site_archiver@lists.apple.com Delivered-To: darwin-kernel@lists.apple.com Dkim-signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=20230601; t=1755294766; x=1755899566; darn=lists.apple.com; h=content-transfer-encoding:cc:to:subject:message-id:date:from :in-reply-to:references:mime-version:from:to:cc:subject:date :message-id:reply-to; bh=CRyrmzaYjzFwq/eubMTek4p6suk7EDDswN65je8KA1w=; b=PtK9pKGM9RYx56xr23BBo5+GIOu7qh/gb/xe/iWiOQigTLbaODM2NVckMCZAQIwBER SlwhEJKTDKoRZSnhEVR7W4mhKL/B/YxYBmq4b1xWPWVQJ4UfYVaSGwC4HYYUKLPYaCVk Sj7u8Prgy4/Y0ar4YRRI+ZlsQcncr3/wPol86oEuOTJPa+bPawwKZlTv3OdP7o/IMlC9 3n31gEkPkaTKmCdpCEkQtCjEKVpnCwItpTRQEgUigdUe/D3DxA3zzodUKcpRIUq36CoM ynK0rEwh1A3otoGhzh9c5OU2ANMKR1BpVKLG1sgD+XLlUqML29scPSqVuBMhnvn8y2yB kZuw== On Fri, Aug 15, 2025 at 2:39 PM Ken Hornstein <kenh@cmf.nrl.navy.mil> wrote:
So, it looks like all I can do is patch the xnu kernel sources and then build a custom kernel from them?
I mean ... no? I gave you an example of an existing network filesystem today that works on MacOS X that is a kext, so it's possible.
Yes, I understood your comment. However, my NFSv4.1/4.2 client is kernel based (it is in FreeBSD), so I have no interest in something that jumps up into userland. Sorry, that I didn't explain that. (My experience with fuse on FreeBSD is that it works but is very slow. The usefulness of NFSv4.1/4.2 is that it can finally perform fairly well in a non-LAN environment, due to various improvements, mainly related to more effective use of delegations.) I am assuming (maybe incorrectly) that the extant NFSv4.0 client is in kernel space and that I would be patching that (maybe by replacing large chunks with the code I currently have in FreeBSD). 15years ago Apple engineering did do quite an extensive rewrite of the FreeBSD code for their client, but I am pretty sure it remains in the kernel, although I have not looked at the xnu sources in 15years. I will be looking at current xnu sources before I decide to go further. (Like finding an old used Macbook that I can afford.)
Digging back into my memories a bit more ... I believe the sticking point was actually that the default firewall configuration really really wants a network process to run from userspace so it can apply security contrls and Auristor wanted to deploy in such environments. I might be misremembering things.
(I remember setting up the toolchain to build a kernel from sources was fun even 20years ago, but I do see a few postings from people who have succeeded in doing it.)
I know Apple Silicon is locked down more when it comes to the kernel so if you are going to go that route you might have better luck with an Intel system.
Yea. I'm also cheap, so finding an old x86-64 Macbook Pro that can at least build an out-of-date xnu may be my only option. (Someone did post that they built a kernel for an M chip, but needed some of the Xcode stuff to do it.) Thanks for the comments, rick
--Ken
_______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-kernel mailing list (Darwin-kernel@lists.apple.com) Help/Unsubscribe/Update your Subscription: https://lists.apple.com/mailman/options/darwin-kernel/site_archiver%40lists.... This email sent to site_archiver@lists.apple.com