Re: [patch 3/5] fuse: fix race in llseek

From: Miklos Szeredi
Date: Mon Apr 28 2008 - 03:51:54 EST


> > +static loff_t fuse_file_llseek(struct file *file, loff_t offset, int origin)
> > +{
> > + long long retval;
>
> I switched this to have a type of loff_t.

Ugh, right. I wonder where that long long came from. I probably
copied the generic_file_llseek() from some earlier kernel...

>
> > + struct inode *inode = file->f_path.dentry->d_inode;
> > +
> > + mutex_lock(&inode->i_mutex);
> > + switch (origin) {
> > + case SEEK_END:
> > + offset += i_size_read(inode);
>
> As we hold i_mutex we could directly read inode->i_size here, save a few
> cycles.

No, that's the whole point of the fuse specific function: fuse doesn't
use i_mutex to protect writing i_size, instead it uses a per
filesystem spinlock.

This is probably worth a comment though, so I'll do that.

Thanks,
Miklos
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/