Gargoyle 1.15.x BETA - 2024-03-11 - Based on OpenWrt 23.05

[pythonic]

Noting that hardware/software offload are somewhat incompatible with some Gargoyle features

Indeed, hardware/software offload are probably akin to a footgun in Gargoyle, which is a good reason for them never to make it into the UI the way they are in OpenWrt.

[JohnPeterson]

I see the Linksys WRT1900ACS is avail again. I get screen message when attempting an upgrade.

Firmware/Device Compatibile:No. Config cannot be migrated from swconfig to DSA

Thanks.

[pythonic]

I see the Linksys WRT1900ACS is avail again. I get screen message when attempting an upgrade.

Firmware/Device Compatibile:No. Config cannot be migrated from swconfig to DSA

That is correct - there is unfortunately no way to preserve the network configuration between swconfig and DSA builds so upgrading will require configuring Gargoyle from scratch (i.e. don't attempt to preserve the configuration). Except for the fact that the mvebu target wasn't supported in Gargoyle v1.14 this transition would have happened with that version.

You may still need to check the "force" option to be able to proceed but as long as you don't attempt to preserve the configuration and the other device/image identity checks are ok you should be able to successfully upgrade.

You should backup your configuration before the upgrade - even though you won't be able to restore it on the new version you can still get access to the information in it (e.g. using 7Zip) for reference purposes.

[oliver44]

Hello all,

@pythonic thanks for the guidance, it seems to work!
There is no one to deal with integrating all this in default for gargoyle and of course to optimize?
with Disable Bandwidth Monitor!
https://ibb.co/hCgwzm2
https://ibb.co/Dp9gDbK
https://ibb.co/QmXR3yJ
https://ibb.co/DCZmL1C


with Bandwidth Monitor activated!

https://ibb.co/7JNrCBm
https://ibb.co/d7xngjL
https://ibb.co/1QpMqzD
https://ibb.co/jzb8NB8

I understand that Bandwidth Monitor increases the consumption of resources, but how does Active Bandwidth Monitor influence a better performance report in the speed test?

nothing can be done to optimize all this?

[oliver44]

I would like to install some plugins that I also used on wrt1900acs v2 but I don't know what the command is for GL-MT6000, for wrt1900acs v2 I use this:
https://lantisproject.com/gargoyle_cust ... bu/default
https://lantisproject.com/gargoyle_cust ... elspecific

[Lantis]

Hello all,

@pythonic thanks for the guidance, it seems to work!
There is no one to deal with integrating all this in default for gargoyle and of course to optimize?
with Disable Bandwidth Monitor!
https://ibb.co/hCgwzm2
https://ibb.co/Dp9gDbK
https://ibb.co/QmXR3yJ
https://ibb.co/DCZmL1C


with Bandwidth Monitor activated!

https://ibb.co/7JNrCBm
https://ibb.co/d7xngjL
https://ibb.co/1QpMqzD
https://ibb.co/jzb8NB8

I understand that Bandwidth Monitor increases the consumption of resources, but how does Active Bandwidth Monitor influence a better performance report in the speed test?

nothing can be done to optimize all this?

I’ll poke around, but the bwmon is what it is. Without a fundamental rewrite it just is slow. The faster you go the more of an impact it has.

[Lantis]

I would like to install some plugins that I also used on wrt1900acs v2 but I don't know what the command is for GL-MT6000, for wrt1900acs v2 I use this:
https://lantisproject.com/gargoyle_cust ... bu/default
https://lantisproject.com/gargoyle_cust ... elspecific

https://lantisproject.com/gargoyle_cust ... k/filogic/
There’s a kernel one as well.

You just need to pick the right subtarget.
The mt6000 is a mediatek/filogic device

[pythonic]

@pythonic thanks for the guidance, it seems to work!

Could you please summarise what you've ended up installing/setting to get these results?

I understand that Bandwidth Monitor increases the consumption of resources, but how does Active Bandwidth Monitor influence a better performance report in the speed test?

The difference is down to 1% or less; even though there seems to be a consistent trend I would suggest this is on the edge of measurement accuracy (especially with web browser based speed tests). It also is more than likely highly hardware dependent (yes, multiple MT6000s might show the same tendency, but Asus TUF-AX6000s might not despite apparently being very similar hardware) and also data pattern dependent (e.g. browser speedtest vs ftp download of an iso image).

One possibility as to why it happens might be that specific patterns of heavy load (bwmon active in your examples) keeps certain CPU/kernel caches "hot" with the right code and data while very light load leaves those same caches "cold". There are no doubt other possibilities that only a serious kernel developer (which I'm not) could contemplate.

Be happy with what you've got - otherwise you're at risk of letting perfect get in the way of good.

[rg66]

A quick test with minimal config

Gargoyle 1.14 official
Gargoyle Version:1.14.0
Model:TP-Link TL-MR3420 v5
Memory Usage:21.8MB / 57.1MB (38.1%)
Root Drive 320.000 kBytes Total, 80.000 kBytes Free

Gargoyle 1.15.x BETA
Gargoyle Version:1.15.X (Built 20240214-1128 git@87065936)
Model:TP-Link TL-MR3420 v5
Memory Usage:23.5MB / 57.6MB (40.8%)
Root Drive 896.000 kBytes Total, 656.000 kBytes Free

Custom lite build from master branch
Gargoyle Version:1.15.X (Built 20240308-1156 git@ae547ce5)
Model:TP-Link TL-MR3420 v5
Memory Usage:17MB / 56.2MB (30.2%)
Root Drive 1.500 MBytes Total, 1.273 MBytes Free

[oliver44]

Could you please summarise what you've ended up installing/setting to get these results?

it works like this only activated with all