# This is a BitKeeper generated patch for the following project:
# Project Name: Linux kernel tree
# This patch format is intended for GNU patch command version 2.5 or higher.
# This patch includes the following deltas:
#                  ChangeSet    1.590   -> 1.591
#         fs/reiserfs/file.c    1.9     -> 1.10
#        fs/reiserfs/inode.c    1.36    -> 1.37
#       include/linux/reiserfs_fs.h     1.22    -> 1.23
#        fs/reiserfs/super.c    1.22    -> 1.23
#       include/linux/reiserfs_fs_sb.h  1.12    -> 1.13
#       fs/reiserfs/bitmap.c    1.16    -> 1.17
#
# The following is the BitKeeper ChangeSet Log
# --------------------------------------------
# 02/09/09      [email protected]       1.591
# Implemented reiserfs_file_write(), to write large amount of data at once into files on reiserfs volumes which should boost write speed somewhat and also should be somewhat more SMP friendly.
# --------------------------------------------
#
diff -Nru a/fs/reiserfs/bitmap.c b/fs/reiserfs/bitmap.c
--- a/fs/reiserfs/bitmap.c      Mon Sep  9 09:28:34 2002
+++ b/fs/reiserfs/bitmap.c      Mon Sep  9 09:28:34 2002
@@ -738,7 +738,7 @@
    int rest = amount_needed;
    int nr_allocated;

-    while (rest > 0) {
+    while (rest > 0 && start <= finish) {
       nr_allocated = scan_bitmap (hint->th, &start, finish, 1,
                                   rest + prealloc_size, !hint->formatted_node,
                                   hint->block);
@@ -884,7 +884,9 @@
    if ( !blocks )
       return;

+    spin_lock(&sb->u.reiserfs_sb.bitmap_lock);
    sb->u.reiserfs_sb.reserved_blocks += blocks;
+    spin_unlock(&sb->u.reiserfs_sb.bitmap_lock);
}

/* Unreserve @blocks amount of blocks in fs pointed by @sb */
@@ -901,6 +903,22 @@
    if ( !blocks )
       return;

+    spin_lock(&sb->u.reiserfs_sb.bitmap_lock);
    sb->u.reiserfs_sb.reserved_blocks -= blocks;
    RFALSE( sb->u.reiserfs_sb.reserved_blocks < 0, "amount of blocks reserved became zero?");
+    spin_unlock(&sb->u.reiserfs_sb.bitmap_lock);
+}
+
+/* This function estimates how much pages we will be able to write to FS
+   used for reiserfs_file_write() purposes for now. */
+int reiserfs_can_fit_pages ( struct super_block *sb /* superblock of filesystem
+                                                      to estimate space */ )
+{
+       unsigned long space;
+
+       spin_lock(&sb->u.reiserfs_sb.bitmap_lock);
+       space = (SB_FREE_BLOCKS(sb) - sb->u.reiserfs_sb.reserved_blocks) / ( PAGE_CACHE_SIZE/sb->s_blocksize);
+       spin_unlock(&sb->u.reiserfs_sb.bitmap_lock);
+
+       return space;
}
diff -Nru a/fs/reiserfs/file.c b/fs/reiserfs/file.c
--- a/fs/reiserfs/file.c        Mon Sep  9 09:28:34 2002
+++ b/fs/reiserfs/file.c        Mon Sep  9 09:28:34 2002
@@ -6,6 +6,9 @@
#include <linux/sched.h>
#include <linux/reiserfs_fs.h>
#include <linux/smp_lock.h>
+#include <asm/uaccess.h>
+#include <linux/pagemap.h>
+#include <linux/locks.h>

/*
** We pack the tails of files on file close, not at the time they are written.
@@ -129,9 +132,1080 @@
    return error ;
}

+/* this function from inode.c would be used here, too */
+extern void restart_transaction(struct reiserfs_transaction_handle *th,
+                                struct inode *inode, struct path *path);
+
+/* I really do not want to play with memory shortage right now, so
+   to simplify the code, we are not going to write more than this much pages at
+   a time. This still should considerably improve performance compared to 4k
+   at a time case. */
+#define REISERFS_WRITE_PAGES_AT_A_TIME 32
+
+/* Allocates blocks for a file to fulfil write request.
+   Maps all unmapped but prepared pages from the list.
+   Updates metadata with newly allocated blocknumbers as needed */
+int reiserfs_allocate_blocks_for_region(
+                               struct inode *inode, /* Inode we work with */
+                               loff_t pos, /* Writing position */
+                               int num_pages, /* number of pages write going
+                                                 to touch */
+                               int write_bytes, /* amount of bytes to write */
+                               struct page **prepared_pages, /* array of
+                                                                prepared pages
+                                                              */
+                               int blocks_to_allocate /* Amount of blocks we
+                                                         need to allocate to
+                                                         fit the data into file
+                                                        */
+                               )
+{
+    struct cpu_key key; // cpu key of item that we are going to deal with
+    struct item_head *ih; // pointer to item head that we are going to deal with
+    struct buffer_head *bh; // Buffer head that contains items that we are going to deal with
+    struct reiserfs_transaction_handle th; // transaction handle for transaction we are going to create.
+    __u32 * item; // pointer to item we are going to deal with
+    INITIALIZE_PATH(path); // path to item, that we are going to deal with.
+    b_blocknr_t allocated_blocks[blocks_to_allocate]; // Pointer to a place where allocated blocknumbers would be stored. Right now statically allocated, later that will change.
+    reiserfs_blocknr_hint_t hint; // hint structure for block allocator.
+    size_t res; // return value of various functions that we call.
+    int curr_block; // current block used to keep track of unmapped blocks.
+    int i; // loop counter
+    int itempos; // position in item
+    unsigned int from = (pos & (PAGE_CACHE_SIZE - 1)); // writing position in
+                                                      // first page
+    unsigned int to = ((pos + write_bytes - 1) & (PAGE_CACHE_SIZE - 1)) + 1; /* last modified byte offset in last page */
+    __u64 hole_size ; // amount of blocks for a file hole, if it needed to be created.
+    int modifying_this_item = 0; // Flag for items traversal code to keep track
+                                // of the fact that we already prepared
+                                // current block for journal
+
+
+    RFALSE(!blocks_to_allocate, "green-9004: tried to allocate zero blocks?");
+
+    /* First we compose a key to point at the writing position, we want to do
+       that outside of any locking region. */
+    make_cpu_key (&key, inode, pos+1, TYPE_ANY, 3/*key length*/);
+
+    /* If we came here, it means we absolutely need to open a transaction,
+       since we need to allocate some blocks */
+    lock_kernel(); // Journaling stuff and we need that.
+    journal_begin(&th, inode->i_sb, JOURNAL_PER_BALANCE_CNT * 3 + 1); // Wish I know if this number enough
+    reiserfs_update_inode_transaction(inode) ;
+
+    /* Look for the in-tree position of our write, need path for block allocator */
+    res = search_for_position_by_key(inode->i_sb, &key, &path);
+    if ( res == IO_ERROR ) {
+       res = -EIO;
+       goto error_exit;
+    }
+
+    /* Allocate blocks */
+    /* First fill in "hint" structure for block allocator */
+    hint.th = &th; // transaction handle.
+    hint.path = &path; // Path, so that block allocator can determine packing locality or whatever it needs to determine.
+    hint.inode = inode; // Inode is needed by block allocator too.
+    hint.search_start = 0; // We have no hint on where to search free blocks for block allocator.
+    hint.key = key.on_disk_key; // on disk key of file.
+    hint.block = inode->i_blocks/(inode->i_sb->s_blocksize/512); // Number of disk blocks this file occupies already.
+    hint.formatted_node = 0; // We are allocating blocks for unformatted node.
+    hint.preallocate = 0; // We do not do any preallocation for now.
+
+    /* Call block allocator to allocate blocks */
+    res = reiserfs_allocate_blocknrs(&hint, allocated_blocks, blocks_to_allocate, blocks_to_allocate);
+    if ( res != CARRY_ON ) {
+       if ( res == NO_DISK_SPACE ) {
+           /* We flush the transaction in case of no space. This way some
+              blocks might become free */
+           SB_JOURNAL(inode->i_sb)->j_must_wait = 1;
+           restart_transaction(&th, inode, &path);
+
+           /* We might have scheduled, so search again */
+           res = search_for_position_by_key(inode->i_sb, &key, &path);
+           if ( res == IO_ERROR ) {
+               res = -EIO;
+               goto error_exit;
+           }
+
+           /* update changed info for hint structure. */
+           res = reiserfs_allocate_blocknrs(&hint, allocated_blocks, blocks_to_allocate, blocks_to_allocate);
+           if ( res != CARRY_ON ) {
+               res = -ENOSPC;
+               pathrelse(&path);
+               goto error_exit;
+           }
+       } else {
+           res = -ENOSPC;
+           pathrelse(&path);
+           goto error_exit;
+       }
+    }
+
+#ifdef __BIG_ENDIAN
+        // Too bad, I have not found any way to convert a given region from
+        // cpu format to little endian format
+    {
+        int i;
+        for ( i = 0; i < blocks_to_allocate ; i++)
+            allocated_blocks[i]=cpu_to_le32(allocated_blocks[i]);
+    }
+#endif
+
+    /* Blocks allocating well might have scheduled and tree might have changed,
+       let's search the tree again */
+    /* find where in the tree our write should go */
+    res = search_for_position_by_key(inode->i_sb, &key, &path);
+    if ( res == IO_ERROR ) {
+       res = -EIO;
+       goto error_exit_free_blocks;
+    }
+
+    bh = get_last_bh( &path ); // Get a bufferhead for last element in path.
+    ih = get_ih( &path );      // Get a pointer to last item head in path.
+    item = get_item( &path );  // Get a pointer to last item in path
+
+    /* Let's see what we have found */
+    if ( res != POSITION_FOUND ) { /* position not found, this means that we
+                                     might need to append file with holes
+                                     first */
+       // Since we are writing past the file's end, we need to find out if
+       // there is a hole that needs to be inserted before our writing
+       // position, and how many blocks it is going to cover (we need to
+       //  populate pointers to file blocks representing the hole with zeros)
+
+       hole_size = (pos + 1 - (le_key_k_offset( get_inode_item_key_version(inode), &(ih->ih_key))+op_bytes_number(ih, inode->i_sb->s_blocksize))) >> inode->i_sb->s_blocksize_bits;
+
+       if ( hole_size > 0 ) {
+           int to_paste = min_t(__u64, hole_size, MAX_ITEM_LEN(inode->i_sb->s_blocksize)/UNFM_P_SIZE ); // How much data to insert first time.
+           /* area filled with zeroes, to supply as list of zero blocknumbers
+              We allocate it outside of loop just in case loop would spin for
+              several iterations. */
+           char *zeros = kmalloc(to_paste*UNFM_P_SIZE, GFP_ATOMIC); // We cannot insert more than MAX_ITEM_LEN bytes anyway.
+           if ( !zeros ) {
+               res = -ENOMEM;
+               goto error_exit_free_blocks;
+           }
+           memset ( zeros, 0, to_paste*UNFM_P_SIZE);
+           do {
+               to_paste = min_t(__u64, hole_size, MAX_ITEM_LEN(inode->i_sb->s_blocksize)/UNFM_P_SIZE );
+               if ( is_indirect_le_ih(ih) ) {
+                   /* Ok, there is existing indirect item already. Need to append it */
+                   /* Calculate position past inserted item */
+                   make_cpu_key( &key, inode, le_key_k_offset( get_inode_item_key_version(inode), &(ih->ih_key)) + op_bytes_number(ih, inode->i_sb->s_blocksize), TYPE_INDIRECT, 3);
+                   res = reiserfs_paste_into_item( &th, &path, &key, (char *)zeros, UNFM_P_SIZE*to_paste);
+                   if ( res ) {
+                       kfree(zeros);
+                       goto error_exit_free_blocks;
+                   }
+               } else if ( is_statdata_le_ih(ih) ) {
+                   /* No existing item, create it */
+                   /* item head for new item */
+                   struct item_head ins_ih;
+
+                   /* create a key for our new item */
+                   make_cpu_key( &key, inode, 1, TYPE_INDIRECT, 3);
+
+                   /* Create new item head for our new item */
+                   make_le_item_head (&ins_ih, &key, key.version, 1,
+                                      TYPE_INDIRECT, to_paste*UNFM_P_SIZE,
+                                      0 /* free space */);
+
+                   /* Find where such item should live in the tree */
+                   res = search_item (inode->i_sb, &key, &path);
+                   if ( res != ITEM_NOT_FOUND ) {
+                       /* item should not exist, otherwise we have error */
+                       if ( res != -ENOSPC ) {
+                           reiserfs_warning ("green-9008: search_by_key (%K) returned %d\n",
+                                              &key, res);
+                       }
+                       res = -EIO;
+                       kfree(zeros);
+                       goto error_exit_free_blocks;
+                   }
+                   res = reiserfs_insert_item( &th, &path, &key, &ins_ih, (char *)zeros);
+               } else {
+                   reiserfs_panic(inode->i_sb, "green-9011: Unexpected key type %K\n", &key);
+               }
+               if ( res ) {
+                   kfree(zeros);
+                   goto error_exit_free_blocks;
+               }
+               /* Now we want to check if transaction is too full, and if it is
+                  we restart it. This will also free the path. */
+               if (journal_transaction_should_end(&th, th.t_blocks_allocated))
+                   restart_transaction(&th, inode, &path);
+
+               /* Well, need to recalculate path and stuff */
+               set_cpu_key_k_offset( &key, cpu_key_k_offset(&key) + to_paste * inode->i_sb->s_blocksize );
+               res = search_for_position_by_key(inode->i_sb, &key, &path);
+               if ( res == IO_ERROR ) {
+                   res = -EIO;
+                   kfree(zeros);
+                   goto error_exit_free_blocks;
+               }
+               bh=get_last_bh(&path);
+               ih=get_ih(&path);
+               item = get_item(&path);
+               hole_size -= to_paste;
+           } while ( hole_size );
+           kfree(zeros);
+       }
+    }
+
+    // Go through existing indirect items first
+    // replace all zeroes with blocknumbers from list
+    // Note that if no corresponding item was found, by previous search,
+    // it means there are no existing in-tree representation for file area
+    // we are going to overwrite, so there is nothing to scan through for holes.
+    for ( curr_block = 0, itempos = path.pos_in_item ; curr_block < blocks_to_allocate && res == POSITION_FOUND ; ) {
+
+       if ( itempos >= ih_item_len(ih)/UNFM_P_SIZE ) {
+           /* We run out of data in this indirect item, let's look for another
+              one. */
+           /* First if we are already modifying current item, log it */
+           if ( modifying_this_item ) {
+               journal_mark_dirty (&th, inode->i_sb, bh);
+               modifying_this_item = 0;
+           }
+           /* Then set the key to look for a new indirect item (offset of old
+              item is added to old item length */
+           set_cpu_key_k_offset( &key, le_key_k_offset( get_inode_item_key_version(inode), &(ih->ih_key)) + op_bytes_number(ih, inode->i_sb->s_blocksize));
+           /* Search ofor position of new key in the tree. */
+           res = search_for_position_by_key(inode->i_sb, &key, &path);
+           if ( res == IO_ERROR) {
+               res = -EIO;
+               goto error_exit_free_blocks;
+           }
+           bh=get_last_bh(&path);
+           ih=get_ih(&path);
+           item = get_item(&path);
+           itempos = path.pos_in_item;
+           continue; // loop to check all kinds of conditions and so on.
+       }
+       /* Ok, we have correct position in item now, so let's see if it is
+          representing file hole (blocknumber is zero) and fill it if needed */
+       if ( !item[itempos] ) {
+           /* Ok, a hole. Now we need to check if we already prepared this
+              block to be journaled */
+           while ( !modifying_this_item ) { // loop until succeed
+               /* Well, this item is not journaled yet, so we must prepare
+                  it for journal first, before we can change it */
+               struct item_head tmp_ih; // We copy item head of found item,
+                                        // here to detect if fs changed under
+                                        // us while we were preparing for
+                                        // journal.
+               int fs_gen; // We store fs generation here to find if someone
+                           // changes fs under our feet
+
+               copy_item_head (&tmp_ih, ih); // Remember itemhead
+               fs_gen = get_generation (inode->i_sb); // remember fs generation
+               reiserfs_prepare_for_journal(inode->i_sb, bh, 1); // Prepare a buffer within which indirect item is stored for changing.
+               if (fs_changed (fs_gen, inode->i_sb) && item_moved (&tmp_ih, &path)) {
+                   // Sigh, fs was changed under us, we need to look for new
+                   // location of item we are working with
+
+                   /* unmark prepaerd area as journaled and search for it's
+                      new position */
+                   reiserfs_restore_prepared_buffer(inode->i_sb, bh);
+                   res = search_for_position_by_key(inode->i_sb, &key, &path);
+                   if ( res == IO_ERROR) {
+                       res = -EIO;
+                       goto error_exit_free_blocks;
+                   }
+                   bh=get_last_bh(&path);
+                   ih=get_ih(&path);
+                   item = get_item(&path);
+                   // Itempos is still the same
+                   continue;
+               }
+               modifying_this_item = 1;
+           }
+           item[itempos] = allocated_blocks[curr_block]; // Assign new block
+           curr_block++;
+       }
+       itempos++;
+    }
+
+    if ( modifying_this_item ) { // We need to log last-accessed block, if it
+                                // was modified, but not logged yet.
+       journal_mark_dirty (&th, inode->i_sb, bh);
+    }
+
+    if ( curr_block < blocks_to_allocate ) {
+       // Oh, well need to append to indirect item, or to create indirect item
+       // if there weren't any
+       if ( is_indirect_le_ih(ih) ) {
+           // Existing indirect item - append. First calculate key for append
+           // position. We do not need to recalculate path as it should
+           // already point to correct place.
+           make_cpu_key( &key, inode, le_key_k_offset( get_inode_item_key_version(inode), &(ih->ih_key)) + op_bytes_number(ih, inode->i_sb->s_blocksize), TYPE_INDIRECT, 3);
+           res = reiserfs_paste_into_item( &th, &path, &key, (char *)(allocated_blocks+curr_block), UNFM_P_SIZE*(blocks_to_allocate-curr_block));
+           if ( res ) {
+               goto error_exit_free_blocks;
+           }
+       } else if (is_statdata_le_ih(ih) ) {
+           // Last found item was statdata. That means we need to create indirect item.
+           struct item_head ins_ih; /* itemhead for new item */
+
+           /* create a key for our new item */
+           make_cpu_key( &key, inode, 1, TYPE_INDIRECT, 3); // Position one,
+                                                           // because that's
+                                                           // where first
+                                                           // indirect item
+                                                           // begins
+           /* Create new item head for our new item */
+           make_le_item_head (&ins_ih, &key, key.version, 1, TYPE_INDIRECT,
+                              (blocks_to_allocate-curr_block)*UNFM_P_SIZE,
+                              0 /* free space */);
+           /* Find where such item should live in the tree */
+           res = search_item (inode->i_sb, &key, &path);
+           if ( res != ITEM_NOT_FOUND ) {
+               /* Well, if we have found such item already, or some error
+                  occured, we need to warn user and return error */
+               if ( res != -ENOSPC ) {
+                   reiserfs_warning ("green-9009: search_by_key (%K) returned %d\n",
+                                     &key, res);
+               }
+               res = -EIO;
+               goto error_exit_free_blocks;
+           }
+           /* Insert item into the tree with the data as its body */
+           res = reiserfs_insert_item( &th, &path, &key, &ins_ih, (char *)(allocated_blocks+curr_block));
+       } else {
+           reiserfs_panic(inode->i_sb, "green-9010: unexpected item type for key %K\n",&key);
+       }
+    }
+
+    /* Now the final thing, if we have grew the file, we must update it's size*/
+    if ( pos + write_bytes > inode->i_size) {
+       inode->i_size = pos + write_bytes; // Set new size
+    }
+
+    /* Amount of on-disk blocks used by file have changed, update it */
+    inode->i_blocks += blocks_to_allocate * (inode->i_sb->s_blocksize / 512);
+    reiserfs_update_sd(&th, inode); // And update on-disk metadata
+    // finish all journal stuff now, We are not going to play with metadata
+    // anymore.
+    pathrelse(&path);
+    journal_end(&th, inode->i_sb, JOURNAL_PER_BALANCE_CNT * 3 + 1);
+    unlock_kernel();
+
+    // go through all the pages/buffers and map the buffers to newly allocated
+    // blocks (so that system knows where to write these pages later).
+    curr_block = 0;
+    for ( i = 0; i < num_pages ; i++ ) {
+       struct page *page=prepared_pages[i]; //current page
+       struct buffer_head *head = page->buffers; // first buffer for a page
+       int block_start, block_end; // in-page offsets for buffers.
+
+       if (!page->buffers)
+           reiserfs_panic(inode->i_sb, "green-9005: No buffers for prepared page???");
+
+       /* For each buffer in page */
+       for(bh = head, block_start = 0; bh != head || !block_start;
+           block_start=block_end, bh = bh->b_this_page) {
+           if (!bh)
+               reiserfs_panic(inode->i_sb, "green-9006: Allocated but absent buffer for a page?");
+           block_end = block_start+inode->i_sb->s_blocksize;
+           if (i == 0 && block_end <= from )
+               /* if this buffer is before requested data to map, skip it */
+               continue;
+           if (i == num_pages - 1 && block_start >= to)
+               /* If this buffer is after requested data to map, abort
+                  processing of current page */
+               break;
+
+           if ( !buffer_mapped(bh) ) { // Ok, unmapped buffer, need to map it
+               bh->b_dev = inode->i_dev;
+               bh->b_blocknr = le32_to_cpu(allocated_blocks[curr_block]);
+               set_bit(BH_Mapped, &bh->b_state);
+               curr_block++;
+           }
+       }
+    }
+
+    RFALSE( curr_block > blocks_to_allocate, "green-9007: Used too many blocks? weird");
+
+    return 0;
+
+// Need to deal with transaction here.
+error_exit_free_blocks:
+    pathrelse(&path);
+    // free blocks
+    for( i = 0; i < blocks_to_allocate; i++ )
+       reiserfs_free_block( &th, le32_to_cpu(allocated_blocks[i]));
+
+error_exit:
+    journal_end(&th, inode->i_sb, JOURNAL_PER_BALANCE_CNT * 3 + 1);
+    unlock_kernel();
+
+    return res;
+}
+
+/* Unlock pages prepared by reiserfs_prepare_file_region_for_write */
+void reiserfs_unprepare_pages(struct page **prepared_pages, /* list of locked pages */
+                             int num_pages /* amount of pages */) {
+    int i; // loop counter
+
+    for (i=0; i < num_pages ; i++) {
+       struct page *page = prepared_pages[i];
+
+       try_to_free_buffers(page,0);
+       kunmap(page);
+       UnlockPage(page);
+       page_cache_release(page);
+    }
+}
+
+/* This function will copy data from userspace to specified pages within
+   supplied byte range */
+int reiserfs_copy_from_user_to_file_region(
+                               loff_t pos, /* In-file position */
+                               int num_pages, /* Number of pages affected */
+                               int write_bytes, /* Amount of bytes to write */
+                               struct page **prepared_pages, /* pointer to
+                                                                array to
+                                                                prepared pages
+                                                               */
+                               const char *buf /* Pointer to user-supplied
+                                                  data*/
+                               )
+{
+    long page_fault=0; // status of copy_from_user.
+    int i; // loop counter.
+    int offset; // offset in page
+
+    for ( i = 0, offset = (pos & (PAGE_CACHE_SIZE-1)); i < num_pages ; i++,offset=0) {
+       int count = min_t(int,PAGE_CACHE_SIZE-offset,write_bytes); // How much of bytes to write to this page
+       struct page *page=prepared_pages[i]; // Current page we process.
+
+       /* Bring in the user page. We need to do it to fight some deadlocks
+          with copying the page to itself and page being inaccessible at
+          the same time.*/
+       { volatile unsigned char dummy;
+           __get_user(dummy, buf);
+           __get_user(dummy, buf+count-1);
+           /* We do getuser for beginning and ending of the region just because
+              userdata may be not page aligned, and therefore data that will
+              go into one page of file may be splitted onto two actual pages
+              in userspace */
+       }
+
+       /* Copy data from userspace to the current page */
+       page_fault = __copy_from_user(page_address(page)+offset, buf, count); // Copy the data.
+       /* Flush processor's dcache for this page */
+       flush_dcache_page(page);
+       buf+=count;
+       write_bytes-=count;
+
+       if (page_fault)
+           break; // Was there a fault? abort.
+    }
+
+    return page_fault?-EFAULT:0;
+}
+
+
+
+/* Submit pages for write. This was separated from actual file copying
+   because we might want to allocate block numbers in-between.
+   This function assumes that caller will adjust file size to correct value. */
+int reiserfs_submit_file_region_for_write(
+                               loff_t pos, /* Writing position offset */
+                               int num_pages, /* Number of pages to write */
+                               int write_bytes, /* number of bytes to write */
+                               struct page **prepared_pages /* list of pages */
+                               )
+{
+    int status; // return status of block_commit_write.
+    int retval = 0; // Return value we are going to return.
+    int i; // loop counter
+    int offset; // Writing offset in page.
+
+    for ( i = 0, offset = (pos & (PAGE_CACHE_SIZE-1)); i < num_pages ; i++,offset=0) {
+       int count = min_t(int,PAGE_CACHE_SIZE-offset,write_bytes); // How much of bytes to write to this page
+       struct page *page=prepared_pages[i]; // Current page we process.
+
+       status = block_commit_write(page, offset, offset+count); // Note this also kunmaps kmapped page.
+       if ( status )
+           retval = status; // To not overcomplicate matters We are going to
+                            // submit all the pages even if there was error.
+                            // we only remember error status to report it on
+                            // exit.
+       write_bytes-=count;
+       SetPageReferenced(page);
+       UnlockPage(page); // We unlock the page as it was locked by earlier call
+                         // to grab_cache_page
+       page_cache_release(page);
+    }
+    return retval;
+}
+
+/* Look if passed writing region is going to touch file's tail
+   (if it is present). And if it is, convert the tail to unformatted node */
+int reiserfs_check_for_tail_and_convert( struct inode *inode, /* inode to deal with */
+                                        loff_t pos, /* Writing position */
+                                        int write_bytes /* amount of bytes to write */
+                                       )
+{
+    INITIALIZE_PATH(path); // needed for search_for_position
+    struct cpu_key key; // Key that would represent last touched writing byte.
+    struct item_head *ih; // item header of found block;
+    int res; // Return value of various functions we call.
+    int cont_expand_offset; // We will put offset for generic_cont_expand here
+                           // This can be int just because tails are created
+                           // only for small files.
+
+/* this embodies a dependency on a particular tail policy */
+    if ( inode->i_size >= inode->i_sb->s_blocksize*4 ) {
+       /* such a big files do not have tails, so we won't bother ourselves
+          to look for tails, simply return */
+       return 0;
+    }
+
+    lock_kernel();
+    /* find the item containing the last byte to be written, or if
+     * writing past the end of the file then the last item of the
+     * file (and then we check its type). */
+    make_cpu_key (&key, inode, pos+write_bytes+1, TYPE_ANY, 3/*key length*/);
+    res = search_for_position_by_key(inode->i_sb, &key, &path);
+    if ( res == IO_ERROR ) {
+        unlock_kernel();
+       return -EIO;
+    }
+    ih = get_ih(&path);
+    res = 0;
+    if ( is_direct_le_ih(ih) ) {
+       /* Ok, closest item is file tail (tails are stored in "direct"
+        * items), so we need to unpack it. */
+       /* To not overcomplicate matters, we just call generic_cont_expand
+          which will in turn call other stuff and finally will boil down to
+           reiserfs_get_block() that would do necessary conversion. */
+       cont_expand_offset = le_key_k_offset(get_inode_item_key_version(inode), &(ih->ih_key));
+       pathrelse(&path);
+       res = generic_cont_expand( inode, cont_expand_offset);
+    } else
+       pathrelse(&path);
+
+    unlock_kernel();
+    return res;
+}
+
+/* This functions kmaps and locks pages starting from @pos for @inode.
+   @num_pages pages are locked and stored in
+   @prepared_pages array. Also buffers are allocated for these pages.
+   First and last page of the region is read if it is overwritten only
+   partially. If last page did not exist before write (file hole or file
+   append), it is zeroed, then.
+   Returns number of unallocated blocks that should be allocated to cover
+   new file data.*/
+int reiserfs_prepare_file_region_for_write(
+                               struct inode *inode /* Inode of the file */,
+                               loff_t pos, /* position in the file */
+                               int num_pages, /* number of pages to
+                                                 prepare */
+                               int write_bytes, /* Amount of bytes to be
+                                                   overwritten from
+                                                   @pos */
+                               struct page **prepared_pages /* pointer to array
+                                                              where to store
+                                                              prepared pages */
+                                          )
+{
+    int res=0; // Return values of different functions we call.
+    unsigned long index = pos >> PAGE_CACHE_SHIFT; // Offset in file in pages.
+    int from = (pos & (PAGE_CACHE_SIZE - 1)); // Writing offset in first page
+    int to = ((pos + write_bytes - 1) & (PAGE_CACHE_SIZE - 1)) + 1;
+                                        /* offset of last modified byte in last
+                                           page */
+    struct address_space *mapping = inode->i_mapping; // Pages are mapped here.
+    int i; // Simple counter
+    int blocks = 0; /* Return value (blocks that should be allocated) */
+    struct buffer_head *bh, *head; // Current bufferhead and first bufferhead
+                                  // of a page.
+    unsigned block_start, block_end; // Starting and ending offsets of current
+                                    // buffer in the page.
+    struct buffer_head *wait[2], **wait_bh=wait; // Buffers for page, if
+                                                // Page appeared to be not up
+                                                // to date. Note how we have
+                                                // at most 2 buffers, this is
+                                                // because we at most may
+                                                // partially overwrite two
+                                                // buffers for one page. One at                                                 // the beginning of write area
+                                                // and one at the end.
+                                                // Everything inthe middle gets                                                 // overwritten totally.
+
+    struct cpu_key key; // cpu key of item that we are going to deal with
+    struct item_head *ih = NULL; // pointer to item head that we are going to deal with
+    struct buffer_head *itembuf=NULL; // Buffer head that contains items that we are going to deal with
+    INITIALIZE_PATH(path); // path to item, that we are going to deal with.
+    __u32 * item=0; // pointer to item we are going to deal with
+
+
+    if ( num_pages < 1 ) {
+       reiserfs_warning("green-9001: reiserfs_prepare_file_region_for_write called with zero number of pages to process\n");
+       return -EFAULT;
+    }
+
+    /* We have 2 loops for pages. In first loop we grab and lock the pages, so
+       that nobody would touch these until we release the pages. Then
+       we'd start to deal with mapping buffers to blocks. */
+    for ( i = 0; i < num_pages; i++) {
+       prepared_pages[i] = grab_cache_page(mapping, index + i); // locks the page
+       if ( !prepared_pages[i]) {
+           res = -ENOMEM;
+           goto failed_page_grabbing;
+       }
+       kmap(prepared_pages[i]); // MAp the page in conventional RAM.
+       if (!prepared_pages[i]->buffers)
+           create_empty_buffers(prepared_pages[i], inode->i_dev, inode->i_sb->s_blocksize);
+    }
+
+    /* Let's count amount of blocks for a case where all the blocks
+       overwritten are new (we will substract already allocated blocks later)*/
+    if ( num_pages > 2 )
+       /* These are full-overwritten pages so we count all the blocks in
+          these pages are counted as needed to be allocated */
+       blocks = (PAGE_CACHE_SIZE/inode->i_sb->s_blocksize) * (num_pages - 2);
+
+    /* count blocks needed for first page (possibly partially written) */
+    blocks += ((PAGE_CACHE_SIZE - from) >> inode->i_sb->s_blocksize_bits) +
+          !!(from & (inode->i_sb->s_blocksize-1)); /* roundup */
+
+    /* Now we account for last page. If last page == first page (we
+       overwrite only one page), we substract all the blocks past the
+       last writing position in a page out of already calculated number
+       of blocks */
+    blocks += (PAGE_CACHE_SIZE/inode->i_sb->s_blocksize) * (num_pages > 1) -
+          ((PAGE_CACHE_SIZE - to) >> inode->i_sb->s_blocksize_bits);
+          /* Note how we do not roundup here since partial blocks still
+                  should be allocated */
+
+    /* Now if all the write area lies past the file end, no point in
+       maping blocks, since there is none, so we just zero out remaining
+       parts of first and last pages in write area (if needed) */
+    if ( (pos & ~(PAGE_CACHE_SIZE - 1)) > inode->i_size ) {
+       if ( from != 0 ) {/* First page needs to be partially zeroed */
+           memset(page_address(prepared_pages[0]), 0, from);
+           SetPageUptodate(prepared_pages[0]);
+       }
+       if ( to != PAGE_CACHE_SIZE ) { /* Last page needs to be partially zeroed */
+           memset(page_address(prepared_pages[num_pages-1])+to, 0, PAGE_CACHE_SIZE - to);
+           SetPageUptodate(prepared_pages[num_pages-1]);
+       }
+
+       /* Since all blocks are new - use already calculated value */
+       return blocks;
+    }
+
+    /* Well, since we write somewhere into the middle of a file, there is
+       possibility we are writing over some already allocated blocks, so
+       let's map these blocks and substract number of such blocks out of blocks
+       we need to allocate (calculated above) */
+    /* Mask write position to start on blocksize, we do it out of the
+       loop for performance reasons */
+    pos &= ~(inode->i_sb->s_blocksize - 1);
+    /* Set cpu key to the starting position in a file (on left block boundary)*/
+    make_cpu_key (&key, inode, 1 + ((pos) & ~(inode->i_sb->s_blocksize - 1)), TYPE_ANY, 3/*key length*/);
+
+    lock_kernel(); // We need that for at least search_by_key()
+    for ( i = 0; i < num_pages ; i++ ) {
+       int item_pos=-1; /* Position in indirect item */
+
+       head = prepared_pages[i]->buffers;
+       /* For each buffer in the page */
+       for(bh = head, block_start = 0; bh != head || !block_start;
+           block_start=block_end, bh = bh->b_this_page) {
+               if (!bh)
+                   reiserfs_panic(inode->i_sb, "green-9002: Allocated but absent buffer for a page?");
+               /* Find where this buffer ends */
+               block_end = block_start+inode->i_sb->s_blocksize;
+               if (i == 0 && block_end <= from )
+                   /* if this buffer is before requested data to map, skip it*/
+                   continue;
+
+               if (i == num_pages - 1 && block_start >= to) {
+                   /* If this buffer is after requested data to map, abort
+                      processing of current page */
+                   break;
+               }
+
+               if ( buffer_mapped(bh) && bh->b_blocknr !=0 ) {
+                   /* This is optimisation for a case where buffer is mapped
+                      and have blocknumber assigned. In case significant amount
+                      of such buffers are present, we may avoid some amount
+                      of search_by_key calls.
+                      Probably it would be possible to move parts of this code
+                      out of BKL, but I afraid that would overcomplicate code
+                      without any noticeable benefit.
+                   */
+                   item_pos++;
+                   /* Update the key */
+                   set_cpu_key_k_offset( &key, cpu_key_k_offset(&key) + inode->i_sb->s_blocksize);
+                   blocks--; // Decrease the amount of blocks that need to be
+                             // allocated
+                   continue; // Go to the next buffer
+               }
+
+               if ( !itembuf || /* if first iteration */
+                    item_pos >= ih_item_len(ih)/UNFM_P_SIZE)
+                                            { /* or if we progressed past the
+                                                 current unformatted_item */
+                       /* Try to find next item */
+                       res = search_for_position_by_key(inode->i_sb, &key, &path);
+                       /* Abort if no more items */
+                       if ( res != POSITION_FOUND )
+                           break;
+
+                       /* Update information about current indirect item */
+                       itembuf = get_last_bh( &path );
+                       ih = get_ih( &path );
+                       item = get_item( &path );
+                       item_pos = path.pos_in_item;
+
+                       RFALSE( !is_indirect_le_ih (ih), "green-9003: indirect item expected");
+               }
+
+               /* See if there is some block associated with the file
+                  at that position, map the buffer to this block */
+               if ( get_block_num(item,item_pos) ) {
+                   bh->b_dev = inode->i_dev;
+                   bh->b_blocknr = get_block_num(item,item_pos);
+                   set_bit(BH_Mapped, &bh->b_state);
+                   blocks--; // Decrease the amount of blocks that need to be
+                             // allocated
+               }
+               item_pos++;
+               /* Update the key */
+               set_cpu_key_k_offset( &key, cpu_key_k_offset(&key) + inode->i_sb->s_blocksize);
+       }
+    }
+    pathrelse(&path); // Free the path
+    unlock_kernel();
+
+       /* Now zero out unmappend buffers for the first and last pages of
+          write area or issue read requests if page is mapped. */
+       /* First page, see if it is not uptodate */
+       if ( !Page_Uptodate(prepared_pages[0]) ) {
+           head = prepared_pages[0]->buffers;
+
+           /* For wach buffer in page */
+           for(bh = head, block_start = 0; bh != head || !block_start;
+               block_start=block_end, bh = bh->b_this_page) {
+
+               if (!bh)
+                   reiserfs_panic(inode->i_sb, "green-9002: Allocated but absent buffer for a page?");
+               /* Find where this buffer ends */
+               block_end = block_start+inode->i_sb->s_blocksize;
+               if ( block_end <= from )
+                   /* if this buffer is before requested data to map, skip it*/
+                   continue;
+               if ( block_start < from ) { /* Aha, our partial buffer */
+                   if ( buffer_mapped(bh) ) { /* If it is mapped, we need to
+                                                 issue READ request for it to
+                                                 not loose data */
+                       ll_rw_block(READ, 1, &bh);
+                       *wait_bh++=bh;
+                   } else { /* Not mapped, zero it */
+                       memset(page_address(prepared_pages[0])+block_start, 0, from-block_start);
+                       set_bit(BH_Uptodate, &bh->b_state);
+                   }
+               }
+           }
+       }
+
+       /* Last page, see if it is not uptodate, or if the last page is past the end of the file. */
+       if ( !Page_Uptodate(prepared_pages[num_pages-1]) ||
+           ((pos+write_bytes)>>PAGE_CACHE_SHIFT) > (inode->i_size>>PAGE_CACHE_SHIFT) ) {
+           head = prepared_pages[num_pages-1]->buffers;
+
+           /* for each buffer in page */
+           for(bh = head, block_start = 0; bh != head || !block_start;
+               block_start=block_end, bh = bh->b_this_page) {
+
+               if (!bh)
+                   reiserfs_panic(inode->i_sb, "green-9002: Allocated but absent buffer for a page?");
+               /* Find where this buffer ends */
+               block_end = block_start+inode->i_sb->s_blocksize;
+               if ( block_start >= to )
+                   /* if this buffer is after requested data to map, skip it*/
+                   break;
+               if ( block_end > to ) { /* Aha, our partial buffer */
+                   if ( buffer_mapped(bh) ) { /* If it is mapped, we need to
+                                                 issue READ request for it to
+                                                 not loose data */
+                       ll_rw_block(READ, 1, &bh);
+                       *wait_bh++=bh;
+                   } else { /* Not mapped, zero it */
+                       memset(page_address(prepared_pages[num_pages-1])+to, 0, block_end-to);
+                       set_bit(BH_Uptodate, &bh->b_state);
+                   }
+               }
+           }
+       }
+
+    /* Wait for read requests we made to happen, if necessary */
+    while(wait_bh > wait) {
+       wait_on_buffer(*--wait_bh);
+       if (!buffer_uptodate(*wait_bh)) {
+           res = -EIO;
+           goto failed_read;
+       }
+    }
+
+    return blocks;
+failed_page_grabbing:
+    num_pages = i;
+failed_read:
+    reiserfs_unprepare_pages(prepared_pages, num_pages);
+    return res;
+}
+
+/* Write @count bytes at position @ppos in a file indicated by @file
+   from the buffer @buf.
+
+   generic_file_write() is only appropriate for filesystems that are not seeking to optimize performance and want
+   something simple that works.  It is not for serious use by general purpose filesystems, excepting the one that it was
+   written for (ext2/3).  This is for several reasons:
+
+   * It has no understanding of any filesystem specific optimizations.
+
+   * It enters the filesystem repeatedly for each page that is written.
+
+   * It depends on reiserfs_get_block() function which if implemented by reiserfs performs costly search_by_key
+   * operation for each page it is supplied with. By contrast reiserfs_file_write() feeds as much as possible at a time
+   * to reiserfs which allows for fewer tree traversals.
+
+   * Each indirect pointer insertion takes a lot of cpu, because it involves memory moves inside of blocks.
+
+   * Asking the block allocation code for blocks one at a time is slightly less efficient.
+
+   All of these reasons for not using only generic file write were understood back when reiserfs was first miscoded to
+   use it, but we were in a hurry to make code freeze, and so it couldn't be revised then.  This new code should make
+   things right finally.
+
+   Future Features: providing search_by_key with hints.
+
+*/
+ssize_t reiserfs_file_write( struct file *file, /* the file we are going to write into */
+                             const char *buf, /*  pointer to user supplied data
+(in userspace) */
+                             size_t count, /* amount of bytes to write */
+                             loff_t *ppos /* pointer to position in file that we start writing at. Should be updated to
+                                           * new current position before returning. */ )
+{
+    size_t already_written = 0; // Number of bytes already written to the file.
+    loff_t pos; // Current position in the file.
+    size_t res; // return value of various functions that we call.
+    unsigned long limit = current->rlim[RLIMIT_FSIZE].rlim_cur; // Current filesize limit
+    struct inode *inode = file->f_dentry->d_inode; // Inode of the file that we are writing to.
+    struct page * prepared_pages[REISERFS_WRITE_PAGES_AT_A_TIME];
+                               /* To simplify coding at this time, we store
+                                  locked pages in array for now */
+    if ( count <= PAGE_CACHE_SIZE || file->f_flags & O_DIRECT)
+        return generic_file_write(file, buf, count, ppos);
+
+    if ( (ssize_t) count < 0 )
+        return -EINVAL;
+
+    if (!access_ok(VERIFY_READ, buf, count))
+        return -EFAULT;
+
+    down(&inode->i_sem); // locks the entire file for just us
+
+    // We are going to duplicate a lot of generic_file_write checks here
+    // for now. I do not have any good idea on how to avoid these now.
+
+    res = -EINVAL;
+    if ( *ppos < 0)
+       goto out;
+
+    res = file->f_error;
+    if ( res ) {
+       file->f_error = 0;
+       goto out;
+    }
+
+    if (file->f_flags & O_APPEND)
+       pos=inode->i_size;
+    else
+       pos=*ppos;
+
+    res = -EFBIG;
+    if ( limit != RLIM_INFINITY) {
+       if (pos >= limit) {
+           send_sig(SIGXFSZ, current, 0);
+           goto out;
+       }
+       if ( pos > 0xFFFFFFFFULL || count > limit - (u32)pos) {
+           count = limit - (u32)pos;
+       }
+    }
+
+    // LFS
+    if ( pos + count > MAX_NON_LFS && !(file->f_flags & O_LARGEFILE) ) {
+       if ( pos >= MAX_NON_LFS ) {
+           send_sig(SIGXFSZ, current, 0);
+           goto out;
+       }
+       if ( count > MAX_NON_LFS - (u32)pos ) {
+           count = MAX_NON_LFS - (u32)pos;
+       }
+    }
+
+    // Check we are not going to exceed block limit
+    if ( pos >= inode->i_sb->s_maxbytes) {
+       if ( count || pos > inode->i_sb->s_maxbytes) {
+           send_sig(SIGXFSZ, current, 0);
+           goto out;
+       }
+       // zero length writes are ok.
+    }
+
+    res = 0;
+    if ( count == 0 )
+       goto out;
+
+    remove_suid(inode);
+    /* Mark inode dirty only in case times are actually changed.
+       This is to speed things up */
+    {
+       time_t now = CURRENT_TIME;
+       if (inode->i_ctime != now || inode->i_mtime != now) {
+           inode->i_ctime = inode->i_mtime = now;
+           mark_inode_dirty_sync(inode);
+       }
+    }
+
+
+    // Ok, we are done with all the checks.
+
+    // Now we should start real work
+
+    /* If we are going to write past the file's packed tail or if we are going
+       to overwrite part of the tail, we need that tail to be converted into
+       unformatted node */
+    res = reiserfs_check_for_tail_and_convert( inode, pos, count);
+    if (res)
+       goto out;
+
+    while ( count > 0) {
+       /* This is the main loop in which we running until some error occures
+          or until we write all of the data. */
+       int num_pages;/* amount of pages we are going to write this iteration */
+       int write_bytes; /* amount of bytes to write during this iteration */
+       int blocks_to_allocate; /* how much blocks we need to allocate for
+                                  this iteration */
+
+        /*  (pos & (PAGE_CACHE_SIZE-1)) is an idiom for offset into a page of pos*/
+       num_pages = !!((pos+count) & (PAGE_CACHE_SIZE - 1)) + /* round up partial
+                                                         pages */
+                   ((count + (pos & (PAGE_CACHE_SIZE-1))) >> PAGE_CACHE_SHIFT);
+                                               /* convert size to amount of
+                                                  pages */
+/* all lock_kernels should be changed to lock_supers in reiserfs, but they need to be changed all at once and all together */
+       lock_kernel();
+       if ( num_pages > REISERFS_WRITE_PAGES_AT_A_TIME
+               || num_pages > reiserfs_can_fit_pages(inode->i_sb) ) {
+           /* If we were asked to write more data than we want to or if there
+              is not that much space, then we shorten amount of data to write
+              for this iteration. */
+           num_pages = min_t(int, REISERFS_WRITE_PAGES_AT_A_TIME, reiserfs_can_fit_pages(inode->i_sb));
+           /* Also we should not forget to set size in bytes accordingly */
+           write_bytes = num_pages * PAGE_CACHE_SIZE -
+                           (pos & (PAGE_CACHE_SIZE-1));
+                                        /* If position is not on the
+                                           start of the page, we need
+                                           to substract the offset
+                                           within page */
+       } else
+           write_bytes = count;
+
+       /* reserve the blocks to be allocated later, so that later on
+          we still have the space to write the blocks to */
+       reiserfs_claim_blocks_to_be_allocated(inode->i_sb, num_pages * (PAGE_CACHE_SIZE/inode->i_sb->s_blocksize));
+       unlock_kernel();
+
+       if ( !num_pages ) { /* If we do not have enough space even for */
+           res = -ENOSPC;  /* single page, return -ENOSPC */
+           if ( pos > (inode->i_size & (inode->i_sb->s_blocksize-1)))
+               break; // In case we are writing past the file end, break.
+           // Otherwise we are possibly overwriting the file, so
+           // let's set write size to be equal or less than blocksize.
+           // This way we get it correctly for file holes.
+           // But overwriting files on absolutelly full volumes would not
+           // be very efficient. Well, people are not supposed to fill
+           // 100% of disk space anyway.
+           write_bytes = min_t(int, count, inode->i_sb->s_blocksize - (pos & (inode->i_sb->s_blocksize - 1)));
+           num_pages = 1;
+       }
+
+       /* Prepare for writing into the region, read in all the
+          partially overwritten pages, if needed. And lock the pages,
+          so that nobody else can access these until we are done.
+          We get number of actual blocks needed as a result.*/
+       blocks_to_allocate = reiserfs_prepare_file_region_for_write(inode, pos, num_pages, write_bytes, prepared_pages);
+       if ( blocks_to_allocate < 0 ) {
+           res = blocks_to_allocate;
+           reiserfs_release_claimed_blocks(inode->i_sb, num_pages * (PAGE_CACHE_SIZE/inode->i_sb->s_blocksize));
+           break;
+       }
+
+       /* First we correct our estimate of how many blocks we need */
+       reiserfs_release_claimed_blocks(inode->i_sb, num_pages * (PAGE_CACHE_SIZE>>inode->i_sb->s_blocksize_bits) - blocks_to_allocate );
+
+       if ( blocks_to_allocate > 0) {/*We only allocate blocks if we need to*/
+           /* Fill in all the possible holes and append the file if needed */
+           res = reiserfs_allocate_blocks_for_region(inode, pos, num_pages, write_bytes, prepared_pages, blocks_to_allocate);
+       } else if ( pos + write_bytes > inode->i_size ) {
+           /* File might have grown even though no new blocks were added */
+           inode->i_size = pos + write_bytes;
+           inode->i_sb->s_op->dirty_inode(inode);
+       }
+
+       /* well, we have allocated the blocks, so it is time to free
+          the reservation we made earlier. */
+       reiserfs_release_claimed_blocks(inode->i_sb, blocks_to_allocate);
+       if ( res ) {
+           reiserfs_unprepare_pages(prepared_pages, num_pages);
+           break;
+       }
+
+/* NOTE that allocating blocks and filling blocks can be done in reverse order
+   and probably we would do that just to get rid of garbage in files after a
+   crash */
+
+       /* Copy data from user-supplied buffer to file's pages */
+       res = reiserfs_copy_from_user_to_file_region(pos, num_pages, write_bytes, prepared_pages, buf);
+       if ( res ) {
+           reiserfs_unprepare_pages(prepared_pages, num_pages);
+           break;
+       }
+
+       /* Send the pages to disk and unlock them. */
+       res = reiserfs_submit_file_region_for_write(pos, num_pages, write_bytes, prepared_pages);
+       if ( res )
+           break;
+
+       already_written += write_bytes;
+       buf += write_bytes;
+       *ppos = pos += write_bytes;
+       count -= write_bytes;
+    }
+
+    if ((file->f_flags & O_SYNC) || IS_SYNC(inode))
+       res = generic_osync_inode(inode, OSYNC_METADATA|OSYNC_DATA);
+
+    up(&inode->i_sem);
+    return (already_written != 0)?already_written:res;
+
+out:
+    up(&inode->i_sem); // unlock the file on exit.
+    return res;
+}
struct file_operations reiserfs_file_operations = {
    read:      generic_file_read,
-    write:     generic_file_write,
+    write:     reiserfs_file_write,
    ioctl:     reiserfs_ioctl,
    mmap:      generic_file_mmap,
    release:   reiserfs_file_release,
diff -Nru a/fs/reiserfs/inode.c b/fs/reiserfs/inode.c
--- a/fs/reiserfs/inode.c       Mon Sep  9 09:28:34 2002
+++ b/fs/reiserfs/inode.c       Mon Sep  9 09:28:34 2002
@@ -913,7 +913,7 @@


    copy_key (INODE_PKEY (inode), &(ih->ih_key));
-    inode->i_blksize = PAGE_SIZE;
+    inode->i_blksize = PAGE_SIZE*32;

    INIT_LIST_HEAD(&inode->u.reiserfs_i.i_prealloc_list) ;

@@ -1605,7 +1605,7 @@

    // these do not go to on-disk stat data
    inode->i_ino = le32_to_cpu (ih.ih_key.k_objectid);
-    inode->i_blksize = PAGE_SIZE;
+    inode->i_blksize = PAGE_SIZE*32;
    inode->i_dev = sb->s_dev;

    // store in in-core inode the key of stat data and version all
diff -Nru a/fs/reiserfs/super.c b/fs/reiserfs/super.c
--- a/fs/reiserfs/super.c       Mon Sep  9 09:28:34 2002
+++ b/fs/reiserfs/super.c       Mon Sep  9 09:28:34 2002
@@ -1261,6 +1261,7 @@
    reiserfs_proc_register( s, "oidmap", reiserfs_oidmap_in_proc );
    reiserfs_proc_register( s, "journal", reiserfs_journal_in_proc );
    init_waitqueue_head (&(s->u.reiserfs_sb.s_wait));
+    s->u.reiserfs_sb.bitmap_lock = SPIN_LOCK_UNLOCKED;

    printk("%s\n", reiserfs_get_version_string()) ;
    return s;
diff -Nru a/include/linux/reiserfs_fs.h b/include/linux/reiserfs_fs.h
--- a/include/linux/reiserfs_fs.h       Mon Sep  9 09:28:34 2002
+++ b/include/linux/reiserfs_fs.h       Mon Sep  9 09:28:34 2002
@@ -2007,6 +2007,7 @@
#endif
void reiserfs_claim_blocks_to_be_allocated( struct super_block *sb, int blocks);
void reiserfs_release_claimed_blocks( struct super_block *sb, int blocks);
+int reiserfs_can_fit_pages(struct super_block *sb);

/* hashes.c */
__u32 keyed_hash (const signed char *msg, int len);
diff -Nru a/include/linux/reiserfs_fs_sb.h b/include/linux/reiserfs_fs_sb.h
--- a/include/linux/reiserfs_fs_sb.h    Mon Sep  9 09:28:34 2002
+++ b/include/linux/reiserfs_fs_sb.h    Mon Sep  9 09:28:34 2002
@@ -466,6 +466,7 @@
    reiserfs_proc_info_data_t s_proc_info_data;
    struct proc_dir_entry *procdir;
    int reserved_blocks; /* amount of blocks reserved for further allocations */
+    spinlock_t bitmap_lock; /* this lock on now only used to protect reserved_blocks variable */
};

/* Definitions of reiserfs on-disk properties: */