The return of RCM Boot-Loop-of-death

[Xemnass]

Hi everyone!
During the summer i've posted about my faithfull switch being... lost >.< instead, after a night having drained all the power from the battery, magically, the day before it came bace to life and i was able to re-enable autoRCM and enjoy it again.
Now, like this summer, after having it sit in the original dock, randomly crashed. Same behaviour like the other time, no boot at all. The strange thing is that i was on the emuMMC but the boot issue appeard even in the untouched stock nand.

Till now i've tried everything. I retain the ability to use hekate so i tried:

• rebuild mmc from scratch
• restore a perfect backup i've done asap i was able to use heakte again during the summer

but after all this attempt, and having Scires read the fatal error, it gave an erro in mesospehre/package 2 if i run from hekate,
if i run it direclty with the payload (fussee) all the screen, on the atmosphere logo gets pixellated and an abort message appears.
And i'm forced to to hard-reset it.
Everything is up-to-date, i even saved the sd setup of this summer but, after every attempt to boot it came back to RCM
with "KRNL 0xF000" issue.

Someone can help me? Pleeeease

 

[thesjaakspoiler]

It sounds like something is corrupted.
Maybe try a clean install of Atmosphere again.
Something like a perfect backup does not exist.
It could have been corrupted as well already.

 

[Xemnass]

Hi ,
thanks for replying.

I've already tried a clean install of atmosphere,
the nand was, i think, good because was a backup made after factory reset with nintendo safe mode.

The question here are plenty, even for the scene itself:

• why does connecting to the docking station, i a bunch of case found here in the forum, and in the internet, mess up badly with the switch, even from the emummc to sysnand?
• what cause it?

Because i even tried to follow all the steps for rebuilding the emmc with the System restore v3 script from tegra explorer and i am not able to get it back to work again. Even after that it still does RCM boot loop, or, in the case of the fusèe payload, the corruption you see.

Any sugestion? Maybe antoher way to rebuild the emmc?

 

[The Real Jdbye]

Hi everyone!
During the summer i've posted about my faithfull switch being... lost >.< instead, after a night having drained all the power from the battery, magically, the day before it came bace to life and i was able to re-enable autoRCM and enjoy it again.
Now, like this summer, after having it sit in the original dock, randomly crashed. Same behaviour like the other time, no boot at all. The strange thing is that i was on the emuMMC but the boot issue appeard even in the untouched stock nand.

Till now i've tried everything. I retain the ability to use hekate so i tried:

• rebuild mmc from scratch
• restore a perfect backup i've done asap i was able to use heakte again during the summer

but after all this attempt, and having Scires read the fatal error, it gave an erro in mesospehre/package 2 if i run from hekate,
if i run it direclty with the payload (fussee) all the screen, on the atmosphere logo gets pixellated and an abort message appears.
And i'm forced to to hard-reset it.
Everything is up-to-date, i even saved the sd setup of this summer but, after every attempt to boot it came back to RCM
with "KRNL 0xF000" issue.

Someone can help me? Pleeeease

You can't boot untouched stock NAND with AutoRCM, it literally corrupts the NAND so CFW is required to even make it boot. So if there is an issue where CFW won't boot it makes sense that would apply to sysNAND as well, might be an issue with your SD card.

 

[Xemnass]

You can't boot untouched stock NAND with AutoRCM, it literally corrupts the NAND so CFW is required to even make it boot. So if there is an issue where CFW won't boot it makes sense that would apply to sysNAND as well, might be an issue with your SD card.

So yes, this is correct, i'm aware of how AutoRCM works. The 'strange thing' is, indeed, that the problem rose with the emuMMC and the got spread even to the cfw on sysMMC. Maybe i express my self wrongly.

Yes i do think, at this point, thah maybe, the problem resides in some kind of corruption of my own SDCard? Can it be possible?
Because it *should* work, it give me no error even if a check disk with terminal etc... but it is the last reasonable thing to think, i actually ordered by amazon a new sd card just in case.

But if anyone has other suggestion in the meantime i'm all listening. Maybe we can found out something "new" or unepected.

 

[linuxares]

I would go 100% stock and try to boot it. If normal OFW works, then your Switch is fine. But this really looks like your NAND is dying.

 

[The Real Jdbye]

So yes, this is correct, i'm aware of how AutoRCM works. The 'strange thing' is, indeed, that the problem rose with the emuMMC and the got spread even to the cfw on sysMMC. Maybe i express my self wrongly.

Yes i do think, at this point, thah maybe, the problem resides in some kind of corruption of my own SDCard? Can it be possible?
Because it *should* work, it give me no error even if a check disk with terminal etc... but it is the last reasonable thing to think, i actually ordered by amazon a new sd card just in case.

But if anyone has other suggestion in the meantime i'm all listening. Maybe we can found out something "new" or unepected.

The quickest way to find out would be to try another SD card with a fresh download of Atmosphere/Hekate and nothing else and see if you can boot into sysNAND.

 

[Xemnass]

The quickest way to find out would be to try another SD card with a fresh download of Atmosphere/Hekate and nothing else and see if you can boot into sysNAND.

Yes. The SDCard should arrive by the end of the day ( i'm in Italy so here it's 10.19 AM).
In the "worst case scenario" have you any other sugestion?

 

[linuxares]

The quickest way to find out would be to try another SD card with a fresh download of Atmosphere/Hekate and nothing else and see if you can boot into sysNAND.

Or remove AutoRCM and boot without SD Card.

 

[Xemnass]

I would go 100% stock and try to boot it. If normal OFW works, then your Switch is fine. But this really looks like your NAND is dying.

mhm this is wat i suspect. I ran the banchmark hopping it gave me some indication but no indication on it. I think replace a NAND is not something easy to do >.<. Teh problem is that, during the summer, even without the RCM dind't boot. I left it, unaware, to dry the battery and the day after it started to work again... I'm afraid this is a nand-dying issue....

 

[linuxares]

mhm this is wat i suspect. I ran the banchmark hopping it gave me some indication but no indication on it. I think replace a NAND is not something easy to do >.<. Teh problem is that, during the summer, even without the RCM dind't boot. I left it, unaware, to dry the battery and the day after it started to work again... I'm afraid this is a nand-dying issue....

Well it is possible to actually "move" to a new NAND. If you got your boot 0/1 and a NAND backup, it's possible to overwrite a hardware nand. Since people have upgraded NANDs here before

Oh and you need your prod.keys

 

[Xemnass]

Well it is possible to actually "move" to a new NAND. If you got your boot 0/1 and a NAND backup, it's possible to overwrite a hardware nand. Since people have upgraded NANDs here before

Very good, i should have everything i need, actually, from this switch i got a full nand backup when it started to work again in the summer, it was the firtst thing i've made, and i have plenty of them with all my keys so if i'm not mistaken i should do the following:

• do however a new backup, just in case, and try to extarct again the keys, just in case
• trhow away in the garbage the eMMC module
• replace it ( i was afrai that was a soldering issue instead looking online i saw it a mere "pop-it" thing)
• flash the full rawnand.bin backup (should restore the prodinfo etc...)
• falsh boot0/1

and at that point, if all is ok, nothing should prevent it from booting again. Right?

 

[linuxares]

1) Try to see if you can get your switch working
2) If Nand is broken/going bad. Keep your backups, especially the clean ones!
3) You don't need to solder if you can find a donor with full card etc. Since it's just to pop out and plug in the "new" one. example from here - https://www.aliexpress.com/item/33007086105.html?spm=a2g0o.productlist.0.0.5d8190368x3StL&algo_pvid=8e35dd4a-4069-4f3d-b6e9-95e858a84cb4&algo_exp_id=8e35dd4a-4069-4f3d-b6e9-95e858a84cb4-7&pdp_ext_f={"sku_id":"12000022914087953"}

Most people that bought this on Aliexpress actually have the same issue you do.

4) I don't normally recommend Video guides, but since the procedure is still the same -

 

[Xemnass]

1) Try to see if you can get your switch working
2) If Nand is broken/going bad. Keep your backups, especially the clean ones!
3) You don't need to solder if you can find a donor with full card etc. Since it's just to pop out and plug in the "new" one. example from here - https://www.aliexpress.com/item/33007086105.html?spm=a2g0o.productlist.0.0.5d8190368x3StL&algo_pvid=8e35dd4a-4069-4f3d-b6e9-95e858a84cb4&algo_exp_id=8e35dd4a-4069-4f3d-b6e9-95e858a84cb4-7&pdp_ext_f={"sku_id":"12000022914087953"}

Most people that bought this on Aliexpress actually have the same issue you do.

4) I don't normally recommend Video guides, but since the procedure is still the same -

So for aliexpress is a guesswork, i had some money to spend and ordered 2 from different vendors, after all were about 15€ each and i had a coupon.

So are there any more trust worthy sites? Or i should find a "broken" switch and try to get the emmc module? >.<

 

[linuxares]

I've used Aliexpress without issues. I guess checking Ebay or local trade page?

 

[Xemnass]

I've used Aliexpress without issues. I guess checking Ebay or local trade page?

yeah good idea, well if yo used aliexpress without issues i think, in the mean time, to wait. I try to go to a local game stop to see if there are any borken switch willing to be "garbage collected" ahah in the meantime. Best have more chance than less!
For now thankyou!

If anyone has other suggestion i'm happy to gather much inforation as possible

I'll update you as soon as i have news!

 

[linuxares]

Honestly, remove AutoRCM and try to boot it.
Remove the SD Card as well.

 

[The Real Jdbye]

Very good, i should have everything i need, actually, from this switch i got a full nand backup when it started to work again in the summer, it was the firtst thing i've made, and i have plenty of them with all my keys so if i'm not mistaken i should do the following:

• do however a new backup, just in case, and try to extarct again the keys, just in case
• trhow away in the garbage the eMMC module
• replace it ( i was afrai that was a soldering issue instead looking online i saw it a mere "pop-it" thing)
• flash the full rawnand.bin backup (should restore the prodinfo etc...)
• falsh boot0/1

and at that point, if all is ok, nothing should prevent it from booting again. Right?

Kind of, the rawnand.bin isn't actually the full NAND, the full NAND is boot0+boot1+rawnand (+ partition layout) IIRC, in order to flash a backup to a blank NAND you need to jump through some hoops since Hekate won't let you flash if it doesn't see a Switch partition layout already on the flash chip. By using this it should allow you to restore rawnand/boot0/boot1 with Hekate afterwards: https://github.com/rajkosto/gptrestore
If you're replacing the NAND anyway you could consider upgrading it to a larger size, 32 GB isn't much after all.

 

[Xemnass]

Hi everyone, whishing you all an happy new year!

So, with the new year finaly came the 2 replacement i've ordered months ago but, unfortunately, after following all the suggestion again gives me the same screens and error. The screens i've posted on the first message. Is there anyone that may have other suggestions?

 

[Xemnass]

So, just to update the topic >.<
Out of the pure randomness, while i was trying to get more logs for the issue i've opened in github for Scires to see, it just booted up again without the glitch

This left a question open... what was it? Due to the fact the in both of the original cases when the switch froze and lead to all of this was connected and left charging at the docking station (the original one) can it be some strange voltage? Somehting that was able to affect the boot process of the EmuMMC and of the SysNand? can it be some kind of issue relating to the battery?