On Mon, Apr 25, 2011 at 4:53 PM, Vaibhav Nagarnaik
<vnagarnaik@xxxxxxxxxx> wrote:
> From: Jiaying Zhang <jiayingz@xxxxxxxxxx>
>
> These few fs tracepoints are useful while debugging latency issues in
> filesystems and were used specifically for debugging various writeback
> subsystem issues. This patch adds entry and exit tracepoints for the
> following functions, viz.:
> wait_on_buffer
> block_write_full_page
> mpage_readpages
> file_read
>
> Signed-off-by: Vaibhav Nagarnaik <vnagarnaik@xxxxxxxxxx>
> ---
> fs/buffer.c | 10 +++
> fs/mpage.c | 3 +
> include/trace/events/fs.h | 162 +++++++++++++++++++++++++++++++++++++++++++++
> mm/filemap.c | 4 +-
> 4 files changed, 178 insertions(+), 1 deletions(-)
> create mode 100644 include/trace/events/fs.h
>
> diff --git a/fs/buffer.c b/fs/buffer.c
> index a08bb8e..1c118f4 100644
> --- a/fs/buffer.c
> +++ b/fs/buffer.c
> @@ -42,6 +42,9 @@
> #include <linux/mpage.h>
> #include <linux/bit_spinlock.h>
>
> +#define CREATE_TRACE_POINTS
> +#include <trace/events/fs.h>
> +
> static int fsync_buffers_list(spinlock_t *lock, struct list_head *list);
>
> #define BH_ENTRY(list) list_entry((list), struct buffer_head, b_assoc_buffers)
> @@ -82,7 +85,9 @@ EXPORT_SYMBOL(unlock_buffer);
> */
> void __wait_on_buffer(struct buffer_head * bh)
> {
> + trace_fs_buffer_wait_enter(bh);
> wait_on_bit(&bh->b_state, BH_Lock, sleep_on_buffer, TASK_UNINTERRUPTIBLE);
> + trace_fs_buffer_wait_exit(bh);
> }
> EXPORT_SYMBOL(__wait_on_buffer);
>
> @@ -1647,6 +1652,8 @@ static int __block_write_full_page(struct inode *inode, struct page *page,
> head = page_buffers(page);
> bh = head;
>
> + trace_block_write_full_page_enter(inode, block, last_block);
> +
> /*
> * Get all the dirty buffers mapped to disk addresses and
> * handle any aliases from the underlying blockdev's mapping.
> @@ -1736,6 +1743,9 @@ done:
> * here on.
> */
> }
> +
> + trace_block_write_full_page_exit(inode, nr_underway, err);
> +
> return err;
>
> recover:
> diff --git a/fs/mpage.c b/fs/mpage.c
> index 0afc809..1c3b8e1 100644
> --- a/fs/mpage.c
> +++ b/fs/mpage.c
> @@ -28,6 +28,7 @@
> #include <linux/backing-dev.h>
> #include <linux/pagevec.h>
>
> +#include <trace/events/fs.h>
> /*
> * I/O completion handler for multipage BIOs.
> *
> @@ -373,6 +374,8 @@ mpage_readpages(struct address_space *mapping, struct list_head *pages,
> for (page_idx = 0; page_idx < nr_pages; page_idx++) {
> struct page *page = list_entry(pages->prev, struct page, lru);
>
> + if (page_idx == 0)
> + trace_mpage_readpages(page, mapping, nr_pages);
> prefetchw(&page->flags);
> list_del(&page->lru);
> if (!add_to_page_cache_lru(page, mapping,
> diff --git a/include/trace/events/fs.h b/include/trace/events/fs.h
> new file mode 100644
> index 0000000..95f7bc8
> --- /dev/null
> +++ b/include/trace/events/fs.h
> @@ -0,0 +1,162 @@
> +#undef TRACE_SYSTEM
> +#define TRACE_SYSTEM fs
> +
> +#if !defined(_TRACE_FS_H) || defined(TRACE_HEADER_MULTI_READ)
> +#define _TRACE_FS_H
> +
> +#include <linux/tracepoint.h>
> +
> +DECLARE_EVENT_CLASS(fs_buffer_wait,
> +
> + TP_PROTO(struct buffer_head *bh),
> +
> + TP_ARGS(bh),
> +
> + TP_STRUCT__entry(
> + __field( void *, bh )
> + ),
> +
> + TP_fast_assign(
> + __entry->bh = bh;
> + ),
> +
> + TP_printk("bh %p", __entry->bh)
> +);
> +
> +DEFINE_EVENT(fs_buffer_wait, fs_buffer_wait_enter,
> +
> + TP_PROTO(struct buffer_head *bh),
> +
> + TP_ARGS(bh)
> +);
> +
> +DEFINE_EVENT(fs_buffer_wait, fs_buffer_wait_exit,
> +
> + TP_PROTO(struct buffer_head *bh),
> +
> + TP_ARGS(bh)
> +);
> +
> +TRACE_EVENT(block_write_full_page_enter,
> +
> + TP_PROTO(struct inode *inode, sector_t block, sector_t last_block),
> +
> + TP_ARGS(inode, block, last_block),
> +
> + TP_STRUCT__entry(
> + __field( dev_t, dev )
> + __field( unsigned long, ino )
> + __field( sector_t, block )
> + __field( sector_t, last_block )
> + ),
> +
> + TP_fast_assign(
> + __entry->dev = inode->i_sb->s_dev;
> + __entry->ino = inode->i_ino;
> + __entry->block = block;
> + __entry->last_block = last_block;
> + ),
> +
> + TP_printk("dev %d,%d ino %lu block %lu last block %lu",
> + MAJOR(__entry->dev), MINOR(__entry->dev),
> + __entry->ino,
> + (unsigned long)__entry->block,
> + (unsigned long)__entry->last_block)
> +);
> +
> +TRACE_EVENT(block_write_full_page_exit,
> +
> + TP_PROTO(struct inode *inode, int nr_underway, int err),
> +
> + TP_ARGS(inode, nr_underway, err),
> +
> + TP_STRUCT__entry(
> + __field( dev_t, dev )
> + __field( unsigned long, ino )
> + __field( int, nr_underway )
> + __field( int, err )
> + ),
> +
> + TP_fast_assign(
> + __entry->dev = inode->i_sb->s_dev;
> + __entry->ino = inode->i_ino;
> + __entry->nr_underway = nr_underway;
> + __entry->err = err;
> + ),
> +
> + TP_printk("dev %d,%d ino %lu nr_underway %d err %d",
> + MAJOR(__entry->dev), MINOR(__entry->dev),
> + __entry->ino, __entry->nr_underway, __entry->err)
> +);
> +
> +DECLARE_EVENT_CLASS(file_read,
> + TP_PROTO(struct inode *inode, loff_t pos, size_t len),
> +
> + TP_ARGS(inode, pos, len),
> +
> + TP_STRUCT__entry(
> + __field( ino_t, ino )
> + __field( dev_t, dev )
> + __field( loff_t, pos )
> + __field( size_t, len )
> + ),
> +
> + TP_fast_assign(
> + __entry->ino = inode->i_ino;
> + __entry->dev = inode->i_sb->s_dev;
> + __entry->pos = pos;
> + __entry->len = len;
> + ),
> +
> + TP_printk("dev %d,%d ino %lu pos %llu len %lu",
> + MAJOR(__entry->dev), MINOR(__entry->dev),
> + (unsigned long) __entry->ino,
> + __entry->pos, __entry->len)
> +);
> +
> +DEFINE_EVENT(file_read, file_read_enter,
> +
> + TP_PROTO(struct inode *inode, loff_t pos, size_t len),
> +
> + TP_ARGS(inode, pos, len)
> +);
> +
> +DEFINE_EVENT(file_read, file_read_exit,
> +
> + TP_PROTO(struct inode *inode, loff_t pos, size_t len),
> +
> + TP_ARGS(inode, pos, len)
> +);
> +
> +TRACE_EVENT(mpage_readpages,
> + TP_PROTO(struct page *page, struct address_space *mapping,
> + unsigned nr_pages),
> +
> + TP_ARGS(page, mapping, nr_pages),
> +
> + TP_STRUCT__entry(
> + __field( pgoff_t, index )
> + __field( ino_t, ino )
> + __field( dev_t, dev )
> + __field( unsigned, nr_pages )
> +
> + ),
> +
> + TP_fast_assign(
> + __entry->index = page->index;
> + __entry->ino = mapping->host->i_ino;
> + __entry->dev = mapping->host->i_sb->s_dev;
> + __entry->nr_pages = nr_pages;
> + ),
> +
> + TP_printk("dev %d,%d ino %lu page_index %lu nr_pages %u",
> + MAJOR(__entry->dev), MINOR(__entry->dev),
> + (unsigned long) __entry->ino,
> + __entry->index, __entry->nr_pages)
> +);
> +
> +#endif /* _TRACE_FS_H */
> +
> +/* This part must be outside protection */
> +#include <trace/define_trace.h>
> +
> diff --git a/mm/filemap.c b/mm/filemap.c
> index c641edf..94e549c 100644
> --- a/mm/filemap.c
> +++ b/mm/filemap.c
> @@ -42,7 +42,7 @@
> #include <linux/buffer_head.h> /* for try_to_free_buffers */
>
> #include <asm/mman.h>
> -
> +#include <trace/events/fs.h>
> /*
> * Shared mappings implemented 30.11.1994. It's not fully working yet,
> * though.
> @@ -1054,6 +1054,7 @@ static void do_generic_file_read(struct file *filp, loff_t *ppos,
> unsigned int prev_offset;
> int error;
>
> + trace_file_read_enter(inode, *ppos, desc->count);
> index = *ppos >> PAGE_CACHE_SHIFT;
> prev_index = ra->prev_pos >> PAGE_CACHE_SHIFT;
> prev_offset = ra->prev_pos & (PAGE_CACHE_SIZE-1);
> @@ -1254,6 +1255,7 @@ out:
> ra->prev_pos <<= PAGE_CACHE_SHIFT;
> ra->prev_pos |= prev_offset;
>
> + trace_file_read_exit(inode, *ppos, desc->written);
> *ppos = ((loff_t)index << PAGE_CACHE_SHIFT) + offset;
> file_accessed(filp);
> }
> --
> 1.7.3.1
>
>
Hi Alexander
Can you provide your thoughts on this patch?
Thanks
Vaibhav Nagarnaik
--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@xxxxxxxxxx For more info on Linux MM,
see: http://www.linux-mm.org/ .
Fight unfair telecom internet charges in Canada: sign http://stopthemeter.ca/
Don't email: <a href