Re: [OpenWrt-Devel] [PATCH] use NTP server received via DHCP
Just a side note while you are working in this area. all the documentation for udhcpc (including it's default config settings) says that it uses one default file path, but when it's run that path gets overridden on the command line and a different file is used instead. The default in the config should be changed to what's actually used and the extra copy should be eliminated. David Lang ___ openwrt-devel mailing list openwrt-devel@lists.openwrt.org https://lists.openwrt.org/cgi-bin/mailman/listinfo/openwrt-devel
Re: [OpenWrt-Devel] Question about custom script at init on Trunk.-
Since it's just two files that you are wanting to override for your images, go
to the top level directory (where you do your make menuconfig) and create the
directory files and put the files that you want to override in there (in the
appropriate directory, so to override /etc/config/wireless, make the file
files/etc/config/wireless)
much easier than maintaining a different version of the package just to change a
couple files.
David Lang
On Tue, 26 Jan 2016, Guillermo Javier Nardoni wrote:
Date: Tue, 26 Jan 2016 21:19:54 -0300
From: Guillermo Javier Nardoni
To: openwrt-devel@lists.openwrt.org, guille...@geryon.com.ar
Subject: [OpenWrt-Devel] Question about custom script at init on Trunk.-
Hello everyone,
At first, I should say that I?ve been away a long time from openwrt-devel
mailing list.
I was using Barrier Braker 14.07 to build our custom firmware, usually for
TP-Link WR740/741 v3/v4, but recently we have received TP-Link WR740 v5
(a.k.a. 741 v5) so I figured out that Barrier Braker don?t have support for
it, at least not native one.
Googling I found a solution which I?ve flashed it first with a stripped
custom firmware found on Internet and then, via console, do a beautifull
reflash with WR741 v4 firmware (build by us) and everything Works perfect!.
Since I?m not the one who flashes the devices I did a new git from trunk
versión of OpenWrt (CHAOS CALMER (15.05)) because it has native support for
TP-Link TL-WR740N versión 5 but when I looked into the code I?ve figured
out, again, that the code changed a lot since the BB 14.07 and I don?t know
how to adapt the custom configurations.
The reason I changed those 2 files it is because of the users sometimes (i
mean always) press the reset button and it won?t allow us to remotely
recofigure the device and to use one port of the switch to bridge it with
the WAN port to use them just because we have ATAs devices that don?t
support NAT etc etc.
When a user presses the reset button, it resets the device and configure it
with a standard configuration which let the user to have Telephone (VoIP),
Internet, Wifi and we can Access remotely to reconfigure it as I said
before.
On BB 14.07 I?ve just modified 2 files:
./package/base-files/files/lib/functions/uci-defaults.sh
??.
ucidef_add_switch() {
local name=$1
local reset=$2
local enable=$3
uci batch
Re: [OpenWrt-Devel] Question about custom script at init on Trunk.-
-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 Eric Schultz wrote: > Guillermo, > > It's always best to use a custom package if you can. Modifying > upstream leads to problems like this. > > I don't think you'd need those changes in uci-defaults.sh. > Instead, I think you should put your changes in your own > uci-default, like 02_network, but your own. Use a number after > 02, like 03_custom and put it in /etc/uci-defaults. It should > run on first boot. As long as this uci-default script completes > successfully, I think the change should occur and the script > itself should be deleted so it doesn't run again. > Along these lines, I've tried putting in something like 80_mystuff, where I want to set some configurations options for various packages that depend on the hardware they're running on. However, the uci-defaults files are run in alphabetical order, so while 80_mystuff might come after the 01_leds, 02_network stuff from core OpenWrt, the uci-defaults files from packages normally don't seem to have a numerical prefix. This then means that my 80_mystuff script tries to set a config option in a file that doesn't exist yet. Now, yes, I can have my 80_mystuff script go and try and create the package config files it might want to modify, but really, I want "mystuff" to run _last_ or at least, after all the packages have run. Is there any better system for this other than just naming my script z_mystuff instead of 80_mystuff? Is that expected to be "good enough" ? I'm trying this, because I've _been_ trying to use packages instead of modifying upstream, but when I want my package to do different things for different targets, it's seemed easier to hook into the existing target code, rather than trying to copy target detection and config into a package Config.in options Sincerely, Karl Palsson -BEGIN PGP SIGNATURE- Version: GnuPG v1.4.11 (GNU/Linux) iQIcBAEBAgAGBQJWqJbVAAoJEBmotQ/U1cr29HwP/3I3TemR4suETncwKdssUHpt 7RP4+SPgSZPENNub4xlViSYz8aB69AQZ+wgwzQZR8p3UbzCtc02BJoa0ikDx8H4F EvYap9GhQ3hCKznTD9GbZCvGLaqAHLn9wmMlPzPcrqQeMQICdsiXhg1c410tICiO JBWwoZvg5bbBStfyh1vX4kvNX3tVMFYgJePSL8r3o/BgX9+q9jCIaZbae2A7Q4xU 1fVuxPXJLlxRDGdA/woFGFX1iokzaZL+yP6jY9+TW/VtA6MiLNBxnYn5X5C78i0Q L9of5vOYkZ4s9FvgsKV99x3Bdqkmnbza0lRt8nGsOKtvWd3CH/sTBX9mpXZKQPsu pSfzgbv3rz5Bmix5HG1vuEqbrA+i6VE7pyS5YstXfIu8GWCtjAlUGSjxLQolrl2x +yaaU7EEIuqa855mhw6dQ5KL5CJQPITiyohsOsB0VsjSaLP6VmUqCPGohNlwK6Zr YslCozrzZf507mieYz5N5F4jRXoLMnO8UuVuBzr0ImZzw7Zy0zZaJm/yLNOaBUvM NGiIFBlXHRsutcHMA4PBwGAn/y3nSC4bZ8V8ip2mmJlNgJpb9mSBDfngahv5n7X5 OboDzgZRWJxAlZgjH06vQNFhP6SuCv7i4pSVifnwydXALUakRmv7A9lk8elR1meE 1Ops4PWevuKM7VURfgaY =Wijj -END PGP SIGNATURE- ___ openwrt-devel mailing list openwrt-devel@lists.openwrt.org https://lists.openwrt.org/cgi-bin/mailman/listinfo/openwrt-devel
Re: [OpenWrt-Devel] Question about custom script at init on Trunk.-
On Wed, 27 Jan 2016, Karl Palsson wrote: -BEGIN PGP SIGNED MESSAGE- Hash: SHA1 Eric Schultz wrote: Guillermo, It's always best to use a custom package if you can. Modifying upstream leads to problems like this. I don't think you'd need those changes in uci-defaults.sh. Instead, I think you should put your changes in your own uci-default, like 02_network, but your own. Use a number after 02, like 03_custom and put it in /etc/uci-defaults. It should run on first boot. As long as this uci-default script completes successfully, I think the change should occur and the script itself should be deleted so it doesn't run again. Along these lines, I've tried putting in something like 80_mystuff, where I want to set some configurations options for various packages that depend on the hardware they're running on. However, the uci-defaults files are run in alphabetical order, so while 80_mystuff might come after the 01_leds, 02_network stuff from core OpenWrt, the uci-defaults files from packages normally don't seem to have a numerical prefix. This then means that my 80_mystuff script tries to set a config option in a file that doesn't exist yet. Now, yes, I can have my 80_mystuff script go and try and create the package config files it might want to modify, but really, I want "mystuff" to run _last_ or at least, after all the packages have run. Is there any better system for this other than just naming my script z_mystuff instead of 80_mystuff? Is that expected to be "good enough" ? I'm trying this, because I've _been_ trying to use packages instead of modifying upstream, but when I want my package to do different things for different targets, it's seemed easier to hook into the existing target code, rather than trying to copy target detection and config into a package Config.in options typically people use 99stuff and figure that the order of 99s doesn't really matter much the other thing is that I believe that rc.local gets run out of the very last of the numbered items, so you can put stuff there. David Lang___ openwrt-devel mailing list openwrt-devel@lists.openwrt.org https://lists.openwrt.org/cgi-bin/mailman/listinfo/openwrt-devel ___ openwrt-devel mailing list openwrt-devel@lists.openwrt.org https://lists.openwrt.org/cgi-bin/mailman/listinfo/openwrt-devel
Re: [OpenWrt-Devel] Question about custom script at init on Trunk.-
On Wed, 27 Jan 2016, Karl Palsson wrote: Along these lines, I've tried putting in something like 80_mystuff, where I want to set some configurations options for various packages that depend on the hardware they're running on. However, the uci-defaults files are run in alphabetical order, so while 80_mystuff might come after the 01_leds, 02_network stuff from core OpenWrt, the uci-defaults files from packages normally don't seem to have a numerical prefix. This then means that my 80_mystuff script tries to set a config option in a file that doesn't exist yet. Now, yes, I can have my 80_mystuff script go and try and create the package config files it might want to modify, but really, I want "mystuff" to run _last_ or at least, after all the packages have run. Is there any better system for this other than just naming my script z_mystuff instead of 80_mystuff? Is that expected to be "good enough" ? what I do is to flash the base image on a device, and then tweak it until I get things working and then create a new image including that file. If I need different versions for different devices, I work on each one separately and then figure a way to detect the different conditions. Another example from Scale, We use Raspberry Pi computers to run self-service registration stations and to provide digital signs when connected to large TVs. I use a single image that checks to see if it has a mouse attached, if so it's a registration station, otherwise it's a sign and then the appropriate thing gets launched. you can use IP addresses/ranges, DHCP parameters, holding a button, or many other things to detect the different conditions. David Lang___ openwrt-devel mailing list openwrt-devel@lists.openwrt.org https://lists.openwrt.org/cgi-bin/mailman/listinfo/openwrt-devel ___ openwrt-devel mailing list openwrt-devel@lists.openwrt.org https://lists.openwrt.org/cgi-bin/mailman/listinfo/openwrt-devel
[OpenWrt-Devel] [malta] CC 15.05 backport request -r46269 malta-add-harddisk-support
Hi all, >From following the list in the past few weeks I believe there may be a CC 15.05.1 planned for release? If so could it be possible to backport the trivial patch to add harddisk support for the malta platform? I tested it yesterday against latest update to CC 15.05 and it worked flawlessly. Commit info: 76bfcd46bd55263947461e40bd02047922755c72 Thnx in advance! - vin ___ openwrt-devel mailing list openwrt-devel@lists.openwrt.org https://lists.openwrt.org/cgi-bin/mailman/listinfo/openwrt-devel
Re: [OpenWrt-Devel] Question about custom script at init on Trunk.-
-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 David Lang wrote: > > > > Now, yes, I can have my 80_mystuff script go and try and create > > the package config files it might want to modify, but really, I > > want "mystuff" to run _last_ or at least, after all the packages > > have run. Is there any better system for this other than just > > naming my script z_mystuff instead of 80_mystuff? Is that > > expected to be "good enough" ? > > > > I'm trying this, because I've _been_ trying to use packages > > instead of modifying upstream, but when I want my package to do > > different things for different targets, it's seemed easier to > > hook into the existing target code, rather than trying to copy > > target detection and config into a package Config.in options > > typically people use 99stuff and figure that the order of 99s > doesn't really matter much > Well, the thing is, 99 isn't last. that's what I was getting at. $ ls -l build_dir/target-mips_34kc_uClibc-0.9.33.2/root-ar71xx/etc/uci-defaults/ total 140 - -rw-r--r-- 1 karlp karlp 128 Jan 26 13:32 00_uhttpd_ubus - -rwxrwxr-x 1 karlp karlp 17158 Jan 20 17:20 01_leds -rwxrwxr-x 1 karlp karlp 10391 Jan 20 17:22 02_network -rwxrwxr-x 1 karlp karlp 1673 Sep 25 11:01 03_network-switchX-migration -rwxrwxr-x 1 karlp karlp 351 Sep 25 11:01 03_network-vlan-migration -rwxrwxr-x 1 karlp karlp 1244 Sep 25 11:01 04_led_migration -rwxrwxr-x 1 karlp karlp 232 Sep 25 11:01 09_fix-seama-header -rw-rw-r-- 1 karlp karlp 193 Sep 25 11:01 09_fix-trx-header -rw-r--r-- 1 karlp karlp 74 Jan 27 10:56 10-fstab -rw-rw-r-- 1 karlp karlp 418 Sep 25 11:01 10_migrate-shadow -rw-rw-r-- 1 karlp karlp 450 Sep 25 11:01 11_migrate-sysctl -rw-rw-r-- 1 karlp karlp 377 Sep 25 11:01 12_network-generate-ula -rw-rw-r-- 1 karlp karlp 377 Sep 25 11:01 99_stuff_here -rw-rw-r-- 1 karlp karlp 211 Aug 21 12:40 agent_etactica -rwxr-xr-x 1 karlp karlp 83 Jan 11 15:05 luci-i18n-english -rwxr-xr-x 1 karlp karlp 1285 Jan 26 13:33 mosquitto.uci-defaults -rw-rw-r-- 1 karlp karlp 479 Aug 21 12:40 mpacker -rwxrwxr-x 1 karlp karlp 291 Aug 21 12:40 mq_delta_gadget - -rwxrwxr-x 1 karlp karlp 970 Aug 21 12:40 mrelay -rwxr-xr-x 1 karlp karlp 300 Jan 26 13:29 odhcpd.defaults -rwxrwxr-x 1 karlp karlp 224 Sep 24 14:24 snmpd Here's what you really have in a new image. Putting things in "99_stuff" just won't work very well. All the package provided uci-defaults files, which are the ones I want to extend/augment with my own files have to come later. They are run at /etc/rc.d/S10boot, -BEGIN PGP SIGNATURE- Version: GnuPG v1.4.11 (GNU/Linux) iQIbBAEBAgAGBQJWqLMzAAoJEBmotQ/U1cr2bW8P93PKOyhaBckgfSaOgeDCI23A T7Tx9N3pRkVoZ9D+DAU+HdmMdWUB8Wt59Qzn37OYSxGyieYTflkgLgCluTX5Mkxg Hew64Vq3TVlZIcoBGiDbDE+YiNBqQ26bO3BTeCtuq6vIdzl2fX9LO8Hv0sKCFXpt fgb/B3eNaUQdrFeZZNj+Ay3/ZxUJz8ofYp4BcV9rY9NF1HdesLrdJ9MIcIfPRWrc cyyTsukrzfMilkeujOqjsNrJdlteU1+jJ844jnMGsOjw9YhB3k+g7sF+z5B2yeOL EmBy5CJo4sDZYJOKdeA5w9LbVjPhS1cqpfRj6u+CEWVJHvHwsmUDR9OroVzfyUw/ GZHzsOwrTClaRrvgKV8Z/sZTCOFD+8STZs+kucD46JBkmUDAmnkMQrAT0F18lQXm PEDWWn/aF9IoWXEOjBbOQNrGhzcjF1emoGpgPgD29o5y30gr0qW/kpFPSkUu0wup 3A1iyT5vwWxbP89Wd6n5PslE0K3e3om6TbNPNHd1s2H1l5dac4LVODfb5sHAkyCs 8YYZzPD7YFVZeN5aNpFmgfzQtPoop3sR/hZb/Ay44P1Y1kAAzRXcglF3iXm7j691 amd4KMMv+NVbWyU2p65ZfKNFCS7pDjfh4MbE6UZIlYkjDipqJRx8ZRQD6Qvmq98k yLIkjtpNL4eW/q997rE= =eSyW -END PGP SIGNATURE- ___ openwrt-devel mailing list openwrt-devel@lists.openwrt.org https://lists.openwrt.org/cgi-bin/mailman/listinfo/openwrt-devel
Re: [OpenWrt-Devel] Question about custom script at init on Trunk.-
On Wed, 27 Jan 2016, Karl Palsson wrote: David Lang wrote: Now, yes, I can have my 80_mystuff script go and try and create the package config files it might want to modify, but really, I want "mystuff" to run _last_ or at least, after all the packages have run. Is there any better system for this other than just naming my script z_mystuff instead of 80_mystuff? Is that expected to be "good enough" ? I'm trying this, because I've _been_ trying to use packages instead of modifying upstream, but when I want my package to do different things for different targets, it's seemed easier to hook into the existing target code, rather than trying to copy target detection and config into a package Config.in options typically people use 99stuff and figure that the order of 99s doesn't really matter much Well, the thing is, 99 isn't last. that's what I was getting at. $ ls -l build_dir/target-mips_34kc_uClibc-0.9.33.2/root-ar71xx/etc/uci-defaults/ total 140 - -rw-r--r-- 1 karlp karlp 128 Jan 26 13:32 00_uhttpd_ubus - -rwxrwxr-x 1 karlp karlp 17158 Jan 20 17:20 01_leds -rwxrwxr-x 1 karlp karlp 10391 Jan 20 17:22 02_network -rwxrwxr-x 1 karlp karlp 1673 Sep 25 11:01 03_network-switchX-migration -rwxrwxr-x 1 karlp karlp 351 Sep 25 11:01 03_network-vlan-migration -rwxrwxr-x 1 karlp karlp 1244 Sep 25 11:01 04_led_migration -rwxrwxr-x 1 karlp karlp 232 Sep 25 11:01 09_fix-seama-header -rw-rw-r-- 1 karlp karlp 193 Sep 25 11:01 09_fix-trx-header -rw-r--r-- 1 karlp karlp 74 Jan 27 10:56 10-fstab -rw-rw-r-- 1 karlp karlp 418 Sep 25 11:01 10_migrate-shadow -rw-rw-r-- 1 karlp karlp 450 Sep 25 11:01 11_migrate-sysctl -rw-rw-r-- 1 karlp karlp 377 Sep 25 11:01 12_network-generate-ula -rw-rw-r-- 1 karlp karlp 377 Sep 25 11:01 99_stuff_here -rw-rw-r-- 1 karlp karlp 211 Aug 21 12:40 agent_etactica -rwxr-xr-x 1 karlp karlp 83 Jan 11 15:05 luci-i18n-english -rwxr-xr-x 1 karlp karlp 1285 Jan 26 13:33 mosquitto.uci-defaults -rw-rw-r-- 1 karlp karlp 479 Aug 21 12:40 mpacker -rwxrwxr-x 1 karlp karlp 291 Aug 21 12:40 mq_delta_gadget - -rwxrwxr-x 1 karlp karlp 970 Aug 21 12:40 mrelay -rwxr-xr-x 1 karlp karlp 300 Jan 26 13:29 odhcpd.defaults -rwxrwxr-x 1 karlp karlp 224 Sep 24 14:24 snmpd badly wrapped. Here's what you really have in a new image. Putting things in "99_stuff" just won't work very well. All the package provided uci-defaults files, which are the ones I want to extend/augment with my own files have to come later. They are run at /etc/rc.d/S10boot, S99_whatever will run after S10_whatever. The S## files are run at startup, in order and the K## files are run at shutdown in reverse order (since shutdown can sometimes require different ordering than startup. If something is in there with just ## I'm not sure if they are even run. David Lang ___ openwrt-devel mailing list openwrt-devel@lists.openwrt.org https://lists.openwrt.org/cgi-bin/mailman/listinfo/openwrt-devel
Re: [OpenWrt-Devel] Question about custom script at init on Trunk.-
-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 David Lang wrote: > On Wed, 27 Jan 2016, Karl Palsson wrote: > > > David Lang wrote: > >>> > >>> Now, yes, I can have my 80_mystuff script go and try and create > >>> the package config files it might want to modify, but really, I > >>> want "mystuff" to run _last_ or at least, after all the packages > >>> have run. Is there any better system for this other than just > >>> naming my script z_mystuff instead of 80_mystuff? Is that > >>> expected to be "good enough" ? > >>> > >>> I'm trying this, because I've _been_ trying to use packages > >>> instead of modifying upstream, but when I want my package to do > >>> different things for different targets, it's seemed easier to > >>> hook into the existing target code, rather than trying to copy > >>> target detection and config into a package Config.in options > >> > >> typically people use 99stuff and figure that the order of 99s > >> doesn't really matter much > >> > > > > Well, the thing is, 99 isn't last. that's what I was getting at. > > > > $ ls -l > > build_dir/target-mips_34kc_uClibc-0.9.33.2/root-ar71xx/etc/uci-defaults/ > > total 140 > > badly wrapped. > > > Here's what you really have in a new image. Putting things in > > "99_stuff" just won't work very well. All the package provided > > uci-defaults files, which are the ones I want to extend/augment > > with my own files have to come later. They are run at > > /etc/rc.d/S10boot, > > S99_whatever will run after S10_whatever. The S## files are run > at startup, in order and the K## files are run at shutdown in > reverse order (since shutdown can sometimes require different > ordering than startup. If something is in there with just ## > I'm not sure if they are even run. I think you may have misunderstood. S10_boot runs the files in /etc/uci-defaults/* The files in _/etc/uci-defaults/_ do not all start with a numerical prefix, and are provided by any and all package that has any config it wants to run. So, for a _package_ that wants to do something "late" to do any sort of special target configuration, such as changing serial ports or mount points for an application based on the hardware it's running on, anything like 99-xxx is not going to reliably be late at all. Naming your uci-defaults file something like z999-something will probably work, but it's pretty gross. If you instead try and add your own _init_ script, with a START=99, you now don't even run until the applications whose config you want to adjust have already attempted startup, at whatever START=xx tag they have in their own provided init script. So, /etc/uci-defaults/z999-something works, but is there a better way? Cheers, Karl P -BEGIN PGP SIGNATURE- Version: GnuPG v1.4.11 (GNU/Linux) iQIcBAEBAgAGBQJWqONhAAoJEBmotQ/U1cr2E7QP/jQz2mZp75xiZDEx8tnSwuku XcYObMug6vtUAO+RAnzAkhLCp+A+37dwxMhvRgCrTMwTncpELncuwoFIJZ7fd5yQ +vHV1e0S0WUgNNEjKOy8iCPnXugatI82F7prUZNNb6LjnwlnD2MQdEUF3prFm0eZ WdQDuE+oShtN/L2V/RcGqF+wnl1c8/PQX+XJ558+1TqBRAjIckkgv+T6E/ysiZ0O 3TfFSpYRftXZ/EQnxeO8J5vCkJb3q7kZ84fZVf9Awr5MuD5zjKnJTIuXpBBcne59 3HM2BCGtlCwn0AWrvg/8lTRSyYN2biubygrsj/w901cEKd+i3Hb9E/VwIY38pEyE 3jXywYvWkKk+rgAw5A7cZk8hYj5Tc3b8E235H1h5tOUJzk/XdDK9lisuTuKK+aib MS6qx8U5gx1cJ2yrUQL9BFv5FWKAh63W/aPweqK18QhQjCcTWMreDRodlasS8TT+ jLptAl10BITtjztRXrlRjaq+BHEu8dU1/jTN97puBApuIiuHpcbjY9e8LoixvGG0 S1w4oT10NkccYyEcXGdHN+u+k0NOKp74sRaLx/vTjvygI2e/BMW2nha5xzhQQEzt oJDrQLh4pOV31IQTa2AzbTKVF9pk40ii7+zZOLLQ8pqJnsi7gTa+MDykdgiuXbHG sEF7mAr2OFE9QF9Ih9uk =HZHN -END PGP SIGNATURE- ___ openwrt-devel mailing list openwrt-devel@lists.openwrt.org https://lists.openwrt.org/cgi-bin/mailman/listinfo/openwrt-devel
Re: [OpenWrt-Devel] Question about custom script at init on Trunk.-
Hi, > So, /etc/uci-defaults/z999-something works, but is there a > better way? Not yet - but we could start changing packages to use numerical prefixes. I do not suppose that sooo many uci-default scripts are used. ~ Jow ___ openwrt-devel mailing list openwrt-devel@lists.openwrt.org https://lists.openwrt.org/cgi-bin/mailman/listinfo/openwrt-devel
Re: [OpenWrt-Devel] [PATCH][ar71xx] Routerboard 951G Switch Fix
Hi, 951G owners! I have RB951G-2HnD device with AR9344 rev 3. It is configured like that: - Balanced 2 WAN ports (mwan3) - ipsec LAN-to-LAN - OpenVPN server - about 30 clients (wi-fi, lan) My results for different ath79_eth0_pll_data.pll_1000 values: 0x3e00 - it worked fine with Barrier Breaker, but sometimes (one or two times per day) it was restarting suddenly. It was happening after I have configured mwan3. 0x0600 - it works wonderful with Chaos Calmer. The configuration is the same. No problem with reboot. So, I would recommend this value. unpatched - of course it doesn't work. :-) If it's useful, please add this information to wiki-page: https://wiki.openwrt.org/toh/mikrotik/rb2011uias#tracking_reported_experience_with_suggested_patch_for_the_5_gige_ports -- Anton ___ openwrt-devel mailing list openwrt-devel@lists.openwrt.org https://lists.openwrt.org/cgi-bin/mailman/listinfo/openwrt-devel
Re: [OpenWrt-Devel] [PATCH][ar71xx] Routerboard 951G Switch Fix
On 27/01/2016 17:03, Anton Kalmykov wrote: > > Hi, 951G owners! > I have RB951G-2HnD device with AR9344 rev 3. It is configured like that: > - Balanced 2 WAN ports (mwan3) > - ipsec LAN-to-LAN > - OpenVPN server > - about 30 clients (wi-fi, lan) > > My results for different ath79_eth0_pll_data.pll_1000 values: > > 0x3e00 - it worked fine with Barrier Breaker, but sometimes (one or > two times per day) it was restarting suddenly. It was happening after I > have configured mwan3. > 0x0600 - it works wonderful with Chaos Calmer. The configuration is so the conensus is that 0x0600 is correct ? John > the same. No problem with reboot. So, I would recommend this value. > unpatched - of course it doesn't work. :-) > > If it's useful, please add this information to wiki-page: > https://wiki.openwrt.org/toh/mikrotik/rb2011uias#tracking_reported_experience_with_suggested_patch_for_the_5_gige_ports > > > -- > Anton > ___ > openwrt-devel mailing list > openwrt-devel@lists.openwrt.org > https://lists.openwrt.org/cgi-bin/mailman/listinfo/openwrt-devel ___ openwrt-devel mailing list openwrt-devel@lists.openwrt.org https://lists.openwrt.org/cgi-bin/mailman/listinfo/openwrt-devel
[OpenWrt-Devel] [PATCH v4] base-files: For sysfixtime use hwclock if RTC available
From: Daniel Dickinson
On systems that have an RTC prefer it to the file-based time fixup (i.e.
use hwclock when there is a permanent clock instead of the faked up time
logic that is needed when there is not RTC).
We can't rely on hctosys kernel feature either as we're usually using
RTC as kernel modules which are usually being loaded after hctosys was
run, leading in the following error:
hctosys: unable to open rtc device (rtc0)
Signed-off-by: Daniel Dickinson
Signed-off-by: Petr Štetiar
---
Changes in v4:
* simplified with using start() in boot() (blogic)
package/base-files/files/etc/init.d/sysfixtime | 14 ++
1 file changed, 14 insertions(+)
diff --git a/package/base-files/files/etc/init.d/sysfixtime
b/package/base-files/files/etc/init.d/sysfixtime
index 4010e06..ab946f6 100755
--- a/package/base-files/files/etc/init.d/sysfixtime
+++ b/package/base-files/files/etc/init.d/sysfixtime
@@ -2,10 +2,24 @@
# Copyright (C) 2013-2014 OpenWrt.org
START=00
+STOP=90
+
+RTC_DEV=/dev/rtc0
+HWCLOCK=/sbin/hwclock
boot() {
+ start && exit 0
+
local curtime="$(date +%s)"
local maxtime="$(find /etc -type f -exec date -r {} +%s \; | sort -nr |
head -n1)"
[ $curtime -lt $maxtime ] && date -s @$maxtime
}
+start() {
+ [ -e "$RTC_DEV" ] && [ -e "$HWCLOCK" ] && $HWCLOCK -s -f $RTC_DEV
+}
+
+stop() {
+ [ -e "$RTC_DEV" ] && [ -e "$HWCLOCK" ] && $HWCLOCK -w -f $RTC_DEV && \
+ logger -t sysfixtime "saved '$(date)' to $RTC_DEV"
+}
--
1.7.9.5
___
openwrt-devel mailing list
openwrt-devel@lists.openwrt.org
https://lists.openwrt.org/cgi-bin/mailman/listinfo/openwrt-devel
Re: [OpenWrt-Devel] [PATCH][ar71xx] Routerboard 951G Switch Fix
Hello John, Of course this has been over a year since I submitted ( http://patchwork.ozlabs.org/patch/419857/ ); however I remember in one of our e-mail threads that you ( or Felix? ) thought it was possible to pull the correct pll_1000 value from the bootloader? Is that still a possibility? If you could somehow acquire the correct value directly from the boot args; that would seem ideal. -- Davey On Wed, Jan 27, 2016 at 9:27 AM, John Crispin wrote: > > > On 27/01/2016 17:03, Anton Kalmykov wrote: >> >> Hi, 951G owners! >> I have RB951G-2HnD device with AR9344 rev 3. It is configured like that: >> - Balanced 2 WAN ports (mwan3) >> - ipsec LAN-to-LAN >> - OpenVPN server >> - about 30 clients (wi-fi, lan) >> >> My results for different ath79_eth0_pll_data.pll_1000 values: >> >> 0x3e00 - it worked fine with Barrier Breaker, but sometimes (one or >> two times per day) it was restarting suddenly. It was happening after I >> have configured mwan3. >> 0x0600 - it works wonderful with Chaos Calmer. The configuration is > > so the conensus is that 0x0600 is correct ? > > John > > > >> the same. No problem with reboot. So, I would recommend this value. >> unpatched - of course it doesn't work. :-) >> >> If it's useful, please add this information to wiki-page: >> https://wiki.openwrt.org/toh/mikrotik/rb2011uias#tracking_reported_experience_with_suggested_patch_for_the_5_gige_ports >> >> >> -- >> Anton >> ___ >> openwrt-devel mailing list >> openwrt-devel@lists.openwrt.org >> https://lists.openwrt.org/cgi-bin/mailman/listinfo/openwrt-devel > ___ > openwrt-devel mailing list > openwrt-devel@lists.openwrt.org > https://lists.openwrt.org/cgi-bin/mailman/listinfo/openwrt-devel ___ openwrt-devel mailing list openwrt-devel@lists.openwrt.org https://lists.openwrt.org/cgi-bin/mailman/listinfo/openwrt-devel
Re: [OpenWrt-Devel] [PATCH 2/2] kernel: other: add PWM controller support
I just realized that this isn't necessary for imx anymore as it's static in the 4.4 kernel. I'll mark it as 'not applicable'. - Pushpal On Tue, Jan 26, 2016 at 2:37 PM, Pushpal Sidhu wrote: > This adds basic PWM controller support and support for i.MX > > Signed-off-by: Tim Harvey > Signed-off-by: Pushpal Sidhu > --- > package/kernel/linux/modules/other.mk | 36 > +++ > 1 file changed, 36 insertions(+) > > diff --git a/package/kernel/linux/modules/other.mk > b/package/kernel/linux/modules/other.mk > index 3039180..8edf8a6 100644 > --- a/package/kernel/linux/modules/other.mk > +++ b/package/kernel/linux/modules/other.mk > @@ -371,6 +371,42 @@ endef > $(eval $(call KernelPackage,mmc)) > > > +define KernelPackage/pwm > + SUBMENU:=$(OTHER_MENU) > + TITLE:=Pulse-Width Modulation support > + KCONFIG:= \ > + CONFIG_PWM=y \ > + CONFIG_REGULATOR_PWM=n \ > + CONFIG_PWM_FSL_FTM=n > +endef > + > +define KernelPackage/pwm/description > + Pulse Width Modulation subsystem. > +endef > + > +$(eval $(call KernelPackage,pwm)) > + > +define AddDepends/pwm > + SUBMENU:=$(OTHER_MENU) > + DEPENDS+=kmod-pwm $(1) > +endef > + > +define KernelPackage/pwm-imx > + SUBMENU:=$(OTHER_MENU) > + TITLE:=IMX Pulse-Width Modulation support > + KCONFIG:=CONFIG_PWM_IMX > + FILES:= $(LINUX_DIR)/drivers/pwm/pwm-imx.ko > + DEPENDS:=@TARGET_imx6 > + AUTOLOAD:=$(call AutoProbe,pwm-imx) > + $(call AddDepends/pwm) > +endef > + > +define KernelPackage/pwm-imx/description > + Freescale IMX Pulse Width Modulation support. > +endef > + > +$(eval $(call KernelPackage,pwm-imx)) > + > define KernelPackage/sdhci >SUBMENU:=$(OTHER_MENU) >TITLE:=Secure Digital Host Controller Interface support > -- > 2.7.0 > ___ openwrt-devel mailing list openwrt-devel@lists.openwrt.org https://lists.openwrt.org/cgi-bin/mailman/listinfo/openwrt-devel
[OpenWrt-Devel] [PATCHv2 5/8] ath9k: enable GPIO buttons
From: Michal Cieslakiewicz
Enable platform-defined GPIO button support for ath9k device. Key poller
is activated for attached platform buttons. Requires ath9k GPIO chip access.
Signed-off-by: Michal Cieslakiewicz
---
(Patch offset adjusted due to lines removal in preceding patch v2 4/8)
.../patches/550-ath9k_enable_gpio_buttons.patch| 164 +
.../ar71xx/files/arch/mips/ath79/dev-ap9x-pci.c| 18 +++
.../ar71xx/files/arch/mips/ath79/dev-ap9x-pci.h| 7 +
.../generic/files/include/linux/ath9k_platform.h | 4 +
4 files changed, 193 insertions(+)
create mode 100644
package/kernel/mac80211/patches/550-ath9k_enable_gpio_buttons.patch
diff --git
a/package/kernel/mac80211/patches/550-ath9k_enable_gpio_buttons.patch
b/package/kernel/mac80211/patches/550-ath9k_enable_gpio_buttons.patch
new file mode 100644
index 000..5d4617a
--- /dev/null
+++ b/package/kernel/mac80211/patches/550-ath9k_enable_gpio_buttons.patch
@@ -0,0 +1,164 @@
+--- a/drivers/net/wireless/ath/ath9k/ath9k.h
b/drivers/net/wireless/ath/ath9k/ath9k.h
+@@ -825,6 +825,13 @@ int ath_create_gpio_led(struct ath_softc
+ void ath9k_register_gpio_chip(struct ath_softc *sc);
+ void ath9k_unregister_gpio_chip(struct ath_softc *sc);
+
++/**/
++/* GPIO Buttons */
++/**/
++
++void ath9k_init_buttons(struct ath_softc *sc);
++void ath9k_deinit_buttons(struct ath_softc *sc);
++
+ #else
+ static inline void ath_init_leds(struct ath_softc *sc)
+ {
+@@ -843,6 +850,13 @@ static inline void ath9k_register_gpio_c
+ static inline void ath9k_unregister_gpio_chip(struct ath_softc *sc)
+ {
+ }
++
++void ath9k_init_buttons(struct ath_softc *sc)
++{
++}
++void ath9k_deinit_buttons(struct ath_softc *sc)
++{
++}
+ #endif
+
+ //
+@@ -1039,6 +1053,7 @@ struct ath_softc {
+ const char *led_default_trigger;
+ struct list_head leds;
+ struct ath9k_gpio_chip *gpiochip;
++ struct platform_device *btnpdev;/* gpio-keys-polled */
+ #endif
+
+ #ifdef CPTCFG_ATH9K_DEBUGFS
+--- a/drivers/net/wireless/ath/ath9k/gpio.c
b/drivers/net/wireless/ath/ath9k/gpio.c
+@@ -24,6 +24,8 @@
+ #ifdef CPTCFG_MAC80211_LEDS
+
+ #include
++#include
++#include
+
+ static void ath_led_brightness(struct led_classdev *led_cdev,
+ enum led_brightness brightness)
+@@ -308,6 +310,83 @@ void ath9k_unregister_gpio_chip(struct a
+ sc->gpiochip = NULL;
+ }
+
++/**/
++/* GPIO Buttons */
++/**/
++
++/* add GPIO buttons */
++void ath9k_init_buttons(struct ath_softc *sc)
++{
++ struct ath9k_platform_data *pdata = sc->dev->platform_data;
++ struct platform_device *pdev;
++ struct gpio_keys_platform_data gkpdata;
++ struct gpio_keys_button *bt;
++ int i;
++
++ if (!sc->gpiochip)
++ return;
++
++ bt = kmemdup(pdata->btns,
++ pdata->num_btns * sizeof(struct gpio_keys_button),
++ GFP_KERNEL);
++
++ if (!bt)
++ return;
++
++ pdev = platform_device_alloc("gpio-keys-polled", PLATFORM_DEVID_AUTO);
++
++ if (!pdev)
++ goto err_bt_free;
++
++ for (i = 0; i < pdata->num_btns; i++) {
++ ath9k_hw_cfg_gpio_input(sc->sc_ah, pdata->btns[i].gpio);
++ bt[i].gpio = sc->gpiochip->gchip.base + pdata->btns[i].gpio;
++ }
++
++ memset(&gkpdata, 0, sizeof(struct gpio_keys_platform_data));
++ gkpdata.buttons = bt;
++ gkpdata.nbuttons = pdata->num_btns;
++ gkpdata.poll_interval = pdata->btn_poll_interval;
++
++ if (platform_device_add_data(pdev, &gkpdata, sizeof(gkpdata)))
++ goto err_pdev_put;
++
++ if (platform_device_add(pdev))
++ goto err_pdev_put;
++
++ sc->btnpdev = pdev;
++
++ return;
++
++err_pdev_put:
++ platform_device_put(pdev);
++
++err_bt_free:
++ kfree(bt);
++}
++
++/* remove GPIO buttons */
++void ath9k_deinit_buttons(struct ath_softc *sc)
++{
++ struct gpio_keys_platform_data *gkpdata;
++ struct gpio_keys_button *bt = NULL;
++
++ if (!sc->gpiochip || !sc->btnpdev)
++ return;
++
++ gkpdata = sc->btnpdev->dev.platform_data;
++ if (gkpdata)
++ bt = gkpdata->buttons;
++
++ platform_device_del(sc->btnpdev);
++ platform_device_put(sc->btnpdev);
++
++ sc->btnpdev = NULL;
++
++ if (bt)
++ kfree(bt);
++}
++
+ #endif
+
+ /***/
+--- a/drivers/net/wireless/ath/ath9k/init.c
b/drivers/net/wireless/ath/ath9k/init.c
+@@ -981,6 +981,7 @@ int ath9k_init_device(u16 devid, struct
+
+ ath9k_register_gpio_chip(sc);
+ ath_init_leds(sc);
++ ath9k_init_buttons(sc);
+ ath_start_rfkill_poll(sc);
+
+ return 0;
+@@ -1026,6 +1027,7 @@ void ath9k_deinit_device(struct ath_soft
+ ath9k_ps_wakeup(sc);
+
+ wiphy_rfkill_stop_polling(sc->hw->wiphy);
++ ath9k_deinit_but
[OpenWrt-Devel] [PATCHv2 4/8] ath9k: enable access to GPIO
From: Michal Cieslakiewicz
Enable access to GPIO chip and its pins for Atheros AR92xx
wireless devices. For now AR9285 and AR9287 are supported.
Signed-off-by: Michal Cieslakiewicz
---
(Removed unneeded pdata pointer check in ath9k_register_gpio_chip() function)
.../patches/549-ath9k_enable_gpio_chip.patch | 229 +
1 file changed, 229 insertions(+)
create mode 100644
package/kernel/mac80211/patches/549-ath9k_enable_gpio_chip.patch
diff --git a/package/kernel/mac80211/patches/549-ath9k_enable_gpio_chip.patch
b/package/kernel/mac80211/patches/549-ath9k_enable_gpio_chip.patch
new file mode 100644
index 000..d49c80f
--- /dev/null
+++ b/package/kernel/mac80211/patches/549-ath9k_enable_gpio_chip.patch
@@ -0,0 +1,229 @@
+--- a/drivers/net/wireless/ath/ath9k/ath9k.h
b/drivers/net/wireless/ath/ath9k/ath9k.h
+@@ -24,6 +24,7 @@
+ #include
+ #include
+ #include
++#include
+
+ #include "common.h"
+ #include "debug.h"
+@@ -817,6 +818,13 @@ void ath_fill_led_pin(struct ath_softc *
+ int ath_create_gpio_led(struct ath_softc *sc, int gpio, const char *name,
+ const char *trigger, bool active_low);
+
++/***/
++/* GPIO Chip */
++/***/
++
++void ath9k_register_gpio_chip(struct ath_softc *sc);
++void ath9k_unregister_gpio_chip(struct ath_softc *sc);
++
+ #else
+ static inline void ath_init_leds(struct ath_softc *sc)
+ {
+@@ -828,6 +836,13 @@ static inline void ath_deinit_leds(struc
+ static inline void ath_fill_led_pin(struct ath_softc *sc)
+ {
+ }
++
++static inline void ath9k_register_gpio_chip(struct ath_softc *sc)
++{
++}
++static inline void ath9k_unregister_gpio_chip(struct ath_softc *sc)
++{
++}
+ #endif
+
+ //
+@@ -963,6 +978,12 @@ struct ath_led {
+ struct led_classdev cdev;
+ };
+
++struct ath9k_gpio_chip {
++ struct ath_softc *sc;
++ char label[32];
++ struct gpio_chip gchip;
++};
++
+ struct ath_softc {
+ struct ieee80211_hw *hw;
+ struct device *dev;
+@@ -1017,6 +1038,7 @@ struct ath_softc {
+ #ifdef CPTCFG_MAC80211_LEDS
+ const char *led_default_trigger;
+ struct list_head leds;
++ struct ath9k_gpio_chip *gpiochip;
+ #endif
+
+ #ifdef CPTCFG_ATH9K_DEBUGFS
+--- a/drivers/net/wireless/ath/ath9k/gpio.c
b/drivers/net/wireless/ath/ath9k/gpio.c
+@@ -22,6 +22,9 @@
+ //
+
+ #ifdef CPTCFG_MAC80211_LEDS
++
++#include
++
+ static void ath_led_brightness(struct led_classdev *led_cdev,
+ enum led_brightness brightness)
+ {
+@@ -60,6 +63,10 @@ static int ath_add_led(struct ath_softc
+ else
+ ath9k_hw_set_gpio(sc->sc_ah, gpio->gpio, gpio->active_low);
+
++ /* If there is GPIO chip configured, reserve LED pin */
++ if (sc->gpiochip)
++ gpio_request(sc->gpiochip->gchip.base + gpio->gpio, gpio->name);
++
+ return 0;
+ }
+
+@@ -116,6 +123,9 @@ void ath_deinit_leds(struct ath_softc *s
+
+ while (!list_empty(&sc->leds)) {
+ led = list_first_entry(&sc->leds, struct ath_led, list);
++ /* If there is GPIO chip configured, free LED pin */
++ if (sc->gpiochip)
++ gpio_free(sc->gpiochip->gchip.base + led->gpio->gpio);
+ list_del(&led->list);
+ ath_led_brightness(&led->cdev, LED_OFF);
+ led_classdev_unregister(&led->cdev);
+@@ -186,6 +196,118 @@ void ath_fill_led_pin(struct ath_softc *
+ /* LED off, active low */
+ ath9k_hw_set_gpio(ah, ah->led_pin, (ah->config.led_active_high) ? 0 :
1);
+ }
++
++/***/
++/* GPIO Chip */
++/***/
++
++/* gpio_chip handler : set GPIO to input */
++static int ath9k_gpio_pin_cfg_input(struct gpio_chip *chip, unsigned offset)
++{
++ struct ath9k_gpio_chip *gc = container_of(chip, struct ath9k_gpio_chip,
++gchip);
++
++ ath9k_hw_cfg_gpio_input(gc->sc->sc_ah, offset);
++
++ return 0;
++}
++
++/* gpio_chip handler : set GPIO to output */
++static int ath9k_gpio_pin_cfg_output(struct gpio_chip *chip, unsigned offset,
++ int value)
++{
++ struct ath9k_gpio_chip *gc = container_of(chip, struct ath9k_gpio_chip,
++gchip);
++
++ ath9k_hw_cfg_output(gc->sc->sc_ah, offset,
++ AR_GPIO_OUTPUT_MUX_AS_OUTPUT);
++ ath9k_hw_set_gpio(gc->sc->sc_ah, offset, value);
++
++ return 0;
++}
++
++/* gpio_chip handler : query GPIO direction (0=out, 1=in) */
++static int ath9k_gpio_pin_get_dir(struct gpio_chip *chip, unsigned offset)
++{
++ struct ath9k_gpio_chip *gc = container_of(chip, struct ath9k_gpio_chip,
++gchip);
++ struct ath_hw *ah = gc->sc->sc_ah;
++
++ return !((REG_READ(ah, AR_GPIO_OE_OUT) >> (offset * 2)) & 3);
++}
++
++/* gpio_chip handler
Re: [OpenWrt-Devel] [PATCH][ar71xx] Routerboard 951G Switch Fix
Hello John, Sorry, I'm mistaken: my value is 0x6f00 (NOT 0x0600) and it works correctly. And, yes, taking value from the bootloader is a nice idea, it would be good! Anton 27.01.2016, 21:28, John Crispin wrote: On 27/01/2016 17:03, Anton Kalmykov wrote: > > Hi, 951G owners! > I have RB951G-2HnD device with AR9344 rev 3. It is configured like that: > - Balanced 2 WAN ports (mwan3) > - ipsec LAN-to-LAN > - OpenVPN server > - about 30 clients (wi-fi, lan) > > My results for different ath79_eth0_pll_data.pll_1000 values: > > 0x3e00 - it worked fine with Barrier Breaker, but sometimes (one or > two times per day) it was restarting suddenly. It was happening after I > have configured mwan3. > 0x0600 - it works wonderful with Chaos Calmer. The configuration is so the conensus is that 0x0600 is correct ? John > the same. No problem with reboot. So, I would recommend this value. > unpatched - of course it doesn't work. :-) > > If it's useful, please add this information to wiki-page: > https://wiki.openwrt.org/toh/mikrotik/rb2011uias#tracking_reported_experience_with_suggested_patch_for_the_5_gige_ports > > > -- > Anton > ___ > openwrt-devel mailing list > openwrt-devel@lists.openwrt.org > https://lists.openwrt.org/cgi-bin/mailman/listinfo/openwrt-devel ___ openwrt-devel mailing list openwrt-devel@lists.openwrt.org https://lists.openwrt.org/cgi-bin/mailman/listinfo/openwrt-devel ___ openwrt-devel mailing list openwrt-devel@lists.openwrt.org https://lists.openwrt.org/cgi-bin/mailman/listinfo/openwrt-devel
Re: [OpenWrt-Devel] Question about custom script at init on Trunk.-
On Wed, 27 Jan 2016, Karl Palsson wrote: I think you may have misunderstood. S10_boot runs the files in /etc/uci-defaults/* The files in _/etc/uci-defaults/_ do not all start with a numerical prefix, and are provided by any and all package that has any config it wants to run. So, for a _package_ that wants to do something "late" to do any sort of special target configuration, such as changing serial ports or mount points for an application based on the hardware it's running on, anything like 99-xxx is not going to reliably be late at all. Naming your uci-defaults file something like z999-something will probably work, but it's pretty gross. If you instead try and add your own _init_ script, with a START=99, you now don't even run until the applications whose config you want to adjust have already attempted startup, at whatever START=xx tag they have in their own provided init script. So, /etc/uci-defaults/z999-something works, but is there a better way? why do you have to change the files from something in uci-defaults? can't you just put the corrected configs in the image (most of them are scripts, so you can add the needed smarts to them to have them pull in variables from elsewhere) you can run a S01 script in init.d (or something late enough to work, early enough to run before whatever you need to change) and have it populate the variables that other things will use. David Lang ___ openwrt-devel mailing list openwrt-devel@lists.openwrt.org https://lists.openwrt.org/cgi-bin/mailman/listinfo/openwrt-devel
[OpenWrt-Devel] kernel patch editing using QUILT=1
Hi,
As quilt is used for patch management it seems that behaviour as
described on the wiki
(https://wiki.openwrt.org/doc/devel/patches#adding_or_editing_kernel_patches)
between packages and kernel is different. For packages it will prepare
the environment when issued:
make package/example/{clean,prepare} V=s QUILT=1
where-as for the kernel it will prepare _and_ apply the patches, when
issued:
make target/linux/{clean,prepare} V=s QUILT=1
Which command to issue to ensure the kernel environment is only prepared
instead?
Best regards,
/Rick
___
openwrt-devel mailing list
openwrt-devel@lists.openwrt.org
https://lists.openwrt.org/cgi-bin/mailman/listinfo/openwrt-devel
[OpenWrt-Devel] Build error for IPQ806X
Hello, I am getting build error while compiling latest OpenWRT trunk for IPQ806x. Below is the error message. Inconsistent kallsyms data Try make KALLSYMS_EXTRA_PASS=1 as a workaround make[4]: *** [vmlinux] Error 1 Regards, Tathagata --- Alumnus Software Ltd Infinity, Tower-II, 2nd Floor, Plot A3, Block GP, Sector V Salt Lake City Kolkata 700091 Phone : +91-33-23575626, +91-33-23575627, +91-33-23575628 +91-33-23571036,+91-33-23571037, +91-33-23571038, +91-33-23571039 Fax : +91-33-23575268 ___ openwrt-devel mailing list openwrt-devel@lists.openwrt.org https://lists.openwrt.org/cgi-bin/mailman/listinfo/openwrt-devel
