[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[Xen-devel] [PATCH 1/4] xen/blkfront: map REQ_FLUSH into a full barrier



From: Jeremy Fitzhardinge <jeremy.fitzhardinge@xxxxxxxxxx>

Implement a flush as a full barrier, since we have nothing weaker.

Signed-off-by: Jeremy Fitzhardinge <jeremy.fitzhardinge@xxxxxxxxxx>
Acked-by: Christoph Hellwig <hch@xxxxxx>
---
 drivers/block/xen-blkfront.c |   19 +++++--------------
 1 files changed, 5 insertions(+), 14 deletions(-)

diff --git a/drivers/block/xen-blkfront.c b/drivers/block/xen-blkfront.c
index 06e2812..3a318d8 100644
--- a/drivers/block/xen-blkfront.c
+++ b/drivers/block/xen-blkfront.c
@@ -245,14 +245,11 @@ static int blkif_ioctl(struct block_device *bdev, fmode_t 
mode,
 }
 
 /*
- * blkif_queue_request
+ * Generate a Xen blkfront IO request from a blk layer request.  Reads
+ * and writes are handled as expected.  Since we lack a loose flush
+ * request, we map flushes into a full ordered barrier.
  *
- * request block io
- *
- * id: for guest use only.
- * operation: BLKIF_OP_{READ,WRITE,PROBE}
- * buffer: buffer to read/write into. this should be a
- *   virtual address in the guest os.
+ * @req: a request struct
  */
 static int blkif_queue_request(struct request *req)
 {
@@ -289,7 +286,7 @@ static int blkif_queue_request(struct request *req)
 
        ring_req->operation = rq_data_dir(req) ?
                BLKIF_OP_WRITE : BLKIF_OP_READ;
-       if (req->cmd_flags & REQ_HARDBARRIER)
+       if (req->cmd_flags & REQ_FLUSH)
                ring_req->operation = BLKIF_OP_WRITE_BARRIER;
 
        ring_req->nr_segments = blk_rq_map_sg(req->q, req, info->sg);
@@ -1069,14 +1066,8 @@ static void blkfront_connect(struct blkfront_info *info)
         */
        info->feature_flush = 0;
 
-       /*
-        * The driver doesn't properly handled empty flushes, so
-        * lets disable barrier support for now.
-        */
-#if 0
        if (!err && barrier)
                info->feature_flush = REQ_FLUSH;
-#endif
 
        err = xlvbd_alloc_gendisk(sectors, info, binfo, sector_size);
        if (err) {
-- 
1.7.2.3


_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-devel


 


Rackspace

Lists.xenproject.org is hosted with RackSpace, monitoring our
servers 24x7x365 and backed by RackSpace's Fanatical Support®.