lkml.org 
[lkml]   [2000]   [Nov]   [17]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
Patch in this message
/
Date
From
Subject[PATCH] get_empty_inode cleanup - part 2
	* Arrgh. Hell knows how, but %s/new_inode/ntfs_&/g in fs/ntfs/inode.c
mentioned in the previous part didn't make it into the patch I've sent.
Mea maxima culpa. Fixed.
* More duplicated initializations removed:
* get_empty_inode() sets i_flags to 0. NFS and UDF did the
reinitialization immediately after new_inode(). To 0.
* Ditto for i_size. Redundant stuff was in UDF, sysvfs, FAT.
* capifs had the following gem:
capifs_read_inode(struct inode *inode)
{
...
inode->i_op = NULL;
... more assignments to fields, then
if (ino == 1) {
inode->i_op = &capifs_root_inode_operations;
...
return;
}
/* need dummy inode operations .... */
inode->i_op = &capifs_inode_operations;
...
}
The first assignment is interesting, indeed, but so is the third
one: capifs_inode_operations is empty. Just as the default one
placed there by caller of ->read_inode() (empty_iops). Crapectomy
performed...
* shm_read_inode() starts with ->i_op = NULL. After that we attempt
shm_lock() and if it fails - return immediately. Correct action:
make_bad_inode(). Every other branch initializes ->i_op, so the
first assignment is unneeded.
* JFFS sets ->i_flags to ->s_flags. Atavism, since s_flags is not
mirrored in ->i_flags anymore (inherited from 2.0 version). Makes
absolutely no sense these days, since nothing (including JFFS)
looks for s_flags bits in i_flags - as the matter of fact they are
completely independent. Removed.
* All callers of ->read_inode() (OK, the only caller ;-) call
clean_inode() immediately before. Several instances have duplicated
initializers for ->i_size (to 0), ->i_nlink (to 1), etc. (obviously,
normal filesystems have meaningful initializations there - I'm
talking only about devfs/capifs/etc.)

Please, apply.
Cheers,
Al
PS: at that point the official tree _never_ has ->i_op == NULL for any
inode. So Tigran's idea about dropping the checks for NULL ->i_op looks
fine, modulo 3rd-party filesystems. For 2.5 it's definitely the right
thing, but it might be OK even now... Comments?

diff -urN rc11-pre6/drivers/isdn/avmb1/capifs.c rc11-6-new_inode/drivers/isdn/avmb1/capifs.c
--- rc11-pre6/drivers/isdn/avmb1/capifs.c Thu Nov 2 22:38:31 2000
+++ rc11-6-new_inode/drivers/isdn/avmb1/capifs.c Fri Nov 17 03:44:58 2000
@@ -119,8 +119,6 @@
lookup: capifs_root_lookup,
};

-struct inode_operations capifs_inode_operations;
-
static struct dentry_operations capifs_dentry_operations = {
d_revalidate: capifs_revalidate,
};
@@ -468,10 +466,8 @@
ino_t ino = inode->i_ino;
struct capifs_sb_info *sbi = SBI(inode->i_sb);

- inode->i_op = NULL;
inode->i_mode = 0;
inode->i_nlink = 0;
- inode->i_size = 0;
inode->i_mtime = inode->i_atime = inode->i_ctime = CURRENT_TIME;
inode->i_blocks = 0;
inode->i_blksize = 1024;
@@ -484,9 +480,6 @@
inode->i_nlink = 2;
return;
}
-
- /* need dummy inode operations .... */
- inode->i_op = &capifs_inode_operations;

ino -= 2;
if ( ino >= sbi->max_ncci )
diff -urN rc11-pre6/fs/autofs/inode.c rc11-6-new_inode/fs/autofs/inode.c
--- rc11-pre6/fs/autofs/inode.c Thu Apr 27 22:09:22 2000
+++ rc11-6-new_inode/fs/autofs/inode.c Fri Nov 17 03:46:09 2000
@@ -211,7 +211,6 @@
inode->i_fop = &autofs_dir_operations;
inode->i_mode = S_IFDIR | S_IRUGO | S_IXUGO;
inode->i_nlink = 2;
- inode->i_size = 0;
inode->i_mtime = inode->i_atime = inode->i_ctime = CURRENT_TIME;
inode->i_blocks = 0;
inode->i_blksize = 1024;
diff -urN rc11-pre6/fs/devfs/base.c rc11-6-new_inode/fs/devfs/base.c
--- rc11-pre6/fs/devfs/base.c Fri Nov 17 02:23:19 2000
+++ rc11-6-new_inode/fs/devfs/base.c Fri Nov 17 03:47:06 2000
@@ -2250,7 +2250,6 @@
printk ("%s: read_inode(%d): VFS inode: %p devfs_entry: %p\n",
DEVFS_NAME, (int) inode->i_ino, inode, de);
#endif
- inode->i_size = 0;
inode->i_blocks = 0;
inode->i_blksize = 1024;
inode->i_op = &devfs_iops;
diff -urN rc11-pre6/fs/fat/inode.c rc11-6-new_inode/fs/fat/inode.c
--- rc11-pre6/fs/fat/inode.c Fri Nov 17 02:23:19 2000
+++ rc11-6-new_inode/fs/fat/inode.c Fri Nov 17 03:14:24 2000
@@ -794,7 +794,6 @@
inode->i_nlink = 1;
}
#endif
- inode->i_size = 0;
if ((nr = MSDOS_I(inode)->i_start) != 0)
while (nr != -1) {
inode->i_size += SECTOR_SIZE*sbi->cluster_size;
@@ -818,7 +817,6 @@
(CF_LE_W(de->starthi) << 16);
}
MSDOS_I(inode)->i_logstart = MSDOS_I(inode)->i_start;
- inode->i_nlink = 1;
inode->i_size = CF_LE_L(de->size);
inode->i_op = &fat_file_inode_operations;
inode->i_fop = &fat_file_operations;
diff -urN rc11-pre6/fs/jffs/inode-v23.c rc11-6-new_inode/fs/jffs/inode-v23.c
--- rc11-pre6/fs/jffs/inode-v23.c Fri Nov 17 02:23:20 2000
+++ rc11-6-new_inode/fs/jffs/inode-v23.c Fri Nov 17 03:40:40 2000
@@ -349,7 +349,6 @@
inode->i_blksize = PAGE_SIZE;
inode->i_blocks = (inode->i_size + 511) >> 9;
inode->i_version = 0;
- inode->i_flags = sb->s_flags;
inode->u.generic_ip = (void *)jffs_find_file(c, raw_inode->ino);

insert_inode_hash(inode);
diff -urN rc11-pre6/fs/nfs/inode.c rc11-6-new_inode/fs/nfs/inode.c
--- rc11-pre6/fs/nfs/inode.c Fri Nov 17 02:23:21 2000
+++ rc11-6-new_inode/fs/nfs/inode.c Fri Nov 17 02:40:51 2000
@@ -723,7 +723,6 @@
struct inode *inode = new_inode(sb);
if (!inode)
goto out;
- inode->i_flags = 0;
inode->i_ino = nfs_fattr_to_ino_t(fattr);
nfs_read_inode(inode);
nfs_fill_inode(inode, fattr);
diff -urN rc11-pre6/fs/ntfs/inode.c rc11-6-new_inode/fs/ntfs/inode.c
--- rc11-pre6/fs/ntfs/inode.c Thu Jul 27 09:36:37 2000
+++ rc11-6-new_inode/fs/ntfs/inode.c Fri Nov 17 02:33:33 2000
@@ -1050,7 +1050,7 @@

/* We have to skip the 16 metafiles and the 8 reserved entries */
static int
-new_inode (ntfs_volume* vol,int* result)
+ntfs_new_inode (ntfs_volume* vol,int* result)
{
int byte,error;
int bit;
@@ -1236,11 +1236,11 @@
ntfs_volume* vol=dir->vol;
int byte,bit;

- error=new_inode (vol,&(result->i_number));
+ error=ntfs_new_inode (vol,&(result->i_number));
if(error==ENOSPC){
error=ntfs_extend_mft(vol);
if(error)return error;
- error=new_inode(vol,&(result->i_number));
+ error=ntfs_new_inode(vol,&(result->i_number));
}
if(error){
ntfs_error ("ntfs_get_empty_inode: no free inodes\n");
diff -urN rc11-pre6/fs/sysv/ialloc.c rc11-6-new_inode/fs/sysv/ialloc.c
--- rc11-pre6/fs/sysv/ialloc.c Fri Nov 17 02:23:22 2000
+++ rc11-6-new_inode/fs/sysv/ialloc.c Fri Nov 17 03:08:30 2000
@@ -142,7 +142,6 @@
mark_inode_dirty(inode);
/* Change directory entry: */
inode->i_mode = 0; /* for sysv_write_inode() */
- inode->i_size = 0; /* ditto */
sysv_write_inode(inode, 0); /* ensure inode not allocated again */
/* FIXME: caller may call this too. */
mark_inode_dirty(inode); /* cleared by sysv_write_inode() */
diff -urN rc11-pre6/fs/udf/ialloc.c rc11-6-new_inode/fs/udf/ialloc.c
--- rc11-pre6/fs/udf/ialloc.c Fri Nov 17 02:23:22 2000
+++ rc11-6-new_inode/fs/udf/ialloc.c Fri Nov 17 03:06:57 2000
@@ -84,7 +84,6 @@
*err = -ENOMEM;
return NULL;
}
- inode->i_flags = 0;
*err = -ENOSPC;

block = udf_new_block(dir, UDF_I_LOCATION(dir).partitionReferenceNum,
@@ -128,7 +127,6 @@
inode->i_ino = udf_get_lb_pblock(sb, UDF_I_LOCATION(inode), 0);
inode->i_blksize = PAGE_SIZE;
inode->i_blocks = 0;
- inode->i_size = 0;
UDF_I_LENEATTR(inode) = 0;
UDF_I_LENALLOC(inode) = 0;
if (UDF_QUERY_FLAG(inode->i_sb, UDF_FLAG_USE_EXTENDED_FE))
diff -urN rc11-pre6/ipc/shm.c rc11-6-new_inode/ipc/shm.c
--- rc11-pre6/ipc/shm.c Wed Oct 4 03:45:11 2000
+++ rc11-6-new_inode/ipc/shm.c Fri Nov 17 03:34:34 2000
@@ -376,13 +376,14 @@
struct shmid_kernel *shp;

id = inode->i_ino;
- inode->i_op = NULL;
inode->i_mode = 0;
inode->i_mtime = inode->i_atime = inode->i_ctime = CURRENT_TIME;

if (id < SEQ_MULTIPLIER) {
- if (!(shp = shm_lock (id)))
+ if (!(shp = shm_lock (id))) {
+ make_bad_inode(inode);
return;
+ }
inode->i_mode = (shp->shm_flags & S_IALLUGO) | S_IFREG;
inode->i_uid = shp->shm_perm.uid;
inode->i_gid = shp->shm_perm.gid;
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
Please read the FAQ at http://www.tux.org/lkml/

\
 
 \ /
  Last update: 2005-03-22 12:47    [W:0.033 / U:3.256 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site