Re: Kernel 3.1.0-rc4 oops when connecting iPod - Hin-Tak Leung

From: Hin-Tak Leung <hintak_leung@yahoo.co.uk>
To: Christoph Hellwig <hch@infradead.org>, Pavel Ivanov <paivanof@gmail.com>
Cc: linux-fsdevel@vger.kernel.org,
	linux-kernel <linux-kernel@vger.kernel.org>,
	Christoph Hellwig <hch@tuxera.com>
Subject: Re: Kernel 3.1.0-rc4 oops when connecting iPod
Date: Mon, 12 Sep 2011 16:57:30 +0100 (BST)	[thread overview]
Message-ID: <1315843050.34041.YahooMailClassic@web29505.mail.ird.yahoo.com> (raw)
In-Reply-To: <CAG1a4rsuYVehnSCM3KxeYQ+g94nrivG1vfss=y8275E4D24qTg@mail.gmail.com>

--- On Mon, 12/9/11, Pavel Ivanov <paivanof@gmail.com> wrote:

> On Mon, Sep 12, 2011 at 10:34 AM,
> Christoph Hellwig <hch@infradead.org>
> wrote:
> > Does this patch fix your issues with large block
> sizes?
> 
> I'll be able to try it in the evening but meanwhile I have
> some comments below.
> 
> >
> >
> > Index: linux-2.6/fs/hfsplus/super.c
> >
> ===================================================================
> > --- linux-2.6.orig/fs/hfsplus/super.c   2011-09-12
> 09:56:58.619988416 -0400
> > +++ linux-2.6/fs/hfsplus/super.c        2011-09-12
> 10:07:18.006651395 -0400
> > @@ -344,6 +344,7 @@ static int
> hfsplus_fill_super(struct sup
> >        struct inode *root, *inode;
> >        struct qstr str;
> >        struct nls_table *nls = NULL;
> > +       u64 last_fs_block, last_fs_page;
> >        int err;
> >
> >        err = -EINVAL;
> > @@ -399,9 +400,13 @@ static int
> hfsplus_fill_super(struct sup
> >        if (!sbi->rsrc_clump_blocks)
> >                sbi->rsrc_clump_blocks = 1;
> >
> > -       err =
> generic_check_addressable(sbi->alloc_blksz_shift,
> > -                                  
>     sbi->total_blocks);
> > -       if (err) {
> > +       err = -EFBIG;
> > +       last_fs_block = sbi->total_blocks - 1;
> > +       last_fs_page = (last_fs_block >>
> sbi->alloc_blksz_shift) <<
> > +                       PAGE_CACHE_SHIFT;
> 
> Did you mix left and right shifts here? Expression doesn't
> make sense to me.
> 
> Also I have a little concern about consistency in using
> PAGE_CACHE_SHIFT and PAGE_SHIFT. hfsplus_read_wrapper()
> limits visible
> block size to PAGE_SIZE, not PAGE_CACHE_SIZE. And although
> now they
> are equal comment in linux/pagemap.h clearly says that
> PAGE_CACHE_SIZE
> can be bigger than PAGE_SIZE. Is it something that should
> be fixed in
> hfsplus_read_wrapper() ?
> 
> > +
> > +       if ((last_fs_block > (sector_t)(~0ULL)
> >> (sbi->alloc_blksz_shift - 9)) ||
> 
> Maybe this 9 should be extracted from here and
> generic_check_addressable() into some macro?
> 
> > +           (last_fs_page > (pgoff_t)(~0ULL)))
> {
> >                printk(KERN_ERR "hfs:
> filesystem size too large.\n");
> >                goto out_free_vhdr;
> >        }
> >

I 2nd that this is kind of ugly. The literal "9". How about abstracting this logic out to say, hfs_check_addressable()?