mirror of https://lore.kernel.org/lkml/
 help / color / mirror / Atom feed
From: "Takashi Sato" <sho@bsd.tnes.nec.co.jp>
To: <linux-kernel@vger.kernel.org>, <linux-fsdevel@vger.kernel.org>
Subject: stat64 for over 2TB file returned invalid st_blocks
Date: Thu, 1 Dec 2005 21:00:26 +0900	[thread overview]
Message-ID: <01e901c5f66e$d4551b70$4168010a@bsd.tnes.nec.co.jp> (raw)

Hi all,

I found a problem at stat64 on 32bit architecture.

When I called stat64 for a file which is larger than 2TB, stat64
returned an invalid number of blocks at st_blocks on 32bit
architecture, although it returned a valid number of blocks on 64bit
architecture(ia64).

The following describes the cause of this issue:
i_blocks in inode is 4bytes on 32bit architecture.  If it receives
more than 2^32 number of blocks, it would overflow and set an
invalid number to st_blocks.

Below describes a sequence of setting overflowed inode.i_blocks
to st_blocks through stat64.

1. generic_fillattr(struct inode *inode, struct kstat *stat)
  - Copy data from overflowed inode.i_blocks to kstat.blocks.

2. vfs_getattr(struct vfsmount *mnt, struct dentry *dentry,
        struct kstat *stat)
  - Return invalid kstat.blocks to sys_stat64().

3. sys_stat64(char __user * filename, struct stat64 __user * statbuf)
  - Copy data from invalid kstat.blocks to stat64.st_blocks.

I also found the following problem.

- ioctl with FIOQSIZE command returns the size of file's data which
  has written to disk.  The size of file's data is calculated as
  follows in inode_get_bytes().
   
   (((loff_t)inode->i_blocks) << 9) + inode->i_bytes

   On the file which is larger than 2TB, the ioctl will return an
   invalid size because i_blocks can't express the right number of
   blocks.

I think the following modification is essential to fix these
problems.

1. Change the type of inode.i_blocks and kstat.blocks from unsigned
   long to unsigned long long.

2. Change the type of architecture dependent stat64.st_blocks in
   include/asm/asm-*/stat.h from unsigned long to unsigned long long.
   I tried modifying only stat64 of 32bit architecture
   (include/asm-i386/stat.h).

I have some tested for a file whose size is 3TB on JFS filesystem.
The following is the patch.

Signed-off-by: Takashi Sato <sho@bsd.tnes.nec.co.jp>

diff -uprN -X linux-2.6.14.org/Documentation/dontdiff linux-2.6.14.or
g/include/asm-i386/stat.h linux-2.6.14-blocks/include/asm-i386/stat.h
--- linux-2.6.14.org/include/asm-i386/stat.h 2005-10-28 09:02:08.000000000 +0900
+++ linux-2.6.14-blocks/include/asm-i386/stat.h 2005-11-18 22:42:37.000000000 +0900
@@ -58,8 +58,7 @@ struct stat64 {
  long long st_size;
  unsigned long st_blksize;
 
- unsigned long st_blocks; /* Number 512-byte blocks allocated. */
- unsigned long __pad4;  /* future possible st_blocks high bits */
+ unsigned long long st_blocks; /* Number 512-byte blocks allocated. */
 
  unsigned long st_atime;
  unsigned long st_atime_nsec;
diff -uprN -X linux-2.6.14.org/Documentation/dontdiff linux-2.6.14.or
g/include/linux/fs.h linux-2.6.14-blocks/include/linux/fs.h
--- linux-2.6.14.org/include/linux/fs.h 2005-10-28 09:02:08.000000000 +0900
+++ linux-2.6.14-blocks/include/linux/fs.h 2005-11-18 17:08:03.000000000 +0900
@@ -438,7 +438,7 @@ struct inode {
  unsigned int  i_blkbits;
  unsigned long  i_blksize;
  unsigned long  i_version;
- unsigned long  i_blocks;
+ unsigned long long i_blocks;
  unsigned short          i_bytes;
  spinlock_t  i_lock; /* i_blocks, i_bytes, maybe i_size */
  struct semaphore i_sem;
diff -uprN -X linux-2.6.14.org/Documentation/dontdiff linux-2.6.14.or
g/include/linux/stat.h linux-2.6.14-blocks/include/linux/stat.h
--- linux-2.6.14.org/include/linux/stat.h 2005-10-28 09:02:08.000000000 +0900
+++ linux-2.6.14-blocks/include/linux/stat.h 2005-11-18 17:08:56.000000000 +0900
@@ -69,7 +69,7 @@ struct kstat {
  struct timespec mtime;
  struct timespec ctime;
  unsigned long blksize;
- unsigned long blocks;
+ unsigned long long blocks;
 };
 
 #endif

Any feedback and comments are welcome.

Best regards, Takashi Sato

             reply	other threads:[~2005-12-01 12:00 UTC|newest]

Thread overview: 31+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2005-12-01 12:00 Takashi Sato [this message]
2005-12-01 12:39 ` Jörn Engel
2005-12-01 12:52 ` Jörn Engel
2005-12-01 13:52   ` Avi Kivity
2005-12-01 14:32 ` Dave Kleikamp
2005-12-02 13:18   ` Takashi Sato
2005-12-02 14:11     ` Dave Kleikamp
2005-12-02 18:58     ` Andreas Dilger
2005-12-03 13:00       ` Takashi Sato
2005-12-05  8:11         ` Andreas Dilger
2005-12-05 12:35           ` Takashi Sato
2005-12-05 13:34           ` Trond Myklebust
2005-12-01 14:53 ` Al Viro
     [not found] <5eVqw-2ug-61@gated-at.bofh.it>
2005-12-03  2:19 ` Bodo Eggert
2005-12-06 12:42 Takashi Sato
2005-12-06 14:30 ` Dave Kleikamp
2005-12-06 14:48   ` Trond Myklebust
2005-12-06 14:51     ` Dave Kleikamp
2005-12-06 21:24     ` Andreas Dilger
2005-12-07  0:59       ` Trond Myklebust
2005-12-07 10:57       ` Takashi Sato
2005-12-07 13:52         ` Trond Myklebust
2005-12-07 15:01           ` Dave Kleikamp
2005-12-07 15:34             ` Trond Myklebust
2005-12-07 16:34               ` Dave Kleikamp
2005-12-07 18:55                 ` Trond Myklebust
2005-12-08 11:38                 ` Takashi Sato
2005-12-08 14:27                   ` Trond Myklebust
2005-12-08 14:50                     ` Anton Altaparmakov
2005-12-08 15:03                       ` Trond Myklebust
2005-12-10 11:22                     ` Takashi Sato

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to='01e901c5f66e$d4551b70$4168010a@bsd.tnes.nec.co.jp' \
    --to=sho@bsd.tnes.nec.co.jp \
    --cc=linux-fsdevel@vger.kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox

all inboxes | Powered by JetHome®