|
|
Log in / Subscribe / Register

OpenWrt "Barrier Breaker" 14.07 released

From:  Steven Barth <cyrus-AT-openwrt.org>
To:  <openwrt-devel-AT-lists.openwrt.org>
Subject:  Barrier Breaker 14.07 Final
Date:  Thu, 02 Oct 2014 14:59:08 +0200
Message-ID:  <8c592d5ef8dabd40295cd2c5f698d1c7@midlink.org>

The OpenWrt developers are proud to announce the final release
of OpenWrt Barrier Breaker.

   _______                     ________        __
  |       |.-----.-----.-----.|  |  |  |.----.|  |_
  |   -   ||  _  |  -__|     ||  |  |  ||   _||   _|
  |_______||   __|_____|__|__||________||__|  |____|
           |__| W I R E L E S S   F R E E D O M
  -----------------------------------------------------
  BARRIER BREAKER (14.07)
  -----------------------------------------------------
   * 1/2 oz Galliano         Pour all ingredients into
   * 4 oz cold Coffee        an irish coffee mug filled
   * 1 1/2 oz Dark Rum       with crushed ice. Stir.
   * 2 tsp. Creme de Cacao
  -----------------------------------------------------

http://downloads.openwrt.org/barrier_breaker/14.07/

Important changes since RC3
* various ath9k related fixes
* a few board related fixes
* fixes for packages depdending on curl
* per feed download folders

Important changes since RC2
* NAT & firewall throughput improvements
* Security updates for OpenSSL & PolarSSL
* Minor fixes in DHCP & DHCPv6 handling
* Configuration support for GRE tunnels
* Various other fixes

Important changes since RC1
* fix a long standing ath9k deadlock bug
* all feeds are now built
* image builder now works and RC2 contains all board specific images
* various board/stability fixes

** Highlights since Attitude Adjustment **
Default configuration and images

* Linux kernel updated to version 3.10
* Procd: new preinit, init, hotplug and event system written in C
* Native IPv6-support
     - RA & DHCPv6+PD client and server
     - Local prefix allocation & source-restricted routes
       (multihoming)
* Filesystem improvements
     - Added support for sysupgrade on NAND-flash
     - Added support for filesystem snapshot and rollback
     - Rewritten mounting system in C for rootfs and block devices
* UCI configuration improvements
     - Support for testing configuration and rollback to working
       last working state
     - Unified change trigger system to restart services on-demand
     - Added a data validation layer
* Networking improvements
     - Netifd now handles setup and configuration reload of
       wireless interfaces
     - Added reworked event support to allow obsoleting network
       hotplug-scripts
     - Added support for dynamic firewall rules and zones
     - Added support for transparent multicast to unicast
       translation for bridges
     - Various other fixes and improvements

Additional highlights selectable in the package feeds or SDK
* Extended IPv6-support
     - Added DS-Lite support and improved 6to4, 6in4 and 6rd-support
     - Experimental support for Lightweight 4over6, MAP-E and MAP-T
     - Draft-support for self-managing home networks (HNCP)
* rpcd: new JSONRPC over HTTP-frontend for remote access to ubus
* mdns: new lightweight mdns daemon (work in progress)
* Initial support for the musl C standard library
* Support for QMI-based 3g/4g modems
* Support for DNSSEC validation
* Added architecture for package signing and SHA256 hashing
* ... and many more cool things

Package feed reorganization
For quite a while already we are not very satisfied with the quality
of the packages-feed. To address this, we decided to do a fresh start
on GitHub. The new feed https://github.com/openwrt/packages should be
used from now on and package maintainers are asked to move their
packages there. For the final release we will still build the old
packages feed but it will be necessary to enable it manually in the
opkg package list to be usable.
Additionally we would like to give a big thank you to all of our 
package
maintainers working on our various feeds.

New build servers
We would like to express our gratitude to Imagination Technology for
funding the 2 build servers that we used for the release.

Whats next ?
We aim at releasing Chaos Calmer (CC) before the end of the year. The
CC release will use 3.14 or a newer LTS kernel as baseline.


Have fun!
     The OpenWrt developer team



to post comments

OpenWRT "Barrier Breaker" 14.07 released

Posted Oct 2, 2014 22:50 UTC (Thu) by cpatulea (guest, #87498) [Link]

Congratulations to the OpenWrt team!

OpenWRT "Barrier Breaker" 14.07 released

Posted Oct 3, 2014 0:05 UTC (Fri) by jg (guest, #17537) [Link] (2 responses)

Congratulations to all.

"Friends don't let friends run factory firmware." Upgrade to OpenWrt today....

OpenWRT "Barrier Breaker" 14.07 released

Posted Oct 3, 2014 19:16 UTC (Fri) by jhoblitt (guest, #77733) [Link] (1 responses)

Unless you have a device with NAND flash in it :)

OpenWRT "Barrier Breaker" 14.07 released

Posted Oct 3, 2014 22:12 UTC (Fri) by rknight (subscriber, #26792) [Link]

Why? Barrier Breaker now even supports sysupgrade on NAND flash!

OpenWRT "Barrier Breaker" 14.07 released

Posted Oct 3, 2014 4:29 UTC (Fri) by zenaan (guest, #3778) [Link] (1 responses)

Anyone know what is required for mainline head branch kernels to become drop-in replaceable? As in, what patches are OpenWRT carrying, or what limitations are not in the mainline kernel which stop this? Linux kernel 3.10 is already a few releases old now...

OpenWRT "Barrier Breaker" 14.07 released

Posted Oct 9, 2014 1:54 UTC (Thu) by kokada (subscriber, #92849) [Link]

Nothing that interesting, really. If you pull the current trunk of OpenWRT it will probably run a more recently kernel (and generally is pretty recent too, like one or at maximum two releases older than current kernel). But for a stable release OpenWRT freezes a kernel version and only pulls bugfixes. And the freeze for Barrier Breaker was some time ago, so this is why the kernel is 3.10.

OpenWRT "Barrier Breaker" 14.07 released

Posted Oct 3, 2014 11:44 UTC (Fri) by jpfrancois (subscriber, #65948) [Link] (1 responses)

Has anybody used procd out of openwrt ? I am looking for an "embedded" init replacement.

embedded 'init' replacement?

Posted Oct 3, 2014 14:08 UTC (Fri) by charlieb (guest, #23340) [Link]

Have a look at s6.

http://skarnet.org/software/s6/

OpenWRT "Barrier Breaker" 14.07 released

Posted Oct 3, 2014 14:51 UTC (Fri) by SEJeff (guest, #51588) [Link] (33 responses)

I guess I'll be "that guy" who wonders why they wrote their own embedded init system, "procd", instead of using systemd. Yocto Linux uses systemd, and that is most definitely an embedded distribution.

Perhaps it is because embedded folks tend to over NIH due to wanting an exact feature set and no more? I'm legitimately unsure. Any insight from the greater lwn community?

OpenWRT "Barrier Breaker" 14.07 released

Posted Oct 3, 2014 15:11 UTC (Fri) by BlueLightning (subscriber, #38978) [Link]

For clarity, OpenEmbedded (which the Yocto Project contributes to) supports both systemd and sysvinit - you can select which one you wish to use.

OpenWRT "Barrier Breaker" 14.07 released

Posted Oct 3, 2014 16:49 UTC (Fri) by ntl (subscriber, #40518) [Link]

Seems to at least partly follow from their distaste for d-bus. OpenWRT has had its own message bus implementation ("ubus") for a while. Some background here, although the "why do we want procd" part of the page seems to be incomplete.

OpenWRT "Barrier Breaker" 14.07 released

Posted Oct 3, 2014 19:20 UTC (Fri) by jhoblitt (guest, #77733) [Link] (1 responses)

With put knowing the technical merits of the decision, on general principals I feel that init system script/config unification is an important issue. I just wish systemd wasn't such a PITA with docker.

OpenWRT "Barrier Breaker" 14.07 released

Posted Oct 5, 2014 16:55 UTC (Sun) by jeff_marshall (subscriber, #49255) [Link]

Could you elaborate on the issues you're facing with systemd inside docker?

So far, my containers only use older distributions (e.g., CentOS 6) so I haven't tried yet. I'd like to know what I'm in for.

OpenWRT "Barrier Breaker" 14.07 released

Posted Oct 4, 2014 18:34 UTC (Sat) by flussence (guest, #85566) [Link] (7 responses)

I'd imagine it's because writing a memory- and size-efficient init system was more realistic than expecting their users to get proficient at soldering additional flash and memory chips into their 4MB routers.

OpenWRT "Barrier Breaker" 14.07 released

Posted Oct 5, 2014 21:30 UTC (Sun) by HelloWorld (guest, #56129) [Link] (6 responses)

Well, the previous release already didn't work properly on devices with that little memory. My WRT54G with 16 MB reboots when I run "opkg update". One can download and install packages manually, but that's not really how it's meant to work…

OpenWRT "Barrier Breaker" 14.07 released

Posted Oct 9, 2014 2:04 UTC (Thu) by kokada (subscriber, #92849) [Link] (5 responses)

Not RAM, flash memory (the equivalent of a HDD for embedded devices). While you can't run OpenWRT reliable with only 16MB of RAM, you can run OpenWRT with only 4MB of Flash if you don't need advance features (like OpenVPN support). My TL-WR740N has 4MB of Flash and 32MB of RAM and runs greatly with Barrier Breaker.

Considering that systemd can only be built with glibc (that is much bigger than uclibc that is used by default on OpenWRT), and only the systemd binary uses half as much Flash space that is available in my Router (yeah, the image is compressed, but the point is still valid), not forgetting the dependency on Dbus and you get something that is impossible to use on the majority of the routers in the market because of space constraints. I think even Routers with 8MB of Flash would be in trouble if systemd was used instead of procd.

Not that I don't like systemd, it would be interesting to see it running on my Router, but unless I put my Router had at least 32MB of Flash I would probably be in trouble.

OpenWRT "Barrier Breaker" 14.07 released

Posted Oct 9, 2014 4:30 UTC (Thu) by raven667 (subscriber, #5198) [Link] (4 responses)

> Not that I don't like systemd, it would be interesting to see it running on my Router, but unless I put my Router had at least 32MB of Flash I would probably be in trouble.

It'd be interesting to see how that would work and what comprimises would need to be made. On my F20 system, /usr/lib/systemd/systemd which has all the options turned on is only 1.2MB and compresses down to 336kb so that may be viable on a 4MB flash system, systemd does have a lot of ./configure options so maybe enough can be stripped to make it work for this case.

OpenWRT "Barrier Breaker" 14.07 released

Posted Oct 10, 2014 9:08 UTC (Fri) by fb (guest, #53265) [Link] (3 responses)

PS: You also need to account for glibc.

Honestly, I really do not get the urge to fit systemd into my router. I use OpenWrt on a router with 8MB Flash. What I want there is a base system which is _as small as possible_ because one of my primary goals with OpenWrt is to able to install more _applications_.

Larger base components would -in my understanding- decrease the utility of OpenWrt for its real life users (assuming in my fantasy that I am somehow representative of the average OpenWrt user).

PPS: it should be easy to 'just' cross-compile glibc+systemd+dbus (+any other dependency?) and measure the size of the resulting compressed & stripped packages.

OpenWRT "Barrier Breaker" 14.07 released

Posted Oct 10, 2014 14:55 UTC (Fri) by raven667 (subscriber, #5198) [Link] (1 responses)

> PPS: it should be easy to 'just' cross-compile glibc+systemd+dbus (+any other dependency?) and measure the size of the resulting compressed & stripped packages.

Yeah I know, I tried downloading the systemd source before responding and playing around with the configure options but ended up making things worse so clearly would need to spend a lot more time tediously checking how much each options changes the size of the resulting tools, more than I wanted to spend just to respond to a comment.

It just seems to me that like Linux itself, systemd should have the options to scale it from large systems to small systems, if there are options missing it would be nice to add them, so far most of the development has been on bigger systems so it hasn't grown these options organically.

OpenWRT "Barrier Breaker" 14.07 released

Posted Oct 10, 2014 16:16 UTC (Fri) by dlang (guest, #313) [Link]

If systemd supported disabling features, it would have far fewer opponents than it does. But for all the claims about how modular it is, when you actually try to turn off features, it doesn't work expected

You are far from the first person to report this sort of problem

OpenWRT "Barrier Breaker" 14.07 released

Posted Oct 10, 2014 18:10 UTC (Fri) by Cyberax (✭ supporter ✭, #52523) [Link]

Somebody compiled systemd using the musl libc. The amount of patches was very small.

Lennart's answer is also illuminating: https://www.mail-archive.com/systemd-devel@lists.freedesk... - he's not opposed to merging good cleanup patches.

OpenWRT "Barrier Breaker" 14.07 released

Posted Oct 5, 2014 3:14 UTC (Sun) by xxiao (guest, #9631) [Link]

Openembedded is the Gentoo for embedded world, a bit too heavy for me, but hi it gets Intel at the back.

Openwrt is more like Debian for embedded(except for the systemd part), it has no major sponsors other than an active community, and no systemd, which is _really_ great.

OpenWRT "Barrier Breaker" 14.07 released

Posted Oct 8, 2014 21:46 UTC (Wed) by xslogic (guest, #97478) [Link] (7 responses)

It probably doesn't help that systemd requires glibc, when embedded systems typically (But not always, I'll grant you) use uclibc.

Which is odd - given that embedded systems are one of the supposed targets.

OpenWRT "Barrier Breaker" 14.07 released

Posted Oct 8, 2014 22:03 UTC (Wed) by raven667 (subscriber, #5198) [Link] (3 responses)

I think it's already used in embedded systems like in-vehicle entertainment and I think of the offspring of the Nokia smartphone uses it. It's probably possible to port it to uclibc if someone was interested in doing the work, so far most of the maintenance has been done by a bunch of the traditional Linux distributors, who don't use uclibc.

OpenWRT "Barrier Breaker" 14.07 released

Posted Oct 8, 2014 23:34 UTC (Wed) by xslogic (guest, #97478) [Link] (2 responses)

The opening gambit for uselessd was that they'd repatched it to work with uclibc and musl.

This bug shows that patching it is unlikely to get into mainline: (And I don't quite see that uclibc is going to start to add the relevant glibc extensions - but hey, anything is possible)

https://bugs.freedesktop.org/show_bug.cgi?id=73729

Apparently the solution is to make uclibc more compatible with glibc. (Off topic: Sometimes I do find it mildly ironic that people say that Microsoft embrace, extend and extinguish...)

It looks like it got heated - and I don't feel that either side particularly covered themselves in glory... (Apartheid was a particularly bad choice of wording - and it kind of degenerated from there)

OpenWRT "Barrier Breaker" 14.07 released

Posted Oct 9, 2014 9:21 UTC (Thu) by cortana (subscriber, #24596) [Link] (1 responses)

But from that link, I see that udev needs locale_t (which is in POSIX) and %m (which is in uclibc). So what's the problem?

OpenWRT "Barrier Breaker" 14.07 released

Posted Oct 9, 2014 21:27 UTC (Thu) by xslogic (guest, #97478) [Link]

To be honest, no idea...

OpenWRT "Barrier Breaker" 14.07 released

Posted Oct 11, 2014 17:21 UTC (Sat) by BlueLightning (subscriber, #38978) [Link] (2 responses)

FWIW, OpenEmbedded as of recently patches systemd to allow it to be built with uclibc. I can't say I've personally used it in that configuration though.

OpenWRT "Barrier Breaker" 14.07 released

Posted Oct 11, 2014 17:39 UTC (Sat) by rahulsundaram (subscriber, #21946) [Link] (1 responses)

Is there a public repository of the patches?

OpenWRT "Barrier Breaker" 14.07 released

Posted Oct 11, 2014 20:23 UTC (Sat) by lsl (subscriber, #86508) [Link]

Apparently, yes:
http://git.yoctoproject.org/cgit.cgi/poky/tree/meta/recip...

Looks pretty minor, if that's all what's needed. They don't seem to apply systemd-specific patches to uclibc, just some configuration items are tagged as "needed by systemd".

Stuff like "#ifdef HAVE_POSIX_FALLOCATE" (two instances in journal code) probably goes away when uclibc people get around to implementing it. Then there's some use of glibc-specific functions with more or less trivial (for non-threaded/threaded programs, respectively, but easy in any case) workarounds available. There's also "ifdef __UCLIBC__" which is not that pretty, but restricted to two instances (secure_getenv and _SC_PHYS_PAGES missing on uclibc). And that's it, basically.

With involvement from the respective upstreams it should be straightforward to clean these up to make them appropriate for inclusion into systemd proper. I'd first try to get rid of the ifdefs and implement the missing functions inside an extra file that the build system only picks up when needed, but that's a matter of taste somewhat.

Do note that I have no idea if this builds or runs, I just grepped through the repo that was sitting on my disk.

OpenWRT "Barrier Breaker" 14.07 released

Posted Oct 10, 2014 16:32 UTC (Fri) by dlang (guest, #313) [Link] (10 responses)

> I guess I'll be "that guy" who wonders why they wrote their own embedded init system, "procd", instead of using systemd.

the systemd advocates can't have it both ways

They try to say that systemd isn't being forced on people, they can just switch to another distro

But then when a distro doesn't use systemd, it gets attacked.

Either it's being forced on people (which the attacks seem to be doing) or it really should be a choice, and if a distro opts not to use it for ANY reason (including just not wanting to give control of large parts of their distro to the systemd developers, the parts to be defined in the future by the systemd developers), then that should be accepted. If it's the "wrong" choice, people will stop using that distro.

OpenWRT "Barrier Breaker" 14.07 released

Posted Oct 10, 2014 16:56 UTC (Fri) by rahulsundaram (subscriber, #21946) [Link] (9 responses)

> But then when a distro doesn't use systemd, it gets attacked.

I didn't see any attacks.

OpenWRT "Barrier Breaker" 14.07 released

Posted Oct 20, 2014 10:44 UTC (Mon) by oldtomas (guest, #72579) [Link] (8 responses)

> I didn't see any attacks.

This is getting tiring. From the reference post:

[SEJeff] > Perhaps it is because embedded folks tend to over NIH due to wanting an exact feature set and no more?

To me, this translates into "if you don't like systemd, this must because of NIH and not because you have legitimate reasons". How would *you* read that?

This is a pattern exhibited by quite a few systemd proponents, and it does contribute quite a bit to the poisoned discussion. After that, this:

[SEJeff] > I'm legitimately unsure. Any insight from the greater lwn community?

just reads like a rethorical question. Yes, I do see the usual attack.

OpenWRT "Barrier Breaker" 14.07 released

Posted Oct 20, 2014 10:49 UTC (Mon) by rahulsundaram (subscriber, #21946) [Link] (7 responses)

"Legitimately unsure" doesn't read remotely like an attack. We just have to disagree.

OpenWRT "Barrier Breaker" 14.07 released

Posted Oct 20, 2014 11:35 UTC (Mon) by oldtomas (guest, #72579) [Link] (6 responses)

> "Legitimately unsure" doesn't read remotely like an attack. We just have to disagree.

Actually we *do* agree on that. This one only reads like a rethorical question. What does read as an attack is:

[SEJeff] > Perhaps it is because embedded folks tend to over NIH due to wanting an exact feature set and no more?

To me it does read as a sugestive question.

OpenWRT "Barrier Breaker" 14.07 released

Posted Oct 20, 2014 13:15 UTC (Mon) by rahulsundaram (subscriber, #21946) [Link] (5 responses)

Suggestion question? Sure. Hardly unfounded. We got some good replies explaining the situation. I am glad that question was asked.

OpenWRT "Barrier Breaker" 14.07 released

Posted Oct 20, 2014 19:56 UTC (Mon) by oldtomas (guest, #72579) [Link] (4 responses)

Suggestion question?

I meant a suggestive question, sorry for the typo.

I don't grok your answer, sorry. Communication is difficult, it seems.

OpenWRT "Barrier Breaker" 14.07 released

Posted Oct 20, 2014 21:03 UTC (Mon) by rahulsundaram (subscriber, #21946) [Link] (3 responses)

Even if the question was phrased in a less than ideal way, the answers clarified a few things and was a useful discussion to have IMO. Thats all.

OpenWRT "Barrier Breaker" 14.07 released

Posted Oct 21, 2014 7:05 UTC (Tue) by oldtomas (guest, #72579) [Link] (2 responses)

Aha. Got it, thanks. Now to our original topic:

[rahulsundaram]> I didn't see any attacks.

Can you see the attack now?

OpenWRT "Barrier Breaker" 14.07 released

Posted Oct 21, 2014 11:47 UTC (Tue) by rahulsundaram (subscriber, #21946) [Link] (1 responses)

I didn't read what was said as an attack, no.

OpenWRT "Barrier Breaker" 14.07 released

Posted Oct 21, 2014 14:10 UTC (Tue) by oldtomas (guest, #72579) [Link]

> I didn't read what was said as an attack, no.

So let's agree to differ on this one, then.

OpenWRT "Barrier Breaker" 14.07 released

Posted Oct 19, 2014 12:42 UTC (Sun) by jch (guest, #51929) [Link]

> I guess I'll be "that guy" who wonders why they wrote their own embedded init system, "procd", instead of using systemd.

$ git clone https://github.com/fingon/procd.git
$ sloccount procd | grep '^Total Physical'
Total Physical Source Lines of Code (SLOC) = 4,149

OpenWRT "Barrier Breaker" 14.07 released

Posted Oct 4, 2014 3:59 UTC (Sat) by anatolik (guest, #73797) [Link]

Congrats to OpenWRT team!!

It is also sad that 11AC support patches were not included into this release :( https://dev.openwrt.org/ticket/17323

OpenWRT "Barrier Breaker" 14.07 released

Posted Oct 4, 2014 8:27 UTC (Sat) by tomgj (guest, #50537) [Link] (2 responses)

I always think it's a bit of a shame that the WRT54G is no longer supported, since that's the device from which this software originally derived its name. This seems to be due to newer versions not meeting the storage / memory constraints inherent in these devices (4 MiB / 16 MiB). Their hardware page shows the most recent version supported (in the "status" column) as 0.9. Though 10.03.1-rc1 seems to be working on the device I have here. The trouble is, I dare not upgrade it, even to the final 10.03.1 release, without some indication it will work. It would have been really nice if they had kept support for the original devices an ongoing goal.

OpenWRT "Barrier Breaker" 14.07 released

Posted Oct 4, 2014 15:08 UTC (Sat) by raven667 (subscriber, #5198) [Link] (1 responses)

I bricked my WRT54Gv4 trying to update a year or so ago, the recovery mode steps don't seem to work, it may be possible to take the unit apart and get onto a serial port and fiddle with the bootloader but I haven't put in the effort. I still use it though, funny side note, the local switching works even when the firmware is b0rked, so it's a dumb switch now.

OpenWRT "Barrier Breaker" 14.07 released

Posted Oct 5, 2014 10:31 UTC (Sun) by tomgj (guest, #50537) [Link]

funny side note, the local switching works even when the firmware is b0rked, so it's a dumb switch now.
Yes, I found this out the hard way. I was using the VLAN functionality and had connections to two separate networks. When the device failed to boot fully, those networks got bridged together. Whenever I've set one up since, I've checked the network situation for graceful degradation: if the thing turns in to a dumb switch at any time (after a power cycle), does it breach a security barrier between networks, or cause address conflicts between two (now one) networks, or have any other ill effects.

libreCMC vs OpenWrt

Posted Nov 22, 2014 9:35 UTC (Sat) by davi (guest, #18853) [Link]

libreCMC is a distribution of GNU/Linux recently endorsed by the FSF.


Copyright © 2014, Eklektix, Inc.
Comments and public postings are copyrighted by their creators.
Linux is a registered trademark of Linus Torvalds