Re: [RFC PATCH V3 2/2] fs : Add sanity checks for block size > PAGE_SIZE

[Raghavendra K T <raghavendra.kt@linux.vnet.ibm.com>]

On 01/22/2014 01:51 AM, Andrew Morton wrote:
> On Tue, 21 Jan 2014 17:00:00 +0530 Raghavendra K T <raghavendra.kt@linux.vnet.ibm.com> wrote:
>
>> We could hit null pointer dereference error during alloc_page_buffers
>> in : (1) block size > PAGE_SIZE (2) low memory.
>> Add sanity check for that.
>>
>> Signed-off-by: Raghavendra K T <raghavendra.kt@linux.vnet.ibm.com>
>> ---
>>   fs/block_dev.c | 1 +
>>   fs/buffer.c    | 6 ++++++
>>   2 files changed, 7 insertions(+)
>>
>> diff --git a/fs/block_dev.c b/fs/block_dev.c
>> index 1e86823..2481d42 100644
>> --- a/fs/block_dev.c
>> +++ b/fs/block_dev.c
>> @@ -1027,6 +1027,7 @@ void bd_set_size(struct block_device *bdev, loff_t size)
>>   			break;
>>   		bsize <<= 1;
>>   	}
>> +	BUG_ON(bsize > PAGE_SIZE);
>>   	bdev->bd_block_size = bsize;
>>   	bdev->bd_inode->i_blkbits = blksize_bits(bsize);
>>   }
>
> alloc_page_buffers() will always return NULL if passed size >=
> PAGE_SIZE.  So if we're going to add a check, it would be better to add
> it to alloc_page_buffers() because that will catch errors from the
> widest range of callsites.

In that case how about converting BUG_ON to setting a default value of
PAGE_SIZE for bs in bd_set_size() itself (with a warning)?

>
> But alloc_page_buffers() is pretty frequently called and I'd be
> inclined to not add any check - most callers will just go oops and that
> will provide basically the same information.

Agree with this concern.

>
>> --- a/fs/buffer.c
>> +++ b/fs/buffer.c
>> @@ -1571,6 +1571,12 @@ void create_empty_buffers(struct page *page,
>>   	struct buffer_head *bh, *head, *tail;
>>
>>   	head = alloc_page_buffers(page, blocksize, 1);
>> +
>> +	/*
>> + 	 * alloc_page_buffers() could return NULL on (1) bs > PAGE_SIZE
>> + 	 * (2) low memory case. Ensure that we don't dereference null ptr
>> + 	 */
>> +	BUG_ON(!head);
>
> This is unneeded.
>
> - bs > PAGE_SIZE can be checked elsewhere in a direct fashion
>
> - low memory case can't happen - we passed retry=1
>
> - create_empty_buffers() will immediately go oops if head==NULL.
>    That oops contains the same info as is presented by a BUG().

Okay.