only enumerate if we don’t update the device tree to the correct line that the reset pin is using
I suspect this will have to do with alternate pin functions of that RESET pin, but before I claim to know the answer, can you share the working and failing device tree snippets?
when the driver is loaded it fails
This is starting to look the same as the errors in Mmc1: Timeout waiting for hardware interrupt.
While the solution in that thread is to disable ADMA, which you can do by adding sdhci-caps-mask = <0x0 0x008000>; to the mmc device tree node, your error is occurring earlier in the initialisation sequence.
Please add this sdhci-caps-mask as you will need it on this platform.
When I saw the mmc timeout with Argument: 0x149a4004 it was due to dips in my supply voltage caused by a badly mod-wired board creating some spurious edges on the SDIO lines - see this post. The response to this command argument is the first time the Morse Micro chip drives the SDIO data lines. Unfortunately the captures you’ve provided are only 20MS/s so I can’t see the same detail for that argument. Are you able to capture at a higher sample rate?