]> git.ipfire.org Git - people/pmueller/ipfire-2.x.git/blobdiff - src/patches/suse-2.6.27.31/patches.suse/ocfs2-Improve-ocfs2_read_xattr_bucket.patch
Reenabled linux-xen, added patches for Xen Kernel Version 2.6.27.31,
[people/pmueller/ipfire-2.x.git] / src / patches / suse-2.6.27.31 / patches.suse / ocfs2-Improve-ocfs2_read_xattr_bucket.patch
diff --git a/src/patches/suse-2.6.27.31/patches.suse/ocfs2-Improve-ocfs2_read_xattr_bucket.patch b/src/patches/suse-2.6.27.31/patches.suse/ocfs2-Improve-ocfs2_read_xattr_bucket.patch
new file mode 100644 (file)
index 0000000..57fa345
--- /dev/null
@@ -0,0 +1,322 @@
+From: Joel Becker <joel.becker@oracle.com>
+Date: Fri, 24 Oct 2008 17:33:40 -0700
+Subject: ocfs2: Improve ocfs2_read_xattr_bucket().
+Patch-mainline: 2.6.29
+
+The ocfs2_read_xattr_bucket() function would read an xattr bucket into a
+list of buffer heads.  However, we have a nice ocfs2_xattr_bucket
+structure.  Let's have it fill that out instead.
+
+In addition, ocfs2_read_xattr_bucket() would initialize buffer heads for
+a bucket that's never been on disk before.  That's confusing.  Let's
+call that functionality ocfs2_init_xattr_bucket().
+
+The functions ocfs2_cp_xattr_bucket() and ocfs2_half_xattr_bucket() are
+updated to use the ocfs2_xattr_bucket structure rather than raw bh
+lists.  That way they can use the new read/init calls.  In addition,
+they drop the wasted read of an existing target bucket.
+
+Signed-off-by: Joel Becker <joel.becker@oracle.com>
+Signed-off-by: Mark Fasheh <mfasheh@suse.com>
+---
+ fs/ocfs2/xattr.c |  165 ++++++++++++++++++++++++++----------------------------
+ 1 files changed, 79 insertions(+), 86 deletions(-)
+
+diff --git a/fs/ocfs2/xattr.c b/fs/ocfs2/xattr.c
+index 3478ad1..fa13fa4 100644
+--- a/fs/ocfs2/xattr.c
++++ b/fs/ocfs2/xattr.c
+@@ -168,6 +168,48 @@ static void ocfs2_xattr_bucket_relse(struct inode *inode,
+       }
+ }
++/*
++ * A bucket that has never been written to disk doesn't need to be
++ * read.  We just need the buffer_heads.  Don't call this for
++ * buckets that are already on disk.  ocfs2_read_xattr_bucket() initializes
++ * them fully.
++ */
++static int ocfs2_init_xattr_bucket(struct inode *inode,
++                                 struct ocfs2_xattr_bucket *bucket,
++                                 u64 xb_blkno)
++{
++      int i, rc = 0;
++      int blks = ocfs2_blocks_per_xattr_bucket(inode->i_sb);
++
++      for (i = 0; i < blks; i++) {
++              bucket->bu_bhs[i] = sb_getblk(inode->i_sb, xb_blkno + i);
++              if (!bucket->bu_bhs[i]) {
++                      rc = -EIO;
++                      mlog_errno(rc);
++                      break;
++              }
++
++              ocfs2_set_new_buffer_uptodate(inode, bucket->bu_bhs[i]);
++      }
++
++      if (rc)
++              ocfs2_xattr_bucket_relse(inode, bucket);
++      return rc;
++}
++
++/* Read the xattr bucket at xb_blkno */
++static int ocfs2_read_xattr_bucket(struct inode *inode,
++                                 struct ocfs2_xattr_bucket *bucket,
++                                 u64 xb_blkno)
++{
++      int rc, blks = ocfs2_blocks_per_xattr_bucket(inode->i_sb);
++
++      rc = ocfs2_read_blocks(inode, xb_blkno, blks, bucket->bu_bhs, 0);
++      if (rc)
++              ocfs2_xattr_bucket_relse(inode, bucket);
++      return rc;
++}
++
+ static inline const char *ocfs2_xattr_prefix(int name_index)
+ {
+       struct xattr_handler *handler = NULL;
+@@ -3097,31 +3139,6 @@ out:
+       return ret;
+ }
+-static int ocfs2_read_xattr_bucket(struct inode *inode,
+-                                 u64 blkno,
+-                                 struct buffer_head **bhs,
+-                                 int new)
+-{
+-      int ret = 0;
+-      u16 i, blk_per_bucket = ocfs2_blocks_per_xattr_bucket(inode->i_sb);
+-
+-      if (!new)
+-              return ocfs2_read_blocks(inode, blkno,
+-                                       blk_per_bucket, bhs, 0);
+-
+-      for (i = 0; i < blk_per_bucket; i++) {
+-              bhs[i] = sb_getblk(inode->i_sb, blkno + i);
+-              if (bhs[i] == NULL) {
+-                      ret = -EIO;
+-                      mlog_errno(ret);
+-                      break;
+-              }
+-              ocfs2_set_new_buffer_uptodate(inode, bhs[i]);
+-      }
+-
+-      return ret;
+-}
+-
+ /*
+  * Find the suitable pos when we divide a bucket into 2.
+  * We have to make sure the xattrs with the same hash value exist
+@@ -3184,7 +3201,7 @@ static int ocfs2_divide_xattr_bucket(struct inode *inode,
+       int ret, i;
+       int count, start, len, name_value_len = 0, xe_len, name_offset = 0;
+       u16 blk_per_bucket = ocfs2_blocks_per_xattr_bucket(inode->i_sb);
+-      struct buffer_head **s_bhs, **t_bhs = NULL;
++      struct ocfs2_xattr_bucket s_bucket, t_bucket;
+       struct ocfs2_xattr_header *xh;
+       struct ocfs2_xattr_entry *xe;
+       int blocksize = inode->i_sb->s_blocksize;
+@@ -3192,37 +3209,34 @@ static int ocfs2_divide_xattr_bucket(struct inode *inode,
+       mlog(0, "move some of xattrs from bucket %llu to %llu\n",
+            (unsigned long long)blk, (unsigned long long)new_blk);
+-      s_bhs = kcalloc(blk_per_bucket, sizeof(struct buffer_head *), GFP_NOFS);
+-      if (!s_bhs)
+-              return -ENOMEM;
++      memset(&s_bucket, 0, sizeof(struct ocfs2_xattr_bucket));
++      memset(&t_bucket, 0, sizeof(struct ocfs2_xattr_bucket));
+-      ret = ocfs2_read_xattr_bucket(inode, blk, s_bhs, 0);
++      ret = ocfs2_read_xattr_bucket(inode, &s_bucket, blk);
+       if (ret) {
+               mlog_errno(ret);
+               goto out;
+       }
+-      ret = ocfs2_journal_access(handle, inode, s_bhs[0],
++      ret = ocfs2_journal_access(handle, inode, s_bucket.bu_bhs[0],
+                                  OCFS2_JOURNAL_ACCESS_WRITE);
+       if (ret) {
+               mlog_errno(ret);
+               goto out;
+       }
+-      t_bhs = kcalloc(blk_per_bucket, sizeof(struct buffer_head *), GFP_NOFS);
+-      if (!t_bhs) {
+-              ret = -ENOMEM;
+-              goto out;
+-      }
+-
+-      ret = ocfs2_read_xattr_bucket(inode, new_blk, t_bhs, new_bucket_head);
++      /*
++       * Even if !new_bucket_head, we're overwriting t_bucket.  Thus,
++       * there's no need to read it.
++       */
++      ret = ocfs2_init_xattr_bucket(inode, &t_bucket, new_blk);
+       if (ret) {
+               mlog_errno(ret);
+               goto out;
+       }
+       for (i = 0; i < blk_per_bucket; i++) {
+-              ret = ocfs2_journal_access(handle, inode, t_bhs[i],
++              ret = ocfs2_journal_access(handle, inode, t_bucket.bu_bhs[i],
+                                          new_bucket_head ?
+                                          OCFS2_JOURNAL_ACCESS_CREATE :
+                                          OCFS2_JOURNAL_ACCESS_WRITE);
+@@ -3232,7 +3246,7 @@ static int ocfs2_divide_xattr_bucket(struct inode *inode,
+               }
+       }
+-      xh = (struct ocfs2_xattr_header *)s_bhs[0]->b_data;
++      xh = bucket_xh(&s_bucket);
+       count = le16_to_cpu(xh->xh_count);
+       start = ocfs2_xattr_find_divide_pos(xh);
+@@ -3245,9 +3259,9 @@ static int ocfs2_divide_xattr_bucket(struct inode *inode,
+                * that of the last entry in the previous bucket.
+                */
+               for (i = 0; i < blk_per_bucket; i++)
+-                      memset(t_bhs[i]->b_data, 0, blocksize);
++                      memset(bucket_block(&t_bucket, i), 0, blocksize);
+-              xh = (struct ocfs2_xattr_header *)t_bhs[0]->b_data;
++              xh = bucket_xh(&t_bucket);
+               xh->xh_free_start = cpu_to_le16(blocksize);
+               xh->xh_entries[0].xe_name_hash = xe->xe_name_hash;
+               le32_add_cpu(&xh->xh_entries[0].xe_name_hash, 1);
+@@ -3257,10 +3271,11 @@ static int ocfs2_divide_xattr_bucket(struct inode *inode,
+       /* copy the whole bucket to the new first. */
+       for (i = 0; i < blk_per_bucket; i++)
+-              memcpy(t_bhs[i]->b_data, s_bhs[i]->b_data, blocksize);
++              memcpy(bucket_block(&t_bucket, i), bucket_block(&s_bucket, i),
++                     blocksize);
+       /* update the new bucket. */
+-      xh = (struct ocfs2_xattr_header *)t_bhs[0]->b_data;
++      xh = bucket_xh(&t_bucket);
+       /*
+        * Calculate the total name/value len and xh_free_start for
+@@ -3325,7 +3340,7 @@ set_num_buckets:
+               xh->xh_num_buckets = 0;
+       for (i = 0; i < blk_per_bucket; i++) {
+-              ocfs2_journal_dirty(handle, t_bhs[i]);
++              ocfs2_journal_dirty(handle, t_bucket.bu_bhs[i]);
+               if (ret)
+                       mlog_errno(ret);
+       }
+@@ -3342,29 +3357,20 @@ set_num_buckets:
+       if (start == count)
+               goto out;
+-      xh = (struct ocfs2_xattr_header *)s_bhs[0]->b_data;
++      xh = bucket_xh(&s_bucket);
+       memset(&xh->xh_entries[start], 0,
+              sizeof(struct ocfs2_xattr_entry) * (count - start));
+       xh->xh_count = cpu_to_le16(start);
+       xh->xh_free_start = cpu_to_le16(name_offset);
+       xh->xh_name_value_len = cpu_to_le16(name_value_len);
+-      ocfs2_journal_dirty(handle, s_bhs[0]);
++      ocfs2_journal_dirty(handle, s_bucket.bu_bhs[0]);
+       if (ret)
+               mlog_errno(ret);
+ out:
+-      if (s_bhs) {
+-              for (i = 0; i < blk_per_bucket; i++)
+-                      brelse(s_bhs[i]);
+-      }
+-      kfree(s_bhs);
+-
+-      if (t_bhs) {
+-              for (i = 0; i < blk_per_bucket; i++)
+-                      brelse(t_bhs[i]);
+-      }
+-      kfree(t_bhs);
++      ocfs2_xattr_bucket_relse(inode, &s_bucket);
++      ocfs2_xattr_bucket_relse(inode, &t_bucket);
+       return ret;
+ }
+@@ -3384,7 +3390,7 @@ static int ocfs2_cp_xattr_bucket(struct inode *inode,
+       int ret, i;
+       int blk_per_bucket = ocfs2_blocks_per_xattr_bucket(inode->i_sb);
+       int blocksize = inode->i_sb->s_blocksize;
+-      struct buffer_head **s_bhs, **t_bhs = NULL;
++      struct ocfs2_xattr_bucket s_bucket, t_bucket;
+       BUG_ON(s_blkno == t_blkno);
+@@ -3392,28 +3398,23 @@ static int ocfs2_cp_xattr_bucket(struct inode *inode,
+            (unsigned long long)s_blkno, (unsigned long long)t_blkno,
+            t_is_new);
+-      s_bhs = kzalloc(sizeof(struct buffer_head *) * blk_per_bucket,
+-                      GFP_NOFS);
+-      if (!s_bhs)
+-              return -ENOMEM;
++      memset(&s_bucket, 0, sizeof(struct ocfs2_xattr_bucket));
++      memset(&t_bucket, 0, sizeof(struct ocfs2_xattr_bucket));
+-      ret = ocfs2_read_xattr_bucket(inode, s_blkno, s_bhs, 0);
++      ret = ocfs2_read_xattr_bucket(inode, &s_bucket, s_blkno);
+       if (ret)
+               goto out;
+-      t_bhs = kzalloc(sizeof(struct buffer_head *) * blk_per_bucket,
+-                      GFP_NOFS);
+-      if (!t_bhs) {
+-              ret = -ENOMEM;
+-              goto out;
+-      }
+-
+-      ret = ocfs2_read_xattr_bucket(inode, t_blkno, t_bhs, t_is_new);
++      /*
++       * Even if !t_is_new, we're overwriting t_bucket.  Thus,
++       * there's no need to read it.
++       */
++      ret = ocfs2_init_xattr_bucket(inode, &t_bucket, t_blkno);
+       if (ret)
+               goto out;
+       for (i = 0; i < blk_per_bucket; i++) {
+-              ret = ocfs2_journal_access(handle, inode, t_bhs[i],
++              ret = ocfs2_journal_access(handle, inode, t_bucket.bu_bhs[i],
+                                          t_is_new ?
+                                          OCFS2_JOURNAL_ACCESS_CREATE :
+                                          OCFS2_JOURNAL_ACCESS_WRITE);
+@@ -3422,22 +3423,14 @@ static int ocfs2_cp_xattr_bucket(struct inode *inode,
+       }
+       for (i = 0; i < blk_per_bucket; i++) {
+-              memcpy(t_bhs[i]->b_data, s_bhs[i]->b_data, blocksize);
+-              ocfs2_journal_dirty(handle, t_bhs[i]);
++              memcpy(bucket_block(&t_bucket, i), bucket_block(&s_bucket, i),
++                     blocksize);
++              ocfs2_journal_dirty(handle, t_bucket.bu_bhs[i]);
+       }
+ out:
+-      if (s_bhs) {
+-              for (i = 0; i < blk_per_bucket; i++)
+-                      brelse(s_bhs[i]);
+-      }
+-      kfree(s_bhs);
+-
+-      if (t_bhs) {
+-              for (i = 0; i < blk_per_bucket; i++)
+-                      brelse(t_bhs[i]);
+-      }
+-      kfree(t_bhs);
++      ocfs2_xattr_bucket_relse(inode, &s_bucket);
++      ocfs2_xattr_bucket_relse(inode, &t_bucket);
+       return ret;
+ }
+-- 
+1.5.6
+