]> git.karo-electronics.de Git - mv-sheeva.git/blobdiff - arch/um/drivers/ubd_kern.c
Merge branch 'for-linus' of git://linux-arm.org/linux-2.6
[mv-sheeva.git] / arch / um / drivers / ubd_kern.c
index f934225fd8ef9c702c869c274df0c271c9a6a20b..aa9e926e13d73dca17015a5bd5f8e5eb76842c63 100644 (file)
@@ -451,23 +451,6 @@ static void do_ubd_request(struct request_queue * q);
 
 /* Only changed by ubd_init, which is an initcall. */
 static int thread_fd = -1;
-
-static void ubd_end_request(struct request *req, int bytes, int error)
-{
-       blk_end_request(req, error, bytes);
-}
-
-/* Callable only from interrupt context - otherwise you need to do
- * spin_lock_irq()/spin_lock_irqsave() */
-static inline void ubd_finish(struct request *req, int bytes)
-{
-       if(bytes < 0){
-               ubd_end_request(req, 0, -EIO);
-               return;
-       }
-       ubd_end_request(req, bytes, 0);
-}
-
 static LIST_HEAD(restart);
 
 /* XXX - move this inside ubd_intr. */
@@ -475,7 +458,6 @@ static LIST_HEAD(restart);
 static void ubd_handler(void)
 {
        struct io_thread_req *req;
-       struct request *rq;
        struct ubd *ubd;
        struct list_head *list, *next_ele;
        unsigned long flags;
@@ -492,10 +474,7 @@ static void ubd_handler(void)
                        return;
                }
 
-               rq = req->req;
-               rq->nr_sectors -= req->length >> 9;
-               if(rq->nr_sectors == 0)
-                       ubd_finish(rq, rq->hard_nr_sectors << 9);
+               blk_end_request(req->req, 0, req->length);
                kfree(req);
        }
        reactivate_fd(thread_fd, UBD_IRQ);
@@ -1243,27 +1222,26 @@ static void do_ubd_request(struct request_queue *q)
 {
        struct io_thread_req *io_req;
        struct request *req;
-       int n, last_sectors;
+       sector_t sector;
+       int n;
 
        while(1){
                struct ubd *dev = q->queuedata;
                if(dev->end_sg == 0){
-                       struct request *req = elv_next_request(q);
+                       struct request *req = blk_fetch_request(q);
                        if(req == NULL)
                                return;
 
                        dev->request = req;
-                       blkdev_dequeue_request(req);
                        dev->start_sg = 0;
                        dev->end_sg = blk_rq_map_sg(q, req, dev->sg);
                }
 
                req = dev->request;
-               last_sectors = 0;
+               sector = blk_rq_pos(req);
                while(dev->start_sg < dev->end_sg){
                        struct scatterlist *sg = &dev->sg[dev->start_sg];
 
-                       req->sector += last_sectors;
                        io_req = kmalloc(sizeof(struct io_thread_req),
                                         GFP_ATOMIC);
                        if(io_req == NULL){
@@ -1272,10 +1250,10 @@ static void do_ubd_request(struct request_queue *q)
                                return;
                        }
                        prepare_request(req, io_req,
-                                       (unsigned long long) req->sector << 9,
+                                       (unsigned long long)sector << 9,
                                        sg->offset, sg->length, sg_page(sg));
 
-                       last_sectors = sg->length >> 9;
+                       sector += sg->length >> 9;
                        n = os_write_file(thread_fd, &io_req,
                                          sizeof(struct io_thread_req *));
                        if(n != sizeof(struct io_thread_req *)){