Re: [PATCH v3 3/7] mm: add find_get_entries_tag()

[Dan Williams <dan.j.williams@xxxxxxxxx>]

On Thu, Dec 10, 2015 at 12:24 PM, Ross Zwisler
<ross.zwisler@xxxxxxxxxxxxxxx> wrote:
> On Wed, Dec 09, 2015 at 11:44:16AM -0800, Dan Williams wrote:
>> On Tue, Dec 8, 2015 at 11:18 AM, Ross Zwisler
>> <ross.zwisler@xxxxxxxxxxxxxxx> wrote:
>> > Add find_get_entries_tag() to the family of functions that include
>> > find_get_entries(), find_get_pages() and find_get_pages_tag().  This is
>> > needed for DAX dirty page handling because we need a list of both page
>> > offsets and radix tree entries ('indices' and 'entries' in this function)
>> > that are marked with the PAGECACHE_TAG_TOWRITE tag.
>> >
>> > Signed-off-by: Ross Zwisler <ross.zwisler@xxxxxxxxxxxxxxx>
> <>
>> Why does this mostly duplicate find_get_entries()?
>>
>> Surely find_get_entries() can be implemented as a special case of
>> find_get_entries_tag().
>
> I'm adding find_get_entries_tag() to the family of functions that already
> exist and include find_get_entries(), find_get_pages(),
> find_get_pages_contig() and find_get_pages_tag().
>
> These functions all contain very similar code with small changes to the
> internal looping based on whether you're looking through all radix slots or
> only the ones that match a certain tag (radix_tree_for_each_slot() vs
> radix_tree_for_each_tagged()).
>
> We already have find_get_page() to get all pages in a range and
> find_get_pages_tag() to get all pages in the range with a certain tag.  We
> have find_get_entries() to get all pages and indices for a given range, but we
> are currently missing find_get_entries_tag() to do that same search based on a
> tag, which is what I'm adding.
>
> I agree that we could probably figure out a way to combine the code for
> find_get_entries() with find_get_entries_tag(), as we could do for the
> existing functions find_get_pages() and find_get_pages_tag().  I think we
> should probably add find_get_entries_tag() per this patch, though, and then
> decide whether to do any combining later as a separate step.

Ok, sounds good to me.