On 12.11.18 23:32, Eric Blake wrote:
> On 8/9/18 5:31 PM, Max Reitz wrote:
>> For completeness' sake, add a function for accessing a node's storage
>> child, too.  For filters, this is there filtered child; for non-filters,
> 
> s/there/their/
> 
>> this is bs->file.
>>
>> Some places are deliberately left unconverted:
>> - BDS opening/closing functions where bs->file is handled specially
>>    (which is basically wrong, but at least simplifies probing)
>> - bdrv_co_block_status_from_file(), because its name implies that it
>>    points to ->file
> 
> I'm wondering if we can clean up block_status to let filters have a NULL
> callback and io.c do the right thing automatically, rather than the
> current approach of filters assigning the callback to the common helper
> routine.  Maybe later in the series.

Hm!  I didn't even think of that.  Yes, everything that uses
bdrv_co_block_status_from_*() seems to be an R/W filter, and which they
use simply depends on whether they use bs->backing or bs->file.  Well,
blkdebug is the exception, because it really wants to put an assertion
there, but inlining the code there shouldn't be the showstopper.

Max