Define legacy hardware

I am sad to see this thread.
https://lists.freebsd.org/pipermail/freebsd-current/2018-May/069401.html
It was like an RFC with the conclusion already decided.
.
"Are you saying there is only one way forward?"
They appeared on the verge of ramrodding this through before this comment.

Thank goodness Rodney spoke up. Being an original FreeBSD developer, along with Kirk, he really is close to a god in my eyes.
Great that he came back to work on FreeBSD.

I just wonder about the attitude of some of these developers.
Do they realize that FreeBSD is used in third world countries.
Not everyone is working for Netflix and drives a Tesla.


Quote:
"So it boils down to the question of allowing old, obsolete hardware, which
nobody's sure works anyway,
to get in the way of moving forward to more
modern hardware? I'm trying to find value from supporting i386, and a
community willing to do that work and I'm not seeing much of either."
 
Maybe someone should look over his window sill sometimes, go out and maybe travel to places on the other side of the globe. Oh, and open the eyes and maybe leave that filter bubble behind for at least 2 weeks a year.
 
I really wish I was in a position to help. Unfortunately I am a lowly user working on my skills.
It is discouraging for someone working on their first port to see such opposite personal opinions.

Sometimes I wonder if we are better with "group think" or a ringleader like Linus.

Thanks to Steve K. from washington.edu. He really put up a good fight.
Gloves came off around here:
"Yes, I recognize that you're clever enough to purposefully
break the API so that you can thumb your nose at those of
us who have older hardware."
 
I'm also using an old T60 on train rides and as a developer system. Great screen, silent and you may use it as a pillow when needed.

I would hate to get rid of it, only because someone wants to mast..^h massage his ego.
 
I have at least a dozen N270's@2.5W_CPU. No reason to replace them.

What I would like to see is a FreBSD hardware testing lab with supported hardware well defined..
I am willing to donate half my stash.
 
I could tell Rodney knows hardware by this comment:

"Also as the Moore's law curve flattens expect the life of these
older, but not so old, machines to live quiet some time. I
believe we are talking sandy bridge and earlier? If that is
correct Sandy bridge is still a very viable system."
 
One of the reasons I use FreeBSD is because it works on almost all i386 hardware. Some built-by-default Linux distros these days are are even going so far as to require not-often-used optional instructions, breaking compatibility with some x86 embedded boards.
 
Tryng to figure out what this is about... AFAIK drm2 is state-of-the-art for intel cpu's with integrated graphics[1], and this is -in my understanding- current hw. Old hw, in my understanding, is stuff built before 2000 - and I have such stuff running as well.

[1] I am not fully sure what "drm2" means in this context, but what I know as drm2 is what I have compiled into my kernel:
Code:
device          agp
device          drm2
device          i915drm
 
Actually, FreeBSD 6.0 (and consequently every version after it) dropped support for the 80386. So it's nothing "new" to have a discussion about dropping support for old stuff. And I really don't see the problem here. It's a discussion about what to do, which is good. Or would you rather have FreeBSD stagnate even further because some people can't or won't upgrade their hardware? Or leave the code in and let it rot because nobody is willing or able to maintain it? Or leave the code in and potentially blocking further development for new and relevant hardware? Or simply remove it and move on?

https://www.freebsd.org/releases/6.0R/relnotes-i386.html#KERNEL
 
What do you expect when FreeBSD developers are using Macs?

I just wonder about the attitude of some of these developers.
Do they realize that FreeBSD is used in third world countries.
Not everyone is working for Netflix and drives a Tesla.

I already went off on the poor TrueOS people about a reality check:

However, I think it imperative to the success of TrueOS for everyone involved in the project to get reacquainted with the real world of people who use desktops (I mean FreeBSD desktops and TrueOS moguls using them. Daily.),

https://forums.freebsd.org/threads/trueos-anyone-using-it-on-a-laptop.63746/page-6#post-371594
Did they? Doubt it.

Prosecution rests. Now I need to.

I haven't been to bed yet and this really flips my lid considering all my hardware is "legacy" or soon will be. The horror...The horror...
 
It's just a matter of time before the whole i386 architecture will become obsolete. PC-BSD/TrueOS already dropped support for it (although for other reasons) and there will come a time for a discussion whether or not FreeBSD should continue to support it.

I haven't been to bed yet and this really flips my lid considering all my hardware is "legacy" or soon will be. The horror...The horror...
Don't worry. If "legacy" hardware gets old enough it will eventually become "vintage" and get a revival ;)
 
Actually, FreeBSD 6.0 (and consequently every version after it) dropped support for the 80386.

So what?

From what I understand up to now, and if I understand it correctly, and if this following quote indeed is a proper description of the issue, then the issue is NOT about old hardware:
One of the reasons for the deprecation and removal of the drm2 bits is
that they prevent us from automatically loading the drm-next/stable-kmod
kernel modules, since the two collide.
https://lists.freebsd.org/pipermail/freebsd-current/2018-May/069443.html

Instead, the issue is about autoloading; in other words, about magically doing things behind the back of the user.
This indeed is a habit that has become custom in recent years. But it is quite the opposite of defined behaviour, and -at least to me- defined behaviour is imperative.
I do NOT expect a system to automagically do things which I might appear to want happen. I expect a system to do exactly what is stated in the configuration stances, and to fail with error if that does not work.

If you run server installations, this is self-explanatory. The habit of doing things automagically does mostly concern end-user gameboys. So, if we run an OS like FreeBSD, which is primarily targeted as a server OS, in a desktop environment, we should not expect things to happen automagically. People expecting such should probably use Apple stuff or similar - there are proper offers for their demand, and so there is no need to mimick that.
I for my part use FreeBSD on desktop exactly because I do NOT want automagic behaviour - I consider the absence of such as an advantage.
 
For the record, I am a massive fan of "legacy" hardware. Especially the Thinkpad *60 range (I have a T60, Z61 and an X61 ;)).

However, I accept (and almost recommend) an OS to drop hardware it no longer has the resources to maintain properly. I would even go so far as to rather FreeBSD only supporting *one* laptop at a time 100% perfectly than all of them half broken! I am not exactly rich but I would certainly still just go out to the shops and buy that laptop outright ;)

The difference with me (and apologies if this is very much against what is recommended in these forums) is that I am perfectly happy running an older OS on my older hardware. For example, I still run FreeBSD 8.4 on some of my machines because something broke in the ACPI / powerd system on my older Pentium 4 causing it not to power scale down (also being able to use Gnome 2 is a very great bonus).

I even risk going online (with a decently configured firewall) and unlike most package managers, the ports system makes it quite easy to backport some crucial software I need to the older release. So running an old version of FreeBSD isn't at all a pain (unlike most Linuxes or Mac OS X)

Basically, if I am using old hardware, I am unlikely to be risking any security breaches or problems in production. On production machines, they are almost always pretty new and so can run the latest version of FreeBSD. It is kind of a win win :)

But just to reiterate, I am not advocating running FreeBSD 6.3 on a T23 in production but it is still perfectly good on the train or for playing Quake during meetings.
 
The point I was trying to make is that discussions about removal or disabling obsolete things isn't new.

Alright, lets discuss it. :)

I for my part don't see much point in joining the rat-race for the always-newest. In fact, I don't even see much breath-taking improvements over the last years.
 
I would even go so far as to rather FreeBSD only supporting *one* laptop at a time 100% perfectly than all of them half broken! I am not exactly rich but I would certainly still just go out to the shops and buy that laptop outright ;)
That would be a macbook, then.
When my current laptop breaks down I'll check used macbook air offers, knowing FreeBSD will work on it.

Other than that, I also got the impression it is about auto loading of the next drm versions.
 
I for my part use FreeBSD on desktop exactly because I do NOT want automagic behaviour - I consider the absence of such as an advantage.

I don't even want mine to automount a USB drive and watch top like a hawk for what processes are running. I want it to bend to my will, not have it do questionable things on its own things behind my back. (Luckily, all bots love jitte, so the upcoming war with intelligent machines doesn't worry me.)

Like kpedersen, I have copies of my older OS going back at least 13 years. (I have PC-BSD v0.75 at FreeBSD 5.3 IIRC.) If it comes down to it I can use computers without the internet so the OS version becomes almost moot. I keep my X61 updated, but it's my .mp3 player and offline except for that.

All my laptops were made in 2007-2008.
 
I really don't see the problem here. It's a discussion about what to do, which is good. Or would you rather have FreeBSD stagnate even further because some people can't or won't upgrade their hardware?
I agree there are stagnant parts of FreeBSD but as a whole I don't think FreeBSD is stagnent.

Some of this divide seems to be along the professional -versus- home user, Whereas a business should have upgraded by now.
 
Last edited:
Some of this divide seems to be along the professional -versus- home user, Whereas a business should have upgraded by now.

That's how I got both my T61 for $50 each. Business lease returns. Both were in excellent condition with Intel Core2 Duo at 2.0GHz and 2.4GHZ, and 4GB RAM, upgradable to 8GB.

While it may have been time for the businesses to upgrade, it was my opportunity to snatch up 2 perfectly good computers in working condition at a steal that can run FreeBSD and all my other dodads with power to... Serve.
 
What do you expect when FreeBSD developers are using Macs?
Objection.
I get the gist of what you might be saying(mac users are aloof)
But at the end of the day they are working on FreeBSD.
So what if they run MacOSX on their laptop to develop. This point is overblown.
I am not that much of a purist. This is many peoples hobby and they should use whatever platform they enjoy working with.

When I give a worker a job in the shop, I don't tell them what machine to use.
One person might use a CNC mill where I might use conventional.
Neither is wrong as long as the part is to specification.

Now for my painful opinion. I would rather see Arm32 go away then i386.
If we took a user poll I wonder what the numbers would look like.

I didn't post this to cause hate and discontent, but I would like see see i386 users speak up or we will be left in the dust.
 
I can tell from this thread that no one has actually finished reading the entire mailing list thread posted by the OP. :) And everyone is getting their panties in a twist over something that won't actually affect them in the end.

It was a discussion thread. "Here's what we'd like to do, what do y'all think?" Various people raise objections, alternatives are discussed, some mud is slung around, a workable plan is devised that keeps everyone happy with only minor changes required.

And life continues as per normal. "Old" hardware remains supported, without impeding development for "new" hardware.

IOW, the process works. You just need to take a step back and see the complete picture.
 
No, I stopped reading at "So do people use i386 for desktop?"

Objection.
I get the gist of what you might be saying(mac users are aloof)
But at the end of the day they are working on FreeBSD.
So what if they run MacOSX on their laptop to develop. This point is overblown.
I am not that much of a purist. This is many peoples hobby and they should use whatever platform they enjoy working with.
*snip*
I didn't post this to cause hate and discontent, but I would like see see i386 users speak up or we will be left in the dust.

If they're a Developer this is not their hobby, is it? Marco isn't getting paid over at TrueOS? He's a nicer guy than I even thought...

Admittedly I am not a developer nor could I even attain to be. I know my limits. But is FreeBSD not able to run the programs they need? Because I really don't know. Is it there aren't powerful enough machines to do the job? I don't know that either. My most powerful is a W520 with Intel Quad Core i7-2760QM @ 2.40GHz, 8GB RAM (upgradable to 32GB), and NVIDIA Quadro 1000M. I'd rather use the T61 I normally do, it's transcoding video now so I'm on my other one.

If it's good enough for me to use it should be good enough for the people who make it. But I never could run fast, so I told Mr. and Mrs. Jones to to go on without me, I could never keep up.


I do try to promote i386, use a screenshot of my 32bit Sony on the index page of my site and pitch it in my tutorial:
If you've got an old PC or laptop laying around that just can't run the bloat that is Windows, FreeBSD can breathe new life into it and is the perfect solution to it collecting dust.
 
Okay, here goes. Yes, my desktop runs a Pentium 4 HT Northwood (circa. 2003). I have rebuilt it myself, so that makes me more of a hobbyist, than professional computer programmer. Yet, I know a fair amount of programming too, Assembler, COBOL, Fortran, Bash,.... The other machine is a Thinkpad SL510 with Core 2 Duo and Centrino wireless card. I basically use the desktop as a beta tester before I install onto the Thinkpad.

I am an independent paralegal who does a limited amount of contract work with attorneys. So, that makes me a professional unrelated to computer programming development.

I use FreeBSD for the BSD licensing and economic reasons of running my own boutique paralegal services. I do not advertise because that puts me in jeopardy with real attorneys who want to claim I am practicing law without a license. I have a family so can not do any such thing as practice law without a license.

I would really appreciate keeping the 32-bit architecture, it saves me alot of money.
I can custom fit some applications to the needs of a paralegal; so FreeBSD is great!

Edit: If you look closely at the avatar picture - book shelf - those are old law books.
 
How about this for comparison. Since FreeBSD 6.0 Ended support for i386.
I see it like this:
i386 introduced =1985
FreeBSD EOL i386 =2005

That's a twenty year span. My i686 machine, the N270 was released in 2009.....

I did read the entire thread and it does show the sausage being made.
It is a thread about drm2 with a side helping of i386 on the desktop.

Like I said I want to make sure the i686 voice is heard.
I saw several "loose facts" in there and get distressed.

This summer I have 3 All-in-One computers playing looping video for the cheapest digital signage system I could put together.
Guess what they use. Pentium4 521 and FreeBSD.
My friend has an electric bike shop and I am helping him with some primitive advertising. My total cost under $100.
Time will will tell if all three rigs make it thought the summer.

So my comments about people in third world countries was deplorable.
I was just trying to emphasize that not all users experiences are equal.
Some users might control million dollar racks while other users struggle with things as simple as reliable power company.
 
Back
Top