Weekly Edition
Return to the Kernel pageRecent Features
| |||||||||||||
AMD and Coreboot
Coreboot (formerly LinuxBIOS) is a
free BIOS implementation; it offers escape from a long list of woes
stemming from poorly-written BIOS's, but it has always suffered from
limited hardware support. AMD has now announced
support for Coreboot on a new set of processors, and more going forward:
"Finally, AMD is now committed to support coreboot for all future
products on the roadmap starting next with support for the upcoming 'Llano'
APU. AMD has come to realize that coreboot is useful in a myriad of
applications and markets, even beyond what was originally considered.
Consequently, AMD plans to continue building its support of coreboot in
both features and roadmap for the foreseeable future."
(Log in to post comments)
AMD and Coreboot Posted May 9, 2011 14:39 UTC (Mon) by robert_s (subscriber, #42402) [Link]
Oh AMD I wish I needed to buy a new computer.
AMD and Coreboot Posted May 9, 2011 17:22 UTC (Mon) by Velmont (guest, #46433) [Link]
I actually need a new computer, but still I think it'll be Intel. Well, with all the good work they've been doing with wireless drivers, graphic drivers (except the abinomion (?) that is pousebo) and the like. I do hope they'll follow tough.
The work AMD does on making FOSS graphic drivers possible after the mess ATI did is very nice although it seems to be lagging a bit. Especially good with the release of documentation, I've heard they've got one up on Intel for that one.
With this announcement my Intel choice is a little less certain, but still, they are not free software enemies by any stretch and, well, Sandy Bridge is damn fast. :)
But these things definately make AMD get a place in my option list.
AMD and Coreboot Posted May 9, 2011 18:08 UTC (Mon) by tzafrir (subscriber, #11501) [Link]
AMD does not have a wireless chipset. ATI have changed their attitude (provided specs and started working on free drivers) right after they were bought by AMD ("right after", give or take an extra lag for the new management to sink in).
AMD and Coreboot Posted May 9, 2011 18:56 UTC (Mon) by svena (guest, #20177) [Link]
Intel does make documentation available:
http://intellinuxgraphics.org/documentation.html
Or did you mean that the docs aren't as good or complete as the ones for AMD/ATI?
AMD and Coreboot Posted May 9, 2011 19:21 UTC (Mon) by ewan (subscriber, #5533) [Link]
AIUI the main problem for Coreboot isn't processor support, it's motherboard support. In this announcement AMD seem to be saying that they'll support it on a tiny number of 'embedded' boards (for what must be an unusually flexible definition of 'embedded').
You may want your next computer to be an EPOS system, but I'm not sure that this is all that significant for most of us.
AMD and Coreboot Posted May 10, 2011 16:08 UTC (Tue) by pgeorgi (guest, #74838) [Link]
Support for chipsets and CPUs is just as important as mainboard support. And compared to mainboards, these are _hard_ (mainboard support is in the days-to-weeks range, chipset support in the weeks-to-months range).
And AMD didn't just build code specifically for coreboot (like they used to), they made a bridge between coreboot and AGESA and opensourced the latter.
Of course, making support for new mainboards (with fully supported chipset) easier is important, and we work on it. My personal goal is that it takes only 1 full day to get a new mainboard to full windows compliance (which generally includes Linux and other ACPI-compliant OSes, too), but we're not there yet.
AMD and Coreboot Posted May 11, 2011 4:52 UTC (Wed) by brouhaha (subscriber, #1698) [Link]
Have they actually said that they will release their AGESA code as open-source? Historically it has been under NDA, which is useless for Free Software like coreboot.
I've never understood why they are so secretive with the AGESA code, and the "Functional Data Sheets" for all of their processors beyond socket 940. :-(
AMD and Coreboot Posted May 11, 2011 6:18 UTC (Wed) by pgeorgi (guest, #74838) [Link]
Feel free to dive in: http://tracker.coreboot.org/trac/coreboot/browser/trunk/s...
AMD and Coreboot Posted May 11, 2011 10:16 UTC (Wed) by brouhaha (subscriber, #1698) [Link]
Wow, that's great!
Now I wonder if AMD can somehow be pursuaded to make the Functional Data Sheets for Socket AM2, AM3, F, S1, etc. available? They published the Functional Data Sheets for Socket 754, 939, and 940, but the later ones are not publicly available. :-(
AMD and Coreboot Posted May 9, 2011 15:01 UTC (Mon) by tajyrink (subscriber, #2750) [Link]
I know that it's not supposed to be easy, but I wish someone would pour their information on the http://www.coreboot.org/Developer_Manual page. Namely file by file in a mainboard subdirectory, what are the files and foremost who to obtain / deduct the contents for each file. I'd bet some of the tinkering needed is in practice easier than others.
Yes, I'm probably looking for a shortcut / "for dummies" instructions that no-one is interested in writing because of the following "I burned my motherboard" woes. I just have a motherboard very similar to http://tracker.coreboot.org/trac/coreboot/browser/trunk/s... regarding components and I've been wondering how to add support besides cp -r + modifying the obvious identifiers.
AMD and Coreboot Posted Jun 23, 2011 7:01 UTC (Thu) by misiu_mp (guest, #41936) [Link]
There is a copy-paste bug in set_gpio40_gfx(void) in that file. The two if cases are identical.
AMD and Coreboot Posted Jun 23, 2011 7:01 UTC (Thu) by misiu_mp (guest, #41936) [Link]
I mean in mainboard.c
AMD and Coreboot Posted May 9, 2011 15:16 UTC (Mon) by SEJeff (subscriber, #51588) [Link]
Now if only Intel would dump the hideous EFI mess and work with coreboot as well.
AMD and Coreboot Posted May 9, 2011 17:21 UTC (Mon) by zlynx (subscriber, #2285) [Link]
Coreboot has an EFI module so Intel would not even need to dump it.
AMD and Coreboot Posted May 9, 2011 17:35 UTC (Mon) by SEJeff (subscriber, #51588) [Link]
That kind of defeats the purpose though. With that argument, why not layer SeaBios ontop of Coreboot everywhere for legacy bios ontop of Coreboot :)
AMD and Coreboot Posted May 9, 2011 17:38 UTC (Mon) by zlynx (subscriber, #2285) [Link]
Because if Intel did that, hackers or people with special needs could do their own versions of CoreBoot and Intel could ship with an EFI enabled version that regular people can use.
AMD and Coreboot Posted May 9, 2011 17:40 UTC (Mon) by SEJeff (subscriber, #51588) [Link]
I dunno, I don't see "server administrators" as "hackers" really. I see coreboot as a superior technical choice for hpc cluster or this whole "cloud" thing the internet seems to be on. Perhaps I'm biased (I work in HPC), but coreboot is a brilliant technical fit.
AMD and Coreboot Posted May 9, 2011 17:42 UTC (Mon) by zlynx (subscriber, #2285) [Link]
HPC cluster administrators fall into the "special needs" category.
AMD and Coreboot Posted May 9, 2011 17:46 UTC (Mon) by SEJeff (subscriber, #51588) [Link]
I would agree, but regular servers are not special needs. Coreboot fits the server use-case flawlessly. Perhaps we should agree to disagree as this isn't /.
AMD and Coreboot Posted May 9, 2011 17:53 UTC (Mon) by zlynx (subscriber, #2285) [Link]
I'm not even sure what we're disagreeing about.
I am in favor of CoreBoot on all Intel machines.
I am also in favor of using an EFI module on top of CoreBoot so that regular people can make use of BIOS features, boot Windows and Windows Server, etc.
Then, since I am in favor of Intel supporting CoreBoot, of course there will be open CoreBoot configurations and documentation so that anyone who doesn't want an EFI layer is free to reflash with a direct to Linux boot.
Special orders of thousands of cluster servers could be shipped with CoreBoot installed in whatever configuration was requested.
I think that this is what you want too.
AMD and Coreboot Posted May 9, 2011 19:51 UTC (Mon) by paravoid (subscriber, #32869) [Link]
I wonder, what kind of servers do you have on your HPC clusters?
AMD and Coreboot Posted May 10, 2011 1:57 UTC (Tue) by pflugstad (subscriber, #224) [Link]
I was really hoping they'd help out in the embedded space either their Fusion devices and chipsets. The state of BIOSes in the ATOM world is simply atrocious.
AMD and Coreboot Posted May 9, 2011 16:34 UTC (Mon) by Hausvib6 (guest, #70606) [Link]
Glad to see this. I'll support this move by using coreboot-capable AMD platform whenever possible.
I hope more and more hardware companies open their products.
AMD and Coreboot Posted May 9, 2011 20:43 UTC (Mon) by fotoba (subscriber, #61150) [Link] It took very long timeIn our local IT news was published 15.12.2006 AMD dokonca uvauje, e v budúcich generáciách základných dosiek umoní OEM výrobcom pouiť LinuxBIOS namiesto klasického BIOS-u. http://www.itnews.sk/spravy/software/2006-12-15/c93395-linuxbios-sa-na-svojich-pocitacoch-dockame-co-nevidiet in english December the 15th, 2006 AMD is even thinking to allow OEM partners to use LinuxBIOS (now Coreboot) instead of clasical BIOS for their future generation motherboards/mainboards. Four and half of the year of thinking in the IT? It is too much.
AMD and Coreboot Posted May 9, 2011 20:56 UTC (Mon) by ebiederm (subscriber, #35028) [Link]
AMD delivered on their promises in 2006. Unfortunately things got scrambled with the ATI acquisition, and this is AMD getting back on track with chipset support.
Frankly coreboot is a weird project that gets just enough support and development effort that it refuses to die. Which makes it amazing in a lot of ways.
AMD and Coreboot Posted May 10, 2011 0:14 UTC (Tue) by ajft (guest, #52749) [Link]
Great news, but why the gibberish business speak of having to put "Going forward" in every sentence or article?
AMD and Coreboot Posted May 10, 2011 6:05 UTC (Tue) by alison (✭ supporter ✭, #63752) [Link]
In a presentation about power management at last August's LinuxCon, Len Brown of Intel bemoaned how a poorly written BIOS had clobbered his shiny new cpu_freq and cpu_idle improvements on the first product with which they were included. "Would Coreboot help?" I asked during the the Q&A. Brown said something to the effect that Coreboot was a joke. I was too surprised to ask why. The video of the talk is up on the web, so you don't have to take my word for it!
Coreboot is a Joke? Posted May 10, 2011 16:05 UTC (Tue) by brugolsky (subscriber, #28) [Link]
And what is APM, ACPI, SMI, EFI, IPMI, PXE, ...? $DEITY save us from Intel's notions of system management. And there is *ZERO* technical reason for this crap; it is all driven by market segmentation plays, the-whole-world-is-Windows mentality, and other stupidity. I've run out of fingers to count the number of platforms on which I've had to bang registers with setpci or a kernel patch just to use the features of the hardware that I bought (virtualization, HPET, AHCI, ... the list is long).
If AMD saw fit to support reasonably competitive hardware that was 100% documented open source from power-on [a goal that I was pitching hard to some AMD VPs early this century while working with a 2nd-tier server vendor], I'd ditch Intel everywhere. AMD did assist LinuxBIOS/CoreBoot in an on-again, off-again fashion, with good processor documentation, less so for their reference chipsets [and understandably so, since that is where the random IP blocks live]. The real problem was, and is, documentation for the Nvidia and other chipsets often found in real motherboards. What a shame to have beautiful Opteron hardware, and then have to struggle to get the on-board SATA and NIC working reliably. :-( Even Nvidia eventually saw that it was in their interest to offer some help, first with enhancements to Manfred Spraul's reverse-engineered NIC driver, and later with the SATA driver.
-Bill
AMD and Coreboot Posted May 11, 2011 4:07 UTC (Wed) by Hausvib6 (guest, #70606) [Link]
With an open BIOS, FLOSS one step closer to bare metal. There are still things like firmware, microcode, etc. Buth with something open, I think it will benefit FLOSS platform in general, perhaps for things like power management.
|
|||||||||||||