Hacking [Pre-release, WIP] Yet another rxTools?

[Madridi]

The background file load occurs every progress bar refresh so it makes addotional load to SD card. EmuNAND dump utilizes sd card more then sysnand but anyway I just dumped sysnand to Class6 card in 4:03. It could be faster than my Class10 because it is my working card so there are fragmentation, while Class6 card FAT partition was just copied file-by-file.

So, less load=less chance that the issue happens?

Also, about the initial setup, I went back and confirmed my suspicions. I know you said it was the opposite, but the firmware decryption does happen on first boot, while font extraction happens on second boot.
This also explains why I don't see anything besides a progress bar for firmware decryption. Because there is no font extracted yet, right?
The good news about this, is that it's stable.. Haven't had a problem with the setup for the last few builds

 

[duke_srg]

So, less load=less chance that the issue happens?

Also, about the initial setup, I went back and confirmed my suspicions. I know you said it was the opposite, but the firmware decryption does happen on first boot, while font extraction happens on second boot.
This also explains why I don't see anything besides a progress bar for firmware decryption. Because there is no font extracted yet, right?
The good news about this, is that it's stable.. Haven't had a problem with the setup for the last few builds

That's right! It is over 450 mbytes of additional reads per one nand dump. With the latest commit it is at least 3 times less, since progress bar updates only once per second. So yes, it is worth to try this commit!

And this is how it works with no firmware and font in /data :

Spoiler

 

[Madridi]

That's right! It is over 450 mbytes of additional reads per one nand dump. With the latest commit it is at least 3 times less, since progress bar updates only once per second. So yes, it is worth to try this commit!

Just tried, sysnand dumped fine this time in 4.43 minutes. But then I got the error with dumping emunand

And this is how it works with no firmware and font in /data :

Spoiler

I don't know what to tell you.. I know this is the behavior I'm supposed to get, but it's not what I'm actually getting. Only included firm folder, no data folder. First boot is firmware decryption with just a progress bar. Main menu is blank so I have to reset. Second boot is font extraction..

I even checked the data folder after first boot (before second boot, I can only see the decrypted firm files. No font

 

[duke_srg]

Just tried, sysnand dumped fine this time in 4.43 minutes. But then I got the error with dumping emunand

Try running checkdisk with your card and also report the allocation unit size, it is displayed as a result of checkdisk also. Then you can backup/restore sdcard FAT partition contents. This will redice fragmentation which also coluld be the issue.
Currently there are no sd card driver implementation with comprehensive error processing, so it'll take a lot more time comparing with backup/restore provedure which will help to localize the issue.

 

[Madridi]

Try running checkdisk with your card and also report the allocation unit size, it is displayed as a result of checkdisk also. Then you can backup/restore sdcard FAT partition contents. This will redice fragmentation which also coluld be the issue.
Currently there are no sd card driver implementation with comprehensive error processing, so it'll take a lot more time comparing with backup/restore provedure which will help to localize the issue.

Wait, FAT? My msd is FAT32..
Also, given that my drive is O for the msd, is this the command I should use: "chkdsk O:"
I dont suppose I should be backing up my msd before trying any of this right? cause I'm running out of space on my PC (lots of pending stuff needs moving)

 

[duke_srg]

Wait, FAT? My msd is FAT32..
Also, given that my drive is O for the msd, is this the command I should use: "chkdsk O:"
I dont suppose I should be backing up my msd before trying any of this right? cause I'm running out of space on my PC (lots of pending stuff needs moving)

For sure it is FAT32, FAT16 dont support such capacities
Yes, running chkdsk with no /f don't affect file system, just checks.

 

[Madridi]

For sure it is FAT32, FAT16 dont support such capacities
Yes, running chkdsk with no /f don't affect file system, just checks.

Spoiler: Result

 

[duke_srg]

Spoiler: Result

Looks good. So for now I only can propose to try to defragment the card or try with another one. Pity we don't have someone else to test

 

[Madridi]

Looks good. So for now I only can propose to try to defragment the card or try with another one. Pity we don't have someone else to test

I know.. I'm disappointed myself

I ran the analysis on my card, and it said 0% fragmented, currently running the defragmentation anyway.

Can't use another card atm. The samsung card is on my other 3ds, which I'm not looking to touch really.

 

[duke_srg]

I know.. I'm disappointed myself
I ran the analysis on my card, and it said 0% fragmented, currently running the defragmentation anyway.
Can't use another card atm. The samsung card is on my other 3ds, which I'm not looking to touch really.

Class6 card is the worst that I have too
Well, just for the test we can reduce load a bit commenting this line
https://github.com/dukesrg/rxTools/blob/master/rxtools/source/lib/progress.c#L54
And then try several times with emunand dump

 

[Madridi]

Class6 card is the worst that I have too
Well, just for the test we can reduce load a bit commenting this line
https://github.com/dukesrg/rxTools/blob/master/rxtools/source/lib/progress.c#L54
And then try several times with emunand dump

I'll try that later tonight. It's taking a while to defragment anyway.

 

[duke_srg]

I'll try that later tonight. It's taking a while to defragment anyway.

Yes, this is a last resort anyway, it notmally should work fine with any fragmentation.

--------------------- MERGED ---------------------------

I'll make a fork with some diagnostics. If fatfs read just ended with timeout error, this is an easy thing to fix.

 

[duke_srg]

Ok, just added a simple error processing, must pring error number after a failed operation and filename. Please check

 

[Madridi]

Ok, just added a simple error processing, must pring error number after a failed operation and filename. Please check

Sysnand once again dumped ok. Emunand error once again
"Top.bgr - 1"

 

[duke_srg]

Sysnand once again dumped ok. Emunand error once again
"Top.bgr - 1"

Yep, this is FR_DISK_ERR. Controller or card responded with error and it was passed to fatfs layer.

 

[duke_srg]

Ok, try the next one. Low level sd/nan read access data timeout masked, not sure it will do the trict, probably it just hangs on that error.

 

[Madridi]

Ok, try the next one. Low level sd/nan read access data timeout masked, not sure it will do the trict, probably it just hangs on that error.

Actually, it's even worse. I get a black screen when loading the code.bin .. Not even the setup starts

 

[duke_srg]

Ok, reverted that commit.

 

[duke_srg]

Not even the setup starts

Please check branch dev-cardboost. Low level read/write operations now repeated on error, so now either resolve the problem or hang in infinite loop. The latter will require adding full card initialization, so if it works once, pleae try several times to ensure it is stable.

 

[Madridi]

Please check branch dev-cardboost. Low level read/write operations now repeated on error, so now either resolve the problem or hang in infinite loop. The latter will require adding full card initialization, so if it works once, pleae try several times to ensure it is stable.

Froze about half way through