OpenWrt Forum Archive

Topic: New OpenWrt Kamikaze Release in August

The content of this topic has been archived between 26 Jul 2017 and 26 Jul 2017. There are no obvious gaps in this topic, but there may still be some posts missing at the end.

The OpenWrt project will over the next two months focus on a new Kamikaze release. The official name is to be: OpenWrt Kamikaze 8.08 Release. The planned schedule will take the following shape:

*Last day in July - final release candidate: 8.08 RC-1. This will be a feature freeze, and all changes after this point will be bug fixes.

*Last day in August - final release: OpenWrt Kamikaze 8.08 Release.

OpenWrt Kamikaze 8.08 Release will, amongst other, focus on bringing the following features:

- Firewall rewrite
- Broadcom 47xx running reliably with the new Kernel, not including wifi
- IMQ and Traffic shaping tested with newer kernels, especially 2.6.25
- Sysupgrade for more platforms (x86 is tested again)
- The new web interface (LuCI, Lua Configuration Interface)
- Full support for new platforms and targets
- Attention towards the integration of security updates
- Package maintaining and updates between releases

The OpenWrt Team

glp wrote:

- Firewall rewrite
- The new web interface (LuCI, Lua Configuration Interface)

I don't see any related commits in trac so far.

forum2008 wrote:

I don't see any related commits in trac so far.

Be patient. Such things take a while, and won't be committed until they are somewhat stable and usable. Besides, the RC1 release is still almost a month ahead.
I rather have one or two weeks longer the "old" webif/firewall than having to deal with brokenness of the trunk (which is almost inevitable on major changes anyways wink.

You know how fast a month is over? A month is nothing.

Btw. I'm also still missing the UCI feature from nbd to make access to unnamed config sections easier from command line.

(Last edited by forum2008 on 1 Jul 2008, 10:58)

glp wrote:

- Firewall rewrite
- The new web interface (LuCI, Lua Configuration Interface)

I think you should pay attention to the documentation of all these new features.

Why is the old firewall broken and how will you fix it? Could you give more information about the new web interface? I'm especially interested in this subject as I would like to contribute.

glp wrote:

- Attention towards the integration of security updates
- Package maintaining and updates between releases

Thanks, it looks great, I was starting to be a litlte bit worried for it has been almost a year since last release. Especially those two points I quoted looks very promising.

glp wrote:

OpenWrt Kamikaze 8.08 Release will, amongst other, focus on bringing the following features:

- IMQ and Traffic shaping tested with newer kernels, especially 2.6.25

The OpenWrt Team

I am curious about why the emphasis is on IMQ for 2.6 kernels when the kernel people seem to want people to use IFB instead?

So you're improving the 2.6 kernal issues, but I cant run 2.6 as I have Broadcom based ASUS WL-500gp.

I would have thought 2.6 broadcom wifi would be top of list. Unfortunately it seems most of the developers who had Broadcom, have upgraded their WiFi cards to Atheros.

Will this continued to be ignored?

BatGnat wrote:

So you're improving the 2.6 kernal issues, but I cant run 2.6 as I have Broadcom based ASUS WL-500gp.

I would have thought 2.6 broadcom wifi would be top of list. Unfortunately it seems most of the developers who had Broadcom, have upgraded their WiFi cards to Atheros.

Will this continued to be ignored?

It's very hard to write drivers for undocumented hardware.  Don't blame the devs, blame Broadcom.  Besides, Atheros cards generally have a larger feature-set and improved range.  For <$20 on eBay, I'd say it's worth it.


For my part, all I have to say is SWEET!  I was beginning to think this project was dead.  Seg-fault fixes on mips will be nice.  And for future developmet, my vote goes to integration with OpenEmbedded.

lamawithonel wrote:

For my part, all I have to say is SWEET!  I was beginning to think this project was dead.  Seg-fault fixes on mips will be nice.  And for future developmet, my vote goes to integration with OpenEmbedded.

Sometimes I'm thinking the same...

forum2008 wrote:
lamawithonel wrote:

For my part, all I have to say is SWEET!  I was beginning to think this project was dead.  Seg-fault fixes on mips will be nice.  And for future developmet, my vote goes to integration with OpenEmbedded.

Sometimes I'm thinking the same...

You both have never used OE and/or have wrong ideas about what's OpenWrt then.

I dont know much about OE but OpenWrt for me is more than "Wireless Freedom".
I am running more than 50 machines (in different projects) in various hardware with Openwrt. None of them has wireless interface.
For me OpenWrt, is a flexible embedded linux platform. Provides toolchain, buildroot and common interface for various architectures and devices.
Keep going good work!

(Last edited by Slammer on 12 Jul 2008, 00:14)

BatGnat wrote:

So you're improving the 2.6 kernal issues, but I cant run 2.6 as I have Broadcom based ASUS WL-500gp.

I would have thought 2.6 broadcom wifi would be top of list. Unfortunately it seems most of the developers who had Broadcom, have upgraded their WiFi cards to Atheros.

Will this continued to be ignored?

openwrt is an open source project.  if you need a feature that is not there, you can simply try to do it your self or hire someone to do it for you (developers have to make a living too) and then kindly contribute it back to the community in order for the project to keep evolving which is a win win case for all.  We are grateful to developers who put so much effort, time and dedication on this project for free.  if we need more out of this project, we just have to contribute more to it instead of criticizing about missing features or unresolved bugs.

openwrt is not a product, it's a power tool!

(Last edited by acoul on 12 Jul 2008, 06:18)

BatGnat wrote:

Will this continued to be ignored?

I'm don't understand. What is being ignored? Are you currently aware of a released, stable, working linux 2.6 mac80211 / nl80211 / bcm43xx / bcm ap-mode software stack somewhere that the rest of us have missed (or are "ignoring")?

You might also like to see http://article.gmane.org/gmane.linux.ke … eral/17564 for a reasonable response to someone else's similar enquiry (but then again, you might not).

I post a simple question. and you go and insult me, very mature.  The people in charge of the project seem to moving away from a solution, and are more inclined to tell people to purchase a new wifi card.

I'm don't understand. What is being ignored? Are you currently aware of a released, stable, working linux 2.6 mac80211 / nl80211 / bcm43xx / bcm ap-mode software stack somewhere that the rest of us have missed (or are "ignoring")?

But that is my point exactly, there is no release, nor from what I have seen, no future plans to do so.  I understand that there are issues with the Broadcom WiFi, but it seems the official answer is, "get an atheros card".

I understand that a lot of the programming is done in peoples spare time, and I do appreciate that. But so far all I have gotten in response to my question is an insult, and do it yourself. Not very helpful.

BatGnat wrote:

I post a simple question. and you go and insult me, very mature.  The people in charge of the project seem to moving away from a solution, and are more inclined to tell people to purchase a new wifi card.

I'm don't understand. What is being ignored? Are you currently aware of a released, stable, working linux 2.6 mac80211 / nl80211 / bcm43xx / bcm ap-mode software stack somewhere that the rest of us have missed (or are "ignoring")?

But that is my point exactly, there is no release, nor from what I have seen, no future plans to do so.  I understand that there are issues with the Broadcom WiFi, but it seems the official answer is, "get an atheros card".

I understand that a lot of the programming is done in peoples spare time, and I do appreciate that. But so far all I have gotten in response to my question is an insult, and do it yourself. Not very helpful.

I fear you have greatly misunderstood both my post and the very nature of openwrt development (and before you launch into a defensive tirade, I'm not saying that to offend you - I'm saying it because it's important that you do understand. Once you have understood, you will realize that no-one is _ignoring_ anything).

Openwrt is (in my opinion) like any other distribution. Like debian, like fedora, like any of those listed on distrowatch. Their purpose is to accumulate individual packages into a coherent whole, which then becomes meaningful to end users such as ourselves. Their purpose is most definitely NOT to develop one of those individual packages. Oh, and by the way, I am already aware that the distinction between the two (distribution vs. package development) is NOT a well-defined line. It is, instead, a blurred grey-area. However, my point still stands, and that is that openwrt is primarily a distribution. Therefore, the people who develop openwrt _must_ limit the amount of time they spend on upstream development - if they don't, their distribution (openwrt) _will_ suffer.

Now let's look at the other side. Where does openwrt get its packages from? Well, from upstream devs of course. And where's the best place to find out about upsteam dev for the linux 2.6 wireless susbsytem? In the case of the broadcom drivers (and many others, come to that), it's the linux wireless mailing list available at linux-wireless@vger.kernel.org. If you would like to search their archives (and follow, and research, all of the bits of it that don't make sense at first), it shouldn't take you more than about 45 hours to realize that no-one in upstream dev-land has a "released, stable, working linux 2.6 mac80211 / nl80211 / bcm43xx / bcm ap-mode software stack" (I've quoted 45 hours since that's about the amount of time that I've spent looking in the last week, and I suspect that it would take us approximately the same amount to time to do the same job).

So, to recap, upstream (for linux wireless) doesn't have a "released, stable, working linux 2.6 mac80211 / nl80211 / bcm43xx / bcm ap-mode software stack" _despite_ the fact that it's all they do (that's their specialist subject) _and_ there's a load of them doing it. I personally conclude, therefore, that it's _difficult_. Not that they (upstream) are "ignoring" anything.

Back to the openwrt devs. The openwrt devs can see that upstream haven't yet produced something that works well enough for the upstream devs to use. So what should they (the openwrt devs) do? Drop their distribution and become linux-wireless devs instead? Of course not. They (the openwrt devs) might be able to help upstream out with a patch or two, but they (the openwrt devs) are developing a distribution, _not_ the wireless susbsystem/drivers. The openwrt devs have got a whole load of other questions to worry about (such as firmware flashing, boot loaders, which package manager to use, etc, etc, etc, the list goes on). If the openwrt devs have in any way suggested that your best chance for working wireless is to get an atheros card, it's no more than a pragmatic reflection of the current state of upstream development.

If I am to explain myself fully, I myself found your original post ("Will this continue to be ignored") to be highly offensive. And no, I'm not just saying that as some form of retaliation. I'm saying that because for me, it implies some form of "not caring", which I believe could not be _further_ from the truth. The simple fact of the matter is that with such an incredibly broad subject as computers, people _must_ specialize in order to achieve anything - the openwrt people have chosen to specialize on their distribution, while the linux-wireless people have chosen to specialize on wireless subsystem/infrastructure design and development. There are not enough resources in a single lifetime to do both. In fact, it's even stronger than that. There aren't enough resources in several people's lives to make the wireless subsystem work properly (yet).

If I have offended you, I will apologize. However I'm still confused. You still haven't told me who (openwrt or upstream) is ignoring something. However, this is my last post on the subject. I've already spent too much of my time trying to explain how "nothing is being ignored" rather than fixing bugs / implementing missing bits of functionality in open source sofware. That, is irony.

Jaime

BatGnat: Perhaps there's a bit of a miscommunication, but your comment seemed to me to be quite insulting to the OpenWrt developers, and not just "a simple question."  In fact your comment seems to me to be more offensive than jaime's reply to you.

The way I see it (and perhaps I'm wrong) is that Broadcom has been very unhelpful wrt. allowing/helping people to write open drivers for their hardware.  Despite this, people are working on the problem and I am sure a future version of OpenWrt using a 2.6 kernel will work well with this hardware.  The reason people say you should get an Atheros card is probably one of the following:  Atheros is much more open to free and open drivers and have even hired an Open Source Atheros driver developer.  Atheros hardware is better supported and so if you want a stable wireless platform NOW, then Atheros is probably a better choice than Broadcom.

If this post has offended anybody, I did not intend it to.

Can this be moved out of the news topic? There are people who just like to read news posts not complete flamewars in their rss reader. Much appriciated hmm

glp wrote:

- Broadcom 47xx running reliably with the new Kernel, not including wifi
- Attention towards the integration of security updates
- Package maintaining and updates between releases

I have a WRT54GL (and GS) with the broadcom chipset.  Will this still be supported with the 2.4 kernel in 8.08?
I'm looking for the security updates and package updates between releases for my WRT54's and would
hate to miss out on the new release (OpenWRT makes these great little devices).

Thanks.

jgmcc wrote:
glp wrote:

- Broadcom 47xx running reliably with the new Kernel, not including wifi
- Attention towards the integration of security updates
- Package maintaining and updates between releases

I have a WRT54GL (and GS) with the broadcom chipset.  Will this still be supported with the 2.4 kernel in 8.08?

yes

Hi,

Just wondering, is the txpower issue solved in kamikaze for broadcom? (2.4 kernel, need wifi)
Since the new wireless driver it seems I'm not able to increase txpower. (on white russian now)

Kind regards,
B

What about these drivers ?

http://linuxwireless.org/en/users/Drivers/b43

It's included in the kernel since 2.6.24 if I'm right. Does it support the broadcom chipset included in our devices ?

If it does, is it stable enough to be included in kamikaze ?

PS : I'm just trying to understand the situation here

Keep up the good work

Pleeeease move this bloody discussion to another thread!!! mad

Or better lock the topic!

Hello,

routers have an internal connection cable - wifi; someone knows if is possible "break" this connection in order that cable works on the one hand and on the other hand wifi?

I have two routers connected with a switch and I want to send traffic of a router to the other router for wifi but the traffic is sending for the cable ethernet.

Help me please.

Thank you.

Kaloz wrote:
forum2008 wrote:
lamawithonel wrote:

...for future developmet, my vote goes to integration with OpenEmbedded.

Sometimes I'm thinking the same...

You both have never used OE and/or have wrong ideas about what's OpenWrt then.

OE is more-or-less just a fancy source-based package management system for embedded GNU/Linux environments.
OpenWRT is more-or-less just a basic embedded GNU/Linux environment.
I don't know where you plan to take the project, but as-is, the two practically scream, "Marry us!"

edit: snipped irrelevant text

(Last edited by lamawithonel on 22 Jul 2008, 01:05)