As an example of what CNAME cloaking is, let's say that a SAAS provider A wants to provide you, company Q, with fancy ad tracking software. In the olden days, they'd tell you to embed a script at e.g. https://A-ads-tracking.example into your website at address https://q-company.example

To block those ads, blocklists that uBlock Origin use have rules then that say "block requests being made to the domain name A-ads-tracking.example", which blocks the ads.

CNAME cloaking is where SAAS provider A sets up their ad-tracking services not on domain A-ads-tracking.example, but instead at a specific IP address of e.g. 29.1.2.3; then (and here's the important part) SAAS A tells you Company Q that you need to set up a subdomain of q-company.example which has a CNAME record pointing to 23.1.2.3, a subdomain with an innocuous name like media.q-company.example; once you've set up that CNAME, you at Company Q add a script tag to your website for `media.q-company.example` and now SAAS A is able to track all the users on your site. This indirection allows for effectively infinite cat-and-mouse on the part of you the owner of the Q Company vs the blocklists that the public assemble.

To get around this CNAME cloaking problem, the software powering extensions like uBlock Origin need to be able to see not only the destination domain of requests by browsers, but the underlying IP addresses of those domains as well. This commit makes that behavior possible, or at least is related to making that code work better.