A Developer Is Going To Focus On Improving GNU Hurd's Hardware Support

[phoronix]

Phoronix: A Developer Is Going To Focus On Improving GNU Hurd's Hardware Support

As explained recently in The Current State Of Debian GNU Hurd, PC hardware support by GNU Hurd is pretty piss poor right now with no support for USB yet, sound support being in its early stages, and 64-bit support just being started. However, a developer is hoping to work on improving hardware support for Hurd...

http://www.phoronix.com/scan.php?page=news_item&px=GNU-Hurd-Dev-Hardware

[Veto]

PC hardware support by GNU Hurd is pretty piss poor right now...

Have you considered changing the name to Prophanix?

[OneTimeShot]

Well, I guess this means that he'll be porting Linux device drivers. It makes sense to create nice clean portable abstraction layers, so I suppose this work will be worthwhile for Linux maintainabilty too. As for Hurd, I like the theory of "running bits as separated service processes" but I don't think that it adds enough end user value to catch on at this point (ditto the BSD kernels)...

[Delgarde]

As for Hurd, I like the theory of "running bits as separated service processes" but I don't think that it adds enough end user value to catch on at this point (ditto the BSD kernels)...

Oh, microkernels are great in theory...nice clean modular architectures, where the core is kept small, and stuff like FS implementations can be pushed out to userspace. The problem is that they just don't work out so well in practice when "beautiful design" meets "real world requirements". Pretty quickly, you find that where you'd drawn up neat divisions of responsibility, your actual hardware doesn't fit neatly into the design, or you don't get decent performance unless you bypass some of the barriers.

Basically, real-world systems are always messy - there's no way around it. All you can do is to try to find a good balance... try to build a good design, but acknowledge that compromises have to happen... that the design *will* get in the way sometimes, and when it does, it's usually the design that has to give way to functionality.

[WorBlux]

Oh, microkernels are great in theory...

L4, a microkernel is the only OS that has been able to provide hard guarantees of isolation and real-time operation. Hurd seems to be behind the times though even as far as microkernels are concerned.


Another benifit from the overly-architected systems is it's easier to push certain test new ideas in a simpler enviroment.

[dragorth]

L4, a microkernel is the only OS that has been able to provide hard guarantees of isolation and real-time operation. Hurd seems to be behind the times though even as far as microkernels are concerned.


Another benifit from the overly-architected systems is it's easier to push certain test new ideas in a simpler enviroment.

To be fair to Hurd, it is based on the Mach kernel, which was a micro kernel implementation on top of the FreeBSD kernel. I.e. it was a microkernel in name only. I actually don't know what state it is in compared to the 4th generation of microkernels such as seL4, but its base was much closer to what is available in iOS/Mac OS X/TV OS than the L4s exist today.

Porting Hurd, which is mostly user land, to seL4 would be interesting, perhaps using some of the Genode Frameworks to fill in the gaps. Genode already has been porting their framework to seL4, and it seems like a more consumer friendly option, or at least ready for prime time, than Hurd, thanks to its already included use of those same Rump Drivers.

[swoorup]

To name great examples of microkernel, shouldn't macosx and windows be ones?

[garegin]

Oh, microkernels are great in theory...nice clean modular architectures, where the core is kept small, and stuff like FS implementations can be pushed out to userspace. The problem is that they just don't work out so well in practice when "beautiful design" meets "real world requirements". Pretty quickly, you find that where you'd drawn up neat divisions of responsibility, your actual hardware doesn't fit neatly into the design, or you don't get decent performance unless you bypass some of the barriers.

Basically, real-world systems are always messy - there's no way around it. All you can do is to try to find a good balance... try to build a good design, but acknowledge that compromises have to happen... that the design *will* get in the way sometimes, and when it does, it's usually the design that has to give way to functionality.

That's not the issue. It's that both Windows, OS X and Linux were designed in an era when microkernels were deficient. This is why both Windows and OS X use hybrid kernels. L4 proved that microkernels are actually better when designed properly. The problems was that by that time others already finished designed their OSes. They couldn't just switch to a microkernel without major rearchitecting.

" In November 2005, NICTA announced that Qualcomm was deploying NICTA's L4 version on their Mobile Station Modem chipsets. This led to the use of L4 in mobile phone handsets on sale from late 2006. In August 2006, ERTOS leader and UNSW professor Gernot Heiser spun out a company called Open Kernel Labs(OK Labs) to support commercial L4 users and further develop L4 for commercial use under the brand name OKL4, in close collaboration with NICTA. OKL4 Version 2.1, released in April 2008, was the first generally available version of L4 which featured capability-based security. OKL4 3.0, released in October 2008, was the last open-source version of OKL4. More recent versions are closed source and based on a rewrite to support a native hypervisor variant called the OKL4 Microvisor. OK Labs also distributed a para-virtualized Linux called OK:Linux, a descendant of Wombat, as well paravirtualized versions of SymbianOS and Android. OK Labs also acquired the rights to seL4 from NICTA.
OKL4 shipments exceeded 1.5 billion in early 2012,^[1] mostly on Qualcomm wireless modem chips. Other deployments include automotive infotainment systems.^[10]
The Apple A7 chip contains a Secure Enclave processor running an L4 operating system,^[11] based on the L4 version used in NICTA's Darbat project. This implies that L4 is now shipping on all iOS devices, the total shipment of which is estimated at 310 million for the year 2015."

[gigaplex]

To name great examples of microkernel, shouldn't macosx and windows be ones?

Neither of those are microkernels, they're both hybrids.

[ncopa]

I expect to use hurd as main OS around year 2059

http://www.xkcd.com/1508/