->ml.node == dlm->node_num);
@@ -250,7 +238,6 @@ void dlm_do_local_bast(struct dlm_ctxt *dlm, struct dlm_lock_resource *res,
{
dlm_bastlockfunc_t *fn = lock->bast;
- mlog_entry_void();
BUG_ON(lock->ml.node != dlm->node_num);
(*fn)(lock->astdata, blocked_type);
@@ -426,9 +413,9 @@ int dlm_send_proxy_ast_msg(struct dlm_ctxt *dlm, struct dlm_lock_resource *res,
size_t veclen = 1;
int status;
- mlog_entry("res %.*s, to=%u, type=%d, blocked_type=%d\n",
- res->lockname.len, res->lockname.name, lock->ml.node,
- msg_type, blocked_type);
+ mlog(0, "res %.*s, to=%u, type=%d, blocked_type=%d\n",
+ res->lockname.len, res->lockname.name, lock->ml.node,
+ msg_type, blocked_type);
memset(&past, 0, sizeof(struct dlm_proxy_ast));
past.node_idx = dlm->node_num;
diff --git a/fs/ocfs2/dlm/dlmconvert.c b/fs/ocfs2/dlm/dlmconvert.c
index 9f30491..29a886d 100644
--- a/fs/ocfs2/dlm/dlmconvert.c
+++ b/fs/ocfs2/dlm/dlmconvert.c
@@ -128,8 +128,8 @@ static enum dlm_status __dlmconvert_master(struct dlm_ctxt *dlm,
assert_spin_locked(&res->spinlock);
- mlog_entry("type=%d, convert_type=%d, new convert_type=%d\n",
- lock->ml.type, lock->ml.convert_type, type);
+ mlog(0, "type=%d, convert_type=%d, new convert_type=%d\n",
+ lock->ml.type, lock->ml.convert_type, type);
spin_lock(&lock->spinlock);
@@ -353,7 +353,7 @@ static enum dlm_status dlm_send_remote_convert_request(struct dlm_ctxt *dlm,
struct kvec vec[2];
size_t veclen = 1;
- mlog_entry("%.*s\n", res->lockname.len, res->lockname.name);
+ mlog(0, "%.*s\n", res->lockname.len, res->lockname.name);
memset(&convert, 0, sizeof(struct dlm_convert_lock));
convert.node_idx = dlm->node_num;
diff --git a/fs/ocfs2/dlm/dlmdomain.c b/fs/ocfs2/dlm/dlmdomain.c
index cc2aaa9..eddbc01 100644
--- a/fs/ocfs2/dlm/dlmdomain.c
+++ b/fs/ocfs2/dlm/dlmdomain.c
@@ -188,7 +188,7 @@ struct dlm_lock_resource * __dlm_lookup_lockres_full(struct dlm_ctxt *dlm,
struct hlist_head *bucket;
struct hlist_node *list;
- mlog_entry("%.*s\n", len, name);
+ mlog(0, "%.*s\n", len, name);
assert_spin_locked(&dlm->spinlock);
@@ -222,7 +222,7 @@ struct dlm_lock_resource * __dlm_lookup_lockres(struct dlm_ctxt *dlm,
{
struct dlm_lock_resource *res = NULL;
- mlog_entry("%.*s\n", len, name);
+ mlog(0, "%.*s\n", len, name);
assert_spin_locked(&dlm->spinlock);
@@ -533,7 +533,7 @@ static int dlm_exit_domain_handler(struct o2net_msg *msg, u32 len, void *data,
unsigned int node;
struct dlm_exit_domain *exit_msg = (struct dlm_exit_domain *) msg->buf;
- mlog_entry("%p %u %p", msg, len, data);
+ mlog(0, "%p %u %p", msg, len, data);
if (!dlm_grab(dlm))
return 0;
@@ -1555,7 +1555,7 @@ static int dlm_try_to_join_domain(struct dlm_ctxt *dlm)
struct domain_join_ctxt *ctxt;
enum dlm_query_join_response_code response = JOIN_DISALLOW;
- mlog_entry("%p", dlm);
+ mlog(0, "%p", dlm);
ctxt = kzalloc(sizeof(*ctxt), GFP_KERNEL);
if (!ctxt) {
diff --git a/fs/ocfs2/dlm/dlmlock.c b/fs/ocfs2/dlm/dlmlock.c
index 69cf369..a0cc7e7 100644
--- a/fs/ocfs2/dlm/dlmlock.c
+++ b/fs/ocfs2/dlm/dlmlock.c
@@ -125,7 +125,7 @@ static enum dlm_status dlmlock_master(struct dlm_ctxt *dlm,
int call_ast = 0, kick_thread = 0;
enum dlm_status status = DLM_NORMAL;
- mlog_entry("type=%d\n", lock->ml.type);
+ mlog(0, "type=%d\n", lock->ml.type);
spin_lock(&res->spinlock);
/* if called from dlm_create_lock_handler, need to
@@ -224,8 +224,8 @@ static enum dlm_status dlmlock_remote(struct dlm_ctxt *dlm,
enum dlm_status status = DLM_DENIED;
int lockres_changed = 1;
- mlog_entry("type=%d\n", lock->ml.type);
- mlog(0, "lockres %.*s, flags = 0x%x\n", res->lockname.len,
+ mlog(0, "type=%d, lockres %.*s, flags = 0x%x\n",
+ lock->ml.type, res->lockname.len,
res->lockname.name, flags);
spin_lock(&res->spinlock);
@@ -305,8 +305,6 @@ static enum dlm_status dlm_send_remote_lock_request(struct dlm_ctxt *dlm,
int tmpret, status = 0;
enum dlm_status ret;
- mlog_entry_void();
-
memset(&create, 0, sizeof(create));
create.node_idx = dlm->node_num;
create.requested_type = lock->ml.type;
@@ -474,8 +472,6 @@ int dlm_create_lock_handler(struct o2net_msg *msg, u32 len, void *data,
BUG_ON(!dlm);
- mlog_entry_void();
-
if (!dlm_grab(dlm))
return DLM_REJECTED;
diff --git a/fs/ocfs2/dlm/dlmmaster.c b/fs/ocfs2/dlm/dlmmaster.c
index f564b0e..dd2e8f0 100644
--- a/fs/ocfs2/dlm/dlmmaster.c
+++ b/fs/ocfs2/dlm/dlmmaster.c
@@ -426,8 +426,6 @@ static void dlm_mle_release(struct kref *kref)
struct dlm_master_list_entry *mle;
struct dlm_ctxt *dlm;
- mlog_entry_void();
-
mle = container_of(kref, struct dlm_master_list_entry, mle_refs);
dlm = mle->dlm;
@@ -3106,8 +3104,6 @@ static int dlm_add_migration_mle(struct dlm_ctxt *dlm,
*oldmle = NULL;
- mlog_entry_void();
-
assert_spin_locked(&dlm->spinlock);
assert_spin_locked(&dlm->master_lock);
@@ -3247,7 +3243,7 @@ void dlm_clean_master_list(struct dlm_ctxt *dlm, u8 dead_node)
struct hlist_node *list;
unsigned int i;
- mlog_entry("dlm=%s, dead node=%u\n", dlm->name, dead_node);
+ mlog(0, "dlm=%s, dead node=%u\n", dlm->name, dead_node);
top:
assert_spin_locked(&dlm->spinlock);
diff --git a/fs/ocfs2/dlm/dlmrecovery.c b/fs/ocfs2/dlm/dlmrecovery.c
index aaaffbc..6602b23 100644
--- a/fs/ocfs2/dlm/dlmrecovery.c
+++ b/fs/ocfs2/dlm/dlmrecovery.c
@@ -2083,8 +2083,6 @@ static void dlm_finish_local_lockres_recovery(struct dlm_ctxt *dlm,
struct hlist_head *bucket;
struct dlm_lock_resource *res, *next;
- mlog_entry_void();
-
assert_spin_locked(&dlm->spinlock);
list_for_each_entry_safe(res, next, &dlm->reco.resources, recovering) {
@@ -2607,8 +2605,6 @@ static int dlm_send_begin_reco_message(struct dlm_ctxt *dlm, u8 dead_node)
int nodenum;
int status;
- mlog_entry("%u\n", dead_node);
-
mlog(0, "%s: dead node is %u\n", dlm->name, dead_node);
spin_lock(&dlm->spinlock);
diff --git a/fs/ocfs2/dlm/dlmthread.c b/fs/ocfs2/dlm/dlmthread.c
index 2211acf..3081362 100644
--- a/fs/ocfs2/dlm/dlmthread.c
+++ b/fs/ocfs2/dlm/dlmthread.c
@@ -122,7 +122,7 @@ int __dlm_lockres_unused(struct dlm_lock_resource *res)
void __dlm_lockres_calc_usage(struct dlm_ctxt *dlm,
struct dlm_lock_resource *res)
{
- mlog_entry("%.*s\n", res->lockname.len, res->lockname.name);
+ mlog(0, "%.*s\n", res->lockname.len, res->lockname.name);
assert_spin_locked(&dlm->spinlock);
assert_spin_locked(&res->spinlock);
@@ -150,7 +150,7 @@ void __dlm_lockres_calc_usage(struct dlm_ctxt *dlm,
void dlm_lockres_calc_usage(struct dlm_ctxt *dlm,
struct dlm_lock_resource *res)
{
- mlog_entry("%.*s\n", res->lockname.len, res->lockname.name);
+ mlog(0, "%.*s\n", res->lockname.len, res->lockname.name);
spin_lock(&dlm->spinlock);
spin_lock(&res->spinlock);
@@ -453,7 +453,7 @@ leave:
/* must have NO locks when calling this with res !=NULL * */
void dlm_kick_thread(struct dlm_ctxt *dlm, struct dlm_lock_resource *res)
{
- mlog_entry("dlm=%p, res=%p\n", dlm, res);
+ mlog(0, "dlm=%p, res=%p\n", dlm, res);
if (res) {
spin_lock(&dlm->spinlock);
spin_lock(&res->spinlock);
@@ -466,7 +466,7 @@ void dlm_kick_thread(struct dlm_ctxt *dlm, struct dlm_lock_resource *res)
void __dlm_dirty_lockres(struct dlm_ctxt *dlm, struct dlm_lock_resource *res)
{
- mlog_entry("dlm=%p, res=%p\n", dlm, res);
+ mlog(0, "dlm=%p, res=%p\n", dlm, res);
assert_spin_locked(&dlm->spinlock);
assert_spin_locked(&res->spinlock);
diff --git a/fs/ocfs2/dlm/dlmunlock.c b/fs/ocfs2/dlm/dlmunlock.c
index 817287c..850aa7e 100644
--- a/fs/ocfs2/dlm/dlmunlock.c
+++ b/fs/ocfs2/dlm/dlmunlock.c
@@ -317,7 +317,7 @@ static enum dlm_status dlm_send_remote_unlock_request(struct dlm_ctxt *dlm,
struct kvec vec[2];
size_t veclen = 1;
- mlog_entry("%.*s\n", res->lockname.len, res->lockname.name);
+ mlog(0, "%.*s\n", res->lockname.len, res->lockname.name);
if (owner == dlm->node_num) {
/* ended up trying to contact ourself. this means
@@ -588,8 +588,6 @@ enum dlm_status dlmunlock(struct dlm_ctxt *dlm, struct dlm_lockstatus *lksb,
struct dlm_lock *lock = NULL;
int call_ast, is_master;
- mlog_entry_void();
-
if (!lksb) {
dlm_error(DLM_BADARGS);
return DLM_BADARGS;
diff --git a/fs/ocfs2/dlmglue.c b/fs/ocfs2/dlmglue.c
index e8d94d7..c2f3fa1 100644
--- a/fs/ocfs2/dlmglue.c
+++ b/fs/ocfs2/dlmglue.c
@@ -397,8 +397,6 @@ static void ocfs2_build_lock_name(enum ocfs2_lock_type type,
{
int len;
- mlog_entry_void();
-
BUG_ON(type >= OCFS2_NUM_LOCK_TYPES);
len = snprintf(name, OCFS2_LOCK_ID_MAX_LEN, "%c%s%016llx%08x",
@@ -729,8 +727,6 @@ void ocfs2_refcount_lock_res_init(struct ocfs2_lock_res *lockres,
void ocfs2_lock_res_free(struct ocfs2_lock_res *res)
{
- mlog_entry_void();
-
if (!(res->l_flags & OCFS2_LOCK_INITIALIZED))
return;
@@ -762,8 +758,6 @@ void ocfs2_lock_res_free(struct ocfs2_lock_res *res)
static inline void ocfs2_inc_holders(struct ocfs2_lock_res *lockres,
int level)
{
- mlog_entry_void();
-
BUG_ON(!lockres);
switch(level) {
@@ -783,8 +777,6 @@ static inline void ocfs2_inc_holders(struct ocfs2_lock_res *lockres,
static inline void ocfs2_dec_holders(struct ocfs2_lock_res *lockres,
int level)
{
- mlog_entry_void();
-
BUG_ON(!lockres);
switch(level) {
@@ -846,8 +838,6 @@ static void lockres_clear_flags(struct ocfs2_lock_res *lockres,
static inline void ocfs2_generic_handle_downconvert_action(struct ocfs2_lock_res *lockres)
{
- mlog_entry_void();
-
BUG_ON(!(lockres->l_flags & OCFS2_LOCK_BUSY));
BUG_ON(!(lockres->l_flags & OCFS2_LOCK_ATTACHED));
BUG_ON(!(lockres->l_flags & OCFS2_LOCK_BLOCKED));
@@ -866,8 +856,6 @@ static inline void ocfs2_generic_handle_downconvert_action(struct ocfs2_lock_res
static inline void ocfs2_generic_handle_convert_action(struct ocfs2_lock_res *lockres)
{
- mlog_entry_void();
-
BUG_ON(!(lockres->l_flags & OCFS2_LOCK_BUSY));
BUG_ON(!(lockres->l_flags & OCFS2_LOCK_ATTACHED));
@@ -895,8 +883,6 @@ static inline void ocfs2_generic_handle_convert_action(struct ocfs2_lock_res *lo
static inline void ocfs2_generic_handle_attach_action(struct ocfs2_lock_res *lockres)
{
- mlog_entry_void();
-
BUG_ON((!(lockres->l_flags & OCFS2_LOCK_BUSY)));
BUG_ON(lockres->l_flags & OCFS2_LOCK_ATTACHED);
@@ -916,7 +902,6 @@ static int ocfs2_generic_handle_bast(struct ocfs2_lock_res *lockres,
int level)
{
int needs_downconvert = 0;
- mlog_entry_void();
assert_spin_locked(&lockres->l_lock);
@@ -1151,8 +1136,6 @@ static void ocfs2_unlock_ast(struct ocfs2_dlm_lksb *lksb, int error)
struct ocfs2_lock_res *lockres = ocfs2_lksb_to_lock_res(lksb);
unsigned long flags;
- mlog_entry_void();
-
mlog(ML_BASTS, "UNLOCK AST fired for lockres %s, action = %d\n",
lockres->l_name, lockres->l_unlock_action);
@@ -1233,7 +1216,6 @@ static inline void ocfs2_recover_from_dlm_error(struct ocfs2_lock_res *lockres,
{
unsigned long flags;
- mlog_entry_void();
spin_lock_irqsave(&lockres->l_lock, flags);
lockres_clear_flags(lockres, OCFS2_LOCK_BUSY);
lockres_clear_flags(lockres, OCFS2_LOCK_UPCONVERT_FINISHING);
@@ -1260,8 +1242,6 @@ static int ocfs2_lock_create(struct ocfs2_super *osb,
unsigned long flags;
unsigned int gen;
- mlog_entry_void();
-
mlog(0, "lock %s, level = %d, flags = %u\n", lockres->l_name, level,
dlm_flags);
@@ -1416,8 +1396,6 @@ static int __ocfs2_cluster_lock(struct ocfs2_super *osb,
unsigned int gen;
int noqueue_attempted = 0;
- mlog_entry_void();
-
ocfs2_init_mask_waiter(&mw);
if (lockres->l_ops->flags & LOCK_TYPE_USES_LVB)
@@ -1605,7 +1583,6 @@ static void __ocfs2_cluster_unlock(struct ocfs2_super *osb,
{
unsigned long flags;
- mlog_entry_void();
spin_lock_irqsave(&lockres->l_lock, flags);
ocfs2_dec_holders(lockres, level);
ocfs2_downconvert_on_unlock(osb, lockres);
@@ -1648,8 +1625,6 @@ int ocfs2_create_new_inode_locks(struct inode *inode)
BUG_ON(!inode);
BUG_ON(!ocfs2_inode_is_new(inode));
- mlog_entry_void();
-
mlog(0, "Inode %llu\n", (unsigned long long)OCFS2_I(inode)->ip_blkno);
/* NOTE: That we don't increment any of the holder counts, nor
@@ -1695,8 +1670,6 @@ int ocfs2_rw_lock(struct inode *inode, int write)
BUG_ON(!inode);
- mlog_entry_void();
-
mlog(0, "inode %llu take %s RW lock\n",
(unsigned long long)OCFS2_I(inode)->ip_blkno,
write ? "EXMODE" : "PRMODE");
@@ -1725,8 +1698,6 @@ void ocfs2_rw_unlock(struct inode *inode, int write)
struct ocfs2_lock_res *lockres = &OCFS2_I(inode)->ip_rw_lockres;
struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
- mlog_entry_void();
-
mlog(0, "inode %llu drop %s RW lock\n",
(unsigned long long)OCFS2_I(inode)->ip_blkno,
write ? "EXMODE" : "PRMODE");
@@ -1748,8 +1719,6 @@ int ocfs2_open_lock(struct inode *inode)
BUG_ON(!inode);
- mlog_entry_void();
-
mlog(0, "inode %llu take PRMODE open lock\n",
(unsigned long long)OCFS2_I(inode)->ip_blkno);
@@ -1776,8 +1745,6 @@ int ocfs2_try_open_lock(struct inode *inode, int write)
BUG_ON(!inode);
- mlog_entry_void();
-
mlog(0, "inode %llu try to take %s open lock\n",
(unsigned long long)OCFS2_I(inode)->ip_blkno,
write ? "EXMODE" : "PRMODE");
@@ -1811,8 +1778,6 @@ void ocfs2_open_unlock(struct inode *inode)
struct ocfs2_lock_res *lockres = &OCFS2_I(inode)->ip_open_lockres;
struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
- mlog_entry_void();
-
mlog(0, "inode %llu drop open lock\n",
(unsigned long long)OCFS2_I(inode)->ip_blkno);
@@ -2043,8 +2008,6 @@ static void ocfs2_downconvert_on_unlock(struct ocfs2_super *osb,
{
int kick = 0;
- mlog_entry_void();
-
/* If we know that another node is waiting on our lock, kick
* the downconvert thread * pre-emptively when we reach a release
* condition. */
@@ -2095,8 +2058,6 @@ static void __ocfs2_stuff_meta_lvb(struct inode *inode)
struct ocfs2_lock_res *lockres = &oi->ip_inode_lockres;
struct ocfs2_meta_lvb *lvb;
- mlog_entry_void();
-
lvb = ocfs2_dlm_lvb(&lockres->l_lksb);
/*
@@ -2145,8 +2106,6 @@ static void ocfs2_refresh_inode_from_lvb(struct inode *inode)
struct ocfs2_lock_res *lockres = &oi->ip_inode_lockres;
struct ocfs2_meta_lvb *lvb;
- mlog_entry_void();
-
mlog_meta_lvb(0, lockres);
lvb = ocfs2_dlm_lvb(&lockres->l_lksb);
@@ -2205,8 +2164,6 @@ static int ocfs2_should_refresh_lock_res(struct ocfs2_lock_res *lockres)
unsigned long flags;
int status = 0;
- mlog_entry_void();
-
refresh_check:
spin_lock_irqsave(&lockres->l_lock, flags);
if (!(lockres->l_flags & OCFS2_LOCK_NEEDS_REFRESH)) {
@@ -2237,7 +2194,6 @@ static inline void ocfs2_complete_lock_res_refresh(struct ocfs2_lock_res *lockre
int status)
{
unsigned long flags;
- mlog_entry_void();
spin_lock_irqsave(&lockres->l_lock, flags);
lockres_clear_flags(lockres, OCFS2_LOCK_REFRESHING);
@@ -2260,8 +2216,6 @@ static int ocfs2_inode_lock_update(struct inode *inode,
struct ocfs2_dinode *fe;
struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
- mlog_entry_void();
-
if (ocfs2_mount_local(osb))
goto bail;
@@ -2374,8 +2328,6 @@ int ocfs2_inode_lock_full_nested(struct inode *inode,
BUG_ON(!inode);
- mlog_entry_void();
-
mlog(0, "inode %llu, take %s META lock\n",
(unsigned long long)OCFS2_I(inode)->ip_blkno,
ex ? "EXMODE" : "PRMODE");
@@ -2517,7 +2469,6 @@ int ocfs2_inode_lock_atime(struct inode *inode,
{
int ret;
- mlog_entry_void();
ret = ocfs2_inode_lock(inode, NULL, 0);
if (ret < 0) {
mlog_errno(ret);
@@ -2556,8 +2507,6 @@ void ocfs2_inode_unlock(struct inode *inode,
struct ocfs2_lock_res *lockres = &OCFS2_I(inode)->ip_inode_lockres;
struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
- mlog_entry_void();
-
mlog(0, "inode %llu drop %s META lock\n",
(unsigned long long)OCFS2_I(inode)->ip_blkno,
ex ? "EXMODE" : "PRMODE");
@@ -2617,8 +2566,6 @@ int ocfs2_super_lock(struct ocfs2_super *osb,
int level = ex ? DLM_LOCK_EX : DLM_LOCK_PR;
struct ocfs2_lock_res *lockres = &osb->osb_super_lockres;
- mlog_entry_void();
-
if (ocfs2_is_hard_readonly(osb))
return -EROFS;
@@ -3054,8 +3001,6 @@ int ocfs2_dlm_init(struct ocfs2_super *osb)
int status = 0;
struct ocfs2_cluster_connection *conn = NULL;
- mlog_entry_void();
-
if (ocfs2_mount_local(osb)) {
osb->node_num = 0;
goto local;
@@ -3119,8 +3064,6 @@ bail:
void ocfs2_dlm_shutdown(struct ocfs2_super *osb,
int hangup_pending)
{
- mlog_entry_void();
-
ocfs2_drop_osb_locks(osb);
/*
@@ -3284,8 +3227,6 @@ int ocfs2_drop_inode_locks(struct inode *inode)
{
int status, err;
- mlog_entry_void();
-
/* No need to call ocfs2_mark_lockres_freeing here -
* ocfs2_clear_inode has done it for us. */
@@ -3352,8 +3293,6 @@ static int ocfs2_downconvert_lock(struct ocfs2_super *osb,
int ret;
u32 dlm_flags = DLM_LKF_CONVERT;
- mlog_entry_void();
-
mlog(ML_BASTS, "lockres %s, level %d => %d\n", lockres->l_name,
lockres->l_level, new_level);
@@ -3385,8 +3324,6 @@ static int ocfs2_prepare_cancel_convert(struct ocfs2_super *osb,
{
assert_spin_locked(&lockres->l_lock);
- mlog_entry_void();
-
if (lockres->l_unlock_action == OCFS2_UNLOCK_CANCEL_CONVERT) {
/* If we're already trying to cancel a lock conversion
* then just drop the spinlock and allow the caller to
@@ -3416,8 +3353,6 @@ static int ocfs2_cancel_convert(struct ocfs2_super *osb,
{
int ret;
- mlog_entry_void();
-
ret = ocfs2_dlm_unlock(osb->cconn, &lockres->l_lksb,
DLM_LKF_CANCEL);
if (ret) {
@@ -3443,8 +3378,6 @@ static int ocfs2_unblock_lock(struct ocfs2_super *osb,
int set_lvb = 0;
unsigned int gen;
- mlog_entry_void();
-
spin_lock_irqsave(&lockres->l_lock, flags);
recheck:
@@ -3859,8 +3792,6 @@ static void ocfs2_set_qinfo_lvb(struct ocfs2_lock_res *lockres)
struct mem_dqinfo *info = sb_dqinfo(oinfo->dqi_gi.dqi_sb,
oinfo->dqi_gi.dqi_type);
- mlog_entry_void();
-
lvb = ocfs2_dlm_lvb(&lockres->l_lksb);
lvb->lvb_version = OCFS2_QINFO_LVB_VERSION;
lvb->lvb_bgrace = cpu_to_be32(info->dqi_bgrace);
@@ -3879,7 +3810,6 @@ void ocfs2_qinfo_unlock(struct ocfs2_mem_dqinfo *oinfo, int ex)
struct ocfs2_super *osb = OCFS2_SB(oinfo->dqi_gi.dqi_sb);
int level = ex ? DLM_LOCK_EX : DLM_LOCK_PR;
- mlog_entry_void();
if (!ocfs2_is_hard_readonly(osb) && !ocfs2_mount_local(osb))
ocfs2_cluster_unlock(osb, lockres, level);
mlog_exit_void();
@@ -3937,8 +3867,6 @@ int ocfs2_qinfo_lock(struct ocfs2_mem_dqinfo *oinfo, int ex)
int level = ex ? DLM_LOCK_EX : DLM_LOCK_PR;
int status = 0;
- mlog_entry_void();
-
/* On RO devices, locking really isn't needed... */
if (ocfs2_is_hard_readonly(osb)) {
if (ex)
@@ -4007,8 +3935,6 @@ static void ocfs2_process_blocked_lock(struct ocfs2_super *osb,
* considered valid until we remove the OCFS2_LOCK_QUEUED
* flag. */
- mlog_entry_void();
-
BUG_ON(!lockres);
BUG_ON(!lockres->l_ops);
@@ -4049,8 +3975,6 @@ unqueue:
static void ocfs2_schedule_blocked_lock(struct ocfs2_super *osb,
struct ocfs2_lock_res *lockres)
{
- mlog_entry_void();
-
assert_spin_locked(&lockres->l_lock);
if (lockres->l_flags & OCFS2_LOCK_FREEING) {
@@ -4080,8 +4004,6 @@ static void ocfs2_downconvert_thread_do_work(struct ocfs2_super *osb)
unsigned long processed;
struct ocfs2_lock_res *lockres;
- mlog_entry_void();
-
spin_lock(&osb->dc_task_lock);
/* grab this early so we know to try again if a state change and
* wake happens part-way through our work */
diff --git a/fs/ocfs2/export.c b/fs/ocfs2/export.c
index 19ad145..2173cc2 100644
--- a/fs/ocfs2/export.c
+++ b/fs/ocfs2/export.c
@@ -56,7 +56,7 @@ static struct dentry *ocfs2_get_dentry(struct super_block *sb,
int status, set;
struct dentry *result;
- mlog_entry("(0x%p, 0x%p)\n", sb, handle);
+ mlog(0, "(0x%p, 0x%p)\n", sb, handle);
if (blkno == 0) {
mlog(0, "nfs wants inode with blkno: 0\n");
@@ -154,8 +154,8 @@ static struct dentry *ocfs2_get_parent(struct dentry *child)
struct dentry *parent;
struct inode *dir = child->d_inode;
- mlog_entry("(0x%p, '%.*s')\n", child,
- child->d_name.len, child->d_name.name);
+ mlog(0, "(0x%p, '%.*s')\n", child,
+ child->d_name.len, child->d_name.name);
mlog(0, "find parent of directory %llu\n",
(unsigned long long)OCFS2_I(dir)->ip_blkno);
@@ -197,9 +197,9 @@ static int ocfs2_encode_fh(struct dentry *dentry, u32 *fh_in, int *max_len,
u32 generation;
__le32 *fh = (__force __le32 *) fh_in;
- mlog_entry("(0x%p, '%.*s', 0x%p, %d, %d)\n", dentry,
- dentry->d_name.len, dentry->d_name.name,
- fh, len, connectable);
+ mlog(0, "(0x%p, '%.*s', 0x%p, %d, %d)\n", dentry,
+ dentry->d_name.len, dentry->d_name.name,
+ fh, len, connectable);
if (len < 3 || (connectable && len < 6)) {
mlog(ML_ERROR, "fh buffer is too small for encoding\n");
diff --git a/fs/ocfs2/extent_map.c b/fs/ocfs2/extent_map.c
index 09e3fdf..ed9916a 100644
--- a/fs/ocfs2/extent_map.c
+++ b/fs/ocfs2/extent_map.c
@@ -841,10 +841,10 @@ int ocfs2_read_virt_blocks(struct inode *inode, u64 v_block, int nr,
u64 p_block, p_count;
int i, count, done = 0;
- mlog_entry("(inode = %p, v_block = %llu, nr = %d, bhs = %p, "
- "flags = %x, validate = %p)\n",
- inode, (unsigned long long)v_block, nr, bhs, flags,
- validate);
+ mlog(0, "(inode = %p, v_block = %llu, nr = %d, bhs = %p, "
+ "flags = %x, validate = %p)\n",
+ inode, (unsigned long long)v_block, nr, bhs, flags,
+ validate);
if (((v_block + nr - 1) << inode->i_sb->s_blocksize_bits) >=
i_size_read(inode)) {
diff --git a/fs/ocfs2/file.c b/fs/ocfs2/file.c
index 77b4c04..480accf 100644
--- a/fs/ocfs2/file.c
+++ b/fs/ocfs2/file.c
@@ -99,8 +99,8 @@ static int ocfs2_file_open(struct inode *inode, struct file *file)
int mode = file->f_flags;
struct ocfs2_inode_info *oi = OCFS2_I(inode);
- mlog_entry("(0x%p, 0x%p, '%.*s')\n", inode, file,
- file->f_path.dentry->d_name.len, file->f_path.dentry->d_name.name);
+ mlog(0, "(0x%p, 0x%p, '%.*s')\n", inode, file,
+ file->f_path.dentry->d_name.len, file->f_path.dentry->d_name.name);
if (file->f_mode & FMODE_WRITE)
dquot_initialize(inode);
@@ -143,9 +143,9 @@ static int ocfs2_file_release(struct inode *inode, struct file *file)
{
struct ocfs2_inode_info *oi = OCFS2_I(inode);
- mlog_entry("(0x%p, 0x%p, '%.*s')\n", inode, file,
- file->f_path.dentry->d_name.len,
- file->f_path.dentry->d_name.name);
+ mlog(0, "(0x%p, 0x%p, '%.*s')\n", inode, file,
+ file->f_path.dentry->d_name.len,
+ file->f_path.dentry->d_name.name);
spin_lock(&oi->ip_lock);
if (!--oi->ip_open_count)
@@ -177,9 +177,9 @@ static int ocfs2_sync_file(struct file *file, int datasync)
struct inode *inode = file->f_mapping->host;
struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
- mlog_entry("(0x%p, %d, 0x%p, '%.*s')\n", file, datasync,
- file->f_path.dentry, file->f_path.dentry->d_name.len,
- file->f_path.dentry->d_name.name);
+ mlog(0, "(0x%p, %d, 0x%p, '%.*s')\n", file, datasync,
+ file->f_path.dentry, file->f_path.dentry->d_name.len,
+ file->f_path.dentry->d_name.name);
if (datasync && !(inode->i_state & I_DIRTY_DATASYNC)) {
/*
@@ -251,8 +251,6 @@ int ocfs2_update_inode_atime(struct inode *inode,
handle_t *handle;
struct ocfs2_dinode *di = (struct ocfs2_dinode *) bh->b_data;
- mlog_entry_void();
-
handle = ocfs2_start_trans(osb, OCFS2_INODE_UPDATE_CREDITS);
if (IS_ERR(handle)) {
ret = PTR_ERR(handle);
@@ -291,7 +289,6 @@ static int ocfs2_set_inode_size(handle_t *handle,
{
int status;
- mlog_entry_void();
i_size_write(inode, new_i_size);
inode->i_blocks = ocfs2_inode_sector_count(inode);
inode->i_ctime = inode->i_mtime = CURRENT_TIME;
@@ -375,8 +372,6 @@ static int ocfs2_orphan_for_truncate(struct ocfs2_super *osb,
struct ocfs2_dinode *di;
u64 cluster_bytes;
- mlog_entry_void();
-
/*
* We need to CoW the cluster contains the offset if it is reflinked
* since we will call ocfs2_zero_range_for_truncate later which will
@@ -442,9 +437,9 @@ static int ocfs2_truncate_file(struct inode *inode,
struct ocfs2_dinode *fe = NULL;
struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
- mlog_entry("(inode = %llu, new_i_size = %llu\n",
- (unsigned long long)OCFS2_I(inode)->ip_blkno,
- (unsigned long long)new_i_size);
+ mlog(0, "(inode = %llu, new_i_size = %llu\n",
+ (unsigned long long)OCFS2_I(inode)->ip_blkno,
+ (unsigned long long)new_i_size);
/* We trust di_bh because it comes from ocfs2_inode_lock(), which
* already validated it */
@@ -578,7 +573,7 @@ static int __ocfs2_extend_allocation(struct inode *inode, u32 logical_start,
struct ocfs2_extent_tree et;
int did_quota = 0;
- mlog_entry("(clusters_to_add = %u)\n", clusters_to_add);
+ mlog(0, "(clusters_to_add = %u)\n", clusters_to_add);
/*
* This function only exists for file systems which don't
@@ -1113,8 +1108,8 @@ int ocfs2_setattr(struct dentry *dentry, struct iattr *attr)
struct dquot *transfer_to[MAXQUOTAS] = { };
int qtype;
- mlog_entry("(0x%p, '%.*s')\n", dentry,
- dentry->d_name.len, dentry->d_name.name);
+ mlog(0, "(0x%p, '%.*s')\n", dentry,
+ dentry->d_name.len, dentry->d_name.name);
/* ensuring we don't even attempt to truncate a symlink */
if (S_ISLNK(inode->i_mode))
@@ -1287,8 +1282,6 @@ int ocfs2_getattr(struct vfsmount *mnt,
struct ocfs2_super *osb = sb->s_fs_info;
int err;
- mlog_entry_void();
-
err = ocfs2_inode_revalidate(dentry);
if (err) {
if (err != -ENOENT)
@@ -1311,8 +1304,6 @@ int ocfs2_permission(struct inode *inode, int mask)
{
int ret;
- mlog_entry_void();
-
ret = ocfs2_inode_lock(inode, NULL, 0);
if (ret) {
if (ret != -ENOENT)
@@ -1336,8 +1327,8 @@ static int __ocfs2_write_remove_suid(struct inode *inode,
struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
struct ocfs2_dinode *di;
- mlog_entry("(Inode %llu, mode 0%o)\n",
- (unsigned long long)OCFS2_I(inode)->ip_blkno, inode->i_mode);
+ mlog(0, "(Inode %llu, mode 0%o)\n",
+ (unsigned long long)OCFS2_I(inode)->ip_blkno, inode->i_mode);
handle = ocfs2_start_trans(osb, OCFS2_INODE_UPDATE_CREDITS);
if (IS_ERR(handle)) {
@@ -2226,10 +2217,10 @@ static ssize_t ocfs2_file_aio_write(struct kiocb *iocb,
int full_coherency = !(osb->s_mount_opt &
OCFS2_MOUNT_COHERENCY_BUFFERED);
- mlog_entry("(0x%p, %u, '%.*s')\n", file,
- (unsigned int)nr_segs,
- file->f_path.dentry->d_name.len,
- file->f_path.dentry->d_name.name);
+ mlog(0, "(0x%p, %u, '%.*s')\n", file,
+ (unsigned int)nr_segs,
+ file->f_path.dentry->d_name.len,
+ file->f_path.dentry->d_name.name);
if (iocb->ki_left == 0)
return 0;
@@ -2425,10 +2416,10 @@ static ssize_t ocfs2_file_splice_write(struct pipe_inode_info *pipe,
.u.file = out,
};
- mlog_entry("(0x%p, 0x%p, %u, '%.*s')\n", out, pipe,
- (unsigned int)len,
- out->f_path.dentry->d_name.len,
- out->f_path.dentry->d_name.name);
+ mlog(0, "(0x%p, 0x%p, %u, '%.*s')\n", out, pipe,
+ (unsigned int)len,
+ out->f_path.dentry->d_name.len,
+ out->f_path.dentry->d_name.name);
if (pipe->inode)
mutex_lock_nested(&pipe->inode->i_mutex, I_MUTEX_PARENT);
@@ -2485,10 +2476,10 @@ static ssize_t ocfs2_file_splice_read(struct file *in,
int ret = 0, lock_level = 0;
struct inode *inode = in->f_path.dentry->d_inode;
- mlog_entry("(0x%p, 0x%p, %u, '%.*s')\n", in, pipe,
- (unsigned int)len,
- in->f_path.dentry->d_name.len,
- in->f_path.dentry->d_name.name);
+ mlog(0, "(0x%p, 0x%p, %u, '%.*s')\n", in, pipe,
+ (unsigned int)len,
+ in->f_path.dentry->d_name.len,
+ in->f_path.dentry->d_name.name);
/*
* See the comment in ocfs2_file_aio_read()
@@ -2516,10 +2507,10 @@ static ssize_t ocfs2_file_aio_read(struct kiocb *iocb,
struct file *filp = iocb->ki_filp;
struct inode *inode = filp->f_path.dentry->d_inode;
- mlog_entry("(0x%p, %u, '%.*s')\n", filp,
- (unsigned int)nr_segs,
- filp->f_path.dentry->d_name.len,
- filp->f_path.dentry->d_name.name);
+ mlog(0, "(0x%p, %u, '%.*s')\n", filp,
+ (unsigned int)nr_segs,
+ filp->f_path.dentry->d_name.len,
+ filp->f_path.dentry->d_name.name);
if (!inode) {
ret = -EINVAL;
diff --git a/fs/ocfs2/inode.c b/fs/ocfs2/inode.c
index f935fd6..7bdfcb6 100644
--- a/fs/ocfs2/inode.c
+++ b/fs/ocfs2/inode.c
@@ -131,7 +131,7 @@ struct inode *ocfs2_iget(struct ocfs2_super *osb, u64 blkno, unsigned flags,
struct super_block *sb = osb->sb;
struct ocfs2_find_inode_args args;
- mlog_entry("(blkno = %llu)\n", (unsigned long long)blkno);
+ mlog(0, "(blkno = %llu)\n", (unsigned long long)blkno);
/* Ok. By now we've either got the offsets passed to us by the
* caller, or we just pulled them off the bh. Lets do some
@@ -192,7 +192,7 @@ static int ocfs2_find_actor(struct inode *inode, void *opaque)
struct ocfs2_inode_info *oi = OCFS2_I(inode);
int ret = 0;
- mlog_entry("(0x%p, %lu, 0x%p)\n", inode, inode->i_ino, opaque);
+ mlog(0, "(0x%p, %lu, 0x%p)\n", inode, inode->i_ino, opaque);
args = opaque;
@@ -218,7 +218,7 @@ static int ocfs2_init_locked_inode(struct inode *inode, void *opaque)
static struct lock_class_key ocfs2_quota_ip_alloc_sem_key,
ocfs2_file_ip_alloc_sem_key;
- mlog_entry("inode = %p, opaque = %p\n", inode, opaque);
+ mlog(0, "inode = %p, opaque = %p\n", inode, opaque);
inode->i_ino = args->fi_ino;
OCFS2_I(inode)->ip_blkno = args->fi_blkno;
@@ -246,8 +246,8 @@ void ocfs2_populate_inode(struct inode *inode, struct ocfs2_dinode *fe,
struct ocfs2_super *osb;
int use_plocks = 1;
- mlog_entry("(0x%p, size:%llu)\n", inode,
- (unsigned long long)le64_to_cpu(fe->i_size));
+ mlog(0, "(0x%p, size:%llu)\n", inode,
+ (unsigned long long)le64_to_cpu(fe->i_size));
sb = inode->i_sb;
osb = OCFS2_SB(sb);
@@ -394,7 +394,7 @@ static int ocfs2_read_locked_inode(struct inode *inode,
int status, can_lock;
u32 generation = 0;
- mlog_entry("(0x%p, 0x%p)\n", inode, args);
+ mlog(0, "(0x%p, 0x%p)\n", inode, args);
status = -EINVAL;
if (inode == NULL || inode->i_sb == NULL) {
@@ -551,8 +551,6 @@ static int ocfs2_truncate_for_delete(struct ocfs2_super *osb,
struct ocfs2_dinode *fe;
handle_t *handle = NULL;
- mlog_entry_void();
-
fe = (struct ocfs2_dinode *) fe_bh->b_data;
/*
@@ -980,7 +978,7 @@ static void ocfs2_delete_inode(struct inode *inode)
sigset_t oldset;
struct buffer_head *di_bh = NULL;
- mlog_entry("(inode->i_ino = %lu)\n", inode->i_ino);
+ mlog(0, "(inode->i_ino = %lu)\n", inode->i_ino);
/* When we fail in read_inode() we mark inode as bad. The second test
* catches the case when inode allocation fails before allocating
@@ -1088,8 +1086,6 @@ static void ocfs2_clear_inode(struct inode *inode)
int status;
struct ocfs2_inode_info *oi = OCFS2_I(inode);
- mlog_entry_void();
-
end_writeback(inode);
mlog(0, "Clearing inode: %llu, nlink = %u\n",
(unsigned long long)OCFS2_I(inode)->ip_blkno, inode->i_nlink);
@@ -1204,8 +1200,6 @@ int ocfs2_drop_inode(struct inode *inode)
struct ocfs2_inode_info *oi = OCFS2_I(inode);
int res;
- mlog_entry_void();
-
mlog(0, "Drop inode %llu, nlink = %u, ip_flags = 0x%x\n",
(unsigned long long)oi->ip_blkno, inode->i_nlink, oi->ip_flags);
@@ -1226,8 +1220,8 @@ int ocfs2_inode_revalidate(struct dentry *dentry)
struct inode *inode = dentry->d_inode;
int status = 0;
- mlog_entry("(inode = 0x%p, ino = %llu)\n", inode,
- inode ? (unsigned long long)OCFS2_I(inode)->ip_blkno : 0ULL);
+ mlog(0, "(inode = 0x%p, ino = %llu)\n", inode,
+ inode ? (unsigned long long)OCFS2_I(inode)->ip_blkno : 0ULL);
if (!inode) {
mlog(0, "eep, no inode!\n");
@@ -1271,8 +1265,8 @@ int ocfs2_mark_inode_dirty(handle_t *handle,
int status;
struct ocfs2_dinode *fe = (struct ocfs2_dinode *) bh->b_data;
- mlog_entry("(inode %llu)\n",
- (unsigned long long)OCFS2_I(inode)->ip_blkno);
+ mlog(0, "(inode %llu)\n",
+ (unsigned long long)OCFS2_I(inode)->ip_blkno);
status = ocfs2_journal_access_di(handle, INODE_CACHE(inode), bh,
OCFS2_JOURNAL_ACCESS_WRITE);
diff --git a/fs/ocfs2/journal.c b/fs/ocfs2/journal.c
index faa2303..7378277 100644
--- a/fs/ocfs2/journal.c
+++ b/fs/ocfs2/journal.c
@@ -303,8 +303,6 @@ static int ocfs2_commit_cache(struct ocfs2_super *osb)
unsigned int flushed;
struct ocfs2_journal *journal = NULL;
- mlog_entry_void();
-
journal = osb->journal;
/* Flush all pending commits and checkpoint the journal. */
@@ -425,7 +423,6 @@ int ocfs2_extend_trans(handle_t *handle, int nblocks)
return 0;
old_nblocks = handle->h_buffer_credits;
- mlog_entry_void();
mlog(0, "Trying to extend transaction by %d blocks\n", nblocks);
@@ -622,12 +619,12 @@ static int __ocfs2_journal_access(handle_t *handle,
BUG_ON(!handle);
BUG_ON(!bh);
- mlog_entry("bh->b_blocknr=%llu, type=%d (\"%s\"), bh->b_size = %zu\n",
- (unsigned long long)bh->b_blocknr, type,
- (type == OCFS2_JOURNAL_ACCESS_CREATE) ?
- "OCFS2_JOURNAL_ACCESS_CREATE" :
- "OCFS2_JOURNAL_ACCESS_WRITE",
- bh->b_size);
+ mlog(0, "bh->b_blocknr=%llu, type=%d (\"%s\"), bh->b_size = %zu\n",
+ (unsigned long long)bh->b_blocknr, type,
+ (type == OCFS2_JOURNAL_ACCESS_CREATE) ?
+ "OCFS2_JOURNAL_ACCESS_CREATE" :
+ "OCFS2_JOURNAL_ACCESS_WRITE",
+ bh->b_size);
/* we can safely remove this assertion after testing. */
if (!buffer_uptodate(bh)) {
@@ -737,8 +734,8 @@ void ocfs2_journal_dirty(handle_t *handle, struct buffer_head *bh)
{
int status;
- mlog_entry("(bh->b_blocknr=%llu)\n",
- (unsigned long long)bh->b_blocknr);
+ mlog(0, "(bh->b_blocknr=%llu)\n",
+ (unsigned long long)bh->b_blocknr);
status = jbd2_journal_dirty_metadata(handle, bh);
BUG_ON(status);
@@ -775,8 +772,6 @@ int ocfs2_journal_init(struct ocfs2_journal *journal, int *dirty)
struct ocfs2_super *osb;
int inode_lock = 0;
- mlog_entry_void();
-
BUG_ON(!journal);
osb = journal->j_osb;
@@ -882,8 +877,6 @@ static int ocfs2_journal_toggle_dirty(struct ocfs2_super *osb,
struct buffer_head *bh = journal->j_bh;
struct ocfs2_dinode *fe;
- mlog_entry_void();
-
fe = (struct ocfs2_dinode *)bh->b_data;
/* The journal bh on the osb always comes from ocfs2_journal_init()
@@ -921,8 +914,6 @@ void ocfs2_journal_shutdown(struct ocfs2_super *osb)
struct inode *inode = NULL;
int num_running_trans = 0;
- mlog_entry_void();
-
BUG_ON(!osb);
journal = osb->journal;
@@ -1024,8 +1015,6 @@ int ocfs2_journal_load(struct ocfs2_journal *journal, int local, int replayed)
int status = 0;
struct ocfs2_super *osb;
- mlog_entry_void();
-
BUG_ON(!journal);
osb = journal->j_osb;
@@ -1070,8 +1059,6 @@ int ocfs2_journal_wipe(struct ocfs2_journal *journal, int full)
{
int status;
- mlog_entry_void();
-
BUG_ON(!journal);
status = jbd2_journal_wipe(journal->j_journal, full);
@@ -1124,8 +1111,6 @@ static int ocfs2_force_read_journal(struct inode *inode)
#define CONCURRENT_JOURNAL_FILL 32ULL
struct buffer_head *bhs[CONCURRENT_JOURNAL_FILL];
- mlog_entry_void();
-
memset(bhs, 0, sizeof(struct buffer_head *) * CONCURRENT_JOURNAL_FILL);
num_blocks = ocfs2_blocks_for_bytes(inode->i_sb, inode->i_size);
@@ -1194,8 +1179,6 @@ void ocfs2_complete_recovery(struct work_struct *work)
struct ocfs2_quota_recovery *qrec;
LIST_HEAD(tmp_la_list);
- mlog_entry_void();
-
mlog(0, "completing recovery from keventd\n");
spin_lock(&journal->j_lock);
@@ -1339,8 +1322,6 @@ static int __ocfs2_recovery_thread(void *arg)
int rm_quota_used = 0, i;
struct ocfs2_quota_recovery *qrec;
- mlog_entry_void();
-
status = ocfs2_wait_on_mount(osb);
if (status < 0) {
goto bail;
@@ -1461,8 +1442,8 @@ bail:
void ocfs2_recovery_thread(struct ocfs2_super *osb, int node_num)
{
- mlog_entry("(node_num=%d, osb->node_num = %d)\n",
- node_num, osb->node_num);
+ mlog(0, "(node_num=%d, osb->node_num = %d)\n",
+ node_num, osb->node_num);
mutex_lock(&osb->recovery_lock);
if (osb->disable_recovery)
@@ -1688,8 +1669,8 @@ static int ocfs2_recover_node(struct ocfs2_super *osb,
struct ocfs2_dinode *la_copy = NULL;
struct ocfs2_dinode *tl_copy = NULL;
- mlog_entry("(node_num=%d, slot_num=%d, osb->node_num = %d)\n",
- node_num, slot_num, osb->node_num);
+ mlog(0, "(node_num=%d, slot_num=%d, osb->node_num = %d)\n",
+ node_num, slot_num, osb->node_num);
/* Should not ever be called to recover ourselves -- in that
* case we should've called ocfs2_journal_load instead. */
diff --git a/fs/ocfs2/localalloc.c b/fs/ocfs2/localalloc.c
index ec6adbf..5f08c70 100644
--- a/fs/ocfs2/localalloc.c
+++ b/fs/ocfs2/localalloc.c
@@ -291,8 +291,6 @@ int ocfs2_load_local_alloc(struct ocfs2_super *osb)
struct inode *inode = NULL;
struct ocfs2_local_alloc *la;
- mlog_entry_void();
-
if (osb->local_alloc_bits == 0)
goto bail;
@@ -388,8 +386,6 @@ void ocfs2_shutdown_local_alloc(struct ocfs2_super *osb)
struct ocfs2_dinode *alloc_copy = NULL;
struct ocfs2_dinode *alloc = NULL;
- mlog_entry_void();
-
cancel_delayed_work(&osb->la_enable_wq);
flush_workqueue(ocfs2_wq);
@@ -502,7 +498,7 @@ int ocfs2_begin_local_alloc_recovery(struct ocfs2_super *osb,
struct inode *inode = NULL;
struct ocfs2_dinode *alloc;
- mlog_entry("(slot_num = %d)\n", slot_num);
+ mlog(0, "(slot_num = %d)\n", slot_num);
*alloc_copy = NULL;
@@ -570,8 +566,6 @@ int ocfs2_complete_local_alloc_recovery(struct ocfs2_super *osb,
struct buffer_head *main_bm_bh = NULL;
struct inode *main_bm_inode;
- mlog_entry_void();
-
main_bm_inode = ocfs2_get_system_file_inode(osb,
GLOBAL_BITMAP_SYSTEM_INODE,
OCFS2_INVALID_SLOT);
@@ -640,8 +634,6 @@ int ocfs2_reserve_local_alloc_bits(struct ocfs2_super *osb,
struct inode *local_alloc_inode;
unsigned int free_bits;
- mlog_entry_void();
-
BUG_ON(!ac);
local_alloc_inode =
@@ -749,7 +741,6 @@ int ocfs2_claim_local_alloc_bits(struct ocfs2_super *osb,
struct ocfs2_dinode *alloc;
struct ocfs2_local_alloc *la;
- mlog_entry_void();
BUG_ON(ac->ac_which != OCFS2_AC_USE_LOCAL);
local_alloc_inode = ac->ac_inode;
@@ -799,8 +790,6 @@ static u32 ocfs2_local_alloc_count_bits(struct ocfs2_dinode *alloc)
u32 count = 0;
struct ocfs2_local_alloc *la = OCFS2_LOCAL_ALLOC(alloc);
- mlog_entry_void();
-
buffer = la->la_bitmap;
for (i = 0; i < le16_to_cpu(la->la_size); i++)
count += hweight8(buffer[i]);
@@ -820,7 +809,7 @@ static int ocfs2_local_alloc_find_clear_bits(struct ocfs2_super *osb,
void *bitmap = NULL;
struct ocfs2_reservation_map *resmap = &osb->osb_la_resmap;
- mlog_entry("(numbits wanted = %u)\n", *numbits);
+ mlog(0, "(numbits wanted = %u)\n", *numbits);
if (!alloc->id1.bitmap1.i_total) {
mlog(0, "No bits in my window!\n");
@@ -903,7 +892,6 @@ static void ocfs2_clear_local_alloc(struct ocfs2_dinode *alloc)
{
struct ocfs2_local_alloc *la = OCFS2_LOCAL_ALLOC(alloc);
int i;
- mlog_entry_void();
alloc->id1.bitmap1.i_total = 0;
alloc->id1.bitmap1.i_used = 0;
@@ -952,9 +940,9 @@ static int ocfs2_sync_local_to_main(struct ocfs2_super *osb,
void *bitmap;
struct ocfs2_local_alloc *la = OCFS2_LOCAL_ALLOC(alloc);
- mlog_entry("total = %u, used = %u\n",
- le32_to_cpu(alloc->id1.bitmap1.i_total),
- le32_to_cpu(alloc->id1.bitmap1.i_used));
+ mlog(0, "total = %u, used = %u\n",
+ le32_to_cpu(alloc->id1.bitmap1.i_total),
+ le32_to_cpu(alloc->id1.bitmap1.i_used));
if (!alloc->id1.bitmap1.i_total) {
mlog(0, "nothing to sync!\n");
@@ -1148,8 +1136,6 @@ static int ocfs2_local_alloc_new_window(struct ocfs2_super *osb,
struct ocfs2_dinode *alloc = NULL;
struct ocfs2_local_alloc *la;
- mlog_entry_void();
-
alloc = (struct ocfs2_dinode *) osb->local_alloc_bh->b_data;
la = OCFS2_LOCAL_ALLOC(alloc);
@@ -1243,8 +1229,6 @@ static int ocfs2_local_alloc_slide_window(struct ocfs2_super *osb,
struct ocfs2_dinode *alloc_copy = NULL;
struct ocfs2_alloc_context *ac = NULL;
- mlog_entry_void();
-
ocfs2_recalc_la_window(osb, OCFS2_LA_EVENT_SLIDE);
/* This will lock the main bitmap for us. */
diff --git a/fs/ocfs2/mmap.c b/fs/ocfs2/mmap.c
index 7e32db9..87438e1 100644
--- a/fs/ocfs2/mmap.c
+++ b/fs/ocfs2/mmap.c
@@ -49,7 +49,7 @@ static int ocfs2_fault(struct vm_area_struct *area, struct vm_fault *vmf)
sigset_t oldset;
int ret;
- mlog_entry("(area=%p, page offset=%lu)\n", area, vmf->pgoff);
+ mlog(0, "(area=%p, page offset=%lu)\n", area, vmf->pgoff);
ocfs2_block_signals(&oldset);
ret = filemap_fault(area, vmf);
diff --git a/fs/ocfs2/namei.c b/fs/ocfs2/namei.c
index ff5744e..f8c8f94 100644
--- a/fs/ocfs2/namei.c
+++ b/fs/ocfs2/namei.c
@@ -106,8 +106,8 @@ static struct dentry *ocfs2_lookup(struct inode *dir, struct dentry *dentry,
struct dentry *ret;
struct ocfs2_inode_info *oi;
- mlog_entry("(0x%p, 0x%p, '%.*s')\n", dir, dentry,
- dentry->d_name.len, dentry->d_name.name);
+ mlog(0, "(0x%p, 0x%p, '%.*s')\n", dir, dentry,
+ dentry->d_name.len, dentry->d_name.name);
if (dentry->d_name.len > OCFS2_MAX_FILENAME_LEN) {
ret = ERR_PTR(-ENAMETOOLONG);
@@ -236,9 +236,9 @@ static int ocfs2_mknod(struct inode *dir,
sigset_t oldset;
int did_block_signals = 0;
- mlog_entry("(0x%p, 0x%p, %d, %lu, '%.*s')\n", dir, dentry, mode,
- (unsigned long)dev, dentry->d_name.len,
- dentry->d_name.name);
+ mlog(0, "(0x%p, 0x%p, %d, %lu, '%.*s')\n", dir, dentry, mode,
+ (unsigned long)dev, dentry->d_name.len,
+ dentry->d_name.name);
dquot_initialize(dir);
@@ -355,9 +355,9 @@ static int ocfs2_mknod(struct inode *dir,
goto leave;
did_quota_inode = 1;
- mlog_entry("(0x%p, 0x%p, %d, %lu, '%.*s')\n", dir, dentry,
- inode->i_mode, (unsigned long)dev, dentry->d_name.len,
- dentry->d_name.name);
+ mlog(0, "(0x%p, 0x%p, %d, %lu, '%.*s')\n", dir, dentry,
+ inode->i_mode, (unsigned long)dev, dentry->d_name.len,
+ dentry->d_name.name);
/* do the real work now. */
status = ocfs2_mknod_locked(osb, dir, inode, dev,
@@ -617,8 +617,8 @@ static int ocfs2_mkdir(struct inode *dir,
{
int ret;
- mlog_entry("(0x%p, 0x%p, %d, '%.*s')\n", dir, dentry, mode,
- dentry->d_name.len, dentry->d_name.name);
+ mlog(0, "(0x%p, 0x%p, %d, '%.*s')\n", dir, dentry, mode,
+ dentry->d_name.len, dentry->d_name.name);
ret = ocfs2_mknod(dir, dentry, mode | S_IFDIR, 0);
mlog_exit(ret);
@@ -632,8 +632,8 @@ static int ocfs2_create(struct inode *dir,
{
int ret;
- mlog_entry("(0x%p, 0x%p, %d, '%.*s')\n", dir, dentry, mode,
- dentry->d_name.len, dentry->d_name.name);
+ mlog(0, "(0x%p, 0x%p, %d, '%.*s')\n", dir, dentry, mode,
+ dentry->d_name.len, dentry->d_name.name);
ret = ocfs2_mknod(dir, dentry, mode | S_IFREG, 0);
mlog_exit(ret);
@@ -654,9 +654,9 @@ static int ocfs2_link(struct dentry *old_dentry,
struct ocfs2_dir_lookup_result lookup = { NULL, };
sigset_t oldset;
- mlog_entry("(inode=%lu, old='%.*s' new='%.*s')\n", inode->i_ino,
- old_dentry->d_name.len, old_dentry->d_name.name,
- dentry->d_name.len, dentry->d_name.name);
+ mlog(0, "(inode=%lu, old='%.*s' new='%.*s')\n", inode->i_ino,
+ old_dentry->d_name.len, old_dentry->d_name.name,
+ dentry->d_name.len, dentry->d_name.name);
if (S_ISDIR(inode->i_mode))
return -EPERM;
@@ -812,8 +812,8 @@ static int ocfs2_unlink(struct inode *dir,
struct ocfs2_dir_lookup_result lookup = { NULL, };
struct ocfs2_dir_lookup_result orphan_insert = { NULL, };
- mlog_entry("(0x%p, 0x%p, '%.*s')\n", dir, dentry,
- dentry->d_name.len, dentry->d_name.name);
+ mlog(0, "(0x%p, 0x%p, '%.*s')\n", dir, dentry,
+ dentry->d_name.len, dentry->d_name.name);
dquot_initialize(dir);
@@ -978,9 +978,9 @@ static int ocfs2_double_lock(struct ocfs2_super *osb,
struct buffer_head **tmpbh;
struct inode *tmpinode;
- mlog_entry("(inode1 = %llu, inode2 = %llu)\n",
- (unsigned long long)oi1->ip_blkno,
- (unsigned long long)oi2->ip_blkno);
+ mlog(0, "(inode1 = %llu, inode2 = %llu)\n",
+ (unsigned long long)oi1->ip_blkno,
+ (unsigned long long)oi2->ip_blkno);
if (*bh1)
*bh1 = NULL;
@@ -1067,10 +1067,10 @@ static int ocfs2_rename(struct inode *old_dir,
/* At some point it might be nice to break this function up a
* bit. */
- mlog_entry("(0x%p, 0x%p, 0x%p, 0x%p, from='%.*s' to='%.*s')\n",
- old_dir, old_dentry, new_dir, new_dentry,
- old_dentry->d_name.len, old_dentry->d_name.name,
- new_dentry->d_name.len, new_dentry->d_name.name);
+ mlog(0, "(0x%p, 0x%p, 0x%p, 0x%p, from='%.*s' to='%.*s')\n",
+ old_dir, old_dentry, new_dir, new_dentry,
+ old_dentry->d_name.len, old_dentry->d_name.name,
+ new_dentry->d_name.len, new_dentry->d_name.name);
dquot_initialize(old_dir);
dquot_initialize(new_dir);
@@ -1501,9 +1501,9 @@ static int ocfs2_create_symlink_data(struct ocfs2_super *osb,
* write i_size + 1 bytes. */
blocks = (bytes_left + sb->s_blocksize - 1) >> sb->s_blocksize_bits;
- mlog_entry("i_blocks = %llu, i_size = %llu, blocks = %d\n",
- (unsigned long long)inode->i_blocks,
- i_size_read(inode), blocks);
+ mlog(0, "i_blocks = %llu, i_size = %llu, blocks = %d\n",
+ (unsigned long long)inode->i_blocks,
+ i_size_read(inode), blocks);
/* Sanity check -- make sure we're going to fit. */
if (bytes_left >
@@ -1610,8 +1610,8 @@ static int ocfs2_symlink(struct inode *dir,
sigset_t oldset;
int did_block_signals = 0;
- mlog_entry("(0x%p, 0x%p, symname='%s' actual='%.*s')\n", dir,
- dentry, symname, dentry->d_name.len, dentry->d_name.name);
+ mlog(0, "(0x%p, 0x%p, symname='%s' actual='%.*s')\n", dir,
+ dentry, symname, dentry->d_name.len, dentry->d_name.name);
dquot_initialize(dir);
@@ -1713,9 +1713,9 @@ static int ocfs2_symlink(struct inode *dir,
goto bail;
did_quota_inode = 1;
- mlog_entry("(0x%p, 0x%p, %d, '%.*s')\n", dir, dentry,
- inode->i_mode, dentry->d_name.len,
- dentry->d_name.name);
+ mlog(0, "(0x%p, 0x%p, %d, '%.*s')\n", dir, dentry,
+ inode->i_mode, dentry->d_name.len,
+ dentry->d_name.name);
status = ocfs2_mknod_locked(osb, dir, inode,
0, &new_fe_bh, parent_fe_bh, handle,
@@ -1845,8 +1845,6 @@ static int ocfs2_blkno_stringify(u64 blkno, char *name)
{
int status, namelen;
- mlog_entry_void();
-
namelen = snprintf(name, OCFS2_ORPHAN_NAMELEN + 1, "%016llx",
(long long)blkno);
if (namelen <= 0) {
@@ -1998,7 +1996,7 @@ static int ocfs2_orphan_add(struct ocfs2_super *osb,
struct ocfs2_dinode *orphan_fe;
struct ocfs2_dinode *fe = (struct ocfs2_dinode *) fe_bh->b_data;
- mlog_entry("(inode->i_ino = %lu)\n", inode->i_ino);
+ mlog(0, "(inode->i_ino = %lu)\n", inode->i_ino);
status = ocfs2_read_inode_block(orphan_dir_inode, &orphan_dir_bh);
if (status < 0) {
@@ -2079,8 +2077,6 @@ int ocfs2_orphan_del(struct ocfs2_super *osb,
int status = 0;
struct ocfs2_dir_lookup_result lookup = { NULL, };
- mlog_entry_void();
-
status = ocfs2_blkno_stringify(OCFS2_I(inode)->ip_blkno, name);
if (status < 0) {
mlog_errno(status);
@@ -2359,8 +2355,8 @@ int ocfs2_mv_orphaned_inode_to_new(struct inode *dir,
struct buffer_head *di_bh = NULL;
struct ocfs2_dir_lookup_result lookup = { NULL, };
- mlog_entry("(0x%p, 0x%p, %.*s')\n", dir, dentry,
- dentry->d_name.len, dentry->d_name.name);
+ mlog(0, "(0x%p, 0x%p, %.*s')\n", dir, dentry,
+ dentry->d_name.len, dentry->d_name.name);
status = ocfs2_inode_lock(dir, &parent_di_bh, 1);
if (status < 0) {
diff --git a/fs/ocfs2/quota_global.c b/fs/ocfs2/quota_global.c
index 4607923..33397bd 100644
--- a/fs/ocfs2/quota_global.c
+++ b/fs/ocfs2/quota_global.c
@@ -343,8 +343,6 @@ int ocfs2_global_read_info(struct super_block *sb, int type)
u64 pcount;
int status;
- mlog_entry_void();
-
/* Read global header */
gqinode = ocfs2_get_system_file_inode(OCFS2_SB(sb), ino[type],
OCFS2_INVALID_SLOT);
@@ -596,7 +594,7 @@ static int ocfs2_sync_dquot_helper(struct dquot *dquot, unsigned long type)
struct ocfs2_super *osb = OCFS2_SB(sb);
int status = 0;
- mlog_entry("id=%u qtype=%u type=%lu device=%s\n", dquot->dq_id,
+ mlog(0, "id=%u qtype=%u type=%lu device=%s\n", dquot->dq_id,
dquot->dq_type, type, sb->s_id);
if (type != dquot->dq_type)
goto out;
@@ -649,7 +647,7 @@ static int ocfs2_write_dquot(struct dquot *dquot)
struct ocfs2_super *osb = OCFS2_SB(dquot->dq_sb);
int status = 0;
- mlog_entry("id=%u, type=%d", dquot->dq_id, dquot->dq_type);
+ mlog(0, "id=%u, type=%d", dquot->dq_id, dquot->dq_type);
handle = ocfs2_start_trans(osb, OCFS2_QWRITE_CREDITS);
if (IS_ERR(handle)) {
@@ -688,7 +686,7 @@ static int ocfs2_release_dquot(struct dquot *dquot)
struct ocfs2_super *osb = OCFS2_SB(dquot->dq_sb);
int status = 0;
- mlog_entry("id=%u, type=%d", dquot->dq_id, dquot->dq_type);
+ mlog(0, "id=%u, type=%d", dquot->dq_id, dquot->dq_type);
mutex_lock(&dquot->dq_lock);
/* Check whether we are not racing with some other dqget() */
@@ -745,7 +743,7 @@ static int ocfs2_acquire_dquot(struct dquot *dquot)
int need_alloc = ocfs2_global_qinit_alloc(sb, type);
handle_t *handle;
- mlog_entry("id=%u, type=%d", dquot->dq_id, type);
+ mlog(0, "id=%u, type=%d", dquot->dq_id, type);
mutex_lock(&dquot->dq_lock);
/*
* We need an exclusive lock, because we're going to update use count
@@ -831,7 +829,7 @@ static int ocfs2_mark_dquot_dirty(struct dquot *dquot)
handle_t *handle;
struct ocfs2_super *osb = OCFS2_SB(sb);
- mlog_entry("id=%u, type=%d", dquot->dq_id, type);
+ mlog(0, "id=%u, type=%d", dquot->dq_id, type);
/* In case user set some limits, sync dquot immediately to global
* quota file so that information propagates quicker */
@@ -879,8 +877,6 @@ static int ocfs2_write_info(struct super_block *sb, int type)
int status = 0;
struct ocfs2_mem_dqinfo *oinfo = sb_dqinfo(sb, type)->dqi_priv;
- mlog_entry_void();
-
status = ocfs2_lock_global_qf(oinfo, 1);
if (status < 0)
goto out;
diff --git a/fs/ocfs2/quota_local.c b/fs/ocfs2/quota_local.c
index dc78764..a71d68d 100644
--- a/fs/ocfs2/quota_local.c
+++ b/fs/ocfs2/quota_local.c
@@ -475,7 +475,7 @@ static int ocfs2_recover_local_quota_file(struct inode *lqinode,
struct ocfs2_recovery_chunk *rchunk, *next;
qsize_t spacechange, inodechange;
- mlog_entry("ino=%lu type=%u", (unsigned long)lqinode->i_ino, type);
+ mlog(0, "ino=%lu type=%u", (unsigned long)lqinode->i_ino, type);
list_for_each_entry_safe(rchunk, next, &(rec->r_list[type]), rc_list) {
chunk = rchunk->rc_chunk;
diff --git a/fs/ocfs2/resize.c b/fs/ocfs2/resize.c
index dacd553..e4a50e3 100644
--- a/fs/ocfs2/resize.c
+++ b/fs/ocfs2/resize.c
@@ -103,8 +103,8 @@ static int ocfs2_update_last_group_and_inode(handle_t *handle,
u16 cl_bpc = le16_to_cpu(cl->cl_bpc);
u16 cl_cpg = le16_to_cpu(cl->cl_cpg);
- mlog_entry("(new_clusters=%d, first_new_cluster = %u)\n",
- new_clusters, first_new_cluster);
+ mlog(0, "(new_clusters=%d, first_new_cluster = %u)\n",
+ new_clusters, first_new_cluster);
ret = ocfs2_journal_access_gd(handle, INODE_CACHE(bm_inode),
group_bh, OCFS2_JOURNAL_ACCESS_WRITE);
@@ -281,8 +281,6 @@ int ocfs2_group_extend(struct inode * inode, int new_clusters)
u32 first_new_cluster;
u64 lgd_blkno;
- mlog_entry_void();
-
if (ocfs2_is_hard_readonly(osb) || ocfs2_is_soft_readonly(osb))
return -EROFS;
@@ -472,8 +470,6 @@ int ocfs2_group_add(struct inode *inode, struct ocfs2_new_group_input *input)
struct ocfs2_chain_rec *cr;
u16 cl_bpc;
- mlog_entry_void();
-
if (ocfs2_is_hard_readonly(osb) || ocfs2_is_soft_readonly(osb))
return -EROFS;
diff --git a/fs/ocfs2/slot_map.c b/fs/ocfs2/slot_map.c
index ab4e017..a6016e6 100644
--- a/fs/ocfs2/slot_map.c
+++ b/fs/ocfs2/slot_map.c
@@ -475,8 +475,6 @@ int ocfs2_find_slot(struct ocfs2_super *osb)
int slot;
struct ocfs2_slot_info *si;
- mlog_entry_void();
-
si = osb->slot_info;
spin_lock(&osb->osb_lock);
diff --git a/fs/ocfs2/suballoc.c b/fs/ocfs2/suballoc.c
index 5fed60d..2713046 100644
--- a/fs/ocfs2/suballoc.c
+++ b/fs/ocfs2/suballoc.c
@@ -389,8 +389,6 @@ static int ocfs2_block_group_fill(handle_t *handle,
struct ocfs2_group_desc *bg = (struct ocfs2_group_desc *) bg_bh->b_data;
struct super_block * sb = alloc_inode->i_sb;
- mlog_entry_void();
-
if (((unsigned long long) bg_bh->b_blocknr) != group_blkno) {
ocfs2_error(alloc_inode->i_sb, "group block (%llu) != "
"b_blocknr (%llu)",
@@ -707,8 +705,6 @@ static int ocfs2_block_group_alloc(struct ocfs2_super *osb,
BUG_ON(ocfs2_is_cluster_bitmap(alloc_inode));
- mlog_entry_void();
-
cl = &fe->id2.i_chain;
status = ocfs2_reserve_clusters_with_limit(osb,
le16_to_cpu(cl->cl_cpg),
@@ -814,8 +810,6 @@ static int ocfs2_reserve_suballoc_bits(struct ocfs2_super *osb,
struct ocfs2_dinode *fe;
u32 free_bits;
- mlog_entry_void();
-
alloc_inode = ocfs2_get_system_file_inode(osb, type, slot);
if (!alloc_inode) {
mlog_errno(-EINVAL);
@@ -1189,8 +1183,6 @@ static int ocfs2_reserve_clusters_with_limit(struct ocfs2_super *osb,
{
int status;
- mlog_entry_void();
-
*ac = kzalloc(sizeof(struct ocfs2_alloc_context), GFP_KERNEL);
if (!(*ac)) {
status = -ENOMEM;
@@ -1357,8 +1349,6 @@ static inline int ocfs2_block_group_set_bits(handle_t *handle,
void *bitmap = bg->bg_bitmap;
int journal_type = OCFS2_JOURNAL_ACCESS_WRITE;
- mlog_entry_void();
-
/* All callers get the descriptor via
* ocfs2_read_group_descriptor(). Any corruption is a code bug. */
BUG_ON(!OCFS2_IS_VALID_GROUP_DESC(bg));
@@ -1888,8 +1878,6 @@ static int ocfs2_claim_suballoc_bits(struct ocfs2_alloc_context *ac,
struct ocfs2_chain_list *cl;
struct ocfs2_dinode *fe;
- mlog_entry_void();
-
BUG_ON(ac->ac_bits_given >= ac->ac_bits_wanted);
BUG_ON(bits_wanted > (ac->ac_bits_wanted - ac->ac_bits_given));
BUG_ON(!ac->ac_bh);
@@ -2201,8 +2189,6 @@ int ocfs2_claim_new_inode(handle_t *handle,
int status;
struct ocfs2_suballoc_result res;
- mlog_entry_void();
-
BUG_ON(!ac);
BUG_ON(ac->ac_bits_given != 0);
BUG_ON(ac->ac_bits_wanted != 1);
@@ -2307,8 +2293,6 @@ int __ocfs2_claim_clusters(handle_t *handle,
struct ocfs2_suballoc_result res = { .sr_blkno = 0, };
struct ocfs2_super *osb = OCFS2_SB(ac->ac_inode->i_sb);
- mlog_entry_void();
-
BUG_ON(ac->ac_bits_given >= ac->ac_bits_wanted);
BUG_ON(ac->ac_which != OCFS2_AC_USE_LOCAL
@@ -2392,8 +2376,6 @@ static int ocfs2_block_group_clear_bits(handle_t *handle,
unsigned int tmp;
struct ocfs2_group_desc *undo_bg = NULL;
- mlog_entry_void();
-
/* The caller got this descriptor from
* ocfs2_read_group_descriptor(). Any corruption is a code bug. */
BUG_ON(!OCFS2_IS_VALID_GROUP_DESC(bg));
@@ -2463,8 +2445,6 @@ static int _ocfs2_free_suballoc_bits(handle_t *handle,
struct buffer_head *group_bh = NULL;
struct ocfs2_group_desc *group;
- mlog_entry_void();
-
/* The alloc_bh comes from ocfs2_free_dinode() or
* ocfs2_free_clusters(). The callers have all locked the
* allocator and gotten alloc_bh from the lock call. This
@@ -2556,11 +2536,8 @@ static int _ocfs2_free_clusters(handle_t *handle,
/* You can't ever have a contiguous set of clusters
* bigger than a block group bitmap so we never have to worry
- * about looping on them. */
-
- mlog_entry_void();
-
- /* This is expensive. We can safely remove once this stuff has
+ * about looping on them.
+ * This is expensive. We can safely remove once this stuff has
* gotten tested really well. */
BUG_ON(start_blk != ocfs2_clusters_to_blocks(bitmap_inode->i_sb, ocfs2_blocks_to_clusters(bitmap_inode->i_sb, start_blk)));
@@ -2756,7 +2733,7 @@ static int ocfs2_get_suballoc_slot_bit(struct ocfs2_super *osb, u64 blkno,
struct buffer_head *inode_bh = NULL;
struct ocfs2_dinode *inode_fe;
- mlog_entry("blkno: %llu\n", (unsigned long long)blkno);
+ mlog(0, "blkno: %llu\n", (unsigned long long)blkno);
/* dirty read disk */
status = ocfs2_read_blocks_sync(osb, blkno, 1, &inode_bh);
@@ -2816,8 +2793,8 @@ static int ocfs2_test_suballoc_bit(struct ocfs2_super *osb,
u64 bg_blkno;
int status;
- mlog_entry("blkno: %llu bit: %u\n", (unsigned long long)blkno,
- (unsigned int)bit);
+ mlog(0, "blkno: %llu bit: %u\n", (unsigned long long)blkno,
+ (unsigned int)bit);
alloc_di = (struct ocfs2_dinode *)alloc_bh->b_data;
if ((bit + 1) > ocfs2_bits_per_group(&alloc_di->id2.i_chain)) {
@@ -2869,7 +2846,7 @@ int ocfs2_test_inode_bit(struct ocfs2_super *osb, u64 blkno, int *res)
struct inode *inode_alloc_inode;
struct buffer_head *alloc_bh = NULL;
- mlog_entry("blkno: %llu", (unsigned long long)blkno);
+ mlog(0, "blkno: %llu", (unsigned long long)blkno);
status = ocfs2_get_suballoc_slot_bit(osb, blkno, &suballoc_slot,
&group_blkno, &suballoc_bit);
diff --git a/fs/ocfs2/super.c b/fs/ocfs2/super.c
index cfeab7c..7572a2f 100644
--- a/fs/ocfs2/super.c
+++ b/fs/ocfs2/super.c
@@ -441,8 +441,6 @@ static int ocfs2_init_global_system_inodes(struct ocfs2_super *osb)
int status = 0;
int i;
- mlog_entry_void();
-
new = ocfs2_iget(osb, osb->root_blkno, OCFS2_FI_FLAG_SYSFILE, 0);
if (IS_ERR(new)) {
status = PTR_ERR(new);
@@ -488,8 +486,6 @@ static int ocfs2_init_local_system_inodes(struct ocfs2_super *osb)
int status = 0;
int i;
- mlog_entry_void();
-
for (i = OCFS2_LAST_GLOBAL_SYSTEM_INODE + 1;
i < NUM_SYSTEM_INODES;
i++) {
@@ -517,8 +513,6 @@ static void ocfs2_release_system_inodes(struct ocfs2_super *osb)
int i;
struct inode *inode;
- mlog_entry_void();
-
for (i = 0; i < NUM_GLOBAL_SYSTEM_INODES; i++) {
inode = osb->global_system_inodes[i];
if (inode) {
@@ -1026,7 +1020,7 @@ static int ocfs2_fill_super(struct super_block *sb, void *data, int silent)
char nodestr[8];
struct ocfs2_blockcheck_stats stats;
- mlog_entry("%p, %p, %i", sb, data, silent);
+ mlog(0, "%p, %p, %i", sb, data, silent);
if (!ocfs2_parse_options(sb, data, &parsed_options, 0)) {
status = -EINVAL;
@@ -1314,8 +1308,8 @@ static int ocfs2_parse_options(struct super_block *sb,
char *p;
u32 tmp;
- mlog_entry("remount: %d, options: \"%s\"\n", is_remount,
- options ? options : "(none)");
+ mlog(0, "remount: %d, options: \"%s\"\n", is_remount,
+ options ? options : "(none)");
mopt->commit_interval = 0;
mopt->mount_opt = OCFS2_MOUNT_NOINTR;
@@ -1611,8 +1605,6 @@ static int __init ocfs2_init(void)
{
int status;
- mlog_entry_void();
-
ocfs2_print_version();
status = init_ocfs2_uptodate_cache();
@@ -1663,8 +1655,6 @@ leave:
static void __exit ocfs2_exit(void)
{
- mlog_entry_void();
-
ocfs2_quota_shutdown();
if (ocfs2_wq) {
@@ -1687,7 +1677,7 @@ static void __exit ocfs2_exit(void)
static void ocfs2_put_super(struct super_block *sb)
{
- mlog_entry("(0x%p)\n", sb);
+ mlog(0, "(0x%p)\n", sb);
ocfs2_sync_blockdev(sb);
ocfs2_dismount_volume(sb, 0);
@@ -1704,7 +1694,7 @@ static int ocfs2_statfs(struct dentry *dentry, struct kstatfs *buf)
struct buffer_head *bh = NULL;
struct inode *inode = NULL;
- mlog_entry("(%p, %p)\n", dentry->d_sb, buf);
+ mlog(0, "(%p, %p)\n", dentry->d_sb, buf);
osb = OCFS2_SB(dentry->d_sb);
@@ -1871,8 +1861,6 @@ static int ocfs2_mount_volume(struct super_block *sb)
int unlock_super = 0;
struct ocfs2_super *osb = OCFS2_SB(sb);
- mlog_entry_void();
-
if (ocfs2_is_hard_readonly(osb))
goto leave;
@@ -1927,7 +1915,7 @@ static void ocfs2_dismount_volume(struct super_block *sb, int mnt_err)
struct ocfs2_super *osb = NULL;
char nodestr[8];
- mlog_entry("(0x%p)\n", sb);
+ mlog(0, "(0x%p)\n", sb);
BUG_ON(!sb);
osb = OCFS2_SB(sb);
@@ -2079,8 +2067,6 @@ static int ocfs2_initialize_super(struct super_block *sb,
struct ocfs2_super *osb;
u64 total_blocks;
- mlog_entry_void();
-
osb = kzalloc(sizeof(struct ocfs2_super), GFP_KERNEL);
if (!osb) {
status = -ENOMEM;
@@ -2384,8 +2370,6 @@ static int ocfs2_verify_volume(struct ocfs2_dinode *di,
{
int status = -EAGAIN;
- mlog_entry_void();
-
if (memcmp(di->i_signature, OCFS2_SUPER_BLOCK_SIGNATURE,
strlen(OCFS2_SUPER_BLOCK_SIGNATURE)) == 0) {
/* We have to do a raw check of the feature here */
@@ -2453,8 +2437,6 @@ static int ocfs2_check_volume(struct ocfs2_super *osb)
* recover
* ourselves. */
- mlog_entry_void();
-
/* Init our journal object. */
status = ocfs2_journal_init(osb->journal, &dirty);
if (status < 0) {
@@ -2549,8 +2531,6 @@ finally:
*/
static void ocfs2_delete_osb(struct ocfs2_super *osb)
{
- mlog_entry_void();
-
/* This function assumes that the caller has the main osb resource */
ocfs2_free_slot_info(osb);
diff --git a/fs/ocfs2/symlink.c b/fs/ocfs2/symlink.c
index 9975457..8a81688 100644
--- a/fs/ocfs2/symlink.c
+++ b/fs/ocfs2/symlink.c
@@ -62,8 +62,6 @@ static char *ocfs2_fast_symlink_getlink(struct inode *inode,
char *link = NULL;
struct ocfs2_dinode *fe;
- mlog_entry_void();
-
status = ocfs2_read_inode_block(inode, bh);
if (status < 0) {
mlog_errno(status);
@@ -88,8 +86,6 @@ static int ocfs2_readlink(struct dentry *dentry,
struct buffer_head *bh = NULL;
struct inode *inode = dentry->d_inode;
- mlog_entry_void();
-
link = ocfs2_fast_symlink_getlink(inode, &bh);
if (IS_ERR(link)) {
ret = PTR_ERR(link);
@@ -117,8 +113,6 @@ static void *ocfs2_fast_follow_link(struct dentry *dentry,
struct inode *inode = dentry->d_inode;
struct buffer_head *bh = NULL;
- mlog_entry_void();
-
BUG_ON(!ocfs2_inode_is_fast_symlink(inode));
target = ocfs2_fast_symlink_getlink(inode, &bh);
if (IS_ERR(target)) {
diff --git a/fs/ocfs2/xattr.c b/fs/ocfs2/xattr.c
index 67cd439..81f98b7 100644
--- a/fs/ocfs2/xattr.c
+++ b/fs/ocfs2/xattr.c
@@ -5538,7 +5538,7 @@ static int ocfs2_xattr_set_entry_bucket(struct inode *inode,
int ret;
struct ocfs2_xa_loc loc;
- mlog_entry("Set xattr %s in xattr bucket\n", xi->xi_name);
+ mlog(0, "Set xattr %s in xattr bucket\n", xi->xi_name);
ocfs2_init_xattr_bucket_xa_loc(&loc, xs->bucket,
xs->not_found ? NULL : xs->here);
@@ -5581,7 +5581,7 @@ static int ocfs2_xattr_set_entry_index_block(struct inode *inode,
{
int ret;
- mlog_entry("Set xattr %s in xattr index block\n", xi->xi_name);
+ mlog(0, "Set xattr %s in xattr index block\n", xi->xi_name);
ret = ocfs2_xattr_set_entry_bucket(inode, xi, xs, ctxt);
if (!ret)
--
1.7.0.4
^ permalink raw reply related [flat|nested] 53+ messages in thread
* [Ocfs2-devel] [PATCH 03/34] ocfs2: Remove EXIT from masklog.
2010-12-23 7:19 [Ocfs2-devel] [PATCH 00/34] OCFS2: Add trace event and replace mlog(0) Tao Ma
2010-12-23 7:30 ` [Ocfs2-devel] [PATCH 01/34] ocfs2: Remove unused truncate function from alloc.c Tao Ma
2010-12-23 7:30 ` [Ocfs2-devel] [PATCH 02/34] ocfs2: Remove ENTRY from masklog Tao Ma
@ 2010-12-23 7:30 ` Tao Ma
2010-12-23 7:30 ` [Ocfs2-devel] [PATCH 04/34] ocfs2: Add ocfs2_trace.h Tao Ma
` (35 subsequent siblings)
38 siblings, 0 replies; 53+ messages in thread
From: Tao Ma @ 2010-12-23 7:30 UTC (permalink / raw)
To: ocfs2-devel
From: Tao Ma <tao.ma@oracle.com>
mlog_exit is used to record the exit status of a function.
But because it is added in so many functions, if we enable it,
the system logs get filled up quickly and cause too much I/O.
So actually no one can open it for a production system or even
for a test.
This patch just try to remove it. So in case of report error, a
mlog_errno is used while if all the errors in the function are
properly reported, mlog_exit is just removed.
for mlog_exit:
1. if all the error paths already use mlog_errno, it is just removed.
Otherwise, it will be replaced by mlog_errno.
2. if it is used to print some return value, it is replaced with
mlog(0,...).
mlog_exit_ptr is changed to mlog(0.
All those mlog(0,...) will be replaced with trace events later.
Signed-off-by: Tao Ma <tao.ma@oracle.com>
---
fs/ocfs2/alloc.c | 26 +++------------
fs/ocfs2/aops.c | 19 ++---------
fs/ocfs2/buffer_head_io.c | 7 ++--
fs/ocfs2/cluster/heartbeat.c | 1 -
fs/ocfs2/cluster/masklog.c | 3 +-
fs/ocfs2/cluster/masklog.h | 44 --------------------------
fs/ocfs2/dcache.c | 3 +-
fs/ocfs2/dir.c | 20 ++++++------
fs/ocfs2/dlm/dlmrecovery.c | 5 +--
fs/ocfs2/dlmglue.c | 71 +++--------------------------------------
fs/ocfs2/export.c | 6 ++--
fs/ocfs2/extent_map.c | 1 -
fs/ocfs2/file.c | 21 +-----------
fs/ocfs2/inode.c | 15 +--------
fs/ocfs2/ioctl.c | 2 -
fs/ocfs2/journal.c | 19 -----------
fs/ocfs2/localalloc.c | 35 ++++++++++++---------
fs/ocfs2/mmap.c | 2 +-
fs/ocfs2/namei.c | 47 ++++++++++++++++++---------
fs/ocfs2/quota_global.c | 17 ++++++----
fs/ocfs2/quota_local.c | 3 +-
fs/ocfs2/resize.c | 6 +--
fs/ocfs2/slot_map.c | 1 -
fs/ocfs2/suballoc.c | 54 +++++++++++++++++++++----------
fs/ocfs2/super.c | 37 ++++++++-------------
fs/ocfs2/symlink.c | 7 ++--
fs/ocfs2/xattr.c | 2 -
27 files changed, 159 insertions(+), 315 deletions(-)
diff --git a/fs/ocfs2/alloc.c b/fs/ocfs2/alloc.c
index c12e9d8..9ab3f27 100644
--- a/fs/ocfs2/alloc.c
+++ b/fs/ocfs2/alloc.c
@@ -985,7 +985,7 @@ int ocfs2_num_free_extents(struct ocfs2_super *osb,
bail:
brelse(eb_bh);
- mlog_exit(retval);
+ mlog(0, "retval = %d\n", retval);
return retval;
}
@@ -1070,8 +1070,8 @@ bail:
brelse(bhs[i]);
bhs[i] = NULL;
}
+ mlog_errno(status);
}
- mlog_exit(status);
return status;
}
@@ -1326,7 +1326,6 @@ bail:
kfree(new_eb_bhs);
}
- mlog_exit(status);
return status;
}
@@ -1407,7 +1406,6 @@ static int ocfs2_shift_tree_depth(handle_t *handle,
bail:
brelse(new_eb_bh);
- mlog_exit(status);
return status;
}
@@ -1493,7 +1491,6 @@ static int ocfs2_find_branch_target(struct ocfs2_extent_tree *et,
bail:
brelse(bh);
- mlog_exit(status);
return status;
}
@@ -4552,7 +4549,7 @@ static int ocfs2_figure_insert_type(struct ocfs2_extent_tree *et,
ocfs2_et_get_last_eb_blk(et),
&bh);
if (ret) {
- mlog_exit(ret);
+ mlog_errno(ret);
goto out;
}
eb = (struct ocfs2_extent_block *) bh->b_data;
@@ -4716,7 +4713,6 @@ int ocfs2_insert_extent(handle_t *handle,
bail:
brelse(last_eb_bh);
- mlog_exit(status);
return status;
}
@@ -4818,7 +4814,6 @@ int ocfs2_add_clusters_in_btree(handle_t *handle,
}
leave:
- mlog_exit(status);
if (reason_ret)
*reason_ret = reason;
return status;
@@ -5029,7 +5024,7 @@ int ocfs2_split_extent(handle_t *handle,
ocfs2_et_get_last_eb_blk(et),
&last_eb_bh);
if (ret) {
- mlog_exit(ret);
+ mlog_errno(ret);
goto out;
}
@@ -5848,7 +5843,6 @@ int ocfs2_truncate_log_append(struct ocfs2_super *osb,
ocfs2_journal_dirty(handle, tl_bh);
bail:
- mlog_exit(status);
return status;
}
@@ -5917,7 +5911,6 @@ static int ocfs2_replay_truncate_records(struct ocfs2_super *osb,
}
bail:
- mlog_exit(status);
return status;
}
@@ -5992,7 +5985,6 @@ out_mutex:
iput(data_alloc_inode);
out:
- mlog_exit(status);
return status;
}
@@ -6020,8 +6012,6 @@ static void ocfs2_truncate_log_worker(struct work_struct *work)
mlog_errno(status);
else
ocfs2_init_steal_slots(osb);
-
- mlog_exit(status);
}
#define OCFS2_TRUNCATE_LOG_FLUSH_INTERVAL (2 * HZ)
@@ -6067,7 +6057,6 @@ static int ocfs2_get_truncate_log_info(struct ocfs2_super *osb,
*tl_inode = inode;
*tl_bh = bh;
bail:
- mlog_exit(status);
return status;
}
@@ -6138,9 +6127,9 @@ bail:
if (status < 0 && (*tl_copy)) {
kfree(*tl_copy);
*tl_copy = NULL;
+ mlog_errno(status);
}
- mlog_exit(status);
return status;
}
@@ -6198,7 +6187,6 @@ int ocfs2_complete_truncate_log_recovery(struct ocfs2_super *osb,
bail_up:
mutex_unlock(&tl_inode->i_mutex);
- mlog_exit(status);
return status;
}
@@ -6218,8 +6206,6 @@ void ocfs2_truncate_log_shutdown(struct ocfs2_super *osb)
brelse(osb->osb_tl_bh);
iput(osb->osb_tl_inode);
}
-
- mlog_exit_void();
}
int ocfs2_truncate_log_init(struct ocfs2_super *osb)
@@ -6243,7 +6229,6 @@ int ocfs2_truncate_log_init(struct ocfs2_super *osb)
osb->osb_tl_bh = tl_bh;
osb->osb_tl_inode = tl_inode;
- mlog_exit(status);
return status;
}
@@ -7109,7 +7094,6 @@ bail:
ocfs2_free_path(path);
- mlog_exit(status);
return status;
}
diff --git a/fs/ocfs2/aops.c b/fs/ocfs2/aops.c
index 9bff526..177b8b4 100644
--- a/fs/ocfs2/aops.c
+++ b/fs/ocfs2/aops.c
@@ -123,7 +123,6 @@ static int ocfs2_symlink_get_block(struct inode *inode, sector_t iblock,
bail:
brelse(bh);
- mlog_exit(err);
return err;
}
@@ -208,7 +207,6 @@ bail:
if (err < 0)
err = -EIO;
- mlog_exit(err);
return err;
}
@@ -323,7 +321,6 @@ out_inode_unlock:
out:
if (unlock)
unlock_page(page);
- mlog_exit(ret);
return ret;
}
@@ -402,8 +399,6 @@ static int ocfs2_writepage(struct page *page, struct writeback_control *wbc)
ret = block_write_full_page(page, ocfs2_get_block, wbc);
- mlog_exit(ret);
-
return ret;
}
@@ -484,8 +479,6 @@ static sector_t ocfs2_bmap(struct address_space *mapping, sector_t block)
bail:
status = err ? 0 : p_blkno;
- mlog_exit((int)status);
-
return status;
}
@@ -613,7 +606,6 @@ static ssize_t ocfs2_direct_IO(int rw,
{
struct file *file = iocb->ki_filp;
struct inode *inode = file->f_path.dentry->d_inode->i_mapping->host;
- int ret;
/*
* Fallback to buffered I/O if we see an inode without
@@ -626,13 +618,10 @@ static ssize_t ocfs2_direct_IO(int rw,
if (i_size_read(inode) <= offset)
return 0;
- ret = __blockdev_direct_IO(rw, iocb, inode, inode->i_sb->s_bdev,
- iov, offset, nr_segs,
- ocfs2_direct_IO_get_blocks,
- ocfs2_dio_end_io, NULL, 0);
-
- mlog_exit(ret);
- return ret;
+ return __blockdev_direct_IO(rw, iocb, inode, inode->i_sb->s_bdev,
+ iov, offset, nr_segs,
+ ocfs2_direct_IO_get_blocks,
+ ocfs2_dio_end_io, NULL, 0);
}
static void ocfs2_figure_cluster_boundaries(struct ocfs2_super *osb,
diff --git a/fs/ocfs2/buffer_head_io.c b/fs/ocfs2/buffer_head_io.c
index 024bd45..c23846d 100644
--- a/fs/ocfs2/buffer_head_io.c
+++ b/fs/ocfs2/buffer_head_io.c
@@ -66,6 +66,7 @@ int ocfs2_write_block(struct ocfs2_super *osb, struct buffer_head *bh,
* can get modified during recovery even if read-only. */
if (ocfs2_is_hard_readonly(osb)) {
ret = -EROFS;
+ mlog_errno(ret);
goto out;
}
@@ -91,11 +92,11 @@ int ocfs2_write_block(struct ocfs2_super *osb, struct buffer_head *bh,
* uptodate. */
ret = -EIO;
put_bh(bh);
+ mlog_errno(ret);
}
ocfs2_metadata_cache_io_unlock(ci);
out:
- mlog_exit(ret);
return ret;
}
@@ -374,7 +375,6 @@ int ocfs2_read_blocks(struct ocfs2_caching_info *ci, u64 block, int nr,
bail:
- mlog_exit(status);
return status;
}
@@ -413,6 +413,7 @@ int ocfs2_write_super_or_backup(struct ocfs2_super *osb,
if (ocfs2_is_hard_readonly(osb) || ocfs2_is_soft_readonly(osb)) {
ret = -EROFS;
+ mlog_errno(ret);
goto out;
}
@@ -432,9 +433,9 @@ int ocfs2_write_super_or_backup(struct ocfs2_super *osb,
if (!buffer_uptodate(bh)) {
ret = -EIO;
put_bh(bh);
+ mlog_errno(ret);
}
out:
- mlog_exit(ret);
return ret;
}
diff --git a/fs/ocfs2/cluster/heartbeat.c b/fs/ocfs2/cluster/heartbeat.c
index 4301242..21dfea8 100644
--- a/fs/ocfs2/cluster/heartbeat.c
+++ b/fs/ocfs2/cluster/heartbeat.c
@@ -1624,7 +1624,6 @@ static int o2hb_populate_slot_data(struct o2hb_region *reg)
}
out:
- mlog_exit(ret);
return ret;
}
diff --git a/fs/ocfs2/cluster/masklog.c b/fs/ocfs2/cluster/masklog.c
index b9ff806..41260fb 100644
--- a/fs/ocfs2/cluster/masklog.c
+++ b/fs/ocfs2/cluster/masklog.c
@@ -30,7 +30,7 @@
struct mlog_bits mlog_and_bits = MLOG_BITS_RHS(MLOG_INITIAL_AND_MASK);
EXPORT_SYMBOL_GPL(mlog_and_bits);
-struct mlog_bits mlog_not_bits = MLOG_BITS_RHS(MLOG_INITIAL_NOT_MASK);
+struct mlog_bits mlog_not_bits = MLOG_BITS_RHS(0);
EXPORT_SYMBOL_GPL(mlog_not_bits);
static ssize_t mlog_mask_show(u64 mask, char *buf)
@@ -80,7 +80,6 @@ struct mlog_attribute {
}
static struct mlog_attribute mlog_attrs[MLOG_MAX_BITS] = {
- define_mask(EXIT),
define_mask(TCP),
define_mask(MSG),
define_mask(SOCKET),
diff --git a/fs/ocfs2/cluster/masklog.h b/fs/ocfs2/cluster/masklog.h
index 72d3599..707e119 100644
--- a/fs/ocfs2/cluster/masklog.h
+++ b/fs/ocfs2/cluster/masklog.h
@@ -82,7 +82,6 @@
/* bits that are frequently given and infrequently matched in the low word */
/* NOTE: If you add a flag, you need to also update mlog.c! */
-#define ML_EXIT 0x0000000000000002ULL /* func call exit */
#define ML_TCP 0x0000000000000004ULL /* net cluster/tcp.c */
#define ML_MSG 0x0000000000000008ULL /* net network messages */
#define ML_SOCKET 0x0000000000000010ULL /* net socket lifetime */
@@ -122,7 +121,6 @@
#define ML_CLUSTER 0x0000001000000000ULL /* cluster stack */
#define MLOG_INITIAL_AND_MASK (ML_ERROR|ML_NOTICE)
-#define MLOG_INITIAL_NOT_MASK (ML_EXIT)
#ifndef MLOG_MASK_PREFIX
#define MLOG_MASK_PREFIX 0
#endif
@@ -220,48 +218,6 @@ extern struct mlog_bits mlog_and_bits, mlog_not_bits;
mlog(ML_ERROR, "status = %lld\n", (long long)_st); \
} while (0)
-#if defined(CONFIG_OCFS2_DEBUG_MASKLOG)
-/*
- * We disable this for sparse.
- */
-#if !defined(__CHECKER__)
-#define mlog_exit(st) do { \
- if (__builtin_types_compatible_p(typeof(st), unsigned long)) \
- mlog(ML_EXIT, "EXIT: %lu\n", (unsigned long) (st)); \
- else if (__builtin_types_compatible_p(typeof(st), signed long)) \
- mlog(ML_EXIT, "EXIT: %ld\n", (signed long) (st)); \
- else if (__builtin_types_compatible_p(typeof(st), unsigned int) \
- || __builtin_types_compatible_p(typeof(st), unsigned short) \
- || __builtin_types_compatible_p(typeof(st), unsigned char)) \
- mlog(ML_EXIT, "EXIT: %u\n", (unsigned int) (st)); \
- else if (__builtin_types_compatible_p(typeof(st), signed int) \
- || __builtin_types_compatible_p(typeof(st), signed short) \
- || __builtin_types_compatible_p(typeof(st), signed char)) \
- mlog(ML_EXIT, "EXIT: %d\n", (signed int) (st)); \
- else if (__builtin_types_compatible_p(typeof(st), long long)) \
- mlog(ML_EXIT, "EXIT: %lld\n", (long long) (st)); \
- else \
- mlog(ML_EXIT, "EXIT: %llu\n", (unsigned long long) (st)); \
-} while (0)
-#else
-#define mlog_exit(st) do { \
- mlog(ML_EXIT, "EXIT: %lld\n", (long long) (st)); \
-} while (0)
-#endif
-
-#define mlog_exit_ptr(ptr) do { \
- mlog(ML_EXIT, "EXIT: %p\n", ptr); \
-} while (0)
-
-#define mlog_exit_void() do { \
- mlog(ML_EXIT, "EXIT\n"); \
-} while (0)
-#else
-#define mlog_exit(...) do { } while (0)
-#define mlog_exit_ptr(...) do { } while (0)
-#define mlog_exit_void(...) do { } while (0)
-#endif /* defined(CONFIG_OCFS2_DEBUG_MASKLOG) */
-
#define mlog_bug_on_msg(cond, fmt, args...) do { \
if (cond) { \
mlog(ML_ERROR, "bug expression: " #cond "\n"); \
diff --git a/fs/ocfs2/dcache.c b/fs/ocfs2/dcache.c
index 351f3d0..866d767 100644
--- a/fs/ocfs2/dcache.c
+++ b/fs/ocfs2/dcache.c
@@ -117,8 +117,7 @@ valid:
ret = 1;
bail:
- mlog_exit(ret);
-
+ mlog(0, "ret = %d\n", ret);
return ret;
}
diff --git a/fs/ocfs2/dir.c b/fs/ocfs2/dir.c
index 7483e9e..65546e3 100644
--- a/fs/ocfs2/dir.c
+++ b/fs/ocfs2/dir.c
@@ -400,7 +400,7 @@ static int inline ocfs2_search_dirblock(struct buffer_head *bh,
}
bail:
- mlog_exit(ret);
+ mlog(0, "ret = %d\n", ret);
return ret;
}
@@ -784,7 +784,7 @@ cleanup_and_exit:
for (; ra_ptr < ra_max; ra_ptr++)
brelse(bh_use[ra_ptr]);
- mlog_exit_ptr(ret);
+ mlog(0, "ret = %p\n", ret);
return ret;
}
@@ -1198,7 +1198,6 @@ static int __ocfs2_delete_entry(handle_t *handle, struct inode *dir,
de = (struct ocfs2_dir_entry *)((char *)de + le16_to_cpu(de->rec_len));
}
bail:
- mlog_exit(status);
return status;
}
@@ -1759,8 +1758,9 @@ int __ocfs2_add_entry(handle_t *handle,
* from ever getting here. */
retval = -ENOSPC;
bail:
+ if (retval)
+ mlog_errno(retval);
- mlog_exit(retval);
return retval;
}
@@ -2045,9 +2045,10 @@ int ocfs2_readdir(struct file * filp, void * dirent, filldir_t filldir)
dirent, filldir, NULL);
ocfs2_inode_unlock(inode, lock_level);
+ if (error)
+ mlog_errno(error);
bail_nolock:
- mlog_exit(error);
return error;
}
@@ -2119,7 +2120,8 @@ int ocfs2_check_dir_for_entry(struct inode *dir,
bail:
ocfs2_free_dir_lookup_result(&lookup);
- mlog_exit(ret);
+ if (ret)
+ mlog_errno(ret);
return ret;
}
@@ -2372,7 +2374,6 @@ static int ocfs2_fill_new_dir_el(struct ocfs2_super *osb,
bail:
brelse(new_bh);
- mlog_exit(status);
return status;
}
@@ -3225,7 +3226,6 @@ static int ocfs2_do_extend_dir(struct super_block *sb,
bail:
if (did_quota && status < 0)
dquot_free_space_nodirty(dir, ocfs2_clusters_to_bytes(sb, 1));
- mlog_exit(status);
return status;
}
@@ -3424,7 +3424,6 @@ bail:
brelse(new_bh);
- mlog_exit(status);
return status;
}
@@ -3571,8 +3570,9 @@ next:
status = 0;
bail:
brelse(bh);
+ if (status)
+ mlog_errno(status);
- mlog_exit(status);
return status;
}
diff --git a/fs/ocfs2/dlm/dlmrecovery.c b/fs/ocfs2/dlm/dlmrecovery.c
index 6602b23..f1beb6f 100644
--- a/fs/ocfs2/dlm/dlmrecovery.c
+++ b/fs/ocfs2/dlm/dlmrecovery.c
@@ -727,7 +727,6 @@ static int dlm_remaster_locks(struct dlm_ctxt *dlm, u8 dead_node)
if (destroy)
dlm_destroy_recovery_area(dlm, dead_node);
- mlog_exit(status);
return status;
}
@@ -1496,9 +1495,9 @@ leave:
kfree(buf);
if (item)
kfree(item);
+ mlog_errno(ret);
}
- mlog_exit(ret);
return ret;
}
@@ -1567,7 +1566,6 @@ leave:
dlm_lockres_put(res);
}
kfree(data);
- mlog_exit(ret);
}
@@ -1986,7 +1984,6 @@ leave:
dlm_lock_put(newlock);
}
- mlog_exit(ret);
return ret;
}
diff --git a/fs/ocfs2/dlmglue.c b/fs/ocfs2/dlmglue.c
index c2f3fa1..7eb4b8d 100644
--- a/fs/ocfs2/dlmglue.c
+++ b/fs/ocfs2/dlmglue.c
@@ -406,8 +406,6 @@ static void ocfs2_build_lock_name(enum ocfs2_lock_type type,
BUG_ON(len != (OCFS2_LOCK_ID_MAX_LEN - 1));
mlog(0, "built lock resource with name: %s\n", name);
-
- mlog_exit_void();
}
static DEFINE_SPINLOCK(ocfs2_dlm_tracking_lock);
@@ -752,7 +750,6 @@ void ocfs2_lock_res_free(struct ocfs2_lock_res *res)
memset(&res->l_lksb, 0, sizeof(res->l_lksb));
res->l_flags = 0UL;
- mlog_exit_void();
}
static inline void ocfs2_inc_holders(struct ocfs2_lock_res *lockres,
@@ -770,8 +767,6 @@ static inline void ocfs2_inc_holders(struct ocfs2_lock_res *lockres,
default:
BUG();
}
-
- mlog_exit_void();
}
static inline void ocfs2_dec_holders(struct ocfs2_lock_res *lockres,
@@ -791,7 +786,6 @@ static inline void ocfs2_dec_holders(struct ocfs2_lock_res *lockres,
default:
BUG();
}
- mlog_exit_void();
}
/* WARNING: This function lives in a world where the only three lock
@@ -850,8 +844,6 @@ static inline void ocfs2_generic_handle_downconvert_action(struct ocfs2_lock_res
lockres_clear_flags(lockres, OCFS2_LOCK_BLOCKED);
}
lockres_clear_flags(lockres, OCFS2_LOCK_BUSY);
-
- mlog_exit_void();
}
static inline void ocfs2_generic_handle_convert_action(struct ocfs2_lock_res *lockres)
@@ -877,8 +869,6 @@ static inline void ocfs2_generic_handle_convert_action(struct ocfs2_lock_res *lo
lockres_or_flags(lockres, OCFS2_LOCK_UPCONVERT_FINISHING);
lockres_clear_flags(lockres, OCFS2_LOCK_BUSY);
-
- mlog_exit_void();
}
static inline void ocfs2_generic_handle_attach_action(struct ocfs2_lock_res *lockres)
@@ -894,8 +884,6 @@ static inline void ocfs2_generic_handle_attach_action(struct ocfs2_lock_res *loc
lockres->l_level = lockres->l_requested;
lockres_or_flags(lockres, OCFS2_LOCK_ATTACHED);
lockres_clear_flags(lockres, OCFS2_LOCK_BUSY);
-
- mlog_exit_void();
}
static int ocfs2_generic_handle_bast(struct ocfs2_lock_res *lockres,
@@ -923,8 +911,7 @@ static int ocfs2_generic_handle_bast(struct ocfs2_lock_res *lockres,
if (needs_downconvert)
lockres_or_flags(lockres, OCFS2_LOCK_BLOCKED);
-
- mlog_exit(needs_downconvert);
+ mlog(0, "needs_downconvert = %d\n", needs_downconvert);
return needs_downconvert;
}
@@ -1145,7 +1132,6 @@ static void ocfs2_unlock_ast(struct ocfs2_dlm_lksb *lksb, int error)
"unlock_action %d\n", error, lockres->l_name,
lockres->l_unlock_action);
spin_unlock_irqrestore(&lockres->l_lock, flags);
- mlog_exit_void();
return;
}
@@ -1169,8 +1155,6 @@ static void ocfs2_unlock_ast(struct ocfs2_dlm_lksb *lksb, int error)
lockres->l_unlock_action = OCFS2_UNLOCK_INVALID;
wake_up(&lockres->l_event);
spin_unlock_irqrestore(&lockres->l_lock, flags);
-
- mlog_exit_void();
}
/*
@@ -1226,7 +1210,6 @@ static inline void ocfs2_recover_from_dlm_error(struct ocfs2_lock_res *lockres,
spin_unlock_irqrestore(&lockres->l_lock, flags);
wake_up(&lockres->l_event);
- mlog_exit_void();
}
/* Note: If we detect another process working on the lock (i.e.,
@@ -1273,7 +1256,6 @@ static int ocfs2_lock_create(struct ocfs2_super *osb,
mlog(0, "lock %s, return from ocfs2_dlm_lock\n", lockres->l_name);
bail:
- mlog_exit(ret);
return ret;
}
@@ -1561,7 +1543,6 @@ out:
caller_ip);
}
#endif
- mlog_exit(ret);
return ret;
}
@@ -1591,7 +1572,6 @@ static void __ocfs2_cluster_unlock(struct ocfs2_super *osb,
if (lockres->l_lockdep_map.key != NULL)
rwsem_release(&lockres->l_lockdep_map, 1, caller_ip);
#endif
- mlog_exit_void();
}
static int ocfs2_create_new_lock(struct ocfs2_super *osb,
@@ -1658,7 +1638,6 @@ int ocfs2_create_new_inode_locks(struct inode *inode)
}
bail:
- mlog_exit(ret);
return ret;
}
@@ -1674,10 +1653,8 @@ int ocfs2_rw_lock(struct inode *inode, int write)
(unsigned long long)OCFS2_I(inode)->ip_blkno,
write ? "EXMODE" : "PRMODE");
- if (ocfs2_mount_local(osb)) {
- mlog_exit(0);
+ if (ocfs2_mount_local(osb))
return 0;
- }
lockres = &OCFS2_I(inode)->ip_rw_lockres;
@@ -1688,7 +1665,6 @@ int ocfs2_rw_lock(struct inode *inode, int write)
if (status < 0)
mlog_errno(status);
- mlog_exit(status);
return status;
}
@@ -1704,8 +1680,6 @@ void ocfs2_rw_unlock(struct inode *inode, int write)
if (!ocfs2_mount_local(osb))
ocfs2_cluster_unlock(OCFS2_SB(inode->i_sb), lockres, level);
-
- mlog_exit_void();
}
/*
@@ -1733,7 +1707,6 @@ int ocfs2_open_lock(struct inode *inode)
mlog_errno(status);
out:
- mlog_exit(status);
return status;
}
@@ -1766,7 +1739,6 @@ int ocfs2_try_open_lock(struct inode *inode, int write)
level, DLM_LKF_NOQUEUE, 0);
out:
- mlog_exit(status);
return status;
}
@@ -1792,7 +1764,7 @@ void ocfs2_open_unlock(struct inode *inode)
DLM_LOCK_EX);
out:
- mlog_exit_void();
+ return;
}
static int ocfs2_flock_handle_signal(struct ocfs2_lock_res *lockres,
@@ -2028,8 +2000,6 @@ static void ocfs2_downconvert_on_unlock(struct ocfs2_super *osb,
if (kick)
ocfs2_wake_downconvert_thread(osb);
-
- mlog_exit_void();
}
#define OCFS2_SEC_BITS 34
@@ -2089,8 +2059,6 @@ static void __ocfs2_stuff_meta_lvb(struct inode *inode)
out:
mlog_meta_lvb(0, lockres);
-
- mlog_exit_void();
}
static void ocfs2_unpack_timespec(struct timespec *spec,
@@ -2136,8 +2104,6 @@ static void ocfs2_refresh_inode_from_lvb(struct inode *inode)
ocfs2_unpack_timespec(&inode->i_ctime,
be64_to_cpu(lvb->lvb_ictime_packed));
spin_unlock(&oi->ip_lock);
-
- mlog_exit_void();
}
static inline int ocfs2_meta_lvb_is_trustable(struct inode *inode,
@@ -2184,7 +2150,7 @@ refresh_check:
status = 1;
bail:
- mlog_exit(status);
+ mlog(0, "status %d\n", status);
return status;
}
@@ -2202,8 +2168,6 @@ static inline void ocfs2_complete_lock_res_refresh(struct ocfs2_lock_res *lockre
spin_unlock_irqrestore(&lockres->l_lock, flags);
wake_up(&lockres->l_event);
-
- mlog_exit_void();
}
/* may or may not return a bh if it went to disk. */
@@ -2284,7 +2248,6 @@ static int ocfs2_inode_lock_update(struct inode *inode,
bail_refresh:
ocfs2_complete_lock_res_refresh(lockres, status);
bail:
- mlog_exit(status);
return status;
}
@@ -2419,7 +2382,6 @@ bail:
if (local_bh)
brelse(local_bh);
- mlog_exit(status);
return status;
}
@@ -2496,7 +2458,6 @@ int ocfs2_inode_lock_atime(struct inode *inode,
} else
*level = 0;
- mlog_exit(ret);
return ret;
}
@@ -2514,8 +2475,6 @@ void ocfs2_inode_unlock(struct inode *inode,
if (!ocfs2_is_hard_readonly(OCFS2_SB(inode->i_sb)) &&
!ocfs2_mount_local(osb))
ocfs2_cluster_unlock(OCFS2_SB(inode->i_sb), lockres, level);
-
- mlog_exit_void();
}
int ocfs2_orphan_scan_lock(struct ocfs2_super *osb, u32 *seqno)
@@ -2597,7 +2556,6 @@ int ocfs2_super_lock(struct ocfs2_super *osb,
ocfs2_track_lock_refresh(lockres);
}
bail:
- mlog_exit(status);
return status;
}
@@ -3057,7 +3015,6 @@ bail:
kthread_stop(osb->dc_task);
}
- mlog_exit(status);
return status;
}
@@ -3086,8 +3043,6 @@ void ocfs2_dlm_shutdown(struct ocfs2_super *osb,
osb->cconn = NULL;
ocfs2_dlm_shutdown_debug(osb);
-
- mlog_exit_void();
}
static int ocfs2_drop_lock(struct ocfs2_super *osb,
@@ -3169,7 +3124,6 @@ static int ocfs2_drop_lock(struct ocfs2_super *osb,
ocfs2_wait_on_busy_lock(lockres);
out:
- mlog_exit(0);
return 0;
}
@@ -3251,7 +3205,6 @@ int ocfs2_drop_inode_locks(struct inode *inode)
if (err < 0 && !status)
status = err;
- mlog_exit(status);
return status;
}
@@ -3314,7 +3267,6 @@ static int ocfs2_downconvert_lock(struct ocfs2_super *osb,
ret = 0;
bail:
- mlog_exit(ret);
return ret;
}
@@ -3362,7 +3314,6 @@ static int ocfs2_cancel_convert(struct ocfs2_super *osb,
mlog(ML_BASTS, "lockres %s\n", lockres->l_name);
- mlog_exit(ret);
return ret;
}
@@ -3552,14 +3503,14 @@ downconvert:
gen);
leave:
- mlog_exit(ret);
+ if (ret)
+ mlog_errno(ret);
return ret;
leave_requeue:
spin_unlock_irqrestore(&lockres->l_lock, flags);
ctl->requeue = 1;
- mlog_exit(0);
return 0;
}
@@ -3800,8 +3751,6 @@ static void ocfs2_set_qinfo_lvb(struct ocfs2_lock_res *lockres)
lvb->lvb_blocks = cpu_to_be32(oinfo->dqi_gi.dqi_blocks);
lvb->lvb_free_blk = cpu_to_be32(oinfo->dqi_gi.dqi_free_blk);
lvb->lvb_free_entry = cpu_to_be32(oinfo->dqi_gi.dqi_free_entry);
-
- mlog_exit_void();
}
void ocfs2_qinfo_unlock(struct ocfs2_mem_dqinfo *oinfo, int ex)
@@ -3812,7 +3761,6 @@ void ocfs2_qinfo_unlock(struct ocfs2_mem_dqinfo *oinfo, int ex)
if (!ocfs2_is_hard_readonly(osb) && !ocfs2_mount_local(osb))
ocfs2_cluster_unlock(osb, lockres, level);
- mlog_exit_void();
}
static int ocfs2_refresh_qinfo(struct ocfs2_mem_dqinfo *oinfo)
@@ -3889,7 +3837,6 @@ int ocfs2_qinfo_lock(struct ocfs2_mem_dqinfo *oinfo, int ex)
ocfs2_qinfo_unlock(oinfo, ex);
ocfs2_complete_lock_res_refresh(lockres, status);
bail:
- mlog_exit(status);
return status;
}
@@ -3968,8 +3915,6 @@ unqueue:
if (ctl.unblock_action != UNBLOCK_CONTINUE
&& lockres->l_ops->post_unlock)
lockres->l_ops->post_unlock(osb, lockres);
-
- mlog_exit_void();
}
static void ocfs2_schedule_blocked_lock(struct ocfs2_super *osb,
@@ -3995,8 +3940,6 @@ static void ocfs2_schedule_blocked_lock(struct ocfs2_super *osb,
osb->blocked_lock_count++;
}
spin_unlock(&osb->dc_task_lock);
-
- mlog_exit_void();
}
static void ocfs2_downconvert_thread_do_work(struct ocfs2_super *osb)
@@ -4027,8 +3970,6 @@ static void ocfs2_downconvert_thread_do_work(struct ocfs2_super *osb)
spin_lock(&osb->dc_task_lock);
}
spin_unlock(&osb->dc_task_lock);
-
- mlog_exit_void();
}
static int ocfs2_downconvert_thread_lists_empty(struct ocfs2_super *osb)
diff --git a/fs/ocfs2/export.c b/fs/ocfs2/export.c
index 2173cc2..190158b 100644
--- a/fs/ocfs2/export.c
+++ b/fs/ocfs2/export.c
@@ -143,7 +143,7 @@ check_gen:
mlog_errno(PTR_ERR(result));
bail:
- mlog_exit_ptr(result);
+ mlog(0, "result = %p\n", result);
return result;
}
@@ -182,7 +182,7 @@ bail_unlock:
ocfs2_inode_unlock(dir, 0);
bail:
- mlog_exit_ptr(parent);
+ mlog(0, "parent %p\n", parent);
return parent;
}
@@ -243,7 +243,7 @@ static int ocfs2_encode_fh(struct dentry *dentry, u32 *fh_in, int *max_len,
*max_len = len;
bail:
- mlog_exit(type);
+ mlog(0, "type = %d\n", type);
return type;
}
diff --git a/fs/ocfs2/extent_map.c b/fs/ocfs2/extent_map.c
index ed9916a..7d74d17 100644
--- a/fs/ocfs2/extent_map.c
+++ b/fs/ocfs2/extent_map.c
@@ -897,7 +897,6 @@ int ocfs2_read_virt_blocks(struct inode *inode, u64 v_block, int nr,
}
out:
- mlog_exit(rc);
return rc;
}
diff --git a/fs/ocfs2/file.c b/fs/ocfs2/file.c
index 480accf..14f335d 100644
--- a/fs/ocfs2/file.c
+++ b/fs/ocfs2/file.c
@@ -135,7 +135,6 @@ static int ocfs2_file_open(struct inode *inode, struct file *file)
}
leave:
- mlog_exit(status);
return status;
}
@@ -154,8 +153,6 @@ static int ocfs2_file_release(struct inode *inode, struct file *file)
ocfs2_free_file_private(inode, file);
- mlog_exit(0);
-
return 0;
}
@@ -195,7 +192,8 @@ static int ocfs2_sync_file(struct file *file, int datasync)
err = jbd2_journal_force_commit(journal);
bail:
- mlog_exit(err);
+ if (err)
+ mlog_errno(err);
return (err < 0) ? -EIO : 0;
}
@@ -278,7 +276,6 @@ int ocfs2_update_inode_atime(struct inode *inode,
out_commit:
ocfs2_commit_trans(OCFS2_SB(inode->i_sb), handle);
out:
- mlog_exit(ret);
return ret;
}
@@ -300,7 +297,6 @@ static int ocfs2_set_inode_size(handle_t *handle,
}
bail:
- mlog_exit(status);
return status;
}
@@ -424,8 +420,6 @@ static int ocfs2_orphan_for_truncate(struct ocfs2_super *osb,
out_commit:
ocfs2_commit_trans(osb, handle);
out:
-
- mlog_exit(status);
return status;
}
@@ -520,7 +514,6 @@ bail:
if (!status && OCFS2_I(inode)->ip_clusters == 0)
status = ocfs2_try_remove_refcount_tree(inode, di_bh);
- mlog_exit(status);
return status;
}
@@ -713,7 +706,6 @@ leave:
brelse(bh);
bh = NULL;
- mlog_exit(status);
return status;
}
@@ -1269,7 +1261,6 @@ bail:
mlog_errno(status);
}
- mlog_exit(status);
return status;
}
@@ -1295,8 +1286,6 @@ int ocfs2_getattr(struct vfsmount *mnt,
stat->blksize = osb->s_clustersize;
bail:
- mlog_exit(err);
-
return err;
}
@@ -1315,7 +1304,6 @@ int ocfs2_permission(struct inode *inode, int mask)
ocfs2_inode_unlock(inode, 0);
out:
- mlog_exit(ret);
return ret;
}
@@ -1356,7 +1344,6 @@ static int __ocfs2_write_remove_suid(struct inode *inode,
out_trans:
ocfs2_commit_trans(osb, handle);
out:
- mlog_exit(ret);
return ret;
}
@@ -2380,7 +2367,6 @@ out_sems:
if (written)
ret = written;
- mlog_exit(ret);
return ret;
}
@@ -2463,7 +2449,6 @@ static ssize_t ocfs2_file_splice_write(struct pipe_inode_info *pipe,
balance_dirty_pages_ratelimited_nr(mapping, nr_pages);
}
- mlog_exit(ret);
return ret;
}
@@ -2494,7 +2479,6 @@ static ssize_t ocfs2_file_splice_read(struct file *in,
ret = generic_file_splice_read(in, ppos, pipe, len, flags);
bail:
- mlog_exit(ret);
return ret;
}
@@ -2570,7 +2554,6 @@ bail:
up_read(&inode->i_alloc_sem);
if (rw_level != -1)
ocfs2_rw_unlock(inode, rw_level);
- mlog_exit(ret);
return ret;
}
diff --git a/fs/ocfs2/inode.c b/fs/ocfs2/inode.c
index 7bdfcb6..f5c0f34 100644
--- a/fs/ocfs2/inode.c
+++ b/fs/ocfs2/inode.c
@@ -172,7 +172,7 @@ bail:
if (!IS_ERR(inode)) {
mlog(0, "returning inode with number %llu\n",
(unsigned long long)OCFS2_I(inode)->ip_blkno);
- mlog_exit_ptr(inode);
+ mlog(0, "inode %p\n", inode);
}
return inode;
@@ -203,7 +203,6 @@ static int ocfs2_find_actor(struct inode *inode, void *opaque)
ret = 1;
bail:
- mlog_exit(ret);
return ret;
}
@@ -235,7 +234,6 @@ static int ocfs2_init_locked_inode(struct inode *inode, void *opaque)
lockdep_set_class(&OCFS2_I(inode)->ip_alloc_sem,
&ocfs2_file_ip_alloc_sem_key);
- mlog_exit(0);
return 0;
}
@@ -381,7 +379,6 @@ void ocfs2_populate_inode(struct inode *inode, struct ocfs2_dinode *fe,
if (S_ISDIR(inode->i_mode))
ocfs2_resv_set_type(&OCFS2_I(inode)->ip_la_data_resv,
OCFS2_RESV_FLAG_DIR);
- mlog_exit_void();
}
static int ocfs2_read_locked_inode(struct inode *inode,
@@ -534,7 +531,6 @@ bail:
if (args && bh)
brelse(bh);
- mlog_exit(status);
return status;
}
@@ -598,7 +594,6 @@ static int ocfs2_truncate_for_delete(struct ocfs2_super *osb,
out:
if (handle)
ocfs2_commit_trans(osb, handle);
- mlog_exit(status);
return status;
}
@@ -1078,7 +1073,7 @@ bail_unlock_nfs_sync:
bail_unblock:
ocfs2_unblock_signals(&oldset);
bail:
- mlog_exit_void();
+ return;
}
static void ocfs2_clear_inode(struct inode *inode)
@@ -1177,8 +1172,6 @@ static void ocfs2_clear_inode(struct inode *inode)
*/
jbd2_journal_release_jbd_inode(OCFS2_SB(inode->i_sb)->journal->j_journal,
&oi->ip_jinode);
-
- mlog_exit_void();
}
void ocfs2_evict_inode(struct inode *inode)
@@ -1208,7 +1201,6 @@ int ocfs2_drop_inode(struct inode *inode)
else
res = generic_drop_inode(inode);
- mlog_exit_void();
return res;
}
@@ -1248,8 +1240,6 @@ int ocfs2_inode_revalidate(struct dentry *dentry)
}
ocfs2_inode_unlock(inode, 0);
bail:
- mlog_exit(status);
-
return status;
}
@@ -1296,7 +1286,6 @@ int ocfs2_mark_inode_dirty(handle_t *handle,
ocfs2_journal_dirty(handle, bh);
leave:
- mlog_exit(status);
return status;
}
diff --git a/fs/ocfs2/ioctl.c b/fs/ocfs2/ioctl.c
index 7a48681..b4f51e9 100644
--- a/fs/ocfs2/ioctl.c
+++ b/fs/ocfs2/ioctl.c
@@ -59,7 +59,6 @@ static int ocfs2_get_inode_attr(struct inode *inode, unsigned *flags)
*flags = OCFS2_I(inode)->ip_attr;
ocfs2_inode_unlock(inode, 0);
- mlog_exit(status);
return status;
}
@@ -125,7 +124,6 @@ bail:
brelse(bh);
- mlog_exit(status);
return status;
}
diff --git a/fs/ocfs2/journal.c b/fs/ocfs2/journal.c
index 7378277..9565091 100644
--- a/fs/ocfs2/journal.c
+++ b/fs/ocfs2/journal.c
@@ -335,7 +335,6 @@ static int ocfs2_commit_cache(struct ocfs2_super *osb)
ocfs2_wake_downconvert_thread(osb);
wake_up(&journal->j_checkpointed);
finally:
- mlog_exit(status);
return status;
}
@@ -450,8 +449,6 @@ int ocfs2_extend_trans(handle_t *handle, int nblocks)
status = 0;
bail:
-
- mlog_exit(status);
return status;
}
@@ -665,7 +662,6 @@ static int __ocfs2_journal_access(handle_t *handle,
mlog(ML_ERROR, "Error %d getting %d access to buffer!\n",
status, type);
- mlog_exit(status);
return status;
}
@@ -739,8 +735,6 @@ void ocfs2_journal_dirty(handle_t *handle, struct buffer_head *bh)
status = jbd2_journal_dirty_metadata(handle, bh);
BUG_ON(status);
-
- mlog_exit_void();
}
#define OCFS2_DEFAULT_COMMIT_INTERVAL (HZ * JBD2_DEFAULT_MAX_COMMIT_AGE)
@@ -854,7 +848,6 @@ done:
}
}
- mlog_exit(status);
return status;
}
@@ -899,7 +892,6 @@ static int ocfs2_journal_toggle_dirty(struct ocfs2_super *osb,
if (status < 0)
mlog_errno(status);
- mlog_exit(status);
return status;
}
@@ -989,7 +981,6 @@ void ocfs2_journal_shutdown(struct ocfs2_super *osb)
done:
if (inode)
iput(inode);
- mlog_exit_void();
}
static void ocfs2_clear_journal_error(struct super_block *sb,
@@ -1048,7 +1039,6 @@ int ocfs2_journal_load(struct ocfs2_journal *journal, int local, int replayed)
osb->commit_task = NULL;
done:
- mlog_exit(status);
return status;
}
@@ -1072,7 +1062,6 @@ int ocfs2_journal_wipe(struct ocfs2_journal *journal, int full)
mlog_errno(status);
bail:
- mlog_exit(status);
return status;
}
@@ -1146,7 +1135,6 @@ static int ocfs2_force_read_journal(struct inode *inode)
bail:
for(i = 0; i < CONCURRENT_JOURNAL_FILL; i++)
brelse(bhs[i]);
- mlog_exit(status);
return status;
}
@@ -1236,7 +1224,6 @@ void ocfs2_complete_recovery(struct work_struct *work)
}
mlog(0, "Recovery completion\n");
- mlog_exit_void();
}
/* NOTE: This function always eats your references to la_dinode and
@@ -1432,7 +1419,6 @@ bail:
if (rm_quota)
kfree(rm_quota);
- mlog_exit(status);
/* no one is callint kthread_stop() for us so the kthread() api
* requires that we call do_exit(). And it isn't exported, but
* complete_and_exit() seems to be a minimal wrapper around it. */
@@ -1469,8 +1455,6 @@ void ocfs2_recovery_thread(struct ocfs2_super *osb, int node_num)
out:
mutex_unlock(&osb->recovery_lock);
wake_up(&osb->recovery_event);
-
- mlog_exit_void();
}
static int ocfs2_read_journal_inode(struct ocfs2_super *osb,
@@ -1646,7 +1630,6 @@ done:
brelse(bh);
- mlog_exit(status);
return status;
}
@@ -1716,7 +1699,6 @@ static int ocfs2_recover_node(struct ocfs2_super *osb,
status = 0;
done:
- mlog_exit(status);
return status;
}
@@ -1826,7 +1808,6 @@ int ocfs2_mark_dead_nodes(struct ocfs2_super *osb)
status = 0;
bail:
- mlog_exit(status);
return status;
}
diff --git a/fs/ocfs2/localalloc.c b/fs/ocfs2/localalloc.c
index 5f08c70..36636e1 100644
--- a/fs/ocfs2/localalloc.c
+++ b/fs/ocfs2/localalloc.c
@@ -364,7 +364,8 @@ bail:
mlog(0, "Local alloc window bits = %d\n", osb->local_alloc_bits);
- mlog_exit(status);
+ if (status)
+ mlog_errno(status);
return status;
}
@@ -478,8 +479,6 @@ out:
if (alloc_copy)
kfree(alloc_copy);
-
- mlog_exit_void();
}
/*
@@ -548,7 +547,8 @@ bail:
iput(inode);
}
- mlog_exit(status);
+ if (status)
+ mlog_errno(status);
return status;
}
@@ -614,7 +614,8 @@ out_mutex:
out:
if (!status)
ocfs2_init_steal_slots(osb);
- mlog_exit(status);
+ if (status)
+ mlog_errno(status);
return status;
}
@@ -724,7 +725,8 @@ bail:
mlog(0, "bits=%d, slot=%d, ret=%d\n", bits_wanted, osb->slot_num,
status);
- mlog_exit(status);
+ if (status)
+ mlog_errno(status);
return status;
}
@@ -779,7 +781,8 @@ int ocfs2_claim_local_alloc_bits(struct ocfs2_super *osb,
ocfs2_journal_dirty(handle, osb->local_alloc_bh);
bail:
- mlog_exit(status);
+ if (status)
+ mlog_errno(status);
return status;
}
@@ -794,7 +797,7 @@ static u32 ocfs2_local_alloc_count_bits(struct ocfs2_dinode *alloc)
for (i = 0; i < le16_to_cpu(la->la_size); i++)
count += hweight8(buffer[i]);
- mlog_exit(count);
+ mlog(0, "count %u\n", count);
return count;
}
@@ -884,7 +887,7 @@ bail:
if (local_resv)
ocfs2_resv_discard(resmap, resv);
- mlog_exit(bitoff);
+ mlog(0, "bitoff %d\n", bitoff);
return bitoff;
}
@@ -898,8 +901,6 @@ static void ocfs2_clear_local_alloc(struct ocfs2_dinode *alloc)
la->la_bm_off = 0;
for(i = 0; i < le16_to_cpu(la->la_size); i++)
la->la_bitmap[i] = 0;
-
- mlog_exit_void();
}
#if 0
@@ -995,7 +996,8 @@ static int ocfs2_sync_local_to_main(struct ocfs2_super *osb,
}
bail:
- mlog_exit(status);
+ if (status)
+ mlog_errno(status);
return status;
}
@@ -1120,7 +1122,8 @@ bail:
*ac = NULL;
}
- mlog_exit(status);
+ if (status)
+ mlog_errno(status);
return status;
}
@@ -1212,7 +1215,8 @@ retry_enospc:
mlog(0, "window bits = %u\n", le32_to_cpu(alloc->id1.bitmap1.i_total));
bail:
- mlog_exit(status);
+ if (status)
+ mlog_errno(status);
return status;
}
@@ -1308,7 +1312,8 @@ bail:
if (ac)
ocfs2_free_alloc_context(ac);
- mlog_exit(status);
+ if (status)
+ mlog_errno(status);
return status;
}
diff --git a/fs/ocfs2/mmap.c b/fs/ocfs2/mmap.c
index 87438e1..49730b5 100644
--- a/fs/ocfs2/mmap.c
+++ b/fs/ocfs2/mmap.c
@@ -55,7 +55,7 @@ static int ocfs2_fault(struct vm_area_struct *area, struct vm_fault *vmf)
ret = filemap_fault(area, vmf);
ocfs2_unblock_signals(&oldset);
- mlog_exit_ptr(vmf->page);
+ mlog(0, "%p\n", vmf->page);
return ret;
}
diff --git a/fs/ocfs2/namei.c b/fs/ocfs2/namei.c
index f8c8f94..f52e6bf 100644
--- a/fs/ocfs2/namei.c
+++ b/fs/ocfs2/namei.c
@@ -183,7 +183,7 @@ bail_unlock:
bail:
- mlog_exit_ptr(ret);
+ mlog(0, "%p\n", ret);
return ret;
}
@@ -468,7 +468,8 @@ leave:
iput(inode);
}
- mlog_exit(status);
+ if (status)
+ mlog_errno(status);
return status;
}
@@ -579,7 +580,8 @@ leave:
}
}
- mlog_exit(status);
+ if (status)
+ mlog_errno(status);
return status;
}
@@ -620,7 +622,8 @@ static int ocfs2_mkdir(struct inode *dir,
mlog(0, "(0x%p, 0x%p, %d, '%.*s')\n", dir, dentry, mode,
dentry->d_name.len, dentry->d_name.name);
ret = ocfs2_mknod(dir, dentry, mode | S_IFDIR, 0);
- mlog_exit(ret);
+ if (ret)
+ mlog_errno(ret);
return ret;
}
@@ -635,7 +638,8 @@ static int ocfs2_create(struct inode *dir,
mlog(0, "(0x%p, 0x%p, %d, '%.*s')\n", dir, dentry, mode,
dentry->d_name.len, dentry->d_name.name);
ret = ocfs2_mknod(dir, dentry, mode | S_IFREG, 0);
- mlog_exit(ret);
+ if (ret)
+ mlog_errno(ret);
return ret;
}
@@ -760,7 +764,8 @@ out:
ocfs2_free_dir_lookup_result(&lookup);
- mlog_exit(err);
+ if (err)
+ mlog_errno(err);
return err;
}
@@ -957,7 +962,8 @@ leave:
ocfs2_free_dir_lookup_result(&orphan_insert);
ocfs2_free_dir_lookup_result(&lookup);
- mlog_exit(status);
+ if (status)
+ mlog_errno(status);
return status;
}
@@ -1025,7 +1031,8 @@ static int ocfs2_double_lock(struct ocfs2_super *osb,
}
bail:
- mlog_exit(status);
+ if (status)
+ mlog_errno(status);
return status;
}
@@ -1476,7 +1483,8 @@ bail:
brelse(old_dir_bh);
brelse(new_dir_bh);
- mlog_exit(status);
+ if (status)
+ mlog_errno(status);
return status;
}
@@ -1579,7 +1587,8 @@ bail:
kfree(bhs);
}
- mlog_exit(status);
+ if (status)
+ mlog_errno(status);
return status;
}
@@ -1836,7 +1845,8 @@ bail:
iput(inode);
}
- mlog_exit(status);
+ if (status)
+ mlog_errno(status);
return status;
}
@@ -1866,7 +1876,8 @@ static int ocfs2_blkno_stringify(u64 blkno, char *name)
status = 0;
bail:
- mlog_exit(status);
+ if (status < 0)
+ mlog_errno(status);
return status;
}
@@ -1979,7 +1990,8 @@ out:
iput(orphan_dir_inode);
}
- mlog_exit(ret);
+ if (ret)
+ mlog_errno(ret);
return ret;
}
@@ -2061,7 +2073,8 @@ static int ocfs2_orphan_add(struct ocfs2_super *osb,
leave:
brelse(orphan_dir_bh);
- mlog_exit(status);
+ if (status)
+ mlog_errno(status);
return status;
}
@@ -2121,7 +2134,8 @@ int ocfs2_orphan_del(struct ocfs2_super *osb,
leave:
ocfs2_free_dir_lookup_result(&lookup);
- mlog_exit(status);
+ if (status)
+ mlog_errno(status);
return status;
}
@@ -2474,7 +2488,8 @@ leave:
ocfs2_free_dir_lookup_result(&lookup);
- mlog_exit(status);
+ if (status)
+ mlog_errno(status);
return status;
}
diff --git a/fs/ocfs2/quota_global.c b/fs/ocfs2/quota_global.c
index 33397bd..10a49e8 100644
--- a/fs/ocfs2/quota_global.c
+++ b/fs/ocfs2/quota_global.c
@@ -402,7 +402,8 @@ int ocfs2_global_read_info(struct super_block *sb, int type)
msecs_to_jiffies(oinfo->dqi_syncms));
out_err:
- mlog_exit(status);
+ if (status)
+ mlog_errno(status);
return status;
out_unlock:
ocfs2_unlock_global_qf(oinfo, 0);
@@ -621,7 +622,6 @@ static int ocfs2_sync_dquot_helper(struct dquot *dquot, unsigned long type)
out_ilock:
ocfs2_unlock_global_qf(oinfo, 1);
out:
- mlog_exit(status);
return status;
}
@@ -660,7 +660,6 @@ static int ocfs2_write_dquot(struct dquot *dquot)
mutex_unlock(&sb_dqopt(dquot->dq_sb)->dqio_mutex);
ocfs2_commit_trans(osb, handle);
out:
- mlog_exit(status);
return status;
}
@@ -722,7 +721,8 @@ out_ilock:
ocfs2_unlock_global_qf(oinfo, 1);
out:
mutex_unlock(&dquot->dq_lock);
- mlog_exit(status);
+ if (status)
+ mlog_errno(status);
return status;
}
@@ -809,7 +809,8 @@ out_dq:
set_bit(DQ_ACTIVE_B, &dquot->dq_flags);
out:
mutex_unlock(&dquot->dq_lock);
- mlog_exit(status);
+ if (status)
+ mlog_errno(status);
return status;
}
@@ -866,7 +867,8 @@ out_dlock:
out_ilock:
ocfs2_unlock_global_qf(oinfo, 1);
out:
- mlog_exit(status);
+ if (status)
+ mlog_errno(status);
return status;
}
@@ -891,7 +893,8 @@ static int ocfs2_write_info(struct super_block *sb, int type)
out_ilock:
ocfs2_unlock_global_qf(oinfo, 1);
out:
- mlog_exit(status);
+ if (status)
+ mlog_errno(status);
return status;
}
diff --git a/fs/ocfs2/quota_local.c b/fs/ocfs2/quota_local.c
index a71d68d..88ec451 100644
--- a/fs/ocfs2/quota_local.c
+++ b/fs/ocfs2/quota_local.c
@@ -575,7 +575,8 @@ out_put_bh:
}
if (status < 0)
free_recovery_list(&(rec->r_list[type]));
- mlog_exit(status);
+ if (status)
+ mlog_errno(status);
return status;
}
diff --git a/fs/ocfs2/resize.c b/fs/ocfs2/resize.c
index e4a50e3..76b3790 100644
--- a/fs/ocfs2/resize.c
+++ b/fs/ocfs2/resize.c
@@ -82,7 +82,6 @@ static u16 ocfs2_calc_new_backup_super(struct inode *inode,
backups++;
}
- mlog_exit_void();
return backups;
}
@@ -176,7 +175,8 @@ out_rollback:
le16_add_cpu(&group->bg_free_bits_count, -1 * num_bits);
}
out:
- mlog_exit(ret);
+ if (ret)
+ mlog_errno(ret);
return ret;
}
@@ -375,7 +375,6 @@ out_mutex:
iput(main_bm_inode);
out:
- mlog_exit_void();
return ret;
}
@@ -585,6 +584,5 @@ out_mutex:
iput(main_bm_inode);
out:
- mlog_exit_void();
return ret;
}
diff --git a/fs/ocfs2/slot_map.c b/fs/ocfs2/slot_map.c
index a6016e6..93b449e 100644
--- a/fs/ocfs2/slot_map.c
+++ b/fs/ocfs2/slot_map.c
@@ -510,7 +510,6 @@ int ocfs2_find_slot(struct ocfs2_super *osb)
mlog_errno(status);
bail:
- mlog_exit(status);
return status;
}
diff --git a/fs/ocfs2/suballoc.c b/fs/ocfs2/suballoc.c
index 2713046..709851f 100644
--- a/fs/ocfs2/suballoc.c
+++ b/fs/ocfs2/suballoc.c
@@ -434,7 +434,8 @@ static int ocfs2_block_group_fill(handle_t *handle,
* allocation time. */
bail:
- mlog_exit(status);
+ if (status)
+ mlog_errno(status);
return status;
}
@@ -792,7 +793,8 @@ bail:
brelse(bg_bh);
- mlog_exit(status);
+ if (status)
+ mlog_errno(status);
return status;
}
@@ -884,7 +886,8 @@ static int ocfs2_reserve_suballoc_bits(struct ocfs2_super *osb,
bail:
brelse(bh);
- mlog_exit(status);
+ if (status)
+ mlog_errno(status);
return status;
}
@@ -1046,7 +1049,8 @@ bail:
*ac = NULL;
}
- mlog_exit(status);
+ if (status)
+ mlog_errno(status);
return status;
}
@@ -1146,7 +1150,8 @@ bail:
*ac = NULL;
}
- mlog_exit(status);
+ if (status)
+ mlog_errno(status);
return status;
}
@@ -1221,7 +1226,8 @@ bail:
*ac = NULL;
}
- mlog_exit(status);
+ if (status)
+ mlog_errno(status);
return status;
}
@@ -1384,7 +1390,8 @@ static inline int ocfs2_block_group_set_bits(handle_t *handle,
ocfs2_journal_dirty(handle, group_bh);
bail:
- mlog_exit(status);
+ if (status)
+ mlog_errno(status);
return status;
}
@@ -1474,7 +1481,8 @@ out_rollback:
prev_bg->bg_next_group = cpu_to_le64(prev_bg_ptr);
}
- mlog_exit(status);
+ if (status)
+ mlog_errno(status);
return status;
}
@@ -1860,7 +1868,8 @@ bail:
brelse(group_bh);
brelse(prev_group_bh);
- mlog_exit(status);
+ if (status)
+ mlog_errno(status);
return status;
}
@@ -1972,7 +1981,8 @@ set_hint:
}
bail:
- mlog_exit(status);
+ if (status)
+ mlog_errno(status);
return status;
}
@@ -2009,7 +2019,8 @@ int ocfs2_claim_metadata(handle_t *handle,
*num_bits = res.sr_bits;
status = 0;
bail:
- mlog_exit(status);
+ if (status)
+ mlog_errno(status);
return status;
}
@@ -2216,7 +2227,8 @@ int ocfs2_claim_new_inode(handle_t *handle,
ocfs2_save_inode_ac_group(dir, ac);
status = 0;
bail:
- mlog_exit(status);
+ if (status)
+ mlog_errno(status);
return status;
}
@@ -2347,7 +2359,8 @@ int __ocfs2_claim_clusters(handle_t *handle,
ac->ac_bits_given += *num_clusters;
bail:
- mlog_exit(status);
+ if (status)
+ mlog_errno(status);
return status;
}
@@ -2491,7 +2504,8 @@ static int _ocfs2_free_suballoc_bits(handle_t *handle,
bail:
brelse(group_bh);
- mlog_exit(status);
+ if (status)
+ mlog_errno(status);
return status;
}
@@ -2563,7 +2577,8 @@ static int _ocfs2_free_clusters(handle_t *handle,
num_clusters);
out:
- mlog_exit(status);
+ if (status)
+ mlog_errno(status);
return status;
}
@@ -2770,7 +2785,8 @@ static int ocfs2_get_suballoc_slot_bit(struct ocfs2_super *osb, u64 blkno,
bail:
brelse(inode_bh);
- mlog_exit(status);
+ if (status)
+ mlog_errno(status);
return status;
}
@@ -2821,7 +2837,8 @@ static int ocfs2_test_suballoc_bit(struct ocfs2_super *osb,
bail:
brelse(group_bh);
- mlog_exit(status);
+ if (status)
+ mlog_errno(status);
return status;
}
@@ -2887,6 +2904,7 @@ int ocfs2_test_inode_bit(struct ocfs2_super *osb, u64 blkno, int *res)
iput(inode_alloc_inode);
brelse(alloc_bh);
bail:
- mlog_exit(status);
+ if (status)
+ mlog_errno(status);
return status;
}
diff --git a/fs/ocfs2/super.c b/fs/ocfs2/super.c
index 7572a2f..9e41894 100644
--- a/fs/ocfs2/super.c
+++ b/fs/ocfs2/super.c
@@ -476,7 +476,8 @@ static int ocfs2_init_global_system_inodes(struct ocfs2_super *osb)
}
bail:
- mlog_exit(status);
+ if (status)
+ mlog_errno(status);
return status;
}
@@ -504,7 +505,8 @@ static int ocfs2_init_local_system_inodes(struct ocfs2_super *osb)
}
bail:
- mlog_exit(status);
+ if (status)
+ mlog_errno(status);
return status;
}
@@ -534,7 +536,7 @@ static void ocfs2_release_system_inodes(struct ocfs2_super *osb)
}
if (!osb->local_system_inodes)
- goto out;
+ return;
for (i = 0; i < NUM_LOCAL_SYSTEM_INODES * osb->max_slots; i++) {
if (osb->local_system_inodes[i]) {
@@ -545,9 +547,6 @@ static void ocfs2_release_system_inodes(struct ocfs2_super *osb)
kfree(osb->local_system_inodes);
osb->local_system_inodes = NULL;
-
-out:
- mlog_exit(0);
}
/* We're allocating fs objects, use GFP_NOFS */
@@ -1196,7 +1195,6 @@ static int ocfs2_fill_super(struct super_block *sb, void *data, int silent)
mlog_errno(status);
atomic_set(&osb->vol_state, VOLUME_DISABLED);
wake_up(&osb->osb_mount_event);
- mlog_exit(status);
return status;
}
}
@@ -1210,7 +1208,6 @@ static int ocfs2_fill_super(struct super_block *sb, void *data, int silent)
/* Start this when the mount is almost sure of being successful */
ocfs2_orphan_scan_start(osb);
- mlog_exit(status);
return status;
read_super_error:
@@ -1225,7 +1222,8 @@ read_super_error:
ocfs2_dismount_volume(sb, 1);
}
- mlog_exit(status);
+ if (status)
+ mlog_errno(status);
return status;
}
@@ -1514,7 +1512,6 @@ static int ocfs2_parse_options(struct super_block *sb,
status = 1;
bail:
- mlog_exit(status);
return status;
}
@@ -1643,10 +1640,9 @@ leave:
ocfs2_quota_shutdown();
ocfs2_free_mem_caches();
exit_ocfs2_uptodate_cache();
+ mlog_errno(status);
}
- mlog_exit(status);
-
if (status >= 0) {
return register_filesystem(&ocfs2_fs_type);
} else
@@ -1671,8 +1667,6 @@ static void __exit ocfs2_exit(void)
unregister_filesystem(&ocfs2_fs_type);
exit_ocfs2_uptodate_cache();
-
- mlog_exit_void();
}
static void ocfs2_put_super(struct super_block *sb)
@@ -1681,8 +1675,6 @@ static void ocfs2_put_super(struct super_block *sb)
ocfs2_sync_blockdev(sb);
ocfs2_dismount_volume(sb, 0);
-
- mlog_exit_void();
}
static int ocfs2_statfs(struct dentry *dentry, struct kstatfs *buf)
@@ -1741,7 +1733,8 @@ bail:
if (inode)
iput(inode);
- mlog_exit(status);
+ if (status)
+ mlog_errno(status);
return status;
}
@@ -1905,7 +1898,6 @@ leave:
if (unlock_super)
ocfs2_super_unlock(osb, 1);
- mlog_exit(status);
return status;
}
@@ -2354,7 +2346,6 @@ static int ocfs2_initialize_super(struct super_block *sb,
}
bail:
- mlog_exit(status);
return status;
}
@@ -2424,7 +2415,8 @@ static int ocfs2_verify_volume(struct ocfs2_dinode *di,
}
out:
- mlog_exit(status);
+ if (status)
+ mlog_errno(status);
return status;
}
@@ -2519,7 +2511,8 @@ finally:
if (local_alloc)
kfree(local_alloc);
- mlog_exit(status);
+ if (status)
+ mlog_errno(status);
return status;
}
@@ -2548,8 +2541,6 @@ static void ocfs2_delete_osb(struct ocfs2_super *osb)
kfree(osb->uuid_str);
ocfs2_put_dlm_debug(osb->osb_dlm_debug);
memset(osb, 0, sizeof(struct ocfs2_super));
-
- mlog_exit_void();
}
/* Put OCFS2 into a readonly state, or (if the user specifies it),
diff --git a/fs/ocfs2/symlink.c b/fs/ocfs2/symlink.c
index 8a81688..56c1eeb 100644
--- a/fs/ocfs2/symlink.c
+++ b/fs/ocfs2/symlink.c
@@ -72,7 +72,6 @@ static char *ocfs2_fast_symlink_getlink(struct inode *inode,
fe = (struct ocfs2_dinode *) (*bh)->b_data;
link = (char *) fe->id2.i_symlink;
bail:
- mlog_exit(status);
return link;
}
@@ -100,7 +99,8 @@ static int ocfs2_readlink(struct dentry *dentry,
brelse(bh);
out:
- mlog_exit(ret);
+ if (ret)
+ mlog_errno(ret);
return ret;
}
@@ -136,7 +136,8 @@ bail:
nd_set_link(nd, status ? ERR_PTR(status) : link);
brelse(bh);
- mlog_exit(status);
+ if (status)
+ mlog_errno(status);
return NULL;
}
diff --git a/fs/ocfs2/xattr.c b/fs/ocfs2/xattr.c
index 81f98b7..d57c794 100644
--- a/fs/ocfs2/xattr.c
+++ b/fs/ocfs2/xattr.c
@@ -5570,7 +5570,6 @@ static int ocfs2_xattr_set_entry_bucket(struct inode *inode,
out:
- mlog_exit(ret);
return ret;
}
@@ -5637,7 +5636,6 @@ static int ocfs2_xattr_set_entry_index_block(struct inode *inode,
mlog_errno(ret);
out:
- mlog_exit(ret);
return ret;
}
--
1.7.0.4
^ permalink raw reply related [flat|nested] 53+ messages in thread
* [Ocfs2-devel] [PATCH 04/34] ocfs2: Add ocfs2_trace.h.
2010-12-23 7:19 [Ocfs2-devel] [PATCH 00/34] OCFS2: Add trace event and replace mlog(0) Tao Ma
` (2 preceding siblings ...)
2010-12-23 7:30 ` [Ocfs2-devel] [PATCH 03/34] ocfs2: Remove EXIT " Tao Ma
@ 2010-12-23 7:30 ` Tao Ma
2010-12-31 12:40 ` Joel Becker
2010-12-23 7:30 ` [Ocfs2-devel] [PATCH 05/34] ocfs2: Remove mlog(0) from fs/ocfs2/alloc.c Tao Ma
` (34 subsequent siblings)
38 siblings, 1 reply; 53+ messages in thread
From: Tao Ma @ 2010-12-23 7:30 UTC (permalink / raw)
To: ocfs2-devel
From: Wengang Wang <wen.gang.wang@oracle.com>
About one year ago, Wengang Wang tried some first steps
to add tracepoints to ocfs2. But at that time we don't
have ocfs2 1.6 introduced, so we delay this work until now
after ocfs2 1.6 and UEK(Unbreakable Enterprise Kernel) are
released.
Wengang's original patch is here:
http://oss.oracle.com/pipermail/ocfs2-devel/2009-November/005512.html
But as Steven Rostedt indicated in his article
http://lwn.net/Articles/383362/, we'd better have our trace
files resides in fs/ocfs2, so I rewrited the patch using the
method Steven mentioned in that article.
Signed-off-by: Wengang Wang <wen.gang.wang@oracle.com>
Signed-off-by: Tao Ma <tao.ma@oracle.com>
---
fs/ocfs2/ocfs2_trace.h | 16 ++++++++++++++++
fs/ocfs2/super.c | 3 +++
2 files changed, 19 insertions(+), 0 deletions(-)
create mode 100644 fs/ocfs2/ocfs2_trace.h
diff --git a/fs/ocfs2/ocfs2_trace.h b/fs/ocfs2/ocfs2_trace.h
new file mode 100644
index 0000000..8d4e49a
--- /dev/null
+++ b/fs/ocfs2/ocfs2_trace.h
@@ -0,0 +1,16 @@
+#undef TRACE_SYSTEM
+#define TRACE_SYSTEM ocfs2
+
+#if !defined(_TRACE_OCFS2_H) || defined(TRACE_HEADER_MULTI_READ)
+#define _TRACE_OCFS2_H
+
+#include <linux/tracepoint.h>
+
+
+#endif /* _TRACE_OCFS2_H */
+
+/* This part must be outside protection */
+#undef TRACE_INCLUDE_PATH
+#define TRACE_INCLUDE_PATH .
+#define TRACE_INCLUDE_FILE ocfs2_trace
+#include <trace/define_trace.h>
diff --git a/fs/ocfs2/super.c b/fs/ocfs2/super.c
index 9e41894..e88013c 100644
--- a/fs/ocfs2/super.c
+++ b/fs/ocfs2/super.c
@@ -42,6 +42,9 @@
#include <linux/seq_file.h>
#include <linux/quotaops.h>
+#define CREATE_TRACE_POINTS
+#include "ocfs2_trace.h"
+
#define MLOG_MASK_PREFIX ML_SUPER
#include <cluster/masklog.h>
--
1.7.0.4
^ permalink raw reply related [flat|nested] 53+ messages in thread
* [Ocfs2-devel] [PATCH 05/34] ocfs2: Remove mlog(0) from fs/ocfs2/alloc.c
2010-12-23 7:19 [Ocfs2-devel] [PATCH 00/34] OCFS2: Add trace event and replace mlog(0) Tao Ma
` (3 preceding siblings ...)
2010-12-23 7:30 ` [Ocfs2-devel] [PATCH 04/34] ocfs2: Add ocfs2_trace.h Tao Ma
@ 2010-12-23 7:30 ` Tao Ma
2010-12-31 13:15 ` Christoph Hellwig
2010-12-23 7:30 ` [Ocfs2-devel] [PATCH 06/34] ocfs2: Remove mlog(0) from fs/ocfs2/localalloc.c Tao Ma
` (33 subsequent siblings)
38 siblings, 1 reply; 53+ messages in thread
From: Tao Ma @ 2010-12-23 7:30 UTC (permalink / raw)
To: ocfs2-devel
From: Tao Ma <tao.ma@oracle.com>
This is the first try of replacing debug mlog(0,...) to
trace events. Wengang has did some work in his original
patch
http://oss.oracle.com/pipermail/ocfs2-devel/2009-November/005513.html
But he didn't finished it.
So this patch removes all mlog(0,...) from alloc.c and adds
the corresponding tracepoints. Different mlogs have different
solutions.
1. Some are replaced with trace event directly.
2. Some are replaced and some new parameters are added since
I think we need to know the btree owner in that case.
3. Some are combined into one trace events.
4. Some redundant mlogs are removed.
What's more, it defines some event classes so that we can use
them later.
Signed-off-by: Wengang Wang <wen.gang.wang@oracle.com>
Signed-off-by: Tao Ma <tao.ma@oracle.com>
---
fs/ocfs2/alloc.c | 165 +++++++++--------
fs/ocfs2/ocfs2_trace.h | 501 ++++++++++++++++++++++++++++++++++++++++++++++++
2 files changed, 589 insertions(+), 77 deletions(-)
diff --git a/fs/ocfs2/alloc.c b/fs/ocfs2/alloc.c
index 9ab3f27..9bdfaf8 100644
--- a/fs/ocfs2/alloc.c
+++ b/fs/ocfs2/alloc.c
@@ -50,6 +50,7 @@
#include "uptodate.h"
#include "xattr.h"
#include "refcounttree.h"
+#include "ocfs2_trace.h"
#include "buffer_head_io.h"
@@ -886,8 +887,7 @@ static int ocfs2_validate_extent_block(struct super_block *sb,
struct ocfs2_extent_block *eb =
(struct ocfs2_extent_block *)bh->b_data;
- mlog(0, "Validating extent block %llu\n",
- (unsigned long long)bh->b_blocknr);
+ trace_ocfs2_validate_extent_block((unsigned long long)bh->b_blocknr);
BUG_ON(!buffer_uptodate(bh));
@@ -985,7 +985,7 @@ int ocfs2_num_free_extents(struct ocfs2_super *osb,
bail:
brelse(eb_bh);
- mlog(0, "retval = %d\n", retval);
+ trace_ocfs2_num_free_extents(retval);
return retval;
}
@@ -1194,8 +1194,11 @@ static int ocfs2_add_branch(handle_t *handle,
* from new_cpos).
*/
if (root_end > new_cpos) {
- mlog(0, "adjust the cluster end from %u to %u\n",
- root_end, new_cpos);
+ trace_ocfs2_adjust_rightmost_branch(
+ (unsigned long long)
+ ocfs2_metadata_cache_owner(et->et_ci),
+ root_end, new_cpos);
+
status = ocfs2_adjust_rightmost_branch(handle, et);
if (status) {
mlog_errno(status);
@@ -1527,7 +1530,10 @@ static int ocfs2_grow_tree(handle_t *handle, struct ocfs2_extent_tree *et,
* another tree level */
if (shift) {
BUG_ON(bh);
- mlog(0, "need to shift tree depth (current = %d)\n", depth);
+ trace_ocfs2_grow_tree(
+ (unsigned long long)
+ ocfs2_metadata_cache_owner(et->et_ci),
+ depth);
/* ocfs2_shift_tree_depth will return us a buffer with
* the new extent block (so we can pass that to
@@ -1557,7 +1563,6 @@ static int ocfs2_grow_tree(handle_t *handle, struct ocfs2_extent_tree *et,
/* call ocfs2_add_branch to add the final part of the tree with
* the new data. */
- mlog(0, "add branch. bh = %p\n", bh);
ret = ocfs2_add_branch(handle, et, bh, last_eb_bh,
meta_ac);
if (ret < 0) {
@@ -1632,8 +1637,9 @@ static void ocfs2_rotate_leaf(struct ocfs2_extent_list *el,
}
insert_index = i;
- mlog(0, "ins %u: index %d, has_empty %d, next_free %d, count %d\n",
- insert_cpos, insert_index, has_empty, next_free, le16_to_cpu(el->l_count));
+ trace_ocfs2_rotate_leaf(insert_cpos, insert_index,
+ has_empty, next_free,
+ le16_to_cpu(el->l_count));
BUG_ON(insert_index < 0);
BUG_ON(insert_index >= le16_to_cpu(el->l_count));
@@ -2046,7 +2052,7 @@ static void ocfs2_complete_edge_insert(handle_t *handle,
left_el = path_leaf_el(left_path);
right_el = path_leaf_el(right_path);
for(i = left_path->p_tree_depth - 1; i > subtree_index; i--) {
- mlog(0, "Adjust records at index %u\n", i);
+ trace_ocfs2_complete_edge_insert(i);
/*
* One nice property of knowing that all of these
@@ -2376,7 +2382,9 @@ static int ocfs2_rotate_tree_right(handle_t *handle,
goto out;
}
- mlog(0, "Insert: %u, first left path cpos: %u\n", insert_cpos, cpos);
+ trace_ocfs2_rotate_tree_right(
+ (unsigned long long)ocfs2_metadata_cache_owner(et->et_ci),
+ insert_cpos, cpos);
/*
* What we want to do here is:
@@ -2405,8 +2413,10 @@ static int ocfs2_rotate_tree_right(handle_t *handle,
* rotating subtrees.
*/
while (cpos && insert_cpos <= cpos) {
- mlog(0, "Rotating a tree: ins. cpos: %u, left path cpos: %u\n",
- insert_cpos, cpos);
+ trace_ocfs2_rotate_tree_right(
+ (unsigned long long)
+ ocfs2_metadata_cache_owner(et->et_ci),
+ insert_cpos, cpos);
ret = ocfs2_find_path(et->et_ci, left_path, cpos);
if (ret) {
@@ -2448,10 +2458,10 @@ static int ocfs2_rotate_tree_right(handle_t *handle,
start = ocfs2_find_subtree_root(et, left_path, right_path);
- mlog(0, "Subtree root at index %d (blk %llu, depth %d)\n",
- start,
- (unsigned long long) right_path->p_node[start].bh->b_blocknr,
- right_path->p_tree_depth);
+ trace_ocfs2_rotate_subtree(start,
+ (unsigned long long)
+ right_path->p_node[start].bh->b_blocknr,
+ right_path->p_tree_depth);
ret = ocfs2_extend_rotate_transaction(handle, start,
orig_credits, right_path);
@@ -2951,8 +2961,7 @@ static int __ocfs2_rotate_tree_left(handle_t *handle,
subtree_root = ocfs2_find_subtree_root(et, left_path,
right_path);
- mlog(0, "Subtree root at index %d (blk %llu, depth %d)\n",
- subtree_root,
+ trace_ocfs2_rotate_subtree(subtree_root,
(unsigned long long)
right_path->p_node[subtree_root].bh->b_blocknr,
right_path->p_tree_depth);
@@ -3976,9 +3985,11 @@ static int ocfs2_append_rec_to_path(handle_t *handle,
goto out;
}
- mlog(0, "Append may need a left path update. cpos: %u, "
- "left_cpos: %u\n", le32_to_cpu(insert_rec->e_cpos),
- left_cpos);
+ trace_ocfs2_append_rec_to_path(
+ (unsigned long long)
+ ocfs2_metadata_cache_owner(et->et_ci),
+ le32_to_cpu(insert_rec->e_cpos),
+ left_cpos);
/*
* No need to worry if the append is already in the
@@ -4665,9 +4676,9 @@ int ocfs2_insert_extent(handle_t *handle,
struct ocfs2_insert_type insert = {0, };
struct ocfs2_extent_rec rec;
- mlog(0, "add %u clusters at position %u to owner %llu\n",
- new_clusters, cpos,
- (unsigned long long)ocfs2_metadata_cache_owner(et->et_ci));
+ trace_ocfs2_insert_extent_start(
+ (unsigned long long)ocfs2_metadata_cache_owner(et->et_ci),
+ cpos, new_clusters);
memset(&rec, 0, sizeof(rec));
rec.e_cpos = cpu_to_le32(cpos);
@@ -4687,11 +4698,9 @@ int ocfs2_insert_extent(handle_t *handle,
goto bail;
}
- mlog(0, "Insert.appending: %u, Insert.Contig: %u, "
- "Insert.contig_index: %d, Insert.free_records: %d, "
- "Insert.tree_depth: %d\n",
- insert.ins_appending, insert.ins_contig, insert.ins_contig_index,
- free_records, insert.ins_tree_depth);
+ trace_ocfs2_insert_extent(insert.ins_appending, insert.ins_contig,
+ insert.ins_contig_index, free_records,
+ insert.ins_tree_depth);
if (insert.ins_contig == CONTIG_NONE && free_records == 0) {
status = ocfs2_grow_tree(handle, et,
@@ -4732,7 +4741,7 @@ int ocfs2_add_clusters_in_btree(handle_t *handle,
struct ocfs2_alloc_context *meta_ac,
enum ocfs2_alloc_restarted *reason_ret)
{
- int status = 0;
+ int status = 0, err = 0;
int free_extents;
enum ocfs2_alloc_restarted reason = RESTART_NONE;
u32 bit_off, num_bits;
@@ -4759,14 +4768,14 @@ int ocfs2_add_clusters_in_btree(handle_t *handle,
* 2) we are so fragmented, we've needed to add metadata too
* many times. */
if (!free_extents && !meta_ac) {
- mlog(0, "we haven't reserved any metadata!\n");
+ err = -1;
status = -EAGAIN;
reason = RESTART_META;
goto leave;
} else if ((!free_extents)
&& (ocfs2_alloc_context_bits_left(meta_ac)
< ocfs2_extend_meta_needed(et->et_root_el))) {
- mlog(0, "filesystem is really fragmented...\n");
+ err = -2;
status = -EAGAIN;
reason = RESTART_META;
goto leave;
@@ -4791,9 +4800,9 @@ int ocfs2_add_clusters_in_btree(handle_t *handle,
}
block = ocfs2_clusters_to_blocks(osb->sb, bit_off);
- mlog(0, "Allocating %u clusters at block %u for owner %llu\n",
- num_bits, bit_off,
- (unsigned long long)ocfs2_metadata_cache_owner(et->et_ci));
+ trace_ocfs2_add_clusters_in_btree(
+ (unsigned long long)ocfs2_metadata_cache_owner(et->et_ci),
+ bit_off, num_bits);
status = ocfs2_insert_extent(handle, et, *logical_offset, block,
num_bits, flags, meta_ac);
if (status < 0) {
@@ -4807,8 +4816,7 @@ int ocfs2_add_clusters_in_btree(handle_t *handle,
*logical_offset += num_bits;
if (clusters_to_add) {
- mlog(0, "need to alloc once more, wanted = %u\n",
- clusters_to_add);
+ err = clusters_to_add;
status = -EAGAIN;
reason = RESTART_TRANS;
}
@@ -4816,6 +4824,7 @@ int ocfs2_add_clusters_in_btree(handle_t *handle,
leave:
if (reason_ret)
*reason_ret = reason;
+ trace_ocfs2_add_clusters_in_btree_ret(status, reason, err);
return status;
}
@@ -5041,9 +5050,9 @@ int ocfs2_split_extent(handle_t *handle,
ctxt.c_has_empty_extent = ocfs2_is_empty_extent(&el->l_recs[0]);
- mlog(0, "index: %d, contig: %u, has_empty: %u, split_covers: %u\n",
- split_index, ctxt.c_contig_type, ctxt.c_has_empty_extent,
- ctxt.c_split_covers_rec);
+ trace_ocfs2_split_extent(split_index, ctxt.c_contig_type,
+ ctxt.c_has_empty_extent,
+ ctxt.c_split_covers_rec);
if (ctxt.c_contig_type == CONTIG_NONE) {
if (ctxt.c_split_covers_rec)
@@ -5177,8 +5186,9 @@ int ocfs2_mark_extent_written(struct inode *inode,
{
int ret;
- mlog(0, "Inode %lu cpos %u, len %u, phys clusters %u\n",
- inode->i_ino, cpos, len, phys);
+ trace_ocfs2_mark_extent_written(
+ (unsigned long long)OCFS2_I(inode)->ip_blkno,
+ cpos, len, phys);
if (!ocfs2_writes_unwritten_extents(OCFS2_SB(inode->i_sb))) {
ocfs2_error(inode->i_sb, "Inode %llu has unwritten extents "
@@ -5497,11 +5507,10 @@ int ocfs2_remove_extent(handle_t *handle,
BUG_ON(cpos < le32_to_cpu(rec->e_cpos) || trunc_range > rec_range);
- mlog(0, "Owner %llu, remove (cpos %u, len %u). Existing index %d "
- "(cpos %u, len %u)\n",
- (unsigned long long)ocfs2_metadata_cache_owner(et->et_ci),
- cpos, len, index,
- le32_to_cpu(rec->e_cpos), ocfs2_rec_clusters(el, rec));
+ trace_ocfs2_remove_extent(
+ (unsigned long long)ocfs2_metadata_cache_owner(et->et_ci),
+ cpos, len, index, le32_to_cpu(rec->e_cpos),
+ ocfs2_rec_clusters(el, rec));
if (le32_to_cpu(rec->e_cpos) == cpos || rec_range == trunc_range) {
ret = ocfs2_truncate_rec(handle, et, path, index, dealloc,
@@ -5780,9 +5789,6 @@ int ocfs2_truncate_log_append(struct ocfs2_super *osb,
struct ocfs2_dinode *di;
struct ocfs2_truncate_log *tl;
- mlog(0, "start_blk = %llu, num_clusters = %u\n",
- (unsigned long long)start_blk, num_clusters);
-
BUG_ON(mutex_trylock(&tl_inode->i_mutex));
start_cluster = ocfs2_blocks_to_clusters(osb->sb, start_blk);
@@ -5819,10 +5825,9 @@ int ocfs2_truncate_log_append(struct ocfs2_super *osb,
goto bail;
}
- mlog(0, "Log truncate of %u clusters starting at cluster %u to "
- "%llu (index = %d)\n", num_clusters, start_cluster,
- (unsigned long long)OCFS2_I(tl_inode)->ip_blkno, index);
-
+ trace_ocfs2_truncate_log_append(
+ (unsigned long long)OCFS2_I(tl_inode)->ip_blkno, index,
+ start_cluster, num_clusters);
if (ocfs2_truncate_log_can_coalesce(tl, start_cluster)) {
/*
* Move index back to the record we are coalescing with.
@@ -5831,9 +5836,10 @@ int ocfs2_truncate_log_append(struct ocfs2_super *osb,
index--;
num_clusters += le32_to_cpu(tl->tl_recs[index].t_clusters);
- mlog(0, "Coalesce with index %u (start = %u, clusters = %u)\n",
- index, le32_to_cpu(tl->tl_recs[index].t_start),
- num_clusters);
+ trace_ocfs2_truncate_log_append(
+ (unsigned long long)OCFS2_I(tl_inode)->ip_blkno,
+ index, le32_to_cpu(tl->tl_recs[index].t_start),
+ num_clusters);
} else {
tl->tl_recs[index].t_start = cpu_to_le32(start_cluster);
tl->tl_used = cpu_to_le16(index + 1);
@@ -5896,8 +5902,9 @@ static int ocfs2_replay_truncate_records(struct ocfs2_super *osb,
/* if start_blk is not set, we ignore the record as
* invalid. */
if (start_blk) {
- mlog(0, "free record %d, start = %u, clusters = %u\n",
- i, le32_to_cpu(rec.t_start), num_clusters);
+ trace_ocfs2_replay_truncate_records(
+ (unsigned long long)OCFS2_I(tl_inode)->ip_blkno,
+ i, le32_to_cpu(rec.t_start), num_clusters);
status = ocfs2_free_clusters(handle, data_alloc_inode,
data_alloc_bh, start_blk,
@@ -5938,8 +5945,9 @@ int __ocfs2_flush_truncate_log(struct ocfs2_super *osb)
tl = &di->id2.i_dealloc;
num_to_flush = le16_to_cpu(tl->tl_used);
- mlog(0, "Flush %u records from truncate log #%llu\n",
- num_to_flush, (unsigned long long)OCFS2_I(tl_inode)->ip_blkno);
+ trace_ocfs2_flush_truncate_log(
+ (unsigned long long)OCFS2_I(tl_inode)->ip_blkno,
+ num_to_flush);
if (!num_to_flush) {
status = 0;
goto out;
@@ -6076,7 +6084,7 @@ int ocfs2_begin_truncate_log_recovery(struct ocfs2_super *osb,
*tl_copy = NULL;
- mlog(0, "recover truncate log from slot %d\n", slot_num);
+ trace_ocfs2_begin_truncate_log_recovery(slot_num);
status = ocfs2_get_truncate_log_info(osb, slot_num, &tl_inode, &tl_bh);
if (status < 0) {
@@ -6093,8 +6101,7 @@ int ocfs2_begin_truncate_log_recovery(struct ocfs2_super *osb,
tl = &di->id2.i_dealloc;
if (le16_to_cpu(tl->tl_used)) {
- mlog(0, "We'll have %u logs to recover\n",
- le16_to_cpu(tl->tl_used));
+ trace_ocfs2_truncate_log_recovery_num(le16_to_cpu(tl->tl_used));
*tl_copy = kmalloc(tl_bh->b_size, GFP_KERNEL);
if (!(*tl_copy)) {
@@ -6151,8 +6158,9 @@ int ocfs2_complete_truncate_log_recovery(struct ocfs2_super *osb,
tl = &tl_copy->id2.i_dealloc;
num_recs = le16_to_cpu(tl->tl_used);
- mlog(0, "cleanup %u records from %llu\n", num_recs,
- (unsigned long long)le64_to_cpu(tl_copy->i_blkno));
+ trace_ocfs2_complete_truncate_log_recovery(
+ (unsigned long long)le64_to_cpu(tl_copy->i_blkno),
+ num_recs);
mutex_lock(&tl_inode->i_mutex);
for(i = 0; i < num_recs; i++) {
@@ -6310,8 +6318,8 @@ static int ocfs2_free_cached_blocks(struct ocfs2_super *osb,
else
bg_blkno = ocfs2_which_suballoc_group(head->free_blk,
head->free_bit);
- mlog(0, "Free bit: (bit %u, blkno %llu)\n",
- head->free_bit, (unsigned long long)head->free_blk);
+ trace_ocfs2_free_cached_blocks(
+ (unsigned long long)head->free_blk, head->free_bit);
ret = ocfs2_free_suballoc_bits(handle, inode, di_bh,
head->free_bit, bg_blkno, 1);
@@ -6364,8 +6372,7 @@ int ocfs2_cache_cluster_dealloc(struct ocfs2_cached_dealloc_ctxt *ctxt,
return ret;
}
- mlog(0, "Insert clusters: (bit %u, blk %llu)\n",
- bit, (unsigned long long)blkno);
+ trace_ocfs2_cache_cluster_dealloc((unsigned long long)blkno, bit);
item->free_blk = blkno;
item->free_bit = bit;
@@ -6440,8 +6447,8 @@ int ocfs2_run_deallocs(struct ocfs2_super *osb,
fl = ctxt->c_first_suballocator;
if (fl->f_first) {
- mlog(0, "Free items: (type %u, slot %d)\n",
- fl->f_inode_type, fl->f_slot);
+ trace_ocfs2_run_deallocs(fl->f_inode_type,
+ fl->f_slot);
ret2 = ocfs2_free_cached_blocks(osb,
fl->f_inode_type,
fl->f_slot,
@@ -6518,8 +6525,9 @@ int ocfs2_cache_block_dealloc(struct ocfs2_cached_dealloc_ctxt *ctxt,
goto out;
}
- mlog(0, "Insert: (type %d, slot %u, bit %u, blk %llu)\n",
- type, slot, bit, (unsigned long long)blkno);
+ trace_ocfs2_cache_block_dealloc(type, slot,
+ (unsigned long long)suballoc,
+ (unsigned long long)blkno, bit);
item->free_bg = suballoc;
item->free_blk = blkno;
@@ -6999,8 +7007,11 @@ start:
goto bail;
}
- mlog(0, "inode->ip_clusters = %u, tree_depth = %u\n",
- OCFS2_I(inode)->ip_clusters, path->p_tree_depth);
+ trace_ocfs2_commit_truncate(
+ (unsigned long long)OCFS2_I(inode)->ip_blkno,
+ new_highest_cpos,
+ OCFS2_I(inode)->ip_clusters,
+ path->p_tree_depth);
/*
* By now, el will point to the extent list on the bottom most
diff --git a/fs/ocfs2/ocfs2_trace.h b/fs/ocfs2/ocfs2_trace.h
index 8d4e49a..2dfc3b0 100644
--- a/fs/ocfs2/ocfs2_trace.h
+++ b/fs/ocfs2/ocfs2_trace.h
@@ -6,6 +6,507 @@
#include <linux/tracepoint.h>
+DECLARE_EVENT_CLASS(ocfs2__int,
+
+ TP_PROTO(int num),
+
+ TP_ARGS(num),
+
+ TP_STRUCT__entry(
+ __field( int, num )
+ ),
+
+ TP_fast_assign(
+ __entry->num = num;
+ ),
+
+ TP_printk("%d", __entry->num)
+);
+
+DECLARE_EVENT_CLASS(ocfs2__int_int,
+
+ TP_PROTO(int value1, int value2),
+
+ TP_ARGS(value1, value2),
+
+ TP_STRUCT__entry(
+ __field( int, value1 )
+ __field( int, value2 )
+ ),
+
+ TP_fast_assign(
+ __entry->value1 = value1;
+ __entry->value2 = value2;
+ ),
+
+ TP_printk("%d %d", __entry->value1, __entry->value2)
+);
+
+DECLARE_EVENT_CLASS(ocfs2__ull_uint,
+
+ TP_PROTO(unsigned long long value1, unsigned int value2),
+
+ TP_ARGS(value1, value2),
+
+ TP_STRUCT__entry(
+ __field( unsigned long long, value1 )
+ __field( unsigned int, value2 )
+ ),
+
+ TP_fast_assign(
+ __entry->value1 = value1;
+ __entry->value2 = value2;
+ ),
+
+ TP_printk("%llu %u", __entry->value1, __entry->value2)
+);
+
+/* Trace events for fs/ocfs2/alloc.c. */
+DECLARE_EVENT_CLASS(ocfs2__btree_ops,
+
+ TP_PROTO(unsigned long long owner,\
+ unsigned int value1, unsigned int value2),
+
+ TP_ARGS(owner, value1, value2),
+
+ TP_STRUCT__entry(
+ __field( unsigned long long, owner )
+ __field( unsigned int, value1 )
+ __field( unsigned int, value2 )
+ ),
+
+ TP_fast_assign(
+ __entry->owner = owner;
+ __entry->value1 = value1;
+ __entry->value2 = value2;
+ ),
+
+ TP_printk("%llu %u %u",
+ __entry->owner, __entry->value1, __entry->value2)
+);
+
+DEFINE_EVENT(ocfs2__btree_ops, ocfs2_adjust_rightmost_branch,
+
+ TP_PROTO(unsigned long long owner,
+ unsigned int root_end, unsigned int new_cpos),
+
+ TP_ARGS(owner, root_end, new_cpos)
+);
+
+DEFINE_EVENT(ocfs2__int, ocfs2_num_free_extents,
+
+ TP_PROTO(int retval),
+
+ TP_ARGS(retval)
+);
+
+DEFINE_EVENT(ocfs2__int, ocfs2_complete_edge_insert,
+
+ TP_PROTO(int index),
+
+ TP_ARGS(index)
+);
+
+TRACE_EVENT(ocfs2_grow_tree,
+
+ TP_PROTO(unsigned long long owner, int depth),
+
+ TP_ARGS(owner, depth),
+
+ TP_STRUCT__entry(
+ __field( unsigned long long, owner )
+ __field( int, depth )
+ ),
+
+ TP_fast_assign(
+ __entry->owner = owner;
+ __entry->depth = depth;
+ ),
+
+ TP_printk("%llu %d", __entry->owner, __entry->depth)
+);
+
+DEFINE_EVENT(ocfs2__btree_ops, ocfs2_rotate_tree_right,
+
+ TP_PROTO(unsigned long long owner,
+ unsigned int insert_cpos, unsigned int cpos),
+
+ TP_ARGS(owner, insert_cpos, cpos)
+);
+
+DEFINE_EVENT(ocfs2__btree_ops, ocfs2_append_rec_to_path,
+
+ TP_PROTO(unsigned long long owner,
+ unsigned int cpos, unsigned int left_cpos),
+
+ TP_ARGS(owner, cpos, left_cpos)
+);
+
+DEFINE_EVENT(ocfs2__btree_ops, ocfs2_insert_extent_start,
+
+ TP_PROTO(unsigned long long owner,
+ unsigned int cpos, unsigned int clusters),
+
+ TP_ARGS(owner, cpos, clusters)
+);
+
+TRACE_EVENT(ocfs2_rotate_subtree,
+
+ TP_PROTO(int subtree_root, unsigned long long blkno,
+ int depth),
+
+ TP_ARGS(subtree_root, blkno, depth),
+
+ TP_STRUCT__entry(
+ __field( int, subtree_root )
+ __field( unsigned long long, blkno )
+ __field( int, depth )
+ ),
+
+ TP_fast_assign(
+ __entry->subtree_root = subtree_root;
+ __entry->blkno = blkno;
+ __entry->depth = depth;
+ ),
+
+ TP_printk("%d %llu %d", __entry->subtree_root,
+ __entry->blkno, __entry->depth)
+);
+
+TRACE_EVENT(ocfs2_insert_extent,
+
+ TP_PROTO(unsigned int ins_appending, unsigned int ins_contig,
+ int ins_contig_index, int free_records, int ins_tree_depth),
+
+ TP_ARGS(ins_appending, ins_contig, ins_contig_index, free_records,
+ ins_tree_depth),
+
+ TP_STRUCT__entry(
+ __field( unsigned int, ins_appending )
+ __field( unsigned int, ins_contig )
+ __field( int, ins_contig_index )
+ __field( int, free_records )
+ __field( int, ins_tree_depth )
+ ),
+
+ TP_fast_assign(
+ __entry->ins_appending = ins_appending;
+ __entry->ins_contig = ins_contig;
+ __entry->ins_contig_index = ins_contig_index;
+ __entry->free_records = free_records;
+ __entry->ins_tree_depth = ins_tree_depth;
+ ),
+
+ TP_printk("%u %u %d %d %d",
+ __entry->ins_appending, __entry->ins_contig,
+ __entry->ins_contig_index, __entry->free_records,
+ __entry->ins_tree_depth)
+);
+
+TRACE_EVENT(ocfs2_split_extent,
+
+ TP_PROTO(int split_index, unsigned int c_contig_type,
+ unsigned int c_has_empty_extent,
+ unsigned int c_split_covers_rec),
+
+ TP_ARGS(split_index, c_contig_type,
+ c_has_empty_extent, c_split_covers_rec),
+
+ TP_STRUCT__entry(
+ __field( int, split_index )
+ __field( unsigned int, c_contig_type )
+ __field( unsigned int, c_has_empty_extent )
+ __field( unsigned int, c_split_covers_rec )
+ ),
+
+ TP_fast_assign(
+ __entry->split_index = split_index;
+ __entry->c_contig_type = c_contig_type;
+ __entry->c_has_empty_extent = c_has_empty_extent;
+ __entry->c_split_covers_rec = c_split_covers_rec;
+ ),
+
+ TP_printk("%d %u %u %u", __entry->split_index, __entry->c_contig_type,
+ __entry->c_has_empty_extent, __entry->c_split_covers_rec)
+);
+
+TRACE_EVENT(ocfs2_remove_extent,
+
+ TP_PROTO(unsigned long long owner, unsigned int cpos,
+ unsigned int len, int index,
+ unsigned int e_cpos, unsigned int clusters),
+
+ TP_ARGS(owner, cpos, len, index, e_cpos, clusters),
+
+ TP_STRUCT__entry(
+ __field( unsigned long long, owner )
+ __field( unsigned int, cpos )
+ __field( unsigned int, len )
+ __field( int, index )
+ __field( unsigned int, e_cpos )
+ __field( unsigned int, clusters)
+ ),
+
+ TP_fast_assign(
+ __entry->owner = owner;
+ __entry->cpos = cpos;
+ __entry->len = len;
+ __entry->index = index;
+ __entry->e_cpos = e_cpos;
+ __entry->clusters = clusters;
+ ),
+
+ TP_printk("%llu %u %u %d %u %u",
+ __entry->owner, __entry->cpos, __entry->len, __entry->index,
+ __entry->e_cpos, __entry->clusters)
+);
+
+TRACE_EVENT(ocfs2_commit_truncate,
+
+ TP_PROTO(unsigned long long ino, unsigned int new_cpos,
+ unsigned int clusters, unsigned int depth),
+
+ TP_ARGS(ino, new_cpos, clusters, depth),
+
+ TP_STRUCT__entry(
+ __field( unsigned long long, ino )
+ __field( unsigned int, new_cpos)
+ __field( unsigned int, clusters)
+ __field( unsigned int, depth )
+ ),
+
+ TP_fast_assign(
+ __entry->ino = ino;
+ __entry->new_cpos = new_cpos;
+ __entry->clusters = clusters;
+ __entry->depth = depth;
+ ),
+
+ TP_printk("%llu %u %u %u",
+ __entry->ino, __entry->new_cpos,
+ __entry->clusters, __entry->depth)
+);
+
+TRACE_EVENT(ocfs2_validate_extent_block,
+
+ TP_PROTO(unsigned long long blkno),
+
+ TP_ARGS(blkno),
+
+ TP_STRUCT__entry(
+ __field( unsigned long long, blkno )
+ ),
+
+ TP_fast_assign(
+ __entry->blkno = blkno;
+ ),
+
+ TP_printk("%llu ", __entry->blkno)
+);
+
+TRACE_EVENT(ocfs2_rotate_leaf,
+
+ TP_PROTO(unsigned int insert_cpos, int insert_index,
+ int has_empty, int next_free,
+ unsigned int l_count),
+
+ TP_ARGS(insert_cpos, insert_index, has_empty,
+ next_free, l_count),
+
+ TP_STRUCT__entry(
+ __field( unsigned int, insert_cpos )
+ __field( int, insert_index )
+ __field( int, has_empty )
+ __field( int, next_free )
+ __field( unsigned int, l_count )
+ ),
+
+ TP_fast_assign(
+ __entry->insert_cpos = insert_cpos;
+ __entry->insert_index = insert_index;
+ __entry->has_empty = has_empty;
+ __entry->next_free = next_free;
+ __entry->l_count = l_count;
+ ),
+
+ TP_printk("%u %d %d %d %u", __entry->insert_cpos,
+ __entry->insert_index, __entry->has_empty,
+ __entry->next_free, __entry->l_count)
+);
+
+DEFINE_EVENT(ocfs2__btree_ops, ocfs2_add_clusters_in_btree,
+
+ TP_PROTO(unsigned long long owner,
+ unsigned int cpos, unsigned int clusters),
+
+ TP_ARGS(owner, cpos, clusters)
+);
+
+TRACE_EVENT(ocfs2_add_clusters_in_btree_ret,
+
+ TP_PROTO(int status, int reason, int err),
+
+ TP_ARGS(status, reason, err),
+
+ TP_STRUCT__entry(
+ __field( int, status )
+ __field( int, reason )
+ __field( int, err )
+ ),
+
+ TP_fast_assign(
+ __entry->status = status;
+ __entry->reason = reason;
+ __entry->err = err;
+ ),
+
+ TP_printk("%d %d %d", __entry->status,
+ __entry->reason, __entry->err)
+);
+
+TRACE_EVENT(ocfs2_mark_extent_written,
+
+ TP_PROTO(unsigned long long owner, unsigned int cpos,
+ unsigned int len, unsigned int phys),
+
+ TP_ARGS(owner, cpos, len, phys),
+
+ TP_STRUCT__entry(
+ __field( unsigned long long, owner )
+ __field( unsigned int, cpos )
+ __field( unsigned int, len )
+ __field( unsigned int, phys )
+ ),
+
+ TP_fast_assign(
+ __entry->owner = owner;
+ __entry->cpos = cpos;
+ __entry->len = len;
+ __entry->phys = phys;
+ ),
+
+ TP_printk("%llu %u %u %u",
+ __entry->owner, __entry->cpos, __entry->len, __entry->phys)
+);
+
+DECLARE_EVENT_CLASS(ocfs2__truncate_log_ops,
+
+ TP_PROTO(unsigned long long blkno, int index,
+ unsigned int start, unsigned int num),
+
+ TP_ARGS(blkno, index, start, num),
+
+ TP_STRUCT__entry(
+ __field( unsigned long long, blkno )
+ __field( int, index )
+ __field( unsigned int, start )
+ __field( unsigned int, num )
+ ),
+
+ TP_fast_assign(
+ __entry->blkno = blkno;
+ __entry->index = index;
+ __entry->start = start;
+ __entry->num = num;
+ ),
+ TP_printk("%llu %d %u %u",
+ __entry->blkno, __entry->index,
+ __entry->start, __entry->num)
+);
+
+DEFINE_EVENT(ocfs2__truncate_log_ops, ocfs2_truncate_log_append,
+
+ TP_PROTO(unsigned long long blkno, int index,
+ unsigned int start, unsigned int num),
+
+ TP_ARGS(blkno, index, start, num)
+);
+
+DEFINE_EVENT(ocfs2__truncate_log_ops, ocfs2_replay_truncate_records,
+
+ TP_PROTO(unsigned long long blkno, int index,
+ unsigned int start, unsigned int num),
+
+ TP_ARGS(blkno, index, start, num)
+);
+
+DEFINE_EVENT(ocfs2__ull_uint, ocfs2_flush_truncate_log,
+
+ TP_PROTO(unsigned long long owner, unsigned int num_to_flush),
+
+ TP_ARGS(owner, num_to_flush)
+);
+
+DEFINE_EVENT(ocfs2__int, ocfs2_begin_truncate_log_recovery,
+
+ TP_PROTO(int slot),
+
+ TP_ARGS(slot)
+);
+
+DEFINE_EVENT(ocfs2__int, ocfs2_truncate_log_recovery_num,
+
+ TP_PROTO(int num),
+
+ TP_ARGS(num)
+);
+
+DEFINE_EVENT(ocfs2__ull_uint, ocfs2_complete_truncate_log_recovery,
+
+ TP_PROTO(unsigned long long owner, unsigned int num_flush),
+
+ TP_ARGS(owner, num_flush)
+);
+
+DEFINE_EVENT(ocfs2__ull_uint, ocfs2_free_cached_blocks,
+
+ TP_PROTO(unsigned long long blkno, unsigned int bit),
+
+ TP_ARGS(blkno, bit)
+);
+
+DEFINE_EVENT(ocfs2__ull_uint, ocfs2_cache_cluster_dealloc,
+
+ TP_PROTO(unsigned long long blkno, unsigned int bit),
+
+ TP_ARGS(blkno, bit)
+);
+
+DEFINE_EVENT(ocfs2__int_int, ocfs2_run_deallocs,
+
+ TP_PROTO(int inode_type, int slot),
+
+ TP_ARGS(inode_type, slot)
+);
+
+TRACE_EVENT(ocfs2_cache_block_dealloc,
+
+ TP_PROTO(int type, int slot, unsigned long long suballoc,
+ unsigned long long blkno, unsigned int bit),
+
+ TP_ARGS(type, slot, suballoc, blkno, bit),
+
+ TP_STRUCT__entry(
+ __field( int, type )
+ __field( int, slot )
+ __field( unsigned long long, suballoc )
+ __field( unsigned long long, blkno )
+ __field( unsigned int, bit )
+ ),
+
+ TP_fast_assign(
+ __entry->type = type;
+ __entry->slot = slot;
+ __entry->suballoc = suballoc;
+ __entry->blkno = blkno;
+ __entry->bit = bit;
+ ),
+
+ TP_printk("%d %d %llu %llu %u",
+ __entry->type, __entry->slot, __entry->suballoc,
+ __entry->blkno, __entry->bit)
+);
+/* End of trace events for fs/ocfs2/alloc.c. */
#endif /* _TRACE_OCFS2_H */
--
1.7.0.4
^ permalink raw reply related [flat|nested] 53+ messages in thread
* [Ocfs2-devel] [PATCH 06/34] ocfs2: Remove mlog(0) from fs/ocfs2/localalloc.c
2010-12-23 7:19 [Ocfs2-devel] [PATCH 00/34] OCFS2: Add trace event and replace mlog(0) Tao Ma
` (4 preceding siblings ...)
2010-12-23 7:30 ` [Ocfs2-devel] [PATCH 05/34] ocfs2: Remove mlog(0) from fs/ocfs2/alloc.c Tao Ma
@ 2010-12-23 7:30 ` Tao Ma
2010-12-23 7:30 ` [Ocfs2-devel] [PATCH 07/34] ocfs2: Remove mlog(0) from fs/ocfs2/suballoc.c Tao Ma
` (32 subsequent siblings)
38 siblings, 0 replies; 53+ messages in thread
From: Tao Ma @ 2010-12-23 7:30 UTC (permalink / raw)
To: ocfs2-devel
From: Tao Ma <tao.ma@oracle.com>
This is the 2nd step to remove the debug info of DISK_ALLOC.
So this patch removes all mlog(0,...) from localalloc.c and adds
the corresponding tracepoints. Different mlogs have different
solutions.
1. Some are replaced with trace event directly.
2. Some are replaced while some new parameters are added.
3. Some are combined into one trace events.
4. Some redundant mlogs are removed.
Signed-off-by: Tao Ma <tao.ma@oracle.com>
---
fs/ocfs2/localalloc.c | 57 +++++++----------
fs/ocfs2/ocfs2_trace.h | 172 ++++++++++++++++++++++++++++++++++++++++++++++++
2 files changed, 195 insertions(+), 34 deletions(-)
diff --git a/fs/ocfs2/localalloc.c b/fs/ocfs2/localalloc.c
index 36636e1..a311b1f 100644
--- a/fs/ocfs2/localalloc.c
+++ b/fs/ocfs2/localalloc.c
@@ -43,6 +43,7 @@
#include "suballoc.h"
#include "super.h"
#include "sysfile.h"
+#include "ocfs2_trace.h"
#include "buffer_head_io.h"
@@ -201,8 +202,7 @@ void ocfs2_la_set_sizes(struct ocfs2_super *osb, int requested_mb)
la_max_mb = ocfs2_clusters_to_megabytes(sb,
ocfs2_local_alloc_size(sb) * 8);
- mlog(0, "requested: %dM, max: %uM, default: %uM\n",
- requested_mb, la_max_mb, la_default_mb);
+ trace_ocfs2_la_set_sizes(requested_mb, la_max_mb, la_default_mb);
if (requested_mb == -1) {
/* No user request - use defaults */
@@ -276,8 +276,8 @@ int ocfs2_alloc_should_use_local(struct ocfs2_super *osb, u64 bits)
ret = 1;
bail:
- mlog(0, "state=%d, bits=%llu, la_bits=%d, ret=%d\n",
- osb->local_alloc_state, (unsigned long long)bits, la_bits, ret);
+ trace_ocfs2_alloc_should_use_local(
+ (unsigned long long)bits, osb->local_alloc_state, la_bits, ret);
spin_unlock(&osb->osb_lock);
return ret;
}
@@ -362,7 +362,7 @@ bail:
if (inode)
iput(inode);
- mlog(0, "Local alloc window bits = %d\n", osb->local_alloc_bits);
+ trace_ocfs2_load_local_alloc(osb->local_alloc_bits);
if (status)
mlog_errno(status);
@@ -497,7 +497,7 @@ int ocfs2_begin_local_alloc_recovery(struct ocfs2_super *osb,
struct inode *inode = NULL;
struct ocfs2_dinode *alloc;
- mlog(0, "(slot_num = %d)\n", slot_num);
+ trace_ocfs2_begin_local_alloc_recovery(slot_num);
*alloc_copy = NULL;
@@ -705,10 +705,6 @@ int ocfs2_reserve_local_alloc_bits(struct ocfs2_super *osb,
goto bail;
}
- if (ac->ac_max_block)
- mlog(0, "Calling in_range for max block %llu\n",
- (unsigned long long)ac->ac_max_block);
-
ac->ac_inode = local_alloc_inode;
/* We should never use localalloc from another slot */
ac->ac_alloc_slot = osb->slot_num;
@@ -722,8 +718,9 @@ bail:
iput(local_alloc_inode);
}
- mlog(0, "bits=%d, slot=%d, ret=%d\n", bits_wanted, osb->slot_num,
- status);
+ trace_ocfs2_reserve_local_alloc_bits(
+ (unsigned long long)ac->ac_max_block,
+ bits_wanted, osb->slot_num, status);
if (status)
mlog_errno(status);
@@ -797,7 +794,7 @@ static u32 ocfs2_local_alloc_count_bits(struct ocfs2_dinode *alloc)
for (i = 0; i < le16_to_cpu(la->la_size); i++)
count += hweight8(buffer[i]);
- mlog(0, "count %u\n", count);
+ trace_ocfs2_local_alloc_count_bits(count);
return count;
}
@@ -812,10 +809,7 @@ static int ocfs2_local_alloc_find_clear_bits(struct ocfs2_super *osb,
void *bitmap = NULL;
struct ocfs2_reservation_map *resmap = &osb->osb_la_resmap;
- mlog(0, "(numbits wanted = %u)\n", *numbits);
-
if (!alloc->id1.bitmap1.i_total) {
- mlog(0, "No bits in my window!\n");
bitoff = -1;
goto bail;
}
@@ -875,8 +869,7 @@ static int ocfs2_local_alloc_find_clear_bits(struct ocfs2_super *osb,
}
}
- mlog(0, "Exiting loop, bitoff = %d, numfound = %d\n", bitoff,
- numfound);
+ trace_ocfs2_local_alloc_find_clear_bits_search_bitmap(bitoff, numfound);
if (numfound == *numbits)
bitoff = startoff - numfound;
@@ -887,7 +880,10 @@ bail:
if (local_resv)
ocfs2_resv_discard(resmap, resv);
- mlog(0, "bitoff %d\n", bitoff);
+ trace_ocfs2_local_alloc_find_clear_bits(*numbits,
+ le32_to_cpu(alloc->id1.bitmap1.i_total),
+ bitoff, numfound);
+
return bitoff;
}
@@ -941,18 +937,16 @@ static int ocfs2_sync_local_to_main(struct ocfs2_super *osb,
void *bitmap;
struct ocfs2_local_alloc *la = OCFS2_LOCAL_ALLOC(alloc);
- mlog(0, "total = %u, used = %u\n",
+ trace_ocfs2_sync_local_to_main(
le32_to_cpu(alloc->id1.bitmap1.i_total),
le32_to_cpu(alloc->id1.bitmap1.i_used));
if (!alloc->id1.bitmap1.i_total) {
- mlog(0, "nothing to sync!\n");
goto bail;
}
if (le32_to_cpu(alloc->id1.bitmap1.i_used) ==
le32_to_cpu(alloc->id1.bitmap1.i_total)) {
- mlog(0, "all bits were taken!\n");
goto bail;
}
@@ -974,8 +968,7 @@ static int ocfs2_sync_local_to_main(struct ocfs2_super *osb,
ocfs2_clusters_to_blocks(osb->sb,
start - count);
- mlog(0, "freeing %u bits starting at local alloc bit "
- "%u (la_start_blk = %llu, blkno = %llu)\n",
+ trace_ocfs2_sync_local_to_main_free(
count, start - count,
(unsigned long long)la_start_blk,
(unsigned long long)blkno);
@@ -1142,12 +1135,9 @@ static int ocfs2_local_alloc_new_window(struct ocfs2_super *osb,
alloc = (struct ocfs2_dinode *) osb->local_alloc_bh->b_data;
la = OCFS2_LOCAL_ALLOC(alloc);
- if (alloc->id1.bitmap1.i_total)
- mlog(0, "asking me to alloc a new window over a non-empty "
- "one\n");
-
- mlog(0, "Allocating %u clusters for a new window.\n",
- osb->local_alloc_bits);
+ trace_ocfs2_local_alloc_new_window(
+ le32_to_cpu(alloc->id1.bitmap1.i_total),
+ osb->local_alloc_bits);
/* Instruct the allocation code to try the most recently used
* cluster group. We'll re-record the group used this pass
@@ -1209,10 +1199,9 @@ retry_enospc:
ocfs2_resmap_restart(&osb->osb_la_resmap, cluster_count,
OCFS2_LOCAL_ALLOC(alloc)->la_bitmap);
- mlog(0, "New window allocated:\n");
- mlog(0, "window la_bm_off = %u\n",
- OCFS2_LOCAL_ALLOC(alloc)->la_bm_off);
- mlog(0, "window bits = %u\n", le32_to_cpu(alloc->id1.bitmap1.i_total));
+ trace_ocfs2_local_alloc_new_window_result(
+ OCFS2_LOCAL_ALLOC(alloc)->la_bm_off,
+ le32_to_cpu(alloc->id1.bitmap1.i_total));
bail:
if (status)
diff --git a/fs/ocfs2/ocfs2_trace.h b/fs/ocfs2/ocfs2_trace.h
index 2dfc3b0..b94458b 100644
--- a/fs/ocfs2/ocfs2_trace.h
+++ b/fs/ocfs2/ocfs2_trace.h
@@ -23,6 +23,23 @@ DECLARE_EVENT_CLASS(ocfs2__int,
TP_printk("%d", __entry->num)
);
+DECLARE_EVENT_CLASS(ocfs2__uint,
+
+ TP_PROTO(unsigned int num),
+
+ TP_ARGS(num),
+
+ TP_STRUCT__entry(
+ __field( unsigned int, num )
+ ),
+
+ TP_fast_assign(
+ __entry->num = num;
+ ),
+
+ TP_printk("%u", __entry->num)
+);
+
DECLARE_EVENT_CLASS(ocfs2__int_int,
TP_PROTO(int value1, int value2),
@@ -61,6 +78,53 @@ DECLARE_EVENT_CLASS(ocfs2__ull_uint,
TP_printk("%llu %u", __entry->value1, __entry->value2)
);
+DECLARE_EVENT_CLASS(ocfs2__uint_uint_uint,
+
+ TP_PROTO(unsigned int value1, unsigned int value2,
+ unsigned int value3),
+
+ TP_ARGS(value1, value2, value3),
+
+ TP_STRUCT__entry(
+ __field( unsigned int, value1 )
+ __field( unsigned int, value2 )
+ __field( unsigned int, value3 )
+ ),
+
+ TP_fast_assign(
+ __entry->value1 = value1;
+ __entry->value2 = value2;
+ __entry->value3 = value3;
+ ),
+
+ TP_printk("%u %u %u", __entry->value1, __entry->value2, __entry->value3)
+);
+
+DECLARE_EVENT_CLASS(ocfs2__ull_int_int_int,
+
+ TP_PROTO(unsigned long long ull, int value1, int value2, int value3),
+
+ TP_ARGS(ull, value1, value2, value3),
+
+ TP_STRUCT__entry(
+ __field( unsigned long long, ull )
+ __field( int, value1 )
+ __field( int, value2 )
+ __field( int, value3 )
+ ),
+
+ TP_fast_assign(
+ __entry->ull = ull;
+ __entry->value1 = value1;
+ __entry->value2 = value2;
+ __entry->value3 = value3;
+ ),
+
+ TP_printk("%llu %d %d %d",
+ __entry->ull, __entry->value1,
+ __entry->value2, __entry->value3)
+);
+
/* Trace events for fs/ocfs2/alloc.c. */
DECLARE_EVENT_CLASS(ocfs2__btree_ops,
@@ -508,6 +572,114 @@ TRACE_EVENT(ocfs2_cache_block_dealloc,
);
/* End of trace events for fs/ocfs2/alloc.c. */
+/* Trace events for fs/ocfs2/localalloc.c. */
+DEFINE_EVENT(ocfs2__uint_uint_uint, ocfs2_la_set_sizes,
+
+ TP_PROTO(unsigned int requested_mb, unsigned int max_mb,
+ unsigned int default_mb),
+
+ TP_ARGS(requested_mb, max_mb, default_mb)
+);
+
+DEFINE_EVENT(ocfs2__ull_int_int_int, ocfs2_alloc_should_use_local,
+
+ TP_PROTO(unsigned long long bit, int state, int la_bits, int ret),
+
+ TP_ARGS(bit, state, la_bits, ret)
+);
+
+DEFINE_EVENT(ocfs2__int, ocfs2_load_local_alloc,
+
+ TP_PROTO(int bit),
+
+ TP_ARGS(bit)
+);
+
+DEFINE_EVENT(ocfs2__int, ocfs2_begin_local_alloc_recovery,
+
+ TP_PROTO(int slot),
+
+ TP_ARGS(slot)
+);
+
+DEFINE_EVENT(ocfs2__ull_int_int_int, ocfs2_reserve_local_alloc_bits,
+
+ TP_PROTO(unsigned long long max_block, int bits_wanted,
+ int slot, int status),
+
+ TP_ARGS(max_block, bits_wanted, slot, status)
+);
+
+DEFINE_EVENT(ocfs2__uint, ocfs2_local_alloc_count_bits,
+
+ TP_PROTO(uint count),
+
+ TP_ARGS(count)
+);
+
+DEFINE_EVENT(ocfs2__int_int, ocfs2_local_alloc_find_clear_bits_search_bitmap,
+
+ TP_PROTO(int bitoff, int numfound),
+
+ TP_ARGS(bitoff, numfound)
+);
+
+DEFINE_EVENT(ocfs2__ull_int_int_int, ocfs2_local_alloc_find_clear_bits,
+
+ TP_PROTO(unsigned long long numbits,
+ int total, int bitoff, int numfound),
+
+ TP_ARGS(numbits, total, bitoff, numfound)
+);
+
+DEFINE_EVENT(ocfs2__int_int, ocfs2_sync_local_to_main,
+
+ TP_PROTO(int total, int used),
+
+ TP_ARGS(total, used)
+);
+
+TRACE_EVENT(ocfs2_sync_local_to_main_free,
+
+ TP_PROTO(int count, int bit, unsigned long long start_blk,
+ unsigned long long blkno),
+
+ TP_ARGS(count, bit, start_blk, blkno),
+
+ TP_STRUCT__entry(
+ __field( int, count )
+ __field( int, bit )
+ __field( unsigned long long, start_blk )
+ __field( unsigned long long, blkno )
+ ),
+
+ TP_fast_assign(
+ __entry->count = count;
+ __entry->bit = bit;
+ __entry->start_blk = start_blk;
+ __entry->blkno = blkno;
+ ),
+
+ TP_printk("%d %d %llu %llu",
+ __entry->count, __entry->bit, __entry->start_blk,
+ __entry->blkno)
+);
+
+DEFINE_EVENT(ocfs2__int_int, ocfs2_local_alloc_new_window,
+
+ TP_PROTO(int total, int alloc_bits),
+
+ TP_ARGS(total, alloc_bits)
+);
+
+DEFINE_EVENT(ocfs2__ull_uint, ocfs2_local_alloc_new_window_result,
+
+ TP_PROTO(unsigned long long bm_off, unsigned int total),
+
+ TP_ARGS(bm_off, total)
+);
+
+/* End of trace events for fs/ocfs2/localalloc.c. */
#endif /* _TRACE_OCFS2_H */
/* This part must be outside protection */
--
1.7.0.4
^ permalink raw reply related [flat|nested] 53+ messages in thread
* [Ocfs2-devel] [PATCH 07/34] ocfs2: Remove mlog(0) from fs/ocfs2/suballoc.c
2010-12-23 7:19 [Ocfs2-devel] [PATCH 00/34] OCFS2: Add trace event and replace mlog(0) Tao Ma
` (5 preceding siblings ...)
2010-12-23 7:30 ` [Ocfs2-devel] [PATCH 06/34] ocfs2: Remove mlog(0) from fs/ocfs2/localalloc.c Tao Ma
@ 2010-12-23 7:30 ` Tao Ma
2010-12-23 7:30 ` [Ocfs2-devel] [PATCH 08/34] " Tao Ma
` (31 subsequent siblings)
38 siblings, 0 replies; 53+ messages in thread
From: Tao Ma @ 2010-12-23 7:30 UTC (permalink / raw)
To: ocfs2-devel
From: Tao Ma <tao.ma@oracle.com>
This is the 3rd step to remove the debug info of DISK_ALLOC.
Signed-off-by: Tao Ma <tao.ma@oracle.com>
---
fs/ocfs2/ocfs2_trace.h | 329 +++++++++++++++++++++++++++++++++++++++++++++++-
fs/ocfs2/suballoc.c | 107 ++++++++--------
2 files changed, 381 insertions(+), 55 deletions(-)
diff --git a/fs/ocfs2/ocfs2_trace.h b/fs/ocfs2/ocfs2_trace.h
index b94458b..196a995 100644
--- a/fs/ocfs2/ocfs2_trace.h
+++ b/fs/ocfs2/ocfs2_trace.h
@@ -40,6 +40,23 @@ DECLARE_EVENT_CLASS(ocfs2__uint,
TP_printk("%u", __entry->num)
);
+DECLARE_EVENT_CLASS(ocfs2__ull,
+
+ TP_PROTO(unsigned long long blkno),
+
+ TP_ARGS(blkno),
+
+ TP_STRUCT__entry(
+ __field( unsigned long long, blkno )
+ ),
+
+ TP_fast_assign(
+ __entry->blkno = blkno;
+ ),
+
+ TP_printk("%llu", __entry->blkno)
+);
+
DECLARE_EVENT_CLASS(ocfs2__int_int,
TP_PROTO(int value1, int value2),
@@ -59,6 +76,25 @@ DECLARE_EVENT_CLASS(ocfs2__int_int,
TP_printk("%d %d", __entry->value1, __entry->value2)
);
+DECLARE_EVENT_CLASS(ocfs2__uint_uint,
+
+ TP_PROTO(unsigned int value1, unsigned int value2),
+
+ TP_ARGS(value1, value2),
+
+ TP_STRUCT__entry(
+ __field( unsigned int, value1 )
+ __field( unsigned int, value2 )
+ ),
+
+ TP_fast_assign(
+ __entry->value1 = value1;
+ __entry->value2 = value2;
+ ),
+
+ TP_printk("%u %u", __entry->value1, __entry->value2)
+);
+
DECLARE_EVENT_CLASS(ocfs2__ull_uint,
TP_PROTO(unsigned long long value1, unsigned int value2),
@@ -78,6 +114,48 @@ DECLARE_EVENT_CLASS(ocfs2__ull_uint,
TP_printk("%llu %u", __entry->value1, __entry->value2)
);
+DECLARE_EVENT_CLASS(ocfs2__ull_ull,
+
+ TP_PROTO(unsigned long long value1, unsigned long long value2),
+
+ TP_ARGS(value1, value2),
+
+ TP_STRUCT__entry(
+ __field( unsigned long long, value1 )
+ __field( unsigned long long, value2 )
+ ),
+
+ TP_fast_assign(
+ __entry->value1 = value1;
+ __entry->value2 = value2;
+ ),
+
+ TP_printk("%llu %llu", __entry->value1, __entry->value2)
+);
+
+DECLARE_EVENT_CLASS(ocfs2__ull_uint_uint,
+
+ TP_PROTO(unsigned long long value1,
+ unsigned int value2, unsigned int value3),
+
+ TP_ARGS(value1, value2, value3),
+
+ TP_STRUCT__entry(
+ __field( unsigned long long, value1 )
+ __field( unsigned int, value2 )
+ __field( unsigned int, value3 )
+ ),
+
+ TP_fast_assign(
+ __entry->value1 = value1;
+ __entry->value2 = value2;
+ __entry->value3 = value3;
+ ),
+
+ TP_printk("%llu %u %u", __entry->value1,
+ __entry->value2, __entry->value3)
+);
+
DECLARE_EVENT_CLASS(ocfs2__uint_uint_uint,
TP_PROTO(unsigned int value1, unsigned int value2,
@@ -125,6 +203,32 @@ DECLARE_EVENT_CLASS(ocfs2__ull_int_int_int,
__entry->value2, __entry->value3)
);
+DECLARE_EVENT_CLASS(ocfs2__ull_uint_uint_uint,
+
+ TP_PROTO(unsigned long long ull, unsigned int value1,
+ unsigned int value2, unsigned int value3),
+
+ TP_ARGS(ull, value1, value2, value3),
+
+ TP_STRUCT__entry(
+ __field( unsigned long long, ull )
+ __field( unsigned int, value1 )
+ __field( unsigned int, value2 )
+ __field( unsigned int, value3 )
+ ),
+
+ TP_fast_assign(
+ __entry->ull = ull;
+ __entry->value1 = value1;
+ __entry->value2 = value2;
+ __entry->value3 = value3;
+ ),
+
+ TP_printk("%llu %u %u %u",
+ __entry->ull, __entry->value1,
+ __entry->value2, __entry->value3)
+);
+
/* Trace events for fs/ocfs2/alloc.c. */
DECLARE_EVENT_CLASS(ocfs2__btree_ops,
@@ -678,8 +782,231 @@ DEFINE_EVENT(ocfs2__ull_uint, ocfs2_local_alloc_new_window_result,
TP_ARGS(bm_off, total)
);
-
/* End of trace events for fs/ocfs2/localalloc.c. */
+
+/* Trace events for fs/ocfs2/suballoc.c. */
+DEFINE_EVENT(ocfs2__ull, ocfs2_validate_group_descriptor,
+
+ TP_PROTO(unsigned long long bg_blkno),
+
+ TP_ARGS(bg_blkno)
+);
+
+DEFINE_EVENT(ocfs2__ull_uint, ocfs2_block_group_alloc_contig,
+
+ TP_PROTO(unsigned long long bg_blkno, unsigned int chain),
+
+ TP_ARGS(bg_blkno, chain)
+);
+
+DEFINE_EVENT(ocfs2__ull_uint, ocfs2_block_group_alloc_discontig,
+
+ TP_PROTO(unsigned long long bg_blkno, unsigned int chain),
+
+ TP_ARGS(bg_blkno, chain)
+);
+
+DEFINE_EVENT(ocfs2__ull, ocfs2_block_group_alloc,
+
+ TP_PROTO(unsigned long long group_blkno),
+
+ TP_ARGS(group_blkno)
+);
+
+DEFINE_EVENT(ocfs2__uint_uint, ocfs2_reserve_suballoc_bits_nospc,
+
+ TP_PROTO(unsigned int wanted, unsigned int free),
+
+ TP_ARGS(wanted, free)
+);
+
+DEFINE_EVENT(ocfs2__uint_uint_uint, ocfs2_reserve_suballoc_bits_no_new_group,
+
+ TP_PROTO(unsigned int slot, unsigned int wanted, unsigned int free),
+
+ TP_ARGS(slot, wanted, free)
+);
+
+DEFINE_EVENT(ocfs2__ull, ocfs2_reserve_new_inode_new_group,
+
+ TP_PROTO(unsigned long long group_blkno),
+
+ TP_ARGS(group_blkno)
+);
+
+DEFINE_EVENT(ocfs2__uint_uint, ocfs2_block_group_set_bits,
+
+ TP_PROTO(unsigned int bit_off, unsigned int num_bits),
+
+ TP_ARGS(bit_off, num_bits)
+);
+
+TRACE_EVENT(ocfs2_relink_block_group,
+
+ TP_PROTO(unsigned long long i_blkno, unsigned int chain,
+ unsigned long long bg_blkno,
+ unsigned long long prev_blkno),
+
+ TP_ARGS(i_blkno, chain, bg_blkno, prev_blkno),
+
+ TP_STRUCT__entry(
+ __field( unsigned long long, i_blkno )
+ __field( unsigned int, chain )
+ __field( unsigned long long, bg_blkno )
+ __field( unsigned long long, prev_blkno )
+ ),
+
+ TP_fast_assign(
+ __entry->i_blkno = i_blkno;
+ __entry->chain = chain;
+ __entry->bg_blkno = bg_blkno;
+ __entry->prev_blkno = prev_blkno;
+ ),
+
+ TP_printk("%llu %u %llu %llu",
+ __entry->i_blkno, __entry->chain, __entry->bg_blkno,
+ __entry->prev_blkno)
+);
+
+DEFINE_EVENT(ocfs2__ull_uint_uint_uint,
+ ocfs2_cluster_group_search_wrong_max_bits,
+
+ TP_PROTO(unsigned long long bg_blkno, unsigned int bg_bits,
+ unsigned int clusters, unsigned int max_bits),
+
+ TP_ARGS(bg_blkno, bg_bits, clusters, max_bits)
+);
+
+DEFINE_EVENT(ocfs2__ull_ull, ocfs2_cluster_group_search_max_block,
+
+ TP_PROTO(unsigned long long blkno,
+ unsigned long long max_block),
+
+ TP_ARGS(blkno, max_block)
+);
+
+DEFINE_EVENT(ocfs2__ull_ull, ocfs2_block_group_search_max_block,
+
+ TP_PROTO(unsigned long long blkno,
+ unsigned long long max_block),
+
+ TP_ARGS(blkno, max_block)
+);
+
+DEFINE_EVENT(ocfs2__ull_uint_uint, ocfs2_search_chain_begin,
+
+ TP_PROTO(unsigned long long blkno,
+ unsigned int chain, unsigned int bits),
+
+ TP_ARGS(blkno, chain, bits)
+);
+
+DEFINE_EVENT(ocfs2__ull_uint, ocfs2_search_chain_succ,
+
+ TP_PROTO(unsigned long long blkno, unsigned int bits),
+
+ TP_ARGS(blkno, bits)
+);
+
+DEFINE_EVENT(ocfs2__ull_uint, ocfs2_search_chain_end,
+
+ TP_PROTO(unsigned long long blkno, unsigned int bits),
+
+ TP_ARGS(blkno, bits)
+);
+
+DEFINE_EVENT(ocfs2__uint, ocfs2_claim_suballoc_bits,
+
+ TP_PROTO(unsigned int chain),
+
+ TP_ARGS(chain)
+);
+
+DEFINE_EVENT(ocfs2__ull_uint, ocfs2_claim_new_inode_at_loc,
+
+ TP_PROTO(unsigned long long blkno, unsigned int bits),
+
+ TP_ARGS(blkno, bits)
+);
+
+DEFINE_EVENT(ocfs2__uint_uint, ocfs2_block_group_clear_bits,
+
+ TP_PROTO(unsigned int bit_off, unsigned int num_bits),
+
+ TP_ARGS(bit_off, num_bits)
+);
+
+TRACE_EVENT(ocfs2_free_suballoc_bits,
+
+ TP_PROTO(unsigned long long inode, unsigned long long group,
+ unsigned int start_bit, unsigned int count),
+
+ TP_ARGS(inode, group, start_bit, count),
+
+ TP_STRUCT__entry(
+ __field( unsigned long long, inode )
+ __field( unsigned long long, group )
+ __field( unsigned int, start_bit )
+ __field( unsigned int, count )
+ ),
+
+ TP_fast_assign(
+ __entry->inode = inode;
+ __entry->group = group;
+ __entry->start_bit = start_bit;
+ __entry->count = count;
+ ),
+
+ TP_printk("%llu %llu %u %u", __entry->inode, __entry->group,
+ __entry->start_bit, __entry->count)
+);
+
+TRACE_EVENT(ocfs2_free_clusters,
+
+ TP_PROTO(unsigned long long bg_blkno, unsigned long long start_blk,
+ unsigned int start_bit, unsigned int count),
+
+ TP_ARGS(bg_blkno, start_blk, start_bit, count),
+
+ TP_STRUCT__entry(
+ __field( unsigned long long, bg_blkno )
+ __field( unsigned long long, start_blk )
+ __field( unsigned int, start_bit )
+ __field( unsigned int, count )
+ ),
+
+ TP_fast_assign(
+ __entry->bg_blkno = bg_blkno;
+ __entry->start_blk = start_blk;
+ __entry->start_bit = start_bit;
+ __entry->count = count;
+ ),
+
+ TP_printk("%llu %llu %u %u", __entry->bg_blkno, __entry->start_blk,
+ __entry->start_bit, __entry->count)
+);
+
+DEFINE_EVENT(ocfs2__ull, ocfs2_get_suballoc_slot_bit,
+
+ TP_PROTO(unsigned long long blkno),
+
+ TP_ARGS(blkno)
+);
+
+DEFINE_EVENT(ocfs2__ull_uint, ocfs2_test_suballoc_bit,
+
+ TP_PROTO(unsigned long long blkno, unsigned int bit),
+
+ TP_ARGS(blkno, bit)
+);
+
+DEFINE_EVENT(ocfs2__ull, ocfs2_test_inode_bit,
+
+ TP_PROTO(unsigned long long blkno),
+
+ TP_ARGS(blkno)
+);
+/* End of trace events for fs/ocfs2/suballoc.c. */
#endif /* _TRACE_OCFS2_H */
/* This part must be outside protection */
diff --git a/fs/ocfs2/suballoc.c b/fs/ocfs2/suballoc.c
index 709851f..f656e81 100644
--- a/fs/ocfs2/suballoc.c
+++ b/fs/ocfs2/suballoc.c
@@ -44,6 +44,7 @@
#include "super.h"
#include "sysfile.h"
#include "uptodate.h"
+#include "ocfs2_trace.h"
#include "buffer_head_io.h"
@@ -308,8 +309,8 @@ static int ocfs2_validate_group_descriptor(struct super_block *sb,
int rc;
struct ocfs2_group_desc *gd = (struct ocfs2_group_desc *)bh->b_data;
- mlog(0, "Validating group descriptor %llu\n",
- (unsigned long long)bh->b_blocknr);
+ trace_ocfs2_validate_group_descriptor(
+ (unsigned long long)bh->b_blocknr);
BUG_ON(!buffer_uptodate(bh));
@@ -476,8 +477,8 @@ ocfs2_block_group_alloc_contig(struct ocfs2_super *osb, handle_t *handle,
/* setup the group */
bg_blkno = ocfs2_clusters_to_blocks(osb->sb, bit_off);
- mlog(0, "new descriptor, record %u, at block %llu\n",
- alloc_rec, (unsigned long long)bg_blkno);
+ trace_ocfs2_block_group_alloc_contig(
+ (unsigned long long)bg_blkno, alloc_rec);
bg_bh = sb_getblk(osb->sb, bg_blkno);
if (!bg_bh) {
@@ -656,8 +657,8 @@ ocfs2_block_group_alloc_discontig(handle_t *handle,
/* setup the group */
bg_blkno = ocfs2_clusters_to_blocks(osb->sb, bit_off);
- mlog(0, "new descriptor, record %u, at block %llu\n",
- alloc_rec, (unsigned long long)bg_blkno);
+ trace_ocfs2_block_group_alloc_discontig(
+ (unsigned long long)bg_blkno, alloc_rec);
bg_bh = sb_getblk(osb->sb, bg_blkno);
if (!bg_bh) {
@@ -727,8 +728,8 @@ static int ocfs2_block_group_alloc(struct ocfs2_super *osb,
}
if (last_alloc_group && *last_alloc_group != 0) {
- mlog(0, "use old allocation group %llu for block group alloc\n",
- (unsigned long long)*last_alloc_group);
+ trace_ocfs2_block_group_alloc(
+ (unsigned long long)*last_alloc_group);
ac->ac_last_group = *last_alloc_group;
}
@@ -851,16 +852,15 @@ static int ocfs2_reserve_suballoc_bits(struct ocfs2_super *osb,
if (bits_wanted > free_bits) {
/* cluster bitmap never grows */
if (ocfs2_is_cluster_bitmap(alloc_inode)) {
- mlog(0, "Disk Full: wanted=%u, free_bits=%u\n",
- bits_wanted, free_bits);
+ trace_ocfs2_reserve_suballoc_bits_nospc(bits_wanted,
+ free_bits);
status = -ENOSPC;
goto bail;
}
if (!(flags & ALLOC_NEW_GROUP)) {
- mlog(0, "Alloc File %u Full: wanted=%u, free_bits=%u, "
- "and we don't alloc a new group for it.\n",
- slot, bits_wanted, free_bits);
+ trace_ocfs2_reserve_suballoc_bits_no_new_group(
+ slot, bits_wanted, free_bits);
status = -ENOSPC;
goto bail;
}
@@ -1117,8 +1117,8 @@ int ocfs2_reserve_new_inode(struct ocfs2_super *osb,
spin_lock(&osb->osb_lock);
osb->osb_inode_alloc_group = alloc_group;
spin_unlock(&osb->osb_lock);
- mlog(0, "after reservation, new allocation group is "
- "%llu\n", (unsigned long long)alloc_group);
+ trace_ocfs2_reserve_new_inode_new_group(
+ (unsigned long long)alloc_group);
/*
* Some inodes must be freed by us, so try to allocate
@@ -1360,8 +1360,7 @@ static inline int ocfs2_block_group_set_bits(handle_t *handle,
BUG_ON(!OCFS2_IS_VALID_GROUP_DESC(bg));
BUG_ON(le16_to_cpu(bg->bg_free_bits_count) < num_bits);
- mlog(0, "block_group_set_bits: off = %u, num = %u\n", bit_off,
- num_bits);
+ trace_ocfs2_block_group_set_bits(bit_off, num_bits);
if (ocfs2_is_cluster_bitmap(alloc_inode))
journal_type = OCFS2_JOURNAL_ACCESS_UNDO;
@@ -1434,10 +1433,10 @@ static int ocfs2_relink_block_group(handle_t *handle,
BUG_ON(!OCFS2_IS_VALID_GROUP_DESC(bg));
BUG_ON(!OCFS2_IS_VALID_GROUP_DESC(prev_bg));
- mlog(0, "Suballoc %llu, chain %u, move group %llu to top, prev = %llu\n",
- (unsigned long long)le64_to_cpu(fe->i_blkno), chain,
- (unsigned long long)le64_to_cpu(bg->bg_blkno),
- (unsigned long long)le64_to_cpu(prev_bg->bg_blkno));
+ trace_ocfs2_relink_block_group(
+ (unsigned long long)le64_to_cpu(fe->i_blkno), chain,
+ (unsigned long long)le64_to_cpu(bg->bg_blkno),
+ (unsigned long long)le64_to_cpu(prev_bg->bg_blkno));
fe_ptr = le64_to_cpu(fe->id2.i_chain.cl_recs[chain].c_blkno);
bg_ptr = le64_to_cpu(bg->bg_next_group);
@@ -1523,10 +1522,10 @@ static int ocfs2_cluster_group_search(struct inode *inode,
if ((gd_cluster_off + max_bits) >
OCFS2_I(inode)->ip_clusters) {
max_bits = OCFS2_I(inode)->ip_clusters - gd_cluster_off;
- mlog(0, "Desc %llu, bg_bits %u, clusters %u, use %u\n",
- (unsigned long long)le64_to_cpu(gd->bg_blkno),
- le16_to_cpu(gd->bg_bits),
- OCFS2_I(inode)->ip_clusters, max_bits);
+ trace_ocfs2_cluster_group_search_wrong_max_bits(
+ (unsigned long long)le64_to_cpu(gd->bg_blkno),
+ le16_to_cpu(gd->bg_bits),
+ OCFS2_I(inode)->ip_clusters, max_bits);
}
ret = ocfs2_block_group_find_clear_bits(OCFS2_SB(inode->i_sb),
@@ -1540,9 +1539,9 @@ static int ocfs2_cluster_group_search(struct inode *inode,
gd_cluster_off +
res->sr_bit_offset +
res->sr_bits);
- mlog(0, "Checking %llu against %llu\n",
- (unsigned long long)blkoff,
- (unsigned long long)max_block);
+ trace_ocfs2_cluster_group_search_max_block(
+ (unsigned long long)blkoff,
+ (unsigned long long)max_block);
if (blkoff > max_block)
return -ENOSPC;
}
@@ -1586,9 +1585,9 @@ static int ocfs2_block_group_search(struct inode *inode,
if (!ret && max_block) {
blkoff = le64_to_cpu(bg->bg_blkno) +
res->sr_bit_offset + res->sr_bits;
- mlog(0, "Checking %llu against %llu\n",
- (unsigned long long)blkoff,
- (unsigned long long)max_block);
+ trace_ocfs2_block_group_search_max_block(
+ (unsigned long long)blkoff,
+ (unsigned long long)max_block);
if (blkoff > max_block)
ret = -ENOSPC;
}
@@ -1754,9 +1753,9 @@ static int ocfs2_search_chain(struct ocfs2_alloc_context *ac,
struct ocfs2_group_desc *bg;
chain = ac->ac_chain;
- mlog(0, "trying to alloc %u bits from chain %u, inode %llu\n",
- bits_wanted, chain,
- (unsigned long long)OCFS2_I(alloc_inode)->ip_blkno);
+ trace_ocfs2_search_chain_begin(
+ (unsigned long long)OCFS2_I(alloc_inode)->ip_blkno,
+ bits_wanted, chain);
status = ocfs2_read_group_descriptor(alloc_inode, fe,
le64_to_cpu(cl->cl_recs[chain].c_blkno),
@@ -1797,8 +1796,8 @@ static int ocfs2_search_chain(struct ocfs2_alloc_context *ac,
goto bail;
}
- mlog(0, "alloc succeeds: we give %u bits from block group %llu\n",
- res->sr_bits, (unsigned long long)le64_to_cpu(bg->bg_blkno));
+ trace_ocfs2_search_chain_succ(
+ (unsigned long long)le64_to_cpu(bg->bg_blkno), res->sr_bits);
res->sr_bg_blkno = le64_to_cpu(bg->bg_blkno);
@@ -1859,8 +1858,9 @@ static int ocfs2_search_chain(struct ocfs2_alloc_context *ac,
goto bail;
}
- mlog(0, "Allocated %u bits from suballocator %llu\n", res->sr_bits,
- (unsigned long long)le64_to_cpu(fe->i_blkno));
+ trace_ocfs2_search_chain_end(
+ (unsigned long long)le64_to_cpu(fe->i_blkno),
+ res->sr_bits);
out_loc_only:
*bits_left = le16_to_cpu(bg->bg_free_bits_count);
@@ -1942,8 +1942,7 @@ static int ocfs2_claim_suballoc_bits(struct ocfs2_alloc_context *ac,
goto bail;
}
- mlog(0, "Search of victim chain %u came up with nothing, "
- "trying all chains now.\n", victim);
+ trace_ocfs2_claim_suballoc_bits(victim);
/* If we didn't pick a good victim, then just default to
* searching each chain in order. Don't allow chain relinking
@@ -2171,8 +2170,8 @@ int ocfs2_claim_new_inode_at_loc(handle_t *handle,
goto out;
}
- mlog(0, "Allocated %u bits from suballocator %llu\n", res->sr_bits,
- (unsigned long long)di_blkno);
+ trace_ocfs2_claim_new_inode_at_loc((unsigned long long)di_blkno,
+ res->sr_bits);
atomic_inc(&OCFS2_SB(ac->ac_inode->i_sb)->alloc_stats.bg_allocs);
@@ -2393,7 +2392,7 @@ static int ocfs2_block_group_clear_bits(handle_t *handle,
* ocfs2_read_group_descriptor(). Any corruption is a code bug. */
BUG_ON(!OCFS2_IS_VALID_GROUP_DESC(bg));
- mlog(0, "off = %u, num = %u\n", bit_off, num_bits);
+ trace_ocfs2_block_group_clear_bits(bit_off, num_bits);
BUG_ON(undo_fn && !ocfs2_is_cluster_bitmap(alloc_inode));
status = ocfs2_journal_access_gd(handle, INODE_CACHE(alloc_inode),
@@ -2466,9 +2465,10 @@ static int _ocfs2_free_suballoc_bits(handle_t *handle,
BUG_ON(!OCFS2_IS_VALID_DINODE(fe));
BUG_ON((count + start_bit) > ocfs2_bits_per_group(cl));
- mlog(0, "%llu: freeing %u bits from group %llu, starting at %u\n",
- (unsigned long long)OCFS2_I(alloc_inode)->ip_blkno, count,
- (unsigned long long)bg_blkno, start_bit);
+ trace_ocfs2_free_suballoc_bits(
+ (unsigned long long)OCFS2_I(alloc_inode)->ip_blkno,
+ (unsigned long long)bg_blkno,
+ start_bit, count);
status = ocfs2_read_group_descriptor(alloc_inode, fe, bg_blkno,
&group_bh);
@@ -2560,10 +2560,9 @@ static int _ocfs2_free_clusters(handle_t *handle,
ocfs2_block_to_cluster_group(bitmap_inode, start_blk, &bg_blkno,
&bg_start_bit);
- mlog(0, "want to free %u clusters starting at block %llu\n",
- num_clusters, (unsigned long long)start_blk);
- mlog(0, "bg_blkno = %llu, bg_start_bit = %u\n",
- (unsigned long long)bg_blkno, bg_start_bit);
+ trace_ocfs2_free_clusters((unsigned long long)bg_blkno,
+ (unsigned long long)start_blk,
+ bg_start_bit, num_clusters);
status = _ocfs2_free_suballoc_bits(handle, bitmap_inode, bitmap_bh,
bg_start_bit, bg_blkno,
@@ -2748,7 +2747,7 @@ static int ocfs2_get_suballoc_slot_bit(struct ocfs2_super *osb, u64 blkno,
struct buffer_head *inode_bh = NULL;
struct ocfs2_dinode *inode_fe;
- mlog(0, "blkno: %llu\n", (unsigned long long)blkno);
+ trace_ocfs2_get_suballoc_slot_bit((unsigned long long)blkno);
/* dirty read disk */
status = ocfs2_read_blocks_sync(osb, blkno, 1, &inode_bh);
@@ -2809,8 +2808,8 @@ static int ocfs2_test_suballoc_bit(struct ocfs2_super *osb,
u64 bg_blkno;
int status;
- mlog(0, "blkno: %llu bit: %u\n", (unsigned long long)blkno,
- (unsigned int)bit);
+ trace_ocfs2_test_suballoc_bit((unsigned long long)blkno,
+ (unsigned int)bit);
alloc_di = (struct ocfs2_dinode *)alloc_bh->b_data;
if ((bit + 1) > ocfs2_bits_per_group(&alloc_di->id2.i_chain)) {
@@ -2863,7 +2862,7 @@ int ocfs2_test_inode_bit(struct ocfs2_super *osb, u64 blkno, int *res)
struct inode *inode_alloc_inode;
struct buffer_head *alloc_bh = NULL;
- mlog(0, "blkno: %llu", (unsigned long long)blkno);
+ trace_ocfs2_test_inode_bit((unsigned long long)blkno);
status = ocfs2_get_suballoc_slot_bit(osb, blkno, &suballoc_slot,
&group_blkno, &suballoc_bit);
--
1.7.0.4
^ permalink raw reply related [flat|nested] 53+ messages in thread
* [Ocfs2-devel] [PATCH 08/34] ocfs2: Remove mlog(0) from fs/ocfs2/suballoc.c
2010-12-23 7:19 [Ocfs2-devel] [PATCH 00/34] OCFS2: Add trace event and replace mlog(0) Tao Ma
` (6 preceding siblings ...)
2010-12-23 7:30 ` [Ocfs2-devel] [PATCH 07/34] ocfs2: Remove mlog(0) from fs/ocfs2/suballoc.c Tao Ma
@ 2010-12-23 7:30 ` Tao Ma
2010-12-23 7:30 ` [Ocfs2-devel] [PATCH 09/34] ocfs2: Remove DISK_ALLOC from masklog Tao Ma
` (30 subsequent siblings)
38 siblings, 0 replies; 53+ messages in thread
From: Tao Ma @ 2010-12-23 7:30 UTC (permalink / raw)
To: ocfs2-devel
From: Tao Ma <tao.ma@oracle.com>
This is the 4th step to remove the debug info of DISK_ALLOC.
Signed-off-by: Tao Ma <tao.ma@oracle.com>
---
fs/ocfs2/ocfs2_trace.h | 25 +++++++++++++++++++++++++
fs/ocfs2/resize.c | 12 +++++++-----
2 files changed, 32 insertions(+), 5 deletions(-)
diff --git a/fs/ocfs2/ocfs2_trace.h b/fs/ocfs2/ocfs2_trace.h
index 196a995..1d518d5 100644
--- a/fs/ocfs2/ocfs2_trace.h
+++ b/fs/ocfs2/ocfs2_trace.h
@@ -784,6 +784,31 @@ DEFINE_EVENT(ocfs2__ull_uint, ocfs2_local_alloc_new_window_result,
);
/* End of trace events for fs/ocfs2/localalloc.c. */
+/* Trace events for fs/ocfs2/resize.c. */
+DEFINE_EVENT(ocfs2__uint_uint, ocfs2_update_last_group_and_inode,
+
+ TP_PROTO(unsigned int new_clusters, unsigned int first_new_cluster),
+
+ TP_ARGS(new_clusters, first_new_cluster)
+);
+
+DEFINE_EVENT(ocfs2__ull_uint, ocfs2_group_extend,
+
+ TP_PROTO(unsigned long long bg_blkno, unsigned int new_clusters),
+
+ TP_ARGS(bg_blkno, new_clusters)
+);
+
+DEFINE_EVENT(ocfs2__ull_uint_uint_uint, ocfs2_group_add,
+
+ TP_PROTO(unsigned long long bg_blkno, unsigned int chain,
+ unsigned int clusters, unsigned int frees),
+
+ TP_ARGS(bg_blkno, chain, clusters, frees)
+);
+
+/* End of trace events for fs/ocfs2/resize.c. */
+
/* Trace events for fs/ocfs2/suballoc.c. */
DEFINE_EVENT(ocfs2__ull, ocfs2_validate_group_descriptor,
diff --git a/fs/ocfs2/resize.c b/fs/ocfs2/resize.c
index 76b3790..761d20d 100644
--- a/fs/ocfs2/resize.c
+++ b/fs/ocfs2/resize.c
@@ -39,6 +39,7 @@
#include "super.h"
#include "sysfile.h"
#include "uptodate.h"
+#include "ocfs2_trace.h"
#include "buffer_head_io.h"
#include "suballoc.h"
@@ -102,8 +103,8 @@ static int ocfs2_update_last_group_and_inode(handle_t *handle,
u16 cl_bpc = le16_to_cpu(cl->cl_bpc);
u16 cl_cpg = le16_to_cpu(cl->cl_cpg);
- mlog(0, "(new_clusters=%d, first_new_cluster = %u)\n",
- new_clusters, first_new_cluster);
+ trace_ocfs2_update_last_group_and_inode(new_clusters,
+ first_new_cluster);
ret = ocfs2_journal_access_gd(handle, INODE_CACHE(bm_inode),
group_bh, OCFS2_JOURNAL_ACCESS_WRITE);
@@ -340,7 +341,8 @@ int ocfs2_group_extend(struct inode * inode, int new_clusters)
goto out_unlock;
}
- mlog(0, "extend the last group at %llu, new clusters = %d\n",
+
+ trace_ocfs2_group_extend(
(unsigned long long)le64_to_cpu(group->bg_blkno), new_clusters);
handle = ocfs2_start_trans(osb, OCFS2_GROUP_EXTEND_CREDITS);
@@ -515,8 +517,8 @@ int ocfs2_group_add(struct inode *inode, struct ocfs2_new_group_input *input)
goto out_unlock;
}
- mlog(0, "Add a new group %llu in chain = %u, length = %u\n",
- (unsigned long long)input->group, input->chain, input->clusters);
+ trace_ocfs2_group_add((unsigned long long)input->group,
+ input->chain, input->clusters, input->frees);
handle = ocfs2_start_trans(osb, OCFS2_GROUP_ADD_CREDITS);
if (IS_ERR(handle)) {
--
1.7.0.4
^ permalink raw reply related [flat|nested] 53+ messages in thread
* [Ocfs2-devel] [PATCH 09/34] ocfs2: Remove DISK_ALLOC from masklog.
2010-12-23 7:19 [Ocfs2-devel] [PATCH 00/34] OCFS2: Add trace event and replace mlog(0) Tao Ma
` (7 preceding siblings ...)
2010-12-23 7:30 ` [Ocfs2-devel] [PATCH 08/34] " Tao Ma
@ 2010-12-23 7:30 ` Tao Ma
2010-12-23 7:30 ` [Ocfs2-devel] [PATCH 10/34] ocfs2: Remove masklog ML_REFCOUNT Tao Ma
` (29 subsequent siblings)
38 siblings, 0 replies; 53+ messages in thread
From: Tao Ma @ 2010-12-23 7:30 UTC (permalink / raw)
To: ocfs2-devel
From: Tao Ma <tao.ma@oracle.com>
Since all 4 files, localalloc.c, suballoc.c, alloc.c and
resize.c, which use DISK_ALLOC are changed to trace events,
Remove masklog DISK_ALLOC totally.
Signed-off-by: Tao Ma <tao.ma@oracle.com>
---
fs/ocfs2/alloc.c | 1 -
fs/ocfs2/cluster/masklog.c | 1 -
fs/ocfs2/cluster/masklog.h | 1 -
fs/ocfs2/localalloc.c | 1 -
fs/ocfs2/resize.c | 1 -
fs/ocfs2/suballoc.c | 1 -
6 files changed, 0 insertions(+), 6 deletions(-)
diff --git a/fs/ocfs2/alloc.c b/fs/ocfs2/alloc.c
index 9bdfaf8..fc2105c 100644
--- a/fs/ocfs2/alloc.c
+++ b/fs/ocfs2/alloc.c
@@ -30,7 +30,6 @@
#include <linux/swap.h>
#include <linux/quotaops.h>
-#define MLOG_MASK_PREFIX ML_DISK_ALLOC
#include <cluster/masklog.h>
#include "ocfs2.h"
diff --git a/fs/ocfs2/cluster/masklog.c b/fs/ocfs2/cluster/masklog.c
index 41260fb..b395c6a 100644
--- a/fs/ocfs2/cluster/masklog.c
+++ b/fs/ocfs2/cluster/masklog.c
@@ -93,7 +93,6 @@ static struct mlog_attribute mlog_attrs[MLOG_MAX_BITS] = {
define_mask(DLM_RECOVERY),
define_mask(AIO),
define_mask(JOURNAL),
- define_mask(DISK_ALLOC),
define_mask(SUPER),
define_mask(FILE_IO),
define_mask(EXTENT_MAP),
diff --git a/fs/ocfs2/cluster/masklog.h b/fs/ocfs2/cluster/masklog.h
index 707e119..70df956 100644
--- a/fs/ocfs2/cluster/masklog.h
+++ b/fs/ocfs2/cluster/masklog.h
@@ -95,7 +95,6 @@
#define ML_DLM_RECOVERY 0x0000000000001000ULL /* dlm master functions */
#define ML_AIO 0x0000000000002000ULL /* ocfs2 aio read and write */
#define ML_JOURNAL 0x0000000000004000ULL /* ocfs2 journalling functions */
-#define ML_DISK_ALLOC 0x0000000000008000ULL /* ocfs2 disk allocation */
#define ML_SUPER 0x0000000000010000ULL /* ocfs2 mount / umount */
#define ML_FILE_IO 0x0000000000020000ULL /* ocfs2 file I/O */
#define ML_EXTENT_MAP 0x0000000000040000ULL /* ocfs2 extent map caching */
diff --git a/fs/ocfs2/localalloc.c b/fs/ocfs2/localalloc.c
index a311b1f..210c352 100644
--- a/fs/ocfs2/localalloc.c
+++ b/fs/ocfs2/localalloc.c
@@ -29,7 +29,6 @@
#include <linux/highmem.h>
#include <linux/bitops.h>
-#define MLOG_MASK_PREFIX ML_DISK_ALLOC
#include <cluster/masklog.h>
#include "ocfs2.h"
diff --git a/fs/ocfs2/resize.c b/fs/ocfs2/resize.c
index 761d20d..ec55add 100644
--- a/fs/ocfs2/resize.c
+++ b/fs/ocfs2/resize.c
@@ -27,7 +27,6 @@
#include <linux/fs.h>
#include <linux/types.h>
-#define MLOG_MASK_PREFIX ML_DISK_ALLOC
#include <cluster/masklog.h>
#include "ocfs2.h"
diff --git a/fs/ocfs2/suballoc.c b/fs/ocfs2/suballoc.c
index f656e81..c4ebfe7 100644
--- a/fs/ocfs2/suballoc.c
+++ b/fs/ocfs2/suballoc.c
@@ -29,7 +29,6 @@
#include <linux/slab.h>
#include <linux/highmem.h>
-#define MLOG_MASK_PREFIX ML_DISK_ALLOC
#include <cluster/masklog.h>
#include "ocfs2.h"
--
1.7.0.4
^ permalink raw reply related [flat|nested] 53+ messages in thread
* [Ocfs2-devel] [PATCH 10/34] ocfs2: Remove masklog ML_REFCOUNT.
2010-12-23 7:19 [Ocfs2-devel] [PATCH 00/34] OCFS2: Add trace event and replace mlog(0) Tao Ma
` (8 preceding siblings ...)
2010-12-23 7:30 ` [Ocfs2-devel] [PATCH 09/34] ocfs2: Remove DISK_ALLOC from masklog Tao Ma
@ 2010-12-23 7:30 ` Tao Ma
2010-12-23 7:30 ` [Ocfs2-devel] [PATCH 11/34] ocfs2: Remove mlog(0) from fs/ocfs2/aops.c Tao Ma
` (28 subsequent siblings)
38 siblings, 0 replies; 53+ messages in thread
From: Tao Ma @ 2010-12-23 7:30 UTC (permalink / raw)
To: ocfs2-devel
From: Tao Ma <tao.ma@oracle.com>
Change all the "mlog(0," in fs/ocfs2/refcounttree.c to trace events.
And finally remove masklog ML_REFCOUNT.
Signed-off-by: Tao Ma <tao.ma@oracle.com>
---
fs/ocfs2/cluster/masklog.c | 1 -
fs/ocfs2/cluster/masklog.h | 1 -
fs/ocfs2/ocfs2_trace.h | 430 ++++++++++++++++++++++++++++++++++++++++++++
fs/ocfs2/refcounttree.c | 158 ++++++++--------
4 files changed, 508 insertions(+), 82 deletions(-)
diff --git a/fs/ocfs2/cluster/masklog.c b/fs/ocfs2/cluster/masklog.c
index b395c6a..b8939c4 100644
--- a/fs/ocfs2/cluster/masklog.c
+++ b/fs/ocfs2/cluster/masklog.c
@@ -108,7 +108,6 @@ static struct mlog_attribute mlog_attrs[MLOG_MAX_BITS] = {
define_mask(EXPORT),
define_mask(XATTR),
define_mask(QUOTA),
- define_mask(REFCOUNT),
define_mask(BASTS),
define_mask(ERROR),
define_mask(NOTICE),
diff --git a/fs/ocfs2/cluster/masklog.h b/fs/ocfs2/cluster/masklog.h
index 70df956..8379361 100644
--- a/fs/ocfs2/cluster/masklog.h
+++ b/fs/ocfs2/cluster/masklog.h
@@ -110,7 +110,6 @@
#define ML_EXPORT 0x0000000010000000ULL /* ocfs2 export operations */
#define ML_XATTR 0x0000000020000000ULL /* ocfs2 extended attributes */
#define ML_QUOTA 0x0000000040000000ULL /* ocfs2 quota operations */
-#define ML_REFCOUNT 0x0000000080000000ULL /* refcount tree operations */
#define ML_BASTS 0x0000001000000000ULL /* dlmglue asts and basts */
/* bits that are infrequently given and frequently matched in the high word */
#define ML_ERROR 0x0000000100000000ULL /* sent to KERN_ERR */
diff --git a/fs/ocfs2/ocfs2_trace.h b/fs/ocfs2/ocfs2_trace.h
index 1d518d5..3671151 100644
--- a/fs/ocfs2/ocfs2_trace.h
+++ b/fs/ocfs2/ocfs2_trace.h
@@ -133,6 +133,29 @@ DECLARE_EVENT_CLASS(ocfs2__ull_ull,
TP_printk("%llu %llu", __entry->value1, __entry->value2)
);
+DECLARE_EVENT_CLASS(ocfs2__ull_ull_uint,
+
+ TP_PROTO(unsigned long long value1,
+ unsigned long long value2, unsigned int value3),
+
+ TP_ARGS(value1, value2, value3),
+
+ TP_STRUCT__entry(
+ __field( unsigned long long, value1 )
+ __field( unsigned long long, value2 )
+ __field( unsigned int, value3 )
+ ),
+
+ TP_fast_assign(
+ __entry->value1 = value1;
+ __entry->value2 = value2;
+ __entry->value3 = value3;
+ ),
+
+ TP_printk("%llu %llu %u",
+ __entry->value1, __entry->value2, __entry->value3)
+);
+
DECLARE_EVENT_CLASS(ocfs2__ull_uint_uint,
TP_PROTO(unsigned long long value1,
@@ -1032,6 +1055,413 @@ DEFINE_EVENT(ocfs2__ull, ocfs2_test_inode_bit,
TP_ARGS(blkno)
);
/* End of trace events for fs/ocfs2/suballoc.c. */
+
+/* Trace events for fs/ocfs2/refcounttree.c. */
+DEFINE_EVENT(ocfs2__ull, ocfs2_validate_refcount_block,
+
+ TP_PROTO(unsigned long long blkno),
+
+ TP_ARGS(blkno)
+);
+
+DEFINE_EVENT(ocfs2__ull, ocfs2_purge_refcount_trees,
+
+ TP_PROTO(unsigned long long tree_blkno),
+
+ TP_ARGS(tree_blkno)
+);
+
+DEFINE_EVENT(ocfs2__ull, ocfs2_create_refcount_tree,
+
+ TP_PROTO(unsigned long long i_blkno),
+
+ TP_ARGS(i_blkno)
+);
+
+DEFINE_EVENT(ocfs2__ull, ocfs2_create_refcount_tree_blkno,
+
+ TP_PROTO(unsigned long long tree_blkno),
+
+ TP_ARGS(tree_blkno)
+);
+
+DEFINE_EVENT(ocfs2__ull_int_int_int, ocfs2_change_refcount_rec,
+
+ TP_PROTO(unsigned long long owner, int index,
+ int count, int change),
+
+ TP_ARGS(owner, index, count, change)
+);
+
+DEFINE_EVENT(ocfs2__ull_uint, ocfs2_expand_inline_ref_root,
+
+ TP_PROTO(unsigned long long owner, unsigned int used),
+
+ TP_ARGS(owner, used)
+);
+
+DEFINE_EVENT(ocfs2__ull_uint_uint, ocfs2_divide_leaf_refcount_block,
+
+ TP_PROTO(unsigned long long blkno,
+ unsigned int count, unsigned int used),
+
+ TP_ARGS(blkno, count, used)
+);
+
+DEFINE_EVENT(ocfs2__ull_uint, ocfs2_new_leaf_refcount_block,
+
+ TP_PROTO(unsigned long long owner, unsigned int new_cpos),
+
+ TP_ARGS(owner, new_cpos)
+);
+
+DECLARE_EVENT_CLASS(ocfs2__refcount_tree_ops,
+
+ TP_PROTO(unsigned long long blkno, int index,
+ unsigned long long cpos,
+ unsigned int clusters, unsigned int refcount),
+
+ TP_ARGS(blkno, index, cpos, clusters, refcount),
+
+ TP_STRUCT__entry(
+ __field( unsigned long long, blkno )
+ __field( int, index )
+ __field( unsigned long long, cpos )
+ __field( unsigned int, clusters )
+ __field( unsigned int, refcount )
+ ),
+
+ TP_fast_assign(
+ __entry->blkno = blkno;
+ __entry->index = index;
+ __entry->cpos = cpos;
+ __entry->clusters = clusters;
+ __entry->refcount = refcount;
+ ),
+
+ TP_printk("%llu %d %llu %u %u", __entry->blkno, __entry->index,
+ __entry->cpos, __entry->clusters, __entry->refcount)
+);
+
+DEFINE_EVENT(ocfs2__refcount_tree_ops, ocfs2_insert_refcount_rec,
+
+ TP_PROTO(unsigned long long blkno, int index,
+ unsigned long long cpos,
+ unsigned int count, unsigned int refcount),
+
+ TP_ARGS(blkno, index, cpos, count, refcount)
+);
+
+TRACE_EVENT(ocfs2_split_refcount_rec,
+
+ TP_PROTO(unsigned long long cpos,
+ unsigned int clusters, unsigned int refcount,
+ unsigned long long split_cpos,
+ unsigned int split_clusters, unsigned int split_refcount),
+
+ TP_ARGS(cpos, clusters, refcount,
+ split_cpos, split_clusters, split_refcount),
+
+ TP_STRUCT__entry(
+ __field( unsigned long long, cpos )
+ __field( unsigned int, clusters )
+ __field( unsigned int, refcount )
+ __field( unsigned long long, split_cpos )
+ __field( unsigned int, split_clusters )
+ __field( unsigned int, split_refcount )
+ ),
+
+ TP_fast_assign(
+ __entry->cpos = cpos;
+ __entry->clusters = clusters;
+ __entry->refcount = refcount;
+ __entry->split_cpos = split_cpos;
+ __entry->split_clusters = split_clusters;
+ __entry->split_refcount = split_refcount;
+ ),
+
+ TP_printk("%llu %u %u %llu %u %u",
+ __entry->cpos, __entry->clusters, __entry->refcount,
+ __entry->split_cpos, __entry->split_clusters,
+ __entry->split_refcount)
+);
+
+DEFINE_EVENT(ocfs2__refcount_tree_ops, ocfs2_split_refcount_rec_insert,
+
+ TP_PROTO(unsigned long long blkno, int index,
+ unsigned long long cpos,
+ unsigned int count, unsigned int refcount),
+
+ TP_ARGS(blkno, index, cpos, count, refcount)
+);
+
+DEFINE_EVENT(ocfs2__ull_ull_uint, ocfs2_increase_refcount_begin,
+
+ TP_PROTO(unsigned long long owner,
+ unsigned long long cpos, unsigned int clusters),
+
+ TP_ARGS(owner, cpos, clusters)
+);
+
+DEFINE_EVENT(ocfs2__ull_uint_uint, ocfs2_increase_refcount_change,
+
+ TP_PROTO(unsigned long long cpos,
+ unsigned int len, unsigned int refcount),
+
+ TP_ARGS(cpos, len, refcount)
+);
+
+DEFINE_EVENT(ocfs2__ull_uint, ocfs2_increase_refcount_insert,
+
+ TP_PROTO(unsigned long long cpos, unsigned int len),
+
+ TP_ARGS(cpos, len)
+);
+
+DEFINE_EVENT(ocfs2__ull_uint_uint, ocfs2_increase_refcount_split,
+
+ TP_PROTO(unsigned long long cpos,
+ unsigned int len, unsigned int refcount),
+
+ TP_ARGS(cpos, len, refcount)
+);
+
+DEFINE_EVENT(ocfs2__ull_ull_uint, ocfs2_remove_refcount_extent,
+
+ TP_PROTO(unsigned long long owner,
+ unsigned long long leaf_blkno, unsigned int cpos),
+
+ TP_ARGS(owner, leaf_blkno, cpos)
+);
+
+DEFINE_EVENT(ocfs2__ull, ocfs2_restore_refcount_block,
+
+ TP_PROTO(unsigned long long owner),
+
+ TP_ARGS(owner)
+);
+
+DEFINE_EVENT(ocfs2__ull_ull_uint, ocfs2_decrease_refcount_rec,
+
+ TP_PROTO(unsigned long long owner,
+ unsigned long long cpos, unsigned int len),
+
+ TP_ARGS(owner, cpos, len)
+);
+
+TRACE_EVENT(ocfs2_decrease_refcount,
+
+ TP_PROTO(unsigned long long owner,
+ unsigned long long cpos,
+ unsigned int len, int delete),
+
+ TP_ARGS(owner, cpos, len, delete),
+
+ TP_STRUCT__entry(
+ __field( unsigned long long, owner )
+ __field( unsigned long long, cpos )
+ __field( unsigned int, len )
+ __field( int, delete )
+ ),
+
+ TP_fast_assign(
+ __entry->owner = owner;
+ __entry->cpos = cpos;
+ __entry->len = len;
+ __entry->delete = delete;
+ ),
+
+ TP_printk("%llu %llu %u %d",
+ __entry->owner, __entry->cpos, __entry->len, __entry->delete)
+);
+
+DEFINE_EVENT(ocfs2__ull_uint_uint_uint, ocfs2_mark_extent_refcounted,
+
+ TP_PROTO(unsigned long long owner, unsigned int cpos,
+ unsigned int len, unsigned int phys),
+
+ TP_ARGS(owner, cpos, len, phys)
+);
+
+DEFINE_EVENT(ocfs2__ull_uint_uint_uint, ocfs2_calc_refcount_meta_credits,
+
+ TP_PROTO(unsigned long long cpos, unsigned int clusters,
+ unsigned int meta_add, unsigned int credits),
+
+ TP_ARGS(cpos, clusters, meta_add, credits)
+);
+
+TRACE_EVENT(ocfs2_calc_refcount_meta_credits_iterate,
+
+ TP_PROTO(int recs_add, unsigned long long cpos,
+ unsigned int clusters, unsigned long long r_cpos,
+ unsigned int r_clusters, unsigned int refcount, int index),
+
+ TP_ARGS(recs_add, cpos, clusters, r_cpos, r_clusters, refcount, index),
+
+ TP_STRUCT__entry(
+ __field( int, recs_add )
+ __field( unsigned long long, cpos )
+ __field( unsigned int, clusters )
+ __field( unsigned long long, r_cpos )
+ __field( unsigned int, r_clusters )
+ __field( unsigned int, refcount )
+ __field( int, index )
+ ),
+
+ TP_fast_assign(
+ __entry->recs_add = recs_add;
+ __entry->cpos = cpos;
+ __entry->clusters = clusters;
+ __entry->r_cpos = r_cpos;
+ __entry->r_clusters = r_clusters;
+ __entry->refcount = refcount;
+ __entry->index = index;
+ ),
+
+ TP_printk("%d %llu %u %llu %u %u %d",
+ __entry->recs_add, __entry->cpos, __entry->clusters,
+ __entry->r_cpos, __entry->r_clusters,
+ __entry->refcount, __entry->index)
+);
+
+DEFINE_EVENT(ocfs2__int_int, ocfs2_add_refcount_flag,
+
+ TP_PROTO(int blocks, int credits),
+
+ TP_ARGS(blocks, credits)
+);
+
+DEFINE_EVENT(ocfs2__int_int, ocfs2_prepare_refcount_change_for_del,
+
+ TP_PROTO(int blocks, int credits),
+
+ TP_ARGS(blocks, credits)
+);
+
+DEFINE_EVENT(ocfs2__int_int, ocfs2_lock_refcount_allocators,
+
+ TP_PROTO(int blocks, int credits),
+
+ TP_ARGS(blocks, credits)
+);
+
+DEFINE_EVENT(ocfs2__ull_uint_uint_uint, ocfs2_duplicate_clusters_by_page,
+
+ TP_PROTO(unsigned long long cpos, unsigned int old_cluster,
+ unsigned int new_cluster, unsigned int new_len),
+
+ TP_ARGS(cpos, old_cluster, new_cluster, new_len)
+);
+
+DEFINE_EVENT(ocfs2__ull_uint_uint_uint, ocfs2_duplicate_clusters_by_jbd,
+
+ TP_PROTO(unsigned long long cpos, unsigned int old_cluster,
+ unsigned int new_cluster, unsigned int new_len),
+
+ TP_ARGS(cpos, old_cluster, new_cluster, new_len)
+);
+
+TRACE_EVENT(ocfs2_clear_ext_refcount,
+
+ TP_PROTO(unsigned long long ino, unsigned int cpos,
+ unsigned int len, unsigned int p_cluster,
+ unsigned int ext_flags),
+
+ TP_ARGS(ino, cpos, len, p_cluster, ext_flags),
+
+ TP_STRUCT__entry(
+ __field( unsigned long long, ino )
+ __field( unsigned int, cpos )
+ __field( unsigned int, len )
+ __field( unsigned int, p_cluster )
+ __field( unsigned int, ext_flags )
+ ),
+
+ TP_fast_assign(
+ __entry->ino = ino;
+ __entry->cpos = cpos;
+ __entry->len = len;
+ __entry->p_cluster = p_cluster;
+ __entry->ext_flags = ext_flags;
+ ),
+
+ TP_printk("%llu %u %u %u %u",
+ __entry->ino, __entry->cpos, __entry->len,
+ __entry->p_cluster, __entry->ext_flags)
+);
+
+TRACE_EVENT(ocfs2_replace_clusters,
+
+ TP_PROTO(unsigned long long ino, unsigned int cpos,
+ unsigned int old, unsigned int new, unsigned int len,
+ unsigned int ext_flags),
+
+ TP_ARGS(ino, cpos, old, new, len, ext_flags),
+
+ TP_STRUCT__entry(
+ __field( unsigned long long, ino )
+ __field( unsigned int, cpos )
+ __field( unsigned int, old )
+ __field( unsigned int, new )
+ __field( unsigned int, len )
+ __field( unsigned int, ext_flags )
+ ),
+
+ TP_fast_assign(
+ __entry->ino = ino;
+ __entry->cpos = cpos;
+ __entry->old = old;
+ __entry->new = new;
+ __entry->len = len;
+ __entry->ext_flags = ext_flags;
+ ),
+
+ TP_printk("%llu %u %u %u %u %u",
+ __entry->ino, __entry->cpos, __entry->old, __entry->new,
+ __entry->len, __entry->ext_flags)
+);
+
+DEFINE_EVENT(ocfs2__ull_uint_uint_uint, ocfs2_make_clusters_writable,
+
+ TP_PROTO(unsigned long long cpos, unsigned int p_cluster,
+ unsigned int num_clusters, unsigned int e_flags),
+
+ TP_ARGS(cpos, p_cluster, num_clusters, e_flags)
+);
+
+TRACE_EVENT(ocfs2_refcount_cow_hunk,
+
+ TP_PROTO(unsigned long long ino, unsigned int cpos,
+ unsigned int write_len, unsigned int max_cpos,
+ unsigned int cow_start, unsigned int cow_len),
+
+ TP_ARGS(ino, cpos, write_len, max_cpos, cow_start, cow_len),
+
+ TP_STRUCT__entry(
+ __field( unsigned long long, ino )
+ __field( unsigned int, cpos )
+ __field( unsigned int, write_len )
+ __field( unsigned int, max_cpos )
+ __field( unsigned int, cow_start )
+ __field( unsigned int, cow_len )
+ ),
+
+ TP_fast_assign(
+ __entry->ino = ino;
+ __entry->cpos = cpos;
+ __entry->write_len = write_len;
+ __entry->max_cpos = max_cpos;
+ __entry->cow_start = cow_start;
+ __entry->cow_len = cow_len;
+ ),
+
+ TP_printk("%llu %u %u %u %u %u",
+ __entry->ino, __entry->cpos, __entry->write_len,
+ __entry->max_cpos, __entry->cow_start, __entry->cow_len)
+);
+
+/* End of trace events for fs/ocfs2/refcounttree.c. */
#endif /* _TRACE_OCFS2_H */
/* This part must be outside protection */
diff --git a/fs/ocfs2/refcounttree.c b/fs/ocfs2/refcounttree.c
index b5f9160..ae1d4ae 100644
--- a/fs/ocfs2/refcounttree.c
+++ b/fs/ocfs2/refcounttree.c
@@ -16,7 +16,6 @@
*/
#include <linux/sort.h>
-#define MLOG_MASK_PREFIX ML_REFCOUNT
#include <cluster/masklog.h>
#include "ocfs2.h"
#include "inode.h"
@@ -34,6 +33,7 @@
#include "aops.h"
#include "xattr.h"
#include "namei.h"
+#include "ocfs2_trace.h"
#include <linux/bio.h>
#include <linux/blkdev.h>
@@ -84,8 +84,7 @@ static int ocfs2_validate_refcount_block(struct super_block *sb,
struct ocfs2_refcount_block *rb =
(struct ocfs2_refcount_block *)bh->b_data;
- mlog(0, "Validating refcount block %llu\n",
- (unsigned long long)bh->b_blocknr);
+ trace_ocfs2_validate_refcount_block((unsigned long long)bh->b_blocknr);
BUG_ON(!buffer_uptodate(bh));
@@ -545,8 +544,8 @@ void ocfs2_purge_refcount_trees(struct ocfs2_super *osb)
while ((node = rb_last(root)) != NULL) {
tree = rb_entry(node, struct ocfs2_refcount_tree, rf_node);
- mlog(0, "Purge tree %llu\n",
- (unsigned long long) tree->rf_blkno);
+ trace_ocfs2_purge_refcount_trees(
+ (unsigned long long) tree->rf_blkno);
rb_erase(&tree->rf_node, root);
ocfs2_free_refcount_tree(tree);
@@ -575,7 +574,8 @@ static int ocfs2_create_refcount_tree(struct inode *inode,
BUG_ON(oi->ip_dyn_features & OCFS2_HAS_REFCOUNT_FL);
- mlog(0, "create tree for inode %lu\n", inode->i_ino);
+ trace_ocfs2_create_refcount_tree(
+ (unsigned long long)OCFS2_I(inode)->ip_blkno);
ret = ocfs2_reserve_new_metadata_blocks(osb, 1, &meta_ac);
if (ret) {
@@ -646,8 +646,7 @@ static int ocfs2_create_refcount_tree(struct inode *inode,
di->i_refcount_loc = cpu_to_le64(first_blkno);
spin_unlock(&oi->ip_lock);
- mlog(0, "created tree for inode %lu, refblock %llu\n",
- inode->i_ino, (unsigned long long)first_blkno);
+ trace_ocfs2_create_refcount_tree_blkno((unsigned long long)first_blkno);
ocfs2_journal_dirty(handle, di_bh);
@@ -1256,8 +1255,9 @@ static int ocfs2_change_refcount_rec(handle_t *handle,
goto out;
}
- mlog(0, "change index %d, old count %u, change %d\n", index,
- le32_to_cpu(rec->r_refcount), change);
+ trace_ocfs2_change_refcount_rec(
+ (unsigned long long)ocfs2_metadata_cache_owner(ci),
+ index, le32_to_cpu(rec->r_refcount), change);
le32_add_cpu(&rec->r_refcount, change);
if (!rec->r_refcount) {
@@ -1353,8 +1353,8 @@ static int ocfs2_expand_inline_ref_root(handle_t *handle,
ocfs2_journal_dirty(handle, ref_root_bh);
- mlog(0, "new leaf block %llu, used %u\n", (unsigned long long)blkno,
- le16_to_cpu(new_rb->rf_records.rl_used));
+ trace_ocfs2_expand_inline_ref_root((unsigned long long)blkno,
+ le16_to_cpu(new_rb->rf_records.rl_used));
*ref_leaf_bh = new_bh;
new_bh = NULL;
@@ -1466,9 +1466,9 @@ static int ocfs2_divide_leaf_refcount_block(struct buffer_head *ref_leaf_bh,
(struct ocfs2_refcount_block *)new_bh->b_data;
struct ocfs2_refcount_list *new_rl = &new_rb->rf_records;
- mlog(0, "split old leaf refcount block %llu, count = %u, used = %u\n",
- (unsigned long long)ref_leaf_bh->b_blocknr,
- le32_to_cpu(rl->rl_count), le32_to_cpu(rl->rl_used));
+ trace_ocfs2_divide_leaf_refcount_block(
+ (unsigned long long)ref_leaf_bh->b_blocknr,
+ le32_to_cpu(rl->rl_count), le32_to_cpu(rl->rl_used));
/*
* XXX: Improvement later.
@@ -1601,8 +1601,8 @@ static int ocfs2_new_leaf_refcount_block(handle_t *handle,
ocfs2_init_refcount_extent_tree(&ref_et, ci, ref_root_bh);
- mlog(0, "insert new leaf block %llu at %u\n",
- (unsigned long long)new_bh->b_blocknr, new_cpos);
+ trace_ocfs2_new_leaf_refcount_block(
+ (unsigned long long)new_bh->b_blocknr, new_cpos);
/* Insert the new leaf block with the specific offset cpos. */
ret = ocfs2_insert_extent(handle, &ref_et, new_cpos, new_bh->b_blocknr,
@@ -1794,11 +1794,10 @@ static int ocfs2_insert_refcount_rec(handle_t *handle,
(le16_to_cpu(rf_list->rl_used) - index) *
sizeof(struct ocfs2_refcount_rec));
- mlog(0, "insert refcount record start %llu, len %u, count %u "
- "to leaf block %llu at index %d\n",
- (unsigned long long)le64_to_cpu(rec->r_cpos),
- le32_to_cpu(rec->r_clusters), le32_to_cpu(rec->r_refcount),
- (unsigned long long)ref_leaf_bh->b_blocknr, index);
+ trace_ocfs2_insert_refcount_rec(
+ (unsigned long long)ref_leaf_bh->b_blocknr, index,
+ (unsigned long long)le64_to_cpu(rec->r_cpos),
+ le32_to_cpu(rec->r_clusters), le32_to_cpu(rec->r_refcount));
rf_list->rl_recs[index] = *rec;
@@ -1850,10 +1849,12 @@ static int ocfs2_split_refcount_rec(handle_t *handle,
BUG_ON(le32_to_cpu(rb->rf_flags) & OCFS2_REFCOUNT_TREE_FL);
- mlog(0, "original r_pos %llu, cluster %u, split %llu, cluster %u\n",
- le64_to_cpu(orig_rec->r_cpos), le32_to_cpu(orig_rec->r_clusters),
- le64_to_cpu(split_rec->r_cpos),
- le32_to_cpu(split_rec->r_clusters));
+ trace_ocfs2_split_refcount_rec(le64_to_cpu(orig_rec->r_cpos),
+ le32_to_cpu(orig_rec->r_clusters),
+ le32_to_cpu(orig_rec->r_refcount),
+ le64_to_cpu(split_rec->r_cpos),
+ le32_to_cpu(split_rec->r_clusters),
+ le32_to_cpu(split_rec->r_refcount));
/*
* If we just need to split the header or tail clusters,
@@ -1967,12 +1968,11 @@ static int ocfs2_split_refcount_rec(handle_t *handle,
if (split_rec->r_refcount) {
rf_list->rl_recs[index] = *split_rec;
- mlog(0, "insert refcount record start %llu, len %u, count %u "
- "to leaf block %llu at index %d\n",
- (unsigned long long)le64_to_cpu(split_rec->r_cpos),
- le32_to_cpu(split_rec->r_clusters),
- le32_to_cpu(split_rec->r_refcount),
- (unsigned long long)ref_leaf_bh->b_blocknr, index);
+ trace_ocfs2_split_refcount_rec_insert(
+ (unsigned long long)ref_leaf_bh->b_blocknr, index,
+ (unsigned long long)le64_to_cpu(split_rec->r_cpos),
+ le32_to_cpu(split_rec->r_clusters),
+ le32_to_cpu(split_rec->r_refcount));
if (merge)
ocfs2_refcount_rec_merge(rb, index);
@@ -1997,7 +1997,7 @@ static int __ocfs2_increase_refcount(handle_t *handle,
struct ocfs2_refcount_rec rec;
unsigned int set_len = 0;
- mlog(0, "Tree owner %llu, add refcount start %llu, len %u\n",
+ trace_ocfs2_increase_refcount_begin(
(unsigned long long)ocfs2_metadata_cache_owner(ci),
(unsigned long long)cpos, len);
@@ -2024,9 +2024,9 @@ static int __ocfs2_increase_refcount(handle_t *handle,
*/
if (rec.r_refcount && le64_to_cpu(rec.r_cpos) == cpos &&
set_len <= len) {
- mlog(0, "increase refcount rec, start %llu, len %u, "
- "count %u\n", (unsigned long long)cpos, set_len,
- le32_to_cpu(rec.r_refcount));
+ trace_ocfs2_increase_refcount_change(
+ (unsigned long long)cpos, set_len,
+ le32_to_cpu(rec.r_refcount));
ret = ocfs2_change_refcount_rec(handle, ci,
ref_leaf_bh, index,
merge, 1);
@@ -2037,7 +2037,7 @@ static int __ocfs2_increase_refcount(handle_t *handle,
} else if (!rec.r_refcount) {
rec.r_refcount = cpu_to_le32(1);
- mlog(0, "insert refcount rec, start %llu, len %u\n",
+ trace_ocfs2_increase_refcount_insert(
(unsigned long long)le64_to_cpu(rec.r_cpos),
set_len);
ret = ocfs2_insert_refcount_rec(handle, ci, ref_root_bh,
@@ -2055,8 +2055,7 @@ static int __ocfs2_increase_refcount(handle_t *handle,
rec.r_clusters = cpu_to_le32(set_len);
le32_add_cpu(&rec.r_refcount, 1);
- mlog(0, "split refcount rec, start %llu, "
- "len %u, count %u\n",
+ trace_ocfs2_increase_refcount_split(
(unsigned long long)le64_to_cpu(rec.r_cpos),
set_len, le32_to_cpu(rec.r_refcount));
ret = ocfs2_split_refcount_rec(handle, ci,
@@ -2095,6 +2094,11 @@ static int ocfs2_remove_refcount_extent(handle_t *handle,
BUG_ON(rb->rf_records.rl_used);
+ trace_ocfs2_remove_refcount_extent(
+ (unsigned long long)ocfs2_metadata_cache_owner(ci),
+ (unsigned long long)ref_leaf_bh->b_blocknr,
+ le32_to_cpu(rb->rf_cpos));
+
ocfs2_init_refcount_extent_tree(&et, ci, ref_root_bh);
ret = ocfs2_remove_extent(handle, &et, le32_to_cpu(rb->rf_cpos),
1, meta_ac, dealloc);
@@ -2137,7 +2141,7 @@ static int ocfs2_remove_refcount_extent(handle_t *handle,
if (!rb->rf_list.l_next_free_rec) {
BUG_ON(rb->rf_clusters);
- mlog(0, "reset refcount tree root %llu to be a record block.\n",
+ trace_ocfs2_restore_refcount_block(
(unsigned long long)ref_root_bh->b_blocknr);
rb->rf_flags = 0;
@@ -2184,6 +2188,10 @@ static int ocfs2_decrease_refcount_rec(handle_t *handle,
BUG_ON(cpos + len >
le64_to_cpu(rec->r_cpos) + le32_to_cpu(rec->r_clusters));
+ trace_ocfs2_decrease_refcount_rec(
+ (unsigned long long)ocfs2_metadata_cache_owner(ci),
+ (unsigned long long)cpos, len);
+
if (cpos == le64_to_cpu(rec->r_cpos) &&
len == le32_to_cpu(rec->r_clusters))
ret = ocfs2_change_refcount_rec(handle, ci,
@@ -2195,12 +2203,6 @@ static int ocfs2_decrease_refcount_rec(handle_t *handle,
le32_add_cpu(&split.r_refcount, -1);
- mlog(0, "split refcount rec, start %llu, "
- "len %u, count %u, original start %llu, len %u\n",
- (unsigned long long)le64_to_cpu(split.r_cpos),
- len, le32_to_cpu(split.r_refcount),
- (unsigned long long)le64_to_cpu(rec->r_cpos),
- le32_to_cpu(rec->r_clusters));
ret = ocfs2_split_refcount_rec(handle, ci,
ref_root_bh, ref_leaf_bh,
&split, index, 1,
@@ -2239,10 +2241,9 @@ static int __ocfs2_decrease_refcount(handle_t *handle,
struct super_block *sb = ocfs2_metadata_cache_get_super(ci);
struct buffer_head *ref_leaf_bh = NULL;
- mlog(0, "Tree owner %llu, decrease refcount start %llu, "
- "len %u, delete %u\n",
- (unsigned long long)ocfs2_metadata_cache_owner(ci),
- (unsigned long long)cpos, len, delete);
+ trace_ocfs2_decrease_refcount(
+ (unsigned long long)ocfs2_metadata_cache_owner(ci),
+ (unsigned long long)cpos, len, delete);
while (len) {
ret = ocfs2_get_refcount_rec(ci, ref_root_bh,
@@ -2352,8 +2353,8 @@ static int ocfs2_mark_extent_refcounted(struct inode *inode,
{
int ret;
- mlog(0, "Inode %lu refcount tree cpos %u, len %u, phys cluster %u\n",
- inode->i_ino, cpos, len, phys);
+ trace_ocfs2_mark_extent_refcounted(OCFS2_I(inode)->ip_blkno,
+ cpos, len, phys);
if (!ocfs2_refcount_tree(OCFS2_SB(inode->i_sb))) {
ocfs2_error(inode->i_sb, "Inode %lu want to use refcount "
@@ -2392,8 +2393,6 @@ static int ocfs2_calc_refcount_meta_credits(struct super_block *sb,
struct buffer_head *ref_leaf_bh = NULL, *prev_bh = NULL;
u32 len;
- mlog(0, "start_cpos %llu, clusters %u\n",
- (unsigned long long)start_cpos, clusters);
while (clusters) {
ret = ocfs2_get_refcount_rec(ci, ref_root_bh,
cpos, clusters, &rec,
@@ -2427,12 +2426,11 @@ static int ocfs2_calc_refcount_meta_credits(struct super_block *sb,
rb = (struct ocfs2_refcount_block *)ref_leaf_bh->b_data;
- mlog(0, "recs_add %d,cpos %llu, clusters %u, rec->r_cpos %llu,"
- "rec->r_clusters %u, rec->r_refcount %u, index %d\n",
- recs_add, (unsigned long long)cpos, clusters,
- (unsigned long long)le64_to_cpu(rec.r_cpos),
- le32_to_cpu(rec.r_clusters),
- le32_to_cpu(rec.r_refcount), index);
+ trace_ocfs2_calc_refcount_meta_credits_iterate(
+ recs_add, (unsigned long long)cpos, clusters,
+ (unsigned long long)le64_to_cpu(rec.r_cpos),
+ le32_to_cpu(rec.r_clusters),
+ le32_to_cpu(rec.r_refcount), index);
len = min((u64)cpos + clusters, le64_to_cpu(rec.r_cpos) +
le32_to_cpu(rec.r_clusters)) - cpos;
@@ -2488,7 +2486,6 @@ static int ocfs2_calc_refcount_meta_credits(struct super_block *sb,
if (!ref_blocks)
goto out;
- mlog(0, "we need ref_blocks %d\n", ref_blocks);
*meta_add += ref_blocks;
*credits += ref_blocks;
@@ -2514,6 +2511,10 @@ static int ocfs2_calc_refcount_meta_credits(struct super_block *sb,
}
out:
+
+ trace_ocfs2_calc_refcount_meta_credits(
+ (unsigned long long)start_cpos, clusters,
+ *meta_add, *credits);
brelse(ref_leaf_bh);
brelse(prev_bh);
return ret;
@@ -2578,8 +2579,7 @@ int ocfs2_prepare_refcount_change_for_del(struct inode *inode,
goto out;
}
- mlog(0, "reserve new metadata %d blocks, credits = %d\n",
- *ref_blocks, *credits);
+ trace_ocfs2_prepare_refcount_change_for_del(*ref_blocks, *credits);
out:
brelse(ref_root_bh);
@@ -2886,8 +2886,7 @@ static int ocfs2_lock_refcount_allocators(struct super_block *sb,
goto out;
}
- mlog(0, "reserve new metadata %d, clusters %u, credits = %d\n",
- meta_add, num_clusters, *credits);
+ trace_ocfs2_lock_refcount_allocators(meta_add, *credits);
ret = ocfs2_reserve_new_metadata_blocks(OCFS2_SB(sb), meta_add,
meta_ac);
if (ret) {
@@ -2937,8 +2936,8 @@ static int ocfs2_duplicate_clusters_by_page(handle_t *handle,
loff_t offset, end, map_end;
struct address_space *mapping = context->inode->i_mapping;
- mlog(0, "old_cluster %u, new %u, len %u at offset %u\n", old_cluster,
- new_cluster, new_len, cpos);
+ trace_ocfs2_duplicate_clusters_by_page(cpos, old_cluster,
+ new_cluster, new_len);
readahead_pages =
(ocfs2_cow_contig_clusters(sb) <<
@@ -3031,8 +3030,8 @@ static int ocfs2_duplicate_clusters_by_jbd(handle_t *handle,
struct buffer_head *old_bh = NULL;
struct buffer_head *new_bh = NULL;
- mlog(0, "old_cluster %u, new %u, len %u\n", old_cluster,
- new_cluster, new_len);
+ trace_ocfs2_duplicate_clusters_by_page(cpos, old_cluster,
+ new_cluster, new_len);
for (i = 0; i < blocks; i++, old_block++, new_block++) {
new_bh = sb_getblk(osb->sb, new_block);
@@ -3085,8 +3084,8 @@ static int ocfs2_clear_ext_refcount(handle_t *handle,
struct super_block *sb = ocfs2_metadata_cache_get_super(et->et_ci);
u64 ino = ocfs2_metadata_cache_owner(et->et_ci);
- mlog(0, "inode %llu cpos %u, len %u, p_cluster %u, ext_flags %u\n",
- (unsigned long long)ino, cpos, len, p_cluster, ext_flags);
+ trace_ocfs2_clear_ext_refcount((unsigned long long)ino,
+ cpos, len, p_cluster, ext_flags);
memset(&replace_rec, 0, sizeof(replace_rec));
replace_rec.e_cpos = cpu_to_le32(cpos);
@@ -3141,8 +3140,8 @@ static int ocfs2_replace_clusters(handle_t *handle,
struct ocfs2_caching_info *ci = context->data_et.et_ci;
u64 ino = ocfs2_metadata_cache_owner(ci);
- mlog(0, "inode %llu, cpos %u, old %u, new %u, len %u, ext_flags %u\n",
- (unsigned long long)ino, cpos, old, new, len, ext_flags);
+ trace_ocfs2_replace_clusters((unsigned long long)ino,
+ cpos, old, new, len, ext_flags);
/*If the old clusters is unwritten, no need to duplicate. */
if (!(ext_flags & OCFS2_EXT_UNWRITTEN)) {
@@ -3236,8 +3235,8 @@ static int ocfs2_make_clusters_writable(struct super_block *sb,
struct ocfs2_caching_info *ref_ci = &context->ref_tree->rf_ci;
struct ocfs2_refcount_rec rec;
- mlog(0, "cpos %u, p_cluster %u, num_clusters %u, e_flags %u\n",
- cpos, p_cluster, num_clusters, e_flags);
+ trace_ocfs2_make_clusters_writable(cpos, p_cluster,
+ num_clusters, e_flags);
ret = ocfs2_lock_refcount_allocators(sb, p_cluster, num_clusters,
&context->data_et,
@@ -3472,9 +3471,9 @@ static int ocfs2_refcount_cow_hunk(struct inode *inode,
goto out;
}
- mlog(0, "CoW inode %lu, cpos %u, write_len %u, cow_start %u, "
- "cow_len %u\n", inode->i_ino,
- cpos, write_len, cow_start, cow_len);
+ trace_ocfs2_refcount_cow_hunk(OCFS2_I(inode)->ip_blkno,
+ cpos, write_len, max_cpos,
+ cow_start, cow_len);
BUG_ON(cow_len == 0);
@@ -3753,8 +3752,7 @@ int ocfs2_add_refcount_flag(struct inode *inode,
goto out;
}
- mlog(0, "reserve new metadata %d, credits = %d\n",
- ref_blocks, credits);
+ trace_ocfs2_add_refcount_flag(ref_blocks, credits);
if (ref_blocks) {
ret = ocfs2_reserve_new_metadata_blocks(OCFS2_SB(inode->i_sb),
--
1.7.0.4
^ permalink raw reply related [flat|nested] 53+ messages in thread
* [Ocfs2-devel] [PATCH 11/34] ocfs2: Remove mlog(0) from fs/ocfs2/aops.c
2010-12-23 7:19 [Ocfs2-devel] [PATCH 00/34] OCFS2: Add trace event and replace mlog(0) Tao Ma
` (9 preceding siblings ...)
2010-12-23 7:30 ` [Ocfs2-devel] [PATCH 10/34] ocfs2: Remove masklog ML_REFCOUNT Tao Ma
@ 2010-12-23 7:30 ` Tao Ma
2010-12-23 7:30 ` [Ocfs2-devel] [PATCH 12/34] ocfs2: Remove FILE_IO from masklog Tao Ma
` (27 subsequent siblings)
38 siblings, 0 replies; 53+ messages in thread
From: Tao Ma @ 2010-12-23 7:30 UTC (permalink / raw)
To: ocfs2-devel
From: Tao Ma <tao.ma@oracle.com>
Remove all the "mlog(0," in fs/ocfs2/aops.c.
Signed-off-by: Tao Ma <tao.ma@oracle.com>
---
fs/ocfs2/aops.c | 55 ++++++++--------
fs/ocfs2/ocfs2_trace.h | 168 ++++++++++++++++++++++++++++++++++++++++++++++++
2 files changed, 197 insertions(+), 26 deletions(-)
diff --git a/fs/ocfs2/aops.c b/fs/ocfs2/aops.c
index 177b8b4..e6153cb 100644
--- a/fs/ocfs2/aops.c
+++ b/fs/ocfs2/aops.c
@@ -29,7 +29,6 @@
#include <linux/mpage.h>
#include <linux/quotaops.h>
-#define MLOG_MASK_PREFIX ML_FILE_IO
#include <cluster/masklog.h>
#include "ocfs2.h"
@@ -45,6 +44,7 @@
#include "super.h"
#include "symlink.h"
#include "refcounttree.h"
+#include "ocfs2_trace.h"
#include "buffer_head_io.h"
@@ -59,8 +59,9 @@ static int ocfs2_symlink_get_block(struct inode *inode, sector_t iblock,
struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
void *kaddr;
- mlog(0, "(0x%p, %llu, 0x%p, %d)\n", inode,
- (unsigned long long)iblock, bh_result, create);
+ trace_ocfs2_symlink_get_block(
+ (unsigned long long)OCFS2_I(inode)->ip_blkno,
+ (unsigned long long)iblock, bh_result, create);
BUG_ON(ocfs2_inode_is_fast_symlink(inode));
@@ -135,8 +136,8 @@ int ocfs2_get_block(struct inode *inode, sector_t iblock,
u64 p_blkno, count, past_eof;
struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
- mlog(0, "(0x%p, %llu, 0x%p, %d)\n", inode,
- (unsigned long long)iblock, bh_result, create);
+ trace_ocfs2_get_block((unsigned long long)OCFS2_I(inode)->ip_blkno,
+ (unsigned long long)iblock, bh_result, create);
if (OCFS2_I(inode)->ip_flags & OCFS2_INODE_SYSTEM_FILE)
mlog(ML_NOTICE, "get_block on system inode 0x%p (%lu)\n",
@@ -198,8 +199,9 @@ int ocfs2_get_block(struct inode *inode, sector_t iblock,
}
past_eof = ocfs2_blocks_for_bytes(inode->i_sb, i_size_read(inode));
- mlog(0, "Inode %lu, past_eof = %llu\n", inode->i_ino,
- (unsigned long long)past_eof);
+
+ trace_ocfs2_get_block_end((unsigned long long)OCFS2_I(inode)->ip_blkno,
+ (unsigned long long)past_eof);
if (create && (iblock >= past_eof))
set_buffer_new(bh_result);
@@ -276,7 +278,8 @@ static int ocfs2_readpage(struct file *file, struct page *page)
loff_t start = (loff_t)page->index << PAGE_CACHE_SHIFT;
int ret, unlock = 1;
- mlog(0, "(0x%p, %lu)\n", file, (page ? page->index : 0));
+ trace_ocfs2_readpage((unsigned long long)oi->ip_blkno,
+ (page ? page->index : 0));
ret = ocfs2_inode_lock_with_page(inode, NULL, 0, page);
if (ret != 0) {
@@ -393,13 +396,11 @@ out_unlock:
*/
static int ocfs2_writepage(struct page *page, struct writeback_control *wbc)
{
- int ret;
-
- mlog(0, "(0x%p)\n", page);
-
- ret = block_write_full_page(page, ocfs2_get_block, wbc);
+ trace_ocfs2_writepage(
+ (unsigned long long)OCFS2_I(page->mapping->host)->ip_blkno,
+ page->index);
- return ret;
+ return block_write_full_page(page, ocfs2_get_block, wbc);
}
/* Taken from ext3. We don't necessarily need the full blown
@@ -445,7 +446,8 @@ static sector_t ocfs2_bmap(struct address_space *mapping, sector_t block)
int err = 0;
struct inode *inode = mapping->host;
- mlog(0, "(block = %llu)\n", (unsigned long long)block);
+ trace_ocfs2_bmap((unsigned long long)OCFS2_I(inode)->ip_blkno,
+ (unsigned long long)block);
/* We don't need to lock journal system files, since they aren't
* accessed concurrently from multiple nodes.
@@ -1518,9 +1520,9 @@ static int ocfs2_try_to_write_inline_data(struct address_space *mapping,
struct ocfs2_inode_info *oi = OCFS2_I(inode);
struct ocfs2_dinode *di = NULL;
- mlog(0, "Inode %llu, write of %u bytes at off %llu. features: 0x%x\n",
- (unsigned long long)oi->ip_blkno, len, (unsigned long long)pos,
- oi->ip_dyn_features);
+ trace_ocfs2_try_to_write_inline_data((unsigned long long)oi->ip_blkno,
+ len, (unsigned long long)pos,
+ oi->ip_dyn_features);
/*
* Handle inodes which already have inline data 1st.
@@ -1682,6 +1684,13 @@ int ocfs2_write_begin_nolock(struct file *filp,
di = (struct ocfs2_dinode *)wc->w_di_bh->b_data;
+ trace_ocfs2_write_begin_nolock(
+ (unsigned long long)OCFS2_I(inode)->ip_blkno,
+ (long long)i_size_read(inode),
+ le32_to_cpu(di->i_clusters),
+ pos, len, flags, mmap_page,
+ clusters_to_alloc, extents_to_split);
+
/*
* We set w_target_from, w_target_to here so that
* ocfs2_write_end() knows which range in the target page to
@@ -1694,12 +1703,6 @@ int ocfs2_write_begin_nolock(struct file *filp,
* ocfs2_lock_allocators(). It greatly over-estimates
* the work to be done.
*/
- mlog(0, "extend inode %llu, i_size = %lld, di->i_clusters = %u,"
- " clusters_to_add = %u, extents_to_split = %u\n",
- (unsigned long long)OCFS2_I(inode)->ip_blkno,
- (long long)i_size_read(inode), le32_to_cpu(di->i_clusters),
- clusters_to_alloc, extents_to_split);
-
ocfs2_init_dinode_extent_tree(&et, INODE_CACHE(inode),
wc->w_di_bh);
ret = ocfs2_lock_allocators(inode, &et,
@@ -1865,8 +1868,8 @@ static void ocfs2_write_end_inline(struct inode *inode, loff_t pos,
memcpy(di->id2.i_data.id_data + pos, kaddr + pos, *copied);
kunmap_atomic(kaddr, KM_USER0);
- mlog(0, "Data written to inode at offset %llu. "
- "id_count = %u, copied = %u, i_dyn_features = 0x%x\n",
+ trace_ocfs2_write_end_inline(
+ (unsigned long long)OCFS2_I(inode)->ip_blkno,
(unsigned long long)pos, *copied,
le16_to_cpu(di->id2.i_data.id_count),
le16_to_cpu(di->i_dyn_features));
diff --git a/fs/ocfs2/ocfs2_trace.h b/fs/ocfs2/ocfs2_trace.h
index 3671151..26d9a05 100644
--- a/fs/ocfs2/ocfs2_trace.h
+++ b/fs/ocfs2/ocfs2_trace.h
@@ -1462,6 +1462,174 @@ TRACE_EVENT(ocfs2_refcount_cow_hunk,
);
/* End of trace events for fs/ocfs2/refcounttree.c. */
+
+/* Trace events for fs/ocfs2/aops.c. */
+DECLARE_EVENT_CLASS(ocfs2__get_block,
+
+ TP_PROTO(unsigned long long ino, unsigned long long iblock,
+ void *bh_result, int create),
+
+ TP_ARGS(ino, iblock, bh_result, create),
+
+ TP_STRUCT__entry(
+ __field( unsigned long long, ino )
+ __field( unsigned long long, iblock )
+ __field( void *, bh_result )
+ __field( int, create )
+ ),
+
+ TP_fast_assign(
+ __entry->ino = ino;
+ __entry->iblock = iblock;
+ __entry->bh_result = bh_result;
+ __entry->create = create;
+ ),
+
+ TP_printk("%llu %llu %p %d",
+ __entry->ino, __entry->iblock,
+ __entry->bh_result, __entry->create)
+);
+
+DEFINE_EVENT(ocfs2__get_block, ocfs2_symlink_get_block,
+
+ TP_PROTO(unsigned long long ino, unsigned long long iblock,
+ void *bh_result, int create),
+
+ TP_ARGS(ino, iblock, bh_result, create)
+);
+
+DEFINE_EVENT(ocfs2__get_block, ocfs2_get_block,
+
+ TP_PROTO(unsigned long long ino, unsigned long long iblock,
+ void *bh_result, int create),
+
+ TP_ARGS(ino, iblock, bh_result, create)
+);
+
+DEFINE_EVENT(ocfs2__ull_ull, ocfs2_get_block_end,
+
+ TP_PROTO(unsigned long long ino, unsigned long long pass_eof),
+
+ TP_ARGS(ino, pass_eof)
+);
+
+DEFINE_EVENT(ocfs2__ull_ull, ocfs2_readpage,
+
+ TP_PROTO(unsigned long long ino, unsigned long long page_index),
+
+ TP_ARGS(ino, page_index)
+);
+
+DEFINE_EVENT(ocfs2__ull_ull, ocfs2_writepage,
+
+ TP_PROTO(unsigned long long ino, unsigned long long page_index),
+
+ TP_ARGS(ino, page_index)
+);
+
+DEFINE_EVENT(ocfs2__ull_ull, ocfs2_bmap,
+
+ TP_PROTO(unsigned long long ino, unsigned long long block),
+
+ TP_ARGS(ino, block)
+);
+
+TRACE_EVENT(ocfs2_try_to_write_inline_data,
+
+ TP_PROTO(unsigned long long ino, unsigned int len,
+ unsigned long long pos, unsigned int flags),
+
+ TP_ARGS(ino, len, pos, flags),
+
+ TP_STRUCT__entry(
+ __field( unsigned long long, ino )
+ __field( unsigned int, len )
+ __field( unsigned long long, pos )
+ __field( unsigned int, flags )
+ ),
+
+ TP_fast_assign(
+ __entry->ino = ino;
+ __entry->len = len;
+ __entry->pos = pos;
+ __entry->flags = flags;
+ ),
+
+ TP_printk("%llu %u %llu 0x%x",
+ __entry->ino, __entry->len, __entry->pos, __entry->flags)
+);
+
+TRACE_EVENT(ocfs2_write_begin_nolock,
+
+ TP_PROTO(unsigned long long ino,
+ long long i_size, unsigned int i_clusters,
+ unsigned long long pos, unsigned int len,
+ unsigned int flags, void *page,
+ unsigned int clusters, unsigned int extents_to_split),
+
+ TP_ARGS(ino, i_size, i_clusters, pos, len, flags,
+ page, clusters, extents_to_split),
+
+ TP_STRUCT__entry(
+ __field( unsigned long long, ino )
+ __field( long long, i_size )
+ __field( unsigned int, i_clusters )
+ __field( unsigned long long, pos )
+ __field( unsigned int, len )
+ __field( unsigned int, flags )
+ __field( void *, page )
+ __field( unsigned int, clusters )
+ __field( unsigned int, extents_to_split)
+ ),
+
+ TP_fast_assign(
+ __entry->ino = ino;
+ __entry->i_size = i_size;
+ __entry->i_clusters = i_clusters;
+ __entry->pos = pos;
+ __entry->len = len;
+ __entry->flags = flags;
+ __entry->page = page;
+ __entry->clusters = clusters;
+ __entry->extents_to_split = extents_to_split;
+ ),
+
+ TP_printk("%llu %lld %u %llu %u %u %p %u %u",
+ __entry->ino, __entry->i_size, __entry->i_clusters,
+ __entry->pos, __entry->len,
+ __entry->flags, __entry->page, __entry->clusters,
+ __entry->extents_to_split)
+);
+
+TRACE_EVENT(ocfs2_write_end_inline,
+
+ TP_PROTO(unsigned long long ino,
+ unsigned long long pos, unsigned int copied,
+ unsigned int id_count, unsigned int features),
+
+ TP_ARGS(ino, pos, copied, id_count, features),
+
+ TP_STRUCT__entry(
+ __field( unsigned long long, ino )
+ __field( unsigned long long, pos )
+ __field( unsigned int, copied )
+ __field( unsigned int, id_count )
+ __field( unsigned int, features )
+ ),
+
+ TP_fast_assign(
+ __entry->ino = ino;
+ __entry->pos = pos;
+ __entry->copied = copied;
+ __entry->id_count = id_count;
+ __entry->features = features;
+ ),
+
+ TP_printk("%llu %llu %u %u %u",
+ __entry->ino, __entry->pos, __entry->copied,
+ __entry->id_count, __entry->features)
+);
+/* End of trace events for fs/ocfs2/aops.c. */
#endif /* _TRACE_OCFS2_H */
/* This part must be outside protection */
--
1.7.0.4
^ permalink raw reply related [flat|nested] 53+ messages in thread
* [Ocfs2-devel] [PATCH 12/34] ocfs2: Remove FILE_IO from masklog.
2010-12-23 7:19 [Ocfs2-devel] [PATCH 00/34] OCFS2: Add trace event and replace mlog(0) Tao Ma
` (10 preceding siblings ...)
2010-12-23 7:30 ` [Ocfs2-devel] [PATCH 11/34] ocfs2: Remove mlog(0) from fs/ocfs2/aops.c Tao Ma
@ 2010-12-23 7:30 ` Tao Ma
2010-12-23 7:30 ` [Ocfs2-devel] [PATCH 13/34] ocfs2: remove INODE from unused files Tao Ma
` (26 subsequent siblings)
38 siblings, 0 replies; 53+ messages in thread
From: Tao Ma @ 2010-12-23 7:30 UTC (permalink / raw)
To: ocfs2-devel
From: Tao Ma <tao.ma@oracle.com>
Change all the "mlog(0," in fs/ocfs2/mmap.c to trace events.
And finally remove masklog FILE_IO.
Signed-off-by: Tao Ma <tao.ma@oracle.com>
---
fs/ocfs2/cluster/masklog.c | 1 -
fs/ocfs2/cluster/masklog.h | 1 -
fs/ocfs2/mmap.c | 7 +++----
fs/ocfs2/ocfs2_trace.h | 28 ++++++++++++++++++++++++++++
4 files changed, 31 insertions(+), 6 deletions(-)
diff --git a/fs/ocfs2/cluster/masklog.c b/fs/ocfs2/cluster/masklog.c
index b8939c4..c654038 100644
--- a/fs/ocfs2/cluster/masklog.c
+++ b/fs/ocfs2/cluster/masklog.c
@@ -94,7 +94,6 @@ static struct mlog_attribute mlog_attrs[MLOG_MAX_BITS] = {
define_mask(AIO),
define_mask(JOURNAL),
define_mask(SUPER),
- define_mask(FILE_IO),
define_mask(EXTENT_MAP),
define_mask(DLM_GLUE),
define_mask(BH_IO),
diff --git a/fs/ocfs2/cluster/masklog.h b/fs/ocfs2/cluster/masklog.h
index 8379361..de412aa 100644
--- a/fs/ocfs2/cluster/masklog.h
+++ b/fs/ocfs2/cluster/masklog.h
@@ -96,7 +96,6 @@
#define ML_AIO 0x0000000000002000ULL /* ocfs2 aio read and write */
#define ML_JOURNAL 0x0000000000004000ULL /* ocfs2 journalling functions */
#define ML_SUPER 0x0000000000010000ULL /* ocfs2 mount / umount */
-#define ML_FILE_IO 0x0000000000020000ULL /* ocfs2 file I/O */
#define ML_EXTENT_MAP 0x0000000000040000ULL /* ocfs2 extent map caching */
#define ML_DLM_GLUE 0x0000000000080000ULL /* ocfs2 dlm glue layer */
#define ML_BH_IO 0x0000000000100000ULL /* ocfs2 buffer I/O */
diff --git a/fs/ocfs2/mmap.c b/fs/ocfs2/mmap.c
index 49730b5..3e9393c 100644
--- a/fs/ocfs2/mmap.c
+++ b/fs/ocfs2/mmap.c
@@ -31,7 +31,6 @@
#include <linux/signal.h>
#include <linux/rbtree.h>
-#define MLOG_MASK_PREFIX ML_FILE_IO
#include <cluster/masklog.h>
#include "ocfs2.h"
@@ -42,6 +41,7 @@
#include "inode.h"
#include "mmap.h"
#include "super.h"
+#include "ocfs2_trace.h"
static int ocfs2_fault(struct vm_area_struct *area, struct vm_fault *vmf)
@@ -49,13 +49,12 @@ static int ocfs2_fault(struct vm_area_struct *area, struct vm_fault *vmf)
sigset_t oldset;
int ret;
- mlog(0, "(area=%p, page offset=%lu)\n", area, vmf->pgoff);
-
ocfs2_block_signals(&oldset);
ret = filemap_fault(area, vmf);
ocfs2_unblock_signals(&oldset);
- mlog(0, "%p\n", vmf->page);
+ trace_ocfs2_fault(OCFS2_I(area->vm_file->f_mapping->host)->ip_blkno,
+ area, vmf->page, vmf->pgoff);
return ret;
}
diff --git a/fs/ocfs2/ocfs2_trace.h b/fs/ocfs2/ocfs2_trace.h
index 26d9a05..efda83e 100644
--- a/fs/ocfs2/ocfs2_trace.h
+++ b/fs/ocfs2/ocfs2_trace.h
@@ -1630,6 +1630,34 @@ TRACE_EVENT(ocfs2_write_end_inline,
__entry->id_count, __entry->features)
);
/* End of trace events for fs/ocfs2/aops.c. */
+
+/* Trace events for fs/ocfs2/mmap.c. */
+TRACE_EVENT(ocfs2_fault,
+
+ TP_PROTO(unsigned long long ino,
+ void *area, void *page, unsigned long pgoff),
+
+ TP_ARGS(ino, area, page, pgoff),
+
+ TP_STRUCT__entry(
+ __field( unsigned long long, ino )
+ __field( void *, area )
+ __field( void *, page )
+ __field( unsigned long, pgoff )
+ ),
+
+ TP_fast_assign(
+ __entry->ino = ino;
+ __entry->area = area;
+ __entry->page = page;
+ __entry->pgoff = pgoff;
+ ),
+
+ TP_printk("%llu %p %p %lu",
+ __entry->ino, __entry->area, __entry->page, __entry->pgoff)
+);
+/* End of trace events for fs/ocfs2/mmap.c. */
+
#endif /* _TRACE_OCFS2_H */
/* This part must be outside protection */
--
1.7.0.4
^ permalink raw reply related [flat|nested] 53+ messages in thread
* [Ocfs2-devel] [PATCH 13/34] ocfs2: remove INODE from unused files.
2010-12-23 7:19 [Ocfs2-devel] [PATCH 00/34] OCFS2: Add trace event and replace mlog(0) Tao Ma
` (11 preceding siblings ...)
2010-12-23 7:30 ` [Ocfs2-devel] [PATCH 12/34] ocfs2: Remove FILE_IO from masklog Tao Ma
@ 2010-12-23 7:30 ` Tao Ma
2010-12-23 7:30 ` [Ocfs2-devel] [PATCH 14/34] ocfs2: Remove mlog(0) from fs/ocfs2/file.c Tao Ma
` (25 subsequent siblings)
38 siblings, 0 replies; 53+ messages in thread
From: Tao Ma @ 2010-12-23 7:30 UTC (permalink / raw)
To: ocfs2-devel
From: Tao Ma <boyu.mt@taobao.com>
As there are no such debug information in fs/ocfs2/ioctl.c,
fs/ocfs2/locks.c and fs/ocfs2/sysfile.c, ML_INODE are also
removed.
Signed-off-by: Tao Ma <boyu.mt@taobao.com>
---
fs/ocfs2/acl.c | 1 -
fs/ocfs2/ioctl.c | 1 -
fs/ocfs2/locks.c | 1 -
3 files changed, 0 insertions(+), 3 deletions(-)
diff --git a/fs/ocfs2/acl.c b/fs/ocfs2/acl.c
index 3919150..ed28db8 100644
--- a/fs/ocfs2/acl.c
+++ b/fs/ocfs2/acl.c
@@ -24,7 +24,6 @@
#include <linux/slab.h>
#include <linux/string.h>
-#define MLOG_MASK_PREFIX ML_INODE
#include <cluster/masklog.h>
#include "ocfs2.h"
diff --git a/fs/ocfs2/ioctl.c b/fs/ocfs2/ioctl.c
index b4f51e9..d9bfa90 100644
--- a/fs/ocfs2/ioctl.c
+++ b/fs/ocfs2/ioctl.c
@@ -9,7 +9,6 @@
#include <linux/mount.h>
#include <linux/compat.h>
-#define MLOG_MASK_PREFIX ML_INODE
#include <cluster/masklog.h>
#include "ocfs2.h"
diff --git a/fs/ocfs2/locks.c b/fs/ocfs2/locks.c
index b5cb3ed..e57c804 100644
--- a/fs/ocfs2/locks.c
+++ b/fs/ocfs2/locks.c
@@ -26,7 +26,6 @@
#include <linux/fs.h>
#include <linux/fcntl.h>
-#define MLOG_MASK_PREFIX ML_INODE
#include <cluster/masklog.h>
#include "ocfs2.h"
--
1.7.0.4
^ permalink raw reply related [flat|nested] 53+ messages in thread
* [Ocfs2-devel] [PATCH 14/34] ocfs2: Remove mlog(0) from fs/ocfs2/file.c
2010-12-23 7:19 [Ocfs2-devel] [PATCH 00/34] OCFS2: Add trace event and replace mlog(0) Tao Ma
` (12 preceding siblings ...)
2010-12-23 7:30 ` [Ocfs2-devel] [PATCH 13/34] ocfs2: remove INODE from unused files Tao Ma
@ 2010-12-23 7:30 ` Tao Ma
2010-12-23 7:30 ` [Ocfs2-devel] [PATCH 15/34] ocfs2: Little refactoring against ocfs2_iget Tao Ma
` (24 subsequent siblings)
38 siblings, 0 replies; 53+ messages in thread
From: Tao Ma @ 2010-12-23 7:30 UTC (permalink / raw)
To: ocfs2-devel
From: Tao Ma <boyu.mt@taobao.com>
This is the 2nd step to remove the debug info of INODE.
Signed-off-by: Tao Ma <boyu.mt@taobao.com>
---
fs/ocfs2/file.c | 190 ++++++++++++-------------
fs/ocfs2/ocfs2_trace.h | 367 ++++++++++++++++++++++++++++++++++++++++++++++++
fs/ocfs2/sysfile.c | 1 -
3 files changed, 460 insertions(+), 98 deletions(-)
diff --git a/fs/ocfs2/file.c b/fs/ocfs2/file.c
index 14f335d..868c686 100644
--- a/fs/ocfs2/file.c
+++ b/fs/ocfs2/file.c
@@ -38,7 +38,6 @@
#include <linux/quotaops.h>
#include <linux/blkdev.h>
-#define MLOG_MASK_PREFIX ML_INODE
#include <cluster/masklog.h>
#include "ocfs2.h"
@@ -61,6 +60,7 @@
#include "acl.h"
#include "quota.h"
#include "refcounttree.h"
+#include "ocfs2_trace.h"
#include "buffer_head_io.h"
@@ -99,8 +99,10 @@ static int ocfs2_file_open(struct inode *inode, struct file *file)
int mode = file->f_flags;
struct ocfs2_inode_info *oi = OCFS2_I(inode);
- mlog(0, "(0x%p, 0x%p, '%.*s')\n", inode, file,
- file->f_path.dentry->d_name.len, file->f_path.dentry->d_name.name);
+ trace_ocfs2_file_open(inode, file, file->f_path.dentry,
+ (unsigned long long)OCFS2_I(inode)->ip_blkno,
+ file->f_path.dentry->d_name.len,
+ file->f_path.dentry->d_name.name, mode);
if (file->f_mode & FMODE_WRITE)
dquot_initialize(inode);
@@ -142,13 +144,15 @@ static int ocfs2_file_release(struct inode *inode, struct file *file)
{
struct ocfs2_inode_info *oi = OCFS2_I(inode);
- mlog(0, "(0x%p, 0x%p, '%.*s')\n", inode, file,
- file->f_path.dentry->d_name.len,
- file->f_path.dentry->d_name.name);
-
spin_lock(&oi->ip_lock);
if (!--oi->ip_open_count)
oi->ip_flags &= ~OCFS2_INODE_OPEN_DIRECT;
+
+ trace_ocfs2_file_release(inode, file, file->f_path.dentry,
+ oi->ip_blkno,
+ file->f_path.dentry->d_name.len,
+ file->f_path.dentry->d_name.name,
+ oi->ip_open_count);
spin_unlock(&oi->ip_lock);
ocfs2_free_file_private(inode, file);
@@ -174,9 +178,11 @@ static int ocfs2_sync_file(struct file *file, int datasync)
struct inode *inode = file->f_mapping->host;
struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
- mlog(0, "(0x%p, %d, 0x%p, '%.*s')\n", file, datasync,
- file->f_path.dentry, file->f_path.dentry->d_name.len,
- file->f_path.dentry->d_name.name);
+ trace_ocfs2_sync_file(inode, file, file->f_path.dentry,
+ OCFS2_I(inode)->ip_blkno,
+ file->f_path.dentry->d_name.len,
+ file->f_path.dentry->d_name.name,
+ (unsigned long long)datasync);
if (datasync && !(inode->i_state & I_DIRTY_DATASYNC)) {
/*
@@ -431,14 +437,14 @@ static int ocfs2_truncate_file(struct inode *inode,
struct ocfs2_dinode *fe = NULL;
struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
- mlog(0, "(inode = %llu, new_i_size = %llu\n",
- (unsigned long long)OCFS2_I(inode)->ip_blkno,
- (unsigned long long)new_i_size);
-
/* We trust di_bh because it comes from ocfs2_inode_lock(), which
* already validated it */
fe = (struct ocfs2_dinode *) di_bh->b_data;
+ trace_ocfs2_truncate_file((unsigned long long)OCFS2_I(inode)->ip_blkno,
+ (unsigned long long)le64_to_cpu(fe->i_size),
+ (unsigned long long)new_i_size);
+
mlog_bug_on_msg(le64_to_cpu(fe->i_size) != i_size_read(inode),
"Inode %llu, inode i_size = %lld != di "
"i_size = %llu, i_flags = 0x%x\n",
@@ -448,19 +454,14 @@ static int ocfs2_truncate_file(struct inode *inode,
le32_to_cpu(fe->i_flags));
if (new_i_size > le64_to_cpu(fe->i_size)) {
- mlog(0, "asked to truncate file with size (%llu) to size (%llu)!\n",
- (unsigned long long)le64_to_cpu(fe->i_size),
- (unsigned long long)new_i_size);
+ trace_ocfs2_truncate_file_error(
+ (unsigned long long)le64_to_cpu(fe->i_size),
+ (unsigned long long)new_i_size);
status = -EINVAL;
mlog_errno(status);
goto bail;
}
- mlog(0, "inode %llu, i_size = %llu, new_i_size = %llu\n",
- (unsigned long long)le64_to_cpu(fe->i_blkno),
- (unsigned long long)le64_to_cpu(fe->i_size),
- (unsigned long long)new_i_size);
-
/* lets handle the simple truncate cases before doing any more
* cluster locking. */
if (new_i_size == le64_to_cpu(fe->i_size))
@@ -566,8 +567,6 @@ static int __ocfs2_extend_allocation(struct inode *inode, u32 logical_start,
struct ocfs2_extent_tree et;
int did_quota = 0;
- mlog(0, "(clusters_to_add = %u)\n", clusters_to_add);
-
/*
* This function only exists for file systems which don't
* support holes.
@@ -584,11 +583,6 @@ static int __ocfs2_extend_allocation(struct inode *inode, u32 logical_start,
restart_all:
BUG_ON(le32_to_cpu(fe->i_clusters) != OCFS2_I(inode)->ip_clusters);
- mlog(0, "extend inode %llu, i_size = %lld, di->i_clusters = %u, "
- "clusters_to_add = %u\n",
- (unsigned long long)OCFS2_I(inode)->ip_blkno,
- (long long)i_size_read(inode), le32_to_cpu(fe->i_clusters),
- clusters_to_add);
ocfs2_init_dinode_extent_tree(&et, INODE_CACHE(inode), bh);
status = ocfs2_lock_allocators(inode, &et, clusters_to_add, 0,
&data_ac, &meta_ac);
@@ -608,6 +602,12 @@ restart_all:
}
restarted_transaction:
+ trace_ocfs2_extend_allocation(
+ (unsigned long long)OCFS2_I(inode)->ip_blkno,
+ (unsigned long long)i_size_read(inode),
+ le32_to_cpu(fe->i_clusters), clusters_to_add,
+ why, restart_func);
+
status = dquot_alloc_space_nodirty(inode,
ocfs2_clusters_to_bytes(osb->sb, clusters_to_add));
if (status)
@@ -654,13 +654,11 @@ restarted_transaction:
if (why != RESTART_NONE && clusters_to_add) {
if (why == RESTART_META) {
- mlog(0, "restarting function.\n");
restart_func = 1;
status = 0;
} else {
BUG_ON(why != RESTART_TRANS);
- mlog(0, "restarting transaction.\n");
/* TODO: This can be more intelligent. */
credits = ocfs2_calc_extend_credits(osb->sb,
&fe->id2.i_list,
@@ -677,11 +675,11 @@ restarted_transaction:
}
}
- mlog(0, "fe: i_clusters = %u, i_size=%llu\n",
+ trace_ocfs2_extend_allocation_end(OCFS2_I(inode)->ip_blkno,
le32_to_cpu(fe->i_clusters),
- (unsigned long long)le64_to_cpu(fe->i_size));
- mlog(0, "inode: ip_clusters=%u, i_size=%lld\n",
- OCFS2_I(inode)->ip_clusters, (long long)i_size_read(inode));
+ (unsigned long long)le64_to_cpu(fe->i_size),
+ OCFS2_I(inode)->ip_clusters,
+ (unsigned long long)i_size_read(inode));
leave:
if (status < 0 && did_quota)
@@ -772,10 +770,11 @@ static int ocfs2_write_zero_page(struct inode *inode, u64 abs_from,
if (!zero_to)
zero_to = PAGE_CACHE_SIZE;
- mlog(0,
- "abs_from = %llu, abs_to = %llu, index = %lu, zero_from = %u, zero_to = %u\n",
- (unsigned long long)abs_from, (unsigned long long)abs_to,
- index, zero_from, zero_to);
+ trace_ocfs2_write_zero_page(
+ (unsigned long long)OCFS2_I(inode)->ip_blkno,
+ (unsigned long long)abs_from,
+ (unsigned long long)abs_to,
+ index, zero_from, zero_to);
/* We know that zero_from is block aligned */
for (block_start = zero_from; block_start < zero_to;
@@ -915,9 +914,10 @@ static int ocfs2_zero_extend_range(struct inode *inode, u64 range_start,
u64 next_pos;
u64 zero_pos = range_start;
- mlog(0, "range_start = %llu, range_end = %llu\n",
- (unsigned long long)range_start,
- (unsigned long long)range_end);
+ trace_ocfs2_zero_extend_range(
+ (unsigned long long)OCFS2_I(inode)->ip_blkno,
+ (unsigned long long)range_start,
+ (unsigned long long)range_end);
BUG_ON(range_start >= range_end);
while (zero_pos < range_end) {
@@ -949,9 +949,9 @@ int ocfs2_zero_extend(struct inode *inode, struct buffer_head *di_bh,
struct super_block *sb = inode->i_sb;
zero_start = ocfs2_align_bytes_to_blocks(sb, i_size_read(inode));
- mlog(0, "zero_start %llu for i_size %llu\n",
- (unsigned long long)zero_start,
- (unsigned long long)i_size_read(inode));
+ trace_ocfs2_zero_extend((unsigned long long)OCFS2_I(inode)->ip_blkno,
+ (unsigned long long)zero_start,
+ (unsigned long long)i_size_read(inode));
while (zero_start < zero_to_size) {
ret = ocfs2_zero_extend_get_range(inode, di_bh, zero_start,
zero_to_size,
@@ -1100,30 +1100,20 @@ int ocfs2_setattr(struct dentry *dentry, struct iattr *attr)
struct dquot *transfer_to[MAXQUOTAS] = { };
int qtype;
- mlog(0, "(0x%p, '%.*s')\n", dentry,
- dentry->d_name.len, dentry->d_name.name);
+ trace_ocfs2_setattr(inode, dentry,
+ (unsigned long long)OCFS2_I(inode)->ip_blkno,
+ dentry->d_name.len, dentry->d_name.name,
+ attr->ia_valid, attr->ia_mode,
+ attr->ia_uid, attr->ia_gid);
/* ensuring we don't even attempt to truncate a symlink */
if (S_ISLNK(inode->i_mode))
attr->ia_valid &= ~ATTR_SIZE;
- if (attr->ia_valid & ATTR_MODE)
- mlog(0, "mode change: %d\n", attr->ia_mode);
- if (attr->ia_valid & ATTR_UID)
- mlog(0, "uid change: %d\n", attr->ia_uid);
- if (attr->ia_valid & ATTR_GID)
- mlog(0, "gid change: %d\n", attr->ia_gid);
- if (attr->ia_valid & ATTR_SIZE)
- mlog(0, "size change...\n");
- if (attr->ia_valid & (ATTR_ATIME | ATTR_MTIME | ATTR_CTIME))
- mlog(0, "time change...\n");
-
#define OCFS2_VALID_ATTRS (ATTR_ATIME | ATTR_MTIME | ATTR_CTIME | ATTR_SIZE \
| ATTR_GID | ATTR_UID | ATTR_MODE)
- if (!(attr->ia_valid & OCFS2_VALID_ATTRS)) {
- mlog(0, "can't handle attrs: 0x%x\n", attr->ia_valid);
+ if (!(attr->ia_valid & OCFS2_VALID_ATTRS))
return 0;
- }
status = inode_change_ok(inode, attr);
if (status)
@@ -1315,8 +1305,9 @@ static int __ocfs2_write_remove_suid(struct inode *inode,
struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
struct ocfs2_dinode *di;
- mlog(0, "(Inode %llu, mode 0%o)\n",
- (unsigned long long)OCFS2_I(inode)->ip_blkno, inode->i_mode);
+ trace_ocfs2_write_remove_suid(
+ (unsigned long long)OCFS2_I(inode)->ip_blkno,
+ inode->i_mode);
handle = ocfs2_start_trans(osb, OCFS2_INODE_UPDATE_CREDITS);
if (IS_ERR(handle)) {
@@ -1522,8 +1513,9 @@ static int ocfs2_zero_partial_clusters(struct inode *inode,
* partial clusters here. There's no need to worry about
* physical allocation - the zeroing code knows to skip holes.
*/
- mlog(0, "byte start: %llu, end: %llu\n",
- (unsigned long long)start, (unsigned long long)end);
+ trace_ocfs2_zero_partial_clusters(
+ (unsigned long long)OCFS2_I(inode)->ip_blkno,
+ (unsigned long long)start, (unsigned long long)end);
/*
* If both edges are on a cluster boundary then there's no
@@ -1547,8 +1539,8 @@ static int ocfs2_zero_partial_clusters(struct inode *inode,
if (tmpend > end)
tmpend = end;
- mlog(0, "1st range: start: %llu, tmpend: %llu\n",
- (unsigned long long)start, (unsigned long long)tmpend);
+ trace_ocfs2_zero_partial_clusters_range1((unsigned long long)start,
+ (unsigned long long)tmpend);
ret = ocfs2_zero_range_for_truncate(inode, handle, start, tmpend);
if (ret)
@@ -1562,8 +1554,8 @@ static int ocfs2_zero_partial_clusters(struct inode *inode,
*/
start = end & ~(osb->s_clustersize - 1);
- mlog(0, "2nd range: start: %llu, end: %llu\n",
- (unsigned long long)start, (unsigned long long)end);
+ trace_ocfs2_zero_partial_clusters_range2(
+ (unsigned long long)start, (unsigned long long)end);
ret = ocfs2_zero_range_for_truncate(inode, handle, start, end);
if (ret)
@@ -1663,6 +1655,11 @@ static int ocfs2_remove_inode_range(struct inode *inode,
ocfs2_init_dinode_extent_tree(&et, INODE_CACHE(inode), di_bh);
ocfs2_init_dealloc_ctxt(&dealloc);
+ trace_ocfs2_remove_inode_range(
+ (unsigned long long)OCFS2_I(inode)->ip_blkno,
+ (unsigned long long)byte_start,
+ (unsigned long long)byte_len);
+
if (byte_len == 0)
return 0;
@@ -1709,11 +1706,6 @@ static int ocfs2_remove_inode_range(struct inode *inode,
trunc_end = (byte_start + byte_len) >> osb->s_clustersize_bits;
cluster_in_el = trunc_end;
- mlog(0, "Inode: %llu, start: %llu, len: %llu, cstart: %u, cend: %u\n",
- (unsigned long long)OCFS2_I(inode)->ip_blkno,
- (unsigned long long)byte_start,
- (unsigned long long)byte_len, trunc_start, trunc_end);
-
ret = ocfs2_zero_partial_clusters(inode, byte_start, byte_len);
if (ret) {
mlog_errno(ret);
@@ -2064,7 +2056,7 @@ static int ocfs2_prepare_inode_for_write(struct file *file,
int ret = 0, meta_level = 0;
struct dentry *dentry = file->f_path.dentry;
struct inode *inode = dentry->d_inode;
- loff_t saved_pos, end;
+ loff_t saved_pos = 0, end;
/*
* We start with a read level meta lock and only jump to an ex
@@ -2103,12 +2095,10 @@ static int ocfs2_prepare_inode_for_write(struct file *file,
/* work on a copy of ppos until we're sure that we won't have
* to recalculate it due to relocking. */
- if (appending) {
+ if (appending)
saved_pos = i_size_read(inode);
- mlog(0, "O_APPEND: inode->i_size=%llu\n", saved_pos);
- } else {
+ else
saved_pos = *ppos;
- }
end = saved_pos + count;
@@ -2179,6 +2169,10 @@ static int ocfs2_prepare_inode_for_write(struct file *file,
*ppos = saved_pos;
out_unlock:
+ trace_ocfs2_prepare_inode_for_write(OCFS2_I(inode)->ip_blkno,
+ saved_pos, appending, count,
+ direct_io, has_refcount);
+
if (meta_level >= 0)
ocfs2_inode_unlock(inode, meta_level);
@@ -2204,10 +2198,11 @@ static ssize_t ocfs2_file_aio_write(struct kiocb *iocb,
int full_coherency = !(osb->s_mount_opt &
OCFS2_MOUNT_COHERENCY_BUFFERED);
- mlog(0, "(0x%p, %u, '%.*s')\n", file,
- (unsigned int)nr_segs,
- file->f_path.dentry->d_name.len,
- file->f_path.dentry->d_name.name);
+ trace_ocfs2_file_aio_write(inode, file, file->f_path.dentry,
+ (unsigned long long)OCFS2_I(inode)->ip_blkno,
+ file->f_path.dentry->d_name.len,
+ file->f_path.dentry->d_name.name,
+ (unsigned int)nr_segs);
if (iocb->ki_left == 0)
return 0;
@@ -2402,10 +2397,11 @@ static ssize_t ocfs2_file_splice_write(struct pipe_inode_info *pipe,
.u.file = out,
};
- mlog(0, "(0x%p, 0x%p, %u, '%.*s')\n", out, pipe,
- (unsigned int)len,
- out->f_path.dentry->d_name.len,
- out->f_path.dentry->d_name.name);
+
+ trace_ocfs2_file_splice_write(inode, out, out->f_path.dentry,
+ (unsigned long long)OCFS2_I(inode)->ip_blkno,
+ out->f_path.dentry->d_name.len,
+ out->f_path.dentry->d_name.name, len);
if (pipe->inode)
mutex_lock_nested(&pipe->inode->i_mutex, I_MUTEX_PARENT);
@@ -2461,10 +2457,10 @@ static ssize_t ocfs2_file_splice_read(struct file *in,
int ret = 0, lock_level = 0;
struct inode *inode = in->f_path.dentry->d_inode;
- mlog(0, "(0x%p, 0x%p, %u, '%.*s')\n", in, pipe,
- (unsigned int)len,
- in->f_path.dentry->d_name.len,
- in->f_path.dentry->d_name.name);
+ trace_ocfs2_file_splice_read(inode, in, in->f_path.dentry,
+ (unsigned long long)OCFS2_I(inode)->ip_blkno,
+ in->f_path.dentry->d_name.len,
+ in->f_path.dentry->d_name.name, len);
/*
* See the comment in ocfs2_file_aio_read()
@@ -2491,10 +2487,11 @@ static ssize_t ocfs2_file_aio_read(struct kiocb *iocb,
struct file *filp = iocb->ki_filp;
struct inode *inode = filp->f_path.dentry->d_inode;
- mlog(0, "(0x%p, %u, '%.*s')\n", filp,
- (unsigned int)nr_segs,
- filp->f_path.dentry->d_name.len,
- filp->f_path.dentry->d_name.name);
+ trace_ocfs2_file_aio_read(inode, filp, filp->f_path.dentry,
+ (unsigned long long)OCFS2_I(inode)->ip_blkno,
+ filp->f_path.dentry->d_name.len,
+ filp->f_path.dentry->d_name.name, nr_segs);
+
if (!inode) {
ret = -EINVAL;
@@ -2537,8 +2534,7 @@ static ssize_t ocfs2_file_aio_read(struct kiocb *iocb,
ocfs2_inode_unlock(inode, lock_level);
ret = generic_file_aio_read(iocb, iov, nr_segs, iocb->ki_pos);
- if (ret == -EINVAL)
- mlog(0, "generic_file_aio_read returned -EINVAL\n");
+ trace_generic_file_aio_read_ret(ret);
/* buffered aio wouldn't have proper lock coverage today */
BUG_ON(ret == -EIOCBQUEUED && !(filp->f_flags & O_DIRECT));
diff --git a/fs/ocfs2/ocfs2_trace.h b/fs/ocfs2/ocfs2_trace.h
index efda83e..79df9e0 100644
--- a/fs/ocfs2/ocfs2_trace.h
+++ b/fs/ocfs2/ocfs2_trace.h
@@ -201,6 +201,29 @@ DECLARE_EVENT_CLASS(ocfs2__uint_uint_uint,
TP_printk("%u %u %u", __entry->value1, __entry->value2, __entry->value3)
);
+DECLARE_EVENT_CLASS(ocfs2__ull_ull_ull,
+
+ TP_PROTO(unsigned long long value1,
+ unsigned long long value2, unsigned long long value3),
+
+ TP_ARGS(value1, value2, value3),
+
+ TP_STRUCT__entry(
+ __field( unsigned long long, value1 )
+ __field( unsigned long long, value2 )
+ __field( unsigned long long, value3 )
+ ),
+
+ TP_fast_assign(
+ __entry->value1 = value1;
+ __entry->value2 = value2;
+ __entry->value3 = value3;
+ ),
+
+ TP_printk("%llu %llu %llu",
+ __entry->value1, __entry->value2, __entry->value3)
+);
+
DECLARE_EVENT_CLASS(ocfs2__ull_int_int_int,
TP_PROTO(unsigned long long ull, int value1, int value2, int value3),
@@ -1658,6 +1681,350 @@ TRACE_EVENT(ocfs2_fault,
);
/* End of trace events for fs/ocfs2/mmap.c. */
+/* Trace events for fs/ocfs2/file.c. */
+DECLARE_EVENT_CLASS(ocfs2__file_ops,
+
+ TP_PROTO(void *inode, void *file, void *dentry,
+ unsigned long long ino,
+ unsigned int d_len, const unsigned char *d_name,
+ unsigned long long para),
+
+ TP_ARGS(inode, file, dentry, ino, d_len, d_name, para),
+
+ TP_STRUCT__entry(
+ __field( void *, inode )
+ __field( void *, file )
+ __field( void *, dentry )
+ __field( unsigned long long, ino )
+ __field( unsigned int, d_len )
+ __field( const unsigned char *, d_name )
+ __field( unsigned long long, para )
+ ),
+
+ TP_fast_assign(
+ __entry->inode = inode;
+ __entry->file = file;
+ __entry->dentry = dentry;
+ __entry->ino = ino;
+ __entry->d_len = d_len;
+ __entry->d_name = d_name;
+ __entry->para = para;
+ ),
+
+ TP_printk("%p %p %p %llu %llu %.*s", __entry->inode, __entry->file,
+ __entry->dentry, __entry->ino, __entry->para,
+ __entry->d_len, __entry->d_name)
+);
+
+DEFINE_EVENT(ocfs2__file_ops, ocfs2_file_open,
+
+ TP_PROTO(void *inode, void *file, void *dentry,
+ unsigned long long ino,
+ unsigned int d_len, const unsigned char *d_name,
+ unsigned long long mode),
+
+ TP_ARGS(inode, file, dentry, ino, d_len, d_name, mode)
+);
+
+DEFINE_EVENT(ocfs2__file_ops, ocfs2_file_release,
+
+ TP_PROTO(void *inode, void *file, void *dentry,
+ unsigned long long ino,
+ unsigned int d_len, const unsigned char *d_name,
+ unsigned long long open_count),
+
+ TP_ARGS(inode, file, dentry, ino, d_len, d_name, open_count)
+);
+
+DEFINE_EVENT(ocfs2__file_ops, ocfs2_sync_file,
+
+ TP_PROTO(void *inode, void *file, void *dentry,
+ unsigned long long ino,
+ unsigned int d_len, const unsigned char *d_name,
+ unsigned long long datasync),
+
+ TP_ARGS(inode, file, dentry, ino, d_len, d_name, datasync)
+);
+
+DEFINE_EVENT(ocfs2__ull_ull_ull, ocfs2_truncate_file,
+
+ TP_PROTO(unsigned long long ip_blkno, unsigned long long size,
+ unsigned long long new_size),
+
+ TP_ARGS(ip_blkno, size, new_size)
+);
+
+DEFINE_EVENT(ocfs2__ull_ull, ocfs2_truncate_file_error,
+
+ TP_PROTO(unsigned long long i_size, unsigned long long new_size),
+
+ TP_ARGS(i_size, new_size)
+);
+
+TRACE_EVENT(ocfs2_extend_allocation,
+
+ TP_PROTO(unsigned long long ip_blkno, unsigned long long size,
+ unsigned int clusters, unsigned int clusters_to_add,
+ int why, int restart_func),
+
+ TP_ARGS(ip_blkno, size, clusters, clusters_to_add, why, restart_func),
+
+ TP_STRUCT__entry(
+ __field( unsigned long long, ip_blkno )
+ __field( unsigned long long, size )
+ __field( unsigned int, clusters )
+ __field( unsigned int, clusters_to_add )
+ __field( int, why )
+ __field( int, restart_func )
+ ),
+
+ TP_fast_assign(
+ __entry->ip_blkno = ip_blkno;
+ __entry->size = size;
+ __entry->clusters = clusters;
+ __entry->clusters_to_add = clusters_to_add;
+ __entry->why = why;
+ __entry->restart_func = restart_func;
+ ),
+
+ TP_printk("%llu %llu %u %u %d %d",
+ __entry->ip_blkno, __entry->size, __entry->clusters,
+ __entry->clusters_to_add, __entry->why, __entry->restart_func)
+);
+
+TRACE_EVENT(ocfs2_extend_allocation_end,
+
+ TP_PROTO(unsigned long long ino,
+ unsigned int di_clusters, unsigned long long di_size,
+ unsigned int ip_clusters, unsigned long long i_size),
+
+ TP_ARGS(ino, di_clusters, di_size, ip_clusters, i_size),
+
+ TP_STRUCT__entry(
+ __field( unsigned long long, ino )
+ __field( unsigned int, di_clusters )
+ __field( unsigned long long, di_size )
+ __field( unsigned int, ip_clusters )
+ __field( unsigned long long, i_size )
+ ),
+
+ TP_fast_assign(
+ __entry->ino = ino;
+ __entry->di_clusters = di_clusters;
+ __entry->di_size = di_size;
+ __entry->ip_clusters = ip_clusters;
+ __entry->i_size = i_size;
+ ),
+
+ TP_printk("%llu %u %llu %u %llu", __entry->ino, __entry->di_clusters,
+ __entry->di_size, __entry->ip_clusters, __entry->i_size)
+);
+
+TRACE_EVENT(ocfs2_write_zero_page,
+
+ TP_PROTO(unsigned long long ino,
+ unsigned long long abs_from, unsigned long long abs_to,
+ unsigned long index, unsigned int zero_from,
+ unsigned int zero_to),
+
+ TP_ARGS(ino, abs_from, abs_to, index, zero_from, zero_to),
+
+ TP_STRUCT__entry(
+ __field( unsigned long long, ino )
+ __field( unsigned long long, abs_from )
+ __field( unsigned long long, abs_to )
+ __field( unsigned long, index )
+ __field( unsigned int, zero_from )
+ __field( unsigned int, zero_to )
+ ),
+
+ TP_fast_assign(
+ __entry->ino = ino;
+ __entry->abs_from = abs_from;
+ __entry->abs_to = abs_to;
+ __entry->index = index;
+ __entry->zero_from = zero_from;
+ __entry->zero_to = zero_to;
+ ),
+
+ TP_printk("%llu %llu %llu %lu %u %u", __entry->ino,
+ __entry->abs_from, __entry->abs_to,
+ __entry->index, __entry->zero_from, __entry->zero_to)
+);
+
+DEFINE_EVENT(ocfs2__ull_ull_ull, ocfs2_zero_extend_range,
+
+ TP_PROTO(unsigned long long ip_blkno, unsigned long long range_start,
+ unsigned long long range_end),
+
+ TP_ARGS(ip_blkno, range_start, range_end)
+);
+
+DEFINE_EVENT(ocfs2__ull_ull_ull, ocfs2_zero_extend,
+
+ TP_PROTO(unsigned long long ip_blkno, unsigned long long zero_start,
+ unsigned long long i_size),
+
+ TP_ARGS(ip_blkno, zero_start, i_size)
+);
+
+TRACE_EVENT(ocfs2_setattr,
+
+ TP_PROTO(void *inode, void *dentry,
+ unsigned long long ino,
+ unsigned int d_len, const unsigned char *d_name,
+ unsigned int ia_valid, unsigned int ia_mode,
+ unsigned int ia_uid, unsigned int ia_gid),
+
+ TP_ARGS(inode, dentry, ino, d_len, d_name,
+ ia_valid, ia_mode, ia_uid, ia_gid),
+
+ TP_STRUCT__entry(
+ __field( void *, inode )
+ __field( void *, dentry )
+ __field( unsigned long long, ino )
+ __field( unsigned int, d_len )
+ __field( const unsigned char *, d_name )
+ __field( unsigned int, ia_valid )
+ __field( unsigned int, ia_mode )
+ __field( unsigned int, ia_uid )
+ __field( unsigned int, ia_gid )
+ ),
+
+ TP_fast_assign(
+ __entry->inode = inode;
+ __entry->dentry = dentry;
+ __entry->ino = ino;
+ __entry->d_len = d_len;
+ __entry->d_name = d_name;
+ __entry->ia_valid = ia_valid;
+ __entry->ia_mode = ia_mode;
+ __entry->ia_uid = ia_uid;
+ __entry->ia_gid = ia_gid;
+ ),
+
+ TP_printk("%p %p %llu %.*s %u %u %u %u", __entry->inode,
+ __entry->dentry, __entry->ino, __entry->d_len,
+ __entry->d_name, __entry->ia_valid, __entry->ia_mode,
+ __entry->ia_uid, __entry->ia_gid)
+);
+
+DEFINE_EVENT(ocfs2__ull_uint, ocfs2_write_remove_suid,
+
+ TP_PROTO(unsigned long long ip_blkno, unsigned int mode),
+
+ TP_ARGS(ip_blkno, mode)
+);
+
+DEFINE_EVENT(ocfs2__ull_ull_ull, ocfs2_zero_partial_clusters,
+
+ TP_PROTO(unsigned long long ip_blkno, unsigned long long start,
+ unsigned long long end),
+
+ TP_ARGS(ip_blkno, start, end)
+);
+
+DEFINE_EVENT(ocfs2__ull_ull, ocfs2_zero_partial_clusters_range1,
+
+ TP_PROTO(unsigned long long start, unsigned long long end),
+
+ TP_ARGS(start, end)
+);
+
+DEFINE_EVENT(ocfs2__ull_ull, ocfs2_zero_partial_clusters_range2,
+
+ TP_PROTO(unsigned long long start, unsigned long long end),
+
+ TP_ARGS(start, end)
+);
+
+DEFINE_EVENT(ocfs2__ull_ull_ull, ocfs2_remove_inode_range,
+
+ TP_PROTO(unsigned long long ip_blkno, unsigned long long start,
+ unsigned long long end),
+
+ TP_ARGS(ip_blkno, start, end)
+);
+
+TRACE_EVENT(ocfs2_prepare_inode_for_write,
+
+ TP_PROTO(unsigned long long ino, unsigned long long saved_pos,
+ int appending, unsigned long count,
+ int *direct_io, int *has_refcount),
+
+ TP_ARGS(ino, saved_pos, appending, count, direct_io, has_refcount),
+
+ TP_STRUCT__entry(
+ __field( unsigned long long, ino )
+ __field( unsigned long long, saved_pos )
+ __field( int, appending )
+ __field( unsigned long, count )
+ __field( int, direct_io )
+ __field( int, has_refcount )
+ ),
+
+ TP_fast_assign(
+ __entry->ino = ino;
+ __entry->saved_pos = saved_pos;
+ __entry->appending = appending;
+ __entry->count = count;
+ __entry->direct_io = direct_io ? *direct_io : -1;
+ __entry->has_refcount = has_refcount ? *has_refcount : -1;
+ ),
+
+ TP_printk("%llu %llu %d %lu %d %d", __entry->ino,
+ __entry->saved_pos, __entry->appending, __entry->count,
+ __entry->direct_io, __entry->has_refcount)
+);
+
+DEFINE_EVENT(ocfs2__file_ops, ocfs2_file_aio_write,
+
+ TP_PROTO(void *inode, void *file, void *dentry,
+ unsigned long long ino,
+ unsigned int d_len, const unsigned char *d_name,
+ unsigned long long nr_segs),
+
+ TP_ARGS(inode, file, dentry, ino, d_len, d_name, nr_segs)
+);
+
+DEFINE_EVENT(ocfs2__file_ops, ocfs2_file_splice_write,
+
+ TP_PROTO(void *inode, void *file, void *dentry,
+ unsigned long long ino,
+ unsigned int d_len, const unsigned char *d_name,
+ unsigned long long len),
+
+ TP_ARGS(inode, file, dentry, ino, d_len, d_name, len)
+);
+
+DEFINE_EVENT(ocfs2__file_ops, ocfs2_file_splice_read,
+
+ TP_PROTO(void *inode, void *file, void *dentry,
+ unsigned long long ino,
+ unsigned int d_len, const unsigned char *d_name,
+ unsigned long long len),
+
+ TP_ARGS(inode, file, dentry, ino, d_len, d_name, len)
+);
+
+DEFINE_EVENT(ocfs2__file_ops, ocfs2_file_aio_read,
+
+ TP_PROTO(void *inode, void *file, void *dentry,
+ unsigned long long ino,
+ unsigned int d_len, const unsigned char *d_name,
+ unsigned long long nr_segs),
+
+ TP_ARGS(inode, file, dentry, ino, d_len, d_name, nr_segs)
+);
+
+DEFINE_EVENT(ocfs2__int, generic_file_aio_read_ret,
+
+ TP_PROTO(int ret),
+
+ TP_ARGS(ret)
+);
+/* End of trace events for fs/ocfs2/file.c. */
+
#endif /* _TRACE_OCFS2_H */
/* This part must be outside protection */
diff --git a/fs/ocfs2/sysfile.c b/fs/ocfs2/sysfile.c
index 902efb2..3d635f4 100644
--- a/fs/ocfs2/sysfile.c
+++ b/fs/ocfs2/sysfile.c
@@ -27,7 +27,6 @@
#include <linux/types.h>
#include <linux/highmem.h>
-#define MLOG_MASK_PREFIX ML_INODE
#include <cluster/masklog.h>
#include "ocfs2.h"
--
1.7.0.4
^ permalink raw reply related [flat|nested] 53+ messages in thread
* [Ocfs2-devel] [PATCH 15/34] ocfs2: Little refactoring against ocfs2_iget.
2010-12-23 7:19 [Ocfs2-devel] [PATCH 00/34] OCFS2: Add trace event and replace mlog(0) Tao Ma
` (13 preceding siblings ...)
2010-12-23 7:30 ` [Ocfs2-devel] [PATCH 14/34] ocfs2: Remove mlog(0) from fs/ocfs2/file.c Tao Ma
@ 2010-12-23 7:30 ` Tao Ma
2010-12-23 7:30 ` [Ocfs2-devel] [PATCH 16/34] ocfs2: Remove masklog ML_INODE Tao Ma
` (23 subsequent siblings)
38 siblings, 0 replies; 53+ messages in thread
From: Tao Ma @ 2010-12-23 7:30 UTC (permalink / raw)
To: ocfs2-devel
From: Tao Ma <boyu.mt@taobao.com>
ocfs2_iget is used to get/create inode. Only iget5_locked
will give us an inode = NULL. So move this check ahead of
ocfs2_read_locked_inode so that we don't need to check
inode before we read and unlock inode. This is also helpful
for trace event(see the next patch).
Signed-off-by: Tao Ma <boyu.mt@taobao.com>
---
fs/ocfs2/inode.c | 10 +++++-----
1 files changed, 5 insertions(+), 5 deletions(-)
diff --git a/fs/ocfs2/inode.c b/fs/ocfs2/inode.c
index f5c0f34..361efd3 100644
--- a/fs/ocfs2/inode.c
+++ b/fs/ocfs2/inode.c
@@ -152,16 +152,16 @@ struct inode *ocfs2_iget(struct ocfs2_super *osb, u64 blkno, unsigned flags,
/* inode was *not* in the inode cache. 2.6.x requires
* us to do our own read_inode call and unlock it
* afterwards. */
- if (inode && inode->i_state & I_NEW) {
- mlog(0, "Inode was not in inode cache, reading it.\n");
- ocfs2_read_locked_inode(inode, &args);
- unlock_new_inode(inode);
- }
if (inode == NULL) {
inode = ERR_PTR(-ENOMEM);
mlog_errno(PTR_ERR(inode));
goto bail;
}
+ if (inode->i_state & I_NEW) {
+ mlog(0, "Inode was not in inode cache, reading it.\n");
+ ocfs2_read_locked_inode(inode, &args);
+ unlock_new_inode(inode);
+ }
if (is_bad_inode(inode)) {
iput(inode);
inode = ERR_PTR(-ESTALE);
--
1.7.0.4
^ permalink raw reply related [flat|nested] 53+ messages in thread
* [Ocfs2-devel] [PATCH 16/34] ocfs2: Remove masklog ML_INODE.
2010-12-23 7:19 [Ocfs2-devel] [PATCH 00/34] OCFS2: Add trace event and replace mlog(0) Tao Ma
` (14 preceding siblings ...)
2010-12-23 7:30 ` [Ocfs2-devel] [PATCH 15/34] ocfs2: Little refactoring against ocfs2_iget Tao Ma
@ 2010-12-23 7:30 ` Tao Ma
2010-12-23 7:31 ` [Ocfs2-devel] [PATCH 17/34] ocfs2: Remove masklog ML_EXTENT_MAP Tao Ma
` (22 subsequent siblings)
38 siblings, 0 replies; 53+ messages in thread
From: Tao Ma @ 2010-12-23 7:30 UTC (permalink / raw)
To: ocfs2-devel
From: Tao Ma <boyu.mt@taobao.com>
Remove mlog(0) from fs/ocfs2/inode.c and the masklog INODE.
Signed-off-by: Tao Ma <boyu.mt@taobao.com>
---
fs/ocfs2/cluster/masklog.c | 1 -
fs/ocfs2/cluster/masklog.h | 1 -
fs/ocfs2/inode.c | 107 +++++++++------------
fs/ocfs2/ocfs2_trace.h | 227 ++++++++++++++++++++++++++++++++++++++++++++
4 files changed, 273 insertions(+), 63 deletions(-)
diff --git a/fs/ocfs2/cluster/masklog.c b/fs/ocfs2/cluster/masklog.c
index c654038..1c1fa9d 100644
--- a/fs/ocfs2/cluster/masklog.c
+++ b/fs/ocfs2/cluster/masklog.c
@@ -99,7 +99,6 @@ static struct mlog_attribute mlog_attrs[MLOG_MAX_BITS] = {
define_mask(BH_IO),
define_mask(UPTODATE),
define_mask(NAMEI),
- define_mask(INODE),
define_mask(VOTE),
define_mask(DCACHE),
define_mask(CONN),
diff --git a/fs/ocfs2/cluster/masklog.h b/fs/ocfs2/cluster/masklog.h
index de412aa..d6dab10 100644
--- a/fs/ocfs2/cluster/masklog.h
+++ b/fs/ocfs2/cluster/masklog.h
@@ -101,7 +101,6 @@
#define ML_BH_IO 0x0000000000100000ULL /* ocfs2 buffer I/O */
#define ML_UPTODATE 0x0000000000200000ULL /* ocfs2 caching sequence #'s */
#define ML_NAMEI 0x0000000000400000ULL /* ocfs2 directory / namespace */
-#define ML_INODE 0x0000000000800000ULL /* ocfs2 inode manipulation */
#define ML_VOTE 0x0000000001000000ULL /* ocfs2 node messaging */
#define ML_DCACHE 0x0000000002000000ULL /* ocfs2 dcache operations */
#define ML_CONN 0x0000000004000000ULL /* net connection management */
diff --git a/fs/ocfs2/inode.c b/fs/ocfs2/inode.c
index 361efd3..c3b9301 100644
--- a/fs/ocfs2/inode.c
+++ b/fs/ocfs2/inode.c
@@ -31,7 +31,6 @@
#include <asm/byteorder.h>
-#define MLOG_MASK_PREFIX ML_INODE
#include <cluster/masklog.h>
#include "ocfs2.h"
@@ -53,6 +52,7 @@
#include "uptodate.h"
#include "xattr.h"
#include "refcounttree.h"
+#include "ocfs2_trace.h"
#include "buffer_head_io.h"
@@ -131,7 +131,8 @@ struct inode *ocfs2_iget(struct ocfs2_super *osb, u64 blkno, unsigned flags,
struct super_block *sb = osb->sb;
struct ocfs2_find_inode_args args;
- mlog(0, "(blkno = %llu)\n", (unsigned long long)blkno);
+ trace_ocfs2_iget_begin((unsigned long long)blkno, flags,
+ sysfile_type);
/* Ok. By now we've either got the offsets passed to us by the
* caller, or we just pulled them off the bh. Lets do some
@@ -157,8 +158,8 @@ struct inode *ocfs2_iget(struct ocfs2_super *osb, u64 blkno, unsigned flags,
mlog_errno(PTR_ERR(inode));
goto bail;
}
+ trace_ocfs2_iget5_locked(inode->i_state);
if (inode->i_state & I_NEW) {
- mlog(0, "Inode was not in inode cache, reading it.\n");
ocfs2_read_locked_inode(inode, &args);
unlock_new_inode(inode);
}
@@ -170,9 +171,8 @@ struct inode *ocfs2_iget(struct ocfs2_super *osb, u64 blkno, unsigned flags,
bail:
if (!IS_ERR(inode)) {
- mlog(0, "returning inode with number %llu\n",
- (unsigned long long)OCFS2_I(inode)->ip_blkno);
- mlog(0, "inode %p\n", inode);
+ trace_ocfs2_iget_end(inode,
+ (unsigned long long)OCFS2_I(inode)->ip_blkno);
}
return inode;
@@ -192,12 +192,12 @@ static int ocfs2_find_actor(struct inode *inode, void *opaque)
struct ocfs2_inode_info *oi = OCFS2_I(inode);
int ret = 0;
- mlog(0, "(0x%p, %lu, 0x%p)\n", inode, inode->i_ino, opaque);
-
args = opaque;
mlog_bug_on_msg(!inode, "No inode in find actor!\n");
+ trace_ocfs2_find_actor(inode, inode->i_ino, opaque, args->fi_blkno);
+
if (oi->ip_blkno != args->fi_blkno)
goto bail;
@@ -217,8 +217,6 @@ static int ocfs2_init_locked_inode(struct inode *inode, void *opaque)
static struct lock_class_key ocfs2_quota_ip_alloc_sem_key,
ocfs2_file_ip_alloc_sem_key;
- mlog(0, "inode = %p, opaque = %p\n", inode, opaque);
-
inode->i_ino = args->fi_ino;
OCFS2_I(inode)->ip_blkno = args->fi_blkno;
if (args->fi_sysfile_type != 0)
@@ -244,9 +242,6 @@ void ocfs2_populate_inode(struct inode *inode, struct ocfs2_dinode *fe,
struct ocfs2_super *osb;
int use_plocks = 1;
- mlog(0, "(0x%p, size:%llu)\n", inode,
- (unsigned long long)le64_to_cpu(fe->i_size));
-
sb = inode->i_sb;
osb = OCFS2_SB(sb);
@@ -298,20 +293,20 @@ void ocfs2_populate_inode(struct inode *inode, struct ocfs2_dinode *fe,
inode->i_nlink = ocfs2_read_links_count(fe);
+ trace_ocfs2_populate_inode(OCFS2_I(inode)->ip_blkno,
+ le32_to_cpu(fe->i_flags));
if (fe->i_flags & cpu_to_le32(OCFS2_SYSTEM_FL)) {
OCFS2_I(inode)->ip_flags |= OCFS2_INODE_SYSTEM_FILE;
inode->i_flags |= S_NOQUOTA;
}
-
+
if (fe->i_flags & cpu_to_le32(OCFS2_LOCAL_ALLOC_FL)) {
OCFS2_I(inode)->ip_flags |= OCFS2_INODE_BITMAP;
- mlog(0, "local alloc inode: i_ino=%lu\n", inode->i_ino);
} else if (fe->i_flags & cpu_to_le32(OCFS2_BITMAP_FL)) {
OCFS2_I(inode)->ip_flags |= OCFS2_INODE_BITMAP;
} else if (fe->i_flags & cpu_to_le32(OCFS2_QUOTA_FL)) {
inode->i_flags |= S_NOQUOTA;
} else if (fe->i_flags & cpu_to_le32(OCFS2_SUPER_BLOCK_FL)) {
- mlog(0, "superblock inode: i_ino=%lu\n", inode->i_ino);
/* we can't actually hit this as read_inode can't
* handle superblocks today ;-) */
BUG();
@@ -391,8 +386,6 @@ static int ocfs2_read_locked_inode(struct inode *inode,
int status, can_lock;
u32 generation = 0;
- mlog(0, "(0x%p, 0x%p)\n", inode, args);
-
status = -EINVAL;
if (inode == NULL || inode->i_sb == NULL) {
mlog(ML_ERROR, "bad inode\n");
@@ -440,6 +433,9 @@ static int ocfs2_read_locked_inode(struct inode *inode,
&& !(args->fi_flags & OCFS2_FI_FLAG_ORPHAN_RECOVERY)
&& !ocfs2_mount_local(osb);
+ trace_ocfs2_read_locked_inode(
+ (unsigned long long)OCFS2_I(inode)->ip_blkno, can_lock);
+
/*
* To maintain backwards compatibility with older versions of
* ocfs2-tools, we still store the generation value for system
@@ -689,8 +685,6 @@ static int ocfs2_check_orphan_recovery_state(struct ocfs2_super *osb,
spin_lock(&osb->osb_lock);
if (ocfs2_node_map_test_bit(osb, &osb->osb_recovering_orphan_dirs, slot)) {
- mlog(0, "Recovery is happening on orphan dir %d, will skip "
- "this inode\n", slot);
ret = -EDEADLK;
goto out;
}
@@ -699,6 +693,7 @@ static int ocfs2_check_orphan_recovery_state(struct ocfs2_super *osb,
osb->osb_orphan_wipes[slot]++;
out:
spin_unlock(&osb->osb_lock);
+ trace_ocfs2_check_orphan_recovery_state(slot, ret);
return ret;
}
@@ -809,6 +804,10 @@ static int ocfs2_inode_is_valid_to_delete(struct inode *inode)
struct ocfs2_inode_info *oi = OCFS2_I(inode);
struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
+ trace_ocfs2_inode_is_valid_to_delete(current, osb->dc_task,
+ (unsigned long long)oi->ip_blkno,
+ oi->ip_flags);
+
/* We shouldn't be getting here for the root directory
* inode.. */
if (inode == osb->root_inode) {
@@ -821,11 +820,8 @@ static int ocfs2_inode_is_valid_to_delete(struct inode *inode)
* have to skip deleting this guy. That's OK though because
* the node who's doing the actual deleting should handle it
* anyway. */
- if (current == osb->dc_task) {
- mlog(0, "Skipping delete of %lu because we're currently "
- "in downconvert\n", inode->i_ino);
+ if (current == osb->dc_task)
goto bail;
- }
spin_lock(&oi->ip_lock);
/* OCFS2 *never* deletes system files. This should technically
@@ -840,11 +836,8 @@ static int ocfs2_inode_is_valid_to_delete(struct inode *inode)
/* If we have allowd wipe of this inode for another node, it
* will be marked here so we can safely skip it. Recovery will
* cleanup any inodes we might inadvertantly skip here. */
- if (oi->ip_flags & OCFS2_INODE_SKIP_DELETE) {
- mlog(0, "Skipping delete of %lu because another node "
- "has done this for us.\n", inode->i_ino);
+ if (oi->ip_flags & OCFS2_INODE_SKIP_DELETE)
goto bail_unlock;
- }
ret = 1;
bail_unlock:
@@ -861,28 +854,27 @@ static int ocfs2_query_inode_wipe(struct inode *inode,
struct buffer_head *di_bh,
int *wipe)
{
- int status = 0;
+ int status = 0, reason = 0;
struct ocfs2_inode_info *oi = OCFS2_I(inode);
struct ocfs2_dinode *di;
*wipe = 0;
+ trace_ocfs2_query_inode_wipe_begin((unsigned long long)oi->ip_blkno,
+ inode->i_nlink);
+
/* While we were waiting for the cluster lock in
* ocfs2_delete_inode, another node might have asked to delete
* the inode. Recheck our flags to catch this. */
if (!ocfs2_inode_is_valid_to_delete(inode)) {
- mlog(0, "Skipping delete of %llu because flags changed\n",
- (unsigned long long)oi->ip_blkno);
+ reason = 1;
goto bail;
}
/* Now that we have an up to date inode, we can double check
* the link count. */
- if (inode->i_nlink) {
- mlog(0, "Skipping delete of %llu because nlink = %u\n",
- (unsigned long long)oi->ip_blkno, inode->i_nlink);
+ if (inode->i_nlink)
goto bail;
- }
/* Do some basic inode verification... */
di = (struct ocfs2_dinode *) di_bh->b_data;
@@ -897,9 +889,7 @@ static int ocfs2_query_inode_wipe(struct inode *inode,
* ORPHANED_FL not.
*/
if (di->i_dyn_features & cpu_to_le16(OCFS2_HAS_REFCOUNT_FL)) {
- mlog(0, "Reflinked inode %llu is no longer orphaned. "
- "it shouldn't be deleted\n",
- (unsigned long long)oi->ip_blkno);
+ reason = 2;
goto bail;
}
@@ -936,8 +926,7 @@ static int ocfs2_query_inode_wipe(struct inode *inode,
status = ocfs2_try_open_lock(inode, 1);
if (status == -EAGAIN) {
status = 0;
- mlog(0, "Skipping delete of %llu because it is in use on "
- "other nodes\n", (unsigned long long)oi->ip_blkno);
+ reason = 3;
goto bail;
}
if (status < 0) {
@@ -946,11 +935,10 @@ static int ocfs2_query_inode_wipe(struct inode *inode,
}
*wipe = 1;
- mlog(0, "Inode %llu is ok to wipe from orphan dir %u\n",
- (unsigned long long)oi->ip_blkno,
- le16_to_cpu(di->i_orphaned_slot));
+ trace_ocfs2_query_inode_wipe_succ(le16_to_cpu(di->i_orphaned_slot));
bail:
+ trace_ocfs2_query_inode_wipe_end(status, reason);
return status;
}
@@ -960,8 +948,8 @@ bail:
static void ocfs2_cleanup_delete_inode(struct inode *inode,
int sync_data)
{
- mlog(0, "Cleanup inode %llu, sync = %d\n",
- (unsigned long long)OCFS2_I(inode)->ip_blkno, sync_data);
+ trace_ocfs2_cleanup_delete_inode(
+ (unsigned long long)OCFS2_I(inode)->ip_blkno, sync_data);
if (sync_data)
write_inode_now(inode, 1);
truncate_inode_pages(&inode->i_data, 0);
@@ -973,15 +961,15 @@ static void ocfs2_delete_inode(struct inode *inode)
sigset_t oldset;
struct buffer_head *di_bh = NULL;
- mlog(0, "(inode->i_ino = %lu)\n", inode->i_ino);
+ trace_ocfs2_delete_inode(inode->i_ino,
+ (unsigned long long)OCFS2_I(inode)->ip_blkno,
+ is_bad_inode(inode));
/* When we fail in read_inode() we mark inode as bad. The second test
* catches the case when inode allocation fails before allocating
* a block for inode. */
- if (is_bad_inode(inode) || !OCFS2_I(inode)->ip_blkno) {
- mlog(0, "Skipping delete of bad inode\n");
+ if (is_bad_inode(inode) || !OCFS2_I(inode)->ip_blkno)
goto bail;
- }
dquot_initialize(inode);
@@ -1082,8 +1070,8 @@ static void ocfs2_clear_inode(struct inode *inode)
struct ocfs2_inode_info *oi = OCFS2_I(inode);
end_writeback(inode);
- mlog(0, "Clearing inode: %llu, nlink = %u\n",
- (unsigned long long)OCFS2_I(inode)->ip_blkno, inode->i_nlink);
+ trace_ocfs2_clear_inode((unsigned long long)oi->ip_blkno,
+ inode->i_nlink);
mlog_bug_on_msg(OCFS2_SB(inode->i_sb) == NULL,
"Inode=%lu\n", inode->i_ino);
@@ -1193,8 +1181,8 @@ int ocfs2_drop_inode(struct inode *inode)
struct ocfs2_inode_info *oi = OCFS2_I(inode);
int res;
- mlog(0, "Drop inode %llu, nlink = %u, ip_flags = 0x%x\n",
- (unsigned long long)oi->ip_blkno, inode->i_nlink, oi->ip_flags);
+ trace_ocfs2_drop_inode((unsigned long long)oi->ip_blkno,
+ inode->i_nlink, oi->ip_flags);
if (oi->ip_flags & OCFS2_INODE_MAYBE_ORPHANED)
res = 1;
@@ -1212,11 +1200,11 @@ int ocfs2_inode_revalidate(struct dentry *dentry)
struct inode *inode = dentry->d_inode;
int status = 0;
- mlog(0, "(inode = 0x%p, ino = %llu)\n", inode,
- inode ? (unsigned long long)OCFS2_I(inode)->ip_blkno : 0ULL);
+ trace_ocfs2_inode_revalidate(inode,
+ inode ? (unsigned long long)OCFS2_I(inode)->ip_blkno : 0ULL,
+ inode ? (unsigned long long)OCFS2_I(inode)->ip_flags : 0);
if (!inode) {
- mlog(0, "eep, no inode!\n");
status = -ENOENT;
goto bail;
}
@@ -1224,7 +1212,6 @@ int ocfs2_inode_revalidate(struct dentry *dentry)
spin_lock(&OCFS2_I(inode)->ip_lock);
if (OCFS2_I(inode)->ip_flags & OCFS2_INODE_DELETED) {
spin_unlock(&OCFS2_I(inode)->ip_lock);
- mlog(0, "inode deleted!\n");
status = -ENOENT;
goto bail;
}
@@ -1255,8 +1242,7 @@ int ocfs2_mark_inode_dirty(handle_t *handle,
int status;
struct ocfs2_dinode *fe = (struct ocfs2_dinode *) bh->b_data;
- mlog(0, "(inode %llu)\n",
- (unsigned long long)OCFS2_I(inode)->ip_blkno);
+ trace_ocfs2_mark_inode_dirty((unsigned long long)OCFS2_I(inode)->ip_blkno);
status = ocfs2_journal_access_di(handle, INODE_CACHE(inode), bh,
OCFS2_JOURNAL_ACCESS_WRITE);
@@ -1328,8 +1314,7 @@ int ocfs2_validate_inode_block(struct super_block *sb,
int rc;
struct ocfs2_dinode *di = (struct ocfs2_dinode *)bh->b_data;
- mlog(0, "Validating dinode %llu\n",
- (unsigned long long)bh->b_blocknr);
+ trace_ocfs2_validate_inode_block((unsigned long long)bh->b_blocknr);
BUG_ON(!buffer_uptodate(bh));
diff --git a/fs/ocfs2/ocfs2_trace.h b/fs/ocfs2/ocfs2_trace.h
index 79df9e0..d9b13c1 100644
--- a/fs/ocfs2/ocfs2_trace.h
+++ b/fs/ocfs2/ocfs2_trace.h
@@ -114,6 +114,25 @@ DECLARE_EVENT_CLASS(ocfs2__ull_uint,
TP_printk("%llu %u", __entry->value1, __entry->value2)
);
+DECLARE_EVENT_CLASS(ocfs2__ull_int,
+
+ TP_PROTO(unsigned long long value1, int value2),
+
+ TP_ARGS(value1, value2),
+
+ TP_STRUCT__entry(
+ __field( unsigned long long, value1 )
+ __field( int, value2 )
+ ),
+
+ TP_fast_assign(
+ __entry->value1 = value1;
+ __entry->value2 = value2;
+ ),
+
+ TP_printk("%llu %d", __entry->value1, __entry->value2)
+);
+
DECLARE_EVENT_CLASS(ocfs2__ull_ull,
TP_PROTO(unsigned long long value1, unsigned long long value2),
@@ -2025,6 +2044,214 @@ DEFINE_EVENT(ocfs2__int, generic_file_aio_read_ret,
);
/* End of trace events for fs/ocfs2/file.c. */
+/* Trace events for fs/ocfs2/inode.c. */
+TRACE_EVENT(ocfs2_iget_begin,
+
+ TP_PROTO(unsigned long long ino, unsigned int flags, int sysfile_type),
+
+ TP_ARGS(ino, flags, sysfile_type),
+
+ TP_STRUCT__entry(
+ __field( unsigned long long, ino )
+ __field( unsigned int, flags )
+ __field( int, sysfile_type )
+ ),
+
+ TP_fast_assign(
+ __entry->ino = ino;
+ __entry->flags = flags;
+ __entry->sysfile_type = sysfile_type;
+ ),
+
+ TP_printk("%llu %u %d", __entry->ino,
+ __entry->flags, __entry->sysfile_type)
+);
+
+DEFINE_EVENT(ocfs2__ull, ocfs2_iget5_locked,
+
+ TP_PROTO(unsigned long long state),
+
+ TP_ARGS(state)
+);
+
+TRACE_EVENT(ocfs2_iget_end,
+
+ TP_PROTO(void *inode, unsigned long long ino),
+
+ TP_ARGS(inode, ino),
+
+ TP_STRUCT__entry(
+ __field( void *, inode )
+ __field( unsigned long long, ino )
+ ),
+
+ TP_fast_assign(
+ __entry->inode = inode;
+ __entry->ino = ino;
+ ),
+
+ TP_printk("%p %llu", __entry->inode, __entry->ino)
+);
+
+TRACE_EVENT(ocfs2_find_actor,
+
+ TP_PROTO(void *inode, unsigned long long ino,
+ void *args, unsigned long long fi_blkno),
+
+ TP_ARGS(inode, ino, args, fi_blkno),
+
+ TP_STRUCT__entry(
+ __field( void *, inode )
+ __field( unsigned long long, ino )
+ __field( void *, args )
+ __field( unsigned long long, fi_blkno )
+ ),
+
+ TP_fast_assign(
+ __entry->inode = inode;
+ __entry->ino = ino;
+ __entry->args = args;
+ __entry->fi_blkno = fi_blkno;
+ ),
+
+ TP_printk("%p %llu %p %llu", __entry->inode, __entry->ino,
+ __entry->args, __entry->fi_blkno)
+);
+
+DEFINE_EVENT(ocfs2__ull_uint, ocfs2_populate_inode,
+
+ TP_PROTO(unsigned long long ino, unsigned int flags),
+
+ TP_ARGS(ino, flags)
+);
+
+DEFINE_EVENT(ocfs2__ull_int, ocfs2_read_locked_inode,
+
+ TP_PROTO(unsigned long long ino, int can_lock),
+
+ TP_ARGS(ino, can_lock)
+);
+
+DEFINE_EVENT(ocfs2__int_int, ocfs2_check_orphan_recovery_state,
+
+ TP_PROTO(int slot, int ret),
+
+ TP_ARGS(slot, ret)
+);
+
+DEFINE_EVENT(ocfs2__ull, ocfs2_validate_inode_block,
+
+ TP_PROTO(unsigned long long ino),
+
+ TP_ARGS(ino)
+);
+
+TRACE_EVENT(ocfs2_inode_is_valid_to_delete,
+
+ TP_PROTO(void *task, void *dc_task, unsigned long long ino,
+ unsigned int flags),
+
+ TP_ARGS(task, dc_task, ino, flags),
+
+ TP_STRUCT__entry(
+ __field( void *, task )
+ __field( void *, dc_task )
+ __field( unsigned long long, ino )
+ __field( unsigned int, flags )
+ ),
+
+ TP_fast_assign(
+ __entry->task = task;
+ __entry->dc_task = dc_task;
+ __entry->ino = ino;
+ __entry->flags = flags;
+ ),
+
+ TP_printk("%p %p %llu %u", __entry->task, __entry->dc_task,
+ __entry->ino, __entry->flags)
+);
+
+DEFINE_EVENT(ocfs2__ull_uint, ocfs2_query_inode_wipe_begin,
+
+ TP_PROTO(unsigned long long ino, unsigned int nlink),
+
+ TP_ARGS(ino, nlink)
+);
+
+DEFINE_EVENT(ocfs2__uint, ocfs2_query_inode_wipe_succ,
+
+ TP_PROTO(unsigned int slot),
+
+ TP_ARGS(slot)
+);
+
+DEFINE_EVENT(ocfs2__int_int, ocfs2_query_inode_wipe_end,
+
+ TP_PROTO(int status, int reason),
+
+ TP_ARGS(status, reason)
+);
+
+DEFINE_EVENT(ocfs2__ull_int, ocfs2_cleanup_delete_inode,
+
+ TP_PROTO(unsigned long long ino, int sync_data),
+
+ TP_ARGS(ino, sync_data)
+);
+
+DEFINE_EVENT(ocfs2__ull_ull_uint, ocfs2_delete_inode,
+
+ TP_PROTO(unsigned long long ino,
+ unsigned long long ip_blkno, unsigned int is_bad_inode),
+
+ TP_ARGS(ino, ip_blkno, is_bad_inode)
+);
+
+DEFINE_EVENT(ocfs2__ull_uint, ocfs2_clear_inode,
+
+ TP_PROTO(unsigned long long ino, unsigned int nlink),
+
+ TP_ARGS(ino, nlink)
+);
+
+DEFINE_EVENT(ocfs2__ull_uint_uint, ocfs2_drop_inode,
+
+ TP_PROTO(unsigned long long ino,
+ unsigned int n_link, unsigned int ip_flags),
+
+ TP_ARGS(ino, n_link, ip_flags)
+);
+
+TRACE_EVENT(ocfs2_inode_revalidate,
+
+ TP_PROTO(void *inode, unsigned long long ino,
+ unsigned int flags),
+
+ TP_ARGS(inode, ino, flags),
+
+ TP_STRUCT__entry(
+ __field( void *, inode )
+ __field( unsigned long long, ino )
+ __field( unsigned int, flags )
+ ),
+
+ TP_fast_assign(
+ __entry->inode = inode;
+ __entry->ino = ino;
+ __entry->flags = flags;
+ ),
+
+ TP_printk("%p %llu %u", __entry->inode, __entry->ino, __entry->flags)
+);
+
+DEFINE_EVENT(ocfs2__ull, ocfs2_mark_inode_dirty,
+
+ TP_PROTO(unsigned long long ino),
+
+ TP_ARGS(ino)
+);
+
+/* End of trace events for fs/ocfs2/inode.c. */
#endif /* _TRACE_OCFS2_H */
/* This part must be outside protection */
--
1.7.0.4
^ permalink raw reply related [flat|nested] 53+ messages in thread
* [Ocfs2-devel] [PATCH 17/34] ocfs2: Remove masklog ML_EXTENT_MAP.
2010-12-23 7:19 [Ocfs2-devel] [PATCH 00/34] OCFS2: Add trace event and replace mlog(0) Tao Ma
` (15 preceding siblings ...)
2010-12-23 7:30 ` [Ocfs2-devel] [PATCH 16/34] ocfs2: Remove masklog ML_INODE Tao Ma
@ 2010-12-23 7:31 ` Tao Ma
2010-12-23 7:31 ` [Ocfs2-devel] [PATCH 18/34] ocfs2: Remove mlog(0) from fs/ocfs2/slot_map.c Tao Ma
` (21 subsequent siblings)
38 siblings, 0 replies; 53+ messages in thread
From: Tao Ma @ 2010-12-23 7:31 UTC (permalink / raw)
To: ocfs2-devel
From: Tao Ma <boyu.mt@taobao.com>
Remove mlog(0) from fs/ocfs2/extent_map.c and the masklog EXTENT_MAP.
Signed-off-by: Tao Ma <boyu.mt@taobao.com>
---
fs/ocfs2/cluster/masklog.c | 1 -
fs/ocfs2/cluster/masklog.h | 1 -
fs/ocfs2/extent_map.c | 5 ++---
fs/ocfs2/ocfs2_trace.h | 32 ++++++++++++++++++++++++++++++++
4 files changed, 34 insertions(+), 5 deletions(-)
diff --git a/fs/ocfs2/cluster/masklog.c b/fs/ocfs2/cluster/masklog.c
index 1c1fa9d..de807d2 100644
--- a/fs/ocfs2/cluster/masklog.c
+++ b/fs/ocfs2/cluster/masklog.c
@@ -94,7 +94,6 @@ static struct mlog_attribute mlog_attrs[MLOG_MAX_BITS] = {
define_mask(AIO),
define_mask(JOURNAL),
define_mask(SUPER),
- define_mask(EXTENT_MAP),
define_mask(DLM_GLUE),
define_mask(BH_IO),
define_mask(UPTODATE),
diff --git a/fs/ocfs2/cluster/masklog.h b/fs/ocfs2/cluster/masklog.h
index d6dab10..fca684c 100644
--- a/fs/ocfs2/cluster/masklog.h
+++ b/fs/ocfs2/cluster/masklog.h
@@ -96,7 +96,6 @@
#define ML_AIO 0x0000000000002000ULL /* ocfs2 aio read and write */
#define ML_JOURNAL 0x0000000000004000ULL /* ocfs2 journalling functions */
#define ML_SUPER 0x0000000000010000ULL /* ocfs2 mount / umount */
-#define ML_EXTENT_MAP 0x0000000000040000ULL /* ocfs2 extent map caching */
#define ML_DLM_GLUE 0x0000000000080000ULL /* ocfs2 dlm glue layer */
#define ML_BH_IO 0x0000000000100000ULL /* ocfs2 buffer I/O */
#define ML_UPTODATE 0x0000000000200000ULL /* ocfs2 caching sequence #'s */
diff --git a/fs/ocfs2/extent_map.c b/fs/ocfs2/extent_map.c
index 7d74d17..23457b4 100644
--- a/fs/ocfs2/extent_map.c
+++ b/fs/ocfs2/extent_map.c
@@ -28,7 +28,6 @@
#include <linux/types.h>
#include <linux/fiemap.h>
-#define MLOG_MASK_PREFIX ML_EXTENT_MAP
#include <cluster/masklog.h>
#include "ocfs2.h"
@@ -39,6 +38,7 @@
#include "inode.h"
#include "super.h"
#include "symlink.h"
+#include "ocfs2_trace.h"
#include "buffer_head_io.h"
@@ -841,8 +841,7 @@ int ocfs2_read_virt_blocks(struct inode *inode, u64 v_block, int nr,
u64 p_block, p_count;
int i, count, done = 0;
- mlog(0, "(inode = %p, v_block = %llu, nr = %d, bhs = %p, "
- "flags = %x, validate = %p)\n",
+ trace_ocfs2_read_virt_blocks(
inode, (unsigned long long)v_block, nr, bhs, flags,
validate);
diff --git a/fs/ocfs2/ocfs2_trace.h b/fs/ocfs2/ocfs2_trace.h
index d9b13c1..af7a87e 100644
--- a/fs/ocfs2/ocfs2_trace.h
+++ b/fs/ocfs2/ocfs2_trace.h
@@ -2252,6 +2252,38 @@ DEFINE_EVENT(ocfs2__ull, ocfs2_mark_inode_dirty,
);
/* End of trace events for fs/ocfs2/inode.c. */
+
+/* Trace events for fs/ocfs2/extent_map.c. */
+TRACE_EVENT(ocfs2_read_virt_blocks,
+
+ TP_PROTO(void *inode, unsigned long long vblock, int nr,
+ void *bhs, unsigned int flags, void *validate),
+
+ TP_ARGS(inode, vblock, nr, bhs, flags, validate),
+
+ TP_STRUCT__entry(
+ __field( void *, inode )
+ __field( unsigned long long, vblock )
+ __field( int, nr )
+ __field( void *, bhs )
+ __field( unsigned int, flags )
+ __field( void *, validate )
+ ),
+
+ TP_fast_assign(
+ __entry->inode = inode;
+ __entry->vblock = vblock;
+ __entry->nr = nr;
+ __entry->bhs = bhs;
+ __entry->flags = flags;
+ __entry->validate = validate;
+ ),
+
+ TP_printk("%p %llu %d %p %x %p", __entry->inode, __entry->vblock,
+ __entry->nr, __entry->bhs, __entry->flags, __entry->validate)
+);
+/* End of trace events for fs/ocfs2/extent_map.c. */
+
#endif /* _TRACE_OCFS2_H */
/* This part must be outside protection */
--
1.7.0.4
^ permalink raw reply related [flat|nested] 53+ messages in thread
* [Ocfs2-devel] [PATCH 18/34] ocfs2: Remove mlog(0) from fs/ocfs2/slot_map.c
2010-12-23 7:19 [Ocfs2-devel] [PATCH 00/34] OCFS2: Add trace event and replace mlog(0) Tao Ma
` (16 preceding siblings ...)
2010-12-23 7:31 ` [Ocfs2-devel] [PATCH 17/34] ocfs2: Remove masklog ML_EXTENT_MAP Tao Ma
@ 2010-12-23 7:31 ` Tao Ma
2010-12-23 7:31 ` [Ocfs2-devel] [PATCH 19/34] ocfs2: Remove mlog(0) from fs/ocfs2/heartbeat.c Tao Ma
` (20 subsequent siblings)
38 siblings, 0 replies; 53+ messages in thread
From: Tao Ma @ 2010-12-23 7:31 UTC (permalink / raw)
To: ocfs2-devel
From: Tao Ma <boyu.mt@taobao.com>
This is the 1st step to remove the debug info of SUPER.
Signed-off-by: Tao Ma <boyu.mt@taobao.com>
---
fs/ocfs2/ocfs2_trace.h | 30 ++++++++++++++++++++++++++++++
fs/ocfs2/slot_map.c | 13 +++++--------
2 files changed, 35 insertions(+), 8 deletions(-)
diff --git a/fs/ocfs2/ocfs2_trace.h b/fs/ocfs2/ocfs2_trace.h
index af7a87e..0856f29 100644
--- a/fs/ocfs2/ocfs2_trace.h
+++ b/fs/ocfs2/ocfs2_trace.h
@@ -2284,6 +2284,36 @@ TRACE_EVENT(ocfs2_read_virt_blocks,
);
/* End of trace events for fs/ocfs2/extent_map.c. */
+/* Trace events for fs/ocfs2/slot_map.c. */
+DEFINE_EVENT(ocfs2__uint, ocfs2_refresh_slot_info,
+
+ TP_PROTO(unsigned int blocks),
+
+ TP_ARGS(blocks)
+);
+
+DEFINE_EVENT(ocfs2__ull_uint, ocfs2_map_slot_buffers,
+
+ TP_PROTO(unsigned long long bytes, unsigned int blocks),
+
+ TP_ARGS(bytes, blocks)
+);
+
+DEFINE_EVENT(ocfs2__ull_uint, ocfs2_map_slot_buffers_block,
+
+ TP_PROTO(unsigned long long blkno, unsigned int i),
+
+ TP_ARGS(blkno, i)
+);
+
+DEFINE_EVENT(ocfs2__int, ocfs2_find_slot,
+
+ TP_PROTO(int slot),
+
+ TP_ARGS(slot)
+);
+
+/* End of trace events for fs/ocfs2/slot_map.c. */
#endif /* _TRACE_OCFS2_H */
/* This part must be outside protection */
diff --git a/fs/ocfs2/slot_map.c b/fs/ocfs2/slot_map.c
index 93b449e..26fc001 100644
--- a/fs/ocfs2/slot_map.c
+++ b/fs/ocfs2/slot_map.c
@@ -27,7 +27,6 @@
#include <linux/slab.h>
#include <linux/highmem.h>
-#define MLOG_MASK_PREFIX ML_SUPER
#include <cluster/masklog.h>
#include "ocfs2.h"
@@ -39,6 +38,7 @@
#include "slot_map.h"
#include "super.h"
#include "sysfile.h"
+#include "ocfs2_trace.h"
#include "buffer_head_io.h"
@@ -142,8 +142,7 @@ int ocfs2_refresh_slot_info(struct ocfs2_super *osb)
BUG_ON(si->si_blocks == 0);
BUG_ON(si->si_bh == NULL);
- mlog(0, "Refreshing slot map, reading %u block(s)\n",
- si->si_blocks);
+ trace_ocfs2_refresh_slot_info(si->si_blocks);
/*
* We pass -1 as blocknr because we expect all of si->si_bh to
@@ -381,8 +380,7 @@ static int ocfs2_map_slot_buffers(struct ocfs2_super *osb,
/* The size checks above should ensure this */
BUG_ON((osb->max_slots / si->si_slots_per_block) > blocks);
- mlog(0, "Slot map needs %u buffers for %llu bytes\n",
- si->si_blocks, bytes);
+ trace_ocfs2_map_slot_buffers(bytes, si->si_blocks);
si->si_bh = kzalloc(sizeof(struct buffer_head *) * si->si_blocks,
GFP_KERNEL);
@@ -400,8 +398,7 @@ static int ocfs2_map_slot_buffers(struct ocfs2_super *osb,
goto bail;
}
- mlog(0, "Reading slot map block %u at %llu\n", i,
- (unsigned long long)blkno);
+ trace_ocfs2_map_slot_buffers_block((unsigned long long)blkno, i);
bh = NULL; /* Acquire a fresh bh */
status = ocfs2_read_blocks(INODE_CACHE(si->si_inode), blkno,
@@ -503,7 +500,7 @@ int ocfs2_find_slot(struct ocfs2_super *osb)
osb->slot_num = slot;
spin_unlock(&osb->osb_lock);
- mlog(0, "taking node slot %d\n", osb->slot_num);
+ trace_ocfs2_find_slot(osb->slot_num);
status = ocfs2_update_disk_slot(osb, si, osb->slot_num);
if (status < 0)
--
1.7.0.4
^ permalink raw reply related [flat|nested] 53+ messages in thread
* [Ocfs2-devel] [PATCH 19/34] ocfs2: Remove mlog(0) from fs/ocfs2/heartbeat.c
2010-12-23 7:19 [Ocfs2-devel] [PATCH 00/34] OCFS2: Add trace event and replace mlog(0) Tao Ma
` (17 preceding siblings ...)
2010-12-23 7:31 ` [Ocfs2-devel] [PATCH 18/34] ocfs2: Remove mlog(0) from fs/ocfs2/slot_map.c Tao Ma
@ 2010-12-23 7:31 ` Tao Ma
2010-12-23 7:31 ` [Ocfs2-devel] [PATCH 20/34] ocfs2: Remove masklog ML_SUPER Tao Ma
` (19 subsequent siblings)
38 siblings, 0 replies; 53+ messages in thread
From: Tao Ma @ 2010-12-23 7:31 UTC (permalink / raw)
To: ocfs2-devel
From: Tao Ma <boyu.mt@taobao.com>
This is the 2nd step to remove the debug info of SUPER.
Signed-off-by: Tao Ma <boyu.mt@taobao.com>
---
fs/ocfs2/heartbeat.c | 4 ++--
fs/ocfs2/ocfs2_trace.h | 9 +++++++++
2 files changed, 11 insertions(+), 2 deletions(-)
diff --git a/fs/ocfs2/heartbeat.c b/fs/ocfs2/heartbeat.c
index 1aa863d..d8208b2 100644
--- a/fs/ocfs2/heartbeat.c
+++ b/fs/ocfs2/heartbeat.c
@@ -28,7 +28,6 @@
#include <linux/types.h>
#include <linux/highmem.h>
-#define MLOG_MASK_PREFIX ML_SUPER
#include <cluster/masklog.h>
#include "ocfs2.h"
@@ -37,6 +36,7 @@
#include "heartbeat.h"
#include "inode.h"
#include "journal.h"
+#include "ocfs2_trace.h"
#include "buffer_head_io.h"
@@ -66,7 +66,7 @@ void ocfs2_do_node_down(int node_num, void *data)
BUG_ON(osb->node_num == node_num);
- mlog(0, "ocfs2: node down event for %d\n", node_num);
+ trace_ocfs2_do_node_down(node_num);
if (!osb->cconn) {
/*
diff --git a/fs/ocfs2/ocfs2_trace.h b/fs/ocfs2/ocfs2_trace.h
index 0856f29..576af58 100644
--- a/fs/ocfs2/ocfs2_trace.h
+++ b/fs/ocfs2/ocfs2_trace.h
@@ -2284,6 +2284,15 @@ TRACE_EVENT(ocfs2_read_virt_blocks,
);
/* End of trace events for fs/ocfs2/extent_map.c. */
+/* Trace events for fs/ocfs2/heartbeat.c. */
+DEFINE_EVENT(ocfs2__int, ocfs2_do_node_down,
+
+ TP_PROTO(int node_num),
+
+ TP_ARGS(node_num)
+);
+/* End of trace events for fs/ocfs2/heartbeat.c. */
+
/* Trace events for fs/ocfs2/slot_map.c. */
DEFINE_EVENT(ocfs2__uint, ocfs2_refresh_slot_info,
--
1.7.0.4
^ permalink raw reply related [flat|nested] 53+ messages in thread
* [Ocfs2-devel] [PATCH 20/34] ocfs2: Remove masklog ML_SUPER.
2010-12-23 7:19 [Ocfs2-devel] [PATCH 00/34] OCFS2: Add trace event and replace mlog(0) Tao Ma
` (18 preceding siblings ...)
2010-12-23 7:31 ` [Ocfs2-devel] [PATCH 19/34] ocfs2: Remove mlog(0) from fs/ocfs2/heartbeat.c Tao Ma
@ 2010-12-23 7:31 ` Tao Ma
2010-12-23 7:31 ` [Ocfs2-devel] [PATCH 21/34] ocfs2: Remove masklog ML_XATTR Tao Ma
` (18 subsequent siblings)
38 siblings, 0 replies; 53+ messages in thread
From: Tao Ma @ 2010-12-23 7:31 UTC (permalink / raw)
To: ocfs2-devel
From: Tao Ma <boyu.mt@taobao.com>
Remove mlog(0) from fs/ocfs2/super.c and the masklog SUPER.
Signed-off-by: Tao Ma <boyu.mt@taobao.com>
---
fs/ocfs2/cluster/masklog.c | 1 -
fs/ocfs2/cluster/masklog.h | 1 -
fs/ocfs2/ocfs2_trace.h | 141 ++++++++++++++++++++++++++++++++++++++++++++
fs/ocfs2/super.c | 29 +++------
4 files changed, 151 insertions(+), 21 deletions(-)
diff --git a/fs/ocfs2/cluster/masklog.c b/fs/ocfs2/cluster/masklog.c
index de807d2..b1e1799 100644
--- a/fs/ocfs2/cluster/masklog.c
+++ b/fs/ocfs2/cluster/masklog.c
@@ -93,7 +93,6 @@ static struct mlog_attribute mlog_attrs[MLOG_MAX_BITS] = {
define_mask(DLM_RECOVERY),
define_mask(AIO),
define_mask(JOURNAL),
- define_mask(SUPER),
define_mask(DLM_GLUE),
define_mask(BH_IO),
define_mask(UPTODATE),
diff --git a/fs/ocfs2/cluster/masklog.h b/fs/ocfs2/cluster/masklog.h
index fca684c..c6fc51f 100644
--- a/fs/ocfs2/cluster/masklog.h
+++ b/fs/ocfs2/cluster/masklog.h
@@ -95,7 +95,6 @@
#define ML_DLM_RECOVERY 0x0000000000001000ULL /* dlm master functions */
#define ML_AIO 0x0000000000002000ULL /* ocfs2 aio read and write */
#define ML_JOURNAL 0x0000000000004000ULL /* ocfs2 journalling functions */
-#define ML_SUPER 0x0000000000010000ULL /* ocfs2 mount / umount */
#define ML_DLM_GLUE 0x0000000000080000ULL /* ocfs2 dlm glue layer */
#define ML_BH_IO 0x0000000000100000ULL /* ocfs2 buffer I/O */
#define ML_UPTODATE 0x0000000000200000ULL /* ocfs2 caching sequence #'s */
diff --git a/fs/ocfs2/ocfs2_trace.h b/fs/ocfs2/ocfs2_trace.h
index 576af58..28c4a05 100644
--- a/fs/ocfs2/ocfs2_trace.h
+++ b/fs/ocfs2/ocfs2_trace.h
@@ -57,6 +57,23 @@ DECLARE_EVENT_CLASS(ocfs2__ull,
TP_printk("%llu", __entry->blkno)
);
+DECLARE_EVENT_CLASS(ocfs2__pointer,
+
+ TP_PROTO(void *pointer),
+
+ TP_ARGS(pointer),
+
+ TP_STRUCT__entry(
+ __field( void *, pointer )
+ ),
+
+ TP_fast_assign(
+ __entry->pointer = pointer;
+ ),
+
+ TP_printk("%p", __entry->pointer)
+);
+
DECLARE_EVENT_CLASS(ocfs2__int_int,
TP_PROTO(int value1, int value2),
@@ -2323,6 +2340,130 @@ DEFINE_EVENT(ocfs2__int, ocfs2_find_slot,
);
/* End of trace events for fs/ocfs2/slot_map.c. */
+
+/* Trace events for fs/ocfs2/super.c. */
+TRACE_EVENT(ocfs2_remount,
+
+ TP_PROTO(unsigned long s_flags, unsigned long osb_flags, int flags),
+
+ TP_ARGS(s_flags, osb_flags, flags),
+
+ TP_STRUCT__entry(
+ __field( unsigned long, s_flags )
+ __field( unsigned long, osb_flags )
+ __field( int, flags )
+ ),
+
+ TP_fast_assign(
+ __entry->s_flags = s_flags;
+ __entry->osb_flags = osb_flags;
+ __entry->flags = flags;
+ ),
+
+ TP_printk("%lu %lu %d", __entry->s_flags,
+ __entry->osb_flags, __entry->flags)
+);
+
+TRACE_EVENT(ocfs2_fill_super,
+
+ TP_PROTO(void *sb, void *data, int silent),
+
+ TP_ARGS(sb, data, silent),
+
+ TP_STRUCT__entry(
+ __field( void *, sb )
+ __field( void *, data )
+ __field( int, silent )
+ ),
+
+ TP_fast_assign(
+ __entry->sb = sb;
+ __entry->data = data;
+ __entry->silent = silent;
+ ),
+
+ TP_printk("%p %p %d", __entry->sb, __entry->data, __entry->silent)
+);
+
+TRACE_EVENT(ocfs2_parse_options,
+
+ TP_PROTO(int is_remount, char *options),
+
+ TP_ARGS(is_remount, options),
+
+ TP_STRUCT__entry(
+ __field( int, is_remount )
+ __field( char *, options )
+ ),
+
+ TP_fast_assign(
+ __entry->is_remount = is_remount;
+ __entry->options = options;
+ ),
+
+ TP_printk("%d %s", __entry->is_remount, __entry->options)
+);
+
+DEFINE_EVENT(ocfs2__pointer, ocfs2_put_super,
+
+ TP_PROTO(void *sb),
+
+ TP_ARGS(sb)
+);
+
+TRACE_EVENT(ocfs2_statfs,
+
+ TP_PROTO(void *sb, void *buf),
+
+ TP_ARGS(sb, buf),
+
+ TP_STRUCT__entry(
+ __field( void *, sb )
+ __field( void *, buf )
+ ),
+
+ TP_fast_assign(
+ __entry->sb = sb;
+ __entry->buf = buf;
+ ),
+
+ TP_printk("%p %p", __entry->sb, __entry->buf)
+);
+
+DEFINE_EVENT(ocfs2__pointer, ocfs2_dismount_volume,
+
+ TP_PROTO(void *sb),
+
+ TP_ARGS(sb)
+);
+
+TRACE_EVENT(ocfs2_initialize_super,
+
+ TP_PROTO(char *label, char *uuid_str, unsigned long long root_dir,
+ unsigned long long system_dir, int cluster_bits),
+
+ TP_ARGS(label, uuid_str, root_dir, system_dir, cluster_bits),
+
+ TP_STRUCT__entry(
+ __field( char *, label )
+ __field( char *, uuid_str )
+ __field( unsigned long long, root_dir )
+ __field( unsigned long long, system_dir )
+ __field( int, cluster_bits )
+ ),
+
+ TP_fast_assign(
+ __entry->label = label;
+ __entry->uuid_str = uuid_str;
+ __entry->root_dir = root_dir;
+ __entry->system_dir = system_dir;
+ __entry->cluster_bits = cluster_bits;
+ ),
+
+ TP_printk("%s %s %llu %llu %d", __entry->label, __entry->uuid_str,
+ __entry->root_dir, __entry->system_dir, __entry->cluster_bits)
+);
+/* End of trace events for fs/ocfs2/super.c. */
#endif /* _TRACE_OCFS2_H */
/* This part must be outside protection */
diff --git a/fs/ocfs2/super.c b/fs/ocfs2/super.c
index e88013c..b8b0553 100644
--- a/fs/ocfs2/super.c
+++ b/fs/ocfs2/super.c
@@ -45,7 +45,6 @@
#define CREATE_TRACE_POINTS
#include "ocfs2_trace.h"
-#define MLOG_MASK_PREFIX ML_SUPER
#include <cluster/masklog.h>
#include "ocfs2.h"
@@ -673,12 +672,9 @@ static int ocfs2_remount(struct super_block *sb, int *flags, char *data)
}
if (*flags & MS_RDONLY) {
- mlog(0, "Going to ro mode.\n");
sb->s_flags |= MS_RDONLY;
osb->osb_flags |= OCFS2_OSB_SOFT_RO;
} else {
- mlog(0, "Making ro filesystem writeable.\n");
-
if (osb->osb_flags & OCFS2_OSB_ERROR_FS) {
mlog(ML_ERROR, "Cannot remount RDWR "
"filesystem due to previous errors.\n");
@@ -696,6 +692,7 @@ static int ocfs2_remount(struct super_block *sb, int *flags, char *data)
sb->s_flags &= ~MS_RDONLY;
osb->osb_flags &= ~OCFS2_OSB_SOFT_RO;
}
+ trace_ocfs2_remount(sb->s_flags, osb->osb_flags, *flags);
unlock_osb:
spin_unlock(&osb->osb_lock);
/* Enable quota accounting after remounting RW */
@@ -1022,7 +1019,7 @@ static int ocfs2_fill_super(struct super_block *sb, void *data, int silent)
char nodestr[8];
struct ocfs2_blockcheck_stats stats;
- mlog(0, "%p, %p, %i", sb, data, silent);
+ trace_ocfs2_fill_super(sb, data, silent);
if (!ocfs2_parse_options(sb, data, &parsed_options, 0)) {
status = -EINVAL;
@@ -1309,8 +1306,7 @@ static int ocfs2_parse_options(struct super_block *sb,
char *p;
u32 tmp;
- mlog(0, "remount: %d, options: \"%s\"\n", is_remount,
- options ? options : "(none)");
+ trace_ocfs2_parse_options(is_remount, options ? options : "(none)");
mopt->commit_interval = 0;
mopt->mount_opt = OCFS2_MOUNT_NOINTR;
@@ -1674,7 +1670,7 @@ static void __exit ocfs2_exit(void)
static void ocfs2_put_super(struct super_block *sb)
{
- mlog(0, "(0x%p)\n", sb);
+ trace_ocfs2_put_super(sb);
ocfs2_sync_blockdev(sb);
ocfs2_dismount_volume(sb, 0);
@@ -1689,7 +1685,7 @@ static int ocfs2_statfs(struct dentry *dentry, struct kstatfs *buf)
struct buffer_head *bh = NULL;
struct inode *inode = NULL;
- mlog(0, "(%p, %p)\n", dentry->d_sb, buf);
+ trace_ocfs2_statfs(dentry->d_sb, buf);
osb = OCFS2_SB(dentry->d_sb);
@@ -1910,7 +1906,7 @@ static void ocfs2_dismount_volume(struct super_block *sb, int mnt_err)
struct ocfs2_super *osb = NULL;
char nodestr[8];
- mlog(0, "(0x%p)\n", sb);
+ trace_ocfs2_dismount_volume(sb);
BUG_ON(!sb);
osb = OCFS2_SB(sb);
@@ -2124,7 +2120,6 @@ static int ocfs2_initialize_super(struct super_block *sb,
status = -EINVAL;
goto bail;
}
- mlog(0, "max_slots for this device: %u\n", osb->max_slots);
ocfs2_orphan_scan_init(osb);
@@ -2263,7 +2258,6 @@ static int ocfs2_initialize_super(struct super_block *sb,
osb->s_clustersize_bits =
le32_to_cpu(di->id2.i_super.s_clustersize_bits);
osb->s_clustersize = 1 << osb->s_clustersize_bits;
- mlog(0, "clusterbits=%d\n", osb->s_clustersize_bits);
if (osb->s_clustersize < OCFS2_MIN_CLUSTERSIZE ||
osb->s_clustersize > OCFS2_MAX_CLUSTERSIZE) {
@@ -2302,11 +2296,10 @@ static int ocfs2_initialize_super(struct super_block *sb,
le64_to_cpu(di->id2.i_super.s_first_cluster_group);
osb->fs_generation = le32_to_cpu(di->i_fs_generation);
osb->uuid_hash = le32_to_cpu(di->id2.i_super.s_uuid_hash);
- mlog(0, "vol_label: %s\n", osb->vol_label);
- mlog(0, "uuid: %s\n", osb->uuid_str);
- mlog(0, "root_blkno=%llu, system_dir_blkno=%llu\n",
- (unsigned long long)osb->root_blkno,
- (unsigned long long)osb->system_dir_blkno);
+ trace_ocfs2_initialize_super(osb->vol_label, osb->uuid_str,
+ (unsigned long long)osb->root_blkno,
+ (unsigned long long)osb->system_dir_blkno,
+ osb->s_clustersize_bits);
osb->osb_dlm_debug = ocfs2_new_dlm_debug();
if (!osb->osb_dlm_debug) {
@@ -2481,8 +2474,6 @@ static int ocfs2_check_volume(struct ocfs2_super *osb)
* ourselves as mounted. */
}
- mlog(0, "Journal loaded.\n");
-
status = ocfs2_load_local_alloc(osb);
if (status < 0) {
mlog_errno(status);
--
1.7.0.4
^ permalink raw reply related [flat|nested] 53+ messages in thread
* [Ocfs2-devel] [PATCH 21/34] ocfs2: Remove masklog ML_XATTR.
2010-12-23 7:19 [Ocfs2-devel] [PATCH 00/34] OCFS2: Add trace event and replace mlog(0) Tao Ma
` (19 preceding siblings ...)
2010-12-23 7:31 ` [Ocfs2-devel] [PATCH 20/34] ocfs2: Remove masklog ML_SUPER Tao Ma
@ 2010-12-23 7:31 ` Tao Ma
2010-12-23 7:31 ` [Ocfs2-devel] [PATCH 22/34] ocfs2: Remove masklog ML_RESERVATIONS Tao Ma
` (17 subsequent siblings)
38 siblings, 0 replies; 53+ messages in thread
From: Tao Ma @ 2010-12-23 7:31 UTC (permalink / raw)
To: ocfs2-devel
From: Tao Ma <boyu.mt@taobao.com>
Remove mlog(0) from fs/ocfs2/xattr.c and the masklog ML_XATTR.
Signed-off-by: Tao Ma <boyu.mt@taobao.com>
---
fs/ocfs2/cluster/masklog.c | 1 -
fs/ocfs2/cluster/masklog.h | 1 -
fs/ocfs2/ocfs2_trace.h | 342 ++++++++++++++++++++++++++++++++++++++++++++
fs/ocfs2/xattr.c | 153 ++++++++++----------
4 files changed, 417 insertions(+), 80 deletions(-)
diff --git a/fs/ocfs2/cluster/masklog.c b/fs/ocfs2/cluster/masklog.c
index b1e1799..d31c5f1 100644
--- a/fs/ocfs2/cluster/masklog.c
+++ b/fs/ocfs2/cluster/masklog.c
@@ -102,7 +102,6 @@ static struct mlog_attribute mlog_attrs[MLOG_MAX_BITS] = {
define_mask(CONN),
define_mask(QUORUM),
define_mask(EXPORT),
- define_mask(XATTR),
define_mask(QUOTA),
define_mask(BASTS),
define_mask(ERROR),
diff --git a/fs/ocfs2/cluster/masklog.h b/fs/ocfs2/cluster/masklog.h
index c6fc51f..736d18c 100644
--- a/fs/ocfs2/cluster/masklog.h
+++ b/fs/ocfs2/cluster/masklog.h
@@ -104,7 +104,6 @@
#define ML_CONN 0x0000000004000000ULL /* net connection management */
#define ML_QUORUM 0x0000000008000000ULL /* net connection quorum */
#define ML_EXPORT 0x0000000010000000ULL /* ocfs2 export operations */
-#define ML_XATTR 0x0000000020000000ULL /* ocfs2 extended attributes */
#define ML_QUOTA 0x0000000040000000ULL /* ocfs2 quota operations */
#define ML_BASTS 0x0000001000000000ULL /* dlmglue asts and basts */
/* bits that are infrequently given and frequently matched in the high word */
diff --git a/fs/ocfs2/ocfs2_trace.h b/fs/ocfs2/ocfs2_trace.h
index 28c4a05..676e75d 100644
--- a/fs/ocfs2/ocfs2_trace.h
+++ b/fs/ocfs2/ocfs2_trace.h
@@ -74,6 +74,23 @@ DECLARE_EVENT_CLASS(ocfs2__pointer,
TP_printk("%p", __entry->pointer)
);
+DECLARE_EVENT_CLASS(ocfs2__string,
+
+ TP_PROTO(const char *name),
+
+ TP_ARGS(name),
+
+ TP_STRUCT__entry(
+ __field( const char *, name )
+ ),
+
+ TP_fast_assign(
+ __entry->name = name;
+ ),
+
+ TP_printk("%s", __entry->name)
+);
+
DECLARE_EVENT_CLASS(ocfs2__int_int,
TP_PROTO(int value1, int value2),
@@ -311,6 +328,32 @@ DECLARE_EVENT_CLASS(ocfs2__ull_uint_uint_uint,
__entry->value2, __entry->value3)
);
+DECLARE_EVENT_CLASS(ocfs2__ull_ull_uint_uint,
+
+ TP_PROTO(unsigned long long value1, unsigned long long value2,
+ unsigned int value3, unsigned int value4),
+
+ TP_ARGS(value1, value2, value3, value4),
+
+ TP_STRUCT__entry(
+ __field( unsigned long long, value1 )
+ __field( unsigned long long, value2 )
+ __field( unsigned int, value3 )
+ __field( unsigned int, value4 )
+ ),
+
+ TP_fast_assign(
+ __entry->value1 = value1;
+ __entry->value2 = value2;
+ __entry->value3 = value3;
+ __entry->value4 = value4;
+ ),
+
+ TP_printk("%llu %llu %u %u",
+ __entry->value1, __entry->value2,
+ __entry->value3, __entry->value4)
+);
+
/* Trace events for fs/ocfs2/alloc.c. */
DECLARE_EVENT_CLASS(ocfs2__btree_ops,
@@ -2464,6 +2507,305 @@ TRACE_EVENT(ocfs2_initialize_super,
__entry->root_dir, __entry->system_dir, __entry->cluster_bits)
);
/* End of trace events for fs/ocfs2/super.c. */
+
+/* Trace events for fs/ocfs2/xattr.c. */
+DEFINE_EVENT(ocfs2__ull, ocfs2_validate_xattr_block,
+
+ TP_PROTO(unsigned long long blkno),
+
+ TP_ARGS(blkno)
+);
+
+DEFINE_EVENT(ocfs2__uint, ocfs2_xattr_extend_allocation,
+
+ TP_PROTO(unsigned int clusters),
+
+ TP_ARGS(clusters)
+);
+
+TRACE_EVENT(ocfs2_init_xattr_set_ctxt,
+
+ TP_PROTO(const char *name, int meta, int clusters, int credits),
+
+ TP_ARGS(name, meta, clusters, credits),
+
+ TP_STRUCT__entry(
+ __field( const char *, name )
+ __field( int, meta )
+ __field( int, clusters )
+ __field( int, credits )
+ ),
+
+ TP_fast_assign(
+ __entry->name = name;
+ __entry->meta = meta;
+ __entry->clusters = clusters;
+ __entry->credits = credits;
+ ),
+
+ TP_printk("%s %d %d %d", __entry->name, __entry->meta,
+ __entry->clusters, __entry->credits)
+);
+
+DECLARE_EVENT_CLASS(ocfs2__xattr_find,
+
+ TP_PROTO(unsigned long long ino, const char *name, int name_index,
+ unsigned int hash, unsigned long long location,
+ int xe_index),
+
+ TP_ARGS(ino, name, name_index, hash, location, xe_index),
+
+ TP_STRUCT__entry(
+ __field( unsigned long long, ino )
+ __field( const char *, name )
+ __field( int, name_index )
+ __field( unsigned int, hash )
+ __field( unsigned long long, location )
+ __field( int, xe_index )
+ ),
+
+ TP_fast_assign(
+ __entry->ino = ino;
+ __entry->name = name;
+ __entry->name_index = name_index;
+ __entry->hash = hash;
+ __entry->location = location;
+ __entry->xe_index = xe_index;
+ ),
+
+ TP_printk("%llu %s %d %u %llu %d", __entry->ino, __entry->name,
+ __entry->name_index, __entry->hash, __entry->location,
+ __entry->xe_index)
+);
+
+DEFINE_EVENT(ocfs2__xattr_find, ocfs2_xattr_bucket_find,
+
+ TP_PROTO(unsigned long long ino, const char *name, int name_index,
+ unsigned int hash, unsigned long long bucket,
+ int xe_index),
+
+ TP_ARGS(ino, name, name_index, hash, bucket, xe_index)
+);
+
+DEFINE_EVENT(ocfs2__xattr_find, ocfs2_xattr_index_block_find,
+
+ TP_PROTO(unsigned long long ino, const char *name, int name_index,
+ unsigned int hash, unsigned long long bucket,
+ int xe_index),
+
+ TP_ARGS(ino, name, name_index, hash, bucket, xe_index)
+);
+
+DEFINE_EVENT(ocfs2__xattr_find, ocfs2_xattr_index_block_find_rec,
+
+ TP_PROTO(unsigned long long ino, const char *name, int name_index,
+ unsigned int hash, unsigned long long blkno,
+ int num_clusters),
+
+ TP_ARGS(ino, name, name_index, hash, blkno, num_clusters)
+);
+
+DEFINE_EVENT(ocfs2__ull_ull_uint, ocfs2_iterate_xattr_buckets,
+
+ TP_PROTO(unsigned long long ino,
+ unsigned long long blkno, unsigned int clusters),
+
+ TP_ARGS(ino, blkno, clusters)
+);
+
+DEFINE_EVENT(ocfs2__ull_uint, ocfs2_iterate_xattr_bucket,
+
+ TP_PROTO(unsigned long long blkno, unsigned int clusters),
+
+ TP_ARGS(blkno, clusters)
+);
+
+DEFINE_EVENT(ocfs2__ull_ull, ocfs2_cp_xattr_block_to_bucket_begin,
+
+ TP_PROTO(unsigned long long blkno, unsigned long long bucket),
+
+ TP_ARGS(blkno, bucket)
+);
+
+DEFINE_EVENT(ocfs2__uint_uint_uint, ocfs2_cp_xattr_block_to_bucket_end,
+
+ TP_PROTO(unsigned int offset, unsigned int size,
+ unsigned int off_change),
+
+ TP_ARGS(offset, size, off_change)
+);
+
+DEFINE_EVENT(ocfs2__ull, ocfs2_xattr_create_index_block_begin,
+
+ TP_PROTO(unsigned long long blkno),
+
+ TP_ARGS(blkno)
+);
+
+DEFINE_EVENT(ocfs2__ull, ocfs2_xattr_create_index_block,
+
+ TP_PROTO(unsigned long long bucket),
+
+ TP_ARGS(bucket)
+);
+
+DEFINE_EVENT(ocfs2__ull_uint_uint_uint, ocfs2_defrag_xattr_bucket,
+
+ TP_PROTO(unsigned long long blkno, unsigned int count,
+ unsigned int free_start, unsigned int name_value_len),
+
+ TP_ARGS(blkno, count, free_start, name_value_len)
+);
+
+DEFINE_EVENT(ocfs2__ull_ull, ocfs2_mv_xattr_bucket_cross_cluster,
+
+ TP_PROTO(unsigned long long last_cluster, unsigned long long new_blkno),
+
+ TP_ARGS(last_cluster, new_blkno)
+);
+
+DEFINE_EVENT(ocfs2__ull_ull, ocfs2_divide_xattr_bucket_begin,
+
+ TP_PROTO(unsigned long long blkno, unsigned long long new_blkno),
+
+ TP_ARGS(blkno, new_blkno)
+);
+
+DEFINE_EVENT(ocfs2__uint_uint_uint, ocfs2_divide_xattr_bucket_move,
+
+ TP_PROTO(unsigned int len, unsigned int from, unsigned int to),
+
+ TP_ARGS(len, from, to)
+);
+
+DEFINE_EVENT(ocfs2__ull_ull_uint, ocfs2_cp_xattr_bucket,
+
+ TP_PROTO(unsigned long long s_blkno,
+ unsigned long long t_blkno, unsigned int is_new),
+
+ TP_ARGS(s_blkno, t_blkno, is_new)
+);
+
+DEFINE_EVENT(ocfs2__ull_ull, ocfs2_mv_xattr_buckets,
+
+ TP_PROTO(unsigned long long s_blkno,
+ unsigned long long t_blkno),
+
+ TP_ARGS(s_blkno, t_blkno)
+);
+
+DEFINE_EVENT(ocfs2__ull_ull_uint, ocfs2_adjust_xattr_cross_cluster,
+
+ TP_PROTO(unsigned long long blkno,
+ unsigned long long new_blk, unsigned int clusters),
+
+ TP_ARGS(blkno, new_blk, clusters)
+);
+
+DEFINE_EVENT(ocfs2__ull_ull_uint_uint, ocfs2_add_new_xattr_cluster_begin,
+
+ TP_PROTO(unsigned long long ino, unsigned long long bucket,
+ unsigned int cpos, unsigned int clusters),
+
+ TP_ARGS(ino, bucket, cpos, clusters)
+);
+
+DEFINE_EVENT(ocfs2__ull_uint, ocfs2_add_new_xattr_cluster,
+
+ TP_PROTO(unsigned long long new_blkno, unsigned int clusters),
+
+ TP_ARGS(new_blkno, clusters)
+);
+
+DEFINE_EVENT(ocfs2__ull_uint_uint, ocfs2_add_new_xattr_cluster_insert,
+
+ TP_PROTO(unsigned long long new_blkno,
+ unsigned int cpos, unsigned int clusters),
+
+ TP_ARGS(new_blkno, cpos, clusters)
+);
+
+DEFINE_EVENT(ocfs2__ull_ull_uint_uint, ocfs2_extend_xattr_bucket,
+
+ TP_PROTO(unsigned long long target, unsigned long long bucket,
+ unsigned int num_clusters, unsigned int num_buckets),
+
+ TP_ARGS(target, bucket, num_clusters, num_buckets)
+);
+
+DEFINE_EVENT(ocfs2__ull, ocfs2_add_new_xattr_bucket,
+
+ TP_PROTO(unsigned long long target),
+
+ TP_ARGS(target)
+);
+
+DEFINE_EVENT(ocfs2__ull_uint_uint, ocfs2_xattr_bucket_value_truncate,
+
+ TP_PROTO(unsigned long long bucket,
+ unsigned int index, unsigned int len),
+
+ TP_ARGS(bucket, index, len)
+);
+
+DEFINE_EVENT(ocfs2__ull_ull_uint_uint, ocfs2_rm_xattr_cluster,
+
+ TP_PROTO(unsigned long long ino, unsigned long long blkno,
+ unsigned int cpos, unsigned int len),
+
+ TP_ARGS(ino, blkno, cpos, len)
+);
+
+DEFINE_EVENT(ocfs2__ull_uint, ocfs2_reflink_xattr_header,
+
+ TP_PROTO(unsigned long long blkno, unsigned int count),
+
+ TP_ARGS(blkno, count)
+);
+
+DEFINE_EVENT(ocfs2__ull_int, ocfs2_create_empty_xattr_block,
+
+ TP_PROTO(unsigned long long blkno, int indexed),
+
+ TP_ARGS(blkno, indexed)
+);
+
+DEFINE_EVENT(ocfs2__string, ocfs2_xattr_set_entry_bucket,
+
+ TP_PROTO(const char *name),
+
+ TP_ARGS(name)
+);
+
+DEFINE_EVENT(ocfs2__string, ocfs2_xattr_set_entry_index_block,
+
+ TP_PROTO(const char *name),
+
+ TP_ARGS(name)
+);
+
+DEFINE_EVENT(ocfs2__ull_uint, ocfs2_xattr_bucket_value_refcount,
+
+ TP_PROTO(unsigned long long blkno, unsigned int count),
+
+ TP_ARGS(blkno, count)
+);
+
+DEFINE_EVENT(ocfs2__ull_uint_uint, ocfs2_reflink_xattr_buckets,
+
+ TP_PROTO(unsigned long long new_blkno,
+ unsigned int num_clusters, unsigned int cpos),
+
+ TP_ARGS(new_blkno, num_clusters, cpos)
+);
+
+DEFINE_EVENT(ocfs2__ull_uint, ocfs2_reflink_xattr_rec,
+
+ TP_PROTO(unsigned long long blkno, unsigned int len),
+
+ TP_ARGS(blkno, len)
+);
+/* End of trace events for fs/ocfs2/xattr.c. */
#endif /* _TRACE_OCFS2_H */
/* This part must be outside protection */
diff --git a/fs/ocfs2/xattr.c b/fs/ocfs2/xattr.c
index d57c794..52e3dad 100644
--- a/fs/ocfs2/xattr.c
+++ b/fs/ocfs2/xattr.c
@@ -37,7 +37,6 @@
#include <linux/string.h>
#include <linux/security.h>
-#define MLOG_MASK_PREFIX ML_XATTR
#include <cluster/masklog.h>
#include "ocfs2.h"
@@ -57,6 +56,7 @@
#include "xattr.h"
#include "refcounttree.h"
#include "acl.h"
+#include "ocfs2_trace.h"
struct ocfs2_xattr_def_value_root {
struct ocfs2_xattr_value_root xv;
@@ -474,8 +474,7 @@ static int ocfs2_validate_xattr_block(struct super_block *sb,
struct ocfs2_xattr_block *xb =
(struct ocfs2_xattr_block *)bh->b_data;
- mlog(0, "Validating xattr block %llu\n",
- (unsigned long long)bh->b_blocknr);
+ trace_ocfs2_validate_xattr_block((unsigned long long)bh->b_blocknr);
BUG_ON(!buffer_uptodate(bh));
@@ -715,11 +714,11 @@ static int ocfs2_xattr_extend_allocation(struct inode *inode,
u32 prev_clusters, logical_start = le32_to_cpu(vb->vb_xv->xr_clusters);
struct ocfs2_extent_tree et;
- mlog(0, "(clusters_to_add for xattr= %u)\n", clusters_to_add);
-
ocfs2_init_xattr_value_extent_tree(&et, INODE_CACHE(inode), vb);
while (clusters_to_add) {
+ trace_ocfs2_xattr_extend_allocation(clusters_to_add);
+
status = vb->vb_access(handle, INODE_CACHE(inode), vb->vb_bh,
OCFS2_JOURNAL_ACCESS_WRITE);
if (status < 0) {
@@ -754,8 +753,6 @@ static int ocfs2_xattr_extend_allocation(struct inode *inode,
*/
BUG_ON(why == RESTART_META);
- mlog(0, "restarting xattr value extension for %u"
- " clusters,.\n", clusters_to_add);
credits = ocfs2_calc_extend_credits(inode->i_sb,
&vb->vb_xv->xr_list,
clusters_to_add);
@@ -3246,8 +3243,8 @@ static int ocfs2_init_xattr_set_ctxt(struct inode *inode,
}
meta_add += extra_meta;
- mlog(0, "Set xattr %s, reserve meta blocks = %d, clusters = %d, "
- "credits = %d\n", xi->xi_name, meta_add, clusters_add, *credits);
+ trace_ocfs2_init_xattr_set_ctxt(xi->xi_name, meta_add,
+ clusters_add, *credits);
if (meta_add) {
ret = ocfs2_reserve_new_metadata_blocks(osb, meta_add,
@@ -3887,8 +3884,10 @@ static int ocfs2_xattr_bucket_find(struct inode *inode,
if (found) {
xs->here = &xs->header->xh_entries[index];
- mlog(0, "find xattr %s in bucket %llu, entry = %u\n", name,
- (unsigned long long)bucket_blkno(xs->bucket), index);
+ trace_ocfs2_xattr_bucket_find(OCFS2_I(inode)->ip_blkno,
+ name, name_index, name_hash,
+ (unsigned long long)bucket_blkno(xs->bucket),
+ index);
} else
ret = -ENODATA;
@@ -3915,8 +3914,10 @@ static int ocfs2_xattr_index_block_find(struct inode *inode,
if (le16_to_cpu(el->l_next_free_rec) == 0)
return -ENODATA;
- mlog(0, "find xattr %s, hash = %u, index = %d in xattr tree\n",
- name, name_hash, name_index);
+ trace_ocfs2_xattr_index_block_find(OCFS2_I(inode)->ip_blkno,
+ name, name_index, name_hash,
+ (unsigned long long)root_bh->b_blocknr,
+ -1);
ret = ocfs2_xattr_get_rec(inode, name_hash, &p_blkno, &first_hash,
&num_clusters, el);
@@ -3927,9 +3928,10 @@ static int ocfs2_xattr_index_block_find(struct inode *inode,
BUG_ON(p_blkno == 0 || num_clusters == 0 || first_hash > name_hash);
- mlog(0, "find xattr extent rec %u clusters from %llu, the first hash "
- "in the rec is %u\n", num_clusters, (unsigned long long)p_blkno,
- first_hash);
+ trace_ocfs2_xattr_index_block_find_rec(OCFS2_I(inode)->ip_blkno,
+ name, name_index, first_hash,
+ (unsigned long long)p_blkno,
+ num_clusters);
ret = ocfs2_xattr_bucket_find(inode, name_index, name, name_hash,
p_blkno, first_hash, num_clusters, xs);
@@ -3955,8 +3957,9 @@ static int ocfs2_iterate_xattr_buckets(struct inode *inode,
return -ENOMEM;
}
- mlog(0, "iterating xattr buckets in %u clusters starting from %llu\n",
- clusters, (unsigned long long)blkno);
+ trace_ocfs2_iterate_xattr_buckets(
+ (unsigned long long)OCFS2_I(inode)->ip_blkno,
+ (unsigned long long)blkno, clusters);
for (i = 0; i < num_buckets; i++, blkno += bucket->bu_blocks) {
ret = ocfs2_read_xattr_bucket(bucket, blkno);
@@ -3972,8 +3975,7 @@ static int ocfs2_iterate_xattr_buckets(struct inode *inode,
if (i == 0)
num_buckets = le16_to_cpu(bucket_xh(bucket)->xh_num_buckets);
- mlog(0, "iterating xattr bucket %llu, first hash %u\n",
- (unsigned long long)blkno,
+ trace_ocfs2_iterate_xattr_bucket((unsigned long long)blkno,
le32_to_cpu(bucket_xh(bucket)->xh_entries[0].xe_name_hash));
if (func) {
ret = func(inode, bucket, para);
@@ -4173,9 +4175,9 @@ static void ocfs2_cp_xattr_block_to_bucket(struct inode *inode,
char *src = xb_bh->b_data;
char *target = bucket_block(bucket, blks - 1);
- mlog(0, "cp xattr from block %llu to bucket %llu\n",
- (unsigned long long)xb_bh->b_blocknr,
- (unsigned long long)bucket_blkno(bucket));
+ trace_ocfs2_cp_xattr_block_to_bucket_begin(
+ (unsigned long long)xb_bh->b_blocknr,
+ (unsigned long long)bucket_blkno(bucket));
for (i = 0; i < blks; i++)
memset(bucket_block(bucket, i), 0, blocksize);
@@ -4211,8 +4213,7 @@ static void ocfs2_cp_xattr_block_to_bucket(struct inode *inode,
for (i = 0; i < count; i++)
le16_add_cpu(&xh->xh_entries[i].xe_name_offset, off_change);
- mlog(0, "copy entry: start = %u, size = %u, offset_change = %u\n",
- offset, size, off_change);
+ trace_ocfs2_cp_xattr_block_to_bucket_end(offset, size, off_change);
sort(target + offset, count, sizeof(struct ocfs2_xattr_entry),
cmp_xe, swap_xe);
@@ -4261,8 +4262,8 @@ static int ocfs2_xattr_create_index_block(struct inode *inode,
struct ocfs2_xattr_tree_root *xr;
u16 xb_flags = le16_to_cpu(xb->xb_flags);
- mlog(0, "create xattr index block for %llu\n",
- (unsigned long long)xb_bh->b_blocknr);
+ trace_ocfs2_xattr_create_index_block_begin(
+ (unsigned long long)xb_bh->b_blocknr);
BUG_ON(xb_flags & OCFS2_XATTR_INDEXED);
BUG_ON(!xs->bucket);
@@ -4295,8 +4296,7 @@ static int ocfs2_xattr_create_index_block(struct inode *inode,
*/
blkno = ocfs2_clusters_to_blocks(inode->i_sb, bit_off);
- mlog(0, "allocate 1 cluster from %llu to xattr block\n",
- (unsigned long long)blkno);
+ trace_ocfs2_xattr_create_index_block((unsigned long long)blkno);
ret = ocfs2_init_xattr_bucket(xs->bucket, blkno);
if (ret) {
@@ -4400,8 +4400,7 @@ static int ocfs2_defrag_xattr_bucket(struct inode *inode,
entries = (char *)xh->xh_entries;
xh_free_start = le16_to_cpu(xh->xh_free_start);
- mlog(0, "adjust xattr bucket in %llu, count = %u, "
- "xh_free_start = %u, xh_name_value_len = %u.\n",
+ trace_ocfs2_defrag_xattr_bucket(
(unsigned long long)blkno, le16_to_cpu(xh->xh_count),
xh_free_start, le16_to_cpu(xh->xh_name_value_len));
@@ -4503,8 +4502,9 @@ static int ocfs2_mv_xattr_bucket_cross_cluster(struct inode *inode,
BUG_ON(le16_to_cpu(bucket_xh(first)->xh_num_buckets) < num_buckets);
BUG_ON(OCFS2_XATTR_BUCKET_SIZE == OCFS2_SB(sb)->s_clustersize);
- mlog(0, "move half of xattrs in cluster %llu to %llu\n",
- (unsigned long long)last_cluster_blkno, (unsigned long long)new_blkno);
+ trace_ocfs2_mv_xattr_bucket_cross_cluster(
+ (unsigned long long)last_cluster_blkno,
+ (unsigned long long)new_blkno);
ret = ocfs2_mv_xattr_buckets(inode, handle, bucket_blkno(first),
last_cluster_blkno, new_blkno,
@@ -4614,8 +4614,8 @@ static int ocfs2_divide_xattr_bucket(struct inode *inode,
struct ocfs2_xattr_entry *xe;
int blocksize = inode->i_sb->s_blocksize;
- mlog(0, "move some of xattrs from bucket %llu to %llu\n",
- (unsigned long long)blk, (unsigned long long)new_blk);
+ trace_ocfs2_divide_xattr_bucket_begin((unsigned long long)blk,
+ (unsigned long long)new_blk);
s_bucket = ocfs2_xattr_bucket_new(inode);
t_bucket = ocfs2_xattr_bucket_new(inode);
@@ -4714,9 +4714,9 @@ static int ocfs2_divide_xattr_bucket(struct inode *inode,
*/
xe = &xh->xh_entries[start];
len = sizeof(struct ocfs2_xattr_entry) * (count - start);
- mlog(0, "mv xattr entry len %d from %d to %d\n", len,
- (int)((char *)xe - (char *)xh),
- (int)((char *)xh->xh_entries - (char *)xh));
+ trace_ocfs2_divide_xattr_bucket_move(len,
+ (int)((char *)xe - (char *)xh),
+ (int)((char *)xh->xh_entries - (char *)xh));
memmove((char *)xh->xh_entries, (char *)xe, len);
xe = &xh->xh_entries[count - start];
len = sizeof(struct ocfs2_xattr_entry) * start;
@@ -4788,9 +4788,9 @@ static int ocfs2_cp_xattr_bucket(struct inode *inode,
BUG_ON(s_blkno == t_blkno);
- mlog(0, "cp bucket %llu to %llu, target is %d\n",
- (unsigned long long)s_blkno, (unsigned long long)t_blkno,
- t_is_new);
+ trace_ocfs2_cp_xattr_bucket((unsigned long long)s_blkno,
+ (unsigned long long)t_blkno,
+ t_is_new);
s_bucket = ocfs2_xattr_bucket_new(inode);
t_bucket = ocfs2_xattr_bucket_new(inode);
@@ -4862,8 +4862,8 @@ static int ocfs2_mv_xattr_buckets(struct inode *inode, handle_t *handle,
int num_buckets = ocfs2_xattr_buckets_per_cluster(osb);
struct ocfs2_xattr_bucket *old_first, *new_first;
- mlog(0, "mv xattrs from cluster %llu to %llu\n",
- (unsigned long long)last_blk, (unsigned long long)to_blk);
+ trace_ocfs2_mv_xattr_buckets((unsigned long long)last_blk,
+ (unsigned long long)to_blk);
BUG_ON(start_bucket >= num_buckets);
if (start_bucket) {
@@ -5013,9 +5013,9 @@ static int ocfs2_adjust_xattr_cross_cluster(struct inode *inode,
{
int ret;
- mlog(0, "adjust xattrs from cluster %llu len %u to %llu\n",
- (unsigned long long)bucket_blkno(first), prev_clusters,
- (unsigned long long)new_blk);
+ trace_ocfs2_adjust_xattr_cross_cluster(
+ (unsigned long long)bucket_blkno(first),
+ (unsigned long long)new_blk, prev_clusters);
if (ocfs2_xattr_buckets_per_cluster(OCFS2_SB(inode->i_sb)) > 1) {
ret = ocfs2_mv_xattr_bucket_cross_cluster(inode,
@@ -5088,10 +5088,10 @@ static int ocfs2_add_new_xattr_cluster(struct inode *inode,
struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
struct ocfs2_extent_tree et;
- mlog(0, "Add new xattr cluster for %llu, previous xattr hash = %u, "
- "previous xattr blkno = %llu\n",
- (unsigned long long)OCFS2_I(inode)->ip_blkno,
- prev_cpos, (unsigned long long)bucket_blkno(first));
+ trace_ocfs2_add_new_xattr_cluster_begin(
+ (unsigned long long)OCFS2_I(inode)->ip_blkno,
+ (unsigned long long)bucket_blkno(first),
+ prev_cpos, prev_clusters);
ocfs2_init_xattr_tree_extent_tree(&et, INODE_CACHE(inode), root_bh);
@@ -5113,8 +5113,7 @@ static int ocfs2_add_new_xattr_cluster(struct inode *inode,
BUG_ON(num_bits > clusters_to_add);
block = ocfs2_clusters_to_blocks(osb->sb, bit_off);
- mlog(0, "Allocating %u clusters at block %u for xattr in inode %llu\n",
- num_bits, bit_off, (unsigned long long)OCFS2_I(inode)->ip_blkno);
+ trace_ocfs2_add_new_xattr_cluster((unsigned long long)block, num_bits);
if (bucket_blkno(first) + (prev_clusters * bpc) == block &&
(prev_clusters + num_bits) << osb->s_clustersize_bits <=
@@ -5130,8 +5129,6 @@ static int ocfs2_add_new_xattr_cluster(struct inode *inode,
*/
v_start = prev_cpos + prev_clusters;
*num_clusters = prev_clusters + num_bits;
- mlog(0, "Add contiguous %u clusters to previous extent rec.\n",
- num_bits);
} else {
ret = ocfs2_adjust_xattr_cross_cluster(inode,
handle,
@@ -5147,8 +5144,8 @@ static int ocfs2_add_new_xattr_cluster(struct inode *inode,
}
}
- mlog(0, "Insert %u clusters at block %llu for xattr@%u\n",
- num_bits, (unsigned long long)block, v_start);
+ trace_ocfs2_add_new_xattr_cluster_insert((unsigned long long)block,
+ v_start, num_bits);
ret = ocfs2_insert_extent(handle, &et, v_start, block,
num_bits, 0, ctxt->meta_ac);
if (ret < 0) {
@@ -5183,9 +5180,9 @@ static int ocfs2_extend_xattr_bucket(struct inode *inode,
u64 end_blk;
u16 new_bucket = le16_to_cpu(bucket_xh(first)->xh_num_buckets);
- mlog(0, "extend xattr bucket in %llu, xattr extend rec starting "
- "from %llu, len = %u\n", (unsigned long long)target_blk,
- (unsigned long long)bucket_blkno(first), num_clusters);
+ trace_ocfs2_extend_xattr_bucket((unsigned long long)target_blk,
+ (unsigned long long)bucket_blkno(first),
+ num_clusters, new_bucket);
/* The extent must have room for an additional bucket */
BUG_ON(new_bucket >=
@@ -5265,8 +5262,8 @@ static int ocfs2_add_new_xattr_bucket(struct inode *inode,
/* The bucket at the front of the extent */
struct ocfs2_xattr_bucket *first;
- mlog(0, "Add new xattr bucket starting from %llu\n",
- (unsigned long long)bucket_blkno(target));
+ trace_ocfs2_add_new_xattr_bucket(
+ (unsigned long long)bucket_blkno(target));
/* The first bucket of the original extent */
first = ocfs2_xattr_bucket_new(inode);
@@ -5382,8 +5379,8 @@ static int ocfs2_xattr_bucket_value_truncate(struct inode *inode,
* modified something. We have to assume they did, and dirty
* the whole bucket. This leaves us in a consistent state.
*/
- mlog(0, "truncate %u in xattr bucket %llu to %d bytes.\n",
- xe_off, (unsigned long long)bucket_blkno(bucket), len);
+ trace_ocfs2_xattr_bucket_value_truncate(
+ (unsigned long long)bucket_blkno(bucket), xe_off, len);
ret = ocfs2_xattr_value_truncate(inode, &vb, len, ctxt);
if (ret) {
mlog_errno(ret);
@@ -5433,8 +5430,9 @@ static int ocfs2_rm_xattr_cluster(struct inode *inode,
ocfs2_init_dealloc_ctxt(&dealloc);
- mlog(0, "rm xattr extent rec at %u len = %u, start from %llu\n",
- cpos, len, (unsigned long long)blkno);
+ trace_ocfs2_rm_xattr_cluster(
+ (unsigned long long)OCFS2_I(inode)->ip_blkno,
+ (unsigned long long)blkno, cpos, len);
ocfs2_remove_xattr_clusters_from_cache(INODE_CACHE(inode), blkno,
len);
@@ -5538,7 +5536,7 @@ static int ocfs2_xattr_set_entry_bucket(struct inode *inode,
int ret;
struct ocfs2_xa_loc loc;
- mlog(0, "Set xattr %s in xattr bucket\n", xi->xi_name);
+ trace_ocfs2_xattr_set_entry_bucket(xi->xi_name);
ocfs2_init_xattr_bucket_xa_loc(&loc, xs->bucket,
xs->not_found ? NULL : xs->here);
@@ -5580,7 +5578,7 @@ static int ocfs2_xattr_set_entry_index_block(struct inode *inode,
{
int ret;
- mlog(0, "Set xattr %s in xattr index block\n", xi->xi_name);
+ trace_ocfs2_xattr_set_entry_index_block(xi->xi_name);
ret = ocfs2_xattr_set_entry_bucket(inode, xi, xs, ctxt);
if (!ret)
@@ -6039,9 +6037,9 @@ static int ocfs2_xattr_bucket_value_refcount(struct inode *inode,
if (ocfs2_meta_ecc(OCFS2_SB(inode->i_sb)))
p = &refcount;
- mlog(0, "refcount bucket %llu, count = %u\n",
- (unsigned long long)bucket_blkno(bucket),
- le16_to_cpu(xh->xh_count));
+ trace_ocfs2_xattr_bucket_value_refcount(
+ (unsigned long long)bucket_blkno(bucket),
+ le16_to_cpu(xh->xh_count));
for (i = 0; i < le16_to_cpu(xh->xh_count); i++) {
xe = &xh->xh_entries[i];
@@ -6337,8 +6335,8 @@ static int ocfs2_reflink_xattr_header(handle_t *handle,
u32 clusters, cpos, p_cluster, num_clusters;
unsigned int ext_flags = 0;
- mlog(0, "reflink xattr in container %llu, count = %u\n",
- (unsigned long long)old_bh->b_blocknr, le16_to_cpu(xh->xh_count));
+ trace_ocfs2_reflink_xattr_header((unsigned long long)old_bh->b_blocknr,
+ le16_to_cpu(xh->xh_count));
last = &new_xh->xh_entries[le16_to_cpu(new_xh->xh_count)];
for (i = 0, j = 0; i < le16_to_cpu(xh->xh_count); i++, j++) {
@@ -6538,8 +6536,8 @@ static int ocfs2_create_empty_xattr_block(struct inode *inode,
goto out;
}
- mlog(0, "create new xattr block for inode %llu, index = %d\n",
- (unsigned long long)fe_bh->b_blocknr, indexed);
+ trace_ocfs2_create_empty_xattr_block(
+ (unsigned long long)fe_bh->b_blocknr, indexed);
ret = ocfs2_create_xattr_block(inode, fe_bh, &ctxt, indexed,
ret_bh);
if (ret)
@@ -6950,8 +6948,8 @@ static int ocfs2_reflink_xattr_buckets(handle_t *handle,
if (ret)
mlog_errno(ret);
- mlog(0, "insert new xattr extent rec start %llu len %u to %u\n",
- (unsigned long long)new_blkno, num_clusters, reflink_cpos);
+ trace_ocfs2_reflink_xattr_buckets((unsigned long long)new_blkno,
+ num_clusters, reflink_cpos);
len -= num_clusters;
blkno += ocfs2_clusters_to_blocks(inode->i_sb, num_clusters);
@@ -6980,8 +6978,7 @@ static int ocfs2_reflink_xattr_rec(struct inode *inode,
struct ocfs2_alloc_context *data_ac = NULL;
struct ocfs2_extent_tree et;
- mlog(0, "reflink xattr buckets %llu len %u\n",
- (unsigned long long)blkno, len);
+ trace_ocfs2_reflink_xattr_rec((unsigned long long)blkno, len);
ocfs2_init_xattr_tree_extent_tree(&et,
INODE_CACHE(args->reflink->new_inode),
--
1.7.0.4
^ permalink raw reply related [flat|nested] 53+ messages in thread
* [Ocfs2-devel] [PATCH 22/34] ocfs2: Remove masklog ML_RESERVATIONS.
2010-12-23 7:19 [Ocfs2-devel] [PATCH 00/34] OCFS2: Add trace event and replace mlog(0) Tao Ma
` (20 preceding siblings ...)
2010-12-23 7:31 ` [Ocfs2-devel] [PATCH 21/34] ocfs2: Remove masklog ML_XATTR Tao Ma
@ 2010-12-23 7:31 ` Tao Ma
2010-12-23 7:31 ` [Ocfs2-devel] [PATCH 23/34] ocfs2: Remove mlog(0) from quota_local.c Tao Ma
` (16 subsequent siblings)
38 siblings, 0 replies; 53+ messages in thread
From: Tao Ma @ 2010-12-23 7:31 UTC (permalink / raw)
To: ocfs2-devel
From: Tao Ma <boyu.mt@taobao.com>
Remove mlog(0) from fs/ocfs2/reservations.c and the masklog RESERVATIONS.
Signed-off-by: Tao Ma <boyu.mt@taobao.com>
---
fs/ocfs2/cluster/masklog.c | 1 -
fs/ocfs2/cluster/masklog.h | 1 -
fs/ocfs2/ocfs2_trace.h | 170 ++++++++++++++++++++++++++++++++++++++++++++
fs/ocfs2/reservations.c | 57 +++++++--------
4 files changed, 196 insertions(+), 33 deletions(-)
diff --git a/fs/ocfs2/cluster/masklog.c b/fs/ocfs2/cluster/masklog.c
index d31c5f1..64bf0b2 100644
--- a/fs/ocfs2/cluster/masklog.c
+++ b/fs/ocfs2/cluster/masklog.c
@@ -107,7 +107,6 @@ static struct mlog_attribute mlog_attrs[MLOG_MAX_BITS] = {
define_mask(ERROR),
define_mask(NOTICE),
define_mask(KTHREAD),
- define_mask(RESERVATIONS),
};
static struct attribute *mlog_attr_ptrs[MLOG_MAX_BITS] = {NULL, };
diff --git a/fs/ocfs2/cluster/masklog.h b/fs/ocfs2/cluster/masklog.h
index 736d18c..57e12a8 100644
--- a/fs/ocfs2/cluster/masklog.h
+++ b/fs/ocfs2/cluster/masklog.h
@@ -110,7 +110,6 @@
#define ML_ERROR 0x0000000100000000ULL /* sent to KERN_ERR */
#define ML_NOTICE 0x0000000200000000ULL /* setn to KERN_NOTICE */
#define ML_KTHREAD 0x0000000400000000ULL /* kernel thread activity */
-#define ML_RESERVATIONS 0x0000000800000000ULL /* ocfs2 alloc reservations */
#define ML_CLUSTER 0x0000001000000000ULL /* cluster stack */
#define MLOG_INITIAL_AND_MASK (ML_ERROR|ML_NOTICE)
diff --git a/fs/ocfs2/ocfs2_trace.h b/fs/ocfs2/ocfs2_trace.h
index 676e75d..b0bb9a6 100644
--- a/fs/ocfs2/ocfs2_trace.h
+++ b/fs/ocfs2/ocfs2_trace.h
@@ -2806,6 +2806,176 @@ DEFINE_EVENT(ocfs2__ull_uint, ocfs2_reflink_xattr_rec,
TP_ARGS(blkno, len)
);
/* End of trace events for fs/ocfs2/xattr.c. */
+
+/* Trace events for fs/ocfs2/reservations.c. */
+DEFINE_EVENT(ocfs2__uint_uint, ocfs2_resv_insert,
+
+ TP_PROTO(unsigned int r_start, unsigned int r_len),
+
+ TP_ARGS(r_start, r_len)
+);
+
+DEFINE_EVENT(ocfs2__ull_uint_uint_uint, ocfs2_resmap_find_free_bits_begin,
+
+ TP_PROTO(unsigned long long search_start, unsigned int search_len,
+ unsigned int wanted, unsigned int len),
+
+ TP_ARGS(search_start, search_len, wanted, len)
+);
+
+DEFINE_EVENT(ocfs2__uint_uint, ocfs2_resmap_find_free_bits_end,
+
+ TP_PROTO(unsigned int best_start, unsigned int best_len),
+
+ TP_ARGS(best_start, best_len)
+);
+
+TRACE_EVENT(ocfs2_resv_find_window_begin,
+
+ TP_PROTO(unsigned int r_start, unsigned int r_end, unsigned int goal,
+ unsigned int wanted, int empty_root),
+
+ TP_ARGS(r_start, r_end, goal, wanted, empty_root),
+
+ TP_STRUCT__entry(
+ __field( unsigned int, r_start )
+ __field( unsigned int, r_end )
+ __field( unsigned int, goal )
+ __field( unsigned int, wanted )
+ __field( int, empty_root )
+ ),
+
+ TP_fast_assign(
+ __entry->r_start = r_start;
+ __entry->r_end = r_end;
+ __entry->goal = goal;
+ __entry->wanted = wanted;
+ __entry->empty_root = empty_root;
+ ),
+
+ TP_printk("%u %u %u %u %d", __entry->r_start, __entry->r_end,
+ __entry->goal, __entry->wanted, __entry->empty_root)
+);
+
+DEFINE_EVENT(ocfs2__uint_uint, ocfs2_resv_find_window_prev,
+
+ TP_PROTO(unsigned int r_start, unsigned int r_end),
+
+ TP_ARGS(r_start, r_end)
+);
+
+DEFINE_EVENT(ocfs2__int_int, ocfs2_resv_find_window_next,
+
+ TP_PROTO(int next_start, int next_end),
+
+ TP_ARGS(next_start, next_end)
+);
+
+DEFINE_EVENT(ocfs2__uint_uint_uint, ocfs2_cannibalize_resv_begin,
+
+ TP_PROTO(unsigned int start, unsigned int len, unsigned int end),
+
+ TP_ARGS(start, len, end)
+);
+
+TRACE_EVENT(ocfs2_cannibalize_resv_end,
+
+ TP_PROTO(unsigned int start, unsigned int end, unsigned int len,
+ unsigned int last_start, unsigned int last_len),
+
+ TP_ARGS(start, end, len, last_start, last_len),
+
+ TP_STRUCT__entry(
+ __field( unsigned int, start )
+ __field( unsigned int, end )
+ __field( unsigned int, len )
+ __field( unsigned int, last_start )
+ __field( unsigned int, last_len )
+ ),
+
+ TP_fast_assign(
+ __entry->start = start;
+ __entry->end = end;
+ __entry->len = len;
+ __entry->last_start = last_start;
+ __entry->last_len = last_len;
+ ),
+
+ TP_printk("%u %u %u %u %u", __entry->start, __entry->end,
+ __entry->len, __entry->last_start, __entry->last_len)
+);
+
+DEFINE_EVENT(ocfs2__uint_uint, ocfs2_resmap_resv_bits,
+
+ TP_PROTO(unsigned int start, unsigned int len),
+
+ TP_ARGS(start, len)
+);
+
+TRACE_EVENT(ocfs2_resmap_claimed_bits_begin,
+
+ TP_PROTO(unsigned int cstart, unsigned int cend, unsigned int clen,
+ unsigned int r_start, unsigned int r_end, unsigned int r_len,
+ unsigned int last_start, unsigned int last_len),
+
+ TP_ARGS(cstart, cend, clen, r_start, r_end,
+ r_len, last_start, last_len),
+
+ TP_STRUCT__entry(
+ __field( unsigned int, cstart )
+ __field( unsigned int, cend )
+ __field( unsigned int, clen )
+ __field( unsigned int, r_start )
+ __field( unsigned int, r_end )
+ __field( unsigned int, r_len )
+ __field( unsigned int, last_start )
+ __field( unsigned int, last_len )
+ ),
+
+ TP_fast_assign(
+ __entry->cstart = cstart;
+ __entry->cend = cend;
+ __entry->clen = clen;
+ __entry->r_start = r_start;
+ __entry->r_end = r_end;
+ __entry->r_len = r_len;
+ __entry->last_start = last_start;
+ __entry->last_len = last_len;
+ ),
+
+ TP_printk("%u %u %u %u %u %u %u %u",
+ __entry->cstart, __entry->cend, __entry->clen,
+ __entry->r_start, __entry->r_end, __entry->r_len,
+ __entry->last_start, __entry->last_len)
+);
+
+TRACE_EVENT(ocfs2_resmap_claimed_bits_end,
+
+ TP_PROTO(unsigned int start, unsigned int end, unsigned int len,
+ unsigned int last_start, unsigned int last_len),
+
+ TP_ARGS(start, end, len, last_start, last_len),
+
+ TP_STRUCT__entry(
+ __field( unsigned int, start )
+ __field( unsigned int, end )
+ __field( unsigned int, len )
+ __field( unsigned int, last_start )
+ __field( unsigned int, last_len )
+ ),
+
+ TP_fast_assign(
+ __entry->start = start;
+ __entry->end = end;
+ __entry->len = len;
+ __entry->last_start = last_start;
+ __entry->last_len = last_len;
+ ),
+
+ TP_printk("%u %u %u %u %u", __entry->start, __entry->end,
+ __entry->len, __entry->last_start, __entry->last_len)
+);
+/* End of trace events for fs/ocfs2/reservations.c. */
#endif /* _TRACE_OCFS2_H */
/* This part must be outside protection */
diff --git a/fs/ocfs2/reservations.c b/fs/ocfs2/reservations.c
index 3e78db3..41ffd36 100644
--- a/fs/ocfs2/reservations.c
+++ b/fs/ocfs2/reservations.c
@@ -30,10 +30,10 @@
#include <linux/bitops.h>
#include <linux/list.h>
-#define MLOG_MASK_PREFIX ML_RESERVATIONS
#include <cluster/masklog.h>
#include "ocfs2.h"
+#include "ocfs2_trace.h"
#ifdef CONFIG_OCFS2_DEBUG_FS
#define OCFS2_CHECK_RESERVATIONS
@@ -321,8 +321,7 @@ static void ocfs2_resv_insert(struct ocfs2_reservation_map *resmap,
assert_spin_locked(&resv_lock);
- mlog(0, "Insert reservation start: %u len: %u\n", new->r_start,
- new->r_len);
+ trace_ocfs2_resv_insert(new->r_start, new->r_len);
while (*p) {
parent = *p;
@@ -423,8 +422,8 @@ static int ocfs2_resmap_find_free_bits(struct ocfs2_reservation_map *resmap,
unsigned int best_start, best_len = 0;
int offset, start, found;
- mlog(0, "Find %u bits within range (%u, len %u) resmap len: %u\n",
- wanted, search_start, search_len, resmap->m_bitmap_len);
+ trace_ocfs2_resmap_find_free_bits_begin(search_start, search_len,
+ wanted, resmap->m_bitmap_len);
found = best_start = best_len = 0;
@@ -463,7 +462,7 @@ static int ocfs2_resmap_find_free_bits(struct ocfs2_reservation_map *resmap,
*rlen = best_len;
*rstart = best_start;
- mlog(0, "Found start: %u len: %u\n", best_start, best_len);
+ trace_ocfs2_resmap_find_free_bits_end(best_start, best_len);
return *rlen;
}
@@ -487,9 +486,8 @@ static void __ocfs2_resv_find_window(struct ocfs2_reservation_map *resmap,
* - our window should be last in all reservations
* - need to make sure we don't go past end of bitmap
*/
-
- mlog(0, "resv start: %u resv end: %u goal: %u wanted: %u\n",
- resv->r_start, ocfs2_resv_end(resv), goal, wanted);
+ trace_ocfs2_resv_find_window_begin(resv->r_start, ocfs2_resv_end(resv),
+ goal, wanted, RB_EMPTY_ROOT(root));
assert_spin_locked(&resv_lock);
@@ -498,9 +496,6 @@ static void __ocfs2_resv_find_window(struct ocfs2_reservation_map *resmap,
* Easiest case - empty tree. We can just take
* whatever window of free bits we want.
*/
-
- mlog(0, "Empty root\n");
-
clen = ocfs2_resmap_find_free_bits(resmap, wanted, goal,
resmap->m_bitmap_len - goal,
&cstart, &clen);
@@ -524,8 +519,6 @@ static void __ocfs2_resv_find_window(struct ocfs2_reservation_map *resmap,
prev_resv = ocfs2_find_resv_lhs(resmap, goal);
if (prev_resv == NULL) {
- mlog(0, "Goal on LHS of leftmost window\n");
-
/*
* A NULL here means that the search code couldn't
* find a window that starts before goal.
@@ -570,13 +563,15 @@ static void __ocfs2_resv_find_window(struct ocfs2_reservation_map *resmap,
next_resv = NULL;
}
+ trace_ocfs2_resv_find_window_prev(prev_resv->r_start,
+ ocfs2_resv_end(prev_resv));
+
prev = &prev_resv->r_node;
/* Now we do a linear search for a window, starting at 'prev_rsv' */
while (1) {
next = rb_next(prev);
if (next) {
- mlog(0, "One more resv found in linear search\n");
next_resv = rb_entry(next,
struct ocfs2_alloc_reservation,
r_node);
@@ -585,7 +580,6 @@ static void __ocfs2_resv_find_window(struct ocfs2_reservation_map *resmap,
gap_end = next_resv->r_start - 1;
gap_len = gap_end - gap_start + 1;
} else {
- mlog(0, "No next node\n");
/*
* We're at the rightmost edge of the
* tree. See if a reservation between this
@@ -596,6 +590,8 @@ static void __ocfs2_resv_find_window(struct ocfs2_reservation_map *resmap,
gap_end = resmap->m_bitmap_len - 1;
}
+ trace_ocfs2_resv_find_window_next(next ? next_resv->r_start: -1,
+ next ? ocfs2_resv_end(next_resv) : -1);
/*
* No need to check this gap if we have already found
* a larger region of free bits.
@@ -654,8 +650,9 @@ static void ocfs2_cannibalize_resv(struct ocfs2_reservation_map *resmap,
lru_resv = list_first_entry(&resmap->m_lru,
struct ocfs2_alloc_reservation, r_lru);
- mlog(0, "lru resv: start: %u len: %u end: %u\n", lru_resv->r_start,
- lru_resv->r_len, ocfs2_resv_end(lru_resv));
+ trace_ocfs2_cannibalize_resv_begin(lru_resv->r_start,
+ lru_resv->r_len,
+ ocfs2_resv_end(lru_resv));
/*
* Cannibalize (some or all) of the target reservation and
@@ -684,10 +681,9 @@ static void ocfs2_cannibalize_resv(struct ocfs2_reservation_map *resmap,
resv->r_len = shrink;
}
- mlog(0, "Reservation now looks like: r_start: %u r_end: %u "
- "r_len: %u r_last_start: %u r_last_len: %u\n",
- resv->r_start, ocfs2_resv_end(resv), resv->r_len,
- resv->r_last_start, resv->r_last_len);
+ trace_ocfs2_cannibalize_resv_end(resv->r_start, ocfs2_resv_end(resv),
+ resv->r_len, resv->r_last_start,
+ resv->r_last_len);
ocfs2_resv_insert(resmap, resv);
}
@@ -748,7 +744,6 @@ int ocfs2_resmap_resv_bits(struct ocfs2_reservation_map *resmap,
if ((resv->r_flags & OCFS2_RESV_FLAG_TMP) || wanted < *clen)
wanted = *clen;
- mlog(0, "empty reservation, find new window\n");
/*
* Try to get a window here. If it works, we must fall
* through and test the bitmap . This avoids some
@@ -757,6 +752,7 @@ int ocfs2_resmap_resv_bits(struct ocfs2_reservation_map *resmap,
* that inode.
*/
ocfs2_resv_find_window(resmap, resv, wanted);
+ trace_ocfs2_resmap_resv_bits(resv->r_start, resv->r_len);
}
BUG_ON(ocfs2_resv_empty(resv));
@@ -813,10 +809,10 @@ void ocfs2_resmap_claimed_bits(struct ocfs2_reservation_map *resmap,
spin_lock(&resv_lock);
- mlog(0, "claim bits: cstart: %u cend: %u clen: %u r_start: %u "
- "r_end: %u r_len: %u, r_last_start: %u r_last_len: %u\n",
- cstart, cend, clen, resv->r_start, ocfs2_resv_end(resv),
- resv->r_len, resv->r_last_start, resv->r_last_len);
+ trace_ocfs2_resmap_claimed_bits_begin(cstart, cend, clen, resv->r_start,
+ ocfs2_resv_end(resv), resv->r_len,
+ resv->r_last_start,
+ resv->r_last_len);
BUG_ON(cstart < resv->r_start);
BUG_ON(cstart > ocfs2_resv_end(resv));
@@ -833,10 +829,9 @@ void ocfs2_resmap_claimed_bits(struct ocfs2_reservation_map *resmap,
if (!ocfs2_resv_empty(resv))
ocfs2_resv_mark_lru(resmap, resv);
- mlog(0, "Reservation now looks like: r_start: %u r_end: %u "
- "r_len: %u r_last_start: %u r_last_len: %u\n",
- resv->r_start, ocfs2_resv_end(resv), resv->r_len,
- resv->r_last_start, resv->r_last_len);
+ trace_ocfs2_resmap_claimed_bits_end(resv->r_start, ocfs2_resv_end(resv),
+ resv->r_len, resv->r_last_start,
+ resv->r_last_len);
ocfs2_check_resmap(resmap);
--
1.7.0.4
^ permalink raw reply related [flat|nested] 53+ messages in thread
* [Ocfs2-devel] [PATCH 23/34] ocfs2: Remove mlog(0) from quota_local.c.
2010-12-23 7:19 [Ocfs2-devel] [PATCH 00/34] OCFS2: Add trace event and replace mlog(0) Tao Ma
` (21 preceding siblings ...)
2010-12-23 7:31 ` [Ocfs2-devel] [PATCH 22/34] ocfs2: Remove masklog ML_RESERVATIONS Tao Ma
@ 2010-12-23 7:31 ` Tao Ma
2010-12-23 7:31 ` [Ocfs2-devel] [PATCH 24/34] ocfs2: Remove masklog ML_QUOTA Tao Ma
` (15 subsequent siblings)
38 siblings, 0 replies; 53+ messages in thread
From: Tao Ma @ 2010-12-23 7:31 UTC (permalink / raw)
To: ocfs2-devel
From: Tao Ma <boyu.mt@taobao.com>
Remove mlog(0) from fs/ocfs2/quota_local.c.
Signed-off-by: Tao Ma <boyu.mt@taobao.com>
---
fs/ocfs2/ocfs2_trace.h | 24 ++++++++++++++++++++++++
fs/ocfs2/quota_local.c | 13 +++++++------
2 files changed, 31 insertions(+), 6 deletions(-)
diff --git a/fs/ocfs2/ocfs2_trace.h b/fs/ocfs2/ocfs2_trace.h
index b0bb9a6..fef6478 100644
--- a/fs/ocfs2/ocfs2_trace.h
+++ b/fs/ocfs2/ocfs2_trace.h
@@ -2976,6 +2976,30 @@ TRACE_EVENT(ocfs2_resmap_claimed_bits_end,
__entry->len, __entry->last_start, __entry->last_len)
);
/* End of trace events for fs/ocfs2/reservations.c. */
+
+/* Trace events for fs/ocfs2/quota_local.c. */
+DEFINE_EVENT(ocfs2__ull_uint, ocfs2_recover_local_quota_file,
+
+ TP_PROTO(unsigned long long ino, unsigned int type),
+
+ TP_ARGS(ino, type)
+);
+
+DEFINE_EVENT(ocfs2__int, ocfs2_finish_quota_recovery,
+
+ TP_PROTO(int slot_num),
+
+ TP_ARGS(slot_num)
+);
+
+DEFINE_EVENT(ocfs2__ull_ull_uint, olq_set_dquot,
+
+ TP_PROTO(unsigned long long space,
+ unsigned long long inode, unsigned int id),
+
+ TP_ARGS(space, inode, id)
+);
+/* End of trace events for fs/ocfs2/quota_local.c. */
#endif /* _TRACE_OCFS2_H */
/* This part must be outside protection */
diff --git a/fs/ocfs2/quota_local.c b/fs/ocfs2/quota_local.c
index 88ec451..dc8007f 100644
--- a/fs/ocfs2/quota_local.c
+++ b/fs/ocfs2/quota_local.c
@@ -8,7 +8,6 @@
#include <linux/quotaops.h>
#include <linux/module.h>
-#define MLOG_MASK_PREFIX ML_QUOTA
#include <cluster/masklog.h>
#include "ocfs2_fs.h"
@@ -23,6 +22,7 @@
#include "quota.h"
#include "uptodate.h"
#include "super.h"
+#include "ocfs2_trace.h"
/* Number of local quota structures per block */
static inline unsigned int ol_quota_entries_per_block(struct super_block *sb)
@@ -475,7 +475,7 @@ static int ocfs2_recover_local_quota_file(struct inode *lqinode,
struct ocfs2_recovery_chunk *rchunk, *next;
qsize_t spacechange, inodechange;
- mlog(0, "ino=%lu type=%u", (unsigned long)lqinode->i_ino, type);
+ trace_ocfs2_recover_local_quota_file((unsigned long)lqinode->i_ino, type);
list_for_each_entry_safe(rchunk, next, &(rec->r_list[type]), rc_list) {
chunk = rchunk->rc_chunk;
@@ -601,7 +601,7 @@ int ocfs2_finish_quota_recovery(struct ocfs2_super *osb,
for (type = 0; type < MAXQUOTAS; type++) {
if (list_empty(&(rec->r_list[type])))
continue;
- mlog(0, "Recovering quota in slot %d\n", slot_num);
+ trace_ocfs2_finish_quota_recovery(slot_num);
lqinode = ocfs2_get_system_file_inode(osb, ino[type], slot_num);
if (!lqinode) {
status = -ENOENT;
@@ -883,9 +883,10 @@ static void olq_set_dquot(struct buffer_head *bh, void *private)
dqblk->dqb_inodemod = cpu_to_le64(od->dq_dquot.dq_dqb.dqb_curinodes -
od->dq_originodes);
spin_unlock(&dq_data_lock);
- mlog(0, "Writing local dquot %u space %lld inodes %lld\n",
- od->dq_dquot.dq_id, (long long)le64_to_cpu(dqblk->dqb_spacemod),
- (long long)le64_to_cpu(dqblk->dqb_inodemod));
+ trace_olq_set_dquot(
+ (unsigned long long)le64_to_cpu(dqblk->dqb_spacemod),
+ (unsigned long long)le64_to_cpu(dqblk->dqb_inodemod),
+ od->dq_dquot.dq_id);
}
/* Write dquot to local quota file */
--
1.7.0.4
^ permalink raw reply related [flat|nested] 53+ messages in thread
* [Ocfs2-devel] [PATCH 24/34] ocfs2: Remove masklog ML_QUOTA.
2010-12-23 7:19 [Ocfs2-devel] [PATCH 00/34] OCFS2: Add trace event and replace mlog(0) Tao Ma
` (22 preceding siblings ...)
2010-12-23 7:31 ` [Ocfs2-devel] [PATCH 23/34] ocfs2: Remove mlog(0) from quota_local.c Tao Ma
@ 2010-12-23 7:31 ` Tao Ma
2010-12-23 7:31 ` [Ocfs2-devel] [PATCH 25/34] ocfs2: remove NAMEI from symlink.c Tao Ma
` (14 subsequent siblings)
38 siblings, 0 replies; 53+ messages in thread
From: Tao Ma @ 2010-12-23 7:31 UTC (permalink / raw)
To: ocfs2-devel
From: Tao Ma <boyu.mt@taobao.com>
Remove mlog(0) from fs/ocfs2/quota_global.c and
the masklog QUOTA.
Signed-off-by: Tao Ma <boyu.mt@taobao.com>
---
fs/ocfs2/cluster/masklog.c | 1 -
fs/ocfs2/cluster/masklog.h | 1 -
fs/ocfs2/ocfs2_trace.h | 110 ++++++++++++++++++++++++++++++++++++++++++++
fs/ocfs2/quota_global.c | 24 +++++-----
4 files changed, 122 insertions(+), 14 deletions(-)
diff --git a/fs/ocfs2/cluster/masklog.c b/fs/ocfs2/cluster/masklog.c
index 64bf0b2..44d22cd 100644
--- a/fs/ocfs2/cluster/masklog.c
+++ b/fs/ocfs2/cluster/masklog.c
@@ -102,7 +102,6 @@ static struct mlog_attribute mlog_attrs[MLOG_MAX_BITS] = {
define_mask(CONN),
define_mask(QUORUM),
define_mask(EXPORT),
- define_mask(QUOTA),
define_mask(BASTS),
define_mask(ERROR),
define_mask(NOTICE),
diff --git a/fs/ocfs2/cluster/masklog.h b/fs/ocfs2/cluster/masklog.h
index 57e12a8..789e735 100644
--- a/fs/ocfs2/cluster/masklog.h
+++ b/fs/ocfs2/cluster/masklog.h
@@ -104,7 +104,6 @@
#define ML_CONN 0x0000000004000000ULL /* net connection management */
#define ML_QUORUM 0x0000000008000000ULL /* net connection quorum */
#define ML_EXPORT 0x0000000010000000ULL /* ocfs2 export operations */
-#define ML_QUOTA 0x0000000040000000ULL /* ocfs2 quota operations */
#define ML_BASTS 0x0000001000000000ULL /* dlmglue asts and basts */
/* bits that are infrequently given and frequently matched in the high word */
#define ML_ERROR 0x0000000100000000ULL /* sent to KERN_ERR */
diff --git a/fs/ocfs2/ocfs2_trace.h b/fs/ocfs2/ocfs2_trace.h
index fef6478..f87a203 100644
--- a/fs/ocfs2/ocfs2_trace.h
+++ b/fs/ocfs2/ocfs2_trace.h
@@ -110,6 +110,25 @@ DECLARE_EVENT_CLASS(ocfs2__int_int,
TP_printk("%d %d", __entry->value1, __entry->value2)
);
+DECLARE_EVENT_CLASS(ocfs2__uint_int,
+
+ TP_PROTO(unsigned int value1, int value2),
+
+ TP_ARGS(value1, value2),
+
+ TP_STRUCT__entry(
+ __field( unsigned int, value1 )
+ __field( int, value2 )
+ ),
+
+ TP_fast_assign(
+ __entry->value1 = value1;
+ __entry->value2 = value2;
+ ),
+
+ TP_printk("%u %d", __entry->value1, __entry->value2)
+);
+
DECLARE_EVENT_CLASS(ocfs2__uint_uint,
TP_PROTO(unsigned int value1, unsigned int value2),
@@ -3000,6 +3019,97 @@ DEFINE_EVENT(ocfs2__ull_ull_uint, olq_set_dquot,
TP_ARGS(space, inode, id)
);
/* End of trace events for fs/ocfs2/quota_local.c. */
+
+/* Trace events for fs/ocfs2/quota_global.c. */
+DEFINE_EVENT(ocfs2__ull, ocfs2_validate_quota_block,
+
+ TP_PROTO(unsigned long long blkno),
+
+ TP_ARGS(blkno)
+);
+
+TRACE_EVENT(ocfs2_sync_dquot,
+
+ TP_PROTO(unsigned int dq_id, long long dqb_curspace,
+ long long spacechange, long long curinodes,
+ long long inodechange),
+
+ TP_ARGS(dq_id, dqb_curspace, spacechange, curinodes, inodechange),
+
+ TP_STRUCT__entry(
+ __field( unsigned int, dq_id )
+ __field( long long, dqb_curspace )
+ __field( long long, spacechange )
+ __field( long long, curinodes )
+ __field( long long, inodechange )
+ ),
+
+ TP_fast_assign(
+ __entry->dq_id = dq_id;
+ __entry->dqb_curspace = dqb_curspace;
+ __entry->spacechange = spacechange;
+ __entry->curinodes = curinodes;
+ __entry->inodechange = inodechange;
+ ),
+
+ TP_printk("%u %lld %lld %lld %lld", __entry->dq_id,
+ __entry->dqb_curspace, __entry->spacechange,
+ __entry->curinodes, __entry->inodechange)
+);
+
+TRACE_EVENT(ocfs2_sync_dquot_helper,
+
+ TP_PROTO(unsigned int dq_id, unsigned int dq_type, unsigned long type,
+ const char *s_id),
+
+ TP_ARGS(dq_id, dq_type, type, s_id),
+
+ TP_STRUCT__entry(
+ __field( unsigned int, dq_id )
+ __field( unsigned int, dq_type )
+ __field( unsigned long, type )
+ __field( const char *, s_id )
+ ),
+
+ TP_fast_assign(
+ __entry->dq_id = dq_id;
+ __entry->dq_type = dq_type;
+ __entry->type = type;
+ __entry->s_id = s_id;
+ ),
+
+ TP_printk("%u %u %lu %s", __entry->dq_id, __entry->dq_type,
+ __entry->type, __entry->s_id)
+);
+
+DEFINE_EVENT(ocfs2__uint_int, ocfs2_write_dquot,
+
+ TP_PROTO(unsigned int id, int type),
+
+ TP_ARGS(id, type)
+);
+
+DEFINE_EVENT(ocfs2__uint_int, ocfs2_release_dquot,
+
+ TP_PROTO(unsigned int id, int type),
+
+ TP_ARGS(id, type)
+);
+
+DEFINE_EVENT(ocfs2__uint_int, ocfs2_acquire_dquot,
+
+ TP_PROTO(unsigned int id, int type),
+
+ TP_ARGS(id, type)
+);
+
+DEFINE_EVENT(ocfs2__uint_int, ocfs2_mark_dquot_dirty,
+
+ TP_PROTO(unsigned int id, int type),
+
+ TP_ARGS(id, type)
+);
+/* End of trace events for fs/ocfs2/quota_global.c. */
#endif /* _TRACE_OCFS2_H */
/* This part must be outside protection */
diff --git a/fs/ocfs2/quota_global.c b/fs/ocfs2/quota_global.c
index 10a49e8..64346f8 100644
--- a/fs/ocfs2/quota_global.c
+++ b/fs/ocfs2/quota_global.c
@@ -11,7 +11,6 @@
#include <linux/writeback.h>
#include <linux/workqueue.h>
-#define MLOG_MASK_PREFIX ML_QUOTA
#include <cluster/masklog.h>
#include "ocfs2_fs.h"
@@ -27,6 +26,7 @@
#include "super.h"
#include "buffer_head_io.h"
#include "quota.h"
+#include "ocfs2_trace.h"
/*
* Locking of quotas with OCFS2 is rather complex. Here are rules that
@@ -132,8 +132,7 @@ int ocfs2_validate_quota_block(struct super_block *sb, struct buffer_head *bh)
struct ocfs2_disk_dqtrailer *dqt =
ocfs2_block_dqtrailer(sb->s_blocksize, bh->b_data);
- mlog(0, "Validating quota block %llu\n",
- (unsigned long long)bh->b_blocknr);
+ trace_ocfs2_validate_quota_block((unsigned long long)bh->b_blocknr);
BUG_ON(!buffer_uptodate(bh));
@@ -509,9 +508,10 @@ int __ocfs2_sync_dquot(struct dquot *dquot, int freeing)
olditime = dquot->dq_dqb.dqb_itime;
oldbtime = dquot->dq_dqb.dqb_btime;
ocfs2_global_disk2memdqb(dquot, &dqblk);
- mlog(0, "Syncing global dquot %u space %lld+%lld, inodes %lld+%lld\n",
- dquot->dq_id, dquot->dq_dqb.dqb_curspace, (long long)spacechange,
- dquot->dq_dqb.dqb_curinodes, (long long)inodechange);
+ trace_ocfs2_sync_dquot(dquot->dq_id, dquot->dq_dqb.dqb_curspace,
+ (long long)spacechange,
+ dquot->dq_dqb.dqb_curinodes,
+ (long long)inodechange);
if (!test_bit(DQ_LASTSET_B + QIF_SPACE_B, &dquot->dq_flags))
dquot->dq_dqb.dqb_curspace += spacechange;
if (!test_bit(DQ_LASTSET_B + QIF_INODES_B, &dquot->dq_flags))
@@ -595,8 +595,8 @@ static int ocfs2_sync_dquot_helper(struct dquot *dquot, unsigned long type)
struct ocfs2_super *osb = OCFS2_SB(sb);
int status = 0;
- mlog(0, "id=%u qtype=%u type=%lu device=%s\n", dquot->dq_id,
- dquot->dq_type, type, sb->s_id);
+ trace_ocfs2_sync_dquot_helper(dquot->dq_id, dquot->dq_type,
+ type, sb->s_id);
if (type != dquot->dq_type)
goto out;
status = ocfs2_lock_global_qf(oinfo, 1);
@@ -647,7 +647,7 @@ static int ocfs2_write_dquot(struct dquot *dquot)
struct ocfs2_super *osb = OCFS2_SB(dquot->dq_sb);
int status = 0;
- mlog(0, "id=%u, type=%d", dquot->dq_id, dquot->dq_type);
+ trace_ocfs2_write_dquot(dquot->dq_id, dquot->dq_type);
handle = ocfs2_start_trans(osb, OCFS2_QWRITE_CREDITS);
if (IS_ERR(handle)) {
@@ -685,7 +685,7 @@ static int ocfs2_release_dquot(struct dquot *dquot)
struct ocfs2_super *osb = OCFS2_SB(dquot->dq_sb);
int status = 0;
- mlog(0, "id=%u, type=%d", dquot->dq_id, dquot->dq_type);
+ trace_ocfs2_release_dquot(dquot->dq_id, dquot->dq_type);
mutex_lock(&dquot->dq_lock);
/* Check whether we are not racing with some other dqget() */
@@ -743,7 +743,7 @@ static int ocfs2_acquire_dquot(struct dquot *dquot)
int need_alloc = ocfs2_global_qinit_alloc(sb, type);
handle_t *handle;
- mlog(0, "id=%u, type=%d", dquot->dq_id, type);
+ trace_ocfs2_acquire_dquot(dquot->dq_id, type);
mutex_lock(&dquot->dq_lock);
/*
* We need an exclusive lock, because we're going to update use count
@@ -830,7 +830,7 @@ static int ocfs2_mark_dquot_dirty(struct dquot *dquot)
handle_t *handle;
struct ocfs2_super *osb = OCFS2_SB(sb);
- mlog(0, "id=%u, type=%d", dquot->dq_id, type);
+ trace_ocfs2_mark_dquot_dirty(dquot->dq_id, type);
/* In case user set some limits, sync dquot immediately to global
* quota file so that information propagates quicker */
--
1.7.0.4
^ permalink raw reply related [flat|nested] 53+ messages in thread
* [Ocfs2-devel] [PATCH 25/34] ocfs2: remove NAMEI from symlink.c
2010-12-23 7:19 [Ocfs2-devel] [PATCH 00/34] OCFS2: Add trace event and replace mlog(0) Tao Ma
` (23 preceding siblings ...)
2010-12-23 7:31 ` [Ocfs2-devel] [PATCH 24/34] ocfs2: Remove masklog ML_QUOTA Tao Ma
@ 2010-12-23 7:31 ` Tao Ma
2010-12-23 7:31 ` [Ocfs2-devel] [PATCH 26/34] ocfs2: Remove mlog(0) from fs/ocfs2/dir.c Tao Ma
` (13 subsequent siblings)
38 siblings, 0 replies; 53+ messages in thread
From: Tao Ma @ 2010-12-23 7:31 UTC (permalink / raw)
To: ocfs2-devel
From: Tao Ma <boyu.mt@taobao.com>
No mlog(0,...) in symlink.c, so just remove NAMEI.
Signed-off-by: Tao Ma <boyu.mt@taobao.com>
---
fs/ocfs2/symlink.c | 1 -
1 files changed, 0 insertions(+), 1 deletions(-)
diff --git a/fs/ocfs2/symlink.c b/fs/ocfs2/symlink.c
index 56c1eeb..d3179a7 100644
--- a/fs/ocfs2/symlink.c
+++ b/fs/ocfs2/symlink.c
@@ -40,7 +40,6 @@
#include <linux/pagemap.h>
#include <linux/namei.h>
-#define MLOG_MASK_PREFIX ML_NAMEI
#include <cluster/masklog.h>
#include "ocfs2.h"
--
1.7.0.4
^ permalink raw reply related [flat|nested] 53+ messages in thread
* [Ocfs2-devel] [PATCH 26/34] ocfs2: Remove mlog(0) from fs/ocfs2/dir.c
2010-12-23 7:19 [Ocfs2-devel] [PATCH 00/34] OCFS2: Add trace event and replace mlog(0) Tao Ma
` (24 preceding siblings ...)
2010-12-23 7:31 ` [Ocfs2-devel] [PATCH 25/34] ocfs2: remove NAMEI from symlink.c Tao Ma
@ 2010-12-23 7:31 ` Tao Ma
2010-12-23 7:31 ` [Ocfs2-devel] [PATCH 27/34] ocfs2: Remove masklog ML_NAMEI Tao Ma
` (12 subsequent siblings)
38 siblings, 0 replies; 53+ messages in thread
From: Tao Ma @ 2010-12-23 7:31 UTC (permalink / raw)
To: ocfs2-devel
From: Tao Ma <boyu.mt@taobao.com>
This is the 2nd step to remove the debug info of NAMEI.
Signed-off-by: Tao Ma <boyu.mt@taobao.com>
---
fs/ocfs2/dir.c | 83 +++++++++-----------
fs/ocfs2/ocfs2_trace.h | 197 ++++++++++++++++++++++++++++++++++++++++++++++++
2 files changed, 234 insertions(+), 46 deletions(-)
diff --git a/fs/ocfs2/dir.c b/fs/ocfs2/dir.c
index 65546e3..2dfb8ad 100644
--- a/fs/ocfs2/dir.c
+++ b/fs/ocfs2/dir.c
@@ -43,7 +43,6 @@
#include <linux/quotaops.h>
#include <linux/sort.h>
-#define MLOG_MASK_PREFIX ML_NAMEI
#include <cluster/masklog.h>
#include "ocfs2.h"
@@ -61,6 +60,7 @@
#include "super.h"
#include "sysfile.h"
#include "uptodate.h"
+#include "ocfs2_trace.h"
#include "buffer_head_io.h"
@@ -400,7 +400,7 @@ static int inline ocfs2_search_dirblock(struct buffer_head *bh,
}
bail:
- mlog(0, "ret = %d\n", ret);
+ trace_ocfs2_search_dirblock(ret);
return ret;
}
@@ -445,8 +445,7 @@ static int ocfs2_validate_dir_block(struct super_block *sb,
* We don't validate dirents here, that's handled
* in-place when the code walks them.
*/
- mlog(0, "Validating dirblock %llu\n",
- (unsigned long long)bh->b_blocknr);
+ trace_ocfs2_validate_dir_block((unsigned long long)bh->b_blocknr);
BUG_ON(!buffer_uptodate(bh));
@@ -784,7 +783,7 @@ cleanup_and_exit:
for (; ra_ptr < ra_max; ra_ptr++)
brelse(bh_use[ra_ptr]);
- mlog(0, "ret = %p\n", ret);
+ trace_ocfs2_find_entry_el(ret);
return ret;
}
@@ -946,11 +945,9 @@ static int ocfs2_dx_dir_search(const char *name, int namelen,
goto out;
}
- mlog(0, "Dir %llu: name: \"%.*s\", lookup of hash: %u.0x%x "
- "returns: %llu\n",
- (unsigned long long)OCFS2_I(dir)->ip_blkno,
- namelen, name, hinfo->major_hash, hinfo->minor_hash,
- (unsigned long long)phys);
+ trace_ocfs2_dx_dir_search((unsigned long long)OCFS2_I(dir)->ip_blkno,
+ namelen, name, hinfo->major_hash,
+ hinfo->minor_hash, (unsigned long long)phys);
ret = ocfs2_read_dx_leaf(dir, phys, &dx_leaf_bh);
if (ret) {
@@ -960,9 +957,9 @@ static int ocfs2_dx_dir_search(const char *name, int namelen,
dx_leaf = (struct ocfs2_dx_leaf *) dx_leaf_bh->b_data;
- mlog(0, "leaf info: num_used: %d, count: %d\n",
- le16_to_cpu(dx_leaf->dl_list.de_num_used),
- le16_to_cpu(dx_leaf->dl_list.de_count));
+ trace_ocfs2_dx_dir_search_leaf_info(
+ le16_to_cpu(dx_leaf->dl_list.de_num_used),
+ le16_to_cpu(dx_leaf->dl_list.de_count));
entry_list = &dx_leaf->dl_list;
@@ -1162,8 +1159,6 @@ static int __ocfs2_delete_entry(handle_t *handle, struct inode *dir,
int i, status = -ENOENT;
ocfs2_journal_access_func access = ocfs2_journal_access_db;
- mlog(0, "(0x%p, 0x%p, 0x%p, 0x%p)\n", handle, dir, de_del, bh);
-
if (OCFS2_I(dir)->ip_dyn_features & OCFS2_INLINE_DATA_FL)
access = ocfs2_journal_access_di;
@@ -1343,8 +1338,8 @@ static int ocfs2_delete_entry_dx(handle_t *handle, struct inode *dir,
}
}
- mlog(0, "Dir %llu: delete entry at index: %d\n",
- (unsigned long long)OCFS2_I(dir)->ip_blkno, index);
+ trace_ocfs2_delete_entry_dx((unsigned long long)OCFS2_I(dir)->ip_blkno,
+ index);
ret = __ocfs2_delete_entry(handle, dir, lookup->dl_entry,
leaf_bh, leaf_bh->b_data, leaf_bh->b_size);
@@ -2022,8 +2017,7 @@ int ocfs2_readdir(struct file * filp, void * dirent, filldir_t filldir)
struct inode *inode = filp->f_path.dentry->d_inode;
int lock_level = 0;
- mlog(0, "dirino=%llu\n",
- (unsigned long long)OCFS2_I(inode)->ip_blkno);
+ trace_ocfs2_readdir((unsigned long long)OCFS2_I(inode)->ip_blkno);
error = ocfs2_inode_lock_atime(inode, filp->f_vfsmnt, &lock_level);
if (lock_level && error >= 0) {
@@ -2064,8 +2058,8 @@ int ocfs2_find_files_on_disk(const char *name,
{
int status = -ENOENT;
- mlog(0, "name=%.*s, blkno=%p, inode=%llu\n", namelen, name, blkno,
- (unsigned long long)OCFS2_I(inode)->ip_blkno);
+ trace_ocfs2_find_files_on_disk(namelen, name, blkno,
+ (unsigned long long)OCFS2_I(inode)->ip_blkno);
status = ocfs2_find_entry(name, namelen, inode, lookup);
if (status)
@@ -2109,8 +2103,8 @@ int ocfs2_check_dir_for_entry(struct inode *dir,
int ret;
struct ocfs2_dir_lookup_result lookup = { NULL, };
- mlog(0, "dir %llu, name '%.*s'\n",
- (unsigned long long)OCFS2_I(dir)->ip_blkno, namelen, name);
+ trace_ocfs2_check_dir_for_entry(
+ (unsigned long long)OCFS2_I(dir)->ip_blkno, namelen, name);
ret = -EEXIST;
if (ocfs2_find_entry(name, namelen, dir, &lookup) == 0)
@@ -2402,9 +2396,9 @@ static int ocfs2_dx_dir_attach_index(struct ocfs2_super *osb,
goto out;
}
- mlog(0, "Dir %llu, attach new index block: %llu\n",
- (unsigned long long)OCFS2_I(dir)->ip_blkno,
- (unsigned long long)dr_blkno);
+ trace_ocfs2_dx_dir_attach_index(
+ (unsigned long long)OCFS2_I(dir)->ip_blkno,
+ (unsigned long long)dr_blkno);
dx_root_bh = sb_getblk(osb->sb, dr_blkno);
if (dx_root_bh == NULL) {
@@ -2502,11 +2496,10 @@ static int ocfs2_dx_dir_format_cluster(struct ocfs2_super *osb,
dx_leaf->dl_list.de_count =
cpu_to_le16(ocfs2_dx_entries_per_leaf(osb->sb));
- mlog(0,
- "Dir %llu, format dx_leaf: %llu, entry count: %u\n",
- (unsigned long long)OCFS2_I(dir)->ip_blkno,
- (unsigned long long)bh->b_blocknr,
- le16_to_cpu(dx_leaf->dl_list.de_count));
+ trace_ocfs2_dx_dir_format_cluster(
+ (unsigned long long)OCFS2_I(dir)->ip_blkno,
+ (unsigned long long)bh->b_blocknr,
+ le16_to_cpu(dx_leaf->dl_list.de_count));
ocfs2_journal_dirty(handle, bh);
}
@@ -2750,12 +2743,11 @@ static void ocfs2_dx_dir_index_root_block(struct inode *dir,
ocfs2_dx_dir_name_hash(dir, de->name, de->name_len, &hinfo);
- mlog(0,
- "dir: %llu, major: 0x%x minor: 0x%x, index: %u, name: %.*s\n",
- (unsigned long long)dir->i_ino, hinfo.major_hash,
- hinfo.minor_hash,
- le16_to_cpu(dx_root->dr_entries.de_num_used),
- de->name_len, de->name);
+ trace_ocfs2_dx_dir_index_root_block(
+ (unsigned long long)dir->i_ino,
+ hinfo.major_hash, hinfo.minor_hash,
+ de->name_len, de->name,
+ le16_to_cpu(dx_root->dr_entries.de_num_used));
ocfs2_dx_entry_list_insert(&dx_root->dr_entries, &hinfo,
dirent_blk);
@@ -3308,8 +3300,8 @@ static int ocfs2_extend_dir(struct ocfs2_super *osb,
down_write(&OCFS2_I(dir)->ip_alloc_sem);
drop_alloc_sem = 1;
dir_i_size = i_size_read(dir);
- mlog(0, "extending dir %llu (i_size = %lld)\n",
- (unsigned long long)OCFS2_I(dir)->ip_blkno, dir_i_size);
+ trace_ocfs2_extend_dir((unsigned long long)OCFS2_I(dir)->ip_blkno,
+ dir_i_size);
/* dir->i_size is always block aligned. */
spin_lock(&OCFS2_I(dir)->ip_lock);
@@ -3803,9 +3795,9 @@ static int ocfs2_dx_dir_rebalance(struct ocfs2_super *osb, struct inode *dir,
struct ocfs2_dx_root_block *dx_root;
struct ocfs2_dx_leaf *tmp_dx_leaf = NULL;
- mlog(0, "DX Dir: %llu, rebalance leaf leaf_blkno: %llu insert: %u\n",
- (unsigned long long)OCFS2_I(dir)->ip_blkno,
- (unsigned long long)leaf_blkno, insert_hash);
+ trace_ocfs2_dx_dir_rebalance((unsigned long long)OCFS2_I(dir)->ip_blkno,
+ (unsigned long long)leaf_blkno,
+ insert_hash);
ocfs2_init_dx_root_extent_tree(&et, INODE_CACHE(dir), dx_root_bh);
@@ -3885,8 +3877,7 @@ static int ocfs2_dx_dir_rebalance(struct ocfs2_super *osb, struct inode *dir,
goto out_commit;
}
- mlog(0, "Split leaf (%u) at %u, insert major hash is %u\n",
- leaf_cpos, split_hash, insert_hash);
+ trace_ocfs2_dx_dir_rebalance_split(leaf_cpos, split_hash, insert_hash);
/*
* We have to carefully order operations here. There are items
@@ -4343,8 +4334,8 @@ int ocfs2_prepare_dir_for_insert(struct ocfs2_super *osb,
unsigned int blocks_wanted = 1;
struct buffer_head *bh = NULL;
- mlog(0, "getting ready to insert namelen %d into dir %llu\n",
- namelen, (unsigned long long)OCFS2_I(dir)->ip_blkno);
+ trace_ocfs2_prepare_dir_for_insert(
+ (unsigned long long)OCFS2_I(dir)->ip_blkno, namelen);
if (!namelen) {
ret = -EINVAL;
diff --git a/fs/ocfs2/ocfs2_trace.h b/fs/ocfs2/ocfs2_trace.h
index f87a203..6456975 100644
--- a/fs/ocfs2/ocfs2_trace.h
+++ b/fs/ocfs2/ocfs2_trace.h
@@ -3110,6 +3110,203 @@ DEFINE_EVENT(ocfs2__uint_int, ocfs2_mark_dquot_dirty,
TP_ARGS(id, type)
);
/* End of trace events for fs/ocfs2/quota_global.c. */
+
+/* Trace events for fs/ocfs2/dir.c. */
+DEFINE_EVENT(ocfs2__int, ocfs2_search_dirblock,
+
+ TP_PROTO(int ret),
+
+ TP_ARGS(ret)
+);
+
+DEFINE_EVENT(ocfs2__ull, ocfs2_validate_dir_block,
+
+ TP_PROTO(unsigned long long blkno),
+
+ TP_ARGS(blkno)
+);
+
+DEFINE_EVENT(ocfs2__pointer, ocfs2_find_entry_el,
+
+ TP_PROTO(void *ret),
+
+ TP_ARGS(ret)
+);
+
+TRACE_EVENT(ocfs2_dx_dir_search,
+
+ TP_PROTO(unsigned long long ino, int namelen, const char *name,
+ unsigned int major_hash, unsigned int minor_hash,
+ unsigned long long blkno),
+
+ TP_ARGS(ino, namelen, name, major_hash, minor_hash, blkno),
+
+ TP_STRUCT__entry(
+ __field( unsigned long long, ino )
+ __field( int, namelen )
+ __field( const char *, name )
+ __field( unsigned int, major_hash )
+ __field( unsigned int, minor_hash )
+ __field( unsigned long long, blkno )
+ ),
+
+ TP_fast_assign(
+ __entry->ino = ino;
+ __entry->namelen = namelen;
+ __entry->name = name;
+ __entry->major_hash = major_hash;
+ __entry->minor_hash = minor_hash;
+ __entry->blkno = blkno;
+ ),
+
+ TP_printk("%llu %.*s %u %u %llu", __entry->ino,
+ __entry->namelen, __entry->name,
+ __entry->major_hash, __entry->minor_hash, __entry->blkno)
+);
+
+DEFINE_EVENT(ocfs2__uint_uint, ocfs2_dx_dir_search_leaf_info,
+
+ TP_PROTO(unsigned int num_used, unsigned int count),
+
+ TP_ARGS(num_used, count)
+);
+
+DEFINE_EVENT(ocfs2__ull_int, ocfs2_delete_entry_dx,
+
+ TP_PROTO(unsigned long long dir, int index),
+
+ TP_ARGS(dir, index)
+);
+
+DEFINE_EVENT(ocfs2__ull, ocfs2_readdir,
+
+ TP_PROTO(unsigned long long dir),
+
+ TP_ARGS(dir)
+);
+
+TRACE_EVENT(ocfs2_find_files_on_disk,
+
+ TP_PROTO(int namelen, const char *name, void *blkno,
+ unsigned long long dir),
+
+ TP_ARGS(namelen, name, blkno, dir),
+
+ TP_STRUCT__entry(
+ __field( int, namelen )
+ __field( const char *, name )
+ __field( void *, blkno )
+ __field( unsigned long long, dir )
+ ),
+
+ TP_fast_assign(
+ __entry->namelen = namelen;
+ __entry->name = name;
+ __entry->blkno = blkno;
+ __entry->dir = dir;
+ ),
+
+ TP_printk("%.*s %p %llu", __entry->namelen, __entry->name,
+ __entry->blkno, __entry->dir)
+);
+
+TRACE_EVENT(ocfs2_check_dir_for_entry,
+
+ TP_PROTO(unsigned long long dir, int namelen, const char *name),
+
+ TP_ARGS(dir, namelen, name),
+
+ TP_STRUCT__entry(
+ __field( unsigned long long, dir )
+ __field( int, namelen )
+ __field( const char *, name )
+ ),
+
+ TP_fast_assign(
+ __entry->dir = dir;
+ __entry->namelen = namelen;
+ __entry->name = name;
+ ),
+
+ TP_printk("%llu %.*s", __entry->dir, __entry->namelen, __entry->name)
+);
+
+DEFINE_EVENT(ocfs2__ull_ull, ocfs2_dx_dir_attach_index,
+
+ TP_PROTO(unsigned long long dir, unsigned long long new_blkno),
+
+ TP_ARGS(dir, new_blkno)
+);
+
+DEFINE_EVENT(ocfs2__ull_ull_uint, ocfs2_dx_dir_format_cluster,
+
+ TP_PROTO(unsigned long long dir, unsigned long long new_blkno,
+ unsigned int count),
+
+ TP_ARGS(dir, new_blkno, count)
+);
+
+TRACE_EVENT(ocfs2_dx_dir_index_root_block,
+
+ TP_PROTO(unsigned long long dir,
+ unsigned int major_hash, unsigned int minor_hash,
+ int namelen, const char *name, unsigned int num_used),
+
+ TP_ARGS(dir, major_hash, minor_hash, namelen, name, num_used),
+
+ TP_STRUCT__entry(
+ __field( unsigned long long, dir )
+ __field( unsigned int, major_hash )
+ __field( unsigned int, minor_hash )
+ __field( int, namelen )
+ __field( const char *, name )
+ __field( unsigned int, num_used )
+ ),
+
+ TP_fast_assign(
+ __entry->dir = dir;
+ __entry->major_hash = major_hash;
+ __entry->minor_hash = minor_hash;
+ __entry->namelen = namelen;
+ __entry->name = name;
+ __entry->num_used = num_used;
+ ),
+
+ TP_printk("%llu %x %x %.*s %u", __entry->dir,
+ __entry->major_hash, __entry->minor_hash,
+ __entry->namelen, __entry->name, __entry->num_used)
+);
+
+DEFINE_EVENT(ocfs2__ull_ull, ocfs2_extend_dir,
+
+ TP_PROTO(unsigned long long dir, unsigned long long i_size),
+
+ TP_ARGS(dir, i_size)
+);
+
+DEFINE_EVENT(ocfs2__ull_ull_uint, ocfs2_dx_dir_rebalance,
+
+ TP_PROTO(unsigned long long dir, unsigned long long leaf_blkno,
+ unsigned int insert_hash),
+
+ TP_ARGS(dir, leaf_blkno, insert_hash)
+);
+
+DEFINE_EVENT(ocfs2__uint_uint_uint, ocfs2_dx_dir_rebalance_split,
+
+ TP_PROTO(unsigned int leaf_cpos, unsigned int split_hash,
+ unsigned int insert_hash),
+
+ TP_ARGS(leaf_cpos, split_hash, insert_hash)
+);
+
+DEFINE_EVENT(ocfs2__ull_int, ocfs2_prepare_dir_for_insert,
+
+ TP_PROTO(unsigned long long dir, int namelen),
+
+ TP_ARGS(dir, namelen)
+);
+/* End of trace events for fs/ocfs2/dir.c. */
#endif /* _TRACE_OCFS2_H */
/* This part must be outside protection */
--
1.7.0.4
^ permalink raw reply related [flat|nested] 53+ messages in thread
* [Ocfs2-devel] [PATCH 27/34] ocfs2: Remove masklog ML_NAMEI.
2010-12-23 7:19 [Ocfs2-devel] [PATCH 00/34] OCFS2: Add trace event and replace mlog(0) Tao Ma
` (25 preceding siblings ...)
2010-12-23 7:31 ` [Ocfs2-devel] [PATCH 26/34] ocfs2: Remove mlog(0) from fs/ocfs2/dir.c Tao Ma
@ 2010-12-23 7:31 ` Tao Ma
2010-12-23 7:31 ` [Ocfs2-devel] [PATCH 28/34] ocfs2: Remove masklog ML_DCACHE Tao Ma
` (11 subsequent siblings)
38 siblings, 0 replies; 53+ messages in thread
From: Tao Ma @ 2010-12-23 7:31 UTC (permalink / raw)
To: ocfs2-devel
From: Tao Ma <boyu.mt@taobao.com>
Remove mlog(0) from fs/ocfs2/namei.c and the masklog NAMEI finally.
Signed-off-by: Tao Ma <boyu.mt@taobao.com>
---
fs/ocfs2/cluster/masklog.c | 1 -
fs/ocfs2/cluster/masklog.h | 1 -
fs/ocfs2/namei.c | 126 +++++++---------
fs/ocfs2/ocfs2_trace.h | 351 ++++++++++++++++++++++++++++++++++++++++++++
4 files changed, 408 insertions(+), 71 deletions(-)
diff --git a/fs/ocfs2/cluster/masklog.c b/fs/ocfs2/cluster/masklog.c
index 44d22cd..89cdbf8 100644
--- a/fs/ocfs2/cluster/masklog.c
+++ b/fs/ocfs2/cluster/masklog.c
@@ -96,7 +96,6 @@ static struct mlog_attribute mlog_attrs[MLOG_MAX_BITS] = {
define_mask(DLM_GLUE),
define_mask(BH_IO),
define_mask(UPTODATE),
- define_mask(NAMEI),
define_mask(VOTE),
define_mask(DCACHE),
define_mask(CONN),
diff --git a/fs/ocfs2/cluster/masklog.h b/fs/ocfs2/cluster/masklog.h
index 789e735..8a8386f 100644
--- a/fs/ocfs2/cluster/masklog.h
+++ b/fs/ocfs2/cluster/masklog.h
@@ -98,7 +98,6 @@
#define ML_DLM_GLUE 0x0000000000080000ULL /* ocfs2 dlm glue layer */
#define ML_BH_IO 0x0000000000100000ULL /* ocfs2 buffer I/O */
#define ML_UPTODATE 0x0000000000200000ULL /* ocfs2 caching sequence #'s */
-#define ML_NAMEI 0x0000000000400000ULL /* ocfs2 directory / namespace */
#define ML_VOTE 0x0000000001000000ULL /* ocfs2 node messaging */
#define ML_DCACHE 0x0000000002000000ULL /* ocfs2 dcache operations */
#define ML_CONN 0x0000000004000000ULL /* net connection management */
diff --git a/fs/ocfs2/namei.c b/fs/ocfs2/namei.c
index f52e6bf..3370616 100644
--- a/fs/ocfs2/namei.c
+++ b/fs/ocfs2/namei.c
@@ -42,7 +42,6 @@
#include <linux/highmem.h>
#include <linux/quotaops.h>
-#define MLOG_MASK_PREFIX ML_NAMEI
#include <cluster/masklog.h>
#include "ocfs2.h"
@@ -63,6 +62,7 @@
#include "uptodate.h"
#include "xattr.h"
#include "acl.h"
+#include "ocfs2_trace.h"
#include "buffer_head_io.h"
@@ -106,17 +106,15 @@ static struct dentry *ocfs2_lookup(struct inode *dir, struct dentry *dentry,
struct dentry *ret;
struct ocfs2_inode_info *oi;
- mlog(0, "(0x%p, 0x%p, '%.*s')\n", dir, dentry,
- dentry->d_name.len, dentry->d_name.name);
+ trace_ocfs2_lookup(dir, dentry, dentry->d_name.len,
+ dentry->d_name.name,
+ (unsigned long long)OCFS2_I(dir)->ip_blkno, 0);
if (dentry->d_name.len > OCFS2_MAX_FILENAME_LEN) {
ret = ERR_PTR(-ENAMETOOLONG);
goto bail;
}
- mlog(0, "find name %.*s in directory %llu\n", dentry->d_name.len,
- dentry->d_name.name, (unsigned long long)OCFS2_I(dir)->ip_blkno);
-
status = ocfs2_inode_lock_nested(dir, NULL, 0, OI_LS_PARENT);
if (status < 0) {
if (status != -ENOENT)
@@ -183,7 +181,7 @@ bail_unlock:
bail:
- mlog(0, "%p\n", ret);
+ trace_ocfs2_lookup_ret(ret);
return ret;
}
@@ -236,9 +234,9 @@ static int ocfs2_mknod(struct inode *dir,
sigset_t oldset;
int did_block_signals = 0;
- mlog(0, "(0x%p, 0x%p, %d, %lu, '%.*s')\n", dir, dentry, mode,
- (unsigned long)dev, dentry->d_name.len,
- dentry->d_name.name);
+ trace_ocfs2_mknod(dir, dentry, dentry->d_name.len, dentry->d_name.name,
+ (unsigned long long)OCFS2_I(inode)->ip_blkno,
+ (unsigned long)dev, mode);
dquot_initialize(dir);
@@ -355,10 +353,6 @@ static int ocfs2_mknod(struct inode *dir,
goto leave;
did_quota_inode = 1;
- mlog(0, "(0x%p, 0x%p, %d, %lu, '%.*s')\n", dir, dentry,
- inode->i_mode, (unsigned long)dev, dentry->d_name.len,
- dentry->d_name.name);
-
/* do the real work now. */
status = ocfs2_mknod_locked(osb, dir, inode, dev,
&new_fe_bh, parent_fe_bh, handle,
@@ -438,9 +432,6 @@ leave:
if (did_block_signals)
ocfs2_unblock_signals(&oldset);
- if (status == -ENOSPC)
- mlog(0, "Disk is full\n");
-
brelse(new_fe_bh);
brelse(parent_fe_bh);
kfree(si.name);
@@ -619,8 +610,8 @@ static int ocfs2_mkdir(struct inode *dir,
{
int ret;
- mlog(0, "(0x%p, 0x%p, %d, '%.*s')\n", dir, dentry, mode,
- dentry->d_name.len, dentry->d_name.name);
+ trace_ocfs2_mkdir(dir, dentry, dentry->d_name.len, dentry->d_name.name,
+ OCFS2_I(dir)->ip_blkno, mode);
ret = ocfs2_mknod(dir, dentry, mode | S_IFDIR, 0);
if (ret)
mlog_errno(ret);
@@ -635,8 +626,8 @@ static int ocfs2_create(struct inode *dir,
{
int ret;
- mlog(0, "(0x%p, 0x%p, %d, '%.*s')\n", dir, dentry, mode,
- dentry->d_name.len, dentry->d_name.name);
+ trace_ocfs2_create(dir, dentry, dentry->d_name.len, dentry->d_name.name,
+ (unsigned long long)OCFS2_I(dir)->ip_blkno, mode);
ret = ocfs2_mknod(dir, dentry, mode | S_IFREG, 0);
if (ret)
mlog_errno(ret);
@@ -658,9 +649,9 @@ static int ocfs2_link(struct dentry *old_dentry,
struct ocfs2_dir_lookup_result lookup = { NULL, };
sigset_t oldset;
- mlog(0, "(inode=%lu, old='%.*s' new='%.*s')\n", inode->i_ino,
- old_dentry->d_name.len, old_dentry->d_name.name,
- dentry->d_name.len, dentry->d_name.name);
+ trace_ocfs2_link((unsigned long long)OCFS2_I(inode)->ip_blkno,
+ old_dentry->d_name.len, old_dentry->d_name.name,
+ dentry->d_name.len, dentry->d_name.name);
if (S_ISDIR(inode->i_mode))
return -EPERM;
@@ -817,19 +808,17 @@ static int ocfs2_unlink(struct inode *dir,
struct ocfs2_dir_lookup_result lookup = { NULL, };
struct ocfs2_dir_lookup_result orphan_insert = { NULL, };
- mlog(0, "(0x%p, 0x%p, '%.*s')\n", dir, dentry,
- dentry->d_name.len, dentry->d_name.name);
+ trace_ocfs2_unlink(dir, dentry, dentry->d_name.len,
+ dentry->d_name.name,
+ (unsigned long long)OCFS2_I(dir)->ip_blkno,
+ (unsigned long long)OCFS2_I(inode)->ip_blkno);
dquot_initialize(dir);
BUG_ON(dentry->d_parent->d_inode != dir);
- mlog(0, "ino = %llu\n", (unsigned long long)OCFS2_I(inode)->ip_blkno);
-
- if (inode == osb->root_inode) {
- mlog(0, "Cannot delete the root directory\n");
+ if (inode == osb->root_inode)
return -EPERM;
- }
status = ocfs2_inode_lock_nested(dir, &parent_node_bh, 1,
OI_LS_PARENT);
@@ -851,9 +840,10 @@ static int ocfs2_unlink(struct inode *dir,
if (OCFS2_I(inode)->ip_blkno != blkno) {
status = -ENOENT;
- mlog(0, "ip_blkno %llu != dirent blkno %llu ip_flags = %x\n",
- (unsigned long long)OCFS2_I(inode)->ip_blkno,
- (unsigned long long)blkno, OCFS2_I(inode)->ip_flags);
+ trace_ocfs2_unlink_noent(
+ (unsigned long long)OCFS2_I(inode)->ip_blkno,
+ (unsigned long long)blkno,
+ OCFS2_I(inode)->ip_flags);
goto leave;
}
@@ -984,9 +974,8 @@ static int ocfs2_double_lock(struct ocfs2_super *osb,
struct buffer_head **tmpbh;
struct inode *tmpinode;
- mlog(0, "(inode1 = %llu, inode2 = %llu)\n",
- (unsigned long long)oi1->ip_blkno,
- (unsigned long long)oi2->ip_blkno);
+ trace_ocfs2_double_lock((unsigned long long)oi1->ip_blkno,
+ (unsigned long long)oi2->ip_blkno);
if (*bh1)
*bh1 = NULL;
@@ -997,7 +986,6 @@ static int ocfs2_double_lock(struct ocfs2_super *osb,
if (oi1->ip_blkno != oi2->ip_blkno) {
if (oi1->ip_blkno < oi2->ip_blkno) {
/* switch id1 and id2 around */
- mlog(0, "switching them around...\n");
tmpbh = bh2;
bh2 = bh1;
bh1 = tmpbh;
@@ -1030,6 +1018,10 @@ static int ocfs2_double_lock(struct ocfs2_super *osb,
mlog_errno(status);
}
+ trace_ocfs2_double_lock_end(
+ (unsigned long long)OCFS2_I(inode1)->ip_blkno,
+ (unsigned long long)OCFS2_I(inode2)->ip_blkno);
+
bail:
if (status)
mlog_errno(status);
@@ -1074,10 +1066,9 @@ static int ocfs2_rename(struct inode *old_dir,
/* At some point it might be nice to break this function up a
* bit. */
- mlog(0, "(0x%p, 0x%p, 0x%p, 0x%p, from='%.*s' to='%.*s')\n",
- old_dir, old_dentry, new_dir, new_dentry,
- old_dentry->d_name.len, old_dentry->d_name.name,
- new_dentry->d_name.len, new_dentry->d_name.name);
+ trace_ocfs2_rename(old_dir, old_dentry, new_dir, new_dentry,
+ old_dentry->d_name.len, old_dentry->d_name.name,
+ new_dentry->d_name.len, new_dentry->d_name.name);
dquot_initialize(old_dir);
dquot_initialize(new_dir);
@@ -1234,16 +1225,15 @@ static int ocfs2_rename(struct inode *old_dir,
if (!new_inode) {
status = -EACCES;
- mlog(0, "We found an inode for name %.*s but VFS "
- "didn't give us one.\n", new_dentry->d_name.len,
- new_dentry->d_name.name);
+ trace_ocfs2_rename_target_exists(new_dentry->d_name.len,
+ new_dentry->d_name.name);
goto bail;
}
if (OCFS2_I(new_inode)->ip_blkno != newfe_blkno) {
status = -EACCES;
- mlog(0, "Inode %llu and dir %llu disagree. flags = %x\n",
+ trace_ocfs2_rename_disagree(
(unsigned long long)OCFS2_I(new_inode)->ip_blkno,
(unsigned long long)newfe_blkno,
OCFS2_I(new_inode)->ip_flags);
@@ -1266,8 +1256,7 @@ static int ocfs2_rename(struct inode *old_dir,
newfe = (struct ocfs2_dinode *) newfe_bh->b_data;
- mlog(0, "aha rename over existing... new_blkno=%llu "
- "newfebh=%p bhblocknr=%llu\n",
+ trace_ocfs2_rename_over_existing(
(unsigned long long)newfe_blkno, newfe_bh, newfe_bh ?
(unsigned long long)newfe_bh->b_blocknr : 0ULL);
@@ -1509,9 +1498,8 @@ static int ocfs2_create_symlink_data(struct ocfs2_super *osb,
* write i_size + 1 bytes. */
blocks = (bytes_left + sb->s_blocksize - 1) >> sb->s_blocksize_bits;
- mlog(0, "i_blocks = %llu, i_size = %llu, blocks = %d\n",
- (unsigned long long)inode->i_blocks,
- i_size_read(inode), blocks);
+ trace_ocfs2_create_symlink_data((unsigned long long)inode->i_blocks,
+ i_size_read(inode), blocks);
/* Sanity check -- make sure we're going to fit. */
if (bytes_left >
@@ -1619,8 +1607,8 @@ static int ocfs2_symlink(struct inode *dir,
sigset_t oldset;
int did_block_signals = 0;
- mlog(0, "(0x%p, 0x%p, symname='%s' actual='%.*s')\n", dir,
- dentry, symname, dentry->d_name.len, dentry->d_name.name);
+ trace_ocfs2_symlink_begin(dir, dentry, symname,
+ dentry->d_name.len, dentry->d_name.name);
dquot_initialize(dir);
@@ -1722,9 +1710,10 @@ static int ocfs2_symlink(struct inode *dir,
goto bail;
did_quota_inode = 1;
- mlog(0, "(0x%p, 0x%p, %d, '%.*s')\n", dir, dentry,
- inode->i_mode, dentry->d_name.len,
- dentry->d_name.name);
+ trace_ocfs2_symlink_create(dir, dentry, dentry->d_name.len,
+ dentry->d_name.name,
+ (unsigned long long)OCFS2_I(dir)->ip_blkno,
+ inode->i_mode);
status = ocfs2_mknod_locked(osb, dir, inode,
0, &new_fe_bh, parent_fe_bh, handle,
@@ -1871,8 +1860,7 @@ static int ocfs2_blkno_stringify(u64 blkno, char *name)
goto bail;
}
- mlog(0, "built filename '%s' for orphan dir (len=%d)\n", name,
- namelen);
+ trace_ocfs2_blkno_stringify(blkno, name, namelen);
status = 0;
bail:
@@ -2008,7 +1996,8 @@ static int ocfs2_orphan_add(struct ocfs2_super *osb,
struct ocfs2_dinode *orphan_fe;
struct ocfs2_dinode *fe = (struct ocfs2_dinode *) fe_bh->b_data;
- mlog(0, "(inode->i_ino = %lu)\n", inode->i_ino);
+ trace_ocfs2_orphan_add_begin(
+ (unsigned long long)OCFS2_I(inode)->ip_blkno);
status = ocfs2_read_inode_block(orphan_dir_inode, &orphan_dir_bh);
if (status < 0) {
@@ -2067,8 +2056,8 @@ static int ocfs2_orphan_add(struct ocfs2_super *osb,
ocfs2_journal_dirty(handle, fe_bh);
- mlog(0, "Inode %llu orphaned in slot %d\n",
- (unsigned long long)OCFS2_I(inode)->ip_blkno, osb->slot_num);
+ trace_ocfs2_orphan_add_end((unsigned long long)OCFS2_I(inode)->ip_blkno,
+ osb->slot_num);
leave:
brelse(orphan_dir_bh);
@@ -2096,9 +2085,9 @@ int ocfs2_orphan_del(struct ocfs2_super *osb,
goto leave;
}
- mlog(0, "removing '%s' from orphan dir %llu (namelen=%d)\n",
- name, (unsigned long long)OCFS2_I(orphan_dir_inode)->ip_blkno,
- OCFS2_ORPHAN_NAMELEN);
+ trace_ocfs2_orphan_del(
+ (unsigned long long)OCFS2_I(orphan_dir_inode)->ip_blkno,
+ name, OCFS2_ORPHAN_NAMELEN);
/* find it's spot in the orphan directory */
status = ocfs2_find_entry(name, OCFS2_ORPHAN_NAMELEN, orphan_dir_inode,
@@ -2332,9 +2321,6 @@ leave:
iput(orphan_dir);
}
- if (status == -ENOSPC)
- mlog(0, "Disk is full\n");
-
if ((status < 0) && inode) {
clear_nlink(inode);
iput(inode);
@@ -2369,8 +2355,10 @@ int ocfs2_mv_orphaned_inode_to_new(struct inode *dir,
struct buffer_head *di_bh = NULL;
struct ocfs2_dir_lookup_result lookup = { NULL, };
- mlog(0, "(0x%p, 0x%p, %.*s')\n", dir, dentry,
- dentry->d_name.len, dentry->d_name.name);
+ trace_ocfs2_mv_orphaned_inode_to_new(dir, dentry,
+ dentry->d_name.len, dentry->d_name.name,
+ (unsigned long long)OCFS2_I(dir)->ip_blkno,
+ (unsigned long long)OCFS2_I(inode)->ip_blkno);
status = ocfs2_inode_lock(dir, &parent_di_bh, 1);
if (status < 0) {
diff --git a/fs/ocfs2/ocfs2_trace.h b/fs/ocfs2/ocfs2_trace.h
index 6456975..fa424a6 100644
--- a/fs/ocfs2/ocfs2_trace.h
+++ b/fs/ocfs2/ocfs2_trace.h
@@ -3307,6 +3307,357 @@ DEFINE_EVENT(ocfs2__ull_int, ocfs2_prepare_dir_for_insert,
TP_ARGS(dir, namelen)
);
/* End of trace events for fs/ocfs2/dir.c. */
+
+/* Trace events for fs/ocfs2/namei.c. */
+DECLARE_EVENT_CLASS(ocfs2__dentry_ops,
+
+ TP_PROTO(void *dir, void *dentry, int name_len, const char *name,
+ unsigned long long dir_blkno, unsigned long long extra),
+
+ TP_ARGS(dir, dentry, name_len, name, dir_blkno, extra),
+
+ TP_STRUCT__entry(
+ __field( void *, dir )
+ __field( void *, dentry )
+ __field( int, name_len )
+ __field( const char *, name )
+ __field( unsigned long long, dir_blkno )
+ __field( unsigned long long, extra )
+ ),
+
+ TP_fast_assign(
+ __entry->dir = dir;
+ __entry->dentry = dentry;
+ __entry->name_len = name_len;
+ __entry->name = name;
+ __entry->dir_blkno = dir_blkno;
+ __entry->extra = extra;
+ ),
+
+ TP_printk("%p %p %.*s %llu %llu", __entry->dir, __entry->dentry,
+ __entry->name_len, __entry->name,
+ __entry->dir_blkno, __entry->extra)
+);
+
+DEFINE_EVENT(ocfs2__dentry_ops, ocfs2_lookup,
+
+ TP_PROTO(void *dir, void *dentry, int name_len, const char *name,
+ unsigned long long dir_blkno, unsigned long long extra),
+
+ TP_ARGS(dir, dentry, name_len, name, dir_blkno, extra)
+);
+
+DEFINE_EVENT(ocfs2__pointer, ocfs2_lookup_ret,
+
+ TP_PROTO(void *ret),
+
+ TP_ARGS(ret)
+);
+
+TRACE_EVENT(ocfs2_mknod,
+
+ TP_PROTO(void *dir, void *dentry, int name_len, const char *name,
+ unsigned long long dir_blkno, unsigned long dev, int mode),
+
+ TP_ARGS(dir, dentry, name_len, name, dir_blkno, dev, mode),
+
+ TP_STRUCT__entry(
+ __field( void *, dir )
+ __field( void *, dentry )
+ __field( int, name_len )
+ __field( const char *, name )
+ __field( unsigned long long, dir_blkno )
+ __field( unsigned long, dev )
+ __field( int, mode )
+ ),
+
+ TP_fast_assign(
+ __entry->dir = dir;
+ __entry->dentry = dentry;
+ __entry->name_len = name_len;
+ __entry->name = name;
+ __entry->dir_blkno = dir_blkno;
+ __entry->dev = dev;
+ __entry->mode = mode;
+ ),
+
+ TP_printk("%p %p %.*s %llu %lu %d", __entry->dir, __entry->dentry,
+ __entry->name_len, __entry->name,
+ __entry->dir_blkno, __entry->dev, __entry->mode)
+);
+
+DEFINE_EVENT(ocfs2__dentry_ops, ocfs2_mkdir,
+
+ TP_PROTO(void *dir, void *dentry, int name_len, const char *name,
+ unsigned long long dir_blkno, unsigned long long mode),
+
+ TP_ARGS(dir, dentry, name_len, name, dir_blkno, mode)
+);
+
+DEFINE_EVENT(ocfs2__dentry_ops, ocfs2_create,
+
+ TP_PROTO(void *dir, void *dentry, int name_len, const char *name,
+ unsigned long long dir_blkno, unsigned long long mode),
+
+ TP_ARGS(dir, dentry, name_len, name, dir_blkno, mode)
+);
+
+TRACE_EVENT(ocfs2_link,
+
+ TP_PROTO(unsigned long long ino, int old_len, const char *old_name,
+ int name_len, const char *name),
+
+ TP_ARGS(ino, old_len, old_name, name_len, name),
+
+ TP_STRUCT__entry(
+ __field( unsigned long long, ino )
+ __field( int, old_len )
+ __field( const char *, old_name )
+ __field( int, name_len )
+ __field( const char *, name )
+ ),
+
+ TP_fast_assign(
+ __entry->ino = ino;
+ __entry->old_len = old_len;
+ __entry->old_name = old_name;
+ __entry->name_len = name_len;
+ __entry->name = name;
+ ),
+
+ TP_printk("%llu %.*s %.*s", __entry->ino,
+ __entry->old_len, __entry->old_name,
+ __entry->name_len, __entry->name)
+);
+
+DEFINE_EVENT(ocfs2__dentry_ops, ocfs2_unlink,
+
+ TP_PROTO(void *dir, void *dentry, int name_len, const char *name,
+ unsigned long long dir_blkno, unsigned long long ino),
+
+ TP_ARGS(dir, dentry, name_len, name, dir_blkno, ino)
+);
+
+DEFINE_EVENT(ocfs2__ull_ull_uint, ocfs2_unlink_noent,
+
+ TP_PROTO(unsigned long long ip_blkno, unsigned long long blkno,
+ unsigned int ip_flags),
+
+ TP_ARGS(ip_blkno, blkno, ip_flags)
+);
+
+DEFINE_EVENT(ocfs2__ull_ull, ocfs2_double_lock,
+
+ TP_PROTO(unsigned long long blkno1, unsigned long long blkno2),
+
+ TP_ARGS(blkno1, blkno2)
+);
+
+DEFINE_EVENT(ocfs2__ull_ull, ocfs2_double_lock_end,
+
+ TP_PROTO(unsigned long long blkno1, unsigned long long blkno2),
+
+ TP_ARGS(blkno1, blkno2)
+);
+
+TRACE_EVENT(ocfs2_rename,
+
+ TP_PROTO(void *old_dir, void *old_dentry,
+ void *new_dir, void *new_dentry,
+ int old_len, const char *old_name,
+ int new_len, const char *new_name),
+
+ TP_ARGS(old_dir, old_dentry, new_dir, new_dentry,
+ old_len, old_name, new_len, new_name),
+
+ TP_STRUCT__entry(
+ __field( void *, old_dir )
+ __field( void *, old_dentry )
+ __field( void *, new_dir )
+ __field( void *, new_dentry )
+ __field( int, old_len )
+ __field( const char *, old_name )
+ __field( int, new_len )
+ __field( const char *, new_name )
+ ),
+
+ TP_fast_assign(
+ __entry->old_dir = old_dir;
+ __entry->old_dentry = old_dentry;
+ __entry->new_dir = new_dir;
+ __entry->new_dentry = new_dentry;
+ __entry->old_len = old_len;
+ __entry->old_name = old_name;
+ __entry->new_len = new_len;
+ __entry->new_name = new_name;
+ ),
+
+ TP_printk("%p %p %p %p %.*s %.*s",
+ __entry->old_dir, __entry->old_dentry,
+ __entry->new_dir, __entry->new_dentry,
+ __entry->old_len, __entry->old_name,
+ __entry->new_len, __entry->new_name)
+);
+
+TRACE_EVENT(ocfs2_rename_target_exists,
+
+ TP_PROTO(int new_len, const char *new_name),
+
+ TP_ARGS(new_len, new_name),
+
+ TP_STRUCT__entry(
+ __field( int, new_len )
+ __field( const char *, new_name )
+ ),
+
+ TP_fast_assign(
+ __entry->new_len = new_len;
+ __entry->new_name = new_name;
+ ),
+
+ TP_printk("%.*s", __entry->new_len, __entry->new_name)
+);
+
+DEFINE_EVENT(ocfs2__ull_ull_uint, ocfs2_rename_disagree,
+
+ TP_PROTO(unsigned long long ip_blkno, unsigned long long new_blkno,
+ unsigned int ip_flags),
+
+ TP_ARGS(ip_blkno, new_blkno, ip_flags)
+);
+
+TRACE_EVENT(ocfs2_rename_over_existing,
+
+ TP_PROTO(unsigned long long new_blkno, void *new_bh,
+ unsigned long long newdi_blkno),
+
+ TP_ARGS(new_blkno, new_bh, newdi_blkno),
+
+ TP_STRUCT__entry(
+ __field( unsigned long long, new_blkno )
+ __field( void *, new_bh )
+ __field( unsigned long long, newdi_blkno )
+ ),
+
+ TP_fast_assign(
+ __entry->new_blkno = new_blkno;
+ __entry->new_bh = new_bh;
+ __entry->newdi_blkno = newdi_blkno;
+ ),
+
+ TP_printk("%llu %p %llu", __entry->new_blkno, __entry->new_bh,
+ __entry->newdi_blkno)
+);
+
+DEFINE_EVENT(ocfs2__ull_ull_uint, ocfs2_create_symlink_data,
+
+ TP_PROTO(unsigned long long iblocks, unsigned long long i_size,
+ unsigned int blocks),
+
+ TP_ARGS(iblocks, i_size, blocks)
+);
+
+TRACE_EVENT(ocfs2_symlink_begin,
+
+ TP_PROTO(void *dir, void *dentry, const char *symname,
+ int len, const char *name),
+
+ TP_ARGS(dir, dentry, symname, len, name),
+
+ TP_STRUCT__entry(
+ __field( void *, dir )
+ __field( void *, dentry )
+ __field( const char *, symname )
+ __field( int, len )
+ __field( const char *, name )
+ ),
+
+ TP_fast_assign(
+ __entry->dir = dir;
+ __entry->dentry = dentry;
+ __entry->symname = symname;
+ __entry->len = len;
+ __entry->name = name;
+ ),
+
+ TP_printk("%p %p %s %.*s", __entry->dir, __entry->dentry,
+ __entry->symname, __entry->len, __entry->name)
+);
+
+DEFINE_EVENT(ocfs2__dentry_ops, ocfs2_symlink_create,
+
+ TP_PROTO(void *dir, void *dentry, int name_len, const char *name,
+ unsigned long long dir_blkno, unsigned long long mode),
+
+ TP_ARGS(dir, dentry, name_len, name, dir_blkno, mode)
+);
+
+TRACE_EVENT(ocfs2_blkno_stringify,
+
+ TP_PROTO(unsigned long long blkno, const char *name, int namelen),
+
+ TP_ARGS(blkno, name, namelen),
+
+ TP_STRUCT__entry(
+ __field( unsigned long long, blkno )
+ __field( const char *, name )
+ __field( int, namelen )
+ ),
+
+ TP_fast_assign(
+ __entry->blkno = blkno;
+ __entry->name = name;
+ __entry->namelen = namelen;
+ ),
+
+ TP_printk("%llu %s %d", __entry->blkno, __entry->name,
+ __entry->namelen)
+);
+
+DEFINE_EVENT(ocfs2__ull, ocfs2_orphan_add_begin,
+
+ TP_PROTO(unsigned long long ino),
+
+ TP_ARGS(ino)
+);
+
+DEFINE_EVENT(ocfs2__ull_uint, ocfs2_orphan_add_end,
+
+ TP_PROTO(unsigned long long ino, unsigned int slot),
+
+ TP_ARGS(ino, slot)
+);
+
+TRACE_EVENT(ocfs2_orphan_del,
+
+ TP_PROTO(unsigned long long dir, const char *name, int namelen),
+
+ TP_ARGS(dir, name, namelen),
+
+ TP_STRUCT__entry(
+ __field( unsigned long long, dir )
+ __field( const char *, name )
+ __field( int, namelen )
+ ),
+
+ TP_fast_assign(
+ __entry->dir = dir;
+ __entry->name = name;
+ __entry->namelen = namelen;
+ ),
+
+ TP_printk("%llu %s %d", __entry->dir, __entry->name,
+ __entry->namelen)
+);
+
+DEFINE_EVENT(ocfs2__dentry_ops, ocfs2_mv_orphaned_inode_to_new,
+
+ TP_PROTO(void *dir, void *dentry, int name_len, const char *name,
+ unsigned long long dir_blkno, unsigned long long ino),
+
+ TP_ARGS(dir, dentry, name_len, name, dir_blkno, ino)
+);
+/* End of trace events for fs/ocfs2/namei.c. */
#endif /* _TRACE_OCFS2_H */
/* This part must be outside protection */
--
1.7.0.4
^ permalink raw reply related [flat|nested] 53+ messages in thread
* [Ocfs2-devel] [PATCH 28/34] ocfs2: Remove masklog ML_DCACHE.
2010-12-23 7:19 [Ocfs2-devel] [PATCH 00/34] OCFS2: Add trace event and replace mlog(0) Tao Ma
` (26 preceding siblings ...)
2010-12-23 7:31 ` [Ocfs2-devel] [PATCH 27/34] ocfs2: Remove masklog ML_NAMEI Tao Ma
@ 2010-12-23 7:31 ` Tao Ma
2010-12-23 7:31 ` [Ocfs2-devel] [PATCH 29/34] ocfs2: Remove masklog ML_EXPORT Tao Ma
` (10 subsequent siblings)
38 siblings, 0 replies; 53+ messages in thread
From: Tao Ma @ 2010-12-23 7:31 UTC (permalink / raw)
To: ocfs2-devel
From: Tao Ma <boyu.mt@taobao.com>
Remove mlog(0) from fs/ocfs2/dcache.c and the masklog DCACHE.
Signed-off-by: Tao Ma <boyu.mt@taobao.com>
---
fs/ocfs2/cluster/masklog.c | 1 -
fs/ocfs2/cluster/masklog.h | 1 -
fs/ocfs2/dcache.c | 44 +++++++-------
fs/ocfs2/ocfs2_trace.h | 143 ++++++++++++++++++++++++++++++++++++++++++++
4 files changed, 165 insertions(+), 24 deletions(-)
diff --git a/fs/ocfs2/cluster/masklog.c b/fs/ocfs2/cluster/masklog.c
index 89cdbf8..15f6f03 100644
--- a/fs/ocfs2/cluster/masklog.c
+++ b/fs/ocfs2/cluster/masklog.c
@@ -97,7 +97,6 @@ static struct mlog_attribute mlog_attrs[MLOG_MAX_BITS] = {
define_mask(BH_IO),
define_mask(UPTODATE),
define_mask(VOTE),
- define_mask(DCACHE),
define_mask(CONN),
define_mask(QUORUM),
define_mask(EXPORT),
diff --git a/fs/ocfs2/cluster/masklog.h b/fs/ocfs2/cluster/masklog.h
index 8a8386f..7e76559 100644
--- a/fs/ocfs2/cluster/masklog.h
+++ b/fs/ocfs2/cluster/masklog.h
@@ -99,7 +99,6 @@
#define ML_BH_IO 0x0000000000100000ULL /* ocfs2 buffer I/O */
#define ML_UPTODATE 0x0000000000200000ULL /* ocfs2 caching sequence #'s */
#define ML_VOTE 0x0000000001000000ULL /* ocfs2 node messaging */
-#define ML_DCACHE 0x0000000002000000ULL /* ocfs2 dcache operations */
#define ML_CONN 0x0000000004000000ULL /* net connection management */
#define ML_QUORUM 0x0000000008000000ULL /* net connection quorum */
#define ML_EXPORT 0x0000000010000000ULL /* ocfs2 export operations */
diff --git a/fs/ocfs2/dcache.c b/fs/ocfs2/dcache.c
index 866d767..02e3998 100644
--- a/fs/ocfs2/dcache.c
+++ b/fs/ocfs2/dcache.c
@@ -28,7 +28,6 @@
#include <linux/slab.h>
#include <linux/namei.h>
-#define MLOG_MASK_PREFIX ML_DCACHE
#include <cluster/masklog.h>
#include "ocfs2.h"
@@ -39,6 +38,7 @@
#include "file.h"
#include "inode.h"
#include "super.h"
+#include "ocfs2_trace.h"
void ocfs2_dentry_attach_gen(struct dentry *dentry)
{
@@ -56,8 +56,8 @@ static int ocfs2_dentry_revalidate(struct dentry *dentry,
int ret = 0; /* if all else fails, just return false */
struct ocfs2_super *osb = OCFS2_SB(dentry->d_sb);
- mlog(0, "(0x%p, '%.*s')\n", dentry,
- dentry->d_name.len, dentry->d_name.name);
+ trace_ocfs2_dentry_revalidate(dentry, dentry->d_name.len,
+ dentry->d_name.name);
/* For a negative dentry -
* check the generation number of the parent and compare with the
@@ -67,9 +67,10 @@ static int ocfs2_dentry_revalidate(struct dentry *dentry,
unsigned long gen = (unsigned long) dentry->d_fsdata;
unsigned long pgen =
OCFS2_I(dentry->d_parent->d_inode)->ip_dir_lock_gen;
- mlog(0, "negative dentry: %.*s parent gen: %lu "
- "dentry gen: %lu\n",
- dentry->d_name.len, dentry->d_name.name, pgen, gen);
+
+ trace_ocfs2_dentry_revalidate_negative(dentry->d_name.len,
+ dentry->d_name.name,
+ pgen, gen);
if (gen != pgen)
goto bail;
goto valid;
@@ -84,8 +85,8 @@ static int ocfs2_dentry_revalidate(struct dentry *dentry,
/* did we or someone else delete this inode? */
if (OCFS2_I(inode)->ip_flags & OCFS2_INODE_DELETED) {
spin_unlock(&OCFS2_I(inode)->ip_lock);
- mlog(0, "inode (%llu) deleted, returning false\n",
- (unsigned long long)OCFS2_I(inode)->ip_blkno);
+ trace_ocfs2_dentry_revalidate_delete(
+ (unsigned long long)OCFS2_I(inode)->ip_blkno);
goto bail;
}
spin_unlock(&OCFS2_I(inode)->ip_lock);
@@ -95,10 +96,9 @@ static int ocfs2_dentry_revalidate(struct dentry *dentry,
* inode nlink hits zero, it never goes back.
*/
if (inode->i_nlink == 0) {
- mlog(0, "Inode %llu orphaned, returning false "
- "dir = %d\n",
- (unsigned long long)OCFS2_I(inode)->ip_blkno,
- S_ISDIR(inode->i_mode));
+ trace_ocfs2_dentry_revalidate_orphaned(
+ (unsigned long long)OCFS2_I(inode)->ip_blkno,
+ S_ISDIR(inode->i_mode));
goto bail;
}
@@ -107,9 +107,8 @@ static int ocfs2_dentry_revalidate(struct dentry *dentry,
* redo it.
*/
if (!dentry->d_fsdata) {
- mlog(0, "Inode %llu doesn't have dentry lock, "
- "returning false\n",
- (unsigned long long)OCFS2_I(inode)->ip_blkno);
+ trace_ocfs2_dentry_revalidate_nofsdata(
+ (unsigned long long)OCFS2_I(inode)->ip_blkno);
goto bail;
}
@@ -117,7 +116,7 @@ valid:
ret = 1;
bail:
- mlog(0, "ret = %d\n", ret);
+ trace_ocfs2_dentry_revalidate_ret(ret);
return ret;
}
@@ -174,8 +173,8 @@ struct dentry *ocfs2_find_local_alias(struct inode *inode,
dentry = list_entry(p, struct dentry, d_alias);
if (ocfs2_match_dentry(dentry, parent_blkno, skip_unhashed)) {
- mlog(0, "dentry found: %.*s\n",
- dentry->d_name.len, dentry->d_name.name);
+ trace_ocfs2_find_local_alias(dentry->d_name.len,
+ dentry->d_name.name);
dget_locked(dentry);
break;
@@ -231,9 +230,8 @@ int ocfs2_dentry_attach_lock(struct dentry *dentry,
struct dentry *alias;
struct ocfs2_dentry_lock *dl = dentry->d_fsdata;
- mlog(0, "Attach \"%.*s\", parent %llu, fsdata: %p\n",
- dentry->d_name.len, dentry->d_name.name,
- (unsigned long long)parent_blkno, dl);
+ trace_ocfs2_dentry_attach_lock(dentry->d_name.len, dentry->d_name.name,
+ (unsigned long long)parent_blkno, dl);
/*
* Negative dentry. We ignore these for now.
@@ -283,7 +281,9 @@ int ocfs2_dentry_attach_lock(struct dentry *dentry,
(unsigned long long)parent_blkno,
(unsigned long long)dl->dl_parent_blkno);
- mlog(0, "Found: %s\n", dl->dl_lockres.l_name);
+ trace_ocfs2_dentry_attach_lock_found(dl->dl_lockres.l_name,
+ (unsigned long long)parent_blkno,
+ (unsigned long long)OCFS2_I(inode)->ip_blkno);
goto out_attach;
}
diff --git a/fs/ocfs2/ocfs2_trace.h b/fs/ocfs2/ocfs2_trace.h
index fa424a6..61197a4 100644
--- a/fs/ocfs2/ocfs2_trace.h
+++ b/fs/ocfs2/ocfs2_trace.h
@@ -3658,6 +3658,149 @@ DEFINE_EVENT(ocfs2__dentry_ops, ocfs2_mv_orphaned_inode_to_new,
TP_ARGS(dir, dentry, name_len, name, dir_blkno, ino)
);
/* End of trace events for fs/ocfs2/namei.c. */
+
+/* Trace events for fs/ocfs2/dcache.c. */
+TRACE_EVENT(ocfs2_dentry_revalidate,
+
+ TP_PROTO(void *dentry, int len, const char *name),
+
+ TP_ARGS(dentry, len, name),
+
+ TP_STRUCT__entry(
+ __field( void *, dentry )
+ __field( int, len )
+ __field( const char *, name )
+ ),
+
+ TP_fast_assign(
+ __entry->dentry = dentry;
+ __entry->len = len;
+ __entry->name = name;
+ ),
+
+ TP_printk("%p %.*s", __entry->dentry, __entry->len, __entry->name)
+);
+
+TRACE_EVENT(ocfs2_dentry_revalidate_negative,
+
+ TP_PROTO(int len, const char *name, unsigned long pgen,
+ unsigned long gen),
+
+ TP_ARGS(len, name, pgen, gen),
+
+ TP_STRUCT__entry(
+ __field( int, len )
+ __field( const char *, name )
+ __field( unsigned long, pgen )
+ __field( unsigned long, gen )
+ ),
+
+ TP_fast_assign(
+ __entry->len = len;
+ __entry->name = name;
+ __entry->pgen = pgen;
+ __entry->gen = gen;
+ ),
+
+ TP_printk("%.*s %lu %lu", __entry->len, __entry->name,
+ __entry->pgen, __entry->gen)
+);
+
+DEFINE_EVENT(ocfs2__ull, ocfs2_dentry_revalidate_delete,
+
+ TP_PROTO(unsigned long long ino),
+
+ TP_ARGS(ino)
+);
+
+DEFINE_EVENT(ocfs2__ull_int, ocfs2_dentry_revalidate_orphaned,
+
+ TP_PROTO(unsigned long long ino, int dir),
+
+ TP_ARGS(ino, dir)
+);
+
+DEFINE_EVENT(ocfs2__ull, ocfs2_dentry_revalidate_nofsdata,
+
+ TP_PROTO(unsigned long long ino),
+
+ TP_ARGS(ino)
+);
+
+DEFINE_EVENT(ocfs2__int, ocfs2_dentry_revalidate_ret,
+
+ TP_PROTO(int ret),
+
+ TP_ARGS(ret)
+);
+
+TRACE_EVENT(ocfs2_find_local_alias,
+
+ TP_PROTO(int len, const char *name),
+
+ TP_ARGS(len, name),
+
+ TP_STRUCT__entry(
+ __field( int, len )
+ __field( const char *, name )
+ ),
+
+ TP_fast_assign(
+ __entry->len = len;
+ __entry->name = name;
+ ),
+
+ TP_printk("%.*s", __entry->len, __entry->name)
+);
+
+TRACE_EVENT(ocfs2_dentry_attach_lock,
+
+ TP_PROTO(int len, const char *name,
+ unsigned long long parent, void *fsdata),
+
+ TP_ARGS(len, name, parent, fsdata),
+
+ TP_STRUCT__entry(
+ __field( int, len )
+ __field( const char *, name )
+ __field( unsigned long long, parent )
+ __field( void *, fsdata )
+ ),
+
+ TP_fast_assign(
+ __entry->len = len;
+ __entry->name = name;
+ __entry->parent = parent;
+ __entry->fsdata = fsdata;
+ ),
+
+ TP_printk("%.*s %llu %p", __entry->len, __entry->name,
+ __entry->parent, __entry->fsdata)
+);
+
+TRACE_EVENT(ocfs2_dentry_attach_lock_found,
+
+ TP_PROTO(const char *name, unsigned long long parent,
+ unsigned long long ino),
+
+ TP_ARGS(name, parent, ino),
+
+ TP_STRUCT__entry(
+ __field( const char *, name )
+ __field( unsigned long long, parent )
+ __field( unsigned long long, ino )
+ ),
+
+ TP_fast_assign(
+ __entry->name = name;
+ __entry->parent = parent;
+ __entry->ino = ino;
+ ),
+
+ TP_printk("%s %llu %llu", __entry->name, __entry->parent, __entry->ino)
+);
+
+/* End of trace events for fs/ocfs2/dcache.c. */
#endif /* _TRACE_OCFS2_H */
/* This part must be outside protection */
--
1.7.0.4
^ permalink raw reply related [flat|nested] 53+ messages in thread
* [Ocfs2-devel] [PATCH 29/34] ocfs2: Remove masklog ML_EXPORT.
2010-12-23 7:19 [Ocfs2-devel] [PATCH 00/34] OCFS2: Add trace event and replace mlog(0) Tao Ma
` (27 preceding siblings ...)
2010-12-23 7:31 ` [Ocfs2-devel] [PATCH 28/34] ocfs2: Remove masklog ML_DCACHE Tao Ma
@ 2010-12-23 7:31 ` Tao Ma
2010-12-23 7:31 ` [Ocfs2-devel] [PATCH 30/34] ocfs2: Remove masklog ML_JOURNAL Tao Ma
` (9 subsequent siblings)
38 siblings, 0 replies; 53+ messages in thread
From: Tao Ma @ 2010-12-23 7:31 UTC (permalink / raw)
To: ocfs2-devel
From: Tao Ma <boyu.mt@taobao.com>
Remove mlog(0) from fs/ocfs2/export.c and the masklog EXPORT.
Signed-off-by: Tao Ma <boyu.mt@taobao.com>
---
fs/ocfs2/cluster/masklog.c | 1 -
fs/ocfs2/cluster/masklog.h | 1 -
fs/ocfs2/export.c | 47 +++++++---------
fs/ocfs2/ocfs2_trace.h | 135 ++++++++++++++++++++++++++++++++++++++++++++
4 files changed, 155 insertions(+), 29 deletions(-)
diff --git a/fs/ocfs2/cluster/masklog.c b/fs/ocfs2/cluster/masklog.c
index 15f6f03..a229ede 100644
--- a/fs/ocfs2/cluster/masklog.c
+++ b/fs/ocfs2/cluster/masklog.c
@@ -99,7 +99,6 @@ static struct mlog_attribute mlog_attrs[MLOG_MAX_BITS] = {
define_mask(VOTE),
define_mask(CONN),
define_mask(QUORUM),
- define_mask(EXPORT),
define_mask(BASTS),
define_mask(ERROR),
define_mask(NOTICE),
diff --git a/fs/ocfs2/cluster/masklog.h b/fs/ocfs2/cluster/masklog.h
index 7e76559..5d358c1 100644
--- a/fs/ocfs2/cluster/masklog.h
+++ b/fs/ocfs2/cluster/masklog.h
@@ -101,7 +101,6 @@
#define ML_VOTE 0x0000000001000000ULL /* ocfs2 node messaging */
#define ML_CONN 0x0000000004000000ULL /* net connection management */
#define ML_QUORUM 0x0000000008000000ULL /* net connection quorum */
-#define ML_EXPORT 0x0000000010000000ULL /* ocfs2 export operations */
#define ML_BASTS 0x0000001000000000ULL /* dlmglue asts and basts */
/* bits that are infrequently given and frequently matched in the high word */
#define ML_ERROR 0x0000000100000000ULL /* sent to KERN_ERR */
diff --git a/fs/ocfs2/export.c b/fs/ocfs2/export.c
index 190158b..c01aee5 100644
--- a/fs/ocfs2/export.c
+++ b/fs/ocfs2/export.c
@@ -26,7 +26,6 @@
#include <linux/fs.h>
#include <linux/types.h>
-#define MLOG_MASK_PREFIX ML_EXPORT
#include <cluster/masklog.h>
#include "ocfs2.h"
@@ -40,6 +39,7 @@
#include "buffer_head_io.h"
#include "suballoc.h"
+#include "ocfs2_trace.h"
struct ocfs2_inode_handle
{
@@ -56,10 +56,9 @@ static struct dentry *ocfs2_get_dentry(struct super_block *sb,
int status, set;
struct dentry *result;
- mlog(0, "(0x%p, 0x%p)\n", sb, handle);
+ trace_ocfs2_get_dentry_begin(sb, handle, (unsigned long long)blkno);
if (blkno == 0) {
- mlog(0, "nfs wants inode with blkno: 0\n");
result = ERR_PTR(-ESTALE);
goto bail;
}
@@ -83,6 +82,7 @@ static struct dentry *ocfs2_get_dentry(struct super_block *sb,
}
status = ocfs2_test_inode_bit(osb, blkno, &set);
+ trace_ocfs2_get_dentry_test_bit(status, set);
if (status < 0) {
if (status == -EINVAL) {
/*
@@ -90,18 +90,14 @@ static struct dentry *ocfs2_get_dentry(struct super_block *sb,
* as an inode, we return -ESTALE to be
* nice
*/
- mlog(0, "test inode bit failed %d\n", status);
status = -ESTALE;
- } else {
+ } else
mlog(ML_ERROR, "test inode bit failed %d\n", status);
- }
goto unlock_nfs_sync;
}
/* If the inode allocator bit is clear, this inode must be stale */
if (!set) {
- mlog(0, "inode %llu suballoc bit is clear\n",
- (unsigned long long)blkno);
status = -ESTALE;
goto unlock_nfs_sync;
}
@@ -114,8 +110,8 @@ unlock_nfs_sync:
check_err:
if (status < 0) {
if (status == -ESTALE) {
- mlog(0, "stale inode ino: %llu generation: %u\n",
- (unsigned long long)blkno, handle->ih_generation);
+ trace_ocfs2_get_dentry_stale((unsigned long long)blkno,
+ handle->ih_generation);
}
result = ERR_PTR(status);
goto bail;
@@ -130,8 +126,9 @@ check_err:
check_gen:
if (handle->ih_generation != inode->i_generation) {
iput(inode);
- mlog(0, "stale inode ino: %llu generation: %u\n",
- (unsigned long long)blkno, handle->ih_generation);
+ trace_ocfs2_get_dentry_generation((unsigned long long)blkno,
+ handle->ih_generation,
+ inode->i_generation);
result = ERR_PTR(-ESTALE);
goto bail;
}
@@ -143,7 +140,7 @@ check_gen:
mlog_errno(PTR_ERR(result));
bail:
- mlog(0, "result = %p\n", result);
+ trace_ocfs2_get_dentry_end(result);
return result;
}
@@ -154,11 +151,8 @@ static struct dentry *ocfs2_get_parent(struct dentry *child)
struct dentry *parent;
struct inode *dir = child->d_inode;
- mlog(0, "(0x%p, '%.*s')\n", child,
- child->d_name.len, child->d_name.name);
-
- mlog(0, "find parent of directory %llu\n",
- (unsigned long long)OCFS2_I(dir)->ip_blkno);
+ trace_ocfs2_get_parent(child, child->d_name.len, child->d_name.name,
+ (unsigned long long)OCFS2_I(dir)->ip_blkno);
status = ocfs2_inode_lock(dir, NULL, 0);
if (status < 0) {
@@ -182,7 +176,7 @@ bail_unlock:
ocfs2_inode_unlock(dir, 0);
bail:
- mlog(0, "parent %p\n", parent);
+ trace_ocfs2_get_parent_end(parent);
return parent;
}
@@ -197,9 +191,9 @@ static int ocfs2_encode_fh(struct dentry *dentry, u32 *fh_in, int *max_len,
u32 generation;
__le32 *fh = (__force __le32 *) fh_in;
- mlog(0, "(0x%p, '%.*s', 0x%p, %d, %d)\n", dentry,
- dentry->d_name.len, dentry->d_name.name,
- fh, len, connectable);
+ trace_ocfs2_encode_fh_begin(dentry, dentry->d_name.len,
+ dentry->d_name.name,
+ fh, len, connectable);
if (len < 3 || (connectable && len < 6)) {
mlog(ML_ERROR, "fh buffer is too small for encoding\n");
@@ -210,8 +204,7 @@ static int ocfs2_encode_fh(struct dentry *dentry, u32 *fh_in, int *max_len,
blkno = OCFS2_I(inode)->ip_blkno;
generation = inode->i_generation;
- mlog(0, "Encoding fh: blkno: %llu, generation: %u\n",
- (unsigned long long)blkno, generation);
+ trace_ocfs2_encode_fh_self((unsigned long long)blkno, generation);
len = 3;
fh[0] = cpu_to_le32((u32)(blkno >> 32));
@@ -236,14 +229,14 @@ static int ocfs2_encode_fh(struct dentry *dentry, u32 *fh_in, int *max_len,
len = 6;
type = 2;
- mlog(0, "Encoding parent: blkno: %llu, generation: %u\n",
- (unsigned long long)blkno, generation);
+ trace_ocfs2_encode_fh_parent((unsigned long long)blkno,
+ generation);
}
*max_len = len;
bail:
- mlog(0, "type = %d\n", type);
+ trace_ocfs2_encode_fh_type(type);
return type;
}
diff --git a/fs/ocfs2/ocfs2_trace.h b/fs/ocfs2/ocfs2_trace.h
index 61197a4..a58008f 100644
--- a/fs/ocfs2/ocfs2_trace.h
+++ b/fs/ocfs2/ocfs2_trace.h
@@ -3800,6 +3800,141 @@ TRACE_EVENT(ocfs2_dentry_attach_lock_found,
TP_printk("%s %llu %llu", __entry->name, __entry->parent, __entry->ino)
);
+TRACE_EVENT(ocfs2_get_dentry_begin,
+
+ TP_PROTO(void *sb, void *handle, unsigned long long blkno),
+
+ TP_ARGS(sb, handle, blkno),
+
+ TP_STRUCT__entry(
+ __field( void *, sb )
+ __field( void *, handle )
+ __field( unsigned long long, blkno )
+ ),
+
+ TP_fast_assign(
+ __entry->sb = sb;
+ __entry->handle = handle;
+ __entry->blkno = blkno;
+ ),
+
+ TP_printk("%p %p %llu", __entry->sb, __entry->handle, __entry->blkno)
+);
+
+DEFINE_EVENT(ocfs2__int_int, ocfs2_get_dentry_test_bit,
+
+ TP_PROTO(int status, int set),
+
+ TP_ARGS(status, set)
+);
+
+DEFINE_EVENT(ocfs2__ull_uint, ocfs2_get_dentry_stale,
+
+ TP_PROTO(unsigned long long blkno, unsigned int generation),
+
+ TP_ARGS(blkno, generation)
+);
+
+DEFINE_EVENT(ocfs2__ull_uint_uint, ocfs2_get_dentry_generation,
+
+ TP_PROTO(unsigned long long blkno, unsigned int h_generation,
+ unsigned int i_generation),
+
+ TP_ARGS(blkno, h_generation, i_generation)
+);
+
+DEFINE_EVENT(ocfs2__pointer, ocfs2_get_dentry_end,
+
+
+ TP_PROTO(void *result),
+
+ TP_ARGS(result)
+);
+
+TRACE_EVENT(ocfs2_get_parent,
+
+ TP_PROTO(void *child, int len, const char *name,
+ unsigned long long ino),
+
+ TP_ARGS(child, len, name, ino),
+
+ TP_STRUCT__entry(
+ __field( void *, child )
+ __field( int, len )
+ __field( const char *, name )
+ __field( unsigned long long, ino )
+ ),
+
+ TP_fast_assign(
+ __entry->child = child;
+ __entry->len = len;
+ __entry->name = name;
+ __entry->ino = ino;
+ ),
+
+ TP_printk("%p %.*s %llu", __entry->child, __entry->len,
+ __entry->name, __entry->ino)
+);
+
+DEFINE_EVENT(ocfs2__pointer, ocfs2_get_parent_end,
+
+
+ TP_PROTO(void *parent),
+
+ TP_ARGS(parent)
+);
+
+TRACE_EVENT(ocfs2_encode_fh_begin,
+
+ TP_PROTO(void *dentry, int name_len, const char *name,
+ void *fh, int len, int connectable),
+
+ TP_ARGS(dentry, name_len, name, fh, len, connectable),
+
+ TP_STRUCT__entry(
+ __field( void *, dentry )
+ __field( int, name_len )
+ __field( const char *, name )
+ __field( void *, fh )
+ __field( int, len )
+ __field( int, connectable )
+ ),
+
+ TP_fast_assign(
+ __entry->dentry = dentry;
+ __entry->name_len = name_len;
+ __entry->name = name;
+ __entry->fh = fh;
+ __entry->len = len;
+ __entry->connectable = connectable;
+ ),
+
+ TP_printk("%p %.*s %p %d %d", __entry->dentry, __entry->name_len,
+ __entry->name, __entry->fh, __entry->len,
+ __entry->connectable)
+);
+
+DEFINE_EVENT(ocfs2__ull_uint, ocfs2_encode_fh_self,
+
+ TP_PROTO(unsigned long long blkno, unsigned int generation),
+
+ TP_ARGS(blkno, generation)
+);
+
+DEFINE_EVENT(ocfs2__ull_uint, ocfs2_encode_fh_parent,
+
+ TP_PROTO(unsigned long long blkno, unsigned int generation),
+
+ TP_ARGS(blkno, generation)
+);
+
+DEFINE_EVENT(ocfs2__int, ocfs2_encode_fh_type,
+
+ TP_PROTO(int type),
+
+ TP_ARGS(type)
+);
+
/* End of trace events for fs/ocfs2/dcache.c. */
#endif /* _TRACE_OCFS2_H */
--
1.7.0.4
^ permalink raw reply related [flat|nested] 53+ messages in thread
* [Ocfs2-devel] [PATCH 30/34] ocfs2: Remove masklog ML_JOURNAL.
2010-12-23 7:19 [Ocfs2-devel] [PATCH 00/34] OCFS2: Add trace event and replace mlog(0) Tao Ma
` (28 preceding siblings ...)
2010-12-23 7:31 ` [Ocfs2-devel] [PATCH 29/34] ocfs2: Remove masklog ML_EXPORT Tao Ma
@ 2010-12-23 7:31 ` Tao Ma
2010-12-23 7:31 ` [Ocfs2-devel] [PATCH 31/34] ocfs2: Remove masklog ML_BH_IO Tao Ma
` (8 subsequent siblings)
38 siblings, 0 replies; 53+ messages in thread
From: Tao Ma @ 2010-12-23 7:31 UTC (permalink / raw)
To: ocfs2-devel
From: Tao Ma <boyu.mt@taobao.com>
Remove mlog(0) from fs/ocfs2/journal.c and the masklog JOURNAL.
Signed-off-by: Tao Ma <boyu.mt@taobao.com>
---
fs/ocfs2/cluster/masklog.c | 1 -
fs/ocfs2/cluster/masklog.h | 1 -
fs/ocfs2/journal.c | 130 ++++++++++--------------
fs/ocfs2/ocfs2_trace.h | 241 ++++++++++++++++++++++++++++++++++++++++++++
4 files changed, 295 insertions(+), 78 deletions(-)
diff --git a/fs/ocfs2/cluster/masklog.c b/fs/ocfs2/cluster/masklog.c
index a229ede..3f1e995 100644
--- a/fs/ocfs2/cluster/masklog.c
+++ b/fs/ocfs2/cluster/masklog.c
@@ -92,7 +92,6 @@ static struct mlog_attribute mlog_attrs[MLOG_MAX_BITS] = {
define_mask(DLM_MASTER),
define_mask(DLM_RECOVERY),
define_mask(AIO),
- define_mask(JOURNAL),
define_mask(DLM_GLUE),
define_mask(BH_IO),
define_mask(UPTODATE),
diff --git a/fs/ocfs2/cluster/masklog.h b/fs/ocfs2/cluster/masklog.h
index 5d358c1..61e8f6c 100644
--- a/fs/ocfs2/cluster/masklog.h
+++ b/fs/ocfs2/cluster/masklog.h
@@ -94,7 +94,6 @@
#define ML_DLM_MASTER 0x0000000000000800ULL /* dlm master functions */
#define ML_DLM_RECOVERY 0x0000000000001000ULL /* dlm master functions */
#define ML_AIO 0x0000000000002000ULL /* ocfs2 aio read and write */
-#define ML_JOURNAL 0x0000000000004000ULL /* ocfs2 journalling functions */
#define ML_DLM_GLUE 0x0000000000080000ULL /* ocfs2 dlm glue layer */
#define ML_BH_IO 0x0000000000100000ULL /* ocfs2 buffer I/O */
#define ML_UPTODATE 0x0000000000200000ULL /* ocfs2 caching sequence #'s */
diff --git a/fs/ocfs2/journal.c b/fs/ocfs2/journal.c
index 9565091..5d3c077 100644
--- a/fs/ocfs2/journal.c
+++ b/fs/ocfs2/journal.c
@@ -31,7 +31,6 @@
#include <linux/time.h>
#include <linux/random.h>
-#define MLOG_MASK_PREFIX ML_JOURNAL
#include <cluster/masklog.h>
#include "ocfs2.h"
@@ -52,6 +51,7 @@
#include "quota.h"
#include "buffer_head_io.h"
+#include "ocfs2_trace.h"
DEFINE_SPINLOCK(trans_inc_lock);
@@ -308,9 +308,10 @@ static int ocfs2_commit_cache(struct ocfs2_super *osb)
/* Flush all pending commits and checkpoint the journal. */
down_write(&journal->j_trans_barrier);
- if (atomic_read(&journal->j_num_trans) == 0) {
+ flushed = atomic_read(&journal->j_num_trans);
+ trace_ocfs2_commit_cache_begin(flushed);
+ if (flushed == 0) {
up_write(&journal->j_trans_barrier);
- mlog(0, "No transactions for me to flush!\n");
goto finally;
}
@@ -329,8 +330,7 @@ static int ocfs2_commit_cache(struct ocfs2_super *osb)
atomic_set(&journal->j_num_trans, 0);
up_write(&journal->j_trans_barrier);
- mlog(0, "commit_thread: flushed transaction %lu (%u handles)\n",
- journal->j_trans_id, flushed);
+ trace_ocfs2_commit_cache_end(journal->j_trans_id, flushed);
ocfs2_wake_downconvert_thread(osb);
wake_up(&journal->j_checkpointed);
@@ -423,7 +423,7 @@ int ocfs2_extend_trans(handle_t *handle, int nblocks)
old_nblocks = handle->h_buffer_credits;
- mlog(0, "Trying to extend transaction by %d blocks\n", nblocks);
+ trace_ocfs2_extend_trans(old_nblocks, nblocks);
#ifdef CONFIG_OCFS2_DEBUG_FS
status = 1;
@@ -436,9 +436,7 @@ int ocfs2_extend_trans(handle_t *handle, int nblocks)
#endif
if (status > 0) {
- mlog(0,
- "jbd2_journal_extend failed, trying "
- "jbd2_journal_restart\n");
+ trace_ocfs2_extend_trans_restart(old_nblocks + nblocks);
status = jbd2_journal_restart(handle,
old_nblocks + nblocks);
if (status < 0) {
@@ -616,12 +614,9 @@ static int __ocfs2_journal_access(handle_t *handle,
BUG_ON(!handle);
BUG_ON(!bh);
- mlog(0, "bh->b_blocknr=%llu, type=%d (\"%s\"), bh->b_size = %zu\n",
- (unsigned long long)bh->b_blocknr, type,
- (type == OCFS2_JOURNAL_ACCESS_CREATE) ?
- "OCFS2_JOURNAL_ACCESS_CREATE" :
- "OCFS2_JOURNAL_ACCESS_WRITE",
- bh->b_size);
+ trace_ocfs2_journal_access(
+ (unsigned long long)ocfs2_metadata_cache_owner(ci),
+ (unsigned long long)bh->b_blocknr, type, bh->b_size);
/* we can safely remove this assertion after testing. */
if (!buffer_uptodate(bh)) {
@@ -730,8 +725,7 @@ void ocfs2_journal_dirty(handle_t *handle, struct buffer_head *bh)
{
int status;
- mlog(0, "(bh->b_blocknr=%llu)\n",
- (unsigned long long)bh->b_blocknr);
+ trace_ocfs2_journal_dirty((unsigned long long)bh->b_blocknr);
status = jbd2_journal_dirty_metadata(handle, bh);
BUG_ON(status);
@@ -809,10 +803,9 @@ int ocfs2_journal_init(struct ocfs2_journal *journal, int *dirty)
goto done;
}
- mlog(0, "inode->i_size = %lld\n", inode->i_size);
- mlog(0, "inode->i_blocks = %llu\n",
- (unsigned long long)inode->i_blocks);
- mlog(0, "inode->ip_clusters = %u\n", OCFS2_I(inode)->ip_clusters);
+ trace_ocfs2_journal_init(inode->i_size,
+ (unsigned long long)inode->i_blocks,
+ OCFS2_I(inode)->ip_clusters);
/* call the kernels journal init function now */
j_journal = jbd2_journal_init_inode(inode);
@@ -822,8 +815,7 @@ int ocfs2_journal_init(struct ocfs2_journal *journal, int *dirty)
goto done;
}
- mlog(0, "Returned from jbd2_journal_init_inode\n");
- mlog(0, "j_journal->j_maxlen = %u\n", j_journal->j_maxlen);
+ trace_ocfs2_journal_init_maxlen(j_journal->j_maxlen);
*dirty = (le32_to_cpu(di->id1.journal1.ij_flags) &
OCFS2_JOURNAL_DIRTY_FL);
@@ -922,10 +914,7 @@ void ocfs2_journal_shutdown(struct ocfs2_super *osb)
BUG();
num_running_trans = atomic_read(&(osb->journal->j_num_trans));
- if (num_running_trans > 0)
- mlog(0, "Shutting down journal: must wait on %d "
- "running transactions!\n",
- num_running_trans);
+ trace_ocfs2_journal_shutdown(num_running_trans);
/* Do a commit_cache here. It will flush our journal, *and*
* release any locks that are still held.
@@ -938,7 +927,7 @@ void ocfs2_journal_shutdown(struct ocfs2_super *osb)
* completely destroy the journal. */
if (osb->commit_task) {
/* Wait for the commit thread */
- mlog(0, "Waiting for ocfs2commit to exit....\n");
+ trace_ocfs2_journal_shutdown_wait(osb->commit_task);
kthread_stop(osb->commit_task);
osb->commit_task = NULL;
}
@@ -1158,7 +1147,7 @@ struct ocfs2_la_recovery_item {
*/
void ocfs2_complete_recovery(struct work_struct *work)
{
- int ret;
+ int ret = 0;
struct ocfs2_journal *journal =
container_of(work, struct ocfs2_journal, j_recovery_work);
struct ocfs2_super *osb = journal->j_osb;
@@ -1167,7 +1156,8 @@ void ocfs2_complete_recovery(struct work_struct *work)
struct ocfs2_quota_recovery *qrec;
LIST_HEAD(tmp_la_list);
- mlog(0, "completing recovery from keventd\n");
+ trace_ocfs2_complete_recovery(
+ (unsigned long long)OCFS2_I(journal->j_inode)->ip_blkno);
spin_lock(&journal->j_lock);
list_splice_init(&journal->j_la_cleanups, &tmp_la_list);
@@ -1176,15 +1166,18 @@ void ocfs2_complete_recovery(struct work_struct *work)
list_for_each_entry_safe(item, n, &tmp_la_list, lri_list) {
list_del_init(&item->lri_list);
- mlog(0, "Complete recovery for slot %d\n", item->lri_slot);
-
ocfs2_wait_on_quotas(osb);
la_dinode = item->lri_la_dinode;
- if (la_dinode) {
- mlog(0, "Clean up local alloc %llu\n",
- (unsigned long long)le64_to_cpu(la_dinode->i_blkno));
+ tl_dinode = item->lri_tl_dinode;
+ qrec = item->lri_qrec;
+ trace_ocfs2_complete_recovery_slot(item->lri_slot,
+ la_dinode ? le64_to_cpu(la_dinode->i_blkno) : 0,
+ tl_dinode ? le64_to_cpu(tl_dinode->i_blkno) : 0,
+ qrec);
+
+ if (la_dinode) {
ret = ocfs2_complete_local_alloc_recovery(osb,
la_dinode);
if (ret < 0)
@@ -1193,11 +1186,7 @@ void ocfs2_complete_recovery(struct work_struct *work)
kfree(la_dinode);
}
- tl_dinode = item->lri_tl_dinode;
if (tl_dinode) {
- mlog(0, "Clean up truncate log %llu\n",
- (unsigned long long)le64_to_cpu(tl_dinode->i_blkno));
-
ret = ocfs2_complete_truncate_log_recovery(osb,
tl_dinode);
if (ret < 0)
@@ -1210,9 +1199,7 @@ void ocfs2_complete_recovery(struct work_struct *work)
if (ret < 0)
mlog_errno(ret);
- qrec = item->lri_qrec;
if (qrec) {
- mlog(0, "Recovering quota files");
ret = ocfs2_finish_quota_recovery(osb, qrec,
item->lri_slot);
if (ret < 0)
@@ -1223,7 +1210,7 @@ void ocfs2_complete_recovery(struct work_struct *work)
kfree(item);
}
- mlog(0, "Recovery completion\n");
+ trace_ocfs2_complete_recovery_end(ret);
}
/* NOTE: This function always eats your references to la_dinode and
@@ -1340,15 +1327,12 @@ restart:
* clear it until ocfs2_recover_node() has succeeded. */
node_num = rm->rm_entries[0];
spin_unlock(&osb->osb_lock);
- mlog(0, "checking node %d\n", node_num);
slot_num = ocfs2_node_num_to_slot(osb, node_num);
+ trace_ocfs2_recovery_thread_node(node_num, slot_num);
if (slot_num == -ENOENT) {
status = 0;
- mlog(0, "no slot for this node, so no recovery"
- "required.\n");
goto skip_recovery;
}
- mlog(0, "node %d was using slot %d\n", node_num, slot_num);
/* It is a bit subtle with quota recovery. We cannot do it
* immediately because we have to obtain cluster locks from
@@ -1375,7 +1359,7 @@ skip_recovery:
spin_lock(&osb->osb_lock);
}
spin_unlock(&osb->osb_lock);
- mlog(0, "All nodes recovered\n");
+ trace_ocfs2_recovery_thread_end(status);
/* Refresh all journal recovery generations from disk */
status = ocfs2_check_journals_nolocks(osb);
@@ -1428,19 +1412,15 @@ bail:
void ocfs2_recovery_thread(struct ocfs2_super *osb, int node_num)
{
- mlog(0, "(node_num=%d, osb->node_num = %d)\n",
- node_num, osb->node_num);
-
mutex_lock(&osb->recovery_lock);
- if (osb->disable_recovery)
- goto out;
- /* People waiting on recovery will wait on
- * the recovery map to empty. */
- if (ocfs2_recovery_map_set(osb, node_num))
- mlog(0, "node %d already in recovery map.\n", node_num);
+ trace_ocfs2_recovery_thread(node_num, osb->node_num,
+ osb->disable_recovery, osb->recovery_thread_task,
+ osb->disable_recovery ?
+ -1 : ocfs2_recovery_map_set(osb, node_num));
- mlog(0, "starting recovery thread...\n");
+ if (osb->disable_recovery)
+ goto out;
if (osb->recovery_thread_task)
goto out;
@@ -1528,7 +1508,7 @@ static int ocfs2_replay_journal(struct ocfs2_super *osb,
* If not, it needs recovery.
*/
if (osb->slot_recovery_generations[slot_num] != slot_reco_gen) {
- mlog(0, "Slot %u already recovered (old/new=%u/%u)\n", slot_num,
+ trace_ocfs2_replay_journal_recovered(slot_num,
osb->slot_recovery_generations[slot_num], slot_reco_gen);
osb->slot_recovery_generations[slot_num] = slot_reco_gen;
status = -EBUSY;
@@ -1539,7 +1519,7 @@ static int ocfs2_replay_journal(struct ocfs2_super *osb,
status = ocfs2_inode_lock_full(inode, &bh, 1, OCFS2_META_LOCK_RECOVERY);
if (status < 0) {
- mlog(0, "status returned from ocfs2_inode_lock=%d\n", status);
+ trace_ocfs2_replay_journal_lock_err(status);
if (status != -ERESTARTSYS)
mlog(ML_ERROR, "Could not lock journal!\n");
goto done;
@@ -1552,7 +1532,7 @@ static int ocfs2_replay_journal(struct ocfs2_super *osb,
slot_reco_gen = ocfs2_get_recovery_generation(fe);
if (!(flags & OCFS2_JOURNAL_DIRTY_FL)) {
- mlog(0, "No recovery required for node %d\n", node_num);
+ trace_ocfs2_replay_journal_skip(node_num);
/* Refresh recovery generation for the slot */
osb->slot_recovery_generations[slot_num] = slot_reco_gen;
goto done;
@@ -1573,7 +1553,6 @@ static int ocfs2_replay_journal(struct ocfs2_super *osb,
goto done;
}
- mlog(0, "calling journal_init_inode\n");
journal = jbd2_journal_init_inode(inode);
if (journal == NULL) {
mlog(ML_ERROR, "Linux journal layer error\n");
@@ -1593,7 +1572,6 @@ static int ocfs2_replay_journal(struct ocfs2_super *osb,
ocfs2_clear_journal_error(osb->sb, journal, slot_num);
/* wipe the journal */
- mlog(0, "flushing the journal.\n");
jbd2_journal_lock_updates(journal);
status = jbd2_journal_flush(journal);
jbd2_journal_unlock_updates(journal);
@@ -1652,8 +1630,7 @@ static int ocfs2_recover_node(struct ocfs2_super *osb,
struct ocfs2_dinode *la_copy = NULL;
struct ocfs2_dinode *tl_copy = NULL;
- mlog(0, "(node_num=%d, slot_num=%d, osb->node_num = %d)\n",
- node_num, slot_num, osb->node_num);
+ trace_ocfs2_recover_node(node_num, slot_num, osb->node_num);
/* Should not ever be called to recover ourselves -- in that
* case we should've called ocfs2_journal_load instead. */
@@ -1662,9 +1639,7 @@ static int ocfs2_recover_node(struct ocfs2_super *osb,
status = ocfs2_replay_journal(osb, node_num, slot_num);
if (status < 0) {
if (status == -EBUSY) {
- mlog(0, "Skipping recovery for slot %u (node %u) "
- "as another node has recovered it\n", slot_num,
- node_num);
+ trace_ocfs2_recover_node_skip(slot_num, node_num);
status = 0;
goto done;
}
@@ -1771,8 +1746,8 @@ int ocfs2_mark_dead_nodes(struct ocfs2_super *osb)
spin_lock(&osb->osb_lock);
osb->slot_recovery_generations[i] = gen;
- mlog(0, "Slot %u recovery generation is %u\n", i,
- osb->slot_recovery_generations[i]);
+ trace_ocfs2_mark_dead_nodes(i,
+ osb->slot_recovery_generations[i]);
if (i == osb->slot_num) {
spin_unlock(&osb->osb_lock);
@@ -1846,11 +1821,12 @@ void ocfs2_queue_orphan_scan(struct ocfs2_super *osb)
os = &osb->osb_orphan_scan;
- mlog(0, "Begin orphan scan\n");
-
if (atomic_read(&os->os_state) == ORPHAN_SCAN_INACTIVE)
goto out;
+ trace_ocfs2_queue_orphan_scan_begin(os->os_count, os->os_seqno,
+ atomic_read(&os->os_state));
+
status = ocfs2_orphan_scan_lock(osb, &seqno);
if (status < 0) {
if (status != -EAGAIN)
@@ -1880,7 +1856,8 @@ void ocfs2_queue_orphan_scan(struct ocfs2_super *osb)
unlock:
ocfs2_orphan_scan_unlock(osb, seqno);
out:
- mlog(0, "Orphan scan completed\n");
+ trace_ocfs2_queue_orphan_scan_end(os->os_count, os->os_seqno,
+ atomic_read(&os->os_state));
return;
}
@@ -1964,8 +1941,7 @@ static int ocfs2_orphan_filldir(void *priv, const char *name, int name_len,
if (IS_ERR(iter))
return 0;
- mlog(0, "queue orphan %llu\n",
- (unsigned long long)OCFS2_I(iter)->ip_blkno);
+ trace_ocfs2_orphan_filldir((unsigned long long)OCFS2_I(iter)->ip_blkno);
/* No locking is required for the next_orphan queue as there
* is only ever a single process doing orphan recovery. */
OCFS2_I(iter)->ip_next_orphan = p->head;
@@ -2081,7 +2057,7 @@ static int ocfs2_recover_orphans(struct ocfs2_super *osb,
struct inode *iter;
struct ocfs2_inode_info *oi;
- mlog(0, "Recover inodes from orphan dir in slot %d\n", slot);
+ trace_ocfs2_recover_orphans(slot);
ocfs2_mark_recovering_orphan_dir(osb, slot);
ret = ocfs2_queue_orphans(osb, slot, &inode);
@@ -2094,7 +2070,8 @@ static int ocfs2_recover_orphans(struct ocfs2_super *osb,
while (inode) {
oi = OCFS2_I(inode);
- mlog(0, "iput orphan %llu\n", (unsigned long long)oi->ip_blkno);
+ trace_ocfs2_recover_orphans_iput(
+ (unsigned long long)oi->ip_blkno);
iter = oi->ip_next_orphan;
@@ -2132,6 +2109,7 @@ static int __ocfs2_wait_on_mount(struct ocfs2_super *osb, int quota)
* MOUNTED flag, but this is set right before
* dismount_volume() so we can trust it. */
if (atomic_read(&osb->vol_state) == VOLUME_DISABLED) {
+ trace_ocfs2_wait_on_mount(VOLUME_DISABLED);
mlog(0, "mount error, exiting!\n");
return -EBUSY;
}
diff --git a/fs/ocfs2/ocfs2_trace.h b/fs/ocfs2/ocfs2_trace.h
index a58008f..736cb40 100644
--- a/fs/ocfs2/ocfs2_trace.h
+++ b/fs/ocfs2/ocfs2_trace.h
@@ -3936,6 +3936,247 @@ DEFINE_EVENT(ocfs2__int, ocfs2_encode_fh_type,
);
/* End of trace events for fs/ocfs2/dcache.c. */
+
+/* Trace events for fs/ocfs2/journal.c. */
+DEFINE_EVENT(ocfs2__uint, ocfs2_commit_cache_begin,
+
+ TP_PROTO(unsigned int flush),
+
+ TP_ARGS(flush)
+);
+
+DEFINE_EVENT(ocfs2__ull_uint, ocfs2_commit_cache_end,
+
+ TP_PROTO(unsigned long long trans_id, unsigned int flushed),
+
+ TP_ARGS(trans_id, flushed)
+);
+
+DEFINE_EVENT(ocfs2__int_int, ocfs2_extend_trans,
+
+ TP_PROTO(int old, int nblocks),
+
+ TP_ARGS(old, nblocks)
+);
+
+DEFINE_EVENT(ocfs2__int, ocfs2_extend_trans_restart,
+
+ TP_PROTO(int nblocks),
+
+ TP_ARGS(nblocks)
+);
+
+DEFINE_EVENT(ocfs2__ull_ull_uint_uint, ocfs2_journal_access,
+
+ TP_PROTO(unsigned long long owner, unsigned long long blkno,
+ unsigned int type, unsigned int b_size),
+
+ TP_ARGS(owner, blkno, type, b_size)
+);
+
+DEFINE_EVENT(ocfs2__ull, ocfs2_journal_dirty,
+
+ TP_PROTO(unsigned long long blkno),
+
+ TP_ARGS(blkno)
+);
+
+DEFINE_EVENT(ocfs2__ull_ull_uint, ocfs2_journal_init,
+
+ TP_PROTO(unsigned long long size, unsigned long long blocks,
+ unsigned int clusters),
+
+ TP_ARGS(size, blocks, clusters)
+);
+
+DEFINE_EVENT(ocfs2__uint, ocfs2_journal_init_maxlen,
+
+ TP_PROTO(unsigned int maxlen),
+
+ TP_ARGS(maxlen)
+);
+
+DEFINE_EVENT(ocfs2__int, ocfs2_journal_shutdown,
+
+ TP_PROTO(int num_running_trans),
+
+ TP_ARGS(num_running_trans)
+);
+
+DEFINE_EVENT(ocfs2__pointer, ocfs2_journal_shutdown_wait,
+
+ TP_PROTO(void *commit_task),
+
+ TP_ARGS(commit_task)
+);
+
+DEFINE_EVENT(ocfs2__ull, ocfs2_complete_recovery,
+
+ TP_PROTO(unsigned long long ino),
+
+ TP_ARGS(ino)
+);
+
+DEFINE_EVENT(ocfs2__int, ocfs2_complete_recovery_end,
+
+ TP_PROTO(int ret),
+
+ TP_ARGS(ret)
+);
+
+TRACE_EVENT(ocfs2_complete_recovery_slot,
+
+ TP_PROTO(int slot, unsigned long long la_ino,
+ unsigned long long tl_ino, void *qrec),
+
+ TP_ARGS(slot, la_ino, tl_ino, qrec),
+
+ TP_STRUCT__entry(
+ __field( int, slot )
+ __field( unsigned long long, la_ino )
+ __field( unsigned long long, tl_ino )
+ __field( void *, qrec )
+ ),
+
+ TP_fast_assign(
+ __entry->slot = slot;
+ __entry->la_ino = la_ino;
+ __entry->tl_ino = tl_ino;
+ __entry->qrec = qrec;
+ ),
+
+ TP_printk("%d %llu %llu %p", __entry->slot, __entry->la_ino,
+ __entry->tl_ino, __entry->qrec)
+);
+
+DEFINE_EVENT(ocfs2__int_int, ocfs2_recovery_thread_node,
+
+ TP_PROTO(int node, int slot),
+
+ TP_ARGS(node, slot)
+);
+
+DEFINE_EVENT(ocfs2__int, ocfs2_recovery_thread_end,
+
+ TP_PROTO(int status),
+
+ TP_ARGS(status)
+);
+
+TRACE_EVENT(ocfs2_recovery_thread,
+
+ TP_PROTO(int node_num, int osb_node_num, int disable,
+ void *recovery_thread, int map_set),
+
+ TP_ARGS(node_num, osb_node_num, disable, recovery_thread, map_set),
+
+ TP_STRUCT__entry(
+ __field( int, node_num )
+ __field( int, osb_node_num )
+ __field( int, disable )
+ __field( void *, recovery_thread )
+ __field( int, map_set )
+ ),
+
+ TP_fast_assign(
+ __entry->node_num = node_num;
+ __entry->osb_node_num = osb_node_num;
+ __entry->disable = disable;
+ __entry->recovery_thread = recovery_thread;
+ __entry->map_set = map_set;
+ ),
+
+ TP_printk("%d %d %d %p %d", __entry->node_num,
+ __entry->osb_node_num, __entry->disable,
+ __entry->recovery_thread, __entry->map_set)
+);
+
+DEFINE_EVENT(ocfs2__uint_uint_uint, ocfs2_replay_journal_recovered,
+
+ TP_PROTO(unsigned int slot_num, unsigned int old_gen,
+ unsigned int new_gen),
+
+ TP_ARGS(slot_num, old_gen, new_gen)
+);
+
+DEFINE_EVENT(ocfs2__int, ocfs2_replay_journal_lock_err,
+
+ TP_PROTO(int status),
+
+ TP_ARGS(status)
+);
+
+DEFINE_EVENT(ocfs2__int, ocfs2_replay_journal_skip,
+
+ TP_PROTO(int slot),
+
+ TP_ARGS(slot)
+);
+
+DEFINE_EVENT(ocfs2__uint_uint_uint, ocfs2_recover_node,
+
+ TP_PROTO(unsigned int slot_num, unsigned int node_num,
+ unsigned int osb_slot_num),
+
+ TP_ARGS(slot_num, node_num, osb_slot_num)
+);
+
+DEFINE_EVENT(ocfs2__uint_uint, ocfs2_recover_node_skip,
+
+ TP_PROTO(unsigned int slot_num, unsigned int node_num),
+
+ TP_ARGS(slot_num, node_num)
+);
+
+DEFINE_EVENT(ocfs2__uint_uint, ocfs2_mark_dead_nodes,
+
+ TP_PROTO(unsigned int slot_num, unsigned int gen),
+
+ TP_ARGS(slot_num, gen)
+);
+
+DEFINE_EVENT(ocfs2__uint_uint_uint, ocfs2_queue_orphan_scan_begin,
+
+ TP_PROTO(unsigned int count, unsigned int seqno, unsigned int state),
+
+ TP_ARGS(count, seqno, state)
+);
+
+DEFINE_EVENT(ocfs2__uint_uint_uint, ocfs2_queue_orphan_scan_end,
+
+ TP_PROTO(unsigned int count, unsigned int seqno, unsigned int state),
+
+ TP_ARGS(count, seqno, state)
+);
+
+DEFINE_EVENT(ocfs2__ull, ocfs2_orphan_filldir,
+
+ TP_PROTO(unsigned long long blkno),
+
+ TP_ARGS(blkno)
+);
+
+DEFINE_EVENT(ocfs2__int, ocfs2_recover_orphans,
+
+ TP_PROTO(int slot),
+
+ TP_ARGS(slot)
+);
+
+DEFINE_EVENT(ocfs2__ull, ocfs2_recover_orphans_iput,
+
+ TP_PROTO(unsigned long long blkno),
+
+ TP_ARGS(blkno)
+);
+
+DEFINE_EVENT(ocfs2__int, ocfs2_wait_on_mount,
+
+ TP_PROTO(int state),
+
+ TP_ARGS(state)
+);
+/* End of trace events for fs/ocfs2/journal.c. */
#endif /* _TRACE_OCFS2_H */
/* This part must be outside protection */
--
1.7.0.4
^ permalink raw reply related [flat|nested] 53+ messages in thread
* [Ocfs2-devel] [PATCH 31/34] ocfs2: Remove masklog ML_BH_IO.
2010-12-23 7:19 [Ocfs2-devel] [PATCH 00/34] OCFS2: Add trace event and replace mlog(0) Tao Ma
` (29 preceding siblings ...)
2010-12-23 7:31 ` [Ocfs2-devel] [PATCH 30/34] ocfs2: Remove masklog ML_JOURNAL Tao Ma
@ 2010-12-23 7:31 ` Tao Ma
2010-12-23 7:31 ` [Ocfs2-devel] [PATCH 32/34] ocfs2: Remove masklog ML_UPTODATE Tao Ma
` (7 subsequent siblings)
38 siblings, 0 replies; 53+ messages in thread
From: Tao Ma @ 2010-12-23 7:31 UTC (permalink / raw)
To: ocfs2-devel
From: Tao Ma <boyu.mt@taobao.com>
Remove mlog(0,...) and mlog(ML_BH_IO,...) from
fs/ocfs2/buffer_head_io.c.
The masklog BH_IO is removed finally.
Signed-off-by: Tao Ma <boyu.mt@taobao.com>
---
fs/ocfs2/buffer_head_io.c | 37 +++++++--------------
fs/ocfs2/cluster/masklog.c | 1 -
fs/ocfs2/cluster/masklog.h | 1 -
fs/ocfs2/ocfs2_trace.h | 76 ++++++++++++++++++++++++++++++++++++++++++++
4 files changed, 89 insertions(+), 26 deletions(-)
diff --git a/fs/ocfs2/buffer_head_io.c b/fs/ocfs2/buffer_head_io.c
index c23846d..d1cf9ab 100644
--- a/fs/ocfs2/buffer_head_io.c
+++ b/fs/ocfs2/buffer_head_io.c
@@ -35,8 +35,8 @@
#include "inode.h"
#include "journal.h"
#include "uptodate.h"
-
#include "buffer_head_io.h"
+#include "ocfs2_trace.h"
/*
* Bits on bh->b_state used by ocfs2.
@@ -55,8 +55,7 @@ int ocfs2_write_block(struct ocfs2_super *osb, struct buffer_head *bh,
{
int ret = 0;
- mlog(0, "(bh->b_blocknr = %llu, ci=%p)\n",
- (unsigned long long)bh->b_blocknr, ci);
+ trace_ocfs2_write_block((unsigned long long)bh->b_blocknr, ci);
BUG_ON(bh->b_blocknr < OCFS2_SUPER_BLOCK_BLKNO);
BUG_ON(buffer_jbd(bh));
@@ -107,10 +106,10 @@ int ocfs2_read_blocks_sync(struct ocfs2_super *osb, u64 block,
unsigned int i;
struct buffer_head *bh;
- if (!nr) {
- mlog(ML_BH_IO, "No buffers will be read!\n");
+ trace_ocfs2_read_blocks_sync((unsigned long long)block, nr);
+
+ if (!nr)
goto bail;
- }
for (i = 0 ; i < nr ; i++) {
if (bhs[i] == NULL) {
@@ -124,10 +123,8 @@ int ocfs2_read_blocks_sync(struct ocfs2_super *osb, u64 block,
bh = bhs[i];
if (buffer_jbd(bh)) {
- mlog(ML_BH_IO,
- "trying to sync read a jbd "
- "managed bh (blocknr = %llu), skipping\n",
- (unsigned long long)bh->b_blocknr);
+ trace_ocfs2_read_blocks_sync_jbd(
+ (unsigned long long)bh->b_blocknr);
continue;
}
@@ -187,8 +184,7 @@ int ocfs2_read_blocks(struct ocfs2_caching_info *ci, u64 block, int nr,
struct buffer_head *bh;
struct super_block *sb = ocfs2_metadata_cache_get_super(ci);
- mlog(0, "(ci=%p, block=(%llu), nr=(%d), flags=%d)\n",
- ci, (unsigned long long)block, nr, flags);
+ trace_ocfs2_read_blocks_begin(ci, (unsigned long long)block, nr, flags);
BUG_ON(!ci);
BUG_ON((flags & OCFS2_BH_READAHEAD) &&
@@ -208,7 +204,6 @@ int ocfs2_read_blocks(struct ocfs2_caching_info *ci, u64 block, int nr,
}
if (nr == 0) {
- mlog(ML_BH_IO, "No buffers will be read!\n");
status = 0;
goto bail;
}
@@ -261,11 +256,10 @@ int ocfs2_read_blocks(struct ocfs2_caching_info *ci, u64 block, int nr,
ignore_cache = 1;
}
+ trace_ocfs2_read_blocks_bh((unsigned long long)bh->b_blocknr,
+ ignore_cache, buffer_jbd(bh), buffer_dirty(bh));
+
if (buffer_jbd(bh)) {
- if (ignore_cache)
- mlog(ML_BH_IO, "trying to sync read a jbd "
- "managed bh (blocknr = %llu)\n",
- (unsigned long long)bh->b_blocknr);
continue;
}
@@ -273,9 +267,6 @@ int ocfs2_read_blocks(struct ocfs2_caching_info *ci, u64 block, int nr,
if (buffer_dirty(bh)) {
/* This should probably be a BUG, or
* at least return an error. */
- mlog(ML_BH_IO, "asking me to sync read a dirty "
- "buffer! (blocknr = %llu)\n",
- (unsigned long long)bh->b_blocknr);
continue;
}
@@ -368,10 +359,8 @@ int ocfs2_read_blocks(struct ocfs2_caching_info *ci, u64 block, int nr,
}
ocfs2_metadata_cache_io_unlock(ci);
- mlog(ML_BH_IO, "block=(%llu), nr=(%d), cached=%s, flags=0x%x\n",
- (unsigned long long)block, nr,
- ((flags & OCFS2_BH_IGNORE_CACHE) || ignore_cache) ? "no" : "yes",
- flags);
+ trace_ocfs2_read_blocks_end((unsigned long long)block, nr,
+ flags, ignore_cache);
bail:
diff --git a/fs/ocfs2/cluster/masklog.c b/fs/ocfs2/cluster/masklog.c
index 3f1e995..137e7e8 100644
--- a/fs/ocfs2/cluster/masklog.c
+++ b/fs/ocfs2/cluster/masklog.c
@@ -93,7 +93,6 @@ static struct mlog_attribute mlog_attrs[MLOG_MAX_BITS] = {
define_mask(DLM_RECOVERY),
define_mask(AIO),
define_mask(DLM_GLUE),
- define_mask(BH_IO),
define_mask(UPTODATE),
define_mask(VOTE),
define_mask(CONN),
diff --git a/fs/ocfs2/cluster/masklog.h b/fs/ocfs2/cluster/masklog.h
index 61e8f6c..5c7e38a 100644
--- a/fs/ocfs2/cluster/masklog.h
+++ b/fs/ocfs2/cluster/masklog.h
@@ -95,7 +95,6 @@
#define ML_DLM_RECOVERY 0x0000000000001000ULL /* dlm master functions */
#define ML_AIO 0x0000000000002000ULL /* ocfs2 aio read and write */
#define ML_DLM_GLUE 0x0000000000080000ULL /* ocfs2 dlm glue layer */
-#define ML_BH_IO 0x0000000000100000ULL /* ocfs2 buffer I/O */
#define ML_UPTODATE 0x0000000000200000ULL /* ocfs2 caching sequence #'s */
#define ML_VOTE 0x0000000001000000ULL /* ocfs2 node messaging */
#define ML_CONN 0x0000000004000000ULL /* net connection management */
diff --git a/fs/ocfs2/ocfs2_trace.h b/fs/ocfs2/ocfs2_trace.h
index 736cb40..9cd8670 100644
--- a/fs/ocfs2/ocfs2_trace.h
+++ b/fs/ocfs2/ocfs2_trace.h
@@ -4177,6 +4177,82 @@ DEFINE_EVENT(ocfs2__int, ocfs2_wait_on_mount,
TP_ARGS(state)
);
/* End of trace events for fs/ocfs2/journal.c. */
+
+/* Trace events for fs/ocfs2/buffer_head_io.c. */
+DEFINE_EVENT(ocfs2__ull_uint, ocfs2_read_blocks_sync,
+
+ TP_PROTO(unsigned long long block, unsigned int nr),
+
+ TP_ARGS(block, nr)
+);
+
+DEFINE_EVENT(ocfs2__ull, ocfs2_read_blocks_sync_jbd,
+
+ TP_PROTO(unsigned long long block),
+
+ TP_ARGS(block)
+);
+
+DEFINE_EVENT(ocfs2__ull_int_int_int, ocfs2_read_blocks_bh,
+
+ TP_PROTO(unsigned long long block, int ignore_cache,
+ int buffer_jbd, int buffer_dirty),
+
+ TP_ARGS(block, ignore_cache, buffer_jbd, buffer_dirty)
+);
+
+DEFINE_EVENT(ocfs2__ull_int_int_int, ocfs2_read_blocks_end,
+
+ TP_PROTO(unsigned long long block, int nr, int flags, int ignore_cache),
+
+ TP_ARGS(block, nr, flags, ignore_cache)
+);
+
+TRACE_EVENT(ocfs2_write_block,
+
+ TP_PROTO(unsigned long long block, void *ci),
+
+ TP_ARGS(block, ci),
+
+ TP_STRUCT__entry(
+ __field( unsigned long long, block )
+ __field( void *, ci )
+ ),
+
+ TP_fast_assign(
+ __entry->block = block;
+ __entry->ci = ci;
+ ),
+
+ TP_printk("%llu %p", __entry->block, __entry->ci)
+);
+
+TRACE_EVENT(ocfs2_read_blocks_begin,
+
+ TP_PROTO(void *ci, unsigned long long block,
+ unsigned int nr, int flags),
+
+ TP_ARGS(ci, block, nr, flags),
+
+ TP_STRUCT__entry(
+ __field( void *, ci )
+ __field( unsigned long long, block )
+ __field( unsigned int, nr )
+ __field( int, flags )
+ ),
+
+ TP_fast_assign(
+ __entry->ci = ci;
+ __entry->block = block;
+ __entry->nr = nr;
+ __entry->flags = flags;
+ ),
+
+ TP_printk("%p %llu %u %d", __entry->ci, __entry->block,
+ __entry->nr, __entry->flags)
+);
+
+/* End of trace events for fs/ocfs2/buffer_head_io.c. */
#endif /* _TRACE_OCFS2_H */
/* This part must be outside protection */
--
1.7.0.4
^ permalink raw reply related [flat|nested] 53+ messages in thread
* [Ocfs2-devel] [PATCH 32/34] ocfs2: Remove masklog ML_UPTODATE.
2010-12-23 7:19 [Ocfs2-devel] [PATCH 00/34] OCFS2: Add trace event and replace mlog(0) Tao Ma
` (30 preceding siblings ...)
2010-12-23 7:31 ` [Ocfs2-devel] [PATCH 31/34] ocfs2: Remove masklog ML_BH_IO Tao Ma
@ 2010-12-23 7:31 ` Tao Ma
2010-12-23 7:31 ` [Ocfs2-devel] [PATCH 33/34] ocfs2: Remove masklog ML_AIO Tao Ma
` (6 subsequent siblings)
38 siblings, 0 replies; 53+ messages in thread
From: Tao Ma @ 2010-12-23 7:31 UTC (permalink / raw)
To: ocfs2-devel
From: Tao Ma <boyu.mt@taobao.com>
Remove mlog(0,...) and mlog(ML_UPTODATE,...) from
fs/ocfs2/uptodate.c and fs/ocfs2/buffer_head_io.c.
The masklog UPTODATE is removed finally.
Signed-off-by: Tao Ma <boyu.mt@taobao.com>
---
fs/ocfs2/buffer_head_io.c | 3 +-
fs/ocfs2/cluster/masklog.c | 1 -
fs/ocfs2/cluster/masklog.h | 1 -
fs/ocfs2/ocfs2_trace.h | 116 ++++++++++++++++++++++++++++++++++++++++++++
fs/ocfs2/uptodate.c | 73 ++++++++++++++--------------
5 files changed, 153 insertions(+), 41 deletions(-)
diff --git a/fs/ocfs2/buffer_head_io.c b/fs/ocfs2/buffer_head_io.c
index d1cf9ab..5d18ad1 100644
--- a/fs/ocfs2/buffer_head_io.c
+++ b/fs/ocfs2/buffer_head_io.c
@@ -247,8 +247,7 @@ int ocfs2_read_blocks(struct ocfs2_caching_info *ci, u64 block, int nr,
*/
if (!ignore_cache && !ocfs2_buffer_uptodate(ci, bh)) {
- mlog(ML_UPTODATE,
- "bh (%llu), owner %llu not uptodate\n",
+ trace_ocfs2_read_blocks_from_disk(
(unsigned long long)bh->b_blocknr,
(unsigned long long)ocfs2_metadata_cache_owner(ci));
/* We're using ignore_cache here to say
diff --git a/fs/ocfs2/cluster/masklog.c b/fs/ocfs2/cluster/masklog.c
index 137e7e8..89235aa 100644
--- a/fs/ocfs2/cluster/masklog.c
+++ b/fs/ocfs2/cluster/masklog.c
@@ -93,7 +93,6 @@ static struct mlog_attribute mlog_attrs[MLOG_MAX_BITS] = {
define_mask(DLM_RECOVERY),
define_mask(AIO),
define_mask(DLM_GLUE),
- define_mask(UPTODATE),
define_mask(VOTE),
define_mask(CONN),
define_mask(QUORUM),
diff --git a/fs/ocfs2/cluster/masklog.h b/fs/ocfs2/cluster/masklog.h
index 5c7e38a..a81b883 100644
--- a/fs/ocfs2/cluster/masklog.h
+++ b/fs/ocfs2/cluster/masklog.h
@@ -95,7 +95,6 @@
#define ML_DLM_RECOVERY 0x0000000000001000ULL /* dlm master functions */
#define ML_AIO 0x0000000000002000ULL /* ocfs2 aio read and write */
#define ML_DLM_GLUE 0x0000000000080000ULL /* ocfs2 dlm glue layer */
-#define ML_UPTODATE 0x0000000000200000ULL /* ocfs2 caching sequence #'s */
#define ML_VOTE 0x0000000001000000ULL /* ocfs2 node messaging */
#define ML_CONN 0x0000000004000000ULL /* net connection management */
#define ML_QUORUM 0x0000000008000000ULL /* net connection quorum */
diff --git a/fs/ocfs2/ocfs2_trace.h b/fs/ocfs2/ocfs2_trace.h
index 9cd8670..482c02a 100644
--- a/fs/ocfs2/ocfs2_trace.h
+++ b/fs/ocfs2/ocfs2_trace.h
@@ -4193,6 +4193,13 @@ DEFINE_EVENT(ocfs2__ull, ocfs2_read_blocks_sync_jbd,
TP_ARGS(block)
);
+DEFINE_EVENT(ocfs2__ull_ull, ocfs2_read_blocks_from_disk,
+
+ TP_PROTO(unsigned long long block, unsigned long long owner),
+
+ TP_ARGS(block, owner)
+);
+
DEFINE_EVENT(ocfs2__ull_int_int_int, ocfs2_read_blocks_bh,
TP_PROTO(unsigned long long block, int ignore_cache,
@@ -4253,6 +4260,115 @@ TRACE_EVENT(ocfs2_read_blocks_begin,
);
/* End of trace events for fs/ocfs2/buffer_head_io.c. */
+
+/* Trace events for fs/ocfs2/uptodate.c. */
+DEFINE_EVENT(ocfs2__ull, ocfs2_purge_copied_metadata_tree,
+
+ TP_PROTO(unsigned long long block),
+
+ TP_ARGS(block)
+);
+
+DEFINE_EVENT(ocfs2__ull_uint_uint, ocfs2_metadata_cache_purge,
+
+ TP_PROTO(unsigned long long owner,
+ unsigned int to_purge, unsigned int tree),
+
+ TP_ARGS(owner, to_purge, tree)
+);
+
+DEFINE_EVENT(ocfs2__ull_ull_uint, ocfs2_buffer_cached_begin,
+
+ TP_PROTO(unsigned long long owner,
+ unsigned long long block, unsigned int inlined),
+
+ TP_ARGS(owner, block, inlined)
+);
+
+TRACE_EVENT(ocfs2_buffer_cached_end,
+
+ TP_PROTO(int index, void *item),
+
+ TP_ARGS(index, item),
+
+ TP_STRUCT__entry(
+ __field( int, index )
+ __field( void *, item )
+ ),
+
+ TP_fast_assign(
+ __entry->index = index;
+ __entry->item = item;
+ ),
+
+ TP_printk("%d %p", __entry->index, __entry->item)
+);
+
+DEFINE_EVENT(ocfs2__ull_ull_uint, ocfs2_append_cache_array,
+
+ TP_PROTO(unsigned long long owner,
+ unsigned long long block, unsigned int index),
+
+ TP_ARGS(owner, block, index)
+);
+
+DEFINE_EVENT(ocfs2__ull_ull_uint, ocfs2_insert_cache_tree,
+
+ TP_PROTO(unsigned long long owner,
+ unsigned long long block, unsigned int num),
+
+ TP_ARGS(owner, block, num)
+);
+
+DEFINE_EVENT(ocfs2__ull_uint_uint, ocfs2_expand_cache,
+
+ TP_PROTO(unsigned long long owner,
+ unsigned int flags, unsigned int num),
+
+ TP_ARGS(owner, flags, num)
+);
+
+DEFINE_EVENT(ocfs2__ull_ull_uint, ocfs2_set_buffer_uptodate,
+
+ TP_PROTO(unsigned long long owner,
+ unsigned long long block, unsigned int expand_tree),
+
+ TP_ARGS(owner, block, expand_tree)
+);
+
+DEFINE_EVENT(ocfs2__ull_ull, ocfs2_set_buffer_uptodate_begin,
+
+ TP_PROTO(unsigned long long owner,
+ unsigned long long block),
+
+ TP_ARGS(owner, block)
+);
+
+DEFINE_EVENT(ocfs2__ull_uint_uint, ocfs2_remove_metadata_array,
+
+ TP_PROTO(unsigned long long owner,
+ unsigned int index, unsigned int num_cached),
+
+ TP_ARGS(owner, index, num_cached)
+);
+
+DEFINE_EVENT(ocfs2__ull_ull, ocfs2_remove_metadata_tree,
+
+ TP_PROTO(unsigned long long owner,
+ unsigned long long block),
+
+ TP_ARGS(owner, block)
+);
+
+DEFINE_EVENT(ocfs2__ull_ull_uint_uint, ocfs2_remove_block_from_cache,
+
+ TP_PROTO(unsigned long long owner, unsigned long long block,
+ unsigned int num_cached, unsigned int flags),
+
+ TP_ARGS(owner, block, num_cached, flags)
+);
+
+/* End of trace events for fs/ocfs2/uptodate.c. */
#endif /* _TRACE_OCFS2_H */
/* This part must be outside protection */
diff --git a/fs/ocfs2/uptodate.c b/fs/ocfs2/uptodate.c
index a0a120e..52eaf33 100644
--- a/fs/ocfs2/uptodate.c
+++ b/fs/ocfs2/uptodate.c
@@ -54,14 +54,13 @@
#include <linux/buffer_head.h>
#include <linux/rbtree.h>
-#define MLOG_MASK_PREFIX ML_UPTODATE
-
#include <cluster/masklog.h>
#include "ocfs2.h"
#include "inode.h"
#include "uptodate.h"
+#include "ocfs2_trace.h"
struct ocfs2_meta_cache_item {
struct rb_node c_node;
@@ -152,8 +151,8 @@ static unsigned int ocfs2_purge_copied_metadata_tree(struct rb_root *root)
while ((node = rb_last(root)) != NULL) {
item = rb_entry(node, struct ocfs2_meta_cache_item, c_node);
- mlog(0, "Purge item %llu\n",
- (unsigned long long) item->c_block);
+ trace_ocfs2_purge_copied_metadata_tree(
+ (unsigned long long) item->c_block);
rb_erase(&item->c_node, root);
kmem_cache_free(ocfs2_uptodate_cachep, item);
@@ -180,9 +179,9 @@ void ocfs2_metadata_cache_purge(struct ocfs2_caching_info *ci)
tree = !(ci->ci_flags & OCFS2_CACHE_FL_INLINE);
to_purge = ci->ci_num_cached;
- mlog(0, "Purge %u %s items from Owner %llu\n", to_purge,
- tree ? "array" : "tree",
- (unsigned long long)ocfs2_metadata_cache_owner(ci));
+ trace_ocfs2_metadata_cache_purge(
+ (unsigned long long)ocfs2_metadata_cache_owner(ci),
+ to_purge, tree);
/* If we're a tree, save off the root so that we can safely
* initialize the cache. We do the work to free tree members
@@ -249,10 +248,10 @@ static int ocfs2_buffer_cached(struct ocfs2_caching_info *ci,
ocfs2_metadata_cache_lock(ci);
- mlog(0, "Owner %llu, query block %llu (inline = %u)\n",
- (unsigned long long)ocfs2_metadata_cache_owner(ci),
- (unsigned long long) bh->b_blocknr,
- !!(ci->ci_flags & OCFS2_CACHE_FL_INLINE));
+ trace_ocfs2_buffer_cached_begin(
+ (unsigned long long)ocfs2_metadata_cache_owner(ci),
+ (unsigned long long) bh->b_blocknr,
+ !!(ci->ci_flags & OCFS2_CACHE_FL_INLINE));
if (ci->ci_flags & OCFS2_CACHE_FL_INLINE)
index = ocfs2_search_cache_array(ci, bh->b_blocknr);
@@ -261,7 +260,7 @@ static int ocfs2_buffer_cached(struct ocfs2_caching_info *ci,
ocfs2_metadata_cache_unlock(ci);
- mlog(0, "index = %d, item = %p\n", index, item);
+ trace_ocfs2_buffer_cached_end(index, item);
return (index != -1) || (item != NULL);
}
@@ -306,8 +305,9 @@ static void ocfs2_append_cache_array(struct ocfs2_caching_info *ci,
{
BUG_ON(ci->ci_num_cached >= OCFS2_CACHE_INFO_MAX_ARRAY);
- mlog(0, "block %llu takes position %u\n", (unsigned long long) block,
- ci->ci_num_cached);
+ trace_ocfs2_append_cache_array(
+ (unsigned long long)ocfs2_metadata_cache_owner(ci),
+ (unsigned long long)block, ci->ci_num_cached);
ci->ci_cache.ci_array[ci->ci_num_cached] = block;
ci->ci_num_cached++;
@@ -324,8 +324,9 @@ static void __ocfs2_insert_cache_tree(struct ocfs2_caching_info *ci,
struct rb_node **p = &ci->ci_cache.ci_tree.rb_node;
struct ocfs2_meta_cache_item *tmp;
- mlog(0, "Insert block %llu num = %u\n", (unsigned long long) block,
- ci->ci_num_cached);
+ trace_ocfs2_insert_cache_tree(
+ (unsigned long long)ocfs2_metadata_cache_owner(ci),
+ (unsigned long long)block, ci->ci_num_cached);
while(*p) {
parent = *p;
@@ -389,9 +390,9 @@ static void ocfs2_expand_cache(struct ocfs2_caching_info *ci,
tree[i] = NULL;
}
- mlog(0, "Expanded %llu to a tree cache: flags 0x%x, num = %u\n",
- (unsigned long long)ocfs2_metadata_cache_owner(ci),
- ci->ci_flags, ci->ci_num_cached);
+ trace_ocfs2_expand_cache(
+ (unsigned long long)ocfs2_metadata_cache_owner(ci),
+ ci->ci_flags, ci->ci_num_cached);
}
/* Slow path function - memory allocation is necessary. See the
@@ -405,9 +406,9 @@ static void __ocfs2_set_buffer_uptodate(struct ocfs2_caching_info *ci,
struct ocfs2_meta_cache_item *tree[OCFS2_CACHE_INFO_MAX_ARRAY] =
{ NULL, };
- mlog(0, "Owner %llu, block %llu, expand = %d\n",
- (unsigned long long)ocfs2_metadata_cache_owner(ci),
- (unsigned long long)block, expand_tree);
+ trace_ocfs2_set_buffer_uptodate(
+ (unsigned long long)ocfs2_metadata_cache_owner(ci),
+ (unsigned long long)block, expand_tree);
new = kmem_cache_alloc(ocfs2_uptodate_cachep, GFP_NOFS);
if (!new) {
@@ -433,7 +434,6 @@ static void __ocfs2_set_buffer_uptodate(struct ocfs2_caching_info *ci,
ocfs2_metadata_cache_lock(ci);
if (ocfs2_insert_can_use_array(ci)) {
- mlog(0, "Someone cleared the tree underneath us\n");
/* Ok, items were removed from the cache in between
* locks. Detect this and revert back to the fast path */
ocfs2_append_cache_array(ci, block);
@@ -490,9 +490,9 @@ void ocfs2_set_buffer_uptodate(struct ocfs2_caching_info *ci,
if (ocfs2_buffer_cached(ci, bh))
return;
- mlog(0, "Owner %llu, inserting block %llu\n",
- (unsigned long long)ocfs2_metadata_cache_owner(ci),
- (unsigned long long)bh->b_blocknr);
+ trace_ocfs2_set_buffer_uptodate_begin(
+ (unsigned long long)ocfs2_metadata_cache_owner(ci),
+ (unsigned long long)bh->b_blocknr);
/* No need to recheck under spinlock - insertion is guarded by
* co_io_lock() */
@@ -542,8 +542,9 @@ static void ocfs2_remove_metadata_array(struct ocfs2_caching_info *ci,
BUG_ON(index >= ci->ci_num_cached);
BUG_ON(!ci->ci_num_cached);
- mlog(0, "remove index %d (num_cached = %u\n", index,
- ci->ci_num_cached);
+ trace_ocfs2_remove_metadata_array(
+ (unsigned long long)ocfs2_metadata_cache_owner(ci),
+ index, ci->ci_num_cached);
ci->ci_num_cached--;
@@ -559,8 +560,9 @@ static void ocfs2_remove_metadata_array(struct ocfs2_caching_info *ci,
static void ocfs2_remove_metadata_tree(struct ocfs2_caching_info *ci,
struct ocfs2_meta_cache_item *item)
{
- mlog(0, "remove block %llu from tree\n",
- (unsigned long long) item->c_block);
+ trace_ocfs2_remove_metadata_tree(
+ (unsigned long long)ocfs2_metadata_cache_owner(ci),
+ (unsigned long long)item->c_block);
rb_erase(&item->c_node, &ci->ci_cache.ci_tree);
ci->ci_num_cached--;
@@ -573,10 +575,10 @@ static void ocfs2_remove_block_from_cache(struct ocfs2_caching_info *ci,
struct ocfs2_meta_cache_item *item = NULL;
ocfs2_metadata_cache_lock(ci);
- mlog(0, "Owner %llu, remove %llu, items = %u, array = %u\n",
- (unsigned long long)ocfs2_metadata_cache_owner(ci),
- (unsigned long long) block, ci->ci_num_cached,
- ci->ci_flags & OCFS2_CACHE_FL_INLINE);
+ trace_ocfs2_remove_block_from_cache(
+ (unsigned long long)ocfs2_metadata_cache_owner(ci),
+ (unsigned long long) block, ci->ci_num_cached,
+ ci->ci_flags);
if (ci->ci_flags & OCFS2_CACHE_FL_INLINE) {
index = ocfs2_search_cache_array(ci, block);
@@ -626,9 +628,6 @@ int __init init_ocfs2_uptodate_cache(void)
if (!ocfs2_uptodate_cachep)
return -ENOMEM;
- mlog(0, "%u inlined cache items per inode.\n",
- OCFS2_CACHE_INFO_MAX_ARRAY);
-
return 0;
}
--
1.7.0.4
^ permalink raw reply related [flat|nested] 53+ messages in thread
* [Ocfs2-devel] [PATCH 33/34] ocfs2: Remove masklog ML_AIO.
2010-12-23 7:19 [Ocfs2-devel] [PATCH 00/34] OCFS2: Add trace event and replace mlog(0) Tao Ma
` (31 preceding siblings ...)
2010-12-23 7:31 ` [Ocfs2-devel] [PATCH 32/34] ocfs2: Remove masklog ML_UPTODATE Tao Ma
@ 2010-12-23 7:31 ` Tao Ma
2010-12-23 7:31 ` [Ocfs2-devel] [PATCH 34/34] ocfs2: Make the left masklogs compat Tao Ma
` (5 subsequent siblings)
38 siblings, 0 replies; 53+ messages in thread
From: Tao Ma @ 2010-12-23 7:31 UTC (permalink / raw)
To: ocfs2-devel
From: Tao Ma <boyu.mt@taobao.com>
There is no user for masklog AIO, so remove it.
Signed-off-by: Tao Ma <boyu.mt@taobao.com>
---
fs/ocfs2/cluster/masklog.c | 1 -
fs/ocfs2/cluster/masklog.h | 1 -
2 files changed, 0 insertions(+), 2 deletions(-)
diff --git a/fs/ocfs2/cluster/masklog.c b/fs/ocfs2/cluster/masklog.c
index 89235aa..aac2251 100644
--- a/fs/ocfs2/cluster/masklog.c
+++ b/fs/ocfs2/cluster/masklog.c
@@ -91,7 +91,6 @@ static struct mlog_attribute mlog_attrs[MLOG_MAX_BITS] = {
define_mask(DLM_THREAD),
define_mask(DLM_MASTER),
define_mask(DLM_RECOVERY),
- define_mask(AIO),
define_mask(DLM_GLUE),
define_mask(VOTE),
define_mask(CONN),
diff --git a/fs/ocfs2/cluster/masklog.h b/fs/ocfs2/cluster/masklog.h
index a81b883..eac025a 100644
--- a/fs/ocfs2/cluster/masklog.h
+++ b/fs/ocfs2/cluster/masklog.h
@@ -93,7 +93,6 @@
#define ML_DLM_THREAD 0x0000000000000400ULL /* dlm domain thread */
#define ML_DLM_MASTER 0x0000000000000800ULL /* dlm master functions */
#define ML_DLM_RECOVERY 0x0000000000001000ULL /* dlm master functions */
-#define ML_AIO 0x0000000000002000ULL /* ocfs2 aio read and write */
#define ML_DLM_GLUE 0x0000000000080000ULL /* ocfs2 dlm glue layer */
#define ML_VOTE 0x0000000001000000ULL /* ocfs2 node messaging */
#define ML_CONN 0x0000000004000000ULL /* net connection management */
--
1.7.0.4
^ permalink raw reply related [flat|nested] 53+ messages in thread
* [Ocfs2-devel] [PATCH 34/34] ocfs2: Make the left masklogs compat.
2010-12-23 7:19 [Ocfs2-devel] [PATCH 00/34] OCFS2: Add trace event and replace mlog(0) Tao Ma
` (32 preceding siblings ...)
2010-12-23 7:31 ` [Ocfs2-devel] [PATCH 33/34] ocfs2: Remove masklog ML_AIO Tao Ma
@ 2010-12-23 7:31 ` Tao Ma
2010-12-23 7:32 ` [Ocfs2-devel] [PATCH 00/34] OCFS2: Add trace event and replace mlog(0) Tao Ma
` (4 subsequent siblings)
38 siblings, 0 replies; 53+ messages in thread
From: Tao Ma @ 2010-12-23 7:31 UTC (permalink / raw)
To: ocfs2-devel
From: Tao Ma <boyu.mt@taobao.com>
Since we have removed almost all of the masklogs in fs/ocfs2/*,
make the left masklogs compat.
Signed-off-by: Tao Ma <boyu.mt@taobao.com>
---
fs/ocfs2/cluster/masklog.h | 40 ++++++++++++++++++++--------------------
1 files changed, 20 insertions(+), 20 deletions(-)
diff --git a/fs/ocfs2/cluster/masklog.h b/fs/ocfs2/cluster/masklog.h
index eac025a..7e90312 100644
--- a/fs/ocfs2/cluster/masklog.h
+++ b/fs/ocfs2/cluster/masklog.h
@@ -82,27 +82,27 @@
/* bits that are frequently given and infrequently matched in the low word */
/* NOTE: If you add a flag, you need to also update mlog.c! */
-#define ML_TCP 0x0000000000000004ULL /* net cluster/tcp.c */
-#define ML_MSG 0x0000000000000008ULL /* net network messages */
-#define ML_SOCKET 0x0000000000000010ULL /* net socket lifetime */
-#define ML_HEARTBEAT 0x0000000000000020ULL /* hb all heartbeat tracking */
-#define ML_HB_BIO 0x0000000000000040ULL /* hb io tracing */
-#define ML_DLMFS 0x0000000000000080ULL /* dlm user dlmfs */
-#define ML_DLM 0x0000000000000100ULL /* dlm general debugging */
-#define ML_DLM_DOMAIN 0x0000000000000200ULL /* dlm domain debugging */
-#define ML_DLM_THREAD 0x0000000000000400ULL /* dlm domain thread */
-#define ML_DLM_MASTER 0x0000000000000800ULL /* dlm master functions */
-#define ML_DLM_RECOVERY 0x0000000000001000ULL /* dlm master functions */
-#define ML_DLM_GLUE 0x0000000000080000ULL /* ocfs2 dlm glue layer */
-#define ML_VOTE 0x0000000001000000ULL /* ocfs2 node messaging */
-#define ML_CONN 0x0000000004000000ULL /* net connection management */
-#define ML_QUORUM 0x0000000008000000ULL /* net connection quorum */
-#define ML_BASTS 0x0000001000000000ULL /* dlmglue asts and basts */
+#define ML_TCP 0x0000000000000001ULL /* net cluster/tcp.c */
+#define ML_MSG 0x0000000000000002ULL /* net network messages */
+#define ML_SOCKET 0x0000000000000004ULL /* net socket lifetime */
+#define ML_HEARTBEAT 0x0000000000000008ULL /* hb all heartbeat tracking */
+#define ML_HB_BIO 0x0000000000000010ULL /* hb io tracing */
+#define ML_DLMFS 0x0000000000000020ULL /* dlm user dlmfs */
+#define ML_DLM 0x0000000000000040ULL /* dlm general debugging */
+#define ML_DLM_DOMAIN 0x0000000000000080ULL /* dlm domain debugging */
+#define ML_DLM_THREAD 0x0000000000000100ULL /* dlm domain thread */
+#define ML_DLM_MASTER 0x0000000000000200ULL /* dlm master functions */
+#define ML_DLM_RECOVERY 0x0000000000000400ULL /* dlm master functions */
+#define ML_DLM_GLUE 0x0000000000000800ULL /* ocfs2 dlm glue layer */
+#define ML_VOTE 0x0000000000001000ULL /* ocfs2 node messaging */
+#define ML_CONN 0x0000000000002000ULL /* net connection management */
+#define ML_QUORUM 0x0000000000004000ULL /* net connection quorum */
+#define ML_BASTS 0x0000000000008000ULL /* dlmglue asts and basts */
/* bits that are infrequently given and frequently matched in the high word */
-#define ML_ERROR 0x0000000100000000ULL /* sent to KERN_ERR */
-#define ML_NOTICE 0x0000000200000000ULL /* setn to KERN_NOTICE */
-#define ML_KTHREAD 0x0000000400000000ULL /* kernel thread activity */
-#define ML_CLUSTER 0x0000001000000000ULL /* cluster stack */
+#define ML_ERROR 0x0000000000010000ULL /* sent to KERN_ERR */
+#define ML_NOTICE 0x0000000000020000ULL /* setn to KERN_NOTICE */
+#define ML_KTHREAD 0x0000000000040000ULL /* kernel thread activity */
+#define ML_CLUSTER 0x0000000000080000ULL /* cluster stack */
#define MLOG_INITIAL_AND_MASK (ML_ERROR|ML_NOTICE)
#ifndef MLOG_MASK_PREFIX
--
1.7.0.4
^ permalink raw reply related [flat|nested] 53+ messages in thread
* [Ocfs2-devel] [PATCH 00/34] OCFS2: Add trace event and replace mlog(0).
2010-12-23 7:19 [Ocfs2-devel] [PATCH 00/34] OCFS2: Add trace event and replace mlog(0) Tao Ma
` (33 preceding siblings ...)
2010-12-23 7:31 ` [Ocfs2-devel] [PATCH 34/34] ocfs2: Make the left masklogs compat Tao Ma
@ 2010-12-23 7:32 ` Tao Ma
2010-12-23 8:44 ` Tao Ma
` (3 subsequent siblings)
38 siblings, 0 replies; 53+ messages in thread
From: Tao Ma @ 2010-12-23 7:32 UTC (permalink / raw)
To: ocfs2-devel
forget to say.
you can view from
http://repo.or.cz/w/taoma-kernel.git/shortlog/refs/heads/mlog_replace
and pull from
git://repo.or.cz/taoma-kernel.git mlog_replace
Regards,
Tao
On 12/23/2010 03:19 PM, Tao Ma wrote:
> Hi all,
>
> This patch set is my attempt to replace our old debug(mlog) with new
> trace events.
>
> Wengang Wang has did some work in this field about one year ago.
> http://oss.oracle.com/pipermail/ocfs2-devel/2009-September/005230.html
> But at that time we don't have ocfs2 1.6 introduced, so we delay this
> work until now after ocfs2 1.6 and UEK(Unbreakable Enterprise Kernel)
> are released.
>
> All patches except one(see below) are only mlog related, so they
> shouldn't affect the normal operations in ocfs2 and they should be easy
> for review(patch 1-5 were already sent to the mail list around one month
> ago).
>
> The only one that changes the codes somehow are patch 15, "ocfs2: Little
> refactoring against ocfs2_iget", but it looks straightforward.
>
> Regards,
> Tao
>
> _______________________________________________
> Ocfs2-devel mailing list
> Ocfs2-devel at oss.oracle.com
> http://oss.oracle.com/mailman/listinfo/ocfs2-devel
^ permalink raw reply [flat|nested] 53+ messages in thread
* [Ocfs2-devel] [PATCH 00/34] OCFS2: Add trace event and replace mlog(0).
2010-12-23 7:19 [Ocfs2-devel] [PATCH 00/34] OCFS2: Add trace event and replace mlog(0) Tao Ma
` (34 preceding siblings ...)
2010-12-23 7:32 ` [Ocfs2-devel] [PATCH 00/34] OCFS2: Add trace event and replace mlog(0) Tao Ma
@ 2010-12-23 8:44 ` Tao Ma
2010-12-31 12:52 ` Joel Becker
` (2 subsequent siblings)
38 siblings, 0 replies; 53+ messages in thread
From: Tao Ma @ 2010-12-23 8:44 UTC (permalink / raw)
To: ocfs2-devel
Oh, I suddenly found that I send an older version. So it contains some
bugs and I will update them after the review is done.
But it doesn't prevent you guys from review, and please don't run it by
now. ;)
Regards,
Tao
On 12/23/2010 03:19 PM, Tao Ma wrote:
> Hi all,
>
> This patch set is my attempt to replace our old debug(mlog) with new
> trace events.
>
> Wengang Wang has did some work in this field about one year ago.
> http://oss.oracle.com/pipermail/ocfs2-devel/2009-September/005230.html
> But at that time we don't have ocfs2 1.6 introduced, so we delay this
> work until now after ocfs2 1.6 and UEK(Unbreakable Enterprise Kernel)
> are released.
>
> All patches except one(see below) are only mlog related, so they
> shouldn't affect the normal operations in ocfs2 and they should be easy
> for review(patch 1-5 were already sent to the mail list around one month
> ago).
>
> The only one that changes the codes somehow are patch 15, "ocfs2: Little
> refactoring against ocfs2_iget", but it looks straightforward.
>
> Regards,
> Tao
>
> _______________________________________________
> Ocfs2-devel mailing list
> Ocfs2-devel at oss.oracle.com
> http://oss.oracle.com/mailman/listinfo/ocfs2-devel
^ permalink raw reply [flat|nested] 53+ messages in thread
* [Ocfs2-devel] [PATCH 01/34] ocfs2: Remove unused truncate function from alloc.c
2010-12-23 7:30 ` [Ocfs2-devel] [PATCH 01/34] ocfs2: Remove unused truncate function from alloc.c Tao Ma
@ 2010-12-28 22:54 ` Mark Fasheh
0 siblings, 0 replies; 53+ messages in thread
From: Mark Fasheh @ 2010-12-28 22:54 UTC (permalink / raw)
To: ocfs2-devel
On Thu, Dec 23, 2010 at 03:30:44PM +0800, Tao Ma wrote:
> From: Tao Ma <tao.ma@oracle.com>
>
> Tristan Ye has done some refactoring against our truncate
> process, so some functions like ocfs2_prepare_truncate and
> ocfs2_free_truncate_context are no use and we'd better
> remove them.
>
> Signed-off-by: Tao Ma <tao.ma@oracle.com>
Hmm, so we're removing dead code here? This should just go in as-is then
IMHO.
Signed-off-by: Mark Fasheh <mfasheh@suse.com>
--
Mark Fasheh
^ permalink raw reply [flat|nested] 53+ messages in thread
* [Ocfs2-devel] [PATCH 04/34] ocfs2: Add ocfs2_trace.h.
2010-12-23 7:30 ` [Ocfs2-devel] [PATCH 04/34] ocfs2: Add ocfs2_trace.h Tao Ma
@ 2010-12-31 12:40 ` Joel Becker
0 siblings, 0 replies; 53+ messages in thread
From: Joel Becker @ 2010-12-31 12:40 UTC (permalink / raw)
To: ocfs2-devel
On Thu, Dec 23, 2010 at 03:30:47PM +0800, Tao Ma wrote:
> From: Wengang Wang <wen.gang.wang@oracle.com>
>
> About one year ago, Wengang Wang tried some first steps
> to add tracepoints to ocfs2. But at that time we don't
> have ocfs2 1.6 introduced, so we delay this work until now
> after ocfs2 1.6 and UEK(Unbreakable Enterprise Kernel) are
> released.
Let's not mention marketing terms. Just say that we had
to delay the work.
Joel
--
"When ideas fail, words come in very handy."
- Goethe
Joel Becker
Senior Development Manager
Oracle
E-mail: joel.becker at oracle.com
Phone: (650) 506-8127
^ permalink raw reply [flat|nested] 53+ messages in thread
* [Ocfs2-devel] [PATCH 00/34] OCFS2: Add trace event and replace mlog(0).
2010-12-23 7:19 [Ocfs2-devel] [PATCH 00/34] OCFS2: Add trace event and replace mlog(0) Tao Ma
` (35 preceding siblings ...)
2010-12-23 8:44 ` Tao Ma
@ 2010-12-31 12:52 ` Joel Becker
2010-12-31 15:11 ` Tao Ma
2010-12-31 13:10 ` Christoph Hellwig
2011-02-14 2:53 ` Tao Ma
38 siblings, 1 reply; 53+ messages in thread
From: Joel Becker @ 2010-12-31 12:52 UTC (permalink / raw)
To: ocfs2-devel
On Thu, Dec 23, 2010 at 03:19:03PM +0800, Tao Ma wrote:
> This patch set is my attempt to replace our old debug(mlog) with new
> trace events.
>
> Wengang Wang has did some work in this field about one year ago.
> http://oss.oracle.com/pipermail/ocfs2-devel/2009-September/005230.html
> But at that time we don't have ocfs2 1.6 introduced, so we delay
> this work until now after ocfs2 1.6 and UEK(Unbreakable Enterprise
> Kernel) are released.
>
> All patches except one(see below) are only mlog related, so they
> shouldn't affect the normal operations in ocfs2 and they should be
> easy for review(patch 1-5 were already sent to the mail list around
> one month ago).
>
> The only one that changes the codes somehow are patch 15, "ocfs2:
> Little refactoring against ocfs2_iget", but it looks
> straightforward.
Tao,
Overall this seems pretty straightforward. There wasn't a lot
of editing of masklog entries; we still have a million tracepoints. I
wonder how much memory that will use. Have you checked the space usage
of all the sysfs files for all of our tracepoints?
Aside from my comment about marketing terms in patch comments
(hehe), I'm OK with going forward with these patches. I would love the
ability to turn specific traces on and off.
Well, one generic comment. It seems a lot of our log messages
have gone from "ocfs2_foo(): Change blah 0x%p to new_blah 0x%p" to
"ocfs2_foo(): %p %p". I wonder if that is going to be hard to read.
Joel
--
Bram's Law:
The easier a piece of software is to write, the worse it's
implemented in practice.
Joel Becker
Senior Development Manager
Oracle
E-mail: joel.becker at oracle.com
Phone: (650) 506-8127
^ permalink raw reply [flat|nested] 53+ messages in thread
* [Ocfs2-devel] [PATCH 00/34] OCFS2: Add trace event and replace mlog(0).
2010-12-23 7:19 [Ocfs2-devel] [PATCH 00/34] OCFS2: Add trace event and replace mlog(0) Tao Ma
` (36 preceding siblings ...)
2010-12-31 12:52 ` Joel Becker
@ 2010-12-31 13:10 ` Christoph Hellwig
2010-12-31 13:14 ` Joel Becker
2011-02-14 2:53 ` Tao Ma
38 siblings, 1 reply; 53+ messages in thread
From: Christoph Hellwig @ 2010-12-31 13:10 UTC (permalink / raw)
To: ocfs2-devel
I really don't like the way this is split into tons of patches, which
makes reviewing what's going on pretty hard. In addition to the
preparatiory patch just one patch is more than enough, or if you really
want two: one to remove the old mlog code and one to add the trace
events.
^ permalink raw reply [flat|nested] 53+ messages in thread
* [Ocfs2-devel] [PATCH 00/34] OCFS2: Add trace event and replace mlog(0).
2010-12-31 13:10 ` Christoph Hellwig
@ 2010-12-31 13:14 ` Joel Becker
2010-12-31 13:18 ` Christoph Hellwig
0 siblings, 1 reply; 53+ messages in thread
From: Joel Becker @ 2010-12-31 13:14 UTC (permalink / raw)
To: ocfs2-devel
On Fri, Dec 31, 2010 at 02:10:20PM +0100, Christoph Hellwig wrote:
> I really don't like the way this is split into tons of patches, which
> makes reviewing what's going on pretty hard. In addition to the
> preparatiory patch just one patch is more than enough, or if you really
> want two: one to remove the old mlog code and one to add the trace
> events.
I have to say, I would have found a single massive patch really
painful to read. The eye glazes over so many repetitive changes. But
I've now read it, and we can certainly push it upstream as one commit if
you like.
Joel
--
Life's Little Instruction Book #497
"Go down swinging."
Joel Becker
Senior Development Manager
Oracle
E-mail: joel.becker at oracle.com
Phone: (650) 506-8127
^ permalink raw reply [flat|nested] 53+ messages in thread
* [Ocfs2-devel] [PATCH 05/34] ocfs2: Remove mlog(0) from fs/ocfs2/alloc.c
2010-12-23 7:30 ` [Ocfs2-devel] [PATCH 05/34] ocfs2: Remove mlog(0) from fs/ocfs2/alloc.c Tao Ma
@ 2010-12-31 13:15 ` Christoph Hellwig
2010-12-31 14:27 ` Tao Ma
0 siblings, 1 reply; 53+ messages in thread
From: Christoph Hellwig @ 2010-12-31 13:15 UTC (permalink / raw)
To: ocfs2-devel
> + (unsigned long long)ocfs2_metadata_cache_owner(et->et_ci),
> + cpos, len, index, le32_to_cpu(rec->e_cpos),
> + ocfs2_rec_clusters(el, rec));
>
> if (le32_to_cpu(rec->e_cpos) == cpos || rec_range == trunc_range) {
> ret = ocfs2_truncate_rec(handle, et, path, index, dealloc,
> @@ -5780,9 +5789,6 @@ int ocfs2_truncate_log_append(struct ocfs2_super *osb,
> struct ocfs2_dinode *di;
> struct ocfs2_truncate_log *tl;
>
> - mlog(0, "start_blk = %llu, num_clusters = %u\n",
> - (unsigned long long)start_blk, num_clusters);
> -
> BUG_ON(mutex_trylock(&tl_inode->i_mutex));
>
> start_cluster = ocfs2_blocks_to_clusters(osb->sb, start_blk);
> @@ -5819,10 +5825,9 @@ int ocfs2_truncate_log_append(struct ocfs2_super *osb,
> goto bail;
> }
>
> - mlog(0, "Log truncate of %u clusters starting at cluster %u to "
> - "%llu (index = %d)\n", num_clusters, start_cluster,
> - (unsigned long long)OCFS2_I(tl_inode)->ip_blkno, index);
> -
> + trace_ocfs2_truncate_log_append(
> + (unsigned long long)OCFS2_I(tl_inode)->ip_blkno, index,
> + start_cluster, num_clusters);
> if (ocfs2_truncate_log_can_coalesce(tl, start_cluster)) {
> /*
> * Move index back to the record we are coalescing with.
> @@ -5831,9 +5836,10 @@ int ocfs2_truncate_log_append(struct ocfs2_super *osb,
> index--;
>
> num_clusters += le32_to_cpu(tl->tl_recs[index].t_clusters);
> - mlog(0, "Coalesce with index %u (start = %u, clusters = %u)\n",
> - index, le32_to_cpu(tl->tl_recs[index].t_start),
> - num_clusters);
> + trace_ocfs2_truncate_log_append(
> + (unsigned long long)OCFS2_I(tl_inode)->ip_blkno,
> + index, le32_to_cpu(tl->tl_recs[index].t_start),
> + num_clusters);
> } else {
> tl->tl_recs[index].t_start = cpu_to_le32(start_cluster);
> tl->tl_used = cpu_to_le16(index + 1);
> @@ -5896,8 +5902,9 @@ static int ocfs2_replay_truncate_records(struct ocfs2_super *osb,
> /* if start_blk is not set, we ignore the record as
> * invalid. */
> if (start_blk) {
> - mlog(0, "free record %d, start = %u, clusters = %u\n",
> - i, le32_to_cpu(rec.t_start), num_clusters);
> + trace_ocfs2_replay_truncate_records(
> + (unsigned long long)OCFS2_I(tl_inode)->ip_blkno,
> + i, le32_to_cpu(rec.t_start), num_clusters);
>
> status = ocfs2_free_clusters(handle, data_alloc_inode,
> data_alloc_bh, start_blk,
> @@ -5938,8 +5945,9 @@ int __ocfs2_flush_truncate_log(struct ocfs2_super *osb)
>
> tl = &di->id2.i_dealloc;
> num_to_flush = le16_to_cpu(tl->tl_used);
> - mlog(0, "Flush %u records from truncate log #%llu\n",
> - num_to_flush, (unsigned long long)OCFS2_I(tl_inode)->ip_blkno);
> + trace_ocfs2_flush_truncate_log(
> + (unsigned long long)OCFS2_I(tl_inode)->ip_blkno,
> + num_to_flush);
> if (!num_to_flush) {
> status = 0;
> goto out;
> @@ -6076,7 +6084,7 @@ int ocfs2_begin_truncate_log_recovery(struct ocfs2_super *osb,
>
> *tl_copy = NULL;
>
> - mlog(0, "recover truncate log from slot %d\n", slot_num);
> + trace_ocfs2_begin_truncate_log_recovery(slot_num);
>
> status = ocfs2_get_truncate_log_info(osb, slot_num, &tl_inode, &tl_bh);
> if (status < 0) {
> @@ -6093,8 +6101,7 @@ int ocfs2_begin_truncate_log_recovery(struct ocfs2_super *osb,
>
> tl = &di->id2.i_dealloc;
> if (le16_to_cpu(tl->tl_used)) {
> - mlog(0, "We'll have %u logs to recover\n",
> - le16_to_cpu(tl->tl_used));
> + trace_ocfs2_truncate_log_recovery_num(le16_to_cpu(tl->tl_used));
>
> *tl_copy = kmalloc(tl_bh->b_size, GFP_KERNEL);
> if (!(*tl_copy)) {
> @@ -6151,8 +6158,9 @@ int ocfs2_complete_truncate_log_recovery(struct ocfs2_super *osb,
>
> tl = &tl_copy->id2.i_dealloc;
> num_recs = le16_to_cpu(tl->tl_used);
> - mlog(0, "cleanup %u records from %llu\n", num_recs,
> - (unsigned long long)le64_to_cpu(tl_copy->i_blkno));
> + trace_ocfs2_complete_truncate_log_recovery(
> + (unsigned long long)le64_to_cpu(tl_copy->i_blkno),
> + num_recs);
>
> mutex_lock(&tl_inode->i_mutex);
> for(i = 0; i < num_recs; i++) {
> @@ -6310,8 +6318,8 @@ static int ocfs2_free_cached_blocks(struct ocfs2_super *osb,
> else
> bg_blkno = ocfs2_which_suballoc_group(head->free_blk,
> head->free_bit);
> - mlog(0, "Free bit: (bit %u, blkno %llu)\n",
> - head->free_bit, (unsigned long long)head->free_blk);
> + trace_ocfs2_free_cached_blocks(
> + (unsigned long long)head->free_blk, head->free_bit);
>
> ret = ocfs2_free_suballoc_bits(handle, inode, di_bh,
> head->free_bit, bg_blkno, 1);
> @@ -6364,8 +6372,7 @@ int ocfs2_cache_cluster_dealloc(struct ocfs2_cached_dealloc_ctxt *ctxt,
> return ret;
> }
>
> - mlog(0, "Insert clusters: (bit %u, blk %llu)\n",
> - bit, (unsigned long long)blkno);
> + trace_ocfs2_cache_cluster_dealloc((unsigned long long)blkno, bit);
>
> item->free_blk = blkno;
> item->free_bit = bit;
> @@ -6440,8 +6447,8 @@ int ocfs2_run_deallocs(struct ocfs2_super *osb,
> fl = ctxt->c_first_suballocator;
>
> if (fl->f_first) {
> - mlog(0, "Free items: (type %u, slot %d)\n",
> - fl->f_inode_type, fl->f_slot);
> + trace_ocfs2_run_deallocs(fl->f_inode_type,
> + fl->f_slot);
> ret2 = ocfs2_free_cached_blocks(osb,
> fl->f_inode_type,
> fl->f_slot,
> @@ -6518,8 +6525,9 @@ int ocfs2_cache_block_dealloc(struct ocfs2_cached_dealloc_ctxt *ctxt,
> goto out;
> }
>
> - mlog(0, "Insert: (type %d, slot %u, bit %u, blk %llu)\n",
> - type, slot, bit, (unsigned long long)blkno);
> + trace_ocfs2_cache_block_dealloc(type, slot,
> + (unsigned long long)suballoc,
> + (unsigned long long)blkno, bit);
>
> item->free_bg = suballoc;
> item->free_blk = blkno;
> @@ -6999,8 +7007,11 @@ start:
> goto bail;
> }
>
> - mlog(0, "inode->ip_clusters = %u, tree_depth = %u\n",
> - OCFS2_I(inode)->ip_clusters, path->p_tree_depth);
> + trace_ocfs2_commit_truncate(
> + (unsigned long long)OCFS2_I(inode)->ip_blkno,
> + new_highest_cpos,
> + OCFS2_I(inode)->ip_clusters,
> + path->p_tree_depth);
>
> /*
> * By now, el will point to the extent list on the bottom most
> diff --git a/fs/ocfs2/ocfs2_trace.h b/fs/ocfs2/ocfs2_trace.h
> index 8d4e49a..2dfc3b0 100644
> --- a/fs/ocfs2/ocfs2_trace.h
> +++ b/fs/ocfs2/ocfs2_trace.h
> @@ -6,6 +6,507 @@
>
> #include <linux/tracepoint.h>
>
> +DECLARE_EVENT_CLASS(ocfs2__int,
> +
> + TP_PROTO(int num),
> +
> + TP_ARGS(num),
> +
> + TP_STRUCT__entry(
> + __field( int, num )
> + ),
> +
> + TP_fast_assign(
> + __entry->num = num;
> + ),
> +
> + TP_printk("%d", __entry->num)
> +);
If you'd split the even up into classes depending on the actual meaning
of num the trace could contain a descriptive name that makes it a lot
more readable. And btw, the above would be a lot more readable if
formatted like normal kernel code:
DECLARE_EVENT_CLASS(ocfs2__int,
TP_PROTO(int num),
TP_ARGS(num),
TP_STRUCT__entry(
__field(int, num)
),
TP_fast_assign(
__entry->num = num;
),
TP_printk("%d", __entry->num)
);
> +DEFINE_EVENT(ocfs2__truncate_log_ops, ocfs2_truncate_log_append,
> +
> + TP_PROTO(unsigned long long blkno, int index,
> + unsigned int start, unsigned int num),
> +
> + TP_ARGS(blkno, index, start, num)
> +);
FYI, if you have a lot of instances ofa given even class using a little
macro makes declaring them a lot simpler, take a look at
fs/xfs/linux-2.6/xfs_trace.h for examples.
^ permalink raw reply [flat|nested] 53+ messages in thread
* [Ocfs2-devel] [PATCH 00/34] OCFS2: Add trace event and replace mlog(0).
2010-12-31 13:14 ` Joel Becker
@ 2010-12-31 13:18 ` Christoph Hellwig
0 siblings, 0 replies; 53+ messages in thread
From: Christoph Hellwig @ 2010-12-31 13:18 UTC (permalink / raw)
To: ocfs2-devel
On Fri, Dec 31, 2010 at 05:14:46AM -0800, Joel Becker wrote:
> On Fri, Dec 31, 2010 at 02:10:20PM +0100, Christoph Hellwig wrote:
> > I really don't like the way this is split into tons of patches, which
> > makes reviewing what's going on pretty hard. In addition to the
> > preparatiory patch just one patch is more than enough, or if you really
> > want two: one to remove the old mlog code and one to add the trace
> > events.
>
> I have to say, I would have found a single massive patch really
> painful to read. The eye glazes over so many repetitive changes. But
> I've now read it, and we can certainly push it upstream as one commit if
> you like.
It doesn't matter too much - in the end it's mostly you guys who have to
deal with it.
^ permalink raw reply [flat|nested] 53+ messages in thread
* [Ocfs2-devel] [PATCH 05/34] ocfs2: Remove mlog(0) from fs/ocfs2/alloc.c
2010-12-31 13:15 ` Christoph Hellwig
@ 2010-12-31 14:27 ` Tao Ma
0 siblings, 0 replies; 53+ messages in thread
From: Tao Ma @ 2010-12-31 14:27 UTC (permalink / raw)
To: ocfs2-devel
Hi Christoph,
Thanks for the review.
On 12/31/2010 09:15 PM, Christoph Hellwig wrote:
>> diff --git a/fs/ocfs2/ocfs2_trace.h b/fs/ocfs2/ocfs2_trace.h
>> index 8d4e49a..2dfc3b0 100644
>> --- a/fs/ocfs2/ocfs2_trace.h
>> +++ b/fs/ocfs2/ocfs2_trace.h
>> @@ -6,6 +6,507 @@
>>
>> #include<linux/tracepoint.h>
>>
>> +DECLARE_EVENT_CLASS(ocfs2__int,
>> +
>> + TP_PROTO(int num),
>> +
>> + TP_ARGS(num),
>> +
>> + TP_STRUCT__entry(
>> + __field( int, num )
>> + ),
>> +
>> + TP_fast_assign(
>> + __entry->num = num;
>> + ),
>> +
>> + TP_printk("%d", __entry->num)
>> +);
>
> If you'd split the even up into classes depending on the actual meaning
> of num the trace could contain a descriptive name that makes it a lot
> more readable. And btw, the above would be a lot more readable if
> formatted like normal kernel code:
>
> DECLARE_EVENT_CLASS(ocfs2__int,
> TP_PROTO(int num),
> TP_ARGS(num),
> TP_STRUCT__entry(
> __field(int, num)
> ),
> TP_fast_assign(
> __entry->num = num;
> ),
>
> TP_printk("%d", __entry->num)
> );
ok, I will change it in the next round.
>
>> +DEFINE_EVENT(ocfs2__truncate_log_ops, ocfs2_truncate_log_append,
>> +
>> + TP_PROTO(unsigned long long blkno, int index,
>> + unsigned int start, unsigned int num),
>> +
>> + TP_ARGS(blkno, index, start, num)
>> +);
>
> FYI, if you have a lot of instances ofa given even class using a little
> macro makes declaring them a lot simpler, take a look at
> fs/xfs/linux-2.6/xfs_trace.h for examples.
OK, I will look at it and use a macro. thanks.
Regards,
Tao
^ permalink raw reply [flat|nested] 53+ messages in thread
* [Ocfs2-devel] [PATCH 00/34] OCFS2: Add trace event and replace mlog(0).
2010-12-31 12:52 ` Joel Becker
@ 2010-12-31 15:11 ` Tao Ma
2010-12-31 22:39 ` Joel Becker
0 siblings, 1 reply; 53+ messages in thread
From: Tao Ma @ 2010-12-31 15:11 UTC (permalink / raw)
To: ocfs2-devel
On 12/31/2010 08:52 PM, Joel Becker wrote:
> On Thu, Dec 23, 2010 at 03:19:03PM +0800, Tao Ma wrote:
>> This patch set is my attempt to replace our old debug(mlog) with new
>> trace events.
>>
>> Wengang Wang has did some work in this field about one year ago.
>> http://oss.oracle.com/pipermail/ocfs2-devel/2009-September/005230.html
>> But at that time we don't have ocfs2 1.6 introduced, so we delay
>> this work until now after ocfs2 1.6 and UEK(Unbreakable Enterprise
>> Kernel) are released.
>>
>> All patches except one(see below) are only mlog related, so they
>> shouldn't affect the normal operations in ocfs2 and they should be
>> easy for review(patch 1-5 were already sent to the mail list around
>> one month ago).
>>
>> The only one that changes the codes somehow are patch 15, "ocfs2:
>> Little refactoring against ocfs2_iget", but it looks
>> straightforward.
>
> Tao,
> Overall this seems pretty straightforward. There wasn't a lot
> of editing of masklog entries; we still have a million tracepoints. I
> wonder how much memory that will use. Have you checked the space usage
> of all the sysfs files for all of our tracepoints?
Sorry, I don't know how to check the space usage of these files. any tips?
> Aside from my comment about marketing terms in patch comments
> (hehe), I'm OK with going forward with these patches. I would love the
> ability to turn specific traces on and off.
OK, I will regenerate the patches with advices from Christoph and your.
> Well, one generic comment. It seems a lot of our log messages
> have gone from "ocfs2_foo(): Change blah 0x%p to new_blah 0x%p" to
> "ocfs2_foo(): %p %p". I wonder if that is going to be hard to read.
Yeah, I used to think of adding all these descriptive strings to trace
buffers, but finally I gave up. The reason is that no matter how we
print the message in the trace buffer, we have to go to read the codes
and with the codes, everything should be clear enough. ;)
Regards,
Tao
^ permalink raw reply [flat|nested] 53+ messages in thread
* [Ocfs2-devel] [PATCH 00/34] OCFS2: Add trace event and replace mlog(0).
2010-12-31 15:11 ` Tao Ma
@ 2010-12-31 22:39 ` Joel Becker
2011-01-04 9:06 ` Tao Ma
0 siblings, 1 reply; 53+ messages in thread
From: Joel Becker @ 2010-12-31 22:39 UTC (permalink / raw)
To: ocfs2-devel
On Fri, Dec 31, 2010 at 11:11:31PM +0800, Tao Ma wrote:
> On 12/31/2010 08:52 PM, Joel Becker wrote:
> > Overall this seems pretty straightforward. There wasn't a lot
> > of editing of masklog entries; we still have a million tracepoints. I
> > wonder how much memory that will use. Have you checked the space usage
> > of all the sysfs files for all of our tracepoints?
> Sorry, I don't know how to check the space usage of these files. any tips?
Just count the number of files and directories added to sysfs.
I believe the files disappear when not open, but the directories
I think have inode and dentry structures around permanently..
> > Well, one generic comment. It seems a lot of our log messages
> > have gone from "ocfs2_foo(): Change blah 0x%p to new_blah 0x%p" to
> > "ocfs2_foo(): %p %p". I wonder if that is going to be hard to read.
> Yeah, I used to think of adding all these descriptive strings to trace
> buffers, but finally I gave up. The reason is that no matter how we
> print the message in the trace buffer, we have to go to read the codes
> and with the codes, everything should be clear enough. ;)
Ok, we'll give it a shot.
Joel
--
"The opposite of a correct statement is a false statement. The
opposite of a profound truth may well be another profound truth."
- Niels Bohr
http://www.jlbec.org/
jlbec at evilplan.org
^ permalink raw reply [flat|nested] 53+ messages in thread
* [Ocfs2-devel] [PATCH 00/34] OCFS2: Add trace event and replace mlog(0).
2010-12-31 22:39 ` Joel Becker
@ 2011-01-04 9:06 ` Tao Ma
2011-01-04 18:43 ` Sunil Mushran
2011-01-04 22:15 ` Joel Becker
0 siblings, 2 replies; 53+ messages in thread
From: Tao Ma @ 2011-01-04 9:06 UTC (permalink / raw)
To: ocfs2-devel
On 01/01/2011 06:39 AM, Joel Becker wrote:
> On Fri, Dec 31, 2010 at 11:11:31PM +0800, Tao Ma wrote:
>> On 12/31/2010 08:52 PM, Joel Becker wrote:
>>> Overall this seems pretty straightforward. There wasn't a lot
>>> of editing of masklog entries; we still have a million tracepoints. I
>>> wonder how much memory that will use. Have you checked the space usage
>>> of all the sysfs files for all of our tracepoints?
>> Sorry, I don't know how to check the space usage of these files. any tips?
>
> Just count the number of files and directories added to sysfs.
> I believe the files disappear when not open, but the directories
> I think have inode and dentry structures around permanently..
OK, so
#find /sys/kernel/debug/tracing/events/ocfs2 -type d|wc -l
give me 319. Every trace event dir has 4 members(enable,filter,format
and id). So it contains about 1600(dir+files).
>
>>> Well, one generic comment. It seems a lot of our log messages
>>> have gone from "ocfs2_foo(): Change blah 0x%p to new_blah 0x%p" to
>>> "ocfs2_foo(): %p %p". I wonder if that is going to be hard to read.
>> Yeah, I used to think of adding all these descriptive strings to trace
>> buffers, but finally I gave up. The reason is that no matter how we
>> print the message in the trace buffer, we have to go to read the codes
>> and with the codes, everything should be clear enough. ;)
>
> Ok, we'll give it a shot.
I have rebased my tree with Linus' latest tree and integrated all the
comments Christoph and you have prompted.
View:
http://repo.or.cz/w/taoma-kernel.git/shortlog/refs/heads/new_mlog_replace
and pull from
git://repo.or.cz/taoma-kernel.git new_mlog_replace
Please consider merging it. Thanks.
Regards,
Tao
^ permalink raw reply [flat|nested] 53+ messages in thread
* [Ocfs2-devel] [PATCH 00/34] OCFS2: Add trace event and replace mlog(0).
2011-01-04 9:06 ` Tao Ma
@ 2011-01-04 18:43 ` Sunil Mushran
2011-01-04 22:15 ` Joel Becker
1 sibling, 0 replies; 53+ messages in thread
From: Sunil Mushran @ 2011-01-04 18:43 UTC (permalink / raw)
To: ocfs2-devel
On 01/04/2011 01:06 AM, Tao Ma wrote:
> #find /sys/kernel/debug/tracing/events/ocfs2 -type d|wc -l
> give me 319. Every trace event dir has 4 members(enable,filter,format
> and id). So it contains about 1600(dir+files).
>
That's a lot. ext4 currently has < 40.
I am wondering whether it is wise to treat this as a mlog replacement.
One suggestion would be to not remove all mlogs (other than remove _enter
and _exit as they are completely useless). Add trace events one by one. Coarse
tracing like ext4 has. Remove only those mlogs.
^ permalink raw reply [flat|nested] 53+ messages in thread
* [Ocfs2-devel] [PATCH 00/34] OCFS2: Add trace event and replace mlog(0).
2011-01-04 9:06 ` Tao Ma
2011-01-04 18:43 ` Sunil Mushran
@ 2011-01-04 22:15 ` Joel Becker
2011-01-04 22:33 ` Sunil Mushran
2011-01-05 2:25 ` Tao Ma
1 sibling, 2 replies; 53+ messages in thread
From: Joel Becker @ 2011-01-04 22:15 UTC (permalink / raw)
To: ocfs2-devel
On Tue, Jan 04, 2011 at 05:06:53PM +0800, Tao Ma wrote:
> On 01/01/2011 06:39 AM, Joel Becker wrote:
> >On Fri, Dec 31, 2010 at 11:11:31PM +0800, Tao Ma wrote:
> >>On 12/31/2010 08:52 PM, Joel Becker wrote:
> >>> Overall this seems pretty straightforward. There wasn't a lot
> >>>of editing of masklog entries; we still have a million tracepoints. I
> >>>wonder how much memory that will use. Have you checked the space usage
> >>>of all the sysfs files for all of our tracepoints?
> >>Sorry, I don't know how to check the space usage of these files. any tips?
> >
> > Just count the number of files and directories added to sysfs.
> >I believe the files disappear when not open, but the directories
> >I think have inode and dentry structures around permanently..
> OK, so
> #find /sys/kernel/debug/tracing/events/ocfs2 -type d|wc -l
> give me 319. Every trace event dir has 4
> members(enable,filter,format and id). So it contains about
> 1600(dir+files).
I think what matters to count is dirs, because files disappear
when you're not using them.
If you have 319 trace event directories, that's 319 inodes. On
my x86, a vanilla struct inode is 360 bytes and a dentry is 136 bytes.
That's 154K always in RAM for these knobs. I gotta say that I'm not too
worried about 154K. On x86_64 (ca-build24) this grows to 241K. I
imagine it will double when cluster and dlm are moved to similar trace
events.
Are we OK with 300K on x86 and 500K on x86_64 always used up by
these tracing entries?
Joel
--
Life's Little Instruction Book #464
"Don't miss the magic of the moment by focusing on what's
to come."
Joel Becker
Senior Development Manager
Oracle
E-mail: joel.becker at oracle.com
Phone: (650) 506-8127
^ permalink raw reply [flat|nested] 53+ messages in thread
* [Ocfs2-devel] [PATCH 00/34] OCFS2: Add trace event and replace mlog(0).
2011-01-04 22:15 ` Joel Becker
@ 2011-01-04 22:33 ` Sunil Mushran
2011-01-05 2:25 ` Tao Ma
1 sibling, 0 replies; 53+ messages in thread
From: Sunil Mushran @ 2011-01-04 22:33 UTC (permalink / raw)
To: ocfs2-devel
On 01/04/2011 02:15 PM, Joel Becker wrote:
>
> I think what matters to count is dirs, because files disappear
> when you're not using them.
> If you have 319 trace event directories, that's 319 inodes. On
> my x86, a vanilla struct inode is 360 bytes and a dentry is 136 bytes.
> That's 154K always in RAM for these knobs. I gotta say that I'm not too
> worried about 154K. On x86_64 (ca-build24) this grows to 241K. I
> imagine it will double when cluster and dlm are moved to similar trace
> events.
> Are we OK with 300K on x86 and 500K on x86_64 always used up by
> these tracing entries?
I'm ok with that.
^ permalink raw reply [flat|nested] 53+ messages in thread
* [Ocfs2-devel] [PATCH 00/34] OCFS2: Add trace event and replace mlog(0).
2011-01-04 22:15 ` Joel Becker
2011-01-04 22:33 ` Sunil Mushran
@ 2011-01-05 2:25 ` Tao Ma
1 sibling, 0 replies; 53+ messages in thread
From: Tao Ma @ 2011-01-05 2:25 UTC (permalink / raw)
To: ocfs2-devel
On 01/05/2011 06:15 AM, Joel Becker wrote:
> On Tue, Jan 04, 2011 at 05:06:53PM +0800, Tao Ma wrote:
>> On 01/01/2011 06:39 AM, Joel Becker wrote:
>>> On Fri, Dec 31, 2010 at 11:11:31PM +0800, Tao Ma wrote:
>>>> On 12/31/2010 08:52 PM, Joel Becker wrote:
>>>>> Overall this seems pretty straightforward. There wasn't a lot
>>>>> of editing of masklog entries; we still have a million tracepoints. I
>>>>> wonder how much memory that will use. Have you checked the space usage
>>>>> of all the sysfs files for all of our tracepoints?
>>>> Sorry, I don't know how to check the space usage of these files. any tips?
>>>
>>> Just count the number of files and directories added to sysfs.
>>> I believe the files disappear when not open, but the directories
>>> I think have inode and dentry structures around permanently..
>> OK, so
>> #find /sys/kernel/debug/tracing/events/ocfs2 -type d|wc -l
>> give me 319. Every trace event dir has 4
>> members(enable,filter,format and id). So it contains about
>> 1600(dir+files).
>
> I think what matters to count is dirs, because files disappear
> when you're not using them.
> If you have 319 trace event directories, that's 319 inodes. On
> my x86, a vanilla struct inode is 360 bytes and a dentry is 136 bytes.
> That's 154K always in RAM for these knobs. I gotta say that I'm not too
> worried about 154K. On x86_64 (ca-build24) this grows to 241K. I
> imagine it will double when cluster and dlm are moved to similar trace
> events.
> Are we OK with 300K on x86 and 500K on x86_64 always used up by
> these tracing entries?
yeah, it isn't much I guess.
And for ext4, yes, it has a small number because they have a very
limited debug tracing than us, maybe because they evolved from
ext2->ext3->ext4. So the code base is really stable now. And I guess we
can remove some of the trace events if we feel that the code is stable
enough. ;)
Regards,
Tao
^ permalink raw reply [flat|nested] 53+ messages in thread
* [Ocfs2-devel] [PATCH 00/34] OCFS2: Add trace event and replace mlog(0).
2010-12-23 7:19 [Ocfs2-devel] [PATCH 00/34] OCFS2: Add trace event and replace mlog(0) Tao Ma
` (37 preceding siblings ...)
2010-12-31 13:10 ` Christoph Hellwig
@ 2011-02-14 2:53 ` Tao Ma
38 siblings, 0 replies; 53+ messages in thread
From: Tao Ma @ 2011-02-14 2:53 UTC (permalink / raw)
To: ocfs2-devel
Hi Joel,
any plan to integrate this patch set for the next merge window?
Regards,
Tao
On 12/23/2010 03:19 PM, Tao Ma wrote:
> Hi all,
>
> This patch set is my attempt to replace our old debug(mlog) with new
> trace events.
>
> Wengang Wang has did some work in this field about one year ago.
> http://oss.oracle.com/pipermail/ocfs2-devel/2009-September/005230.html
> But at that time we don't have ocfs2 1.6 introduced, so we delay this
> work until now after ocfs2 1.6 and UEK(Unbreakable Enterprise Kernel)
> are released.
>
> All patches except one(see below) are only mlog related, so they
> shouldn't affect the normal operations in ocfs2 and they should be easy
> for review(patch 1-5 were already sent to the mail list around one month
> ago).
>
> The only one that changes the codes somehow are patch 15, "ocfs2: Little
> refactoring against ocfs2_iget", but it looks straightforward.
>
> Regards,
> Tao
>
> _______________________________________________
> Ocfs2-devel mailing list
> Ocfs2-devel at oss.oracle.com
> http://oss.oracle.com/mailman/listinfo/ocfs2-devel
>
^ permalink raw reply [flat|nested] 53+ messages in thread
end of thread, other threads:[~2011-02-14 2:53 UTC | newest]
Thread overview: 53+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2010-12-23 7:19 [Ocfs2-devel] [PATCH 00/34] OCFS2: Add trace event and replace mlog(0) Tao Ma
2010-12-23 7:30 ` [Ocfs2-devel] [PATCH 01/34] ocfs2: Remove unused truncate function from alloc.c Tao Ma
2010-12-28 22:54 ` Mark Fasheh
2010-12-23 7:30 ` [Ocfs2-devel] [PATCH 02/34] ocfs2: Remove ENTRY from masklog Tao Ma
2010-12-23 7:30 ` [Ocfs2-devel] [PATCH 03/34] ocfs2: Remove EXIT " Tao Ma
2010-12-23 7:30 ` [Ocfs2-devel] [PATCH 04/34] ocfs2: Add ocfs2_trace.h Tao Ma
2010-12-31 12:40 ` Joel Becker
2010-12-23 7:30 ` [Ocfs2-devel] [PATCH 05/34] ocfs2: Remove mlog(0) from fs/ocfs2/alloc.c Tao Ma
2010-12-31 13:15 ` Christoph Hellwig
2010-12-31 14:27 ` Tao Ma
2010-12-23 7:30 ` [Ocfs2-devel] [PATCH 06/34] ocfs2: Remove mlog(0) from fs/ocfs2/localalloc.c Tao Ma
2010-12-23 7:30 ` [Ocfs2-devel] [PATCH 07/34] ocfs2: Remove mlog(0) from fs/ocfs2/suballoc.c Tao Ma
2010-12-23 7:30 ` [Ocfs2-devel] [PATCH 08/34] " Tao Ma
2010-12-23 7:30 ` [Ocfs2-devel] [PATCH 09/34] ocfs2: Remove DISK_ALLOC from masklog Tao Ma
2010-12-23 7:30 ` [Ocfs2-devel] [PATCH 10/34] ocfs2: Remove masklog ML_REFCOUNT Tao Ma
2010-12-23 7:30 ` [Ocfs2-devel] [PATCH 11/34] ocfs2: Remove mlog(0) from fs/ocfs2/aops.c Tao Ma
2010-12-23 7:30 ` [Ocfs2-devel] [PATCH 12/34] ocfs2: Remove FILE_IO from masklog Tao Ma
2010-12-23 7:30 ` [Ocfs2-devel] [PATCH 13/34] ocfs2: remove INODE from unused files Tao Ma
2010-12-23 7:30 ` [Ocfs2-devel] [PATCH 14/34] ocfs2: Remove mlog(0) from fs/ocfs2/file.c Tao Ma
2010-12-23 7:30 ` [Ocfs2-devel] [PATCH 15/34] ocfs2: Little refactoring against ocfs2_iget Tao Ma
2010-12-23 7:30 ` [Ocfs2-devel] [PATCH 16/34] ocfs2: Remove masklog ML_INODE Tao Ma
2010-12-23 7:31 ` [Ocfs2-devel] [PATCH 17/34] ocfs2: Remove masklog ML_EXTENT_MAP Tao Ma
2010-12-23 7:31 ` [Ocfs2-devel] [PATCH 18/34] ocfs2: Remove mlog(0) from fs/ocfs2/slot_map.c Tao Ma
2010-12-23 7:31 ` [Ocfs2-devel] [PATCH 19/34] ocfs2: Remove mlog(0) from fs/ocfs2/heartbeat.c Tao Ma
2010-12-23 7:31 ` [Ocfs2-devel] [PATCH 20/34] ocfs2: Remove masklog ML_SUPER Tao Ma
2010-12-23 7:31 ` [Ocfs2-devel] [PATCH 21/34] ocfs2: Remove masklog ML_XATTR Tao Ma
2010-12-23 7:31 ` [Ocfs2-devel] [PATCH 22/34] ocfs2: Remove masklog ML_RESERVATIONS Tao Ma
2010-12-23 7:31 ` [Ocfs2-devel] [PATCH 23/34] ocfs2: Remove mlog(0) from quota_local.c Tao Ma
2010-12-23 7:31 ` [Ocfs2-devel] [PATCH 24/34] ocfs2: Remove masklog ML_QUOTA Tao Ma
2010-12-23 7:31 ` [Ocfs2-devel] [PATCH 25/34] ocfs2: remove NAMEI from symlink.c Tao Ma
2010-12-23 7:31 ` [Ocfs2-devel] [PATCH 26/34] ocfs2: Remove mlog(0) from fs/ocfs2/dir.c Tao Ma
2010-12-23 7:31 ` [Ocfs2-devel] [PATCH 27/34] ocfs2: Remove masklog ML_NAMEI Tao Ma
2010-12-23 7:31 ` [Ocfs2-devel] [PATCH 28/34] ocfs2: Remove masklog ML_DCACHE Tao Ma
2010-12-23 7:31 ` [Ocfs2-devel] [PATCH 29/34] ocfs2: Remove masklog ML_EXPORT Tao Ma
2010-12-23 7:31 ` [Ocfs2-devel] [PATCH 30/34] ocfs2: Remove masklog ML_JOURNAL Tao Ma
2010-12-23 7:31 ` [Ocfs2-devel] [PATCH 31/34] ocfs2: Remove masklog ML_BH_IO Tao Ma
2010-12-23 7:31 ` [Ocfs2-devel] [PATCH 32/34] ocfs2: Remove masklog ML_UPTODATE Tao Ma
2010-12-23 7:31 ` [Ocfs2-devel] [PATCH 33/34] ocfs2: Remove masklog ML_AIO Tao Ma
2010-12-23 7:31 ` [Ocfs2-devel] [PATCH 34/34] ocfs2: Make the left masklogs compat Tao Ma
2010-12-23 7:32 ` [Ocfs2-devel] [PATCH 00/34] OCFS2: Add trace event and replace mlog(0) Tao Ma
2010-12-23 8:44 ` Tao Ma
2010-12-31 12:52 ` Joel Becker
2010-12-31 15:11 ` Tao Ma
2010-12-31 22:39 ` Joel Becker
2011-01-04 9:06 ` Tao Ma
2011-01-04 18:43 ` Sunil Mushran
2011-01-04 22:15 ` Joel Becker
2011-01-04 22:33 ` Sunil Mushran
2011-01-05 2:25 ` Tao Ma
2010-12-31 13:10 ` Christoph Hellwig
2010-12-31 13:14 ` Joel Becker
2010-12-31 13:18 ` Christoph Hellwig
2011-02-14 2:53 ` Tao Ma
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.