Re: i_size misuses

[Mikulas Patocka <mpatocka@redhat.com>]

On Wed, 20 Oct 2010, Neil Brown wrote:

> On Tue, 19 Oct 2010 14:55:31 -0400 (EDT)
> Mikulas Patocka <mpatocka@redhat.com> wrote:
> 
> > 
> > 
> > On Sun, 12 Sep 2010, Neil Brown wrote:
> > 
> > > On Wed, 8 Sep 2010 09:32:13 -0400 (EDT)
> > > Mikulas Patocka <mpatocka@redhat.com> wrote:
> > > 
> > > > Hi
> > > > 
> > > > when reviewing some i_size problem, I searched the kernel for i_size usage 
> > > > (the variable should really be written with i_size_write and read with 
> > > > i_size_read).
> > > > 
> > > > Properly locked direct use of "i_size" inside memory management or 
> > > > filesystems may not be a problem, but there are many problems in general 
> > > > code outside mm.
> > > > 
> > > > The misuses are:
> > > > SOUND/SOUND_FIRMWARE.C:l = filp->f_path.dentry->d_inode->i_size;
> > > > KERNEL/RELAY.C:buf->dentry->d_inode->i_size = buf->early_bytes;
> > > > KERNEL/RELAY.C:buf->dentry->d_inode->i_size += buf->chan->subbuf_size 
> > > > -buf->padding[old_subbuf];
> > > > DRIVERS/USB/CORE/INODE.C:dev->usbfs_dentry->d_inode->i_size = i_size;
> > > > DRIVERS/MTD/DEVICES/BLOCK2MTD.C:dev->mtd.size = 
> > > > dev->blkdev->bd_inode->i_size & PAGE_MASK;
> > > > DRIVERS/MD/MD.C: many reads of i_size 
> > > > DRIVERS/BLOCK/NBD.C: many writes to i_size without i_size_write
> > > > DRIVERS/BLOCK/DRBD/DRBD_INT.H: return bdev ? bdev->bd_inode->i_size >> 9 : 0;
> > > > DRIVERS/BLOCK/DRBD/DRBD_WRAPPERS.H: mdev->this_bdev->bd_inode->i_size = 
> > > > (loff_t)size << 9;
> > > > BLOCK/BLK-CORE.C:printk(KERN_INFO "%s: rw=%ld, want=%Lu, limit=%Lu\n",
> > > >                 bdevname(bio->bi_bdev, b),
> > > >                 bio->bi_rw,
> > > >                 (unsigned long long)bio->bi_sector + bio_sectors(bio),
> > > >                 (long long)(bio->bi_bdev->bd_inode->i_size >> 9));
> > > > maxsector = bio->bi_bdev->bd_inode->i_size >> 9;
> > > > BLOCK/COMPAT_IOCTL.C: size = bdev->bd_inode->i_size;
> > > > return compat_put_u64(arg, bdev->bd_inode->i_size);
> > > > BLOCK/IOCTL.C: if (start + len > (bdev->bd_inode->i_size >> 9))
> > > > size = bdev->bd_inode->i_size;
> > > > return put_u64(arg, bdev->bd_inode->i_size);
> > > > 
> > > > The problem with this code is that if you read i_size without i_size_read 
> > > > and the size wraps around 32 bits, for example from 0xffffffff to 
> > > > 0x100000000 , there is a possibility on 32-bit machines to read an invalid 
> > > > value (either 0 or 0x1ffffffff). Similarly, if you write i_size without 
> > > > i_size_write, the readers can see intermediate invalid values.
> > > > 
> > > > 
> > > > The original problem that caused this investigation is the question, how a 
> > > > block device driver can change the size of its device. Normal method (used 
> > > > in a few drivers, including dm), consists of
> > > > mutex_lock(&inode->i_mutex);
> > > > i_size_write(inode, new_size);
> > > > mutex_unlock(&inode->i_mutex);
> > > 
> > > Don't you just do
> > > 
> > >   set_capacity(gendisk, sectors);
> > >   revalidate(gendisk);
> > > 
> > > ??
> > > 
> > > NeilBrown
> > 
> > revalidate_disk() has still the problem that it changes i_size without 
> > holding i_mutex and other kernel parts (for example generic_file_llseek) 
> > assume that if we hold the lock, i_size_can't be changed.
> 
> generic_file_llseek is not used for block devices.  They use block_llseek
> which uses i_size_read, so I think it is safe.
> 
> > 
> > The rules for accessing i_size must be specified and followed.
> 
> I agree.  However the rules can be different for different file systems and
> file types.
> A filesystem that used the generic_* function would need to only change
> i_size under i_mutex as you say.
> For block devices it appears that the rule is that it can only be changed
> under bd_mutex.
> For 'relay' (which you mentioned above), it seem the relevant mutex is the
> global relay_channels_mutex, though I didn't read the code thoroughly to be
> sure.

You are right.

> It still would probably be useful to review all the i_size related code to
> ensure that it is safe, but you should not assume that everything follows the
> same rules.  So first you need to work out the rule for a given subsystem,
> then audit it against that rule (and maybe document that rule if it isn't
> already documented!)
> 
> NeilBrown

If you specify complex rules (i_size can be read directly under i_mutex 
for files, but can't be read for block devices under i_mutex), it will 
only complicate review and it may introduce bugs in the future --- 
remember that Linux doesn't have a formal specification and the 
specification is inferred from the code --- thus, you use 
i_size_read(inode) somewhere and inode->i_size elsewhere, people will 
infer a wrong specification --- as shown above on the few examples of 
i_size abusers.

I'd change i_size to __i_size (this will warn people not to touch it) and 
convert all accesses to i_size_read and i_size_write.

Mikulas