Julio M. Merino Vidal’s Google
Summer of Code Project, a new memory-based file system called
tmpfs has been
imported into the NetBSD source tree.
See Julio’s
message to the tech-kern mailinglist for more details.
Julio M. Merino Vidal’s Google
Summer of Code Project, a new memory-based file system called
tmpfs has been
imported into the NetBSD source tree.
See Julio’s
message to the tech-kern mailinglist for more details.
Well, google is doing great job for free software community. They have summer of code, IM based on jabber etc. It seems that we will have first “open source” monopoly in few years
.
tmpfs has been in Solaris for years, and is present in the recent OpenSolaris release (http://cvs.opensolaris.org/source/xref/usr/src/uts/common/fs/tmpfs/) although this looks like a new implementation. Good to see though.
Sounds broadly the same as tmpfs on Linux too, though I’ve no idea how similiar the implementations are…
Off-topic a bit, but I heard that gtalk was actually developed for Google’s SoC.
…the more I am interested. It might not be the most popular nor the most performant but I definitely like what I see! It’s unfortunate that the documentation is rather scarce but I might still try it soon (and might try to change this).
It’s a bit late for including tmpfs in 2.1, but do they plan to ship 3.0 with it?
It’s a bit late for including tmpfs in 2.1, but do they plan to ship 3.0 with it?
More likely, it’ll be a 4.0 thing. The 3.0_BETA has already been branched, and is scheduled for release next month.
What I’m wondering is……where’s the HFS+? That was also a NetBSD SoC project….
FFS really needs a good replacement now. 🙂
As for trying NetBSD, I can’t recommend it enough. It’s just a solid, no-nonsense OS. You’re not going to find anything particularly cutting-edge, but you’re not going to find any nasty hacks, either.
And it works almost exactly the same on every single platform.
what do you mean nebsd is not performant – after an overhaul of its internal algorithms for 2.0 it beat freebsd on many points and gave linux a run for its money .. i suspect 3.0 will push the envelope further.
the myth that freebsd is faster/performant is just that – a math. sure, freebsd does do a bit of micro-optimisation … but “correct by design” and no “hacks” is a more robust, correct and maintainable approach.
if netbsd scored 8/10 and freebsd 9/10 for speed – i’d go for netbsd knowning that its internals are clean and maintainable and clear – its implementation done by considered design.
but choice is choice …
Yeah, I remember the overhaul, but I wonder if they did it for a specific benchmark suite[1]. Anyway, it doesn’t matter that much since I value stability over performance or the lastest bling-bling.
Will probably give it a try soon.
[1]: http://bulk.fefe.de/scalability/#netbsd2
the myth that freebsd is faster/performant is just that – a math. sure, freebsd does do a bit of micro-optimisation … but “correct by design” and no “hacks” is a more robust, correct and maintainable approach.
Actually, NetBSD is more into micro-optimizations than FreeBSD. Those benchmarks (and later others) are of micro operations.
FreeBSD tend to focus more on macro-optimizations such as queuing operations for processing under a common lock to decrease locking overhead.
Back on topic tho, tmpfs is quite nice to have. GSoC have brought several nice things into the BSD sphere and I expect several of them will be ported between the BSD:s
Thanks for the info. While at it, is it easy to upgrade from one version to another? I am wondering if I should wait for trying 3.0 instead of 2.1…
> While at it, is it easy to upgrade from one version to
> another? I am wondering if I should wait for trying 3.0
> instead of 2.1…
Yes, upgrading is easy, both from source as from binary sets. But if you want to hop on 3.0 right away, you may want to install 3.0_BETA (and help testing the Release Candidates in a few weeks :-). I’m typing this from a 3.0_BETA machine which has been running smooth for months.
Watch the ftp://ftp.netbsd.org/pub/NetBSD-daily/ space for regular bulk builds of all branches.