I'd like to propose a sister patch on the write part. It may not be
as easy to measure any performance impacts of it, but I'll try.
---
Subject: remove plugging at buffered write time
Date: Tue Jan 31 18:25:48 CST 2012
Buffered write(2) is not directly tied to IO, so no need to handle plug
in generic_file_aio_write().
CC: Jens Axboe <axboe@xxxxxxxxx>
CC: Li Shaohua <shaohua.li@xxxxxxxxx>
Signed-off-by: Wu Fengguang <fengguang.wu@xxxxxxxxx>
---
mm/filemap.c | 6 +++---
1 file changed, 3 insertions(+), 3 deletions(-)
--- linux-next.orig/mm/filemap.c 2012-01-31 18:23:52.000000000 +0800
+++ linux-next/mm/filemap.c 2012-01-31 18:25:38.000000000 +0800
@@ -2267,6 +2267,7 @@ generic_file_direct_write(struct kiocb *
struct file *file = iocb->ki_filp;
struct address_space *mapping = file->f_mapping;
struct inode *inode = mapping->host;
+ struct blk_plug plug;
ssize_t written;
size_t write_len;
pgoff_t end;
@@ -2301,7 +2302,9 @@ generic_file_direct_write(struct kiocb *
}
}
+ blk_start_plug(&plug);
written = mapping->a_ops->direct_IO(WRITE, iocb, iov, pos, *nr_segs);
+ blk_finish_plug(&plug);
/*
* Finally, try again to invalidate clean pages which might have been
@@ -2610,13 +2613,11 @@ ssize_t generic_file_aio_write(struct ki
{
struct file *file = iocb->ki_filp;
struct inode *inode = file->f_mapping->host;
- struct blk_plug plug;
ssize_t ret;
BUG_ON(iocb->ki_pos != pos);
mutex_lock(&inode->i_mutex);
- blk_start_plug(&plug);
ret = __generic_file_aio_write(iocb, iov, nr_segs, &iocb->ki_pos);
mutex_unlock(&inode->i_mutex);
@@ -2627,7 +2628,6 @@ ssize_t generic_file_aio_write(struct ki
if (err < 0 && ret > 0)
ret = err;
}
- blk_finish_plug(&plug);
return ret;
}
EXPORT_SYMBOL(generic_file_aio_write);
--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@xxxxxxxxx. 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=mailto:"dont@xxxxxxxxx";> email@xxxxxxxxx </a>