FreeBSD is deprecating 32-bit platforms over the next couple of major releases. We anticipate FreeBSD 15.0 will not include the armv6, i386, and powerpc platforms, and FreeBSD 16.0 will not include armv7. Support for executing 32-bit binaries on 64-bit kernels will be retained through at least the lifetime of the stable/16 branch if not longer. (There is currently no plan to remove support for 32-bit binaries on 64-bit kernels.)
↫ John Baldwin on freebsd-announce
I don’t think this is too egregious of a timeline, but there’s always someone with some weird edge case that gets bit hard by deprecations like these.
“there’s always someone with some weird edge case”
This is an odd phrasing in this instance as the “weird edge case” is simply having 32 bit hardware already in use that continues to perform its function ( whatever that is ). I mean, it is getting a little unlikely that people are trying to be productive on 32 bit desktops these days ( an old Pentium perhaps ) but there are plenty of other things that these machines might be doing.
Think of all the things that people use SBCs for these days: firewalls / routers, DNS servers, print servers, audio servers, email servers, and web servers as examples. In the past, these took “real” computers and while these “real” computers are now viewed as e-Waste, many of them are still up to the task or still in use. The same is true of industrial control systems, point-of -sales systems and the like. I would not reach for a 32 bit computer to run a database or host an application server for a line-of-business app today but I see no reason why one that was installed years ago cannot keep faithfully serving up an application if it continues to meet the need. I mean, even DOS is use for many of these things to this day.
The point is that the actual uses are do not have to be “edge cases”. They may be very mainstream use cases that have simply been in place a while. If it ain’t broke, don’t fix it. That said, we all know that an operating system that has not been updated is in fact broken even if only for the security implications.
If I have a Pentium based email or ftp server that is happily chugging along on FreeBSD today ( fully configured and working in production ), why not leave it in place to do its thing? Of course, since both of these services are network facing, it is important that I keep the software up to date. So continued support of my hardware by modern ( and more importantly maintained ) software is critical.
Now, I am not going to get angry and cry fowl because people that give me software for free don’t want to maintain support for my old hardware. It makes sense. And there are fewer and fewer of these old systems doing useful work every day.
Did you know that the Intel 386 ( their first 32 bit processor ) was sold up until 2007. That is a long time ago now but perhaps not as long ago as people think.
That does not make the use cases themselves edge or fringe though. And so, it is in some ways a shame. The 32 bit “servers” will become eWaste. Things like industrial controls will simply not get upgraded and will become increasingly insecure. Thankfully, many of those kinds of systems will not be networked anyway. One benefit of Open Source at least is that the software remains available even if unmaintained. When that old hardware dies, you may be able to redeploy. It is not a guarantee though because, while you can now replace that ancient hardware with something far superior for cheap, your now ancient version of the OS may not be supported on that hardware. And if you have not been staying up-to-date, you may discover that your ancient application may no longer work on a newer version of the OS.
Even in Open Source though, time marches on.
FreeBSD is a pig to update.
Good thing I decided not to update my OSes anymore.