DS203 missing shell.elf

[Compycompy]

Hello,
I have a DS203 currently with wildcat 6.5 hw 2.81 sys 1.64
I went in DSU mode and copied the ds203.hex file, all went fine.
Disconnected, restarted and asked for shell.elf, I copied the elf files but still keeps on asking for the shell.elf file.
I believe there is some issue with the memory location as it seems to look in another part but I am not sure.
Any suggestion?

[gabonator]

Try this firmware:
system_ds203.hex.zip

And copy these files to devices:
ds203apps.zip

If it won't work, please send here screenshot of the root of the device's filesystem

[Compycompy]

Hi gab,
I tried to flash clean with the RS232 port, all was smooth, I copied the sys_b164.hex and all started correctly.
I loaded the system_ds203.hex, it was accepted. Turned off, then on asking for shell.elf (which at this point is fine).
I copied the ds203apps (uncompressed), turned off then on but still asking for shell.elf :-(
On the device directory I have the seven elf files, CALIB.DAT and CONFIG.DAT
Thanks for the help

[gabonator]

I just tried the files I provided and with the same result as in your case. For some strange reason, the filesystem gets corrupted when I copied all the files at once. After reconnecting I saw only empty root folder. But when I copied only the shell.elf to the disk, it worked. And then when the shell was loaded I copied remaining elf files. So please try to delete all *.elf files you copied before and copy only "shell.elf". Or try copying the file until the device will find it, possibly device format could help (dont forget to backup your CALIB.DAT and CONFIG.DAT)

[Compycompy]

Unfortunately no luck always missing shell.elf (tried to format both quick and standard). I tried either pressing f1 directly after copying or doing power cycle (several times).
I will try with another PC/OS maybe it helps.
Thank you!

[gabonator]

Just out of curiosity, please drop here screenshot of the disk root and hex view of the shell.elf

[Compycompy]

Hi,
this is the shell.elf hex view and the folder.

thank you!

[gabonator]

try running "fsutil fsinfo sectorInfo e:" and "fsutil fsinfo ntfsinfo e:", this should show us how the disk is formatted, the disk should have 4096 sectors and every sector should have 512 bytes, on OSX:

$ sudo diskutil info /dev/disk2                                      ↑
   Device Identifier:         disk2
   Device Node:               /dev/disk2
   Whole:                     Yes
   Part of Whole:             disk2
   Device / Media Name:       Flash Disk

   Volume Name:               DS203
   Mounted:                   No

   Content (IOContent):       None
   File System Personality:   MS-DOS FAT12
   Type (Bundle):             msdos
   Name (User Visible):       MS-DOS (FAT12)

   OS Can Be Installed:       No
   Media Type:                Generic
   Protocol:                  USB
   SMART Status:              Not Supported
   Volume UUID:               FC397130-DDE8-3700-94A9-F8B250F2079E

   Disk Size:                 2.1 MB (2097152 Bytes) (exactly 4096 512-Byte-Units)
   Device Block Size:         512 Bytes

   Volume Total Space:        0 B (0 Bytes) (exactly 0 512-Byte-Units)
   Volume Free Space:         0 B (0 Bytes) (exactly 0 512-Byte-Units)

   Read-Only Media:           No
   Read-Only Volume:          Not applicable (not mounted)

   Device Location:           External
   Removable Media:           Removable
   Media Removal:             Software-Activated

   Solid State:               Info not available
   Virtual:                   No

[Compycompy]

Hello gab,
my disk size is 8mb not 2, could that be a problem?
I try to run the command later today from linux.
thank you

[gabonator]

Yes that is a problem, the disk parameters are hard coded in the firmware. I do not own DS203 with 8MB disk, but if you can tell me what is the sector size and how many sectors are in that disk, I can prepare a build for you (and for others who own this newer version of DSO203), on linux try this command "fdisk -l"

[Compycompy]

From what I can see from windows it is exactly 8.388.608 byte with 4096 byte/sector and 2048 total sectors.
I can look with the command under linux if needed.

[gabonator]

Try this
system_ds203hd.hex.zip
ds203apps.zip

[Compycompy]

Copied in DFU mode the new system, now it goes directly in file manager mode.
Reboot and in normal mode I copied the shell.elf
Reboot and goes into file manager, if I choose the shell.elf some text appears quickly at the bottom (cannot see what is coming up) and then goes back to the file manager like if nothing happened.
I copied all the other elf files, they all work with the exception of miniosc.elf which gives an hardware error.

[gabonator]

Try one of these, maybe it will work. Their (minidso's) bios use different memory layout, because it needs bigger buffer for the filesystem (4k vs. 512), so I tried to build the app for different memory locations:
Archive.zip

[Compycompy]

No luck always hardware fault :-(
Btw I noticed that since I flashed the new system file it quickly says at boot FPGA configuration error, but then it continues without problems.

[gabonator]

Without having access to your device it is very difficult to fix this issue. And when it says that there is problem with FPGA, then also the oscilloscope functions wont work even when the application was running.

[MotoMaxis]

Happy New 2021 Year everyone.

@gabonator, Thank you for the tip. Originally I installed WebUSB driver using zadig-2.5.exe but now installed CDCSerial and see Com34 in device manager. I no longer get the popup when plugging in the DSO.

Here is what chrome says.

So next I used HyperTerminal to connect to that port. As soon as I press a key, it Hard Faults. Happens with stock and relocated webosc.

I do not have info about other devices. But do own DS201, DS202 and DS203.

Thank you for your hard work.

Update:
Some info from disassembler. Here is Unstripped elf.
weboscunstripped.zip

I don't understand why failing at 080233A8. A Push command. Maybe stack overflow?

Segmentation from elf

[gabonator]

GetRpcAddress is called when a line was received over serial link (line ended with \n). After typing a single letter, it should not call this function. I looked at the disassembly and got different one. Mine says 080233A8 is in _ZN4BIOS3LOW8SetParamEii after relocating ROM begin to 0x08023000 (compiling with gcc-arm-none-eabi-9-2020-q2-update on osx)

In the build folder, there is also assembly listing file, just for such cases when we need to trace the error. It could be caused by stack size. Hard to tell, you can try changing the stack size here, if it helps:

system/os_host/app_ds203.ld:

/* Define memory regions. */
MEMORY
{
  rom (rx) : ORIGIN = 0x0800C000, LENGTH = 32K
  ram (rwx) : ORIGIN = 0x20003000, LENGTH = 34K
}

_estack = ORIGIN(ram)+LENGTH(ram)-0x100;

Reading the documentation about HFSR/CFSR (https://www.keil.com/appnotes/files/apnt209.pdf), it says this:

[MotoMaxis]

I came across a nice small FLDump utility by bobtidey that lets you make a full image dump of your DSO. I modified it to dump 512KB instead of 256KB. You simply copy FL_DUMP4.HEX or (FL_DUMP4.ADR then FL_DUMP4.BIN) to device.

Reboot with Button 4 pressed. When it's done you should see FLASH.BIN on the drive.

credits: bobtidey
FlDumpHD.zip

[szmolnar]

Hi!

Have you got this fw running on DS203 2.82 hw?
Me too I was facing the same issue, whatever I did I'm just getting missing shell.elf, so I went back to MotoMaxis fw, which runs ok.

In spite the fact that I could not use our fw. I'm still interested in an potion how I could use it on my device. I need a solution to be able to make long time aquisitions, and I hope taht I will be able to do that with your fw or at least that It I will be able to write an app to to that.

Another question would be if it is possible to run four fw. in slot 3 or 4?
Slots 3,4 are open in w6.5, so theoretically I could be possible to run both fw. on this device.

Thanks!

[MotoMaxis]

Hi,

I was not able to make webosc.elf to work on DS203 2.82 hw. Always get HARDFAULT. I tried moving it to different area but it didn't help. We are waiting for gabonator to look into it.

On a side note, Wildcat fw has the Chart mode with extended time base of 100mS/div to 10min/div and the ability to Auto save to .csv until memory full. Check out -CHART MODE: in https://github.com/MotoMaxis/DS203-DSOQuad/blob/master/WildcatV6.5/User%20Guide%20V5.6-6.5.pdf

I am not sure if we can squeeze both firmwares on same device. It will require recompiling Wildcat to shift it to unused areas. But gabonator is one to know the answer.

[Turro75]

If I remember well wildcat firmware uses slot 1,2 and (a lot of) the spare flash memory over slot 4, so slot 3 and 4 are free

[szmolnar]

On a side note, Wildcat fw has the Chart mode with extended time base of 100mS/div to 10min/div and the ability to Auto save to .csv until memory full. Check out -CHART MODE: in https://github.com/MotoMaxis/DS203-DSOQuad/blob/master/WildcatV6.5/User%20Guide%20V5.6-6.5.pdf

Oh, thanks for that!
I totally missed that somehow. What is the longest period we can capture in practice?

[MotoMaxis]

On screen with meters off we see 13 divisions so that's 130 minutes. But if you enable auto log to .csv files then its until the flash drive becomes full.

I think its over 6 full days on a 8MB device. Each file has 4094 samples that are about 100KB in size.

[szmolnar]

I got that but not idea on how much fits on an empty flash drive? Can I store for example 24 hours?

…

On Sun, Feb 7, 2021 at 10:52 PM MotoMaxis ***@***.***> wrote: On screen with meters off we see 13 divisions so that's 130 minutes. But if log to .csv then its until flash drive becomes full. — You are receiving this because you commented. Reply to this email directly, view it on GitHub <#28 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AA224X3HY7LHEPRFCOIM4ILS54DQTANCNFSM4TJ32HXA> .

[MotoMaxis]

Oh 24 hours is easy to do.
It will depend on how fast you tell it to acquire a reading. The fastest time in Chart mode is 100ms/div and slowest is 10min/div which is a range of 13sec to about 20 hours to fill in the buffer. This will use up about 100KB of space. So I think you could max our in 20 minutes with fastest speed to something like 60 days with slowest speed. .

Chart mode works only in Full size buffer mode orange with scrolling rectangle at the bottom. . Which is 13 divisions times about 10 screens. That's about 4094 samples I think.

[szmolnar]

I might be wrong but I just try to understand the math behind :-) We have 13 divisions, so the buffer is enough for: @ 100ms/div (0.1sec/div) we should get 0.1 x 13 = 1.3sec (not 13) @ 10min/div x 13 = 130min = aprox 2.1hr (not 20) Assuming that the buffer space buffering the 13 divisions uses 100kb then the 8Mb will last for 80 x 1.3sec (approx 100sec) to 80 x 2.1hr (approx 168 hours) Am I wrong?

…

On Mon, Feb 8, 2021 at 9:28 AM MotoMaxis ***@***.***> wrote: Oh 24 hours is easy to do. It will depend on how fast you tell it to acquire a reading. The fastest time in Chart mode is 100ms/div and slowest is 10min/div which is a range of 13sec to about 20 hours to fill in the buffer. This will use up about 100KB of space. So I think you could max our in 20 minutes with fastest speed to something like 60 days with slowest speed. . — You are receiving this because you commented. Reply to this email directly, view it on GitHub <#28 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AA224X44EFI3VH6KGMRMO4TS56OCTANCNFSM4TJ32HXA> .

[MotoMaxis]

Once the screen fills up it will start scrolling through 10 more screens then loop back to the beginning. .

[gabonator]

Hello guys, this issue become a live discussion thread so I converted to into a discussion to cleanup the issues list a little bit. Feel free to add new discussion topics. Hope you wont mind this change

[Turro75]

Hello Gabonator,

I'm an owner of a DSO203 2.72 8Mb, still no chance to get la104 running?

This is the fdisk -l of my 8mb disk

Disk /dev/sdi: 8 MiB, 8388608 bytes, 2048 sectors
Disk model: Flash Disk      
Units: sectors of 1 * 4096 = 4096 bytes
Sector size (logical/physical): 4096 bytes / 4096 bytes
I/O size (minimum/optimal): 4096 bytes / 4096 bytes
Disklabel type: dos
Disk identifier: 0x00000000

not sure they are useful to determine the memory mapping, those are the command I use to flash/backup/unbrick my dso203:

#backup full image
#./stm32flash /dev/ttyUSB0 -g 0x08000000 -r WILDCAT56.BIN

#original flash full image from minidso (SYS162,PAV112,FPGA261,DFU314c) requires license activation
#./stm32flash /dev/ttyUSB0 -v -g 0x8000000 -w FULL272.BIN

#DFU.
#./stm32flash /dev/ttyUSB0 -S 0x08000000:0x3fff -w DFU345C.BIN
#./stm32flash /dev/ttyUSB0 -S 0x08000000:0x3fff -w DFU314C.BIN

#latest operating system file from minidso
#./stm32flash /dev/ttyUSB0 -S 0x0803d800:0x27ff -w Logo_001.BIN
#./stm32flash /dev/ttyUSB0 -S 0x0802c000:0x0117ff -w FPGA.BIN
#./stm32flash /dev/ttyUSB0 -S 0x08004000:0x7fff -w SYS164.BIN

#latest official minidso firmware
#./stm32flash /dev/ttyUSB0 -S 0x0800c000:0xfffe -v -w PA113.BIN

#wildcat 5.6 firmware
#./stm32flash /dev/ttyUSB0 -S 0x0800c000:0xfffe -v -w APP1A.BIN
#./stm32flash /dev/ttyUSB0 -S 0x08048000:0x1ffff -v -w APP1B.BIN

[Turro75]

For those who are interested I successfully ported SYS 1.64 to gcc here

https://github.com/Turro75/DSO203_SYS164_GCC

[accept]

Hello. Nice to meet you.
I am Japanese and not fluent in English, so I am using DeepL translation.
I apologize if there are sentences that do not make sense or if there are rude phrases.
I also own HW version:2.82, 8MB disk model.
When I follow the procedure to write the file, I still get an error that shell.elf is not found.
Is there anything I can do to help in supporting this device?