Hi!

On 10/21/2013 07:47 PM, [email protected] wrote:
> Hi!
>> -static int tbio_open(struct inode *inode, struct file *filep)
>> +static int tbio_open(struct block_device *blk, fmode_t mode)
>>   {
>> -    if (!Device.bdev) {
>> -            Device.bdev = inode->i_bdev;
>> -            //atomic_inc((atomic_t)&Device.bdev->bd_part_count);
>> -    }
>> -
>>      return 0;
>>   }
>>   
>> -static int tbio_release(struct inode *inode, struct file *filep)
>> +static int tbio_release(struct gendisk *gd, fmode_t mode)
> This function is declared as void since 
> db2a144bedd58b3dcf19950c2f476c58c9f39d18
>
> I guess that we should fix this by adding ifdefs for kernel versions
> around the function declaration and around the return.
I see, I will add ifdefs.
> The rest of the code looks good to me.
>
> Apart from that I'm wondering if it would be a good idea to add read()
> and write() tests to the userspace program as well. Something as writing
> to the device, issuing fsync() and checking that the request made it to
> the block device driver once we have returned from it... But that can be
> done in subsequent patches.
>
I agree, some new test-cases can be implemented, I think, the above with 
fsync() is a good one.

Thanks,
Alexey

------------------------------------------------------------------------------
October Webinars: Code for Performance
Free Intel webinars can help you accelerate application performance.
Explore tips for MPI, OpenMP, advanced profiling, and more. Get the most from 
the latest Intel processors and coprocessors. See abstracts and register >
http://pubads.g.doubleclick.net/gampad/clk?id=60135991&iu=/4140/ostg.clktrk
_______________________________________________
Ltp-list mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/ltp-list

Reply via email to