That's what I thought But I wanted to verify as I do not like acronyms. Too many have the same acronym but are completely different.Out of bound.
Well, in a correct implementation, trimmer ROMs should still return valid padding ( 00s ? or whatever the padding is ) data when reads beyond the trimmed point are done. Being a trimmed ROM doesn' mean the MiG Flash shouldn't behave as if the ROM was a verbatim copy...That's what I thought But I wanted to verify as I do not like acronyms. Too many have the same acronym but are completely different.
Obviously, trimmed ROMs would throw out of bounds errors.
Is there a possibility that Nintendo can see that XCI itself doesn't contain the certificate as it's separate? I also wonder if there is more inside a legit cart than just the ROM image.
outside of the boundariesOOB Reads?
I mean, you can detect MIG "just" using nxdumptool too... try to produce an untrimmed dump of a 1.2.0 MIG...Considering that the homebrew "sphaira" can easily detect if the current ROM loaded in the MIG Flash is trimmed or not, there is no reason to believe that Nintendo can't do the same.
That was in 1.1.9, now, with 1.2.0, it correctly reports all MIG as trimmed, regardless of the used XCI... which is worrying to me!Considering that the homebrew "sphaira" can easily detect if the current ROM loaded in the MIG Flash is trimmed or not, there is no reason to believe that Nintendo can't do the same.
You don't need to wonder, everything is on here https://switchbrew.org/wiki/XCI.That's what I thought But I wanted to verify as I do not like acronyms. Too many have the same acronym but are completely different.
Obviously, trimmed ROMs would throw out of bounds errors.
Is there a possibility that Nintendo can see that XCI itself doesn't contain the certificate as it's separate? I also wonder if there is more inside a legit cart than just the ROM image.
What I meant was maybe there is stuff inside the chip that we don't know about yet that dumpers aren't able to properly dump. Something like a fake challenge where the MIG reports TRUE when it's always FALSE or should not be responded to. I know this is a gross oversimplification.You don't need to wonder, everything is on here https://switchbrew.org/wiki/XCI.
Why would it report trimmed, even on a full dump? That makes no sense. Shouldn't it always repot untrimmed?That was in 1.1.9, now, with 1.2.0, it correctly reports all MIG as trimmed, regardless of the used XCI... which is worrying to me
Because it reports trimmed when the cart fails with specific OOB reads. With 1.2.0 MIG tries to fake a correct response when the Switch does that and they partially succeeds, in fact it seems to confuse the actual Switch 2... but this new implementation fails when you try to read a large quantity of OOB data like sphaira and nxdumptool do.What I meant was maybe there is stuff inside the chip that we don't know about yet that dumpers aren't able to properly dump. Something like a fake challenge where the MIG reports TRUE when it's always FALSE or should not be responded to. I know this is a gross oversimplification.
Post automatically merged:
Why would it report trimmed, even on a full dump? That makes no sense. Shouldn't it always repot untrimmed?
Shouldn't a FULL untrimmed dump be reported as full and not trimmed. What is the expected response on a real cart?Because it reports trimmed when the cart fails with specific OOB reads. With 1.2.0 MIG tries to fake a correct response when the Switch does that and they partially succeeds, in fact it seems to confuse the actual Switch 2... but this new implementation fails when you try to read a large quantity of OOB data like sphaira and nxdumptool do.
outside of the boundaries
Post automatically merged:
I mean, you can detect MIG "just" using nxdumptool too... try to produce an untrimmed dump of a 1.2.0 MIG...
Post automatically merged:
That was in 1.1.9, now, with 1.2.0, it correctly reports all MIG as trimmed, regardless of the used XCI... which is worrying to me!
The untrimmed dumps come in 2,4,8,16 and 32gb. The exact size of the flash memory in the official carts, it's like an image including all blank unused areas. I imagine that this will stop reading at the end of the data, whereas a mig possibly leaks something that can be read in the additional storage the SD card has or from one of the other files.Shouldn't a FULL untrimmed dump be reported as full and not trimmed. What is the expected response on a real cart?
They’re easy to find: they’re in your original cart.Just bought one of these ( just to have it / potentially use it with Switch 2 if I feel like getting banned) I already have a banned Switch V1 with atmosphere... didn't realize this would be such a pain in the ass to use as you need to find .bin and initial data files apparently.
Just insert the cart (either the original you dumped or the MIG) and update online, they (should) behave the same.Updating games / DLC on a stock console seems like it's not possible either. Better than nothing I suppose.
I wouldn't worry too much about that oob issue, just use untrimmed ROMs and you'll be fine there.They’re easy to find: they’re in your original cart.
Post automatically merged:
Just insert the cart (either the original you dumped or the MIG) and update online, they (should) behave the same.
But I’d be cautious until the OOB reading issue isn’t further investigated.
It has been proved with different experiments you can replicate that using untrimmed dumps worked until firmware 1.1.9 (well except N was able to detect the MIG on S2)… with 1.2.0 they tried to implement OOB reads regardless of the completeness of the dump. And unfortunately their attempt can be detected, so with 1.2.0 OOB reads can unveil the MIG regardless of using an untrimmed or trimmed dump.I wouldn't worry too much about that oob issue, just use untrimmed ROMs and you'll be fine there.
The bigger issue is the sector access speed measurement that was introduced in a recent FS version.
If MIG can't fix that via software, then you can be 100% sure that you will always be flagged for ban.
Nintendo knows the speeds of the flash chips they use for game cartridge production.Regarding the FS speed, is that a conjecture or is it a proven thing?
I mean when dumping my OG carts and testing the MIG I noticed that first party games can be read quicker than the flash cart, but my third party games (Cave Shmups) are equivalent in speed to a MIG. So testing speed seems unreliable to me unless using a proper per game database.
I see, I mean, we can only speculate what Nintendo uses for effectively flagging MIGs (maybe speed isn't so accurate, I'm thinking to temperature variations, to give a ban)... but OOB reads are a speculation such as speeds. We only know that 1.1.9 were detected by Switch 2, 1.2.0 seems to not be detected. And we know 1.2.0 altered OOB reads, but we don't know if that's the only change and if that's the culrpit.Nintendo knows the speeds of the flash chips they use for game cartridge production.
There are probably only a handfull of possibilities, but if the MIG executes anything slower or faster (micro SD card is also a factor), then it's easy sailing for Nintendo.
The code is implemented amd the results are probably already part of their captured telemetry.
It's MIGs turn now to do countermeassures, from a user POV there is nothing you can do.
I mean sure if you are using as intended. Nintendo seems to be banning people for using their backups so they can't update games though, at least on switch 2.They’re easy to find: they’re in your original cart.
Post automatically merged:
Just insert the cart (either the original you dumped or the MIG) and update online, they (should) behave the same.
But I’d be cautious until the OOB reading issue isn’t further investigated.