Lack of maintainers with write access and time to review, especially the core (openwrt) package is more "picky". So PR from none "core" members tend to take a long time or they need to draw attention to them via IRC.
There was a discussion to let PR go through after a minimal test + wait time, but seems all those proposals did go nowhere, so new PR's still need to "campaign".
Many reasons; maybe the pull is not up to the project standards and needs changes, people are busy (this project is entirely run by volunteers, for free), or a technical reason.
Please see point number 8 on this page
master has improved dramatically over the past couple years. Part if this is due to high standards for code review and acceptance.
Not to forget, some core developers doesn't want to use Github despite it being agreed on as a viable path for submission. Having said that, as there's no "official" response I don't see how you'd attract more people since many patches more or less bitrots which in turn makes people lose interest. The overall review load would also most likely be a lot lower if the toolchain and its options were documented.