On Fri, Apr 15, 2016 at 10:37 AM, Verma, Vishal L
<vishal.l.verma@intel.com> wrote:
> On Fri, 2016-04-15 at 13:11 -0400, Jeff Moyer wrote:
[..]
>> >
>> > But, how does _EIOCBQUEUED work? Maybe we need an exception for it?
>> For async direct I/O, only the setup phase of the I/O is performed
>> and
>> then we return to the caller.  -EIOCBQUEUED signifies this.
>>
>> You're heading towards code that looks like this:
>>
>>         if (IS_DAX(inode)) {
>>                 ret = dax_do_io(iocb, inode, iter, offset,
>> blkdev_get_block,
>>                                 NULL, DIO_SKIP_DIO_COUNT);
>>                 if (ret == -EIO && (iov_iter_rw(iter) == WRITE))
>>                         ret_saved = ret;
>>                 else
>>                         return ret;
>>         }
>>
>>         ret = __blockdev_direct_IO(iocb, inode, I_BDEV(inode), iter,
>> offset,
>>                                     blkdev_get_block, NULL, NULL,
>>                                     DIO_SKIP_DIO_COUNT);
>>         if (ret < 0 && ret != -EIOCBQUEUED && ret_saved)
>>                 return ret_saved;
>>
>> There's a lot of special casing here, so you might consider adding
>> comments.
>
> Correct - maybe we should reconsider wrapper-izing this? :)

Another option is just to skip dax_do_io() and this special casing
fallback entirely if errors are present.  I.e. only attempt dax_do_io
when: IS_DAX() && gendisk->bb && bb->count == 0.