Country & Revision Setting Problem with E3200 (5 GHZ -> USB Device) / mfg_wait=on problem

[maverick3005]

On actual mips trunk the Country Setting is not working for the 5 GHZ USB-AP Device.

Based on a fresh flash and before a complete nvram clear/erase -> cfe defaults i have this results:

When setted Country & Revision via Webinterface nvram shows this variables:

sb/1/regrev=7
usb/0xBD17/regrev=0 <- wrong
wl0_country_rev=7
wl1_country_rev=7
wl_corerev=
wl_country_rev=0

and:

sb/1/ccode=DE
usb/0xBD17/ccode=0 <- wrong

after i set manually the nvram variables, all is working:

usb/0xBD17/regrev=7
and
usb/0xBD17/ccode=DE

So the problem is, that usualy wl0/wl1 is configured via web, but not internel USB-connected devices.
I´am expect, that other Linksys Ex2xx devices, which have internal USB/5GHz-Modules, have the same issue, when only configured by webinterface.

The webinterface shows "cosmetic" the correct country/revision, but the usb-device itself needs the extra usb variables (the ID can differ!)
configured, too.

The very long bootdelay:

One more issue is the very long boot time. Checked in serial console, and after power on, there are 5 trys with a specific TFTP Request to
192.168.1.254 which will not be answered. Followed by "boot_wait" & the setted delay and then router boots from internal flash.

After a nvram clear the default nvram setting is "mfg_wait=on" (!) - a better option is to set it to "mfg_wait=off" to speed up the boot time
of 10 seconds and more !

One cosmetic issue in the Overview is the missing "CFE-Version: -"
Workaround: By adding: "bl_version=1.0.37" via nvram command, the variable is used in the overview.
I asume there is no "bl_version" setted after a factory nvram clear for the E3200 ?

[txnative777]

A question, what are you setting for country/region, country/rev is it "DE/7"?

The "mfg_wait" is optional to set or disable, I am sure you know that it there for TFTP recovery? Either way the user can disable it but know that if you need TFTP and your router is in a boot loop, recovery with TFTP is disabled and so you have two options, serial or maybe 30-30-30.

One cosmetic issue in the Overview is the missing "CFE-Version: -"
Workaround: By adding: "bl_version=1.0.37" via nvram command, the variable is used in the overview.
I asume there is no "bl_version" setted after a factory nvram clear for the E3200 ?

Not all models will show their "CFE Version" the "bl_version" is in the code something might need to be added or if additional code can be added to search in NVRAM, currently it is searching in the CFE, /dev/mtd0ro, while some models may support it and the code finds it, but not for the E3200 as well as other models, it isn't going to find the bl_version?

[txnative777]

On actual mips trunk the Country Setting is not working for the 5 GHZ USB-AP Device.

Based on a fresh flash and before a complete nvram clear/erase -> cfe defaults i have this results:

When setted Country & Revision via Webinterface nvram shows this variables:

sb/1/regrev=7 usb/0xBD17/regrev=0 <- wrong wl0_country_rev=7 wl1_country_rev=7 wl_corerev= wl_country_rev=0

and:

sb/1/ccode=DE usb/0xBD17/ccode=0 <- wrong

after i set manually the nvram variables, all is working:

usb/0xBD17/regrev=7 and usb/0xBD17/ccode=DE

So the problem is, that usualy wl0/wl1 is configured via web, but not internel USB-connected devices. I´am expect, that other Linksys Ex2xx devices, which have internal USB/5GHz-Modules, have the same issue, when only configured by webinterface.

The webinterface shows "cosmetic" the correct country/revision, but the usb-device itself needs the extra usb variables (the ID can differ!) configured, too.

I believe the E3200 CFE doesn't set any specific ccode/regrev, it mostly done through upgrades then set in nvram.

For example: On my E3200 the CFE sets ccode=0, regrev=0.

In FreshTomato, it is important to set the proper country/region, country/rev within the "advanced-wireless".

Looking over "DE/7" country/rev=7 doesn't appear to be in the wlc_channels file.

{"DE",	2, LOCALES(b,    3,  bn,  3n)},	/* Germany */
{"DE",	3, LOCALES(b,    3c, bn, 3cn)},	/* Germany */
{"DE",	4, LOCALES(b_11, 3d, bn_9, 3dn)},	/* Germany Apple X21 */
{"DE",	5, LOCALES(b,    3l_1, bn1_1, 3ln_1)},	/* Germany */

[maverick3005]

I know the tricks with direct nvram commands very well ;-)

The "bl-trick for CPE 1.0.37" was the first i made with NVRAM direct command to have a better "cosmetic"

More importent is for the Devices with 5 GHz USB-Modules, that the changes from the Webinterface are still made for a
classic Onboard-Modul - but here these are not used. For the E3200 it is importent that the USB specific Device-Name get the
same Variables as the Onboard Device(s). In my case was missing 5 GHz Channels the reason for closer checking the NVRAM
Variables. (I actualy set CC and REV via NVRAM command manually to have a real effect on the USB-Device)

Checking the other Settings for mips-devices is a good idea, i took this "DE/7" from my EA6xxx-arm Routers quickly
and basicly is working and have all channels avail. I will check the mips-typical rev-codes, too shortly

Patching the CFE and change the "missing" variables is the easier solution, of course.

When i remember correctly for the E3200 is an EU and US Firmware avaibale - but booth use the same CFE.
Linksys set the variables in his stockfirmware after initial reset. In my EU-based Device are still US variables
used comming from the CFE after an nvram erase/clear.

The "mfg_wait=on" as default is known with tftp and a specific recovery filename. I saw this all in the serial
console. Setting it to OFF speed up the boot process massive. When mfg_wait=off and a normal wait=X
is settet, it is still possible to still use the tftp client mode with just 3 seconds timeout. "Recovery" is only a 2nd
Option with a huge pre boot delay of 10-15 seconds extra.

One thing more i figured out, is the wrong "Port-State Info":

The WAN0-Port is shown as "LAN3", and "LAN0" is not used. The "Reverse Port function" doesn´t work, and just
show the active Port order again wrongly. (so LAN0 becomes WAN....)

The VLAN-Port-Settings instead are correct (so WAN IS WAN). The table for the Stats is still buggy.
I can remember that in an older Version this was fixed for years - and now the issue is back.

"robocfg" gives me this:

Port 0: DOWN enabled stp: none vlan: 1 jumbo: off mac: 00:00:00:00:00:00
Port 1: DOWN enabled stp: none vlan: 1 jumbo: off mac: 00:00:00:00:00:00
Port 2: DOWN enabled stp: none vlan: 1 jumbo: off mac: 00:00:00:00:00:00
Port 3: DOWN enabled stp: none vlan: 1 jumbo: off mac: 00:00:00:00:00:00
Port 4: 1000FD enabled stp: none vlan: 1 jumbo: off mac: 24:be:05:08:xx:xx
Port 5: DOWN enabled stp: none vlan: 1 jumbo: off mac: 00:00:00:00:00:00
Port 8: 1000FD enabled stp: none vlan: 1 jumbo: off mac: 58:6d:8f:df:xxxx

So Port 8 is the usual CPU-Port and Port 4 is the real WAN-Port, Ports 0-3 are the LAN-Ports on E3200

One more interessting finding on this device was by modding the internal soldered 2,4 GHz wires by adding
Pigtail-Support for external antennas. By using the Settings "Antenna A" and "Antenna B" i figured out,
that the middle Antenna-Point is basicly not used for TX (and in my case it can be unconnected/unused).
I asume the middle Ports works only in receive direction and is not configurable. (actualy by web)

Adding an external Antenna and play with "AUTO, A or B" doesn´t change anything on the TX-situation.
Only the Left and Right Antenna-Plugs are in use. The same is for the 5 GHz Module the case.
(This makes Sense for 2x2 MIMO) - I have added 2 Panel-Antennas and have the setting "AUTO" to
divide RX/TX equaly to booth external Antennas which works perfectly in a sectorized config with a full
metal shielded new outdoor case.

Good to know is, that the device has no problems with 21+22db / 126+158mW on 2.4 GHz. Still a cool device
without problems. The WRT610/E3000 has here a big lower maximum Limit (84 mW max when i remember correctly)

Basicly, when the 5 GHz USB-Module is not in use, this old router is perfect for some outdoor-Applications in a new Case
to drive external Antennas very stable. Newer Models are mostly limited in maximum TX-Power for the 2.4 GHz band.

Only downside is the not avail "g/n-only" Mode by web interface to get rid of the "b-support".
(is the same problem with arm-routers, too). BTW: dd-wrt has support for g/n-only on the same hardware.

Checking "wl", there is an option available to disable B-modes with one command, instead using "AUTO" which
always include b/g/n. I fix this with a wl command in the script section actualy.

[txnative777]

The "bl-trick for CPE 1.0.37" was the first i made with NVRAM direct command to have a better "cosmetic"

If you would like to do a "pr", you have that opportunity to do so at your convenience.

Checking the other Settings for mips-devices is a good idea, i took this "DE/7" from my EA6xxx-arm Routers quickly
and basicly is working and have all channels avail. I will check the mips-typical rev-codes, too shortly

Driver for sdk6 is much different to sdk5, what I had shown you were the list of channels for "DE".
If you get a chance you can check the "wl -i eth1 or eth2 country" the output will show you country/region, country/rev. "DE/7" but if you're have issues with channels then use what I sent in the previous message.
The sdk5 driver recognizes country and rev from wlc_channels.c file

Only downside is the not avail "g/n-only" Mode by web interface to get rid of the "b-support".
(is the same problem with arm-routers, too). BTW: dd-wrt has support for g/n-only on the same hardware.

"auto" works just fine with any wireless devices that support 802.11/B/G/N, dd-wrt uses a much different driver so "auto" may work you can select G+N for best desired connectivity to each wireless client device if you're expecting better performance.

Checking "wl", there is an option available to disable B-modes with one command, instead using "AUTO" which
always include b/g/n. I fix this with a wl command in the script section actualy.

I had seen this as well, if you feel the need to disable it, its your choice remember its open source.

[maverick3005]

Thx again for the hint with SDK5, with DE/7 i got:

DE (DE/0) GERMANY

After changing 2,4+5 GHz via web to "rev 3" i got:

wl -i eth1 country
DE (DE/3) GERMANY

wl -i eth2 country
DE (DE/0) GERMANY

So, again, i have to manually change the revision for 5 GHz USB-Module directly via NVRAM twice:

nvram set "usb/0xBD17/regrev=3"
reboot

and this time the result is now:

wl -i eth2 country
DE (DE/3) GERMANY

I´am not sure, which Routines still works with the default "wl1" Parameters, but importent can be to take a look
for:

usb/0xBD17/boardrev=0x1100
usb/0xBD17/regrev=3
usb/0xBD17/sromrev=8

... when it comes to specific commands to the 5 GHz Module. As long as the wl commands are send to "eth2",
all is fine, but for the device "0xBD17" there is a huge register list, pre setted from the cfe/default.

Comming to G/N-Mode:

This option is NOT available via the web interface for the 2.4 GHz Band, as i have stated before.
There is only available: AUTO, B-only, G-Only, B/G-Mixed and N-only - so N/G-Mixed is missing :-(
Setting a minimum Transmission Rate > 11 MBit to "disable B-Mode" is a very bad idea.
My scenario just needs "G/N-only", without the option for clients to use B-Mode as Fallback.
This have further todo with the Preamble "short" instead "long". N-Mode alone is still no good option,
because some clients (IoT) works better with G-Mode.

NB: G/N-only is on arm/AC-Routers the same problem: only B/G-Mixed is available. i have the same
"issue" with the AC-Routers, too - i like to finaly completly "clean" disable B-Mode on all my Routers.
Eventually this was forgotten in the Webinterface/Pull-Down Menu - in the "wl tool" such option is available.

It looks like it make sense to open an extra ticket for missing G/N-Mixed mode for all Routers :-)

[txnative777]

Thx again for the hint with SDK5, with DE/7 i got:

DE (DE/0) GERMANY

After changing 2,4+5 GHz via web to "rev 3" i got:

wl -i eth1 country DE (DE/3) GERMANY

wl -i eth2 country DE (DE/0) GERMANY

So, again, i have to manually change the revision for 5 GHz USB-Module directly via NVRAM twice:

nvram set "usb/0xBD17/regrev=3" reboot

and this time the result is now:

wl -i eth2 country DE (DE/3) GERMANY

Where are you making changes at?
I am using "advanced-wireless" -> country/region -> "Germany", country/rev -> "3"

`root@unknown:/tmp/home/root# wl -i eth1 country
DE (DE/3) GERMANY

root@unknown:/tmp/home/root# wl -i eth2 country
DE (DE/3) GERMANY
`
The code is working just fine, I am using 2025.3 build.

I´am not sure, which Routines still works with the default "wl1" Parameters, but importent can be to take a look for:

usb/0xBD17/boardrev=0x1100 usb/0xBD17/regrev=3 usb/0xBD17/sromrev=8

... when it comes to specific commands to the 5 GHz Module. As long as the wl commands are send to "eth2", all is fine, but for the device "0xBD17" there is a huge register list, pre setted from the cfe/default.

Those that have usb/0xBD17 are all associated with the 5GHz, except boardrev, srom, FreshTomato uses the boardrev as part of detection as well other driver related functions there is nothing to be done using an old broadcom blob for wl support on these usb 5GHz radio.

Comming to G/N-Mode:

This option is NOT available via the web interface for the 2.4 GHz Band, as i have stated before. There is only available: AUTO, B-only, G-Only, B/G-Mixed and N-only - so N/G-Mixed is missing :-( Setting a minimum Transmission Rate > 11 MBit to "disable B-Mode" is a very bad idea. My scenario just needs "G/N-only", without the option for clients to use B-Mode as Fallback. This have further todo with the Preamble "short" instead "long". N-Mode alone is still no good option, because some clients (IoT) works better with G-Mode.

NB: G/N-only is on arm/AC-Routers the same problem: only B/G-Mixed is available. i have the same "issue" with the AC-Routers, too - i like to finaly completly "clean" disable B-Mode on all my Routers. Eventually this was forgotten in the Webinterface/Pull-Down Menu - in the "wl tool" such option is available.

It looks like it make sense to open an extra ticket for missing G/N-Mixed mode for all Routers :-)

The optimal choice to use is "auto" IMHO using FreshTomato, older devices depending on their driver optimization to perform on one or more the 802.11 spectrum heavily relies on the wireless device intending usage on supporting one or more A.B/G/N/AC protocols, using an older wireless adapter for example may slow other wireless clients down in favor of supporting a client using 802.11/B.

If you would like to donate to one of the developers to look into adding the "G+N" mix selection added to the GUI in basic-network for MIPS, you may inquire then donate.

IMHO, I am able to reproduce the Country & Revision on my E3200 and I am not sure if the router can boot faster than under 90 seconds, there needs to a moments for a router to not be accessible for startup, nvram to settle before user intervention is done in 90 seconds.

[maverick3005]

Before i fix the USB-Device parameters equaly via NVRAM command to the same vales as with the webinterface,
my results was "unsetted" or rev 0 - when only use the webinterface, because it set the NVRAM variables only for wl1.
For correct setting the 5 GHz USB Module, basicly the data needs to "copy" to the USB-variables, too.

I saw today, that the other Modes on E3200 not "really working". N-only still advertise b/g/n - and is nearly the same as Auto.
G-only still have B-Mode enabled, so AUTO is more or less the only option which works correctly.
The other option is B/G-mixed, when supress N-Modes. So in the driver/interface for this device some options are still
missing or not compatible actualy. maybee the wl-tool doesn´t reflect all options the used brcm-driver really have ?

I use the actual trunk, far newer as 2025.3
Possibly with the branch-version the webinterface only works as expected without the usb-device extra settings.

As i say, the correct Port-Order with the wrong WAN-Port on the Info site is borked in the actual trunk a little bit.

90 seconds boot time is still huge. My Router comes up in less then 45-55 seconds from cold start. mfg_wait=OFF really makes a
huge difference. As i say: the standard 3 seconds tftp-client delay is enough for a manual tftp session in case of
"recovery".

Very interessting are the "undocumented" default register-settings for booth modules coming from the cfe and are untouched
from the running system. Playing here a little bit with some settings can disable the MIMO-Mode, when there is only
one Antenna physicly connected, or to overcome some other "hard" limitations. "Diversity-Switch-OFF" is such an option...

[txnative777]

Before i fix the USB-Device parameters equaly via NVRAM command to the same vales as with the webinterface,
my results was "unsetted" or rev 0 - when only use the webinterface, because it set the NVRAM variables only for wl1.
For correct setting the 5 GHz USB Module, basicly the data needs to "copy" to the USB-variables, too.

Are you actually using "advanced-wireless" to change any of the country/region, country/rev?
I am not seeing a problem on my E3200 setting Germany or 2, 3, 4, 5 rev.

If you're having to manually set these to nvram then I would suggest to perform a 30-30-30 or hard reset by pressing & hold the reset button for 10, then release.

2025.3 is working fine using "advanced-wireless", I had configured US/10 for obvious reason the variables recognize these updated changes as well.
wl1_country=US, wl1_country_code=US, wl1_country_rev=10

As i say, the correct Port-Order with the wrong WAN-Port on the Info site is borked in the actual trunk a little bit.

Understood this is a known issue.

90 seconds boot time is still huge. My Router comes up in less then 45-55 seconds from cold start. mfg_wait=OFF really makes a
huge difference. As i say: the standard 3 seconds tftp-client delay is enough for a manual tftp session in case of
"recovery".

Default is all that it is, users can make their own modification as they see fit. Now if you would like to create a how to and post it in the forum for anyone to use is entirely up to you to share.

Very interessting are the "undocumented" default register-settings for booth modules coming from the cfe and are untouched
from the running system. Playing here a little bit with some settings can disable the MIMO-Mode, when there is only
one Antenna physicly connected, or to overcome some other "hard" limitations. "Diversity-Switch-OFF" is such an option...

Remember FreshTomato is not the oem or the holder of the Broadcom drivers, these are closed source, what FreshTomato does give is the ability for the user to set one of the other options instead MIMO it also implies the importants to set the proper country/region, country/rev to work properly.

Any other concerns can be done in the forum, if you want to request the "G+N" mixed mode is entirely up to you or if you have experience in some coding clone Pedro311's repo, following the README.md to getting started setting up the build environment create your proposed fixes and push them to a "pr" for them to look over.

[maverick3005]

Oh, good idea, actualy i have the classic webinterface active, not the advanced. Will give it a try ;-)

For the other things i make some more testings the next days.

[txnative777]

i have the classic webinterface active, not the advanced.

I am using the classic theme, I am referring to "advanced-wireless.asp"

[maverick3005]

of course "advanced-wireless". Here is actually the problem that only wl1 is configured and ignored at the 5 GHz USB-Module,
because the "wl1-variables" for country+rev are NOT copied to the USB-Device-ID.

When you say in 2025.3-branch it is working without modify nvram variables for the usb modul, then there is a new bug in the actual
trunk, means this function is no longer working out of the Box ;-)

[txnative777]

of course "advanced-wireless". Here is actually the problem that only wl1 is configured and ignored at the 5 GHz USB-Module,
because the "wl1-variables" for country+rev are NOT copied to the USB-Device-ID.

Can you briefly describe the USB-Device-ID?

When you say in 2025.3-branch it is working without modify nvram variables for the usb modul, then there is a new bug in the actual
trunk, means this function is no longer working out of the Box ;-)

Looking at the nvram parameters over the usb/0xBD17/regrev=0 the driver is handling the both country/region, country/rev properly.

nvram set "usb/0xBD17/regrev=3"
reboot

There is no need to add or modify the nvram "usb/0xBD17/regrev" the wl driver sets regrev related WiFi sb/1/regrev=xx, usb/0xBD17/regrev=xx. there is nothing to do here, what your seeing is the default value for "usb/0xBD17/regrev=0"

[maverick3005]

In my case with "wl -i eth2 country" i got DE/0 when i not set the USB device by hand. Strange thing.

[txnative777]

In my case with "wl -i eth2 country" i got DE/0 when i not set the USB device by hand. Strange thing.

Had you performed a 30-30-30 or hard reset to clear out any bugs that may had accumulated?

[maverick3005]

To be sure on such things i always retest with entire cleaning to verify the findings ;-)