HI,

 

Im tring to print a long long int to hex but i never get the wirte answer

long long  IMSI = 206108601156779

when i do a print("Imsi: 0x%Lx\n", IMSI); it returns me 0x9430

im using a happy ghecko board 

 

regards

I have seen numerous post from users having issues with Simplicity Studio hanging and/or crashing. Unfortunately, running the configuration tool to regenerate RAIL configuration files no longer completes for our project. It seem to worse as a given project grows in number of files, and we may have hit that tipping point.

I found out yesterday, from another forum post, that a workaround is to create a new workspace and project, update radio configuration, and regenerate RAIL config. Then copy this into our project. While this works, it is very tedious and needs to be fixed.

I have seen in the the past 2 releases, but seems to be worse in the latest (SV4.1.11.8). I also saw a forum post on this, saying it was fixed. I had created a Windows machine JUST to run simplicity, and it fails just the same. So the hang is independent of host OS. I have seen it both in OSX and Windows.

 

 

Hi,

I am trying to replace a call to the "deprecated" arm_rfft_f32() in some existing code (for now, the SLSTK3701A_spectrum_analyzer example) for its replacement, arm_rfft_fast_f32(), and hit a wall. The latter depends on sqrtf(), which...isn't actually defined anywhere. According to all of the 3e+08 Google hits for this error, as well as a grep of the entire CMSIS directory, it is defined in the standard pre-compiled math library and is this error is solved by remembering to link it ("-lm"). However, the library is specified in the project as shown below.

I am using the default GCC toolchain of this project (GNU ARM v7.2.1 as of this writing). Digging into arm_math.h, it seems it should not be calling sqrtf() at all, but rather __builtin_sqrtf() for a device with an FPU:

  CMSIS_INLINE __STATIC_INLINE arm_status arm_sqrt_f32(
  float32_t in,
  float32_t * pOut)
  {
    if (in >= 0.0f)
    {

#if   (__FPU_USED == 1) && defined ( __CC_ARM   )
      *pOut = __sqrtf(in);
#elif (__FPU_USED == 1) && (defined(__ARMCC_VERSION) && (__ARMCC_VERSION >= 6010050))
      *pOut = __builtin_sqrtf(in);
#elif (__FPU_USED == 1) && defined(__GNUC__)
      *pOut = __builtin_sqrtf(in);
#elif (__FPU_USED == 1) && defined ( __ICCARM__ ) && (__VER__ >= 6040000)
      __ASM("VSQRT.F32 %0,%1" : "=t"(*pOut) : "t"(in));
#else
      *pOut = sqrtf(in);
#endif
...

As a sanity check, I confirmed at (at least in main.c) that __GNUC__  is defined and __FPU_USED==1, so I would not expect to see the fallback 'sqrtf()' used at all.

// Sanity checking FPU-related compiler flags
#if (__FPU_USED == 0)
      #error "FPU is not used"
#endif

#if !defined(__FPU_PRESENT)
      #error "FPU_PRESENT is not defined"
#endif
#if (__FPU_PRESENT == 0)
      #error "FPU is not present"
#endif

#if !defined(__VFP_FP__)
      #error "VFP_FP is not defined"
#endif
#if defined(__SOFTFP__)
      #error "SOFTFP defined"
#endif

//More sanity: Confirmed __GNUC__ is defined
//#if defined(__GNUC__)
//	#error "GCC defined!"
//#endif

// Sanity check the sanity check: Confirmed __FPU_USED == 1
//#if (__FPU_USED == 1)
//      #error "FPU is used, sanity check OK"
//#endif

Sooo...

1) Can anyone point out what I may be missing to get sqrtf() visible to the compiler?

2) Should I be concerned that it is looking for "sqrtf()" specifically and not another variant of the function? If so, why would *this* be happening?

Thanks!

I'm struggling to use Simplicity Studio on Fedora 29 (64 bit).   The problem comes when trying to install the Bluetooth SDK (also when opening the internal web browser.)

Initially the error message was the oft reported:

Failed to execute runnable (org.eclipse.swt.SWTError: No more handles [Unknown Mozilla path (MOZILLA_FIVE_HOME not set)])

I saw the advice about using libwebkit.  Both versions 1 and 2 are installed on my OS:

ls -l /usr/lib64/libwebkit*
lrwxrwxrwx. 1 root root       28 Feb 11 19:49 /usr/lib64/libwebkit2gtk-4.0.so.37 -> libwebkit2gtk-4.0.so.37.33.7
-rwxr-xr-x. 1 root root 46471648 Feb 11 20:05 /usr/lib64/libwebkit2gtk-4.0.so.37.33.7
lrwxrwxrwx. 1 root root       44 Feb 15 15:00 /usr/lib64/libwebkitgtk-1.0.so -> /usr/local/lib64/libwebkitgtk-1.0.so.0.22.17
lrwxrwxrwx. 1 root root       44 Feb 15 15:00 /usr/lib64/libwebkitgtk-1.0.so.0 -> /usr/local/lib64/libwebkitgtk-1.0.so.0.22.17
lrwxrwxrwx. 1 root root       44 Feb 15 14:59 /usr/lib64/libwebkitgtk-1.0.so.0.22.17 -> /usr/local/lib64/libwebkitgtk-1.0.so.0.22.17

However, SimplicityStudio doesn't seem to be able to find these library files.  The error message remains the same.

So, I thought I'd try downloading xulrunner from here:

http://ftp.mozilla.org/pub/xulrunner/releases/1.9.2.19/sdk/

However, no 64 bit version exists.  Just to confirm I pointed MOZILLA_FIVE_HOME at the xulrunner and added the xulrunner bin directory to the library path.  The result was this error message:

[snip]/SimplicityStudio_v4/configuration/org.eclipse.osgi/196/0/.cp/libswt-mozilla-gtk-4528.so: libxpcom.so: wrong ELF class: ELFCLASS32

So three questions:

1) How do I get SimplicityStudio to work?

2) How is a "manual" installation of the SDK, as suggested actually done?

2) When will there be a version of SimplicityStudio that doesn't rely on these libraries? After all xulrunner has been dead for a long time, and version 1 of libwebkit is considered a security risk by the maintainers of version 2.

Thanks,

Jon

 

Hello,

I have a custom board with on it an MGM111A256V2. I have access to the SWD_CLK, the SWD_IO and the nReset. I cannot seem to flash it.

In Simplicity Studio I have two projects, one for the bootloader, which is just the SPI-ezsp project and which should generate the bootloader file. The second project should generate the actual project. So, I need to flash these two files.

I used to work with the EFR32MG12P dev kit, and I used JLink CLI for flashing. When I flashed the program, I just called loadfile twice, once for the bootloader and once for the actual project. JLink CLI would then figure out on its own that the bootloader needed to go to 0x0FE10000:

Comparing flash   [100%] Done.
J-Link: Flash download: Bank 3 @ 0x0FE10000: Skipped. Contents already match
O.K.

And the actual program is placed at 0x0.

So back to my actual board, the one with MGM111A256V2: I cannot flash it properly. I use the old red Ember Debug Adapter ISA3, but that does not seem to be a good idea:

*** J-Link V6.46k J-Link V6.46k Warning ***
You are using a J-Link which does not have intelligence
for the selected CPU core (Cortex-M4) in the firmware.

Intelligence in the firmware enables J-Link
to generate sequences for the CPU core.
Without this feature, all sequences are generated by the PC.

Intelligence in the firmware allows higher target interface speeds
and significantly enhances both speed and stability of the
communication with the target CPU.

You can use this J-Link with your target CPU, but we recommend
using a newer model of J-Link / J-Trace.
*** J-Link V6.46k J-Link V6.46k Warning ***

But it does not seem to be a dealbreaker. Still though: should I get a new one? And which one do I need to get?

Moving on: flashing does not work as nice, at all. My custom bootloader is flashed to address 0x0, just like the actual program and the board remains dead. At this point, I do not see what I am doing wrong. Could it be that I need a different bootloader? Or do I need to configure it differently? And many, many more questions.

So, is there someone here with the information I need to properly create and flash the bootloader and the actual program?

I am happy to answer any questions.

Hey

I have a problem to debug a BGM113 bluetooth module with the EFR32 on it. My IDE is Simplicity Studio V4.

With my own code it doesn't work. The debugger don't have any breakpoints and it seems that it stucks in a loop. It comes out of the blue because yesterday it works fine.

The debugger works fine with a code example for adc for the EFM32PG1B on the same device with the same setup. I think it will be a settings problem.

When I try to generate a new empty-soc the debugger also skipp breakpoints and I have the same issue like I have in my own code. I tried to attach my code and the working example code but it does not load and the attached file appears in grey.

My problem is very similar to this problem but I set the SWD and the debug level to -g.

https://www.silabs.com/community/software/simplicity-studio/forum.topic.html/can_t_debug_withsim-0pUy

What could it be that the debugger not even works with the generated empty-soc example code?

Hello Team,

We are designing a custom WF200 based solution , in this we have ported simplicity micrium example code given at following link:

https://github.com/SiliconLabs/wfx-fullMAC-tools

The code is working fine till now, i have a question related to the licensing of this code?

We would like to use this example code and develop our application  code further, but do we need to pay any licencing fee for commercial use?

If yes than , we may have to switch to some other RTOS,

Let me know if any other input is needed.

Regards ,

Santosh

Hi,

 

I am importing a project and keep getting this build error:

 

c:\siliconlabs\simplicitystudio\v4_2\developer\sdks\gecko_sdk_suite\v2.5\platform\emlib\inc\em_device.h:67:2: error: #error "em_device.h: PART NUMBER undefined"

 

I have defined the symbols as instructed in previous forum posts.  I can also get it to work if I use an updated SDK.  Any other ideas would be helpful.

 

Thanks!

 

M

Hi,

I am trying to use the code correlation feature in Energy Profiler, starting off with the SLSTK3701A_spectrum_analyzer code example on the STK3701A (GG11) starter kit, and followed the instructions in UG343 to enable SWO output and start a profiler session. The session runs and displays the current consumption graph, but the Energy Profile (live) or (range) views display a blank grid, as follows:

 

Specifically, if I start the session via the right-click Project context menu (->Profile as -> Simplicity Energy Profiler Target) to start the session, the Correlation grid is present, but unpopulated, and remains so after explicitly selecting the (only) J-Link device. If I start it via the "Run -> Profile Configurations..." menu, selecting the J-Link device causes it to display the "Select an executable" prompt as shown in the UG. After selecting the appropriate *.axf file (or the inappropriate one, e.g. the Debug binary while in Release configuration or vice versa), this prompt is again replaced by a blank (unpopulated) grid. I tried profiling the Debug and Release configurations with the same results.

Notes:

• Studio version "SV4.1.11.8" (updated yesterday), Gecko SDK Suite v2.6.3: MCU 5.8.3.0
• Target board: SLSTK3701A with up to date firmware (also updated yesterday)
• Debug configuration: Debug level Maximum (-g3), optimization None (-o0), "Generate debugger-friendly prologs" and "Always branch to builtin functions" checked, "Omit all symbol information" unchecked.
• bsp_trace.c included in project (by me); contents of ".../kit/SLSTK3701A_EFM32GG11/config" (including hal-config-board.h) included (by default)

Did I miss a trick? How do I get the Code Correlation view to show the expected results, and/or a helpful error message if something is wrong?

Thanks!

hi,

I have the EFR32MG21 dev board, (BRD4180A) connected to my wireless starter kit mainboard BRD4001A Rev 01. I just did a fresh install of SS last week Friday to see if it would solve another issue I was having.

 

Ok so scenerio: I just built a Z3SwitchSoc project and modified it a bit and was planning on uploading it to my WSTK. I built the project and then from the Debug Adapters window, right clicked the WSTK and clicked "Upload application." Just like I had done about a couple dozen times already today. I selected the .gbl file, selected a bootloader, checked the 'erase flash' box and then hi ok and got my first error of this terrible evening. 

Error: Could not erase chip, error code -13 (Invalid MCU Info.).

I tried the 3 solutions suggested in forums, using commander and trying the Unlock Debug Access, but that just told me: "Could not unlock chip. this error is usually a result of not being able to connect to the debug interface. Error: device recovery failed" 

Then I tried to do the same thing in every debug mode option, OFF, IN, OUT, MCU, they all didnt work

the commander window tells me "Could not connect to target device"

Then I tried the same thing in the Flash Programmer, and get a "DP write failed" error. "Could not access Debug challenge interface"

Then I tried the command line approach and it still didnt work. perhaps I did not specify the correct device name? I put: 

cd C:\SiliconLabs\SimplicityStudio\v4\developer\adapter_packs\commander
commander adapter dbgmode OUT
commander device recover --device efr32mg21a020f1024im32

I also tried changing the adapter firmware version (1v4p3b1062). And had one heck of a time trying to find a download link. but if you go into Package Manager and go to the Assets tab, you can find things to download. When you select things to install the links don’t work because that would be too easy. So instead you have to tailor the link that is spit out in the error box to trigger a download. This is the link for the BRD4001A Rev A01 board [link](https://developer.silabs.com/studio/v4/assets/hwtools/firmware/S1015B_wireless_stk_firmware_package.zip).

I tried that and then restarting my computer and now it tells me Error, could not connect debugger when i try to flash it. I also tried to see if i could connect the commander before I flashed it, but also, no success. 

So I am losing hope because this [link](https://www.silabs.com/community/wireless/bluetooth/forum.topic.html/error_could_not_con-OvJT) here suggests that after following all that you basically just have a really expensive brick. So I am hoping I can solve this with SS, and possibly even with some soldering.

I am very hesitant to proceed because I have bricked 2 of these boards now and if i cannot fix them there goes 100 bucks.

I should also mention that the software is fully functional on one of the boards. (the other one had a bug in it that i needed to fix that caused a perpetual reset) I can open a console session and get all the CLI commands to print presing help. The LEDs on the board flash too. 

 

Any advice?

Thanks,

Alex

 

Edit:

I was previously programming the Mighty gecko EFR32mg12 board on the same WSTK main board, maybe this has something to do with it?

I also am able to swap out the radio board with a different one and the WSTK main board recognizes it with no issues.

 

This should be a general question, not specific to any EFM32 (or EFM8) device. It's an SS4 question, not a device question.

I want to use a specific em_xxx library, say em_csen, in a design. How do I add it to a project in SS4? I mean, other than finding it in the magic location in the file system and copying it.

I have the Project Properties open, with C/C++ Build -> Project Modules open. EFM32->Common are both checked, as is EFM Peripheral Support. If I toggle EFM Peripheral Support, it'll add ALL of the libraries!

Similarly, checking energyAware Drivers now includes ALL drivers in the project, whether or not you need them. This usually causes the build to fail, as unneeded sources still have dependencies that are not defined.

This has been broken for quite awhile now. Can it be fixed, please?

(Also, I'll add a plea for better module descriptions. CMSIS Examples has a description CMSIS Examples. Really! See also General Board Support and General Part Support.)

 

 

 

When using SLWSTK6020B, what are the main differences between USB debugging and using simplicity debugger adapter board? 

Hello,

my code size is about 72kb and i get the error: ERROR L107: ADDRESS SPACE OVERFLOW. Since there are some parts that are rarely used lookup tables i want to put the in another banks. I did the following in Simplicity Studio(!):

- enable Code Banking in Settings > Keil 8051 Linker > Banking

- set "Bank2" for the selected .OBJ Files

- set "F580=1", "USE_BANKING=1" and "SILABS_STARTUP=1" in Settings > Keil 8051 Assembler > Control Symbols

- set "NUMBANKS=4" in Settings > Keil 8051 Assembler > Symbols

- set Settings > Keil 8051 Compiler > General > Code size: Large

I still get the same error message. The selected .obj Files are about 20kb so it must fit in a single Bank.

How do i tell the compiler i want to use banking on that files?

 

Thank you!