Do you plan to persist the urls collected by WasUnableToDownloadFavicon calls or something different?

I was planning on storing a mapping between a 404 page URL to an entry in the favicon database with no favicon_bitmaps.
This is to handle the case of us getting 404s for the default favicon (favicon.ico) which we try to fetch if the page does not have any <link rel="icon"> URLs

This would enable removing WasUnableToDownloadFavicon()

I was wondering because WasUnableToDownloadFavicon() is now being used from the new function in LargeIconService for downloading from the Google server. 

If I get it right, that case could also be handled using your DB-based solution.

Yes, this could be handled in the database based solution too