[PLUG] FOLLOWUP: Portland Linux/Unix Group General Meeting: OpenWrt Clinic, Part 2

[Ted Mittelstaedt]

I'm following up on this I know it's old.

I think the warm boot hang affects other Linksys MediaTek products.  I think I saw the same exact thing with a Linksys EA7500v2 but I didn't realize it at the time.  I had flashed the device with the latest firmware OpenWRT 24.10 on it and did a few tests with it, used it for a week or so then set it aside.  A week later I wanted to config it - but like an idiot I had forgotten what IP address I had left on it - so I pin reset the thing.  It did not reboot.  I immediately suspected the new firmware (of course - a tech always assumes the most complicated thing broke instead of the simplest thing - dumb mistake, that) and through power-cycling got it to boot off it's alternate partition where it still had the factory firmware on that, then got it to boot of 24.10 - finally.

The Linksys EA7500v2 uses the MediaTek MT7615N, and it does have a 160Mhz wide radio in the 5Ghz band, while the Linksys E8450 has a MediaTek MT7622BV/MT7915E chip.  The EA7500v2 is a wifi-5 chip while the E8405 is a Wifi-6 chip.  While you would think the architectures of the devices are different enough that the bug wouldn't be present in both - I suspect Linksys/Belkin is just a cheap company that uses crummy NAND suppliers.

Ted

-----Original Message-----
From: PLUG <plug-bounces at lists.pdxlinux.org> On Behalf Of Russell Senior
Sent: Friday, January 3, 2025 9:19 AM
To: plug at lists.pdxlinux.org
Subject: Re: [PLUG] FOLLOWUP: Portland Linux/Unix Group General Meeting: OpenWrt Clinic, Part 2



On 1/3/25 08:22, Ted Mittelstaedt wrote:
> Any way to detect the problematical ram chips so the scaling_min_freq 
> is a conditional setting based on the presense or absence of the 
> chip?you

I don't think so. The memory is all initialized by the bootloading process before the kernel gets any control, so (i think) the kernel just sees a range of addresses. I think the primary clue that a device is affected is going to be the early vendor firmware version and the fact that it hangs during the sysupgrade process. So, you probably find out about it before it goes out the door and can modify it as needed.

--
Russell Senior
russell at pdxlinux.org