Announcement

Collapse
No announcement yet.

AMD To Support Coreboot On All Future CPUs

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

  • #16
    Also have a Gigabyte board here. The bios is full of bugs (or lacks chipset specific bugfixes) that never got fixed. One of these bugs makes the thing unable to boot from usb. Unfortunately my board isn't listed as supported on the coreboot website otherwise I would have flashed it a long time ago.

    Comment


    • #17
      Originally posted by curaga View Post
      @89c51

      Much faster boot, as said above
      anything funkier ????


      i am not saying this is not important but i am not the kind of guy that turns its computer off or reboots

      Comment


      • #18
        Originally posted by 89c51 View Post
        anything funkier ????
        One can build small apps (e.g. for diagnostics or backup/restore) or more featureful bootloaders into the firmware image without changing any of the core hardware setup code. The payload can basically be anything; it doesn't have to be Linux (which is why they changed the name from "LinuxBIOS"). It's a lot like building a custom kernel; most people never need to do it, but it's nice to have the option.

        Comment


        • #19
          Originally posted by droidhacker View Post
          The chipset is the chipset. The difference is the extra crap that gets "bolted on" to the chipset. Having the chipset supported by the manufacturer means that the community can focus more on the rest of the crap than in reverse engineering the chipset.
          Oh really?

          http://www.coreboot.org/Supported_Chipsets_and_Devices

          http://www.coreboot.org/Supported_Motherboards

          Sure looks like most major chipsets are already supported, yet the list of motherboards is a tiny fraction of boards using said chipsets.

          For example, the 780/785 is a supported chipset. I only see two consumer boards even mentioned as partially supported.

          There is far more than simply getting chipset support.

          At this point though, you start to get the advantage of having all or most of the extra crap being pulled into the chipset. At this point, you can pretty much just expect it to work, except for the few components that are added outside of the chipset.
          Except this isn't how it works. This isn't modprobing some device drivers into a kernel. Write something to a bad memory address due to a customization, motherboard doesn't boot, dead in the water. If you don't have a removable bios chip, your board is essentially trashed.

          The mainboard manufacturers will eventually be forced to produce superior generic boards.
          After you figure out how that is going to happen, go ahead and tackle global climate change, world food production shortages, middle east peace, and peak oil.

          Comment


          • #20
            "Support" is Coreboot jargon for "OK, PCI*16 doesn't work, but we can now run assembly code so who needs the GPU anyway?!" ;-)

            Comment


            • #21
              I had always used Intel chipsets in the past. They've converted me.

              Definitely going with llano now for my next rig! Open source BIOS and graphics drivers! Now if only we could get OpenCL via Gallium3D.

              Comment


              • #22
                Fascinating, no hardcoded BIOS?.. I'm new to this, but wouldn't that be a little bit risky, since BIOS is usually the last fallback and thus has to work right under nearly any condition?

                Comment


                • #23
                  Originally posted by GreatEmerald View Post
                  Fascinating, no hardcoded BIOS?.. I'm new to this, but wouldn't that be a little bit risky, since BIOS is usually the last fallback and thus has to work right under nearly any condition?
                  I'm not sure how a stock BIOS/UEFI image is any more "hardcoded" than coreboot is. In either case, they can be overwritten with a new image. The difference is that with coreboot, you can easily build your own image rather than relying solely on the board vendor (or pseudonymous folks on web forums who are either really handy with a hex editor or somehow got the Award/AMI/Phoenix/etc. OEM tools). Generally, unless you're porting coreboot to new hardware, you would just be changing the payload(s) and not the actual coreboot part.

                  Also, even if you totally hose the firmware, most boards have some way to recover. It might take some stuff found on the bench of the average electronics geek rather than the bench of the average PC geek, but nothing too exotic.

                  Comment


                  • #24
                    Originally posted by Kano View Post
                    Coreboot is really interesting, maybe they should tell the board vendors not to solder the bios chip directly for better recovery. If somebody has a spare fusion board let me know...
                    Eh, I'd wait for the Llano cchips as mentioned, they are the Fuzion A series, quad core CPU and 400 Shader GPU based on the HD5670, so it should be plenty of grunt to use as a main box even with the R600G drivers.

                    Now if only they would put a CrystalHD chip onboard as well for video playback till they release something around their next implementation of UVD which will have the DRM pushed aside for us to have an oss driver...

                    Comment


                    • #25
                      ...if only "all future CPUs" also meant "all future motherboards".

                      Unfortunately, they can only support their CPUs and GPUs, since they have no control over what others decide to put on their motherboards. I don't blame AMD, it's not their fault, but it sure is frustrating.

                      Comment


                      • #26
                        Originally posted by 89c51 View Post
                        can someone outline potential benefits to this????
                        As others have mentioned, it's a hell of a lot faster. Although for me the true advantages are:

                        1) Choice of "payload" - the payload is the thing that's force-loaded into memory on boot. With CoreBoot you can load a BIOS equivalent (called SeaBIOS) to load up legacy systems that require a BIOS (namely Windows). But there are many other things that can be loaded instead - hardware detection tools that run straight from coreboot (diagnose faulty hardware without loading an OS, for example), as well as other boot mechanisms (instant network booting for diskless systems, etc). You can also embed telnet/SSH to get direct hardware access below the OS (remote access to a terminal even when the OS has crashed, which is a godsend for sysadmins, and much cheaper than "iLO" Lights Out style systems sold by big vendors).

                        2) It's open source, which means the community can fix problems, and not just the motherboard manufacturer. It's 100% customisable for anyone to enhance as they see fit. Coreboot's original design was created by people who maintained clusters, and were sick and tired of taking a whole week to reboot a thousand machines (not to mention needing to push "F1" on each system that wasn't configured to boot without a keyboard). Coreboot let them boot their clusters in hours instead of days, and saved a hell of a lot of time. Speaking as a sysadmin (and one who does deal with remote clusters), this is great news.

                        Originally posted by _txf_ View Post
                        Also in terms of consumers how is this better than uefi? After all is coreboot still a BIOS with all the inherent limitations like long hardware initialisation times and antiquated hardware support?
                        UEFI still isn't 100% open source, and you're still at the mercy of the hardware vendor's implementation.

                        Worse, Intel have started considering to take IBM's approach to selling high-end systems - you buy a high end CPU, but "lease" time on the CPU. So you pay a fee per year to access 10% of the CPU (with the other 90% taken up by "sleep" statements). If you want a faster CPU, you pay more and get another 10% or so.

                        IBM have been doing this for years on mainframe, system p and system i, but now Intel (and HP) are considering it on both Itanium and Xeon. And it'll all be handled by the UEFI layer that the end user can't get access to.

                        Coreboot is as important to the hardware market as Linux is to the software market. It means people finally get control over the systems they purchased, and can do what they like regardless of what a vendor stipulates.

                        Again, speaking as a sysadmin, for me that's a big deal. A lot of my working day is spent arguing with vendors about what they've promised their systems can deliver and what they actually do deliver. To be handed back some of that control and be allowed to configure systems to do what the businesses I support actually want (rather than what the vendor thinks they want) is a good thing.

                        It was only a matter of time until one of the big motherboard vendors adopted Coreboot. I hope that we see some of the consumer guys start to offer the same. The benefits are there for everyone (not just Linux users either - there's huge benefit in it for anyone regardless of their running OS).

                        Comment


                        • #27
                          Originally posted by elvis View Post
                          It was only a matter of time until one of the big motherboard vendors adopted Coreboot. I hope that we see some of the consumer guys start to offer the same. The benefits are there for everyone (not just Linux users either - there's huge benefit in it for anyone regardless of their running OS).
                          What "big motherboard vendor" has adopted coreboot? Nobody I know of to date. If you're referring to AMD, AMD doesn't make consumer nor server motherboards.

                          Coreboot (and LinuxBios) has been around since '99. In 12 years they haven't had anything remotely close to a usable release. Coreboot (LinuxBios) v1 hung around for a while, then they started the v2 line which went nowhere, then v3 came out and now that's deprecated as a "development branch" in lieu of v4.

                          When you have notes like this:

                          "This board is very pick about what ROM chips it can use and it will fry any chip that doesn't work."

                          then your project is no where near release for anything close to enthusiast, let alone mainstream.

                          Comment


                          • #28
                            Originally posted by locovaca View Post
                            What "big motherboard vendor" has adopted coreboot? Nobody I know of to date.
                            ISTR some mentions that Tyan was offering it to direct customers on their supported AMD boards. Whether that meets your definition of "big" and/or "adopted" is another matter.

                            Comment


                            • #29
                              Originally posted by Ex-Cyber View Post
                              ISTR some mentions that Tyan was offering it to direct customers on their supported AMD boards. Whether that meets your definition of "big" and/or "adopted" is another matter.
                              From wikipedia:

                              However, Tyan seems to have dropped support of coreboot.
                              Wikipedia also mentions Gigabyte and MSI and google searches only point to CoreBoot/LinuxBIOS offering support for boards from those manufacturers, not the other way round. Those articles are from 2006/2007, and also proclaim that "LinuxBIOS is about to go mainstream" as well.

                              From http://www.linux.com/archive/articles/58781

                              After seven years of work, the LinuxBIOS project is on the brink of making a free BIOS a standard option for computers. Serious obstacles remain, including a lack of resources and resistance from some proprietary chipset manufacturers and OEMs, but the advantages of LinuxBIOS indicate that its availability to the average computer buyer may be only months away.
                              However, AMD is expected to offer LinuxBIOS as an alternative to vendors in its next generation of high-end boards, and Minnich is willing to say that, with the help of the Free Software Foundation, he hopes that at least one vendor will support LinuxBIOS in workstation machines within the next year, and possibly on a laptop, he hints.
                              This doesn't sound familiar at all...

                              Comment


                              • #30
                                What would convince mfrs to switch?

                                I wonder what the per-unit savings (if any) would be for a motherboard manufacturer to switch from Phoenix or ... whoever else is left in the bios business.
                                I'm wondering whether the decision to stay with traditional BIOS is financial, or if they are just risk averse, since they have something that works for the majority of their customers.
                                Like a lot of things, the benefits would be cumulative, as motherboard advances are mostly incremental, other than the occasional major CPU/bus changes.

                                Comment

                                Working...
                                X