xfs: remove old vmap cache
[linux-2.6.git] / fs / direct-io.c
index b05d1b2..e82adc2 100644 (file)
@@ -5,11 +5,11 @@
  *
  * O_DIRECT
  *
- * 04Jul2002   akpm@zip.com.au
+ * 04Jul2002   Andrew Morton
  *             Initial version
  * 11Sep2002   janetinc@us.ibm.com
  *             added readv/writev support.
- * 29Oct2002   akpm@zip.com.au
+ * 29Oct2002   Andrew Morton
  *             rewrote bio_add_page() support.
  * 30Oct2002   pbadari@us.ibm.com
  *             added support for non-aligned IO.
@@ -27,6 +27,7 @@
 #include <linux/slab.h>
 #include <linux/highmem.h>
 #include <linux/pagemap.h>
+#include <linux/task_io_accounting_ops.h>
 #include <linux/bio.h>
 #include <linux/wait.h>
 #include <linux/err.h>
  *
  * If blkfactor is zero then the user's request was aligned to the filesystem's
  * blocksize.
- *
- * lock_type is DIO_LOCKING for regular files on direct-IO-naive filesystems.
- * This determines whether we need to do the fancy locking which prevents
- * direct-IO from being able to read uninitialised disk blocks.  If its zero
- * (blockdev) this locking is not done, and if it is DIO_OWN_LOCKING i_mutex is
- * not held for the entire direct write (taken briefly, initially, during a
- * direct read though, but its never held for the duration of a direct-IO).
  */
 
 struct dio {
@@ -67,7 +61,7 @@ struct dio {
        struct inode *inode;
        int rw;
        loff_t i_size;                  /* i_size when submitted */
-       int lock_type;                  /* doesn't change */
+       int flags;                      /* doesn't change */
        unsigned blkbits;               /* doesn't change */
        unsigned blkfactor;             /* When we're using an alignment which
                                           is finer than the filesystem's soft
@@ -103,6 +97,18 @@ struct dio {
        unsigned cur_page_len;          /* Nr of bytes at cur_page_offset */
        sector_t cur_page_block;        /* Where it starts */
 
+       /* BIO completion state */
+       spinlock_t bio_lock;            /* protects BIO fields below */
+       unsigned long refcount;         /* direct_io_worker() and bios */
+       struct bio *bio_list;           /* singly linked via bi_private */
+       struct task_struct *waiter;     /* waiting task (NULL if none) */
+
+       /* AIO related stuff */
+       struct kiocb *iocb;             /* kiocb */
+       int is_async;                   /* is IO async ? */
+       int io_error;                   /* IO error in completion path */
+       ssize_t result;                 /* IO result */
+
        /*
         * Page fetching state. These variables belong to dio_refill_pages().
         */
@@ -114,23 +120,16 @@ struct dio {
         * Page queue.  These variables belong to dio_refill_pages() and
         * dio_get_page().
         */
-       struct page *pages[DIO_PAGES];  /* page buffer */
        unsigned head;                  /* next page to process */
        unsigned tail;                  /* last valid page + 1 */
        int page_errors;                /* errno from get_user_pages() */
 
-       /* BIO completion state */
-       spinlock_t bio_lock;            /* protects BIO fields below */
-       int bio_count;                  /* nr bios to be completed */
-       int bios_in_flight;             /* nr bios in flight */
-       struct bio *bio_list;           /* singly linked via bi_private */
-       struct task_struct *waiter;     /* waiting task (NULL if none) */
-
-       /* AIO related stuff */
-       struct kiocb *iocb;             /* kiocb */
-       int is_async;                   /* is IO async ? */
-       int io_error;                   /* IO error in completion path */
-       ssize_t result;                 /* IO result */
+       /*
+        * pages[] (and any fields placed after it) are not zeroed out at
+        * allocation time.  Don't add new fields after pages[] unless you
+        * wish that they not be zeroed.
+        */
+       struct page *pages[DIO_PAGES];  /* page buffer */
 };
 
 /*
@@ -150,20 +149,14 @@ static int dio_refill_pages(struct dio *dio)
        int nr_pages;
 
        nr_pages = min(dio->total_pages - dio->curr_page, DIO_PAGES);
-       down_read(&current->mm->mmap_sem);
-       ret = get_user_pages(
-               current,                        /* Task for fault acounting */
-               current->mm,                    /* whose pages? */
+       ret = get_user_pages_fast(
                dio->curr_user_address,         /* Where from? */
                nr_pages,                       /* How many pages? */
                dio->rw == READ,                /* Write to memory? */
-               0,                              /* force (?) */
-               &dio->pages[0],
-               NULL);                          /* vmas */
-       up_read(&current->mm->mmap_sem);
+               &dio->pages[0]);                /* Put results here */
 
-       if (ret < 0 && dio->blocks_available && (dio->rw == WRITE)) {
-               struct page *page = ZERO_PAGE(dio->curr_user_address);
+       if (ret < 0 && dio->blocks_available && (dio->rw & WRITE)) {
+               struct page *page = ZERO_PAGE(0);
                /*
                 * A memory fault, but the filesystem has some outstanding
                 * mapped blocks.  We need to use those blocks up to avoid
@@ -209,91 +202,82 @@ static struct page *dio_get_page(struct dio *dio)
        return dio->pages[dio->head++];
 }
 
-/*
- * Called when all DIO BIO I/O has been completed - let the filesystem
- * know, if it registered an interest earlier via get_block.  Pass the
- * private field of the map buffer_head so that filesystems can use it
- * to hold additional state between get_block calls and dio_complete.
- */
-static void dio_complete(struct dio *dio, loff_t offset, ssize_t bytes)
-{
-       if (dio->end_io && dio->result)
-               dio->end_io(dio->iocb, offset, bytes, dio->map_bh.b_private);
-       if (dio->lock_type == DIO_LOCKING)
-               up_read(&dio->inode->i_alloc_sem);
-}
-
-/*
- * Called when a BIO has been processed.  If the count goes to zero then IO is
- * complete and we can signal this to the AIO layer.
+/**
+ * dio_complete() - called when all DIO BIO I/O has been completed
+ * @offset: the byte offset in the file of the completed operation
+ *
+ * This releases locks as dictated by the locking type, lets interested parties
+ * know that a DIO operation has completed, and calculates the resulting return
+ * code for the operation.
+ *
+ * It lets the filesystem know if it registered an interest earlier via
+ * get_block.  Pass the private field of the map buffer_head so that
+ * filesystems can use it to hold additional state between get_block calls and
+ * dio_complete.
  */
-static void finished_one_bio(struct dio *dio)
+static int dio_complete(struct dio *dio, loff_t offset, int ret)
 {
-       unsigned long flags;
+       ssize_t transferred = 0;
 
-       spin_lock_irqsave(&dio->bio_lock, flags);
-       if (dio->bio_count == 1) {
-               if (dio->is_async) {
-                       ssize_t transferred;
-                       loff_t offset;
+       /*
+        * AIO submission can race with bio completion to get here while
+        * expecting to have the last io completed by bio completion.
+        * In that case -EIOCBQUEUED is in fact not an error we want
+        * to preserve through this call.
+        */
+       if (ret == -EIOCBQUEUED)
+               ret = 0;
 
-                       /*
-                        * Last reference to the dio is going away.
-                        * Drop spinlock and complete the DIO.
-                        */
-                       spin_unlock_irqrestore(&dio->bio_lock, flags);
+       if (dio->result) {
+               transferred = dio->result;
 
-                       /* Check for short read case */
-                       transferred = dio->result;
-                       offset = dio->iocb->ki_pos;
+               /* Check for short read case */
+               if ((dio->rw == READ) && ((offset + transferred) > dio->i_size))
+                       transferred = dio->i_size - offset;
+       }
 
-                       if ((dio->rw == READ) &&
-                           ((offset + transferred) > dio->i_size))
-                               transferred = dio->i_size - offset;
+       if (dio->end_io && dio->result)
+               dio->end_io(dio->iocb, offset, transferred,
+                           dio->map_bh.b_private);
 
-                       /* check for error in completion path */
-                       if (dio->io_error)
-                               transferred = dio->io_error;
+       if (dio->flags & DIO_LOCKING)
+               /* lockdep: non-owner release */
+               up_read_non_owner(&dio->inode->i_alloc_sem);
 
-                       dio_complete(dio, offset, transferred);
+       if (ret == 0)
+               ret = dio->page_errors;
+       if (ret == 0)
+               ret = dio->io_error;
+       if (ret == 0)
+               ret = transferred;
 
-                       /* Complete AIO later if falling back to buffered i/o */
-                       if (dio->result == dio->size ||
-                               ((dio->rw == READ) && dio->result)) {
-                               aio_complete(dio->iocb, transferred, 0);
-                               kfree(dio);
-                               return;
-                       } else {
-                               /*
-                                * Falling back to buffered
-                                */
-                               spin_lock_irqsave(&dio->bio_lock, flags);
-                               dio->bio_count--;
-                               if (dio->waiter)
-                                       wake_up_process(dio->waiter);
-                               spin_unlock_irqrestore(&dio->bio_lock, flags);
-                               return;
-                       }
-               }
-       }
-       dio->bio_count--;
-       spin_unlock_irqrestore(&dio->bio_lock, flags);
+       return ret;
 }
 
 static int dio_bio_complete(struct dio *dio, struct bio *bio);
 /*
  * Asynchronous IO callback. 
  */
-static int dio_bio_end_aio(struct bio *bio, unsigned int bytes_done, int error)
+static void dio_bio_end_aio(struct bio *bio, int error)
 {
        struct dio *dio = bio->bi_private;
-
-       if (bio->bi_size)
-               return 1;
+       unsigned long remaining;
+       unsigned long flags;
 
        /* cleanup the bio */
        dio_bio_complete(dio, bio);
-       return 0;
+
+       spin_lock_irqsave(&dio->bio_lock, flags);
+       remaining = --dio->refcount;
+       if (remaining == 1 && dio->waiter)
+               wake_up_process(dio->waiter);
+       spin_unlock_irqrestore(&dio->bio_lock, flags);
+
+       if (remaining == 0) {
+               int ret = dio_complete(dio, dio->iocb->ki_pos, 0);
+               aio_complete(dio->iocb, ret, 0);
+               kfree(dio);
+       }
 }
 
 /*
@@ -303,22 +287,17 @@ static int dio_bio_end_aio(struct bio *bio, unsigned int bytes_done, int error)
  * During I/O bi_private points at the dio.  After I/O, bi_private is used to
  * implement a singly-linked list of completed BIOs, at dio->bio_list.
  */
-static int dio_bio_end_io(struct bio *bio, unsigned int bytes_done, int error)
+static void dio_bio_end_io(struct bio *bio, int error)
 {
        struct dio *dio = bio->bi_private;
        unsigned long flags;
 
-       if (bio->bi_size)
-               return 1;
-
        spin_lock_irqsave(&dio->bio_lock, flags);
        bio->bi_private = dio->bio_list;
        dio->bio_list = bio;
-       dio->bios_in_flight--;
-       if (dio->waiter && dio->bios_in_flight == 0)
+       if (--dio->refcount == 1 && dio->waiter)
                wake_up_process(dio->waiter);
        spin_unlock_irqrestore(&dio->bio_lock, flags);
-       return 0;
 }
 
 static int
@@ -328,8 +307,6 @@ dio_bio_alloc(struct dio *dio, struct block_device *bdev,
        struct bio *bio;
 
        bio = bio_alloc(GFP_KERNEL, nr_vecs);
-       if (bio == NULL)
-               return -ENOMEM;
 
        bio->bi_bdev = bdev;
        bio->bi_sector = first_sector;
@@ -346,6 +323,8 @@ dio_bio_alloc(struct dio *dio, struct block_device *bdev,
  * In the AIO read case we speculatively dirty the pages before starting IO.
  * During IO completion, any of these pages which happen to have been written
  * back will be redirtied by bio_check_pages_dirty().
+ *
+ * bios hold a dio reference between submit_bio and ->end_io.
  */
 static void dio_bio_submit(struct dio *dio)
 {
@@ -353,12 +332,14 @@ static void dio_bio_submit(struct dio *dio)
        unsigned long flags;
 
        bio->bi_private = dio;
+
        spin_lock_irqsave(&dio->bio_lock, flags);
-       dio->bio_count++;
-       dio->bios_in_flight++;
+       dio->refcount++;
        spin_unlock_irqrestore(&dio->bio_lock, flags);
+
        if (dio->is_async && dio->rw == READ)
                bio_set_pages_dirty(bio);
+
        submit_bio(dio->rw, bio);
 
        dio->bio = NULL;
@@ -375,28 +356,37 @@ static void dio_cleanup(struct dio *dio)
 }
 
 /*
- * Wait for the next BIO to complete.  Remove it and return it.
+ * Wait for the next BIO to complete.  Remove it and return it.  NULL is
+ * returned once all BIOs have been completed.  This must only be called once
+ * all bios have been issued so that dio->refcount can only decrease.  This
+ * requires that that the caller hold a reference on the dio.
  */
 static struct bio *dio_await_one(struct dio *dio)
 {
        unsigned long flags;
-       struct bio *bio;
+       struct bio *bio = NULL;
 
        spin_lock_irqsave(&dio->bio_lock, flags);
-       while (dio->bio_list == NULL) {
-               set_current_state(TASK_UNINTERRUPTIBLE);
-               if (dio->bio_list == NULL) {
-                       dio->waiter = current;
-                       spin_unlock_irqrestore(&dio->bio_lock, flags);
-                       blk_run_address_space(dio->inode->i_mapping);
-                       io_schedule();
-                       spin_lock_irqsave(&dio->bio_lock, flags);
-                       dio->waiter = NULL;
-               }
-               set_current_state(TASK_RUNNING);
+
+       /*
+        * Wait as long as the list is empty and there are bios in flight.  bio
+        * completion drops the count, maybe adds to the list, and wakes while
+        * holding the bio_lock so we don't need set_current_state()'s barrier
+        * and can call it after testing our condition.
+        */
+       while (dio->refcount > 1 && dio->bio_list == NULL) {
+               __set_current_state(TASK_UNINTERRUPTIBLE);
+               dio->waiter = current;
+               spin_unlock_irqrestore(&dio->bio_lock, flags);
+               io_schedule();
+               /* wake up sets us TASK_RUNNING */
+               spin_lock_irqsave(&dio->bio_lock, flags);
+               dio->waiter = NULL;
+       }
+       if (dio->bio_list) {
+               bio = dio->bio_list;
+               dio->bio_list = bio->bi_private;
        }
-       bio = dio->bio_list;
-       dio->bio_list = bio->bi_private;
        spin_unlock_irqrestore(&dio->bio_lock, flags);
        return bio;
 }
@@ -425,34 +415,24 @@ static int dio_bio_complete(struct dio *dio, struct bio *bio)
                }
                bio_put(bio);
        }
-       finished_one_bio(dio);
        return uptodate ? 0 : -EIO;
 }
 
 /*
- * Wait on and process all in-flight BIOs.
+ * Wait on and process all in-flight BIOs.  This must only be called once
+ * all bios have been issued so that the refcount can only decrease.
+ * This just waits for all bios to make it through dio_bio_complete.  IO
+ * errors are propagated through dio->io_error and should be propagated via
+ * dio_complete().
  */
-static int dio_await_completion(struct dio *dio)
+static void dio_await_completion(struct dio *dio)
 {
-       int ret = 0;
-
-       if (dio->bio)
-               dio_bio_submit(dio);
-
-       /*
-        * The bio_lock is not held for the read of bio_count.
-        * This is ok since it is the dio_bio_complete() that changes
-        * bio_count.
-        */
-       while (dio->bio_count) {
-               struct bio *bio = dio_await_one(dio);
-               int ret2;
-
-               ret2 = dio_bio_complete(dio, bio);
-               if (ret == 0)
-                       ret = ret2;
-       }
-       return ret;
+       struct bio *bio;
+       do {
+               bio = dio_await_one(dio);
+               if (bio)
+                       dio_bio_complete(dio, bio);
+       } while (bio);
 }
 
 /*
@@ -535,21 +515,24 @@ static int get_more_blocks(struct dio *dio)
                map_bh->b_state = 0;
                map_bh->b_size = fs_count << dio->inode->i_blkbits;
 
-               create = dio->rw == WRITE;
-               if (dio->lock_type == DIO_LOCKING) {
+               /*
+                * For writes inside i_size on a DIO_SKIP_HOLES filesystem we
+                * forbid block creations: only overwrites are permitted.
+                * We will return early to the caller once we see an
+                * unmapped buffer head returned, and the caller will fall
+                * back to buffered I/O.
+                *
+                * Otherwise the decision is left to the get_blocks method,
+                * which may decide to handle it or also return an unmapped
+                * buffer head.
+                */
+               create = dio->rw & WRITE;
+               if (dio->flags & DIO_SKIP_HOLES) {
                        if (dio->block_in_file < (i_size_read(dio->inode) >>
                                                        dio->blkbits))
                                create = 0;
-               } else if (dio->lock_type == DIO_NO_LOCKING) {
-                       create = 0;
                }
 
-               /*
-                * For writes inside i_size we forbid block creations: only
-                * overwrites are permitted.  We fall back to buffered writes
-                * at a higher level for inside-i_size block-instantiating
-                * writes.
-                */
                ret = (*dio->get_block)(dio->inode, fs_startblk,
                                                map_bh, create);
        }
@@ -674,6 +657,13 @@ submit_page_section(struct dio *dio, struct page *page,
 {
        int ret = 0;
 
+       if (dio->rw & WRITE) {
+               /*
+                * Read accounting is performed in submit_bio()
+                */
+               task_io_account_write(len);
+       }
+
        /*
         * Can we just grow the current page's presence in the dio?
         */
@@ -768,7 +758,7 @@ static void dio_zero_block(struct dio *dio, int end)
 
        this_chunk_bytes = this_chunk_blocks << dio->blkbits;
 
-       page = ZERO_PAGE(dio->curr_user_address);
+       page = ZERO_PAGE(0);
        if (submit_page_section(dio, page, 0, this_chunk_bytes, 
                                dio->next_block_for_io))
                return;
@@ -863,11 +853,10 @@ static int do_direct_IO(struct dio *dio)
 do_holes:
                        /* Handle holes */
                        if (!buffer_mapped(map_bh)) {
-                               char *kaddr;
                                loff_t i_size_aligned;
 
                                /* AKPM: eargh, -ENOTBLK is a hack */
-                               if (dio->rw == WRITE) {
+                               if (dio->rw & WRITE) {
                                        page_cache_release(page);
                                        return -ENOTBLK;
                                }
@@ -884,11 +873,8 @@ do_holes:
                                        page_cache_release(page);
                                        goto out;
                                }
-                               kaddr = kmap_atomic(page, KM_USER0);
-                               memset(kaddr + (block_in_page << blkbits),
-                                               0, 1 << blkbits);
-                               flush_dcache_page(page);
-                               kunmap_atomic(kaddr, KM_USER0);
+                               zero_user(page, block_in_page << blkbits,
+                                               1 << blkbits);
                                dio->block_in_file++;
                                block_in_page++;
                                goto next_block;
@@ -952,49 +938,28 @@ direct_io_worker(int rw, struct kiocb *iocb, struct inode *inode,
        struct dio *dio)
 {
        unsigned long user_addr; 
+       unsigned long flags;
        int seg;
        ssize_t ret = 0;
        ssize_t ret2;
        size_t bytes;
 
-       dio->bio = NULL;
        dio->inode = inode;
        dio->rw = rw;
        dio->blkbits = blkbits;
        dio->blkfactor = inode->i_blkbits - blkbits;
-       dio->start_zero_done = 0;
-       dio->size = 0;
        dio->block_in_file = offset >> blkbits;
-       dio->blocks_available = 0;
-       dio->cur_page = NULL;
 
-       dio->boundary = 0;
-       dio->reap_counter = 0;
        dio->get_block = get_block;
        dio->end_io = end_io;
-       dio->map_bh.b_private = NULL;
        dio->final_block_in_bio = -1;
        dio->next_block_for_io = -1;
 
-       dio->page_errors = 0;
-       dio->io_error = 0;
-       dio->result = 0;
        dio->iocb = iocb;
        dio->i_size = i_size_read(inode);
 
-       /*
-        * BIO completion state.
-        *
-        * ->bio_count starts out at one, and we decrement it to zero after all
-        * BIOs are submitted.  This to avoid the situation where a really fast
-        * (or synchronous) device could take the count to zero while we're
-        * still submitting BIOs.
-        */
-       dio->bio_count = 1;
-       dio->bios_in_flight = 0;
        spin_lock_init(&dio->bio_lock);
-       dio->bio_list = NULL;
-       dio->waiter = NULL;
+       dio->refcount = 1;
 
        /*
         * In case of non-aligned buffers, we may need 2 more
@@ -1002,8 +967,6 @@ direct_io_worker(int rw, struct kiocb *iocb, struct inode *inode,
         */
        if (unlikely(dio->blkfactor))
                dio->pages_in_io = 2;
-       else
-               dio->pages_in_io = 0;
 
        for (seg = 0; seg < nr_segs; seg++) {
                user_addr = (unsigned long)iov[seg].iov_base;
@@ -1045,7 +1008,7 @@ direct_io_worker(int rw, struct kiocb *iocb, struct inode *inode,
                }
        } /* end iovec loop */
 
-       if (ret == -ENOTBLK && rw == WRITE) {
+       if (ret == -ENOTBLK && (rw & WRITE)) {
                /*
                 * The remaining part of the request will be
                 * be handled by buffered I/O when we return
@@ -1079,107 +1042,75 @@ direct_io_worker(int rw, struct kiocb *iocb, struct inode *inode,
         * we can let i_mutex go now that its achieved its purpose
         * of protecting us from looking up uninitialized blocks.
         */
-       if ((rw == READ) && (dio->lock_type == DIO_LOCKING))
+       if (rw == READ && (dio->flags & DIO_LOCKING))
                mutex_unlock(&dio->inode->i_mutex);
 
        /*
-        * OK, all BIOs are submitted, so we can decrement bio_count to truly
-        * reflect the number of to-be-processed BIOs.
+        * The only time we want to leave bios in flight is when a successful
+        * partial aio read or full aio write have been setup.  In that case
+        * bio completion will call aio_complete.  The only time it's safe to
+        * call aio_complete is when we return -EIOCBQUEUED, so we key on that.
+        * This had *better* be the only place that raises -EIOCBQUEUED.
         */
-       if (dio->is_async) {
-               int should_wait = 0;
+       BUG_ON(ret == -EIOCBQUEUED);
+       if (dio->is_async && ret == 0 && dio->result &&
+           ((rw & READ) || (dio->result == dio->size)))
+               ret = -EIOCBQUEUED;
 
-               if (dio->result < dio->size && rw == WRITE) {
-                       dio->waiter = current;
-                       should_wait = 1;
-               }
-               if (ret == 0)
-                       ret = dio->result;
-               finished_one_bio(dio);          /* This can free the dio */
+       if (ret != -EIOCBQUEUED) {
+               /* All IO is now issued, send it on its way */
                blk_run_address_space(inode->i_mapping);
-               if (should_wait) {
-                       unsigned long flags;
-                       /*
-                        * Wait for already issued I/O to drain out and
-                        * release its references to user-space pages
-                        * before returning to fallback on buffered I/O
-                        */
-
-                       spin_lock_irqsave(&dio->bio_lock, flags);
-                       set_current_state(TASK_UNINTERRUPTIBLE);
-                       while (dio->bio_count) {
-                               spin_unlock_irqrestore(&dio->bio_lock, flags);
-                               io_schedule();
-                               spin_lock_irqsave(&dio->bio_lock, flags);
-                               set_current_state(TASK_UNINTERRUPTIBLE);
-                       }
-                       spin_unlock_irqrestore(&dio->bio_lock, flags);
-                       set_current_state(TASK_RUNNING);
-                       kfree(dio);
-               }
-       } else {
-               ssize_t transferred = 0;
-
-               finished_one_bio(dio);
-               ret2 = dio_await_completion(dio);
-               if (ret == 0)
-                       ret = ret2;
-               if (ret == 0)
-                       ret = dio->page_errors;
-               if (dio->result) {
-                       loff_t i_size = i_size_read(inode);
+               dio_await_completion(dio);
+       }
 
-                       transferred = dio->result;
-                       /*
-                        * Adjust the return value if the read crossed a
-                        * non-block-aligned EOF.
-                        */
-                       if (rw == READ && (offset + transferred > i_size))
-                               transferred = i_size - offset;
-               }
-               dio_complete(dio, offset, transferred);
-               if (ret == 0)
-                       ret = transferred;
+       /*
+        * Sync will always be dropping the final ref and completing the
+        * operation.  AIO can if it was a broken operation described above or
+        * in fact if all the bios race to complete before we get here.  In
+        * that case dio_complete() translates the EIOCBQUEUED into the proper
+        * return code that the caller will hand to aio_complete().
+        *
+        * This is managed by the bio_lock instead of being an atomic_t so that
+        * completion paths can drop their ref and use the remaining count to
+        * decide to wake the submission path atomically.
+        */
+       spin_lock_irqsave(&dio->bio_lock, flags);
+       ret2 = --dio->refcount;
+       spin_unlock_irqrestore(&dio->bio_lock, flags);
 
-               /* We could have also come here on an AIO file extend */
-               if (!is_sync_kiocb(iocb) && rw == WRITE &&
-                   ret >= 0 && dio->result == dio->size)
-                       /*
-                        * For AIO writes where we have completed the
-                        * i/o, we have to mark the the aio complete.
-                        */
-                       aio_complete(iocb, ret, 0);
+       if (ret2 == 0) {
+               ret = dio_complete(dio, offset, ret);
                kfree(dio);
-       }
+       } else
+               BUG_ON(ret != -EIOCBQUEUED);
+
        return ret;
 }
 
 /*
  * This is a library function for use by filesystem drivers.
- * The locking rules are governed by the dio_lock_type parameter.
- *
- * DIO_NO_LOCKING (no locking, for raw block device access)
- * For writes, i_mutex is not held on entry; it is never taken.
- *
- * DIO_LOCKING (simple locking for regular files)
- * For writes we are called under i_mutex and return with i_mutex held, even
- * though it is internally dropped.
- * For reads, i_mutex is not held on entry, but it is taken and dropped before
- * returning.
  *
- * DIO_OWN_LOCKING (filesystem provides synchronisation and handling of
- *     uninitialised data, allowing parallel direct readers and writers)
- * For writes we are called without i_mutex, return without it, never touch it.
- * For reads we are called under i_mutex and return with i_mutex held, even
- * though it may be internally dropped.
+ * The locking rules are governed by the flags parameter:
+ *  - if the flags value contains DIO_LOCKING we use a fancy locking
+ *    scheme for dumb filesystems.
+ *    For writes this function is called under i_mutex and returns with
+ *    i_mutex held, for reads, i_mutex is not held on entry, but it is
+ *    taken and dropped again before returning.
+ *    For reads and writes i_alloc_sem is taken in shared mode and released
+ *    on I/O completion (which may happen asynchronously after returning to
+ *    the caller).
  *
- * Additional i_alloc_sem locking requirements described inline below.
+ *  - if the flags value does NOT contain DIO_LOCKING we don't use any
+ *    internal locking but rather rely on the filesystem to synchronize
+ *    direct I/O reads/writes versus each other and truncate.
+ *    For reads and writes both i_mutex and i_alloc_sem are not held on
+ *    entry and are never taken.
  */
 ssize_t
 __blockdev_direct_IO(int rw, struct kiocb *iocb, struct inode *inode,
        struct block_device *bdev, const struct iovec *iov, loff_t offset, 
        unsigned long nr_segs, get_block_t get_block, dio_iodone_t end_io,
-       int dio_lock_type)
+       int flags)
 {
        int seg;
        size_t size;
@@ -1190,14 +1121,12 @@ __blockdev_direct_IO(int rw, struct kiocb *iocb, struct inode *inode,
        ssize_t retval = -EINVAL;
        loff_t end = offset;
        struct dio *dio;
-       int release_i_mutex = 0;
-       int acquire_i_mutex = 0;
 
        if (rw & WRITE)
-               current->flags |= PF_SYNCWRITE;
+               rw = WRITE_ODIRECT_PLUG;
 
        if (bdev)
-               bdev_blkbits = blksize_bits(bdev_hardsect_size(bdev));
+               bdev_blkbits = blksize_bits(bdev_logical_block_size(bdev));
 
        if (offset & blocksize_mask) {
                if (bdev)
@@ -1225,43 +1154,37 @@ __blockdev_direct_IO(int rw, struct kiocb *iocb, struct inode *inode,
        retval = -ENOMEM;
        if (!dio)
                goto out;
-
        /*
-        * For block device access DIO_NO_LOCKING is used,
-        *      neither readers nor writers do any locking at all
-        * For regular files using DIO_LOCKING,
-        *      readers need to grab i_mutex and i_alloc_sem
-        *      writers need to grab i_alloc_sem only (i_mutex is already held)
-        * For regular files using DIO_OWN_LOCKING,
-        *      neither readers nor writers take any locks here
+        * Believe it or not, zeroing out the page array caused a .5%
+        * performance regression in a database benchmark.  So, we take
+        * care to only zero out what's needed.
         */
-       dio->lock_type = dio_lock_type;
-       if (dio_lock_type != DIO_NO_LOCKING) {
+       memset(dio, 0, offsetof(struct dio, pages));
+
+       dio->flags = flags;
+       if (dio->flags & DIO_LOCKING) {
                /* watch out for a 0 len io from a tricksy fs */
                if (rw == READ && end > offset) {
-                       struct address_space *mapping;
+                       struct address_space *mapping =
+                                       iocb->ki_filp->f_mapping;
 
-                       mapping = iocb->ki_filp->f_mapping;
-                       if (dio_lock_type != DIO_OWN_LOCKING) {
-                               mutex_lock(&inode->i_mutex);
-                               release_i_mutex = 1;
-                       }
+                       /* will be released by direct_io_worker */
+                       mutex_lock(&inode->i_mutex);
 
                        retval = filemap_write_and_wait_range(mapping, offset,
                                                              end - 1);
                        if (retval) {
+                               mutex_unlock(&inode->i_mutex);
                                kfree(dio);
                                goto out;
                        }
-
-                       if (dio_lock_type == DIO_OWN_LOCKING) {
-                               mutex_unlock(&inode->i_mutex);
-                               acquire_i_mutex = 1;
-                       }
                }
 
-               if (dio_lock_type == DIO_LOCKING)
-                       down_read(&inode->i_alloc_sem);
+               /*
+                * Will be released at I/O completion, possibly in a
+                * different thread.
+                */
+               down_read_non_owner(&inode->i_alloc_sem);
        }
 
        /*
@@ -1270,22 +1193,28 @@ __blockdev_direct_IO(int rw, struct kiocb *iocb, struct inode *inode,
         * even for AIO, we need to wait for i/o to complete before
         * returning in this case.
         */
-       dio->is_async = !is_sync_kiocb(iocb) && !((rw == WRITE) &&
+       dio->is_async = !is_sync_kiocb(iocb) && !((rw & WRITE) &&
                (end > i_size_read(inode)));
 
        retval = direct_io_worker(rw, iocb, inode, iov, offset,
                                nr_segs, blkbits, get_block, end_io, dio);
 
-       if (rw == READ && dio_lock_type == DIO_LOCKING)
-               release_i_mutex = 0;
+       /*
+        * In case of error extending write may have instantiated a few
+        * blocks outside i_size. Trim these off again for DIO_LOCKING.
+        *
+        * NOTE: filesystems with their own locking have to handle this
+        * on their own.
+        */
+       if (flags & DIO_LOCKING) {
+               if (unlikely((rw & WRITE) && retval < 0)) {
+                       loff_t isize = i_size_read(inode);
+                       if (end > isize)
+                               vmtruncate(inode, isize);
+               }
+       }
 
 out:
-       if (release_i_mutex)
-               mutex_unlock(&inode->i_mutex);
-       else if (acquire_i_mutex)
-               mutex_lock(&inode->i_mutex);
-       if (rw & WRITE)
-               current->flags &= ~PF_SYNCWRITE;
        return retval;
 }
 EXPORT_SYMBOL(__blockdev_direct_IO);