Hacking [RCM Payload] Hekate - CTCaer mod

[whyyes]

Got to 91% this time. It didn't write a partial.idx this time though and attempted to restart from 0%. When I ran the 1.3 version it did create a partial though I deleted it before trying your 1.5 beta. So I'm going to try 1.3 until it errors then continue with 1.5 I guess since I know it can recognize the partial. It just does not like something in part 13 at all.

Edit: Unbelievably this finally worked. I feel like it must time out at some point which caused the error for me at 90%ish. Let me know if this looks right

 

[jetracer]

I was getting constant error 1 messages on my 4.0.1. Have a 64gig sandisc formatted in exfat. Ater doing a full format not a quick format it worked 100% and was MUCH faster than before. It was done in maybe a half hour or so. It took over an hour to get to 10% before the error the last 4 times i tried backing up. So for anyone who has issues fully wipe the card and try again.

 

[kamo]

@CTCaer it wasn't easy to get a good pic. Hope it helps. 32GB class 4 (32KB a. size)

 

[CTCaer]

Got to 91% this time. It didn't write a partial.idx this time though and attempted to restart from 0%. When I ran the 1.3 version it did create a partial though I deleted it before trying your 1.5 beta. So I'm going to try 1.3 until it errors then continue with 1.5 I guess since I know it can recognize the partial. It just does not like something in part 13 at all.

Edit: Unbelievably this finally worked. I feel like it must time out at some point which caused the error for me at 90%ish. Let me know if this looks right

Yes they look correct in size. Check the forum for the total size in bytes. It ends in 320.

I was getting constant error 1 messages on my 4.0.1. Have a 64gig sandisc formatted in exfat. Ater doing a full format not a quick format it worked 100% and was MUCH faster than before. It was done in maybe a half hour or so. It took over an hour to get to 10% before the error the last 4 times i tried backing up. So for anyone who has issues fully wipe the card and try again.

Normally that's the reason. FatFS, the small size filesystem, cannot operate like windows or fat-fuse, which after many retries can skip the bad sectors and write elsewhere. So your full format, either unblocked these bad areas, or they got marked for avoidance in the FS records.

@CTCaer it wasn't easy to get a good pic. Hope it helps. 32GB class 4 (32KB a. size)

I can see the text clearly.

 

[hippy dave]

Oh good question about the cluster size. Mine was whatever MacOS Disk Utility used as default for ExFAT on a 32gb card, but I can check later.

FWIW my cluster size was 32K(i)B.

 

[CTCaer]

@CTCaer it wasn't easy to get a good pic. Hope it helps. 32GB class 4 (32KB a. size)

So this helped a lot. Thanks

If I followed it through correctly, it's because of this:
/* Could not allocate a new cluster (disk full) */
So, is this the case? how much space left after that failed message? And what size was the part it failed?

Additionally, I will create a more thorough (with messages, instead of codes this time) debug version on this part of code.
(I'll also make it print what was the starting available space, what's the new one, how many parts it was going to dump, and at which LBA was currently trying, before spitting the error message).
Are you willing to try one more time?

 

[kamo]

So this helped a lot. Thanks

If I followed it through correctly, it's because of this:
/* Could not allocate a new cluster (disk full) */
So, is this the case? how much space left after that failed message? And what size was the part it failed?

Additionally, I will create a more thorough (with messages, instead of codes this time) debug version on this part of code.
(I'll also make it print what was the starting available space, what's the new one, how many parts it was going to dump, and at which LBA was currently trying, before spitting the error message).
Are you willing to try one more time?

The card wasn't full at that point. I'm pretty sure it was part 9 (it could be 8 but I'm 99% sure it was 9). It gave me a proper "card full" message at part 10. It wasn't really full at that point either being a 32gb card and 10 parts should be around 23gb. I did copy the 10 parts and deleted them from the card. After that part 11 - 14 worked without any error. I think I have a good backup now.

 

[CTCaer]

The card wasn't full at that point. I'm pretty sure it was part 9 (it could be 8 but I'm 99% sure it was 9). It gave me a proper "card full" message at part 10. It wasn't really full at that point either being a 32gb card and 10 parts should be around 23gb. I did copy the 10 parts and deleted them from the card. After that part 11 - 14 worked without any error. I think I have a good backup now.

Yeah that was the intended work flow of 1.3. Move files and continue.
Even if an error occurs. But unfortunately, in 1.3 I coded it to write the partial.idx after the error occured. And ofcourse it could not write it.

Yeah your backup is OK now. Make sure to check the filesize after joining it.

The proper card full message is triggered when you have less than 2GB space available and you try to start or continue a dump.

So when you started the backup, you had ~18GB free space. So it should only dump the first 9 parts (.00 to .08).
If you saw a broken part .09 inside, this means the limit is not honored and it continued with the 10th (.09) part.
And that's why it stopped at [ff] 4 (which the check for available cluster) and got the error (1).

Anyway, I'll prepare a proper debug version, just in case, anyone else want to try.

 

[kamo]

Yeah that was the intended work flow of 1.3. Move files and continue.
Even if an error occurs. But unfortunately, in 1.3 I coded it to write the partial.idx after the error occured. And ofcourse it could not write it.

Yeah your backup is OK now. Make sure to check the filesize after joining it.

The proper card full message is triggered when you have less than 2GB space available and you try to start or continue a dump.

So when you started the backup, you had ~18GB free space. So it should only dump the first 9 parts (.00 to .08).
If you saw a broken part .09 inside, this means the limit is not honored and it continued with the 10th (.09) part.
And that's why it stopped at [ff] 4 (which the check for available cluster) and got the error (1).

Anyway, I'll prepare a proper debug version, just in case, anyone else want to try.

Sounds right to me, I think this is exactly what happened. I'm also pretty sure now that I confused "nine parts" and "part.09".

 

[CTCaer]

Sounds right to me, I think this is exactly what happened. I'm also pretty sure now that I confused "nine parts" and "part.09".

It's strange though. This "if (currPartIdx >= maxSplitParts)" is explicit. Your maxsplit was 9. It should have written the .08 part and then change currpartidx to 9 and hit this check.

On another note, in the FatFS, it is explained that this error does not always mean full disk.
The good thing is that the final 1.5 will be more explicit with the error codes. And it will not output error (1) for all disk errors.

-----------------------------

Just to remind anyone with Switch 1.0.0 and 2.X.X, please try the version here, with the classic sdfiles.zip. And report back, if you can launch hbmenu.
Thanks.

 

[kamo]

It's strange though. This "if (currPartIdx >= maxSplitParts)" is explicit. Your maxsplit was 9. It should have written the .08 part and then change currpartidx to 9 and hit this check.

On another note, in the FatFS, it is explained that this error does not always mean full disk.
The good thing is that the final 1.5 will be more explicit with the error codes. And it will not output error (1) for all disk errors.

-----------------------------

Just to remind anyone with Switch 1.0.0 and 2.X.X, please try the version here, with the classic sdfiles.zip. And report back, if you can launch hbmenu.
Thanks.

Couldn't be my card be at fault here? Maybe reporting the wrong space left size because of bad sectors? chkdsk didn't find anything but that too isn't 100% reliable all the time. I'm still wondering why this old card even worked considering my almost new class 10 didn't.

I'm also willing to try the final 1.5 on my class 10 card.

 

[Deletedmember448668]

So to dump everything to my 64gb SD card I need to choose Raw emmc right? And how about after, if I wasnt to change inject my emmc back on to my switch, what tool would i go to, or is that not available yet?

 

[chronoss]

Hello, here is the pic of my dump. I got 13 part : Switch 5.0.2

Is that correct ?

 

[Deletedmember448668]

Hello, here is the pic of my dump. I got 13 part : Switch 5.0.2

Is that correct ?

Detailed release notes:
This release lets you dump the USER partition, or the whole RAW eMMC.

The automatic partial dump is only activated if you don't have enough space in your SD card to dump the whole eMMC.

The partial dumping is done in 15 2GB parts.
This way you can even use a 4GB sd card to dump the whole eMMC.

Might being missing 2 parts? Idk, what size is your sd card?

 

[chronoss]

Detailed release notes:
This release lets you dump the USER partition, or the whole RAW eMMC.

The automatic partial dump is only activated if you don't have enough space in your SD card to dump the whole eMMC.

The partial dumping is done in 15 2GB parts.
This way you can even use a 4GB sd card to dump the whole eMMC.

Ok, I repeat the dump now !

 

[King_Knoedel]

I'm on 4.1 and tried to dump my nand 2 times now.
It ends in a fatal error 1 at 90 percent.
When is the 1.5 version going to be released?

 

[Deletedmember448668]

I'm on 4.1 and tried to dump my nand 2 times now.
It ends in a fatal error 1 at 90 percent.
When is the 1.5 version going to be released?

What size is your SD card?

 

[King_Knoedel]

What size is your SD card?

128 gb sdxc Samsung evo plus formatted with fat32 gui formatter.

 

[CTCaer]

Couldn't be my card be at fault here? Maybe reporting the wrong space left size because of bad sectors? chkdsk didn't find anything but that too isn't 100% reliable all the time. I'm still wondering why this old card even worked considering my almost new class 10 didn't.

I'm also willing to try the final 1.5 on my class 10 card.

Dunno actually

So to dump everything to my 64gb SD card I need to choose Raw emmc right? And how about after, if I wasnt to change inject my emmc back on to my switch, what tool would i go to, or is that not available yet?

You need eMMC BOOT and the Raw eMMC.
Restore option will not be available for now. It exists, but it's too dangerous for the general public. Too many corrupted dumps out there, because of broken old hekate binaries, that did not report errors on write.

Hello, here is the pic of my dump. I got 13 part : Switch 5.0.2

Is that correct ?

You've done the wrong procedure. You selected "Dump raw eMMC" on the first session, and when you moved the rawnand.bin.XX files, you pressed "Dump eMMC USER".
It's either raw->raw or user->user, until it's finished. Not mixed.

I'm on 4.1 and tried to dump my nand 2 times now.
It ends in a fatal error 1 at 90 percent.
When is the 1.5 version going to be released?

If you can't continue, because the partial.idx file cannot be written because of the previous failure, yeah better wait for 1.5.
It is still baking. I have some more things to code before releasing it.
Probably tomorrow.

 

[Stoned]

Hey @CTCaer any News for 3.0.1? :/