Linux 3.1 Released With Support for the OpenRISC CPU 165
diegocg writes "Linux 3.1 has been released. The changes include support for the OpenRISC opensource CPU; performance improvements to the writeback throttling; some speedups in the slab allocator; a new iSCSI implementation; support for NFC chips; bad block management in the generic software RAID layer; a new 'cpupowerutils' utility for power management; filesystem barriers enabled by default in Ext3; Wii Controller support; and [the usual] new drivers and many small improvements."
3 series (Score:2)
There have been so many major improvements during the life of the 2 series. I wonder what finally through them over the line to go into the 3 series.
Marketing (Score:2)
Basically, it's the beginning of the end.
Re: (Score:2, Funny)
In January 2012 they will release Linux 3.11
In March they will release Linux 95
In May they will release Linux 98
In July they will release Linux Me
In September they will release Linux XP
In November they will release Linux Vista
In January 2013 they will release Linux 7....
and then it will be the year of Linux on the Desktop!
Re: (Score:2)
Re:3 series (Score:4, Informative)
Re:3 series (Score:5, Informative)
"So what are the big changes?
"NOTHING. Absolutely nothing. Sure, we have the usual two thirds driver changes, and a lot of random fixes, but the point is that 3.0 is just about renumbering, we are very much not doing a KDE-4 or a Gnome-3 here. No breakage, no special scary new features, nothing at all like that. We've been doing time-based releases for many years now, this is in no way about features. If you want an excuse for the renumbering, you really should look at the time-based one ('20 years') instead.
tl;dr - Nothing happened.
Re: (Score:2)
"NOTHING. Absolutely nothing."
Linux (huh, good god), what is it good for?
Re: (Score:2)
Which means the version number becomes completely meaningless.
Unlike "2.6", which was incredibly meaningful.
Really, for the last decade, the only part of the version number that has had any meaning was the last number. They got rid of one out of two meaningless numbers. They might as well have gotten rid of the first number too.
Re: (Score:2)
The version number is useful as long as it's unique to each version, so you don't want them to drop it.
It's more useful if it increments with each version, though by how much doesn't matter.
A really useful version number would be a multipart that tracks independently incrementing versions of each of GUI, APIs, and file formats. That way users (and admins, and developers) could use the number to see whether a version is compatible with their people, SW or data. But nobody ever does that, though it's the onl
Re: (Score:3)
The model of incrementing the versions had long been incompatible (as far as getting reasonable numbers) with the changed development model. Since the development was now more like a steady stream of whatever features that happened to be developed instead of the old one where bunches of big features were released at once and less often, the mere silliness of a .39 release (as 2.6 had in practice become the "major" version (as opposed to major being 2 and minor being 6) with the aforementioned shift) and th
Re: (Score:2)
Gotcha. So now the minor is going to go up faster?
September 1991 (Score:3)
Linux entered its 3rd decade with 3.0.
"This is a free minix-like kernel for i386(+) based AT-machines," began the Linux version 0.01 release notes in September of 1991 for the first release of the Linux kernel.
Re: (Score:2)
Oh good time to make the flip. IC
Where can I get one? (Score:2, Interesting)
Where can I get an OpenRISC CPU and a motherboard that will support it, and how much do they cost compared to Intel/AMD CPUs of similar performance?
Re: (Score:2)
how much do they do they cost compared to Intel/AMD CPUs of similar performance?
Depends on where you live and what 10 year old PCs go for at your local garage sales.
Re: (Score:2)
Were ARM motherboards (I assume you don't mean embedded stuff) available when Linux added support for it?
For that matter, are any available now?
Re: (Score:2)
For that matter, are any available now?
http://www.embeddedarm.com/ [embeddedarm.com] aka technologic systems
http://beagleboard.org/ [beagleboard.org]
http://www.friendlyarm.net/ [friendlyarm.net]
Look in the embedded market, not the FPS gamer enthusiast market.
Re: (Score:3)
TI's OMAP platform uses ARM cores. You can get eval boards for under $200.
Re: (Score:1)
You can't get one of similar performance to a modern Intel CPU. Think orders of magnitude slower.
Re: (Score:3)
OpenRisc is a soft-cpu, defined in the Verilog language, suitable for implementing in many different types of FPGA's of varying price/performance/power.
Here is one source for boards of all types:
--jeffk++
Re: (Score:2)
Can I implement it on a Zynq-7000 [wikipedia.org]? How many gates does it consume (and so how many left over on the Zynq)?
Re: (Score:2)
I haven't tried the OpenRISC cpu but apparently it takes more gates than the MicroBlaze soft-cpu that Xilinx provides. BTW you can already run linux on the Microblaze in many Xilinx devices; see my github repo: https://github.com/jdkoftinoff/mb-linux-msli [github.com]
--jeffk++
Re: (Score:2)
Hey, you don't happen to know where I can get a PIC18F core that'll run on a Zynq-7000? I've got an embedded PIC board with code I'd like to try turning "soft", possibly as a peripheral to a soft CPU running Linux, on the Zynq as a "virtual HW" host. Once it's all in gates I want to try porting Linux and PIC code into straight circuits. Possible?
Re: (Score:2)
I think it would be interesting to run Linux on one of the Zynq's ARMs, and run Linux on OpenRISC on the FPGA, and port Linux functions from SW into HW while leaving a complete SW Linux on the ARM. The tight integration on the chip of the two Linux instances could make the tools more effective. But really just fun, I think.
Re: (Score:2)
You can't get one with even 10% of a modern intel/amd's performance, and they cost more.
Re:Where can I get one? (Score:4, Interesting)
Where can I get an OpenRISC CPU and a motherboard that will support it,
http://opencores.org/or1k/FPGA_Development_Boards [opencores.org]
I have not bothered to research why its listed as supporting the Spartan-3A DSP 1800 and not the spartan3 dev board I have sitting at home, probably needs more gates? Or depends on some part of the DSP1800's innards? Or simply the dude who did it owned a DSP1800 as opposed to the board I have at home?
and how much do they cost compared to Intel/AMD CPUs of similar performance?
Which vegetable has similar price to an apple or an orange? Perhaps a potatoe?
The thing with FPGAs is... how much do you wanna spend? I know there are simply gigantic FPGA arrays out there, so on one FPGA chip you could probably put a whole Beowulf cluster of a dozen of these things on one chip complete with the ethernet switch that interconnects them. So its kind of meaningless, like debating the weight of a soul.
The goal of a FPGA system is not to be a generic processor, but to use the field programability... you use the embedded CPU for generic "who cares how fast" stuff like a user interface, or a TCP network stack, or a DHCP client. You do all the heavy lifting inside FPGA hardware itself. If you used this CPU for a FPGA based mythtv frontend, you would not write a H.264 decoder in the emulated RISC processor assembly language, you'd use a hardware one (or at least hardware accelerated one) inside the FPGA written in verilog or VHDL. If you're running benchmarks on the FPGA processor trying to optimize it, you're probably doing it Very wrong, or trying some insane level of optimization / price cutting.
Re: (Score:2)
Re: (Score:2)
I know they are still trying to get funded for the 'open' project, but didn't some 3rd party already do this about a year ago? Or did that end up just being vapor?
Cool, Now Fix Sandy Bridge (Score:3, Interesting)
Re: (Score:2)
Y U NO FIX? requests are pretty useless without knowing what's broken.
Even older versions of Linux (such as the kernels included with Ubuntu 11.04) work just fine on Sandy Bridge - I just upgraded and it's great.
So whatever's broken for you is obviously some specific corner case which you haven't bothered to specify.
Re: (Score:1)
?? It's been working fine on my wife's Sandy Bridge i3 since 3.0 (including the onboard graphics). Actually the stock 2.6.39 released with Natty ran okay on it; I built a 3.0 kernel and ran that which worked better, particularly in conjunction with a recent Mesa and libva. I've since dist-upgraded her to Oneiric and everything still runs fine (still on 3.0, and I gather there are some further improvements planned for 3.1) - but I wouldn't exactly say it was broken before.
There are serious power usage regressions, making many laptop users suffer. Compare, e.g., https://bugs.launchpad.net/ubuntu/+source/linux/+bug/834037 [launchpad.net]
Re: (Score:2)
I'm guessing that GP was referring to support for QuickSync, which is only available in the i5 and i7.
How well do openrisc cpus compare? (Score:3)
Against other open cores such as the SPARC cores?
Re: (Score:2)
Re: (Score:2)
3.11 (Score:5, Funny)
I for one am holding out for 3.11. I heard it will be for Workgroups!
Re: (Score:3)
LUSER ... I'm holding out for 3.14159265...
Re: (Score:2)
More vibrant colors? (Score:2)
I for one am holding out for 3.11. I heard it will be for Workgroups!
It also has wonderful improvements to video compression, I hear.
Will it have more vibrant colors like the Intel CPUs?
Wiimote support built-in (Score:2)
Whoa, I didn't expect that.
Some can argue it's unnecessary and that stuff, but I have a classic controller and it's damn good to use with my computer. (I actually use it more with it than my Wii......).
What is that "barrier" for ext3, btw?
Re: (Score:3)
From the Changelog linked to in the article...
1.3. Filesystem barriers enabled by default in Ext3
Hard disks have a memory buffer were they temporally store the instructions and data issued from the OS while the disk processes it. The internal software of modern disks changes the order of the instructions to improve performance, which means that instructions may or may not be committed to the disk in the same order the OS issued them. This breaks many of the assumptions that filesystems need to reliably impl
Re: (Score:2)
Re: (Score:3)
This, >9000 times. I don't understand why this would need to be in the official kernel... if somebody really wants/needs it in the kernel, shouldn't they be compiling it in themselves? Why should people have to choose to exclude it?
Doesn't this mean that future security audits have to include this wii driver? Do bloat-conscious or security-minded people have to cut this out?
I'm not trying to be sarcastic, I'm genuinely curious, and I'm well aware of how wrong 'common sense' can be when one steps outsi
Re: (Score:2)
Since it is a modular driver, it will IIRC execute nothing at all until the module is pulled in by udev. So there's no need to audit much if you're not using the hardware in question. And if you have physical access to the server, there are ways of subverting it other than hooking up hardware that has security holes in the drivers. So no need for paranoia.
Re: (Score:2)
Ummm. Isn't this sort of like saying "Don't worry about the screen door on this submarine. As long as nobody uses it, we are OK."
Re: (Score:2)
Nope. Not at all. You can't really randomly run into this issue without attaching a device first, and if you can do that you may as well own the server in other ways.
Re: (Score:2)
Maybe, but the point still stands: don't throw stuff in the kernel without a very good reason to do so.
Re: (Score:2)
Since there's on the order of a 100 million wiimotes out there, I'd think just that is a good enough reason. There's plenty of drivers for way less popular hardware in the linux kernel! Less as in an order of magnitude or two less popular.
Re: (Score:2)
There is a reason to do this. The hardware exists. People want to use that hardware with the Linux kernel. The inclusion of the wiimote driver within Linus' development linux kernel tree just means that driver has become good enough for Linus to redistribute it.
Re: (Score:2)
Re: (Score:2)
Well, if it concerns you the module can be removed from the system completely. I would also expect the driver to not be built as part of RHEL or SLES distributions, seeing as how it probably has little to no use on a server platform.
Re: (Score:2)
Yes, if you're in a sub that doesn't have a screen door installed, then you don't have to worry if it can hold water out/air in, which is basically whats going on here.
The code doesn't get pulled in without the hardware being detected. No hardware, no udev, no exploit. If the hardware is there, then you'll have a udev and an active driver, at which point you can do something about it, since you are aware of the hardware in your machines generally when you're that paranoid about security.
Re: (Score:2)
Typically, the module will be deleted from the server since it isn't supposed to be used. That done, there's no need to further audit it.
Re: (Score:2)
This is how the Linux kernel development process works. If someone writes a Linux driver for a piece of hardware they can usually get that driver into the main kernel tree if they follow the proper process. The Linux Kernel Mailing List FAQ covers this here: http://vger.kernel.org/lkml/#s2 [kernel.org] It says that the driver must be tested successfully by other people. The code has be written against the latest kernel. Coding standards and best practices have to be followed. This driver has just as much right to be the
Re: (Score:2)
How is this any different from an ethernet card driver you aren't using?
Re: (Score:2)
Try this link [kernel.org]
Re: (Score:2)
The Wii controller is a very nice set of sensors in a cheap package. Researchers have already done cool stuff with it.
I'd say it's anything but unnecessary, even if most people likely won't use it.
Re: (Score:2)
True but I would love for Linux to add a standard interfaces for SPI, GPIO, and ican. That could help embedded developers a lot.
Re: (Score:2)
SPI and GPIO support exist (i2c too), it's just very hardware dependent and needs to be configured with board-specific bits at compile time.
GPIO [kernel.org]
SPI [kernel.org]
Not exactly easy at first. There are userland extensions to all of the above too, which is fine for blinking LEDs and such, but has some limitations... spidev only has certain modes and data lengths, /SS lines are defined in board config (so kernel compile to change), things like that.
Re: (Score:2)
I knew that i2c was around but the dev board I used all had proprietary interfaces to the GPIO and SPI.
Re: (Score:2)
Sorry to double post but that is really handy. Now IAMFT I just need to write code to convert a Centronics port to GPIO and and bit banged SPI
OpenRISC on FPGA? (Score:2)
Is there an FPGA big enough to implement the OpenRISC on it? Has anyone done that yet?
Re: (Score:2)
There have been FPGA's big enough to implement OpenRisc on now for at least 13-14 years. I designed an OpenRISC based system that included additional logic many times the size of the OpenRISC on an FPGA in 2002 and designed and manufactured an ASIC with most of the same logic in 2003. These repeated "Someone just built the first opensource chip/system/board/hardware etc" threads on Slashdot make me laugh. Does anyone do any research anymore?
Re: (Score:2)
OpenRISC/FPGA seems like an interesting platform on which to study converting OS functions from iterated CPU code into parallel circuits. Device drivers come to mind, and all the OS functions that lookup DBMs or /etc config params, but any kernel module might be a candidate. Is anyone grinding away at the OS to move it off of the CPU and into circuits?
Re: (Score:2)
I have coded device drivers, though I have not coded Verilog or VHDL for devices. I know there will be a HW/SW interface, but I'd like it to be FPGA configs that put most of the logic into circuits that don't consume CPU cycles - but can be changed digitally as the devices and OS change around the needed interface. I'd like the OS driver to be the minimum interface to the processes accessing the device. Mapped registers that "virtualize" the device into just an API that's almost all reconfigurable HW would
Donations needed (Score:2)
They desperately need more funding to produce a ASIC version required for Linux support to mean much:
http://hardware.slashdot.org/story/11/04/30/172214/help-build-the-worlds-first-community-funded-cpu-asic [slashdot.org]
Re: (Score:3, Interesting)
Don't feel too bad, from what I remember Linus just randomly decided that a minor number of the 2.6 series was now 3.0 a few months back.
Re:3.1! and I'm still stuck on 2.6... (Score:5, Informative)
Re: (Score:3)
Re: (Score:2)
That would be a rather silly assumption, since the "." in the version number is not a decimal separator. The version after 3.9 will presumably be 3.10.
If the 3.x series lasts as long as the 2.6.x series, then 4.0 would happen around 2019. Or if Linus decides to stick with a time-based approach to incrementing the major number, then a sensible schedule might be to incremement the major number every five years. In any case
Re: (Score:2)
That would be a rather silly assumption, since the "." in the version number is not a decimal separator. The version after 3.9 will presumably be 3.10.
If major.minor was going to mean something, sure. As it is, the only reason to go with 3.10 over 4.0 would be for that 3.11 joke.
Re: (Score:3)
Actually they've been on a steady quarterly release schedule now for years, one month merge window, two months = 8 weeks of release candidates (RC). If they need more or fewer they simply adjust the length of the merge window, that's the period when all new development of the last months is pulled into the main tree. The release candidates are for QA, bug fixing and regression testing, meanwhile new development continues in branches. The merge window and rc1 can be a little hairy, but rc2+ is normally fine,
Re:3.1! and I'm still stuck on 2.6... (Score:4, Insightful)
Can opensource projects stop with this utterly terrible use of the major.minor numbering...
You're right. The sky is in fact falling.
Re:3.1! and I'm still stuck on 2.6... (Score:4, Insightful)
Re: (Score:2)
My first employer insisted that all software start at version 3.1.
His theory was that it would be easier to sell software that was version 3.1 or later than starting at version 1.
Every application we wrote therefore began at Version 3.1 when brand new.
Re: (Score:1)
Re: (Score:3)
He's not the first to think along those lines. One of the most well-known examples is the Kaypro II computer. It was actually their first product but they wanted that magic "II" after the name.
Re: (Score:2)
Or Solaris [wikipedia.org], which was released by Sun as a successor to SunOS 4.x as "Solaris 2", while retroactively renaming SunOS 4.x as "Solaris 1". Then after Solaris 2.6 Sun called the next version Solaris 7.
There's other reasons other than marketing saying "higher version numbers are more advanced!", but they're even stupider.
Windows, for all its BS in numbering, was doing the right thing by naming versions after the release year. But I guess the further they got from the magically futuristic "2000", the more they w
Re: (Score:2)
of course everyone calls the Ubuntu version by its alliterative animal name
I, for one, prefer numbers and often cannot map them to animal names (especially given that names are different for English and localized Ubuntu versions), at least for past releases. Also, I don't remember the last time when someone else around me used animal names in spoken language (well, this can be partially explained by the fact that those names aren't easy to remember for non-English speakers).
Re: (Score:2)
Followed by version 95, 98 and ME?
Re: (Score:2)
You forgot 3.11.
Re: (Score:2)
Re: (Score:2)
Yeah, because Microsoft Windows NT 3.5 had 3 releases behind it, at least, right?
Oh wait, it didn't? They called the first release 3.5 because it was a larger number than 3.11? Even though it had nothing in common with 3.11 or 95 or anything else but VMS and OS/2?
Wow, those closed source version numbering methods are so superior.
Yeah.
Moron.
--
BMO
Re: (Score:2)
http://en.wikipedia.org/wiki/Windows_NT_3.1 [wikipedia.org]
Re:3.1 (Score:5, Funny)
Re: (Score:1)
Re: (Score:2)
Can't wait for Linux 3.11 for Workgroups
followed by Linux NT, Linux XP, Linux Vista and reverting back to numbers w// Linux 7 & Linux 8
Slow down there, cowboys! You forgot Linux 95, 98, and 2000.
Re: (Score:1)
Re: (Score:1)
And don't forget the "Signed by Linus Torvalds himself" edition
Re: (Score:2)
I am waiting for the LSDN
Re: (Score:2)
Don't forget the home versions with Networking disabled.
Re: (Score:2)
Re: (Score:2)
Re: (Score:2)
Word for Linux 3.11 for Workgroups for Pen Computing FTW!
Re: (Score:2)
Re: (Score:2)
Okay, those are all important things but nothing you described has to do with the kernel.
plus I want me Email client to have full Exchange 2010 support.
And did you not want to download your email client as a third-party application along with that? But the graphical framework, the desktop manager, and the widgets framework will all be third-party apps as well.
The kernel takes care of hardware support and basic I/O such as file systems and things. Everything else is a "third-party" app.
Re: (Score:2)
Re: (Score:2)
gitweb links broken, but git repo still available (Score:2)
The stable repo appears to be at:
git://git.kernel.org/pub/scm/linux/kernel/git/stable/linux-stable.git
Re: (Score:2)