Skip to content

Modified RAM Requirements for F/W Flash Process#480

Merged
ExtremeFiretop merged 1 commit intoExtremeFiretop:devfrom
Martinski4GitHub:dev
Jun 3, 2025
Merged

Modified RAM Requirements for F/W Flash Process#480
ExtremeFiretop merged 1 commit intoExtremeFiretop:devfrom
Martinski4GitHub:dev

Conversation

@Martinski4GitHub
Copy link
Collaborator

Made changes in the memory requirements for routers with 512MB RAM to be less restrictive and allow the process to continue with the F/W flash.

Made changes in the memory requirements for routers with 512MB RAM to be less restrictive and allow to proceed with the F/W flash.
@Martinski4GitHub
Copy link
Collaborator Author

@ExtremeFiretop,

Please take a look and review the PR changes when you have time. The changes mainly affect routers with 512 MB RAM, but I did make slight modifications for routers with 1GB RAM or more.

Essentially, for routers with 512MB RAM, the implementation has 2 fronts:

a) Lower the initial required RAM overhead from 50% to 35% before the ZIP file is downloaded, and then lower it again to 25% after the F/W image is extracted.

I tested this with my own RT-A86U, and it worked fine, although I did not reach the low memory condition. I also tested with my cousin's router, but I didn't do the full F/W flash as I didn't want to interfere with their home network (one of his teenage kids was playing some game very late at night - summer vacation and out of school - so I was careful not to overstay my welcome).

b) When trying to acquire more available RAM, the code attempts to disable some ASUS/TrendMicro services:
AsusNat Tunnel, AiProtection, Traffic Analyzer/Monitor, and Bandwidth Monitor

This part I did not test with my cousin's router. On my RT-AC86U, I tested with AiProtection services enabled, and it worked to disable the service, kill the processes, flash the F/W image, and reboot.

I'll notify the user @DrEthan77 so that they can test this latest implementation on their own RT-AX86S router.

@ExtremeFiretop
Copy link
Owner

I tested it on my Gnuton router since it's the only one with 512Mb of RAM I have.
This was the results to see the new RAM requirements:
image

Then I put the memory reporting in debug and forced a specific value to see the process complete which looked like this:
image

Approved!

@ExtremeFiretop ExtremeFiretop merged commit 3df2e77 into ExtremeFiretop:dev Jun 3, 2025
1 check passed
@ExtremeFiretop
Copy link
Owner

ExtremeFiretop commented Jun 3, 2025

I like the fact that this is a hybrid of my suggestion and your suggestion, to essentially lower the required overhead once the extraction and download is completed.

But also to shutdown the TrendMicro stuff for memory

@Martinski4GitHub
Copy link
Collaborator Author

I tested it on my Gnuton router since it's the only one with 512Mb of RAM I have. This was the results to see the new RAM requirements: image

Then I put the memory reporting in debug and forced a specific value to see the process complete which looked like this: image

Approved!

Great tests and validation!!

@Martinski4GitHub
Copy link
Collaborator Author

I like the fact that this is a hybrid of my suggestion and your suggestion, to essentially lower the required overhead once the extraction and download is completed.

But also to shutdown the TrendMicro stuff for memory

Yes, the implementation attempts to solve the low-memory condition on the 2 fronts. I think the key is to lower the overhead after we've already downloaded the ZIP file and extracted its contents, because at that point we've already used quite a bit of the overhead, especially for routers with only 512 MB RAM.

@ExtremeFiretop
Copy link
Owner

I like the fact that this is a hybrid of my suggestion and your suggestion, to essentially lower the required overhead once the extraction and download is completed.
But also to shutdown the TrendMicro stuff for memory

Yes, the implementation attempts to solve the low-memory condition on the 2 fronts. I think the key is to lower the overhead after we've already downloaded the ZIP file and extracted its contents, because at that point we've already used quite a bit of the overhead, especially for routers with only 512 MB RAM.

I agree, if somehow this comes up again; I'm going to claim the guy is simply too close to the edge because this issue has been investigated on all fronts at this point (or so I think?)

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants