[Cerowrt-devel] cerowrt-3.10.44-5 released

[Cerowrt-devel] cerowrt-3.10.44-5 released

[Dave Taht]

+ resync with openwrt head
   updates to iw, mac80211
   various routing table fixes in netifd

+ dnsmasq 2.71 with mini-gmp and libnettle mainlined
   also moved into procd for better automagic restart
   (this leaves babel as the only major daemon not managed by procd. sigh)

- totally untested (I tested -4 pretty thoroughly though)
  I won't have time for this personally til later this week.
- still no answer for bug 442 - I do get bad things to happen on a
ubnt device now
- left off on the homewrt integration for now

I need to get around to submitting sqm upstream again, but am busy on
other tasks.
IETF is coming up, also.

-- 
Dave Täht

NSFW: https://w2.eff.org/Censorship/Internet_censorship_bills/russell_0296_indecent.article

Re: [Cerowrt-devel] cerowrt-3.10.44-5 released

[Stephen Hemminger]

On Tue, 24 Jun 2014 15:33:59 -0700
Dave Taht <dave.taht@gmail.com> wrote:

> + resync with openwrt head
>    updates to iw, mac80211
>    various routing table fixes in netifd
> 
> + dnsmasq 2.71 with mini-gmp and libnettle mainlined
>    also moved into procd for better automagic restart
>    (this leaves babel as the only major daemon not managed by procd. sigh)
> 
> - totally untested (I tested -4 pretty thoroughly though)
>   I won't have time for this personally til later this week.
> - still no answer for bug 442 - I do get bad things to happen on a
> ubnt device now
> - left off on the homewrt integration for now
> 
> I need to get around to submitting sqm upstream again, but am busy on
> other tasks.
> IETF is coming up, also.
> 

Is this going to be any more stable? I almost about to give up
on cerowrt for home use and go to dd-wrt or something else.
Wireless still crashes once a week with 3.10.36-6

Re: [Cerowrt-devel] cerowrt-3.10.44-5 released

[Dave Taht]

On Tue, Jun 24, 2014 at 4:09 PM, Stephen Hemminger
<stephen@networkplumber.org> wrote:
> On Tue, 24 Jun 2014 15:33:59 -0700
> Dave Taht <dave.taht@gmail.com> wrote:
>
>> + resync with openwrt head
>>    updates to iw, mac80211
>>    various routing table fixes in netifd
>>
>> + dnsmasq 2.71 with mini-gmp and libnettle mainlined
>>    also moved into procd for better automagic restart
>>    (this leaves babel as the only major daemon not managed by procd. sigh)
>>
>> - totally untested (I tested -4 pretty thoroughly though)
>>   I won't have time for this personally til later this week.
>> - still no answer for bug 442 - I do get bad things to happen on a
>> ubnt device now
>> - left off on the homewrt integration for now
>>
>> I need to get around to submitting sqm upstream again, but am busy on
>> other tasks.
>> IETF is coming up, also.
>>
>
> Is this going to be any more stable? I almost about to give up
> on cerowrt for home use and go to dd-wrt or something else.
> Wireless still crashes once a week with 3.10.36-6

I think bug 442 is universal across openwrt derived versions for the
last 5+ months. There is (no longer) any real difference between the
network stacks of cero and openwrt. I do think cero triggers it more
often by using adhoc mode, multiple wifi interfaces, and routing, and
I'd suggest to those using it to disable the guest and adhoc
interfaces entirely and see what happens. (definately kill adhoc
unless you are using it)

otherwise...

I too, despair. What I have been doing is developing means to saturate
links for long periods, take big packet captures, and for all that
I've only seen that bug 3 times, all 3 uncaptured. Yet others get it
all the time.

It is costing users hair, me time, and money I cannot spend on this
project, and hurting other efforts. I would have released a "stable"
version months ago if it wasn't for this.

I'll miss you if you leave! but if you had a stabler version that you
liked, you can regress to that - or by all means! give modern dd-wrt a
shot - it's got most of the core features cero has by now, and I
haven't tried it lately.

-- 
Dave Täht

NSFW: https://w2.eff.org/Censorship/Internet_censorship_bills/russell_0296_indecent.article

Re: [Cerowrt-devel] cerowrt-3.10.44-5 released

[Dave Taht]

sigh. I gotta admit that an install was not encouraging for basic
functionality. My *principal* need for doing a new build is to keep
trying to find and fix bug 442. That's it. All the other stuff
changing out underneath is a distraction.

I really need to find more time to integrate new stuff and test than I
have at the moment. A ton of things are landing in openwrt mainline
done slightly differently to how cero does them.

0) yes, it does basically "work"
1) it looks like the ntp monitoring thing toke did is not working
and/or we're running the wrong ntp now
    but dnsmasq does not run with timechecks enabled by default, so we
work with invalid time
    it does look like two versions are running, too...
2) mdnsd got enabled again (my fault, I guess) and conflicts with avahi
3) the new version of miniupnpd is doing new and unusual things like
listening on port 5000...
4) it's firing off netserver all of a sudden by default instead of out
of xinetd, where it is firewalled

All these bugs are survivable (disable mdnsd, firewall off port 5000,
delete /etc/init.d/netserver) but I'm going to cut another release
tomorrow anyway. I don't know what's going on on the miniupnpd front:

Wed Jun 25 01:55:40 2014 daemon.notice miniupnpd[2706]: shutting down MiniUPnPd
Wed Jun 25 01:55:41 2014 daemon.notice miniupnpd[2909]: HTTP listening
on port 5000
Wed Jun 25 01:55:41 2014 daemon.notice miniupnpd[2909]: HTTP IPv6
address given to control points :
[2001:470:8236:261:2ac6:8eff:febb:9ff1]
Wed Jun 25 01:55:41 2014 daemon.err miniupnpd[2909]:
OpenAndConfNATPMPSocket: bind(): Address already in use
Wed Jun 25 01:55:41 2014 daemon.err miniupnpd[2909]: Failed to open
sockets for NAT-PMP/PCP.
Wed Jun 25 01:55:47 2014 daemon.notice miniupnpd[2909]: shutting down MiniUPnPd
Wed Jun 25 01:55:48 2014 daemon.notice miniupnpd[3415]: HTTP listening
on port 5000
Wed Jun 25 01:55:48 2014 daemon.notice miniupnpd[3415]: HTTP IPv6
address given to control points :
[2001:470:8236:261:2ac6:8eff:febb:9ff1]
Wed Jun 25 01:55:48 2014 daemon.err miniupnpd[3415]:
OpenAndConfNATPMPSocket: bind(): Address already in use
Wed Jun 25 01:55:48 2014 daemon.err miniupnpd[3415]: Failed to open
sockets for NAT-PMP/PCP.


On Tue, Jun 24, 2014 at 3:33 PM, Dave Taht <dave.taht@gmail.com> wrote:
> + resync with openwrt head
>    updates to iw, mac80211
>    various routing table fixes in netifd
>
> + dnsmasq 2.71 with mini-gmp and libnettle mainlined
>    also moved into procd for better automagic restart
>    (this leaves babel as the only major daemon not managed by procd. sigh)
>
> - totally untested (I tested -4 pretty thoroughly though)
>   I won't have time for this personally til later this week.
> - still no answer for bug 442 - I do get bad things to happen on a
> ubnt device now
> - left off on the homewrt integration for now
>
> I need to get around to submitting sqm upstream again, but am busy on
> other tasks.
> IETF is coming up, also.
>
> --
> Dave Täht
>
> NSFW: https://w2.eff.org/Censorship/Internet_censorship_bills/russell_0296_indecent.article



-- 
Dave Täht

NSFW: https://w2.eff.org/Censorship/Internet_censorship_bills/russell_0296_indecent.article