[PATCH 1/3] fs: allow short direct-io reads to be completed via buffered IO

[PATCH 1/3] fs: allow short direct-io reads to be completed via buffered IO

[Josef Bacik]

This is similar to what already happens in the write case.  If we have a short
read while doing O_DIRECT, instead of just returning, fallthrough and try to
read the rest via buffered IO.  BTRFS needs this because if we encounter a
compressed or inline extent during DIO, we need to fallback on buffered.  If the
extent is compressed we need to read the entire thing into memory and
de-compress it into the users pages.  I have tested this with fsx and everything
works great.  Thanks,

Signed-off-by: Josef Bacik <josef@redhat.com>
---
 mm/filemap.c |   23 ++++++++++++++++++-----
 1 files changed, 18 insertions(+), 5 deletions(-)

diff --git a/mm/filemap.c b/mm/filemap.c
index 140ebda..423b439 100644
--- a/mm/filemap.c
+++ b/mm/filemap.c
@@ -1263,7 +1263,7 @@ generic_file_aio_read(struct kiocb *iocb, const struct iovec *iov,
 {
 	struct file *filp = iocb->ki_filp;
 	ssize_t retval;
-	unsigned long seg;
+	unsigned long seg = 0;
 	size_t count;
 	loff_t *ppos = &iocb->ki_pos;
 
@@ -1290,21 +1290,34 @@ generic_file_aio_read(struct kiocb *iocb, const struct iovec *iov,
 				retval = mapping->a_ops->direct_IO(READ, iocb,
 							iov, pos, nr_segs);
 			}
-			if (retval > 0)
+			if (retval > 0) {
 				*ppos = pos + retval;
-			if (retval) {
+				count -= retval;
+			}
+			if (retval < 0 || !count) {
 				file_accessed(filp);
 				goto out;
 			}
 		}
 	}
 
+	count = retval;
 	for (seg = 0; seg < nr_segs; seg++) {
 		read_descriptor_t desc;
+		loff_t offset = 0;
+
+		if (count) {
+			if (count > iov[seg].iov_len) {
+				count -= iov[seg].iov_len;
+				continue;
+			}
+			offset = count;
+			count = 0;
+		}
 
 		desc.written = 0;
-		desc.arg.buf = iov[seg].iov_base;
-		desc.count = iov[seg].iov_len;
+		desc.arg.buf = iov[seg].iov_base + offset;
+		desc.count = iov[seg].iov_len - offset;
 		if (desc.count == 0)
 			continue;
 		desc.error = 0;
-- 
1.6.6.1

Re: [PATCH 1/3] fs: allow short direct-io reads to be completed via buffered IO

[Dave Chinner]

On Mon, May 03, 2010 at 01:27:02PM -0400, Josef Bacik wrote:
> This is similar to what already happens in the write case.  If we have a short
> read while doing O_DIRECT, instead of just returning, fallthrough and try to
> read the rest via buffered IO.  BTRFS needs this because if we encounter a
> compressed or inline extent during DIO, we need to fallback on buffered.  If the
> extent is compressed we need to read the entire thing into memory and
> de-compress it into the users pages.  I have tested this with fsx and everything
> works great.  Thanks,

Won't this mean that any direct IO read that spans EOF  (i.e. get a
short read) now attempt a buffered IO (that will fail) before returning?

Cheers,

Dave.
-- 
Dave Chinner
david@fromorbit.com

Re: [PATCH 1/3] fs: allow short direct-io reads to be completed via buffered IO

[Josef Bacik]

On Tue, May 04, 2010 at 10:14:18AM +1000, Dave Chinner wrote:
> On Mon, May 03, 2010 at 01:27:02PM -0400, Josef Bacik wrote:
> > This is similar to what already happens in the write case.  If we have a short
> > read while doing O_DIRECT, instead of just returning, fallthrough and try to
> > read the rest via buffered IO.  BTRFS needs this because if we encounter a
> > compressed or inline extent during DIO, we need to fallback on buffered.  If the
> > extent is compressed we need to read the entire thing into memory and
> > de-compress it into the users pages.  I have tested this with fsx and everything
> > works great.  Thanks,
> 
> Won't this mean that any direct IO read that spans EOF  (i.e. get a
> short read) now attempt a buffered IO (that will fail) before returning?
> 

Hmm yeah you are right.  What would be an acceptable way to avoid this, do a

if (retval || !count || ppos >= i_size_read(inode))
	goto out;

type thing?  Thanks,

Josef

Re: [PATCH 1/3] fs: allow short direct-io reads to be completed via buffered IO

[Dave Chinner]

On Tue, May 04, 2010 at 11:27:50AM -0400, Josef Bacik wrote:
> On Tue, May 04, 2010 at 10:14:18AM +1000, Dave Chinner wrote:
> > On Mon, May 03, 2010 at 01:27:02PM -0400, Josef Bacik wrote:
> > > This is similar to what already happens in the write case.  If we have a short
> > > read while doing O_DIRECT, instead of just returning, fallthrough and try to
> > > read the rest via buffered IO.  BTRFS needs this because if we encounter a
> > > compressed or inline extent during DIO, we need to fallback on buffered.  If the
> > > extent is compressed we need to read the entire thing into memory and
> > > de-compress it into the users pages.  I have tested this with fsx and everything
> > > works great.  Thanks,
> > 
> > Won't this mean that any direct IO read that spans EOF  (i.e. get a
> > short read) now attempt a buffered IO (that will fail) before returning?
> > 
> 
> Hmm yeah you are right.  What would be an acceptable way to avoid this, do a
> 
> if (retval || !count || ppos >= i_size_read(inode))
> 	goto out;
> 
> type thing?  Thanks,

Yes, that looks like it would work to me. Might be worth a comment,
though.

Cheers,

Dave.
-- 
Dave Chinner
david@fromorbit.com

[PATCH 1/3] fs: allow short direct-io reads to be completed via buffered IO

[Josef Bacik]

This is similar to what already happens in the write case.  If we have a short
read while doing O_DIRECT, instead of just returning, fallthrough and try to
read the rest via buffered IO.  BTRFS needs this because if we encounter a
compressed or inline extent during DIO, we need to fallback on buffered.  If the
extent is compressed we need to read the entire thing into memory and
de-compress it into the users pages.  I have tested this with fsx and everything
works great.  Thanks,

Signed-off-by: Josef Bacik <josef@redhat.com>
---
 mm/filemap.c |   19 +++++++++++++++----
 1 files changed, 15 insertions(+), 4 deletions(-)

diff --git a/mm/filemap.c b/mm/filemap.c
index 140ebda..cc804d9 100644
--- a/mm/filemap.c
+++ b/mm/filemap.c
@@ -1263,7 +1263,7 @@ generic_file_aio_read(struct kiocb *iocb, const struct iovec *iov,
 {
 	struct file *filp = iocb->ki_filp;
 	ssize_t retval;
-	unsigned long seg;
+	unsigned long seg = 0;
 	size_t count;
 	loff_t *ppos = &iocb->ki_pos;
 
@@ -1292,19 +1292,30 @@ generic_file_aio_read(struct kiocb *iocb, const struct iovec *iov,
 			}
 			if (retval > 0)
 				*ppos = pos + retval;
-			if (retval) {
+			if (retval < 0 || !count) {
 				file_accessed(filp);
 				goto out;
 			}
 		}
 	}
 
+	count = retval;
 	for (seg = 0; seg < nr_segs; seg++) {
 		read_descriptor_t desc;
+		loff_t offset = 0;
+
+		if (count) {
+			if (count > iov[seg].iov_len) {
+				count -= iov[seg].iov_len;
+				continue;
+			}
+			offset = count;
+			count = 0;
+		}
 
 		desc.written = 0;
-		desc.arg.buf = iov[seg].iov_base;
-		desc.count = iov[seg].iov_len;
+		desc.arg.buf = iov[seg].iov_base + offset;
+		desc.count = iov[seg].iov_len - offset;
 		if (desc.count == 0)
 			continue;
 		desc.error = 0;
-- 
1.6.6.1