Announcement

Collapse
No announcement yet.

Debian Dropping A Number Of Old Linux Drivers Is Angering Vintage Hardware Users

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • #81
    Originally posted by DMJC View Post
    I strongly disagree. I have a working Commodore 64 here that hasn't required maintenance in its entire lifespan. Still boots/runs without issue. Computers can last decades with appropriate care. Sometimes capacitors/voltage regulators need replacement but good hardware lasts for decades.
    Commodore 64 is a cool running system. Even so where I live there is no functional Commodore 64 left as the PCB split. So even the Commodore 64 has it mechanical limits where I live has brought out those mechanical limits way sooner..

    The machines that take the UMS drivers debian has just removed is not as durable hardware as the Commodore 64. Yes some computers can last decades dependably others cannot. 20 year old PC are coming right up on their mechanical limits. This is like silicon migration failures where silicon chips die, PCB failures.....

    Originally posted by DMJC View Post
    Regarding old hardware support, at a minimum they shouldn't be getting rid of anything that doesn't virtualise properly. QEMU/KVM are good products but they're not good enough yet. There's no Voodoo1/2/3 Emulation for QEMU/KVM yet so it's not possible to GLIDE on Windows 95/98 with accelerated 3D graphics in a Virtual Machine with any sort of useful performance,
    Really for Voodoo 1/2/3 emulation is mostly not required glide to opengl wrappers exist for windows 95/98 people did not only buy Voodoo hardware in that time frame but people want to use windows 3dfx drivers in emulation for other reasons and the other reason being direct x support.

    bochs does have Voodoo 1 and 2 emulation yes can run Windows 95/98

    dosbox-x also has voodoo support and is working in Windows 95/98 and ME support..

    https://github.com/kjliew/qemu-3dfx

    Yes Voodoo emulation is not mainline qemu but this third party well and truly supporting Windows 95/98/ME in qemu/kvm with opengl 2.0 passthough and Voodoo 1/2/3 emulation on modern AMD and intel based graphics.

    This is more a man hours problem past this point there are multi different projects that are really close to nuking requirement of having a real voodoo 1/2/3 hardware for anything.

    Originally posted by DMJC View Post
    hell the state of regular 3D performance in Virtual Machines is pretty crap too (no 3D acceleration driver for DirectX/OpenGL on Windows 95/98 under KVM/QEMU either).
    Yes the mainline versions it still no with qemu/kvm. But outside the mainline versions this is not the case. The only area that need work under Windows 95/98 is Direct X with the qemu-3dfx yes the 3dfx/voodoo direct x driver for windows 95/98 does work on top of qemu-3dfx.

    Wine is used to run a lot Windows 95/98 programs. This is why I know about a lot of the glide wrappers because they are commonly installed in wine to use modern graphics drivers yes qemu-3dfx uses glide to get to modern graphical output processing.

    Originally posted by DMJC View Post
    I know that's slightly off topic since we're talking X11, but there are always funky use cases for old hardware that regular people will not think of. I agree that a Legacy fork of Debian might be required to support these things but nothing should be removed until there's a firm plan in place for a transition.
    There is a basic problem to maintain a legacy fork you still need the legacy hardware to test with to make sure you are not just release something that can never work so causing people lot of headaches debugging something that they don't know if it the software or the hardware.

    Originally posted by DMJC View Post
    IThe biggest strength of Linux is the support for things that properietary software refuses to handle anymore. The entire reason I switched to linux 20+ years ago forever was because of Microsoft Sidewinder Force Feedback support being dropped in the Windows 98SE to XP move. Early XP didn't support a $200 product I bought from Microsoft just a year before XP came out. Linux supported it, and even though the Linux support caused a kernel panic when using it, a kernel dev and I fixed the problem and that joystick has been stable on Linux for 20 years. This is part of the magic that makes Linux users fanatical about Linux. It's a key differentiator to Windows and proprietary software.
    I guess the Linux kernel developer asked you to test a patch and you tested the patch so the Linux kernel developer knew what he changed worked. Remember in the current case with the UMS drivers there has been no one with that old hardware putting up their time to test that any changes made to the UMS drivers work so developer makes change so driver still builds and maintainer builds driver while totally crossing fingers that it works. This is why support is being dropped.

    Comment


    • #82
      Originally posted by oiaohm
      https://www.phoronix.com/forums/foru...96#post1174496

      There is a basic problem to maintain a legacy fork you still need the legacy hardware to test with to make sure you are not just release something that can never work so causing people lot of headaches debugging something that they don't know if it the software or the hardware.
      I would agree...if people were trying to maintain and update that code.

      What if people just want to run the code for whatever reason? Nostalgia? It does what they need? Whatever? Most people with Commodore 64 hardware and software do just that, though maybe a few still program on it because they like/know it well.

      If the old code will do what they want/need, why not simply include it in a "Tombstone" edition?

      In other words, an edition or release that says something like, "Here's the code. Have fun. No updates. No warranty. No fixes. Don't call us or whine all over the Interwebs like a L00zer if it breaks your stuff." <<-- Most people might be satisified with that.

      One of the major fears I think people have when they read that XYZ distro is dropping support for [blank] is a reasonable fear that the code for [blank] will simply disappear into random bits on the Interwebs, no longer to be found nor will an a suitable OS be available to run it.
      Last edited by NotMine999; 25 April 2020, 05:32 PM.

      Comment


      • #83
        Originally posted by NotMine999 View Post
        In other words, an edition or release that says something like, "Here's the code. Have fun. No updates. No warranty. No fixes. Don't call us or whine all over the Interwebs like a L00zer if it breaks your stuff." <<-- Most people might be satisified with that.
        Debian does make Tombstone editions a lot of them https://snapshot.debian.org/ yes almost daily is a new forever never changing debian edition made.

        Originally posted by NotMine999 View Post
        One of the major fears I think people have when they read that XYZ distro is dropping support for [blank] is a reasonable fear that the code for [blank] will simply disappear into random bits on the Interwebs, no longer to be found nor will an a suitable OS be available to run it.
        Yes that might be a fear but that fear is not justified with Debian due to its snapshot system.

        The big complain with to debian those are doing wiith the dropping UMS drivers is that they will not be able to use modern firefox and other things. Lets get real here modern applications like firefox and chrome require more ram to work well than a 20 year computer has. Modern applications will be requiring more GPU acceleration than 20 year old GPU can provide to work well.

        Only thing I can think of is they have been scrapping computers they have brought 20+ year old GPUs and put them into 15 year or younger computers because they still work and are failing on people because they are technically past the end of life for those cards. This is the only way you will have enough ram to run firefox well. Remember a lot of these second hand machines are given to people with basically no warranty so as long as they run when you give them out you are done right??

        Of course they don't want the poor with second rate security from using old software either. So no updates play does not work when you want security. Reason why UMS drivers are being got rid of is they are a security problem.

        Sometimes be you hated or not you have to make the hard choices. Its not wise for Debian to keep on providing UMS drivers in the updating releases that are not being maintained properly and are forcing the kernel to have a security weakness. Now if parties want to keep on using that hardware someone with that hardware has to step up to be testing it and someone need to find a developer willing to help out to convert it to KMS.

        Whining about Linux Distributions maintainers dropping stuff like this is being selfish. Distribution maintainers technically should have dropped the UMS stuff 2 years ago but they waited at least a year to see if anyone with the hardware would step up and do the need testing. This is why I believe people have no right to be upset about this. This is not a Microsoft like style change where they change something without notice and you have to live with it. Debian and other Linux distributions have given their end users notice and its been ignored by the media and by those using it.

        Basically the false belief that parties like Debian will maintain current hardware support for current software for really old hardware with those using the hardware having a requirement to-do nothing to maintain those drivers has to die. There is a requirement for those wanting to run really old hardware with current software to step up and volunteer some time to testing and in some cases being the maintainers of those drivers.

        The arguement we don't have the time or the means to-do that really don't cut it because why should anyone who is not getting any belief from that old hardware be putting their time in.

        Comment


        • #84
          Originally posted by elvis View Post
          Perhaps there could be a middle ground where this code is moved out of the main Linux Git project and into some side project? Tools like DKMS and others exist to make compiling, inserting and maintaining kernel modules in parallel with a modern running system less cumbersome.
          I developed a system called Driver on Demand in 2001 which could have been used for this purpose. Unfortunately, despite having a mention on many major news sites including Slashdot, O'reilly's and Linux Format magazine (the hardcopy, which I only found out about years later when a mate was flipping through them), I gave up after the community compared it to copying windows and basically called me an idiot. At times, the community is their own worst enemy. The community has gotten a bit better though..

          Originally posted by elvis View Post
          The nature of open source is accessibility. Writing driver-level, kernel-level code is not within the realm of all users, and sometimes money is hard to come by too. "Pay up or shut up" isn't a valid motto for the open source community.
          Yes, but if people don't want to program, or perform testing (which takes a LOT of work, and doesn't neccessarily require extensive program knowledge to assist with), financial incentive is required to bring it to production standard, or someone needs to want to do it (nobody does though). Including untested code increases the maintenance burden. Even if it's not official supported, people will still complain when they don't work properly and expect results. Stripping out unmaintained code is better than keeping it in, and pouring resources into telling users it's not supported (and dealing with people complaining). The Linux kernel has grown so big, that they often come across drivers which hadn't been maintained (and didn't work for ages).

          Originally posted by elvis View Post
          Keeping old hardware working is about more than performance. I have old gear lying around because of obscure data transfer ports that don't (and can't) exist on modern platforms or Raspberry Pis.
          Such as? A lot of old data transfer ports either have adapters, or in many cases, are fairly simple in design.

          Originally posted by elvis View Post
          I can certainly run older Linux distributions on said hardware, but that also brings with it challenges over time (security, compatibility with other tools and hardware, etc). I'm already seeing that where I'd love to run F2FS on a very old embedded system, but am struggling to find a kernel that will support both it and the limitations of the hardware. And that's not even "very old" kit. (Acknowledging that this particular example is merely a want and not a need).
          Which hardware?


          Comment


          • #85
            Originally posted by oiaohm View Post
            The arguement we don't have the time or the means to-do that really don't cut it because why should anyone who is not getting any belief from that old hardware be putting their time in.
            Exactly this.. People often say they don't know how, but when you ask them if they'd be willing to run extensive testing of the hardware to help iron out the bugs to a production standard, they also aren't interested in doing that either. Even testing the hardware takes a HUGE amount of time in some cases to do it properly.

            Furthermore, these people aren't often willing to learn to do it either (and learning is within the realm of almost everyone).

            Comment


            • #86
              Originally posted by TemplarGR View Post

              LOL. Yeah i am an evil fascist who wants to force people still on early Pentium 90mhz PCs to upgrade to something more recent if they want to use the latest year 2020 Debian version...
              So you claim others are retarded but you recommend them insecure software and you seem to think this is about Pentium PCs @90MHz? That's not clever either.

              R128 lived in the Pentium 4 era and a bare minimum i386 distro will accomplish plenty of tasks for people who don't have anything else. It's 2020 and we still have people dying of hunger, so any assumption that everyone can afford a cheap second-hand PC to learn is nothing but a wrong assumption.

              The main problem is not that the Debian team decided to pull the plug on those drivers. The main issue is that we don't have maintainers for those drivers (R128 being a popular one) and instead of trying to find maintainers "we" crap on those people who still make use of their Pentium 4. That's inconsiderate to say the least. It's only a matter of attitude to find solutions for such things. Bountysource could be used to fund driver and other software updates that enables old PCs to be used for modern tasks. They're not supposed to offer high performance, but simply a reliable working environment for a few more years.

              Comment


              • #87
                Originally posted by kneekoo View Post
                R128 lived in the Pentium 4 era and a bare minimum i386 distro will accomplish plenty of tasks for people who don't have anything else. It's 2020 and we still have people dying of hunger, so any assumption that everyone can afford a cheap second-hand PC to learn is nothing but a wrong assumption.
                That is not the assumption here.

                Originally posted by kneekoo View Post
                The main problem is not that the Debian team decided to pull the plug on those drivers. The main issue is that we don't have maintainers for those drivers (R128 being a popular one) and instead of trying to find maintainers "we" crap on those people who still make use of their Pentium 4.
                Debian choice at moment can be reversed if maintainers step forwards to work with developers.


                Originally posted by kneekoo View Post
                That's inconsiderate to say the least. It's only a matter of attitude to find solutions for such things.
                No you are also being inconsiderate to the developer side.

                Originally posted by kneekoo View Post
                Bountysource could be used to fund driver and other software updates that enables old PCs to be used for modern tasks.
                The problem is not money. X.org developers/build system/CI system does not have a single R128 left let alone any of the other UMS drivers cards. Fedora maintainer fo the UMS drivers was the first to run out of cards. Debian maintainer has now run out of cards.

                Bountysource or the like technically could help to fund to convert UMS drivers to KMS but you still need the stock pile of cards and systems to test with that currently does not exist.

                Originally posted by kneekoo View Post
                They're not supposed to offer high performance, but simply a reliable working environment for a few more years.
                If these cards were still reliable the maintainers and x.org would not be out of cards. Maintainers and x.org in quality control push cards running test suites so when a card comes up on it end of life it fails faster in the development environment than production but remember they did not start of with a single card the start was over 10 cards of each card at x.org and that complete stockpile is gone and was gone 2 years ago. Distribution maintainers did manage to scrap up some more but those are gone as well now.

                Now the big question who was the cards is going to step up to be the maintainer who can test this is the first problem that need to be solved. Remember single instance of these cards will not last long they are old and failing. Next of course is the security problem UMS drivers is getting funding to port to KMS but that not possible if you don't have the test cards.

                Comment


                • #88
                  Originally posted by oiaohm View Post
                  That is not the assumption here.
                  It is the general theme all over the web - people assuming it's very easy to get a cheap second hand, so it's not worth keeping 20 year old hardware alive.

                  Originally posted by oiaohm View Post
                  Debian choice at moment can be reversed if maintainers step forwards to work with developers.
                  True, but maintainers stepping forward and developers searching for maintainers are different things. Debian removed the packages. Is there any news about Debian looking for maintainers? That was my point.

                  Originally posted by oiaohm View Post
                  No you are also being inconsiderate to the developer side.
                  What is inconsiderate of me to the developer side? Expecting them to consider announcing that they need maintainers and/or hardware to keep some drivers alive?

                  Originally posted by oiaohm View Post
                  The problem is not money. X.org developers/build system/CI system does not have a single R128 left let alone any of the other UMS drivers cards. Fedora maintainer fo the UMS drivers was the first to run out of cards. Debian maintainer has now run out of cards.
                  Ok, but is there any announcement that the X.org and Debian people need a few cards to be able to keep their drivers alive? I bet there are enough "relics" with working cards just sitting around that could end up in the right hands to keep this going for a while longer, should the developers and maintainers ask people for donations.

                  Originally posted by oiaohm View Post
                  Bountysource or the like technically could help to fund to convert UMS drivers to KMS but you still need the stock pile of cards and systems to test with that currently does not exist.
                  Converting the drivers would be great!

                  Originally posted by oiaohm View Post
                  If these cards were still reliable the maintainers and x.org would not be out of cards.
                  Regular people don't run extensive testing that shorten the lives of their cards' electronics like the developers use them, so asking for hardware donations could help. I hope someone will take this into consideration. Maybe not for all the cards, but R128 was especially popular and it's worth asking the community.

                  Now of course that when it comes to numbers, it's virtually impossible to get a real figure of people still using some graphics card. I think we can safely assume that most users don't pay attention to such discussions and I wouldn't be surprised if the majority of them use some old Windows for that matter. So this is a tricky matter with a lot of question marks. But asking the community is still worth it if the hardware itself is the real problem.

                  Comment


                  • #89
                    Originally posted by kneekoo View Post
                    Ok, but is there any announcement that the X.org and Debian people need a few cards to be able to keep their drivers alive? I bet there are enough "relics" with working cards just sitting around that could end up in the right hands to keep this going for a while longer, should the developers and maintainers ask people for donations.
                    The bad news here is history of these events say don't waste time on a nice request. Fastest way to get the hardware is threaten to remove the drivers you don't have the hardware to.
                    1. If no one complains no one is using the hardware.
                    2. If someone complains ask them for hardware to fix the lack of hardware problem.

                    https://wiki.debian.org/Hardware/Wanted
                    As you can see on the debian hardware wanted and other sites like it those groups making computer for the poor did not ear mark any hardware for development of the OS they were using. By the way a lot of those making computers for the poor a developer attempts to get hardware from them they get sorry you are not poor you are not entitled to a free computer/hardware and you cannot even have it if you pay for it so supply from this source is basically nil.

                    X.org at the X11 conference did ask nicely 2 years ago and got nothing. They also announced then that all long term supported drivers would have to come KMS. Distributions are now moving on to brutal acquirement method since nice method failed.

                    Originally posted by kneekoo View Post
                    Regular people don't run extensive testing that shorten the lives of their cards' electronics like the developers use them, so asking for hardware donations could help. I hope someone will take this into consideration. Maybe not for all the cards, but R128 was especially popular and it's worth asking the community.
                    When you think about this closely most of the supply should not have been coming from regular people but from those who are doing refurbishment systems should have been ear marking a percentage to go to development. Of course the arguement doing refurbishmentis they did not have the resources to put into this so now hell breaks lose as supplies run out and demand for particular drivers is not in face.

                    Originally posted by kneekoo View Post
                    Now of course that when it comes to numbers, it's virtually impossible to get a real figure of people still using some graphics card. I think we can safely assume that most users don't pay attention to such discussions and I wouldn't be surprised if the majority of them use some old Windows for that matter. So this is a tricky matter with a lot of question marks. But asking the community is still worth it if the hardware itself is the real problem.
                    The best way to ask community to get somewhere near real number is threaten to remove the item or intentionally break it and see who notices of course no one notices no one is using it is the theory. Also most users don't pay attention to such discussions and most users you ask them what video card is in their computer they have no clue as well so you cannot simple survey your way out this problem. Horrible as it sounds best way to survey is break it next best is threaten to break it.

                    Also think about it if those doing refurbishment was offering X types of hardware up to development this would be direct evidence that that hardware was still in use by someone.

                    Basically the system is broken. No point blaming the Debian people for doing what is required in the current system to get results.

                    Comment


                    • #90
                      Originally posted by oiaohm View Post
                      Also most users don't pay attention to such discussions and most users you ask them what video card is in their computer they have no clue as well so you cannot simple survey your way out this problem. Horrible as it sounds best way to survey is break it next best is threaten to break it.
                      Touche! And there's an abnormal number of developers who don't seem to make the distinction between running 32-bit software on a RAM-limited 64-bit machine compared to a 32-bit only machine, citing "the developers no longer have 32-bit machines for testing" when trying to justify the drop of 32-bit support.

                      Originally posted by oiaohm View Post
                      No point blaming the Debian people for doing what is required in the current system to get results.
                      I didn't blame them. I was curious if they considered turning to the community for maintainers or hardware. Apparently they did it, sort of, but I never heard of it. There are better ways to do it, apart from removing the drivers, and that's with news items to address this specific issue.

                      Now talking about the Debian website, it's so big and inefficiently presented/structured that it's hard to get a particular image, depending on the use case. But this is a deep rabbit hole that's not worth getting into in this thread. My point is, I had no idea there's a wanted hardware page - which doesn't mention these graphics cards anyway, to be fair. The Debian website has no news item entirely dedicated to such matters, that the community can read and share. Communication can clearly be improved. What would be the results, we don't know because it hasn't been done yet. But I think a few news websites would share the news and that should increase the chance of a good outcome.

                      How big was that X11 conference? Did they publish anything on the topic? Their most recent item on their Press page dates back to 2015. Did they at least reach out to the high profile distros, open source/free software organizations (FSF, LF, OSI, OS, etc) and specialized mass-media to spread the news? These are rhetorical questions, of course. If they did, then I missed the buzz in the community. But I suspect they haven't, because I usually catch this kind of news.

                      This is an old problem that most free software projects have - suboptimal communication. The community as a whole should come up with a better way to handle communication for highly important matters, in both directions, towards finding solutions.

                      Comment

                      Working...
                      X