Solved GNP NVRAM setting question and Script advice

[kstamand]

Looking for scripting advice on how best to parse the Guest Network Pro (GNP) nvram vlan_rl value??

nvram vlan_rl value question
My setup has 3 custom GNP VLAN where the nvram setting vlan_rl has the value <1>52>0><2>53>0><3>54>0>
My understanding is that vlan_rl consists of columns of information >> SDN # (1, 2, or 3 in my case), Bridge Interface number (52, 53, or 54), and I'm not sure what the third field is for (0)
Question 1 - what is the third column of information in the nvram setting vlan_rl in each of the three SDN (0 in my case)?

Script question/advice, related to the nvram vlan_rl setting
My understanding is that the SDN number can be 1 - 2 characters, as can the Bridge Interface number, and I'm going to assume that third value can be 1 - 2 characters as well. I also understand there can be a variable number of SDN (1 - 9 or more)
Question 2 (I'm not sure how best to ask, but ..), is there some pattern matching magic that allows me to dynamically loop through the nvram vlan_rl string and pull out each SDN combination of information (SDN #, Bridge Interface #, third ??) into an array or separate variables - Something like SDN#: value, BR#: value, XX#, value??

Goal is to have a "somewhat" dynamic DNSMASQ-SDN.POSTCONF for dnsmasq customizations
Currently, my dnsmasq-sdn.postconf has hard coded SDN# references for adding different domain names (and potential other information) for each VLAN (e.g., vlan 1 = home.lan, vlan 2 = adults.lan, ...) and it's working well.
I'm just looking to enhance my scripting skills.

 

[dave14305]

Use the get_mtlan command to see how the nvram is broken up into relatable fields.

 

[rung]

Not sure if this is helpful, but when I needed the bridge identifier in a dnsmasq sdn post conf script, I used this kludge:

BR=$(nvram get subnet_rl | tr '<' '\n' | awk -F'>' '{print $1 ":" $2}' | grep $SDN":" | awk -F':' '{print $2}')

Where I had previously set the variable SDN to the second received parameter $2.

 

[kstamand]

Use the get_mtlan command to see how the nvram is broken up into relatable fields.

Interesting command with useful information for future reference - thank you

 

[kstamand]

Not sure if this is helpful, but when I needed the bridge identifier in a dnsmasq sdn post conf script, I used this kludge:

BR=$(nvram get subnet_rl | tr '<' '\n' | awk -F'>' '{print $1 ":" $2}' | grep $SDN":" | awk -F':' '{print $2}')

Where I had previously set the variable SDN to the second received parameter $2.

Thanks for the suggestion!!

 

[kstamand]

I ended up creating the following script to parse the nvram vlan_rl setting for my current needs, in case it may be helpful for others.

#!/bin/sh
# set -x # uncomment for debugging purposes

# read the list of vlan from nvram variable vlan_rl
vlan_list="$(nvram get vlan_rl)"

# count the number of vlan in the vlan_list string (each occurrence of the string "<" marks the beginning of a new set of vlan info)
vlan_count=$(echo $vlan_list | grep -o "<" | wc -l)

# get the length of the vlan_list string
vlan_list_length=${#vlan_list}

# calculate the length of each string of vlan info (sdn#, vlanid, ??) by dividing total vlan_list length by the number of vlan in list
let vlan_str_length=$vlan_list_length/$vlan_count

# set substring start position, for extracting individual vlan info set from vlan_list
substring_start=0

# loop through the vlan_list string, for the count of vlans identified
for i in 0 .. $vlan_count; do
    substring="${vlan_list:$substring_start:$vlan_str_length}" # extract a set of vlan info from the vlan_list string

    # strip out the '<' and '>' characters from the substring
    vlan_info="$(echo $substring | tr '<' ' ' | tr '>' ' ')"

    #now pick up each piece of information
    sdn_idx="$(echo $vlan_info | cut -f1 -d ' ')" # capture sdn index number
    vlan_id="$(echo $vlan_info | cut -f2 -d ' ')" # capture vlan id
    unknown_id="$(echo $vlan_info | cut -f3 -d ' ')" # capture unknown id

    echo "sdn_idx - $sdn_idx ; vlan_id - $vlan_id ; unknown_id - $unknown_id"

    #move the substring start position over to the next set of vlan info
    let substring_start=$substring_start+$vlan_str_length
done

 

[Martinski]

Just a couple of observations:

...
#now pick up each piece of information
sdn_idx="$(echo $vlan_info | cut -f1 -d ' ')" # capture sdn index number
vlan_id="$(echo $vlan_info | cut -f2 -d ' ')" # capture vlan id
unknown_id="$(echo $vlan_info | cut -f3 -d ' ')" # capture unknown id

The 3rd parameter (i.e. unknown_id) in each VLAN entry from the NVRAM vlan_rl variable indicates the VLAN Port Isolation (Enabled=1, Disabled=0). IOW, each VLAN entry consists of:

<SDN_Index>VLAN_ID>VLAN_Port_Isolation>

...
# calculate the length of each string of vlan info (sdn#, vlanid, ??) by dividing total vlan_list length by the number of vlan in list
let vlan_str_length=$vlan_list_length/$vlan_count
...
#move the substring start position over to the next set of vlan info
let substring_start=$substring_start+$vlan_str_length

The above code is incorrect because it assumes that all VLAN entries will always have equal length, which is not true for all possible scenarios. For example, the VLAN_ID can be between 1 to 4 digits (i.e. ID value between 1 to 4093).

My 2 cents.

 

[kstamand]

Just a couple of observations:


The 3rd parameter (i.e. unknown_id) in each VLAN entry from the NVRAM vlan_rl variable indicates the VLAN Port Isolation (Enabled=1, Disabled=0). IOW, each VLAN entry consists of:

<SDN_Index>VLAN_ID>VLAN_Port_Isolation>

The above code is incorrect because it assumes that all VLAN entries will always have equal length, which is not true for all possible scenarios. For example, the VLAN_ID can be between 1 to 4 digits (i.e. ID value between 1 to 4093).

My 2 cents.

Thank you for filling in the “unknown_id” blank for me and catching my vlan info length miscalculation. I’ll redo the code, unless you have an alternate suggestion on how to parse that type of sting + encoding(<sdn_index>vlan_id>vlan_port_isolation><..>..>..><..>..>..> ) ??

 

[rung]

Thank you for filling in the “unknown_id” blank for me and catching my vlan info length miscalculation. I’ll redo the code, unless you have an alternate suggestion on how to parse that type of sting + encoding(<sdn_index>vlan_id>vlan_port_isolation><..>..>..><..>..>..> ) ??

Not sure if this is correct, but I view the "<" as a next row indicator and the ">" as a next column indicator.

 

[kstamand]

Not sure if this is correct, but I view the "<" as a next row indicator and the ">" as a next column indicator.

That is my understanding as well. I just don’t know if it’s easier / clearer to attempt it the way I am In the sample code or if there is some parsing code technique I could be using (something like a JSON parser)

 

[rung]

That is my understanding as well. I just don’t know if it’s easier / clearer to attempt it the way I am In the sample code or if there is some parsing code technique I could be using (something like a JSON parser)

I use this to make these variables human readable:

nvram get vlan_rl | tr '<' '\n' | tr '>' '\t'

You can use tricks with awk and grep to pick out the parts you are looking for.

 

[kstamand]

I use this to make these variables human readable:

nvram get vlan_rl | tr '<' '\n' | tr '>' '\t'

You can use tricks with awk and grep to pick out the parts you are looking for.

That’s a hell of a lot easier - THANKS!!!

 

[dave14305]

You can also infer the nvram fields from the GUI code linked below.

asuswrt-merlin.ng/release/src/router/www/sysdep/FUNCTION/SDN/SDN/sdn.js at 8f2cb746b176952bb08031f38c2f6d08d655b722 · RMerl/asuswrt-merlin.ng

Third party firmware for Asus routers (newer codebase) - RMerl/asuswrt-merlin.ng

github.com

 

[kstamand]

You can also infer the nvram fields from the GUI code linked below.

asuswrt-merlin.ng/release/src/router/www/sysdep/FUNCTION/SDN/SDN/sdn.js at 8f2cb746b176952bb08031f38c2f6d08d655b722 · RMerl/asuswrt-merlin.ng

Third party firmware for Asus routers (newer codebase) - RMerl/asuswrt-merlin.ng

github.com

VERY helpful, much appreciated

 

[Martinski]

Thank you for filling in the “unknown_id” blank for me and catching my vlan info length miscalculation. I’ll redo the code, unless you have an alternate suggestion on how to parse that type of sting + encoding(<sdn_index>vlan_id>vlan_port_isolation><..>..>..><..>..>..> ) ??

Here's a simple function that parses the NVRAM value and iterates through each VLAN entry:

_Extract_NVRAM_vlan_rl_INFO_()
{
    echo "VLAN_INDX,VLAN_ID,PORT_ISOLATION"
    while read -r VLAN_INDX VLAN_ID PORT_ISOLATION && [ -n "$VLAN_INDX" ]
    do  echo "${VLAN_INDX},${VLAN_ID},$PORT_ISOLATION"
    done <<EOT
$(nvram get vlan_rl | sed 's/></\n/g;s/</\n/g;s/>$//g;s/>/ /g' | sed '/^$/d')
EOT
}

You could modify the function to output the results to a temporary CSV file so other functions down the line can easily read and process the extracted information as needed.

HTH.

 

[kstamand]

Here's a simple function that parses the NVRAM value and iterates through each VLAN entry:

_Extract_NVRAM_vlan_rl_INFO_()
{
    echo "VLAN_INDX,VLAN_ID,PORT_ISOLATION"
    while read -r VLAN_INDX VLAN_ID PORT_ISOLATION && [ -n "$VLAN_INDX" ]
    do  echo "${VLAN_INDX},${VLAN_ID},$PORT_ISOLATION"
    done <<EOT
$(nvram get vlan_rl | sed 's/></\n/g;s/</\n/g;s/>$//g;s/>/ /g' | sed '/^$/d')
EOT
}

You could modify the function to output the results to a temporary CSV file so other functions down the line can easily read and process the extracted information as needed.

HTH.

Having a function, with results output to a file, was my end state goal of my initial code hack.

Interesting and educational for me, how this sample code accomplishes what I was attempting, but in a much more elegant fashion.

Thank you @Martinski, @dave14305, and @rung for all the tips and guidance. This forum and members like yourself are the best.

 

[Martinski]

Having a function, with results output to a file, was my end state goal of my initial code hack.

In that case, here's a leaner version of the function that does exactly what you want and nothing else:

# Global variable #
vlan_rl_CSVfilePath="$HOME/NVRAM_vlan_rl_INFO.CSV"

_Extract_NVRAM_vlan_rl_INFO_()
{
   {
      echo "VLAN_INDX,VLAN_ID,PORT_ISOLATION"
      nvram get vlan_rl | sed 's/></\n/g;s/</\n/g;s/>$//g;s/>/,/g' | sed '/^$/d'
   } > "$vlan_rl_CSVfilePath"
}

If you prefer, you could instead pass the CSV file path as a parameter to the function.,