For context, he's saying that to promote his own commercial linux security patches. Has been havig a disagreement with mainline people for a long time. I wish it was easier for people like him to just fork linux (this is one reason why microkernels are better)
I mean, if the mainline kernel really did have a substantial security regression and his variant would have prevented it, that would be an interesting angle to explore. Although, the bias is a good reason to question whether the impact is actually that bad; could anyone more qualified comment on whether the regression was real and meaningful?
I don't agree with that except linux is FOSS and he took grsec commercial only a few years ago, so that really isn't relevant. I mean, it's easier to find vulns and critique than to be part of the team responsible for finding/fixing it in mainline. None of us pay for linux so complaining about someone not finding it faster makes no sense. It would be nice if he contributed his solutions to Linux. Trovalds could have also made commercial Linux for the same reasons as grsec.
Breaking it down into smaller parts with stable interfaces would allow a fork to only touch what it needs to while being able to take the rest as-is. Of course, that would also have significant costs with it; there's a reason Linux doesn't have a stable ABI or API for modules, nor support out of tree changes.
Readit News
grsecurity?[1]
[1]: https://web.archive.org/web/20200201055409/https://en.wikipe...
Why did you use archive.org to get to wikipedia??